MASS SPECTROMETRY AND GEOCHRONOLOGY

John R. de Laeter
Department of Applied Physics, Curtin University of Technology, Perth,
Western Australia
Received 14 June 1998; accepted 11 August 1998
‘‘About this time, Rutherford, walking in the Campus with a Rutherford had determined the amount of uranium and ra-
small black rock in his hand, met the Professor of Geology; dium in the rock, calculated the annual output of alpha
‘‘Adams,’’ he said, ‘‘how old is the Earth supposed to be?’’ particles, was confident that these were helium, measured
The answer was that the various methods led to an estimate the amount of helium in the rock and by simple division
of 100 million years. ‘‘I know,’’ said Rutherford quietly, found the period during which the rock had existed in a
‘‘that this piece of pitchblende is 700 million years old.’’ compacted form. He was the pioneer in this method and his
This was the first occasion when so large a value was large value surprised and delighted both geologists and
given, based too on evidence of a reliable character; for biologists.’’ (Eve, 1939).

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
II. Birth of Modern Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
III. Development of Geochronology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
A. Uranium, Thorium-Lead Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
B. Potassium-Argon Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
C. Rubidium-Strontium Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
D. Samarium-Neodymium Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
E. Lutetium-Hafnium Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
F. Rhenium-Osmium Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
G. Potassium-Calcium Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
H. Uranium-Series Disequilibrium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
I. Dating by Cosmogenic Isotopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
IV. Advances in Sector-field Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
A. Solid Source Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
1. Ionization Enhancement Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
2. Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
3. Abundance Sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
4. Multicollectors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5. Mass Fractionation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
B. Gas Source Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
V. Recent Developments in Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
A. Secondary Ion Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
B. Accelerator Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
C. Inductively Coupled Plasma Mass Spectrometry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
D. Other Developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
VI. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

The physical method of geochronology is based on the measure- over the past 50 years. In fact, the mutual development of these
ment of isotopic ratios by mass spectrometry. Geochronology, two fields is an excellent example of the interplay between sci-
as an essential field of endeavor in the earth sciences, would ence and technology. The demand for improved mass spectro-
not have been possible without mass spectrometry, and these metric instrumentation to satisfy the increasingly exacting re-
two endeavors have developed in a mutually cooperative manner quirements of age determinations has led to a succession of

Contract grant sponsor: Australian Research Council.

Mass Spectrometry Reviews, 1998, 17, 97–125

q 1999 by John Wiley & Sons, Inc. CCC 0277-7037/99/020097-29

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

n DE LAETER

technological advances in mass spectrometry. These advances if one could measure either of the decay products He or
have, in turn, enabled geochronology to reach a level of scientific Pb, then one could measure geological time by this nuclear
achievement of impressive dimensions. The deceptively simple clock (Rutherford, 1906).
design of the 607 sector field mass spectrometer by A. O. Nier Unfortunately, the only means available at that time
in the 1940s was the foundation on which geochronology was
to measure He and Pb were chemical techniques, and both
established. Over the past 50 years, technological advances have
enabled the sector field instrument to meet most of the require-
suffered serious shortcomings. Helium, being a gas, leaked
ments of the various geochronological techniques that have been out of U-rich ores because of radiation damage, and geo-
developed during this period of time. During the past 20 years logical and weathering effects. Thus, the calculated ages
or so, however, radically new mass spectrometers have been were very much minimum estimates. On the other hand,
developed. Secondary ion mass spectrometry (SIMS) enables the measurement of the quantity of Pb could not distin-
geochronologists to analyze U- or Th-bearing minerals in situ guish radiogenic Pb from primordial Pb, and so the ages
to obtain accurate U, Th-Pb ages. These double-focusing mass were overestimates.
spectrometers have revolutionized mineral age determinations. In fact, no real progress was made in measuring accu-
Accelerator mass spectrometry (AMS) has applied techniques rate ages until the discovery of isotopes by J. J. Thomson
developed in nuclear physics to measure short-lived cosmogenic in the Cavendish Laboratory in Cambridge University in
nuclides such as 14C, 26Al, and 129I to provide information on
1912 laid the foundation for physical geochronology. In
geologically short-term events in a more efficient manner than
by radioactive-counting methods. Inductively coupled plasma
1929, F. W. Aston, using a mass spectrometer, was able
mass spectrometry (ICPMS) allows for the rapid determination to show that a U-rich sample of broggerite had at least
of the abundances of almost all elements in solid and liquid three isotopes of Pb and that the sample was enriched in
206
samples, and is now being applied to geochronology systems. Pb, a decay product of 238U. This finding enabled an
This review examines the interplay between the evolution of mass age of 909 Ma to be calculated for the broggerite sample
spectrometric instrumentation and the development of geochro- (Aston, 1929). A companion article by Rutherford gave
nology over the past 50 years. q 1999 John Wiley & Sons, Inc., an estimate of the age of the earth as 3.4 1 109 y (Ruther-
Mass Spec Rev 17, 97–125, 1998 ford, 1929).
This experiment by Aston (1929) heralded a new era
in geochronology, an era based on isotopic rather than
I. INTRODUCTION elemental measurements. Since that time, the mass spec-
trometer has been an indispensable tool for every geochro-
The concept of time has never ceased to intrigue those nologist—a veritable ‘‘time machine’’ that enables us to
who have thought about it. In particular, the age of the explore the past with a variety of radioactive decay sys-
earth has been studied by scientists over the centuries. tems and with ever-increasing accuracy. A significant step
Lord Kelvin estimated the age of the earth by heat flow, in validating the isotopic method of age determinations
George Darwin by tidal interactions, Edmund Halley from was provided by A. O. Nier, who, in 1938, demonstrated
the saltiness of the oceans, and Charles Walcott by the that the isotopic abundances of Pb samples varied consid-
accumulation of sediments. However, none of these erably, depending on the chemical composition and age
‘‘clocks’’ was particularly accurate and the physicist, Kel- of the ores (Nier, 1938). Nier showed that U-rich ores
vin, fell into disrepute with geologists, because his esti- gave Pb enriched in 206Pb and 207Pb, whereas Pb derived
mated age of 10 million years, which he obtained by exam- from Th-rich ores were enriched in 208Pb. The only Pb
ining the cooling of the earth from a molten body, was isotope that is unaffected by radioactive decay is 204Pb,
far too short as compared to geological estimates. Unfortu- because it is not the end-product of a radioactive parent
nately for Kelvin, radioactivity was unknown at that time. and, therefore, represents primordial Pb that was inherited
The discovery of radioactivity by Henri Becquerel in by the ore at its time of formation. 204Pb can, therefore,
1896, coupled with Marie and Pierre Curie’s demonstra- be used to correct for primordial or ‘‘common’’ Pb con-
tion that radioactive atoms transmute into atoms at prede- tamination for the isotopes 206,207,208Pb. Using this tech-
termined and essentially immutable rates, provided the nique, Nier, Thompson, and Murphy (1941) measured a
basis for a new clock to measure geological time. In 1902, U-Pb age for a sample of monazite to be 2570 { 70 Ma.
two scientists at McGill University in Canada, the New Nier also pioneered the K-Ar technique of age dating
Zealand physicist Ernest Rutherford and the English chem- by proving the existence of the long-lived radionuclide
ist Frederick Soddy, investigated the radioactive decay of 40
K, using a 1807 mass spectrometer (Nier, 1935). He later
U and showed that it decayed to the end-product Pb and measured the absolute isotopic abundance of 40K as 0.0119
also produced He from the alpha particles emitted in the { 0.0001% (Nier, 1950), a figure that compares remark-
decay scheme. In 1905, in the Silliman lectures at Yale ably well with modern values (Coplen, 1995). The first
University, Rutherford suggested the possibility of using age determination using the K-Ar technique was achieved
radioactivity as a geological time-keeper on the basis that, by Aldrich and Nier (1948).

98

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

 MASS SPECTROMETRY AND GEOCHRONOLOGY n

Thus, at the end of World War II, when nonmilitary

scientific research began its phenomenal growth, geochro-
nology was poised to revolutionize our knowledge of geo-
logical events. One ingredient, however, was missing. Al-
though the theoretical concepts of radioactive dating had
been developed, no instrument existed outside physics re-
search laboratories that would enable geologists to exploit
the powerful new geochronological tools available to
them. However, this situation was soon to be changed.

II. BIRTH OF MODERN MASS

SPECTROMETRY

Until 1940, mass spectrometers were either sophisticated

double-focusing machines used primarily for measuring
the atomic masses of nuclides, or large, complicated 1807
machines that were difficult to build and operate success-
fully. However, in 1940 A. O. Nier described a distinc-
tively different type of mass spectrometer that was des-
tined to be the tool whereby geochronology could be ex-
ploited. Nier (1940) designed a 607 sector field mass
spectrometer, which not only reduced the weight and
power consumption of the electromagnet, but also enabled
the ion source and detector to be removed from the influ-
ence of the magnetic field. This single-stage instrument
was significantly simpler to construct and operate than
previous mass spectrometers. FIGURE 1. A schematic diagram of the Nier 607 magnetic sector mass
This new mass spectrometer was used extensively spectrometer (Nier, 1947).
during the Manhattan Project from 1940–1945 (Nier,
1989). It was, in fact, the instrument used by Nier to
separate microgram quantities of 235U from 238U, thus en- field mass spectrometer was the beginning of modern mass
abling the fissionable nature of 235U to be established (Nier spectrometry. In the case of geochronology, it is undoubt-
et al., 1940). This design became the prototype for all edly true that the development of this field in the 1950s
subsequent magnetic deflection mass spectrometers. These and 1960s was based almost entirely on the sector field
607 mass spectrometers were used to measure the enrich- instrument, and that the majority of geochronological anal-
ment in U in the diffusion plant at Oak Ridge National yses in the 1990s are still undertaken using this basic
Laboratory, in measuring H/D ratios in ‘‘heavy’’ water instrumental design, albeit with numerous technological
plants, and as helium leak detectors. improvements.
In 1947, using the extensive experience of building Another factor in the development of geochronology
and using these mass spectrometers in the Manhattan Proj- incorporating mass spectrometry was an analytical tech-
ect, and incorporating technological improvements that nique named isotope dilution mass spectrometry (IDMS).
had occurred since 1940, Nier described a modified ver- One of the outcomes of the Manhattan Project was the
sion of the 607 sector field instrument (Nier, 1947). Figure availability of separated isotopes of most of the chemical
1 is a schematic diagram of Nier’s 1947 mass spectrome- elements. These isotopically enriched isotopes enabled
ter. The importance of the 607 sector field instrument was IDMS to be developed to measure quantitatively the ele-
that the elegance of its simple design enabled scientists mental abundance of trace elements in a wide range of
other than physicists to utilize the power of isotopes in samples (Inghram, 1954). The accuracy and sensitivity of
biology, chemistry, and geology, and this development IDMS is approached only by neutron activation analysis,
was achieved against the prevailing opinion of the day, and over 70 elements can be determined using this tech-
as expressed by Francis Aston, who believed that mass nique (Herzog, 1970). IDMS can be applied to all elements
spectrometry would die away as an active research field that have two or more isotopes, and can also be applied
(Svec, 1985). to mononuclidic elements, provided that a long-lived ra-
It can be argued that the development of the sector dioisotope of the element exists (Webster, 1960).

99

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

n DE LAETER

The principle of the method is to mix a sample that can be monitored relative to the invariant 204Pb isotope.
contains the element whose abundance is required, with a Although attempts had been made to date rocks by the U,
known quantity of an enriched tracer of that element. The Th-Pb method, an important milestone was achieved by
isotopic composition of the natural element and the tracer Nier, Thompson, and Murphy (1941), who measured a
207
must be determined by mass spectrometry, and the isotopic Pb/206Pb age of 2570 { 70 Ma from a monazite sample
composition of the mixture measured as accurately as pos- from the Huron district in Canada.
sible under similar mass spectrometric conditions in order In the 1950s, the availability of isotopically enriched
to minimize variations in mass fractionation. Once these isotopes and the availability of sector field mass spectrom-
parameters have been measured, the abundance of the ele- eters provided the base on which the U, Th-Pb method
ment in the sample can be determined. An analysis of the could be developed. IDMS could be used to measure the
advantages and limitations of the IDMS technique is given elemental abundances of U, Th, and Pb, even at low con-
by Heumann (1992). centrations. When combined with analyses of the isotopic
The development of mass spectrometry has paralleled, composition of Pb in the same sample, one could calculate
206
to a great extent, the development of vacuum techniques Pb/238U, 207Pb/235U, and 208Pb/232Th ages. Because the
207
and electronics. Prior to World War II, mass spectrographs Pb/206Pb ratio changes systematically with time due to
were used to measure atomic masses, and these instru- the differing decay rates of the U parent, a 207Pb/206Pb age
ments incorporated a photographic plate as the means of could also be determined, which was more reliable than
detection of the various isotopic species. The sector field the other ages, because the 207Pb/206Pb system is less af-
mass spectrometer of Nier (1947), however, utilized an fected by recent radiogenic Pb loss. The magnitude of
204
electronic mode of detection, and this mode of detection Pb was used to correct for primordial or common Pb
became the accepted methodology in geochronological ap- contamination (Tilton et al., 1955), and clean-room facili-
plications, and was an important factor in the success of ties were established to minimize the amount of Pb con-
this new field of science. It was fortuitous that these instru- tamination introduced in the chemical processing. Tilton
mental and analytical advances were available to be uti- et al. (1955) also showed that isotopic U-Pb ages could
lized by geochronologists in the 1950s. be determined for minerals such as zircon and sphene.
Ahrens (1955) analyzed old U-rich and Th-rich miner-
als from Rhodesia, and showed that the 207Pb/235U,
206
III. DEVELOPMENT OF Pb/238U,and 208Pb/232Th ages were discordant with
GEOCHRONOLOGY respect to 207Pb/206Pb ages. He found that a plot of
207
Pb/206Pb ages against 207Pb/235U ages gave a definite
Developments in mass spectrometry (the 607 sector-field pattern in which the data points for the sample converged
instrument), geochronological techniques (notably the U, to an age that he took to be the age of crystallization.
Th-Pb, and the K-Ar techniques), and the availability of Wetherill (1955) realized that the Ahrens (1955) data
isotopically enriched materials (IDMS) provided the basic could be explained by a partial loss of radiogenic lead.
building blocks for geochronology to blossom as a field Wetherill (1955) plotted the 206Pb/238U against 207Pb/235U
of study in the 1950s. At that time, gas source mass spec- ratios, and showed that the Ahrens (1955) data fell on a
trometry was used almost exclusively, not only for K-Ar straight line that intersected an ‘‘age equality curve’’ at
age determination, but also for U, Th-Pb geochronology, two points. The upper intersection gave the primary crys-
because of its excellent sensitivity and the relative ease of tallization age of the samples, which fitted the straight
the chemical separation involved. However, as solid line, whereas the point of lower intersection gave the age
source mass spectrometry developed, it was applied to a of episodic lead loss. The ‘‘age equality curve’’ was given
large range of geochronological systems with end-prod- the name of ‘concordia’ by Wetherill (1955) (see Fig. 2).
ucts that are solids. If the isotopic composition of Pb extracted from a
number of rocks that have the same age and initial isotopic
composition are measured, then they may be plotted on a
A. Uranium, Thorium-Lead Method 207
Pb/204Pb vs. 206Pb/204Pb isochron diagram. The slope of
In many respects, age determination techniques based on the straight-line array is dependent on the age of the sam-
the decay of U and Th to Pb is the best known, most ples and on the U and Pb concentrations of the samples.
versatile, and most valuable of all the geochronological It is of interest to note that the first application of this
techniques. Primordial lead consists of four isotopes— common Pb-Pb method was applied by Patterson (1956)
204,206,207,
and 208Pb. 206Pb is the radiogenic end-product of to meteoritic material to obtain an age of 4.55 1 109 y,
238
U, Pb is the radiogenic end-product of 235U, and 208Pb
207
and he showed that the meteoritic troilite represents the
is the radiogenic end-product of 232Th. Changes in the primordial isotopic composition of Pb in the solar system.
isotopic composition of lead caused by radiogenic decay Patterson (1956) also showed that the earth has the same

100

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

 MASS SPECTROMETRY AND GEOCHRONOLOGY n

more, the ability of certain minerals to remain closed sys-

tems despite the fact that the rocks from which they are
extracted may have been isotopically disturbed, enables
U, Th-Pb dating to be a powerful geochronological system.
In most respects, zircon is the ideal crystal for U, Th-
Pb dating, because it has a lattice structure that often en-
ables U, Th and Pb to remain a closed geochemical system,
excludes initial (common) Pb from its structure, and is
widely distributed in silicate rocks. Other U- and/or Th-
containing minerals such as monazite, sphene (titanite),
and apatite can be used for dating. Krogh et al. (1987)
have shown that baddeleyite can be used to date basic
rocks, which, because of their paucity of zircons, have
proved difficult to date by other means.
Krogh (1982a) showed that it is possible to separate
radiation-damaged metamict grains in a high flux magnetic
FIGURE 2. D/P diagram, showing the Witwatersrand and Rhodesia separator to improve the quality of the geochronological
regularities found by Ahrens (1955). D1 and D2 are the mol concentrations data. An alternative method to improve the concordance
of the end-product isotopes, Pb206 and Pb207, respectively, and P1 and P2
of the zircon data is to remove the surface of the crystals
refer to the mol concentrations of the parent isotopes, U238 and U235,
respectively. The curve marked ‘‘concordia’’ is the locus of all points by abrasion (Krogh, 1982b). Another major advance was
such that tu-235 Å tu-238 (Wetherill, 1955). to mount single zircons in situ in a mass spectrometer ion
source. Kober (1986) showed that these single zircons
enabled discordant lead to be evaporated at low tempera-
age (and origin) as meteorites. This work was the basis ture conditions, whereas the concordant fraction was emit-
of terrestrial Pb isotope evolution models. ted at higher filament temperatures. Kober (1987) argued
In the pioneering article by Nier, Thompson, and Mur- that the concordant Pb fraction is incorporated into the
phy (1941), mass spectrometric analysis showed that sig- zircon lattice itself, whereas the discordant fraction is from
nificant variations occurred in the isotopic composition of Pb, which occupies defect sites. Scharer and Allegre
Pb in various lead ore minerals, such as galena. Nier, (1982) showed that it was possible to date zircons with
Thompson, and Murphy (1941) argued that present-day complex geochronological histories by breaking up large
Pb measured in galenas is a mixture of two components— single crystals of zircons and analyzing individual frag-
ments. Corfu (1988) developed a system of breaking the
primeval Pb of fixed isotopic composition that existed
tips off large zircon crystals in order to distinguish recent
at the time of formation of the earth, and radiogenic Pb
events from the crystallization age of the older cores.
generated by U and Th decay. Holmes (1946) and Houter-
Although innovative, these bulk techniques were lim-
mans (1946) independently suggested that model galena
ited in utilizing zircon geochronology to the fullest extent.
ages could be determined by assuming that the galena
An alternative approach is microanalysis, which can pro-
source rock is a closed system with a single stage Pb
vide unique information about the timing, growth, and
isotope history. A growth curve for the galena source can
recrystallization of minerals and on the origin of rocks and
be calibrated for different values for the age of the galena.
their subsequent geological history. The in situ analysis of
For common Pb ores of the same age, the radiogenic Pb from U- and Th-bearing minerals by ion microprobe
207
Pb/206Pb will be constant, whereas the concentrations of has revolutionized U, Th-Pb geochronology. Secondary
207
Pb and 206Pb will depend on the U concentration in the ion mass spectrometry was applied to the determination
source region. If the Pb isotopic data are measured relative of zircon ages by Hinton and Long (1979), but a major
to the invariant 204Pb, then the measured 207Pb/204Pb and problem was the presence of interferences in the mass
206
Pb/204Pb ratios will plot on a linear array, the slope of region of Pb. These interferences implied the necessity of
which is related to the time of mineralization and to the using a double-focusing machine of large resolution, one
age of the earth. Houtermans (1947) called this line an of which was designed and built at the Australian National
‘‘isochrone’’—although it is now referred to as an isochron. University (Compston, Williams, & Meyer, 1984). This
Although U, Th-Pb systems do not always remain sensitive high resolution ion micro probe mass spectrome-
closed in silicate rocks, the special relationships between ter (SHRIMP) is described in a later section of this review.
235
U and 238U with their end-products 207Pb and 206Pb, re-
spectively, in terms of geochemical behavior, enables age- B. Potassium-Argon Method
information to be obtained even from systems that have Potassium possesses three isotopes, 39, 40, and 41, of
undergone metamorphism and/or weathering. Further- which 40K is radioactive and decays to two end-products,

101

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

n DE LAETER

40
Ca and 40Ar, by a branched nuclear decay scheme. Al- the irradiated sample so that an age of retention can be
though 89% of the disintegrating nuclei decay to 40Ca, this calculated for each step of the heating regime. Further-
potential geochronological method was not developed in more, because K and Ar can be measured simultaneously
the 1950s because of limitations in thermal ionization mass by mass spectrometry, the measuring systematics can be
spectrometry (TIMS). On the other hand, the sensitivity simplified. The progressive heating technique provides a
of gas source mass spectrometry was sufficient to enable wealth of information on the geological history of the rock
Aldrich and Nier (1948) to demonstrate that the 40Ar/36Ar under examination, and is, therefore, a powerful method
ratio from potassium-rich minerals was several times in geochronology.
higher than the terrestrial (atmospheric) value; this value York et al. (1981) demonstrated the enormous poten-
opened the way for the K-Ar method to be exploited. tial of using a laser beam in K-Ar geochronology by mea-
In addition to the fact that the isotopic measurements suring the age of a whole rock sample of slate by per-
of Ar were relatively straightforward by gas source mass forming a step-wise heating procedure similar to 39Ar-40Ar
spectrometry, potassium is widely distributed in the crust, analysis. Layer, Hall, and York (1987) showed that the
so that the method is applicable to most rock units. The age of a hornblende standard Hb3GR as determined by
fundamental question that concerns the validity of the laser heating was identical, to within experimental limits,
technique was whether the gaseous radiogenic 40Ar was of the conventional step heating age. The technique was
retained quantitatively in potassium minerals. Wetherill, applied to biotites by Layer, Hall, and York (1987), who
Aldrich, and Davis (1955) measured the 40Ar/40K ratios of showed that the measured age was dependent on the size
coexisting feldspars and micas from the same set of rocks of the biotite grains. The laser has now become an ac-
and found that, in each case, the feldspar yielded a lower cepted device for step heating and spot dating in K-Ar
40
Ar/40K ratio than the mica. Reynolds (1957) demon- geochronology.
strated experimentally that Ar diffused from feldspars
more readily at high temperatures than micas, and further-
C. Rubidium-Strontium Method
more, that even at low temperatures, there would be Ar
loss by diffusion over geologically long periods of time. In 1949, Ahrens published the results of an extensive in-
The K-Ar method was, therefore, developed by using mica vestigation into the Rb-Sr geochronological method,
rather than feldspar, although it was shown that reliable which built on earlier work by Strassman and Walling
ages could also be obtained from muscovite and biotite. (1938). Ahrens (1949) concluded that the Rb-Sr method,
Lipson (1958) showed that it was possible to use the K-Ar based on the decay of 87Rb to 87Sr with a half-life of 4.88
method to determine the time of deposition of sedimentary 1 109 y, could produce reliable ages for a variety of min-
rocks by using the minerals glauconite and sylvite. eral and whole rock samples. Aldrich, Doak, and Davis
In the early 1960s, efforts were undertaken to apply (1953) demonstrated that ion exchange techniques could
the K-Ar method to a variety of other minerals. Hart be used to chemically separate Rb and Sr from other ele-
(1964) showed that hornblende and pyroxene would be ments, to enable small concentrations of these elements
suitable for K-Ar analyses, thus providing the possibility to be measured with the IDMS technique. Initially, Rb-Sr
of dating mafic igneous rocks. Evernden and Curtis (1965) dating was confined to Rb-rich and Sr-poor minerals, such
developed improved methods for K-Ar dating of low po- as lepidolite and biotite, where most of the 87Sr measured
tassium minerals in young rocks, and were able to measure was radiogenic; but, as TIMS techniques improved, the
the age of volcanic minerals associated with the fossil method was applied to a wide range of geological samples.
hominids in Africa. It is of importance to note that these Schreiner (1958) found that whole rock samples could be
same authors pointed out the unreliability of K-Ar ages used as well as mineral separates. Compston and Jeffery
measured on whole rock samples of basalt due to the (1959) established the Rb-Sr technique as a genuine geo-
possibility of argon loss. chronological method by describing how the age of pri-
An ingenious method of overcoming the concerns re- mary crystallization in isotopically disturbed systems
lating to the validity of K-Ar dates was developed by could be measured by analyzing several whole-rock sam-
Merrihue and Turner (1966). This method was called the ples from the one rock unit. Metamorphic events could be
Ar-39 method, and involved the conversion of a proportion obtained by analyzing mineral separates from the same
of the 39K atoms in the sample to 39Ar by fast neutron rock unit.
capture. When heated, the sample releases the 39Ar to- Nicolaysen (1961) suggested a graphical representa-
gether with 40Ar. If the proportion of 39K atoms to 39Ar is tion of the Rb-Sr isotopic data by plotting the measured
known, then the measured 39Ar/40Ar ratio can give the age 87
Sr/86Sr ratios on the ordinate and the 87Rb/86Sr ratios on
of the sample. The Ar-39 method is now the definitive the abscissa. This technique has the advantage that samples
methodology in K-Ar (or Ar/Ar) dating, because argon with the same age and initial 87Sr/86Sr ratio define a line
loss (or gain) can be observed from step-wise heating of whose slope is proportional to its age, and whose intercept

102

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

 MASS SPECTROMETRY AND GEOCHRONOLOGY n

laboratories had published some data on their IDMS values

of Rb and Sr for geochemical reference samples, it was
not until the early 1970s that the importance of reporting
Rb/Sr ratios on a routine basis was established (De
Laeter & Abercrombie, 1970). The timeliness of this pro-
cedure was illustrated with respect to the controversy that
arose over the Rb-Sr ages of lunar rocks reported by differ-
ent laboratories. The discrepancies in the measured ages
resulted from errors in the concentrations of the spike
solutions that had been used for Rb and Sr IDMS analyses
in at least one laboratory (De Laeter, Vernon, & Com-
pston, 1973). The National Institute of Science and Tech-
nology (NIST), previously known as the United States
National Bureau of Standards (NBS), played an important
role in producing reference materials for geochronology.
FIGURE 3. Isotopic analyses of biotites from the Baltimore gneiss NIST, for example, introduced a potassium feldspar Refer-
(Nicolayson, 1961). ence Material (SRM 607), which possessed a large Rb-Sr
ratio, as a specific response to the requirements of Rb-Sr
geochronology.
on the ordinate of the line is the initial 87Sr/86Sr ratio. The Mineral separates can be used to great advantage in
initial ratio could be used to distinguish between reworked, geochronology, because different minerals are affected by
Rb-rich, sialic material, which has a high initial ratio at metamorphism by differing amounts. In the case of Rb-
the time of emplacement, compared to the lower initial Sr geochronology, biotite is a valuable mineral in that it
ratios, which are associated with material from the Rb- is common in granites and gneisses and can be separated
depleted mantle (Hurley et al., 1963). This graphical repre- into concentrates of high purity. The Sr closure tem-
sentation is now referred to as an ‘‘isochron diagram’’ perature of biotite is approximately 3207C (Harrison &
(see Fig. 3), and the same concept has been used to good McDougall, 1980), and the biotite Rb-Sr system, therefore,
advantage in other geochronological systems. The iso- reflects age events that are different from whole-rock Rb-
chron concept ranks with the concordia concept as two of Sr ages. De Laeter and Libby (1993) have demonstrated
geochronology’s most valuable tools. that Rb-Sr biotite dates give a completely different geo-
The Rb-Sr method was one of the first geochronologi- chronological pattern from whole-rock Rb-Sr dates near
cal methods exploited by the sector field mass spectrome- the western margin of the Yilgarn Craton in Western Aus-
ter of Nier (1947), because Rb and Sr are both of low tralia, and that the younger ages probably represent reset-
ionization potential; that factor enabled TIMS to provide ting during uplift in the early Palaeozoic era.
sufficient sensitivity for essentially all geological samples.
The Rb-Sr method has provided a large body of geochro-
D. Samarium-Neodymium Method
nological information over a wide time-scale, and has
proved a very useful tool in the earth sciences. It has The Sm-Nd geochronological method is based on the de-
also provided the framework for the development of a cay of 147Sm to 143Nd with the emission of an alpha particle
standardization methodology in geochronology. and a half-life of 1.06 1 1011 y. The original application
Geochronologists who use the Rb-Sr technique nor- of the Sm-Nd method was in cosmochemical research with
malize their Sr isotopic ratios to a 88Sr/86Sr value of 8.3752, meteorites and lunar samples (Lugmuir, Scheinin, &
because the abundances of 88Sr and 86Sr are both invariant Marti, 1975). The long half-life of 147Sm makes it more
in nature. A strontium carbonate ‘standard’, named the suited to Precambrian age determinations rather than dat-
‘‘Eimer and Amend standard,’’ was circulated to many ing younger samples. Even for old samples, the quantity
laboratories in the 1960s, and it became customary to of radiogenic 143Nd produced is comparatively small, and
quote the Sr isotopic ratios of this standard to provide necessitates highly precise isotope ratio measurements of
some indication of the precision of the mass spectrometric Nd to obtain an accurate age. The requirements of Sm-
measurements carried out in that laboratory. However, no Nd geochronology were important factors in the develop-
such comparative method was used routinely to assess a ment of more advanced solid source mass spectrometers.
laboratory’s values for their 87Rb/86Sr ratios despite the The distinguishing feature of the Sm-Nd method is
fact that these values were an integral part of the isochron that Sm and Nd are both rare earth elements (REE) and,
methodology to determine the age and initial 87Sr/86Sr ratio hence, are not fractionated to any great extent in many
of the samples concerned. Although a few geochronology geochemical environments, particularly with respect to

103

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

n DE LAETER

thermal effects. Rare earth elements are found in all rocks because of its high ionization potential. The combined
as trace elements, but may be concentrated in certain min- effect of these natural characteristics implies laborious an-
erals (e.g., monazite), and can be fractionated in some alytical procedures to obtain ages of sufficient accuracy
minerals, so that sampling and chemical dissolution must to compete with other geochronological methods. In fact,
be given special attention. In fact, the chemical separation the 176Lu-176Hf system has proved to be more useful as a
of Sm from Nd and from other REEs is not a trivial exer- geochemical tracer than in geochronological studies (Pat-
cise, and great care must be taken to obtain good separa- chett & Tatsumoto, 1980).
tions by ion exchange chemistry.
The isochron technique can be applied to the reduction
F. Rhenium-Osmium Method
of Sm-Nd data. Because the Sm-Nd system is difficult to
187
perturb, on a whole-rock scale even high-grade metamor- Re decays to 187Os by beta decay with a half-life of 4.47
phism does not affect the system, so that whole-rock iso- 1 109 y. Re and Os are both enriched in metals so that,
chrons can give the age of magmatic differentiation of in this respect, it is unique among the commonly used
metamorphosed as well as undisturbed rocks (in contrast to geochronometers. The Re-Os method was introduced by
the Rb-Sr and U-Pb systems). However, the geochemical Herr et al. (1961), who showed that iron meteorites and
similarity of Sm and Nd implies that whole-rock isochrons terrestrial samples such as molybdenite could be dated.
often do not yield precise ages because of the lack of Re and Os both occur in ppb concentrations in rocks, and
spread in 147Sm/143Nd ratios within a suite of samples. the high ionization potential of these elements presented
Igneous minerals can give the age of crystallization, sensitivity problems for TIMS at that time; however, the
whereas metamorphic minerals give the age of recrystalli- method is now used in a number of geochronological labo-
zation. In this respect, the Sm-Nd method has considerable ratories (Luck, 1989). The SIMS technique has been used
advantages over the Rb-Sr method. to good effect for the Re-Os method (Luck, 1989). The
An alternative application of Sm-Nd is the use of large fractionation between Re and Os in magmatic pro-
model Sm-Nd ages. For the simplest model, these ages cesses makes this method of value not only for age deter-
are referred to as CHUR (chrondritic uniform reservoir) minations, but also for petrogenetic studies in silicate sys-
ages. The significance of these ages depends on whether tems (Luck & Arndt, 1985). The large variation in Os
the rocks are basaltic or acidic, that is, whether they are isotope abundances between crust and mantle makes Os
derived fairly directly from the mantle (basaltic) or have a valuable geochemical tracer. As will be discussed later,
been involved with crustal materials (intermediate to negative thermal ionization has also been successfully
acidic), which may themselves have resided in the crust used for Re-Os mass spectrometry. The Re-Os method has
for a significant time after fractionating from the mantle. also been used as a cosmochronometer to determine the
Again, the assumption of a chrondritic reservoir in the mean age of nucleosynthesis of the elements (Fowler,
mantle is not always correct, and other models must be 1972).
considered (De Paolo, 1981).
G. Potassium-Calcium Method
E. Lutetium-Hafnium Method
As noted earlier, 40K is radioactive and decays to 40Ar
176 176 10
Lu decays to Hf with a half-life of 3.57 1 10 y and dominantly by electron capture, and to 40Ca by beta decay.
can, therefore, be used as a geochronometer. The 176Lu-176Hf The total half-life of 40K is 1.26 1 109 y, with the branch
method also promised to be an excellent cosmochrono- favoring 40Ca decay by 88.8% to the 40Ar branch of 11.2%.
meter, but the nature of the nuclear systematics for 176Lu The value of the decay of 40K to 40Ar in geochronology
rendered it unreliable in estimating the mean life-time of was first demonstrated by Aldrich and Nier (1948) to form
the synthesis of the chemical elements (De Laeter et al., the basis of the K-Ar method; however, the 40Ca decay
1988). The isochron technique can be applied to the Lu- branch was not exploited for age-dating at that time due
Hf system in much the same way as for the Rb-Sr and to limitations in mass spectrometric instrumentation. 40Ca
Sm-Nd methods. is the most abundant naturally occuring isotope of Ca
Although the 176Lu-176Hf method is intrinsically sim- (96.941%), which is itself an abundant element in the
ple, it suffers from the serious disadvantage that, as a REE, earth’s crust, whereas 40K is a very minor isotope of K in
the concentration of Lu in most rocks and minerals is terms of abundance (0.0117%). Thus, the enhancement in
40
extremely low. This low concentration is exacerbated by Ca due to the radioactive decay of 40K is very small and
the fact that 176Lu represents only 2.58% of Lu. Variations is difficult to distinguish vs. ‘‘common’’ 40Ca.
in 176Lu/176Hf are small, and these factors limit the accu- The first researchers to produce high-quality mass
racy with which the age of ‘‘ordinary’’ rocks can be mea- spectrometric data for Ca were Russell, Papanastassiou,
sured. In addition, the mass spectrometry of Hf is difficult and Tombrello (1978); however, their study was directed

104

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

 MASS SPECTROMETRY AND GEOCHRONOLOGY n

primarily to resolve the extent of isotopic fractionation in and calcium carbonate deposition rates in marine environ-
Ca rather than to exploit its use as a geochronometer. ments.
Marshall and De Paolo (1982) were the first to successfully The U-series geochronometers have traditionally been
use the K-Ca method in age-dating and petrogenetic stud- based on measuring the activity ratios of the radiometric
ies. They showed that the K-Ca geochronometer could be pairs by counting techniques. However, in the light of
treated in a similar manner to the Rb-Sr geochronometer, improvements in mass spectrometry, greater precision is
particularly with respect to the isochron technique. Nelson now possible by measuring the amount of the accumulated
and McCulloch (1989) undertook a substantial investiga- decay product rather than by counting directly the radioac-
tion of the potential usefulness of the K-Ca method in tive decay for some radioactive pairs. For example, 234U
petrogenetic studies, using a multicollector methodology and 230Th are alpha emitters, and the traditional method of
in which a double-jump measurement procedure was used analysis has been by isotope dilution alpha spectrometry.
to overcome the large relative mass range in the abundance However, the applicability of the technique to dating ma-
of the Ca isotopes. They concluded that the K-Ca method rine precipitates and studying the characteristics of the
can be applied to a variety of petrogenetic and geochrono- source region during magma generation has been limited
logical problems. by analytical uncertainty and diagenetic changes in the
Fletcher et al. (1997a) have examined the problems host minerals. The importance of these geological prob-
associated with the K-Ca method, namely large relative lems is highlighted by the fact that other geochronological
mass range, beam dispersion, and mass fractionation. A techniques, such as 14C dating, cannot give useful informa-
thermal ionization mass spectrometer fitted with a modi- tion for ages approaching 105 y.
fied analyzer tube produced a dispersion of ú4 amu in An alternative to radioactive counting is high-preci-
the Ca mass range, permitting static multicollector data sion isotope dilution TIMS. Edwards, Chen, and Wasser-
acquisition. Mass fractionation effects were reduced by burg (1986) have described their IDMS technique for mea-
loading the Ca sample in the iodide form, and by control- suring the abundance of 230Th, which, coupled with previ-
ling the operating procedures as much as possible to give ous improvements in measuring 234U, 238U, and 232Th by
reproducible fractionation patterns. The reduced fraction- mass spectrometric methods (Chen, Edwards, & Wasser-
ation renders the data insensitive to the choice of fraction- burg, 1986), significantly reduce analytical uncertainties
ation law used for internal corrections. Fletcher et al. and enable the sample size to be reduced. Edwards, Chen,
(1997a) loaded K as the iodide in the same mass spectrom- and Wasserburg (1986) demonstrated that it is possible to
eter, and showed that the procedure also enhanced the measure 6 1 108 230 Th atoms to {3% and 2 1 1010 230 Th
reproducibility of the K isotope ratios. For single K analy- atoms to {0.2%, using isotope dilution mass spectrometry.
ses, a fractionation limit of approximately 0.8‰ per amu The mass spectrometric technique established by Ed-
was demonstrated. However, the K-Ca geochronometer is wards, Chen, and Wasserburg (1987) has been used exten-
less robust than the Rb-Sr system due to the greater mobil- sively in recent years to determine the age of corals. A 3-g
ity of 40Ca with respect to 87Sr (Fletcher et al., 1997b). sample of coral can be dated to 180 { 5 ky, and older
Thus, it seems unlikely that the K-Ca method will be corals (231.1 ky) to {1 ky with a 250 mg sample. This
used extensively in geochronology, except for specialized technique can give valuable information on climate change
applications. and sea level variations (Eisenhauer et al., 1993; McCul-
loch et al., 1996). U-Th mass spectrometric ages are now
coupled with Sr/Ca and 18O measurements to enable
H. Uranium-Series Disequilibrium ocean-atmosphere interactions to be deciphered from the
coral record. Another application of the U-Th ages on
The decay chains of 238U to 206Pb and 235U to 207Pb com- corals has been to provide a calibration of 14C ages mea-
prise a series of radioactive isotopes of many elements, sured on the same samples by accelerator mass spectrome-
until the stable Pb isotopes are reached. The many decay try. Bard et al. (1990) have shown that the U-Th ages
products of 235U and 238U can be separated from their accord with the dendochronological record and are, there-
parents by geological processes because of their differing fore, accurate to at least 30,000 y ago. Before 9000 y ago,
geochemical properties that interrupt the radioactive decay the 14C ages are systematically younger than the U-Th
series. These separations lead to the possibility of measur- ages, with a maximum difference of approximately 3500
ing geological time, albeit of the order of the last million y at 20,000 y B.P. Thus, the effectiveness of 14C dating
years or less, depending on the half-life of the particular has been improved by this system of U-Th age calibration.
radionuclide. These disequilibrium methods of dating fill
an important gap in the geological time frame between I. Dating by Cosmogenic Isotopes
14
C dating and the K-Ar method. These U-series methods The estimation of the time of formation of the Solar Sys-
have proved their importance in dating sedimentation ages tem was achieved by a consideration of a number of radio-

105

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

n DE LAETER

nuclides with geologically short half-lives that were syn- TABLE 1. Cosmogenic isotopes produced
thesized during nucleosynthesis in stars. Although the par- in the atmosphere by cosmic ray
ent radionuclides have long since decayed away, their end- bombardment (Raisbeck & Yiou, 1989).
products can be identified in primitive meteorites. The
Isotopes Half-life
existence of these so-called ‘‘extinct radionuclides,’’ such
as 129I, 26Al, 107Pd, and 41Ca, were distinguished by identi- 10
4Be 1.5 1 106 y
fying excesses in the stable decay products 129Xe, 26Mg, 26
Al
13 7.16 1 105 y
36
107
Ag, and 41K, respectively, from primitive meteorites us- Cl
17 3.0 1 105 y
81
ing mass spectrometry (De Laeter, 1990). Although these Kr
56 2.1 1 105 y
41
Ca 105 y
radionuclides that were produced in nucleosynthesis are 20
14
6C 5730 y
now extinct, they are also produced as a result of the 39
Ar 270 y
18
earth’s bombardment by cosmic rays, and to some extent 32
Si
14 100 y
the solar wind. The cosmic rays, in their interaction with 22
Na
11 2.6 y
terrestrial matter, particularly the atmosphere, produce nu- 3
1H 12.3 y
35
clides that are collectively known as cosmogenic isotopes. S
16 87 d
7
Some of these cosmogenic isotopes have half-lives that 4B 53.3 d
37
make them of potential value in geochronology. Ar
18 35 d
33
The most well-known cosmogenic isotope is undoubt- P
15 25 d
32
P 14.3 d
edly 14C, which has a half-life of 5568 y. Carbon has three 15

isotopes—two stable isotopes 12C and 13C, together with

the radioactive isotope 14C. 14C is absorbed from the atmo-
sphere by plants and animals, and retains this atmospheric ences as a dating tool, especially in environmental studies.
14
C value until the organism dies, when the existing 14C Raisbeck and Yiou (1989) discuss the applications of 10Be,
26
decays with its characteristic half-life. It is possible to Al, and 41Ca in the earth sciences, and emphasize that
determine the time since the death of the organism if the the advent of AMS has been the driving force behind the
activity of the 14C in the organism is measured, assuming development of dating using these cosmogenic isotopes.
that the cosmic ray production of 14C has remained con- A number of cosmogenic nuclides are produced in
stant. Because the formation and decay of 14C is in radioac- situ in surface rocks. Although low in abundance, they
tive equilibrium, it is possible to determine the natural have opened up a wide variety of geological processes to
radioactivity of 14C from a living organism as 13.56 { dating in the 10 ka to 1 Ma timescale (Nishiizumi et al.,
0.07 disintegrations per minute per gram of C (Karlen et 1986; Cerling & Craig, 1994). Timing of glaciation, ero-
al., 1966). sion, soil formation, volcanic eruption, floods, landslides,
A variety of radioactive-counting techniques have and stone pavement formation have all been studied, open-
been developed to measure the activity of 14C decay, but, ing up a new quantitative era in geomorphology.
as the age of the sample increases, the count rate decreases The dating of very young igneous activity has been
to the stage where it is impossible to distinguish the 14C hampered by the lack of suitable chronometers. The cos-
count rate from background noise. This situation imposes mogenic nuclide 3He provides a method of dating surface
an upper limit of approximately 40,000 y on the 14C exposure ages by conventional mass spectrometry (Kurtz,
method. An alternative measurement technique is to distin- 1986; Craig & Poreda, 1986). 3He extracted from mineral
guish 14C by its mass, using a mass spectrometer. Unfortu- separates from a lava is measured by static gas source
nately, sector-field mass spectrometers are limited to an mass spectrometry, using ion counting (Anthony & Poths,
abundance sensitivity of approximately 109, whereas the 1992). The high-performance, double-focusing mass spec-
concentrations of 14C are at a level from 1 1 1012 to 1 1 trometer of Nier and Schlutter (1985) is an ideal instru-
1015. Furthermore, classical mass spectrometric techniques ment for this purpose. Mineral separates (such as olivine),
are unable to resolve 14C from other isobaric interferences are crushed and heated while on-line to the mass spectrom-
such as 14N, 12CH2, and 13CH (Duplessy & Arnold, 1989). eter. The heavy noble gases are removed, enabling He and
The advent of accelerator mass spectrometry enabled the Ne to both be isotopically analyzed. Cosmogenic 3He/21Ne
latter alternative of measuring 14C to be a practical reality. measurements can confirm that 3He has been retained by
Details of AMS and its application to 14C dating are dis- the samples.
cussed in Section V.
The situation with respect to 14C is mirrored with respect IV. ADVANCES IN SECTOR-FIELD
to other cosmogenic isotopes, a list of which is given in MASS SPECTROMETRY
Table 1 (Raisbeck & Yiou, 1989). Many of these cosmogenic The basic sector-field mass spectrometer of Nier (1940;
isotopes have found increasing application in the earth sci- 1947) provided an instrument to the scientific community

106

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

 MASS SPECTROMETRY AND GEOCHRONOLOGY n

that could be utilized by various scientific disciplines to tion. For example, DC amplifiers were used in the de-
fulfill their own objectives. This situation was certainly tecting systems of many mass spectrometers in the early
true for the earth science community, particularly for geo- 1950s, and these amplifiers required considerable attention
chronology. The measurement of time, on a scale appro- to produce adequate data. Chemists were also involved in
priate for the history of planet Earth, and the interpretation evolving chemical separation methods for the U, Th-Pb,
of those dated events, in terms of the origin and subsequent and the Rb-Sr methods.
history of the materials involved, is the essence of geology.
It is for this reason that modern geochronology, based on
A. Solid-Source Mass Spectrometry
a study of naturally occurring, time-dependent isotopic
systems and on the law of radioactive decay, and the inter- In the 1960s, a demand was created for more reliable mass
pretation of such ages in terms of geologic events, has spectrometers, so that scientists would not be dependent
become such an essential component of the earth sciences on ‘‘home-made’’ instruments. The first commercial mass
(Harper, 1973). spectrometers comprised an ion source, in which only one
The availability of separated isotopes following work sample could be analyzed at any one time, a 607 sector
at the Oak Ridge National Laboratory in the USA during magnet often mounted in the vertical plane, and a single
World War II, together with the development of the sector- detector, usually a Faraday Cup collector. By the 1960s,
field mass spectrometer, enabled geochronology to de- vibrating reed electrometers had replaced DC amplifiers,
velop quickly in the 1950s. The U, Th-Pb method could but the stability of the detection system was critically de-
be applied to a range of U- or Th-bearing minerals such pendent on the quality of the input resistor across which
as zircon, apatite, and sphene, whose elemental concentra- ion-induced voltages were measured. To change a sample,
tions were extremely small compared to earlier work. The it was necessary to vent the whole system up to atmo-
IDMS technique enabled these small amounts of U, Th, spheric pressure before loading another sample, and wait-
and Pb to be measured with good accuracy, and new chem- ing for the system to pump down by an oil or mercury
ical techniques were developed to separate these elements diffusion pump backed by a rotary pump. The magnetic
from other contaminants. field was swept to and fro across the isotopes in question,
The K-Ar technique, which had been developed in and data reduction was achieved by measuring the height
the late 1940s, was also applied to numerous geological of the various isotopic peaks from a chart recorder. If an
problems in the 1950s, and by the end of that decade had electron multiplier was used instead of a Faraday Cup in
become a well-established geochronological method as the the detecting system, then the shape of the peaks usually
result of major progress in gas-handling techniques. An deteriorated, so that there was a ‘‘play off’’ between in-
important advance in geochronology was made by the creased sensitivity and a decline in the quality of the isoto-
introduction of the Rb-Sr method during the 1950s, be- pic data.
cause Rb and Sr could both be efficiently ionized by TIMS, Three major advances occurred in the mid 1960s—
provided that these elements could be analyzed in a rela- reliable high-vacuum gate valves became available to iso-
tively pure form. This development was made possible by late the ion source from the remainder of the spectrometer,
ion-exchange chemistry, and the IDMS technique again digital voltmeters replaced chart recorders as the method
provided the ability to measure the concentration of these of data analysis, and peak-switching rather than magnetic
elements at the trace level. scanning became the accepted method of data collection.
Thus, by the end of the decade, the earth science The advent of digital computers enabled the isotopic data
community possessed two age-dating methods based on to be reduced off-line, until Hagan and De Laeter (1966)
the radioactive decay of the alkali elements Rb and K. demonstrated that on-line data collection and reduction
This situation represented a major advance, because Rb could be achieved with a time-sharing computer. Thus,
and K are both widely distributed in nature, so that a range the 1960s witnessed a revolution in mass spectrometry
of rock-forming minerals could be dated together with that enabled the output of geochronological data to be
whole-rock samples. No longer would geochronologists increased significantly. The availability of commercial
be restricted to rare or accessory minerals that contain mass spectrometers was a key factor in the establishment
high concentrations of U or Th (Harper, 1973). of new geochronological laboratories, and, in turn, pro-
The basic sector-field mass spectrometer of Nier vided the incentive for commercial manufacturers of mass
(1940; 1947) proved to be quite adequate for the early spectrometers to develop more reliable and versatile in-
geochronological studies of the U, Th-Pb, K-Ar, and Rb- struments in the 1970s. One of the innovations was to
Sr systems. In the 1950s, physicists working together with introduce turret ion sources, which enabled up to 16 sam-
geologists were involved in geochronology laboratories, ples to be loaded at a time.
because mass spectrometers were often unreliable due to The advent of the Apollo lunar program at the end of
vacuum problems and inadequate electronic instrumenta- the 1960s had a significant impact on the development of

107

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

n DE LAETER

geochronology. The availability of returned lunar samples gium has produced another set of certified U-isotopic ref-
provided the opportunity for selected laboratories to under- erence materials with a conservative uncertainty of 0.05%
take an age-dating program of these rare, unusual samples. on the abundance ratios (Rosman et al., 1987). There is a
A competitive atmosphere was engendered by the selec- tendency to believe that, because one possesses a new,
tion of international laboratories, all of which wanted to sophisticated, modern mass spectrometer, it is unnecessary
prove their capability. One of the outcomes was to estab- to calibrate the machine on a regular basis. This state of
lish newly designed clean-room facilities, where contami- mind is dangerous, because no matter how ‘‘impressive’’
nation could be kept to a minimum, and chemical tech- the mass spectrometer, there is no guarantee that it is
niques developed, which minimized blanks. Another out- performing to specifications unless systematic calibration
come was the necessity to calibrate the performance of a checks are made.
mass spectrometer against internationally accepted stan- Another outcome of the Apollo program was that a
dards, so that ages measured in different laboratories could number of laboratories designed and built new TIMS. One
be compared against each other on a systematic basis. of the most successful of these endeavors was undertaken
In fact, a controversy arose over the values of Rb-Sr by G. J. Wasserburg and his colleagues at the California
ages of lunar rocks reported by different laboratories. The Institute of Technology. The ‘‘Lunatic Asylum’’ mass
discrepancies in these ages were the result of errors in the spectrometer laboratory at California Institute of Technol-
concentration of the spike solutions that had been used ogy comprised two solid-source mass spectrometers of
to determine the Rb and Sr IDMS analyses (De Laeter, advanced design and high performance (Wasserburg et al.,
Vernon, & Compston, 1973). It was possible to resolve 1969).
the discrepant ages because of the availability of certified
stoichiometric salts, RbCl SRM984 and Sr CO3 SRM 987,
which were released by the U.S. National Bureau of Stan- 1. Ionization Enhancement Methods
dards in 1970 and 1971, respectively. This experience rein- As smaller samples are used and elements with higher
forced the importance of certified isotopic reference mate- ionization potential are analyzed, it becomes necessary
rials and elemental reference materials certified by IDMS, to examine ways of improving the sensitivity of thermal
made available to the international mass spectrometric ionization instruments. The efficiency of a thermal ioniza-
community to ensure the compatibility of isotopic data. tion source is described by the Langmuir equation:
It is of interest to note that the U.S. National Bureau
of Standards established an Isotopic Reference Materials
program in 1954 (Mohler, 1955), with the aim of ensuring
in-sample and between-sample homogeneity in reference
n//noexp S D
f0I
KT
, (1)

materials. Over a period of some 25 years, this work was

driven by mass spectrometric applications in isotopic geo- where n/ and no represent the number of positive ions
chemistry and nuclear science. A comprehensive descrip- that leave the filament and the number of neutral atoms,
tion of reference materials, certified by isotope ratio mass respectively; f is the work function of the filament; I is
spectrometry, is given by De Bièvre et al. (1993). the ionization potential of the element being analyzed; T
It was also realized that, in order to ensure interna- is the temperature of the filament; and k is the Boltzmann
tional comparability, so that an age measured in one local- constant. High work-function metals such as Re, Pt, Ta,
ity by a certain laboratory could be compared directly with or W are used for the filament materials; however, as the
a similar sample measured in another laboratory, particu- ionization potential of the element being analyzed in-
larly for time stratigraphic studies, that the characteristics creases, the ionization efficiency decreases. In geochronol-
of the mass spectrometer itself needed to be monitored. ogy, a lack of ionization efficiency does not present a
The U.S. National Bureau of Standards understood the difficulty for the alkali, alkali earth, and rare earth elements
need to validate the accuracy of mass spectrometers, and or Pb; however, when one attempts to analyze Hf, Re, and
introduced a systems calibration methodology that became Os, for example, insufficient sensitivity may be a problem.
widely accepted in the 1970s (Garner, Machlan, & Shields, The introduction of a multiple filament assembly (In-
1971). The procedure involves the analysis of a set of U- ghram & Chupka, 1953) gave greater control over evapo-
reference materials, the 238U abundances of which extend ration rates and in optimizing preferential evaporations;
from 0.02–97%. The calibration is designed to test the however, it was unable to cope with the poor ionization
total mass spectrometric system, including the operator. efficiencies for Hf, Re, and Os, and other similar elements.
Although time-consuming, it is essential that this method- Thus, other techniques were adopted to improve the ion-
ology be carried out to ensure that the mass spectrometer ization efficiency, to control the rate of evaporation, or to
is capable of producing the correct isotope ratios. The maximize the yield of a particular ion. Cameron, Smith,
Institute of Reference Materials and Measurements in Bel- and Walker (1969) showed that small quantities of certain

108

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

 MASS SPECTROMETRY AND GEOCHRONOLOGY n

high ionization potential elements could be successfully which enables relative precisions in the range 0.04–0.01%
analyzed if the filament was coated with a silica gel activa- for the Os ratios to be obtained. By coating Ni filaments
tor, together with phosphoric acid. This technique is partic- with a thin layer of V2O5 particles, it is possible to elimi-
ularly effective for the isotope analysis of Pb. The tech- nate the Re blank from Pt filaments (Heumann, 1996).
nique can be applied to a number of elements, and Ros- Concurrent with these developments in instrumenta-
man, Loss, and De Laeter (1984) have shown that the tion has been the necessity to improve chemical-pro-
efficiency of ionization for tin can be improved by adding cessing procedures in order to reduce chemical blanks and
approximately 80% by weight of boric acid to the silica to improve the purity of the extracted samples, so that the
gel as a replacement for phosphoric acid, which causes enhanced sensitivity in mass spectrometric instrumenta-
isobaric interferences in the mass spectrum of tin. tion can be fully utilized. Good clean-room facilities and
Loading the sample with graphite has proved effective reagent purity are mandatory, if ideal samples for thermal
in U and Th analysis (Edwards, Chen, & Wasserburg, ionization mass spectrometry are to be produced.
1987), and Noble, Lightfoot, and Scharer (1989) have used
a similar technique for Nd isotopes. Corfu and Noble 2. Resolution
(1992) showed that Hf ion production could be improved
by using fine-grained Ir metal on a single filament. Molyb- The conventional single-stage, sector-field mass spectrom-
denum powder also promotes ionization, whereas graphite eter of Nier (1940; 1947) had a typical mass resolution of
suppresses REE oxide interference (Dickin, 1995). 400 (defined as M/DM at 10% valley), which is adequate
Walker, Carter, and Smith (1981) used resin beads for for most geochronological applications. Improvements in
similar reasons. It should be noted that Nd analyses can ion optical design include the use of non-normal beam
be made by using Nd O/ ions from a single filament source incidence and curved pole pieces, which enable the effec-
in the presence of oxygen, although loading Nd with silica tive radius of curvature of the mass spectrometer to be
gel can achieve similar results without sacrificing the extended without affecting the size of the magnet. This
source vacuum (Thirlwell, 1991). Uranium can be mea- increase in dispersion improves the abundance sensitivity
and provides for a higher transmission efficiency without
sured as the oxide by placing TaO2 on the filament. Micro-
a deterioration in resolving power.
electrodeposition of the sample directly onto the filament
prior to the addition of a silica gel/phosphoric acid activa-
tor is an effective means of purifying the sample without 3. Abundance Sensitivity
increasing the blank (Loss, Rosman, & De Laeter, 1984).
One of the most important instrumental factors that affects
The technique is used for a number of elements, includ-
the ultimate accuracy of the measured isotope abundance
ing Pb.
ratios is abundance sensitivity. The scattering of ion beams
Negative thermal ionization can be used as an alterna- in a mass spectrometer results in the ‘‘tailing’’ of the mass
tive to the more commonly used positive ionization tech- peaks in the detector, in part due to collisions of the ions
nique, because it can produce high ionization efficiencies with gas molecules. Optimum vacuum conditions help,
for certain elements, and has the added advantage that but this effect cannot be eliminated, because some of the
organic impurities are reduced. In negative thermal ioniza- tail is produced by collisions of the ions with the spectrom-
tion mass spectrometry, the formation of negative thermal eter, by energy spread in the ion beam, and by electrostatic
ions is preferred by high electron affinities of the analyte repulsion within the ion beam itself. In order to measure
coupled with low electron work functions of the filaments a small number of ions of one isotope in the presence of
(Heumann, 1996). The compounds to be analyzed should a much larger ion beam from an adjacent isotope of the
have at least an electron affinity of 2 eV. Atomic ions same or different element, a mass spectrometer must pos-
or oxide ions are formed preferentially for many of the sess a high abundance sensitivity. The abundance sensitiv-
transition metals, nonmetals, and semi-metals. The work ity can be defined as the contribution to one peak by the
function of the filament material is reduced by the addition tail from the peak of a neighboring isotope. Conventional
of lanthanum or barium salts to the filament’s surface. As single magnetic-sector thermal ionization mass spectrome-
discussed above, the Re-Os geochronological method is ters, with a radius of curvature of 30 cm, typically have
extremely limited in positive thermal ionization mass spec- an abundance sensitivity in the 238 mass range of ca. 5
trometry, and it was not until Heumann (1996) demon- ppm (Belshaw & O’Nions, 1990). This factor limits the
strated that Os isotope ratios could be measured with high dynamic range in isotope abundance ratio measurements
precision by the negative mode that real advances in Re- of neighboring species in the mass spectrum, and the appli-
Os geochronology could be made. Heumann (1996) states cability of mass spectrometry to problems that require
that, for a 1-ng Os sample, an ion current of approximately accurate measurements of minor isotopes. High abundance
10011 A can be obtained for the most abundant OsO30 ions, sensitivity in TIMS instruments is essential for measuring

109

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

n DE LAETER

FIGURE 4. Finnigan MAT 262 RPQ extended geometry optics, giving a mass dispersion equivalent to
that of a standard 64-cm radius instrument. The center beam can be deflected into either a Faraday cup,
SEM, or RPQ.

230
Th/232Th ratios in U-series disequilibrium studies (Esat, 4. Multicollectors
1995), and for 210Pb/206Pb measurement of U-Pb ages of
zircon (Wendt, Wendt, & Tuttas, 1993). One of the most significant enhancements in mass spec-
Two approaches have been taken to improve abun- trometer design has been the advent of multicollection.
dance sensitivity. The traditional approach has been to Multicollection enables simultaneous measurements to be
construct multiple-stage instruments in which high resolu- made of a number of ion beams, thus removing errors
tion can be obtained by discriminating against stray ions caused by fluctuations of beam intensity that occur in the
by the use of two or more analyzers. Belshaw and O’Nions peak-jumping mode. Samples can, therefore, be measured
(1990) describe a sophisticated instrument that consists of with greater precision than is the case for single collector
an electrostatic and magnetic analyzer, followed by a sec- instruments in a given amount of time. Modern commer-
ond electrostatic analyzer. Results indicate that this mass cial mass spectrometers equipped with a computer-con-
spectrometer has an abundance sensitivity of 5 1 1009, an trolled, fully automated, multicollector system allow the
improvement of approximately three orders of magnitude analysis of a variety of elements with minimal operator
as compared to conventional single-stage instruments. The intervention. The collector system can also be fitted with
second approach to the problem of abundance sensitivity an electron multiplier or Daly Collector system for ion
has been to use an extended geometry 907 magnetic-sector beams that are too small to be measured with Faraday Cup
coupled with a static quadrupole mass filter (Fig. 4). The collectors. It is possible to achieve an internal precision
interface between the two mass analyzers is a decelerating of approximately 0.01 per mil on isotope ratios on modern
immersion lens and a beam-shaping static quadrupole lens. commercial instruments, although the external precision
De Bièvre and Perrin (1991) have shown that the abun- is limited by isotope fractionation effects and the gain/
dance sensitivity of an instrument of this type is approxi- response/electron suppression factors of the individual de-
mately 2 1 1008 at mass 237. tector (Aggarwal & Jain, 1997). Good mass spectrometric
Another factor in the improvement of the abundance practice demands the necessity to systematically determine
sensitivity in a modern mass spectrometer is that vacuum the gain of each Faraday Cup on a regular basis.
systems are now so much more efficient than in the past.
Turbomolecular pumps have replaced diffusion pumps, 5. Mass Fractionation
together with a cold finger located at the ion source region
to absorb gases that are produced during filament outgas- Isotopic compositions are normally expressed in terms of
sing. Ion getter pumps and cryo-pumps may also be used. abundances, but are measured as abundance ratios. An
The net result is that an operating vacuum better than 1007 isotope abundance ratio R is defined as the ratio of a
torr is now the norm for TIMS, and in some cases pressures number N2 of the atoms of isotope 2 compared to the
of approximately 1008 torr are routinely attained. number N1 of the atoms of isotope 1 of the same element:

110

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

 MASS SPECTROMETRY AND GEOCHRONOLOGY n

R Å N2/N1. (2) the ionization yield throughout the duration of the analysis.
The small amount of sample that needs to be used in this
Accuracy is defined as the uncertainty interval stated with technique necessitates excellent chemical separation to
a measurement, such that the interval contains the absolute avoid contamination. Another technique is to monitor the
or true value with a certain probability. An ‘‘accurate’’ isotope fractionation during the measurement of each sam-
isotope abundance ratio is one in which the uncertainty ple and to take the ratio for each sample at precisely the
interval is likely to contain the true value. Thus, the mea- same time after the commencement of the analysis in the
surement of R must contain an assessment and correction sample and reference material measurement. It is, of
for all of the systematic errors associated with it, and the course, necessary to ensure that samples in such a compari-
uncertainty on R must contain all the uncertainties of the son sequence are analyzed under the same experimental
corrections for all these systematic errors (De Laeter, De conditions. This technique is suited to gas-source mass
Bièvre, & Peiser, 1992). spectrometry.
Precision is defined as the uncertainty associated with
the reproducibility of an isotope abundance ratio measure-
B. Gas-Source Mass Spectrometry
ment. The precision can be expressed in two ways: as the
standard deviation of a number of measurements on a The title of the Nier (1947) article, which was the basis
single sample (the internal standard deviation), or as the of modern sector-field mass spectrometry, was ‘‘A mass
standard deviation of the mean of a number of measure- spectrometer for isotope and gas analysis.’’ In fact, the
ments of several samples of the same material (the external motivation of the article was ‘‘to describe a new mass
standard deviation). In both cases, a ‘‘standard deviation of spectrometer which was designed particularly for the iso-
the mean’’ can also be calculated (De Laeter, De Bièvre, & topic analysis of elements such as carbon, oxygen or nitro-
Peiser, 1992). gen or for making gas analyses in general’’ (Nier, 1947).
Mass fractionation represents a fundamental limitation Nier, Ney, and Inghram (1947) introduced the null method
to the accuracy of isotopic ratio measurements in mass of determining isotopic abundance ratios into gas-source
spectrometry. Mass fractionation can arise from a variety mass spectrometry, in which dual collections of the two
of sources—from evaporation processes in thermal ioniza- isotopes under study were undertaken simultaneously to
tion sources, in sputtering processes in ion microprobe optimize the quality of the isotopic measurements by elim-
spectrometers, and in collector systems. Furthermore, the inating fluctuations in the ion beams. This dynamic mode
measured isotopic ratios may change progressively with of gas-source mass spectrometry, coupled with rapid inter-
time during an analysis. It is also important to distinguish change between a sample and a corresponding standard,
between instrumental-induced mass fractionation, and enabled very precise data to be determined. The source
fractionation caused by natural processes. The double- construction of the Nier (1947) mass spectrometer also
spike technique can be used to eliminate machine-induced permitted the insertion of a small furnace for studying
fractionation, and, therefore, enables physiochemical iso- substances that have low vapor pressures.
topic effects to be determined unambiguously (Russell, Apart from K-Ar geochronology, which was ideally
1971). Light elements are particularly prone to mass spec- suited to gas-source mass spectrometry, many of the early
trometer-induced mass fractionation. This mass fraction- age determinations using the U-Pb method were carried
ation can be greatly reduced by measuring a heavy molecu- out by analyzing U and Pb in the gaseous form. However,
lar species that contains the light element, rather than by the development of the Rb-Sr method quickly led to the
measuring the isotopic composition of the element itself development of TIMS, and, because most of the geochro-
(e.g., Fletcher et al., 1997a). nologically useful radiometric pairs are solids, gas-source
Isotope fractionation effects can be overcome, for mass spectrometry has not played as important a role in
some elements, by internal normalization. This technique geochronology since the early 1950s except for K-Ar geo-
is used in Rb-Sr geochronology by normalizing the isoto- chronology.
pic data to a 88Sr/86Sr ratio of 8.3752. In Sm-Nd geochro- The dynamic mode in gas-source mass spectrometry
nology, 146Nd/144Nd and 147Sm/152Sm ratios are used for provides for the sample being analyzed to be held in a
internal normalization of the Nd and Sm isotopes, respec- reservoir connected to the changeover valve by a capilliary
tively. However, this technique is not possible for elements leak, as is the standard (Pillinger, 1984). The sample or
with only two isotopes (such as Rb). standard is allowed to leak into the ion source, which is
One approach to the problem of time-dependent frac- continuously pumped during the analysis. This method is
tionation is to integrate the total ion beam of each isotope excellent, provided that sample size is not a limiting factor,
over the duration of the ‘‘life’’ of the sample. This method- which is the situation for the majority of K-Ar analyses.
ology has been applied by Fiedler (1996), but an assump- The sensitivity of these dynamic instruments was not par-
tion in this total evaporation approach is the constancy of ticularly good, and only a factor of 50-fold improvement

111

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

n DE LAETER

in sensitivity was achieved in this technique from 1947 to the measured isotopic data obey gas kinetic theory to better
1975 (Pillinger, 1984). A major advance was the introduc- than 1 part in 1000 (De Bièvre, 1996).
tion of an ion-counting system, which produced an im- One of the problems with gas-source mass spectrome-
provement in sensitivity of approximately 200-fold ters that operate at a high accelerating potential is that the
(Schoeller & Hayes, 1975). Dynamic instruments are used ions may be implanted into metallic components in the
primarily for C, N2, O2, and H2 systems, where sample vacuum system, so that memory effects are a potential
size is rarely a limiting factor. problem. Care must be taken to outgas the mass spectrom-
A quantum jump in improvement in the sensitivity of eter and to monitor any memory effects in succeeding
gas-source mass spectrometry was achieved by Reynolds analyses. An alternative is to use a low accelerating poten-
(1956), who introduced the static method, in which the tial in the ion source in conjunction with a small permanent
gases of interest are introduced into the mass spectrometer magnet. Peak-switching is achieved by varying the accel-
after its isolation from the pumping line. Reynolds (1956) erating potential; however, this method leads to poorer
and his colleagues used the static methodology for noble precision in the isotopic ratios even though it minimizes
gases, where effective gettering can ensure an adequate memory effects.
vacuum. Pillinger (1992) extended the static methodology Another advance that has increased the sensitivity of
to other gases—primarily N2 and C. Ar analyses are now static noble gas mass spectrometry to near the theoretical
made on static systems. Another major advance in the limit, detecting approximately one-third of the atoms in the
improvement in the quality of isotopic information from mass spectrometer, is the Baur–Signer ion source (Baur &
gas-source mass spectrometry was the concept of step- Signer, 1980). It is widely used for K-Ar geochronology
wise heating. Merrihue and Turner (1966) demonstrated and in state-of-the-art noble gas mass spectrometers (Ho-
the effectiveness of this technique in Ar-Ar dating of mete- henberg, 1980).
orites, and it is now an essential tool in K-Ar dating. The
great advantage of step-wise heating is that it enables one
to unravel anomalous subsystems of the sample from the
V. RECENT DEVELOPMENTS
major component. In the K-Ar method, it can be success-
IN MASS SPECTROMETRY
fully applied to samples with an inherited Ar component
or ones in which Ar loss has occurred. Step-wise heating
can be applied to mineral or whole-rock samples. Laser A. Secondary Ion Mass Spectrometry
heating of samples has been used successfully in such
Secondary Ion Mass Spectrometry (SIMS) is a powerful
step-wise heating experiments.
analytical technique for investigating the surfaces of mate-
Figure 5 is a schematic diagram of an argon extraction
rials. Ion bombardment of the surface under examination
line coupled to a static gas mass spectrometer (Dalrym- enables information about its compositional characteristics
ple & Lanphere, 1969). The amount of Ar in the sample to be obtained, and how this composition varies with
is determined by introducing a known volume of 38Ar depth, as controlled erosion of the surface takes place. The
spike to the sample by opening valve 1. The amount of sample is bombarded with high-energy ions such as O0,
38
Ar spike is determined by the volume of the capilliary O20, Cs/, or neutral noble gas atoms. A fraction of the
tube between valves 1 and 2. Aliquots of the spike are sputtered ions from the specimen are emitted into the mass
calibrated on a periodic basis against a known volume of spectrometer, which can be a time-of-flight, quadrupole,
atmospheric Ar. or magnetic sector-field design.
Mass discrimination in gas-source mass spectrometry The SIMS technique has the advantage of high sensi-
was investigated by Honig (1945), who showed that the tivity and excellent depth resolution and can analyze light
gas reaching the ion source could have a different isotopic elements such as H, C, O, and N, as well as heavier ele-
composition from that in the reservoir, if the gas pressure ments. SIMS has been particularly valuable for the study
in the reservoir was high or the dimensions of the capilliary of materials (Tyagi, Dash, & Rajagopalan, 1997), but has
leak tube were small. These changes in isotopic composi- also been indispensable to cosmochemical studies of mete-
tion correspond to a change from viscous to molecular oritic inclusions (Ireland, 1996). Ion microprobe analysis
flow, in which the mass discrimination is equal to the allows the measurement of the isotopic composition of
square root of the mass ratio. De Bolle, Valkiers, and De single grains down to the micron size. Anomalous isotopic
Bièvre (1990) have described a procedure that enables the material may be related to astrophysical settings, which
errors in gas-source mass spectrometry to be minimized are possible sites for grain formation (Ireland, 1996). The
by extrapolating the observed isotope abundance ratio val- benefit of ion probe analysis is that a single inclusion can
ues to time zero; that is, at the time the gas is introduced be analyzed for a variety of elements while maintaining
into the mass spectrometer. It is of interest to note that the petrographic context of the sample. Zinner and his

112

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

 MASS SPECTROMETRY AND GEOCHRONOLOGY n

FIGURE 5. Schematic diagram of an argon extraction line coupled to a static gas mass spectrometer
(Dalrymple & Lanphere, 1969).

colleagues at the University of Washington in St. Louis, field. This configuration also possessed good vertical fo-
Missouri, have carried out some exciting work of astro- cusing characteristics. The design process included an ex-
physical significance on presolar grains using a Cameca amination of the influence of fringing fields with an accu-
Ion Probe (Nittler et al., 1997). racy effectively of third order.
The ability to isotopically analyze minerals in situ, In the early 1970s, an ion microprobe of high sensitiv-
without the need of chemical processing, in a mass spec- ity and with a mass resolution M/DM of at least 5000
trometer with high sensitivity and good resolution, has (using a DM value measured at the 1% peak height) was
long been a goal for geochronology. Compston and his commercially unavailable, so that the decision was made
colleagues at the Australian National University realized by the Australian National University to build a sensitive
that a SIMS-type mass spectrometer would be required, high resolution ion microprobe (SHRIMP) mass spectrom-
incorporating an efficient ionizing source (such as a duo- eter based on the design of Matsuda (1974). Details of the
plasmatron), which could analyze elements such as Pb and design and fabrication of the instrument are described by
U with high sensitivity. These requirements imply the need Compston (1996).
for a double-focusing mass spectrometer with energy fo- The configuration of SHRIMP is shown schematically
cusing to compensate for the large energy spread of ions
in Fig. 6. The production of secondary ions in SHRIMP
sputtered from a sample by the duoplasmatron ion
is achieved by ion-impact sputtering. A primary beam of
source. The mass spectrometer would also need to pro-
negatively charged ions is extracted from the oxygen
vide high resolution to resolve interfering isobars in the
plasma of a duoplasmatron ion source and accelerated to
vicinity of Pb and U so that U, Th-Pb geochronology
a potential of 10 KV. After focusing and steering, this
can be carried out.
Double-focusing mass spectrometers have been suc- beam is passed through a Wien magnet to produce a mass-
cessively refined since the early work of Aston (1919). filtered O20 beam that impacts at an angle of 457 onto the
These spectrometers were used to measure the atomic target. The primary beam size is defined by a circular
masses of the nuclides. Second-order ion optical designs, aperture, the dimension of which can be selected as appro-
which permit double-focusing machines of ever-increasing priate, and Köhler illumination provides a constant ion
resolving power to be constructed, are described by De density. The image of the aperture is demagnified by a
Laeter, De Bièvre, and Peiser (1992). Matsuda and Matsuo factor of six with a pair of Einzel lenses, and the final
(1971) described a method for calculating the second-order result is a flat-bottomed, 15 1 25 mm, elliptical sputter
image aberrations of a double-focusing mass spectrometer, crater, the size of which can be varied by altering the size
and subsequently Matsuda (1974) showed that it was pos- of the aperture in the Köhler lens segment of the ionization
sible to design such an instrument, which attained com- column. The primary beam current is typically 4 nA.
plete second-order focusing. This instrument consisted of In order to use SHRIMP for geochronological pur-
a cylindrical electrostatic analyzer plus an electrostatic poses, it is necessary to monitor biases and machine drift
quadrupole lens, together with a homogeneous magnetic during the analysis with a suitable reference material. In

113

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

n DE LAETER

can be used for geochronological purposes. However, a

problem in developing geochronological techniques for
these minerals is in the availability of suitable standards,
which must be characterized by conventional TIMS and
be present in sufficient quantity to overcome the difficulty
of finding new standards from time to time. A review of
the applications of SHRIMP to various geochronological
systems is given by Compston (1996). The performance
characteristics of SHRIMP have been described by De
Laeter and Kennedy (1998), who argue that SHRIMP si-
multaneously meets the requirements of high mass resolu-
tion and high ion transmission efficiency together with
good abundance sensitivity. In terms of spatial and mass
resolution coupled with high sensitivity, it is ideally suited
FIGURE 6. Schematic of the SHRIMP II. The spectrometer comprises
for in situ analysis of U- and/or Th-bearing minerals and
an 857 cylindrical electrostatic analyzer of turning radius 1.27 meters
followed by a 72.57 sector magnet of turning radius 1 meter (Compston, enables high-quality geochronological information to be
1996). routinely obtained.

B. Accelerator Mass Spectrometry

the case of zircon geochronology, gem-quality zircons
from Sri Lanka are used. Conventional thermal ionization Mass spectrometry has progressively adopted techniques
mass spectrometry U/Pb isotopic data on this zircon stan- developed in other areas of nuclear science to improve the
dard gives an U concentration of approximately 550 ppm sensitivity and precision with which the isotopic composi-
and a concordant age of 564 Ma. A piece of the zircon tion of stable or radioactive nuclides can be measured.
standard is always mounted together with zircon grains The use of accelerators in mass spectrometric measure-
extracted from each sample to enable determination of ments has provided a ‘‘quantum jump’’ in sensitivity for
the Pb/U ratios of the sample by using comparative data some isotopic systems. The extension of high sensitivity
analysis. A reproducibility of approximately 2% can be mass spectrometry to isotope ratios below 10012 has been
routinely achieved for 206Pb/238U for 15 analyses of the called ‘‘accelerator mass spectrometry’’ (AMS), because
standard zircon; the uncertainty is largely determined by of the use of an additional stage of acceleration to remove
the homogeneity of the standard. molecular interferences and the separation of isobars by
SHRIMP is ideally suited to the U, Th-Pb geochrono- rate of energy loss or range methods.
logical system, and this system has been the dominant use The basic idea of AMS is to discriminate between
of the instrument. A major reason for the development of atomic species in the ion source by using negative ions to
SHRIMP was the prospect of detecting age differences destroy interfering molecular species, and then to count
between different growth zones within a single mineral identified individual accelerated ions. The detection of ul-
(Compston, 1996). In fact, SHRIMP has now enabled the tratrace quantities of elements by conventional mass spec-
frontier of zircon geochronology to be within individual trometry is limited by the ability of the spectrometer to
minerals. The ability to focus an ion beam on a single discriminate between a variety of molecules and atomic
zircon crystal with a spot size less than 20 mm in situ, species of nearly the same mass as the isotopes under
and to measure the isotopic characteristics of that ‘‘spot’’ investigation. However, because multiply charged mole-
before moving to another nearby location, can provide an cules decompose rapidly, any process that can eliminate
understanding of geological processes on a near to micron two or more electrons from a molecule can be used to
scale. Crystals such as zircons with complex structures remove these molecules from the mass spectrum
can, therefore, be examined to reveal their geological his- (Litherland, 1980).
tory. SHRIMP can also be used to measure the ages of Accelerator mass spectrometry has the ability to accel-
Palaeozoic zircons, such as the analyses of 2 Ma old zir- erate singly or doubly charged ions to a velocity such that,
cons from the Western Himalaya (Zeitler et al., 1989). on the average, three or more electrons are removed from
The instrument has also been successfully applied to lunar the atom or molecule after traversing a thin foil or corre-
zircons (Compston, Williams, & Meyer, 1983), and to sponding thickness of gas. An ion energy of approximately
zircons from meteorites (Ireland & Wlotzka, 1992). 6 MeV is required for AMS to reach optimum efficiency.
SHRIMP is a versatile machine, because numerous Although small electrostatic accelerators with positive ion
other U- and Th-bearing accessory minerals such as mona- sources can be used, they have the disadvantage of requir-
zite, apatite, titanite, perovskite, rutile, and baddeleyite ing the ion source or the detector system to be at a high

114

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

 MASS SPECTROMETRY AND GEOCHRONOLOGY n

voltage. On the other hand, tandem accelerators have the with a high 14C/12C ratio would require an abundance sen-
advantage of changing the negative ions to positive ions sitivity of the order of 1 1 10012, so that radioactive count-
at the central positive electrode followed by subsequent ing was the only viable method of 14C dating, even though
acceleration of the positive ions. The ion source and detec- mass spectrometry’s ability to measure the integrated de-
tor systems are both located outside the high voltage elec- cay product is a powerful alternative. It was not until AMS
trode. However, in using multiple-charged atoms to elimi- was introduced (Muller, 1977) that a viable alternative to
nate molecular interferences, all charge states õ 10 are radioactive counting was available. Because AMS has an
present. Consequently, mass and charge spectrometry are abundance sensitivity of the order of 1 1 10015, it is able
both required to eliminate unwanted charge states. The to accommodate even old 14C samples. The efficiency of
use of energy, mass, and charge signatures at energies AMS 14C measurements is illustrated by the fact that it
such that charge states /3 or higher are dominant is the yields the same count rate from 1 mg of C as the beta
basis for this type of ultra-sensitive mass spectrometry, count rate does from 1 g of C (Dickin, 1995).
which can be applied to essentially all the stable isotopes. One of the most publicized experiments carried out
In fact, AMS can be used for all elements that form nega- by AMS was that of the determination of the age of the
tive ions, with the notable exception of the noble gases. Shroud of Turin. The Shroud of Turin, which was popu-
One of the most important applications of accelerator larly believed to be the cloth that was used to wrap Christ’s
mass spectrometry is the carbon-14 dating technique. Car- body after the crucifixion, has been housed in the Royal
bon-14, which has a half-life of 5730 years, is generated in Chapel of Turin Cathedral since 1694 (Damon et al.,
the atmosphere by cosmic rays. The observed beta particle 1989). Although many scientific tests were performed on
count rate from 1 g of contemporary carbon of biological the Shroud in this century, the most definitive test as to
origin is only about 15 counts per minute, although there its authenticity was to measure its age. This measurement
are some 6.5 1 1010 atoms of 14C in the 1 g sample (Muller, was not a practical possibility at a time when 14C dating
1977). The advantage of a mass spectrometric technique required large samples to provide an accurate age; how-
to determine the 14C content is significant as compared to ever, AMS provided the means of enabling 14C dates to be
conventional counting techniques in that the accumulated determined with a minimum amount of destructive testing.
decay products can be measured rather than a count rate Three AMS laboratories (Tucson in the USA, Oxford in
(provided that a high sensitivity mass spectrometric tech- the UK, and Zurich in Switzerland) were selected for the
nique can be established). Accelerator mass spectrometry study. Three samples, whose ages had been determined
offers this possibility, because 14C can be separated from independently, were also dated as controls. The labora-
isobaric 14N because of the instability of the N0 ions in tories did not know which of the four samples to be ana-
the ground and isomeric states. The optimum condition lyzed was that of the Shroud, and strict controls on the
for 14C/// detection occurs at 12 Mev, which is ideal for experiment were instituted. The 14C mean age as deter-
a tandem accelerator. A detection limit of 0–1 micrograms mined by the three laboratories was 691 { 31 years, which
of C can be achieved with AMS, and the 14C method can corresponds to a calibrated calendar age range from 1260–
be extended back to approximately 50,000 y with this 1390 AD at the 95% confidence level. These results pro-
technique. vide conclusive evidence that the linen of the Shroud of
Figure 7 is a schematic diagram of the Rochester Uni- Turin is mediaeval (Damon et al., 1989). A compilation
versity accelerator mass spectrometer (Litherland, 1980). of the applications of AMS is given by Jull (1995).
The negative ion sputter source produces up to 10 mA of Accelerator Mass Spectrometry is also the preferred
12 0
C ions at 20 keV. These ions are magnetically analyzed, method of measuring the short-lived radionuclides 10Be,
accelerated, and focused onto aperture AP1. The C0 ions 36
Cl, 129I, 26Al, and 41Ca. Beryllium-10 is produced when
are accelerated to 8 meV, and five electrons are removed cosmic rays interact with N2 and O2 in the atmosphere.
to produce C/4 ions, which are further accelerated and The resulting spallation products include 10Be, which has
focused onto aperture AP2. The C/4 ions are analyzed a half-life of 1.51 1 106y. 10Be has a short residence time
by two magnets, followed by electrostatic analyzer and in the atmosphere, because it attaches itself to aerosols
detector stages. The 12C and 13C ions are separated from and becomes part of the precipitation cycle. It can be used
14
C by the electrostatic analyzer. The 14C count rate can as a tracer of climatic change and, in conjunction with
be compared to the injected 12C0 ion current measured on 14
C, to understand the processes of formation in the atmo-
Faraday Cage collector FCl. The ratio of these two count sphere by cosmic rays. Studies of 10Be by AMS in ice
rates is related to the 14C/12C ratio, which can be used to cores (Raisbeck et al., 1981) give a good correlation with
determine the age of the sample undergoing analysis. d18O, which supports the solar modulation model of cos-
As discussed above, the abundance sensitivity of a mic ray intensity. 10Be can be used for soil profile studies,
TIMS fitted with a retarding lens is typically of the order because Be is absorbed by soil particles (Pavich et al.,
of 1 1 1009. In the case of 14C dating, even modern carbon 1985). 10Be has also been used successfully for dating

115

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

n DE LAETER

FIGURE 7. The ion beam transport system of the Rochester University tandem accelerator (Litherland,
1980) used as an ultra-sensitive mass spectrometer is shown schematically. Ion-beam defining apertures
are designated by AP, and Faraday cups for ion beam current measurements are designated FC.

sediment burial by the radioactive decay of 10Be and 26Al C. Inductively Coupled Plasma Mass
in quartz (Granger, Kirchner, & Finkel, 1997). Spectrometry
Chlorine-36, with a half-life of 3 1 105 y, is similar
to Be in terms of its production (by spallation on 40Ar)
10
One of the limitations of TIMS is the lack of sensitivity
and its short atmospheric residence time. Because chlo- for high-ionization potential elements. One of the most
rine-36 is not absorbed by soil particles, it is a good tracer significant advances in mass spectrometry in recent years
of Quaternary ground water systems. For example, Bentley has been to use an inductively coupled plasma ion source,
et al. (1986) analyzed 36Cl/total Cl ratios in a study of the in conjunction with a quadrupole mass spectrometer (ICP-
movement of ground water samples in the Great Artesian MS), to produce a versatile analytical technique, especially
Basin in Australia. This experiment enabled the ages of for elemental analysis. ICP-MS is a sensitive multielement
these waters to be estimated. analysis technique in which samples can be introduced
Iodine-129 is formed in the atmosphere as a spallation into the ICP source with relative ease. The ICP source not
product of Xe, and has been used to study the entry of only has a high ionization efficiency for most elements,
anthropogenic material into natural systems (Dickin, but also produces singly charged ions with low energy
1995). 129I has a half-life of 15.7 1 106 y. 129I is also a spread, which can, therefore, be matched to a variety of
fission product of nuclear explosions, so that care must be mass spectrometers such as a quadrupole, a time-of-flight,
taken to separate the two sources of 129I in hydrological or a magnetic sector field instrument.
studies (Fabryka–Martin, Davis, & Elmore, 1987). The ICP-MS enables solution aerosols to be injected
Aluminium-26, which has a half-life of 7 1 105 y, is into an inductively coupled argon plasma to generate a
produced in measurable quantities by cosmic rays imping- high density of positive ions derived from the elemental
ing on the earth’s surface. An important role of 26Al has constituents. Alternatively, solid samples can be analyzed
been to determine terrestrial ages of meteorites, a topic by laser ablation, where the ablated particulate matter is
that is outside the range of this review. However, 26Al can swept by argon gas into the ICP source. A fraction of the
be used as a geochronometer of rock exposure ages, often resulting ions is extracted through a sampling orifice into
in conjunction with another cosmogenic radio-nuclide a differentially pumped vacuum system that contains the
(Middleton & Klein, 1987). In addition, 41Ca, which has mass spectrometer.
a half-life of 1.04 1 105 y, shows great promise as a dating The dominant use of the ICP-MS has been in measur-
tool (Fink et al., 1990). ing elemental concentrations, because the precision and

116

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

 MASS SPECTROMETRY AND GEOCHRONOLOGY n

FIGURE 8. A schematic diagram of the multiple-collection ICP-MS (Halliday et al., 1995).

accuracy of isotope ratio measurements is restricted by a wide flight tube and magnet, with an additional Faraday
the lack of detector linearity, background contributions, detector located on the high-mass side, so that U can be
and/or mass discrimination effects (Russ & Bazan, 1987). detected simultaneously with Pb (Halliday et al., 1995).
The use of ICP-MS instruments fitted with quadrupole Russ and Bazan (1987) have demonstrated that the
mass spectrometers for isotope abundance measurements mass discrimination in the source of the multiple collector
is, therefore, inferior to the precision and accuracy attained ICP-MS is a simple function of mass, and that the mass
by thermal ionization instruments, which, for several de- bias in the ICP-MS source approximates to a power law:
cades, have been accepted as the benchmark technique for
isotopic analysis (Hirata, 1996). RTrue Å Rmeasured(1 / F)n, (3)
However, the combination of using an ICP ion source,
together with a magnetic sector field mass spectrometer where R is the ratio of the heavier isotope m1 to a lighter
fitted with a multicollector, has given a new perspective to isotope, m2; F is the mass bias per a.m.u.; and n is the
isotope abundance measurements by ICP-MS. The major difference in mass of the two isotopes (m1 0 m2). Although
limiting factor on the precision of ICP-MS has been the F is not constant over the entire mass range, it is a good
instability of the plasma; however, this instability can model for limited mass ranges. This mass bias relationship
be overcome by simultaneous static multicollection is in contrast to TIMS, where time-dependent mass dis-
(Walder & Freedman, 1992). Figure 8 is a schematic dia- crimination cannot always be accurately modeled. Taylor
gram of a multiple collector ICP-MS (Halliday et al., et al. (1995) examined mass discrimination effects in a
1995). The ICP-MS extracts the ions at high voltage and multiple collector ICP-MS by measuring the isotopic ratios
uses a quadrupole lens to focus the ion beam profile onto of a set of synthetic U standards. They concluded that
the analyzer entrance slit. An electrostatic analyzer a power law and exponential function result in the best
matches the energy dispersion of the plasma source with correction for mass discrimination.
that of the magnetic analyzer, which, in turn, provides Hirata (1996) has analyzed the lead standard reference
double dispersion with an effective 54-cm extended geom- materials NIST SRM 981 and 982 with the multicollector
etry. The collector assembly consists of eight indepen- ICP-MS, and has shown that the resultant Pb isotopic
dently adjustable Faraday Cups and a fixed axial Faraday ratios are in excellent agreement with the standard values,
Cup, which can be moved to enable the ion beam to im- using a 205Tl/203Tl value of 2.3871 to correct for mass
pinge directly onto a Daly collector. Thus, the multicollec- discrimination and the power law equation described
tor ICP-MS allows the simultaneous collection of up to above. Similar analyses were carried out for other geochro-
nine masses. A recent innovation has been the design of nological-important elements Rb, Sr, Nd, Hf, Re, and Os.

117

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

n DE LAETER

Although the magnitude of the mass discrimination for Os sure a few atoms of the noble gases with isotopic selectiv-
is significantly higher than for the other elements, the ity. Resonance ionization mass spectrometry combines
cause of this effect is currently not well understood. Hirata resonance ionization spectroscopy with mass spectrome-
(1996) argues that the ability to correct for mass discrimi- try. When the laser is tuned to a discrete resonant elec-
nation effects in a multicollector ICP-MS enables the pre- tronic transition, the probability of ionization increases
cision and accuracy of isotope ratios comparable to those enormously because of the step-wise absorption of pho-
of TIMS to be attained. tons. Fassett, et al. (1986) described a pulsed, tuned, UV
In terms of geochronology, multicollector ICP-MS of- laser in conjunction with a magnetic sector-field mass
fers the great advantage, through laser ablation, of analyz- spectrometer and a detector system capable of measuring
ing minerals in situ, without the need for time-consuming the pulsed ion currents produced in the experiment. This
chemical separations. Laser ablation is valuable for U, Th- principle of time-resolved magnetic dispersion improves
Pb geochronology, using U- or Th-rich minerals such as the abundance sensitivity for Re and Ta compared to con-
zircon, monazite, baddeleyite, and sphenes. However the ventional mass spectrometry without the necessity of ex-
SHRIMP mass spectrometer has been specifically de- tensive chemical processing. One of the most attractive
signed for this purpose, and at this time is a superior elements for RIMS analysis is Nd, an element that is useful
instrument; however, this situation may change in the fu- for geological applications and for reactor burn-up studies
ture. One of the problems of the ICP-MS method is that in the nuclear industry; however, isobaric interferences
it is necessary to control the mass discrimination from arise with Ce and Sm, and chemical separation of the rare
204
Pb to 238U in U, Th-Pb geochronology. Although Hirata earth elements is difficult. Smith et al. (1985) analyzed
(1996) has demonstrated the ability to control mass dis- nanogram-sized samples of Nd with RIMS, and evaluated
crimination for the Pb isotopes, where the mass range is the potential usefulness of the technique. Laser microprobe
from 204Pb– 208Pb, there is no convincing evidence that mass spectrometry (LAMMA) makes use of a high-inten-
mass discrimination can be controlled from 204Pb– 238U. sity laser pulse to vaporize and ionize small amounts of
ICP-MS may also have some advantage over conventional a sample, which are analyzed in a time-of-flight mass spec-
TIMS for La-Hf geochronology (Hirata, 1996). trometer. Simons (1983) traced the history of the develop-
ICPMS is not suited to laser ablation age studies, using ment of LAMMA-type instrumentation, and evaluated the
the Rb-Sr method, because isobaric interferences occur in performance characteristics of the machine. The primary
that mass range. In the case of the Sm-Nd method, it is advantage of a laser microprobe is its ability to analyze
unlikely that ICPMS can attain the high precision that is rapidly and with high sensitivity organic and inorganic
required for Sm-Nd analysis. On the other hand, ICP-MS, species present in the microregion of a sample by using
with its high sensitivity, has been very effective in Re-Os an ionization source in either the positive or negative de-
geochronology (Matsuda et al., 1986), in contrast to posi- tection mode. However, neither RIMS nor LAMMA has
tive ionization TIMS. However, negative ionization TIMS so far found any application in geochronology to any sig-
is an effective method to analyze Re-Os isotopes. One of nificant extent, and it is unlikely that they will in the
the differences between ICP-MS and SHRIMP is that the foreseeable future.
size of a laser ablation spot is larger than that of a SHRIMP It is appropriate that the scientist who was responsible
sputtering spot, which is typically less than or equal to 20 for the design of the 607 sector field mass spectrometer,
micron in diameter. The amount of sample consumed in an which heralded the birth of modern mass spectrometry
ICP-MS analysis is, therefore, larger than for a SHRIMP and enabled geochronology to be placed on a firm founda-
analysis of corresponding precision. tion in the early 1950s (Nier, 1947), was also responsible
for an innovative new mass spectrometer some 40 years
later (Nier & Schlutter, 1985). After his pioneering work in
D. Other Developments
developing the U-Pb and K-Ar geochronological methods,
In addition to SIMS, AMS, and ICP-MS, all of which have Nier designed a large radius of curvature, double-focusing
characteristics that are of value to geochronology, other mass mass spectrometer for precise atomic mass determinations
spectrometric techniques have also been developed such as (Johnson & Nier, 1953), and later became involved in
resonance ionization mass spectrometry (RIMS) and laser space physics experiments, which required the miniatur-
microprobe mass spectrometry (LAMMA). ization of mass spectrometer instrumentation (Nier & Hay-
Resonance ionization spectroscopy is a multistep ab- den, 1971). On the basis of this experience, a high-perfor-
sorption process in which the final state is the ionization mance, double-focusing mass spectrometer, which was
continuum of an atom. The process can be saturated with well suited for geochronology, was designed (Nier &
a pulsed laser, so that one electron can be removed from Schlutter, 1985). This 9.5-cm magnetic deflection mass
each atom of the selected type. Hurst et al. (1985) de- spectrometer employs Mattauch–Herzog geometry, and
scribed how this method can be used, in principle, to mea- the design features electrostatic and magnetic deflection

118

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

 MASS SPECTROMETRY AND GEOCHRONOLOGY n

FIGURE 9. A schematic diagram of the high-performance, double-focusing mass spectrometer of Nier

and Schlutter (1985).

in tandem to achieve the optimum resolution (Fig. 9). In chronological laboratories were developed by physicists
the single collector mode, repeated analyses of Nd and Sr working in collaboration with geologists.
standards show a precision of better than 0.005% for Nd The first geochronological techniques used were U,
and 0.01% for Sr. It has excellent performance for 3He/4He Th-Pb, K-Ar, and Rb-Sr dating; however, the subsequent
measurements and its high abundance sensitivity allows it diversification in the number and variety of these tech-
to be used for specialized applications for Kr and Xe. niques was matched by the growth in the number of labora-
The instrument can be used with either a thermal ion- tories around the world that became involved in geochro-
ization or electron impact source. Although designed pri- nological studies. This growth in the number of labora-
marily for 3He isotope analyses, this mass spectrometer tories enabled the geochronological community to demand
can be utilized for a wide range of geochemical problems. mass spectrometers that could meet the increasing needs of
The instrument has been particularly effective for surface reliability, sensitivity, resolution, and sample throughput.
exposure dating using 3He and 21Ne analyses (Anthony & Manufacturers of commercial mass spectrometers pro-
Poths, 1992). It can also be used to analyze Xe with an duced successive generations of instruments that incorpo-
overlap of less than 1 in 10,000 for adjacent mass numbers rated the best technology available.
(Nier & Schlutter, 1992). The availability of enriched isotopes enabled the sta-
ble isotope dilution mass spectrometry to be established
in the 1950s. This analytical technique provided the means
VI. CONCLUSIONS to determine absolute elemental abundances for nuclides
of geochronological importance at trace levels by means
Mass spectrometry and geochronology have been inextri- of mass spectrometry.
cably related since the discovery of isotopes led to the The lunar program had a major influence on the devel-
physical method of age determination. It is doubtful opment of geochronology because one of the program’s
whether any geochronological measurement has been objectives was to measure the age of lunar samples. This
made over the past fifty years that has not involved mass international effort brought into sharp focus the need for
spectrometry in one form or another. The onset of this better chemical extraction procedures to handle the rela-
period coincided with the development of the 607 sector- tively small, rare samples that had been returned from the
field mass spectrometer and its subsequent transition from lunar surface, and the need for certified reference materials
a research tool in physics to an instrument that could be that could provide a better control on ages measured in
used by earth scientists; however, in the 1950s, many geo- different laboratories. The resulting injection of funds into

119

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

n DE LAETER

geochronology laboratories enabled improved mass spec- ples in situ have long been objectives of geochronologists.
trometers to be designed, a new generation of clean rooms With a single-stage mass spectrometer, it was impossible
to be built, and reference materials to be provided. to achieve this goal, because in order to obtain the sensitiv-
One of the major geochronological ‘‘breakthroughs’’ ity required for small samples, one requires an ion source
in the 1960s was the development of the Ar-Ar technique, with higher efficiency than available by thermal ionization.
which largely eliminated a basic deficiency in K-Ar dating, The resulting energy spread in the emitted ions is greater
namely the loss of radiogenic Ar, particularly from old than can be handled in a single-stage instrument. A duo-
and/or weathered samples. The use of the ‘‘whole-rock’’ plasmatron source that uses sputtering as the mode of
technique in Rb-Sr dating was established together with ionization in a sensitive, high-resolution, ion-microprobe
the isochron method, which proved to be a versatile tool mass spectrometer (SHRIMP) incorporating an electro-
in geochronological studies. The use of mineral separates static and a magnetic analyzer has not only proved ex-
became a valuable adjunct to the geochronologists’ arse- tremely valuable for measuring the age of zircon crystals,
nal, particularly the use of zircons in U, Th-Pb geochronol- but can also be utilized in a wide variety of minerals
ogy. The concordia concept was a significant advance in containing U and/or Th. This instrument has revolution-
our understanding of the U-Pb method. ized U, Th-Pb geochronology, provided zircon ages with
The advent of Sm-Nd dating in the mid-1970s placed small uncertainties, and enabled the detailed geochrono-
greater demands on the precision of isotopic measure- logical history of old single zircons to be deciphered.
ments, and, together with the general demand for greater Microanalysis is undoubtedly the new frontier of geochro-
sample throughput, led, in part, to the development of nology.
multicollector systems. In a similar manner, the introduc- The realization that nuclear accelerators could be
tion of Lu-Hf and Re-Os dating provided the impetus for modified to become sophisticated mass spectrometers,
sensitivity improvements by using a variety of ion-enhanc- whose high abundance sensitivity could be used to mea-
ing techniques in thermal ionization ion sources, including sure 14C production and other cosmogenic nuclides, led to
the use of negative thermal ionization mass spectrometry. the establishment of the field of accelerator mass spec-
The use of lasers as an ionization mechanism was particu- trometry, which is suited to measuring ages of samples
larly effective in K-Ar studies and in ICP-MS. Improve- with low concentrations of relatively short-lived radioiso-
ments in ion optical design, in conjunction with widely topes. The technique has been exploited to good advantage
spaced multicollectors, provided the impetus to develop in archeological, environmental, and petrogenic studies.
the K-Ca geochronometer. Meanwhile, innovative single The most recent development in the evolution of mass
crystal zircon techniques offered a new level in the accu- spectrometry, which is applicable to age determinations,
racy with which ages could be determined by the U, Th- is inductively coupled plasma mass spectrometry, particu-
Pb method. larly when the instrument is fitted with a multicollector
Over the past 30 years, mass spectrometer instrumen- detector system. The ICP-MS has the advantage of being
tation has been revolutionized by the use of computers, able to analyze samples in liquid or solid phases, but suf-
multicollector systems, solid state electronics, improve- fers from the disadvantages of possible background con-
ments in the efficiency of ion production and vacuum tamination and mass discrimination effects. A great deal of
technology, and advances in ion optical design. Mass spec- effort is being expended to correct for mass discrimination
trometry has progressively adopted technological ad- effects so that ICP-MS can compete with TIMS in terms
vances developed in other scientific areas to improve the of the precision and accuracy of isotope ratio measure-
sensitivity and precision with which isotopic ratios can be ments for age determinations.
measured. The advent of modern computer-controlled Whereas TIMS has been the benchmark technique in
mass spectrometers that can routinely measure isotope geochronology since the 607 sector-field mass spectrome-
abundance ratios with a precision of better than 1 part in ter was developed (Nier, 1940; 1947), over the past decade
104 creates the illusion of high accuracy, which may not be three new mass spectrometric techniques—the sensitive
warranted. A comprehensive calibration regime, including high-resolution ion microprobe, accelerator mass spec-
testing the linearity of the measuring system, needs to be trometry, and inductively coupled plasma mass spectrome-
adopted to ensure that the abundance ratios are properly try—have been developed to the stage where their impact
corrected for mass discrimination. Isotope reference mate- on geochronology is significant. SHRIMP has revolution-
rials provide the means of subjecting mass spectrometers ized zircon geochronology and is now being applied to
to a rigorous ‘‘systems calibration’’ procedure. The sci- other minerals, AMS is uniquely suited to measure the
ence of accurate isotope abundance measurements is a effects of short-lived cosmogenic nuclides, and ICP-MS
complex, but often undervalued, activity. is being refined in order to compete with conventional
The desire to eliminate the necessity of chemical pro- mass spectrometry in geochronology.
cessing, and the ability to measure the age of small sam- The physical method of geochronology was possible

120

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

 MASS SPECTROMETRY AND GEOCHRONOLOGY n

because of the ability of mass spectrometers to measure 8. Aston, F. W. ‘‘The mass spectrum of uranium lead and the
isotopic ratios. The development of the 607 sector-field atomic weight of proactinium.’’ Nature 1929, 123, 313.
mass spectrometer provided an instrument of great power 9. Bard, E.; Hamelin, B.; Fairbanks, R. G.; Zindler, A. ‘‘Cali-
and versatility to the earth science community, which de- bration of the 14C timescale over the past 30,000 years
pended on this instrument as a variety of geochronological using mass spectrometric U-Th ages from the Barbados
techniques were established and more and more de- corals.’’ Nature 1990, 345, 405–410.
manding scientific tasks were investigated. It is, therefore, 10. Baur, H.; Signer, P. ‘‘Numerical simulation of ion source
pleasing to see that new and radically different mass spec- for gas mass spectrometer.’’ Helv. Phys. Acta 1980, 52,
trometers have now been established to enable even more 395.
demanding geochronological projects to be attempted. 11. Belshaw, N. S.; O’Nions, R. K. ‘‘The Cambridge ISOLAB
The intimate interplay between mass spectrometry and 120, an assessment of abundance sensitivity.’’ Advances
geochronology, which has been so evident in the past, will Inorg. Mass Spectrom. 1990, 1, 1–3.
continue with the innovative advances that have been 12. Bentley, H. W.; Phillips, F. M.; Davis, S. N.; Habermehl,
made in mass spectrometric instrumentation over the past M. A.; Airey, P. L.; Calf, G. E.; Elmore, D.; Gove, H. E.;
decade. The subject of this review provides an excellent Torgerson, T. ‘‘Chlorine-36 dating of very old ground wa-
ters. The Great Artesian Basin, Australia.’’ Water Re-
example of the interplay between science and technology.
sources Res. 1986, 22, 1991–2002.
This two-way interaction exhibited by geochronology on
13. Cameron, A. E.; Smith, D. H.; Walker, R. L. ‘‘Mass spec-
the one hand, and mass spectrometric instrumentation on
trometry of nanogram-size samples of lead.’’ Anal. Chem.
the other, is an exemplar of the manner in which science
1969, 41, 525–526.
progresses.
14. Cerling, T. E.; Craig, H. ‘‘Geomorphology and in situ cos-
mogenic isotopes.’’ Ann. Rev. Earth Planet. Sci. 1994, 317,
273–317.
ACKNOWLEDGMENTS 15. Chen, J. H.; Edwards, R. L.; Wasserburg, G. J. ‘‘238U, 234U
and 232Th in seawater.’’ Earth Planet. Sci. Lett. 1986, 80,
I would like to thank the following scientists who have 241–251.
provided input into this review: I. R. Fletcher, A. K. Ken- 16. Compston, W. ‘‘SHRIMP: Origins, impact and continuing
nedy, D. R. Nelson, K. J. R. Rosman, and M. Zadnik. Mrs. evolution.’’ J. Royal Soc. W. Aust. 1996, 79, 109–117.
S. Shephard prepared the manuscript with skill and effi- 17. Compston, W.; Jeffery, P. M. ‘‘Anomalous ‘‘common
ciency. strontium’’ in granite.’’ Nature 1959, 184, 1792–1793.
18. Compston, W.; Williams, I. S.; Meyer, C. ‘‘U-Pb geochro-
nology of zircons from lunar breccia 73217 using a sensi-
tive high mass resolution ion microprobe.’’ Proc. 14th Lu-
REFERENCES nar Planet. Sci. Conf.; J. Geophys. Res. 1984, 89, B525–
534.
1. Aggarwal, S. K.; Jain, H. C. ‘‘Thermal ionization mass 19. Compston, W.; Williams, I. S.; Meyer, C. ‘‘Ages deter-
spectrometry.’’ In Introduction to Mass Spectrometry, Ag- mined on lunar zircons in thin section.’’ Research School
garwal, S. K.; Jain, H. C., Eds.; Indian Society for Mass of Earth Sciences Annual Report, Australian National Uni-
Spectrometry, Trombay, 1997, 233–250. versity, 1983, 116–119.
2. Ahrens, L. H. ‘‘Measuring geologic time by the strontium 20. Coplen, T. B. ‘‘Atomic weights of the elements 1995.’’
method.’’ Bull. Geol. Soc. Amer. 1949, 60, 217–266. Pure Appl. Chem. 1996, 68, 2339–2359.
3. Ahrens, L. H. ‘‘The convergent lead ages of the oldest 21. Corfu, F. ‘‘Differential response of U-Pb systems in co-
monazites and uraninites.’’ Geochim. Cosmochim. Acta. existing accessory minerals, Winnepeg River Sub Prov-
1955, 7, 294–300. ince, Canadian Shield: Implication for Archaean crustal
4. Aldrich, L. T.; Doak, J. B.; Davis, G. L. ‘‘The use of ion growth and stabilization.’’ Contrib. Mineral. Petrol. 1988,
exchange columns in mineral analysis for age determina- 98, 312–315.
tions.’’ Amer. J. Sci. 1953, 251, 377–387. 22. Corfu, F.; Noble, S. R. ‘‘Genesis of the southern Abitibi
5. Aldrich, L. T.; Nier, A. O. ‘‘Argon 40 in potassium miner- greenstone belt, Superior Province, Canada; evidence from
als.’’ Phys. Rev. 1948, 74, 876–877. zircon Hf isotopic analysis using a single filament tech-
6. Anthony, E. Y.; Poths, J. ‘‘3He surface exposure dating nique.’’ Geochim. Cosmochim. Acta 1992, 56, 2081–2097.
and its implications for magma evolution in the Potrillo 23. Craig, H.; Poreda, R. J. ‘‘Cosmogenic 3He in terrestrial
volcanic field, Rio Grande Rift, New Mexico, USA.’’ Geo- rocks: the summit lavas of Maui.’’ Proc. Natl. Acad. Sci.
chim. Cosmochim. Acta 1992, 56, 4105–4108. USA 1986, 83, 1970–1974.
7. Aston, F. W. ‘‘A positive ray spectrograph.’’ Phil. Mag. 24. Dalrymple, G. B.; Lanphere, M. A. ‘‘Potassium-argon dat-
1919, 38, 707–714. ing.’’ Freeman. 1969, 258.

121

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

n DE LAETER

25. Damon, P. E. and 20 others, ‘‘Radiocarbon dating of the etry of thorium.’’ Int. J. Mass Spect. Ion Proc. 1995, 148,
Shroud of Turin.’’ Science 1989, 337, 611–615. 159–170.
26. De Bièvre, P.; De Laeter, J. R.; Peiser, H. S.; Reed, W. P. 43. Eve, A. S. Rutherford, McMillan: New York, 1939.
‘‘Reference materials by isotope ratio mass spectrometry.’’ 44. Evernden, J. F.; Curtis, G. H. ‘‘Potassium-argon dating of
Mass Spectrom. Rev. 1993, 12, 143–172. late Cenozoic rocks in East Africa and Italy.’’ Current
27. De Bièvre, P.; Perrin, R. E. ‘‘Why higher abundance sensi- Anthropol. 1965, 6, 343–385.
tivities on commercial mass spectrometers?’’ A. O. Nier 45. Fabryka–Martin, J.; Davis, S. N.; Elmore, D. ‘‘Applica-
Symposium, LA-12522-C, Rokop, D. J., Ed.; Los Alamos tions of 129I and 36Cl in hydrology.’’ Nucl. Instrum. Meth.
National Laboratory, 1991, 5–21. in Phys. Res. 1987, B29, 361–371.
28. De Bièvre, P. ‘‘Ultra-high accuracy isotopic measure- 46. Fassett, J. D.; Zeininger, N. J.; Moore, L. J. ‘‘Time-re-
ments: Avogadros constant is up.’’ J. Royal. Soc. W. Aust. solved magnetic dispersion for large isotopic ratio mea-
1996, 79, 11–19. surements in resonance ionization mass spectrometry.’’ Int.
29. De Bolle, W.; Valkiers, S.; De Bièvre, P. CBNM Report No J. Mass Spectrom. Ion Processes 1986, 69, 285–296.
GE/R/MS/09/90, Central Bureau of Nuclear Measurements, 47. Fiedler, R. K. ‘‘Total evaporation measurements, a key to
Geel, Belgium, 1990. more accurate mass spectrometry.’’ Proc. of the Second
30. De Laeter, J. R. ‘‘Mass spectrometry in cosmochemistry.’’ A.O. Nier Symposium. Green, L. W., Ed.; 1996, AECL
Mass Spectrom. Rev. 1990, 9, 453–497. 11342, 19–23.
31. De Laeter, J. R.; Abercrombie, I. D. ‘‘Mass spectrometric 48. Fink, D.; Middleton, R.; Klein, J.; Sharma, P. ‘‘41Ca mea-
isotope dilution analyses of Rb and Sr in standard rocks.’’ surement by accelerator mass spectrometry and applica-
Earth Planet. Sci. Lett. 1970, 9, 327–330. tions.’’ Nucl. Instrum. Meth. in Phys. Res. 1990, B47, 79–
32. De Laeter, J. R.; Allen, B. J.; Lowenthal, G. C.; Bolde- 96.
mann, J. W. ‘‘Constraints on the 176Lu Cosmochronom- 49. Fletcher, I. R.; Maggi, A. L.; Rosman, K. J. R.; McNaugh-
eter.’’ J. Ast. Astrophys. 1988, 9, 7–15. ton, N. J. ‘‘Isotopic abundance measurements of K and Ca
33. De Laeter, J. R.; De Bièvre, P.; Peiser, H. S. ‘‘Mass spec- using a wide dispersion multi-collector mass spectrometer
trometry in metrology.’’ Mass Spectrom. Rev. 1992, 11, and low fractionation ionization techniques.’’ Intern. J.
193–245. Mass Spectrom. Ion Proc. 1997a, 163, 1–17.
34. De Laeter, J. R.; Vernon, M. J.; Compston, W. ‘‘Revision 50. Fletcher, I. R.; McNaughton, N. J.; Pidgeon, R. T.; Ros-
of lunar Rb-Sr ages.’’ Geochim. Cosmochim. Acta 1973, man, K. J. R. ‘‘Sequential closure of K-Ca and Rb-Sr isoto-
37, 700–702. pic systems in Archaean micas.’’ Chem. Geol. 1997b, 138,
35. De Laeter, J. R.; Libby, W. G. ‘‘Early palaeozoic biotite 289–301.
Rb-Sr dates in the Yilgarn Craton near Harvey, Western 51. Fowler, W. A. ‘‘New observations and old nucleocosmo-
Australia.’’ Aust. J. Earth Sci. 1993, 40, 445–453. chronologies.’’ In Cosmology, Fusion and Other Matters,
36. De Laeter, J. R.; Kennedy, A. K. ‘‘A double focusing mass Reines, F., Ed.; Adam Hilger: London, 1972, 67–123.
spectrometer for geochronology.’’ Int. J. Mass Spectrom. 52. Garner, E. L.; Machlan, L. A.; Shields, W. R. Nat. Bur.
Ion Proc., to appear. Stand. (US) Spec. Publ. No 260-27, Washington, D.C.,
37. De Paolo, D. J. ‘‘Neodymium isotopes in the Colorado 1971.
Front Range and crust-mantle evolution in the Protero- 53. Granger, D. E.; Kirchner, J. W.; Finkel, R. C. ‘‘Quaternary
zoic.’’ Nature 1981, 291, 193–196. downcutting rate of the New River, Virginia measured
38. Dickin, A. P. Radiogenic Isotope Geology, Cambridge from differential decay of cosmogenic 26Al and 10Be in
University Press: Cambridge, U.K., 1995. cave-deposited alluvium.’’ Geology 1997, 25, 107–110.
39. Duplessy, J.-C.; Arnold, M. ‘‘Radiocarbon dating by accel- 54. Hagan, P. J.; De Laeter, J. R. ‘‘Mass spectrometric data
erator mass spectrometry.’’ In Nuclear Methods of Dating, processing using a time-shared computer.’’ J. Sci. Instrum.
Roth, E.; Poty, B., Eds.; Kluwer Academic: London, 1989, 1966, 43, 662–664.
437–453. 55. Halliday, A. N.; Lee, D–C.; Christensen, J. N.; Walder,
40. Edwards, R. L.; Chen, J. H.; Wasserburg, G. J. ‘‘238U-234U- A. J.; Freedman, P. A.; Jones, C. E.; Hall, C. M.; Yi, W.;
230
Th-232Th systematics and the precise measurement of Teagle, D. ‘‘Recent developments in inductively coupled
time over the past 500,000 years.’’ Earth Planet Sci. Lett. plasma magnetic sector multiple collector mass spectrome-
1987, 81, 175–192. try.’’ Int. J. Mass Spectrom. Ion Proc. 1995, 146/147, 21–
41. Eisenhauer, A.; Wasserburg, G. J.; Chen, J. H.; Bonani, G.; 33.
Collins, L. B.; Zhu, Z. R.; Wyrwoll, K. H. ‘‘Holocene sea- 56. Harper, C. T. ‘‘Geochronology: Radiometric dating of
level determination relative to the Australian continent: U/ rocks and minerals.’’ Benchmark Papers in Geology, Dow-
Th (TIMS) and 14C (AMS) dating of coral cores from the den, Hutchison and Ross: Pennsylvania, 1973.
Abrolhos Islands.’’ Earth Planet. Sci. Lett. 1993, 114, 529– 57. Harrison, T. M.; McDougall, I. ‘‘Investigations of an intru-
547. sive contact, northwest Nelson, New Zealand.’’ Geochim.
42. Esat, T. M. ‘‘Charge collection thermal ion mass spectrom- Cosmochim. Acta 1980, 44, 1985–2003.

122

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

 MASS SPECTROMETRY AND GEOCHRONOLOGY n

58. Hart, S. R. ‘‘The use of hornblendes and pyroxenes for K- shaped electromagnetic lenses for focusing beams of
Ar dating.’’ J. Geophys. Res. 1964, 66, 2995–3001. charged particles.’’ Phys. Rev. 1953, 91, 10–17.
59. Herr, W.; Hoffmeister, W.; Hirt, B.; Houtermans, F. ‘‘Ver- 77. Jull, A. T. J., Ed.; ‘‘7th Intern. Conf. on Accelerator Mass
such zur datierung von eisin meteoriten nach Re-Os meth- Spectrometry,’’ Radiocarbon 1995, 38.
ode.’’ Z. Naturforsch. 1961, 16a, 1053–1058. 78. Karlen, I.; Olsson, I. U.; Kalberg, P.; Kilicci, S. ‘‘Absolute
60. Herzog, L. F. ‘‘Mass spectrometry in geo- and cosmo- determination of the activity of two 14C dating standards.’’
chemistry: An overview.’’ Intern. J. Mass Spectrom. Ion Arkiv. Geofysik. 1966, 6, 465–471.
Phys. 1970, 4, 253–265. 79. Kober, B. ‘‘Whole grain evaporation for 207Pb/206Pb age
61. Heumann, K. G. ‘‘New developments and applications in investigations on single zircons using a double-filament
negative thermal ionization mass spectrometry.’’ Proc. ion source.’’ Contrib. Mineral. Petrol. 1986, 93, 482–490.
A.O. Nier Symposium, AECL-11342, Green, L. W., Ed.; 80. Kober, B. ‘‘Single zircon evaporation combined with Pb/
1996, 2–7. emitter bedding for 207Pb/206Pb age investigations using
62. Heumann, K. G. ‘‘Isotope dilution mass spectrometry thermal ion mass spectrometry and implications to zir-
(IDMS) of the elements.’’ Mass Spectrom. Rev. 1992, 11, conology.’’ Contrib. Mineral. Petrol. 1987, 96, 63–71.
41–67. 81. Krogh, T. E. ‘‘Improved accuracy of U-Ph zircon dating
by selection of more concordant fractions using a high
63. Hinton, R. W.; Long, J. V. ‘‘High resolution ion-micro-
gradient magnetic separation technique.’’ Geochim. Cos-
probe measurement of lead isotopes: variations within sin-
mochim. Acta. 1982a, 46, 631–635.
gle zircon from Lae Seul, north-west Ontario.’’ Earth
Planet, Sci. Lett. 1979, 45, 309–325. 82. Krogh, T. E. ‘‘Improved accuracy of U-Pb zircon ages
by the creation of more concordant systems using the air
64. Hirata, T. ‘‘Lead isotopic analyses of NIST Standard Ref-
abrasion technique.’’ Geochim. Cosmochim Acta. 1982b,
erence Materials using multiple collector inductively cou-
46, 637–649.
pled plasma mass spectrometry coupled with a modified
external correction method for the mass discrimination ef- 83. Krogh, T. E.; Corfu, F.; Davis, D. W.; Dunning, G. R.;
fect.’’ Analyst 1996, 121, 1407–1411. Heaman, L. M.; Kamo, S. L.; Machado, N. ‘‘Precise isoto-
pic ages of diabase dykes and mafic to ultra mafic rocks
65. Hohenberg, C. M. ‘‘A high sensitivity pulse counting mass using trace amounts of baddeleyite and zircon.’’ In: Mafic
spectrometer system for noble gas analysis.’’ Rev. Sci. In- Dyke Swarms. Geol. Assoc. Canada Spec. Paper, Halls,
strum. 1980, 51, 1075–1082. H. C.; Fahrig, W. F., Eds.; 1987, 34, 147–152.
66. Holmes, A. ‘‘An estimate of the age of the Earth.’’ Nature 84. Kurtz, M. D. ‘‘In-situ production of terrestrial cosmogenic
1946, 157, 680–684. helium and some applications to geochronology.’’ Geo-
67. Honig, R. E. J. Appl. Phys. 1945, 16, 646–654. chim. Cosmochim. Acta. 1986, 50, 2855–2862.
68. Houtermans, F. G. ‘‘Die isotopen-haufigkeiten im natur- 85. Layer, P. W.; Hall, C. M.; York, D. ‘‘The derivation of
40
lichen Blei and das Aalter des Urans.’’ Naturwissenschaf- Ar/39Ar age spectra of single grains of hornblende and
ten 1946, 33, 185–187. biotite by laser step-heating.’’ Geophys. Res. Lett. 1987,
69. Houtermans, F. G. ‘‘Das Alter des Urans.’’ Z. Naturforsch 14, 757–760.
1947, 29, 322–328. 86. Lipson, J. ‘‘Potassium-argon dating of sedimentary rocks.’’
Geol. Soc. Amer. Bull. 1958, 69, 137–150.
70. Hurley, P. M.; Fairbairn, H. W.; Faure, G.; Pinson, W. H.
‘‘New approaches to geochronology by strontium isotope 87. Litherland, A. E. ‘‘Ultra sensitive mass spectrometry with
variations in whole rocks.’’ Radiometric Dating 1963, accelerators.’’ Ann. Rev. Nucl. Part. Sci. 1980, 30, 437–
201–217. 473.
71. Hurst, G. S.; Payne, M. G.; Kramer, S. D.; Chen, C. H.; 88. Loss, R. D.; Rosman, K. J. R.; De Laeter, J. R. ‘‘Micro-
Phillips, R. C.; Allman, S. L.; Alton, G. D.; Dabbs, electro deposition: A sample purification and loading tech-
J. W. T.; Willis, R. D.; Lehmann, B. E. ‘‘Method for count- nique for thermal ionization mass spectrometry.’’ Int. J.
ing noble gas atoms with isotopic selectivity.’’ Rept. Prog. Mass Spectrom. Ion Proc. 1984, 57, 201–209.
Phys. 1985, 48, 1333–1370. 89. Luck, J. M.; Arndt, N. T. ‘‘Re-Os study of archaean koma-
talites from Alexo (Abitibi): the first extra-terrestrial os-
72. Inghram, M. G.; Chupka, W. A. ‘‘Surface ionization source
mium isochron.’’ B.I.G.G., Cambridge, 1985.
using multiple filaments.’’ Rev. Sci. Instrum. 1953, 24,
518–520. 90. Luck, J. M. Nuclear Methods of Dating; Roth, E.; Poty,
B., Eds.; Kluwer Academic: London, 1989, pp. 95–104.
73. Inghram, M. G. ‘‘Stable isotope dilution as an analytical
tool.’’ Ann. Rev. Nucl. Sci. 1954, 4, 81–92. 91. Lugmair, G. W.; Scheinin, N. B.; Marti, K. ‘‘Sm-Nd age
and history of Apollo 17 basalt 75075: evidence for early
74. Ireland, T. R.; Wlotzka, F. ‘‘The oldest zircons in the solar differentiation of the lunar exterior.’’ Proc. 6th Sci. Conf.
system.’’ Earth and Plan. Sci. Lett. 1992, 109, 1–10. Geochim. Cosmochim. Acta. Suppl. 1975, 6, 1419–1429.
75. Ireland, T. R. ‘‘Isotopic anomalies in extra terrestrial 92. Marshall, B. D.; De Paolo, D. J. ‘‘Precise age determina-
grains.’’ J. Roy. Soc. W. Aust. 1996, 79, 43–50. tions and petrogenetic studies using the K-Ca method.’’
76. Johnson, E. G.; Nier, A. O. ‘‘Angular aberrations in sector Geochim. Cosmochim. Acta. 1982, 46, 2537–2545.

123

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

n DE LAETER

93. Masuda, A.; Hirata, T.; Shimizu, H. ‘‘Determination of pic constitution of lead and the measurement of geological
osmium isotope ratios in iron meteorites and iridosmines age.’’ Phys. Rev. 1941, 60, 112–116.
by ICP-MS.’’ Geochem. J. 1986, 20, 233–239. 113. Nier, A. O.; Hayden, J. L. ‘‘A miniature Mattauch–Herzog
94. Matsuda, H. ‘‘Double focusing mass spectrometers of sec- mass spectrometer for the investigation of planetary atmo-
ond-order.’’ Int. J. Mass Spect. Ion Phys. 1974, 14, 219– spheres.’’ Intern. J. Mass Spectrom. Ion Phys. 1971, 6,
233. 339–346.
95. Matsuda, H.; Matsuo, T. ‘‘Influences of fringing fields on 114. Nier, A. O.; Schlutter, D. J. ‘‘High performance double-
image aberrations of double-focusing mass spectrome- focusing mass spectrometer.’’ Rev. Sci. Instrum. 1985, 56,
ters.’’ Int. J. Mass Spect. Ion Phys. 1971, 6, 385–392. 214–219.
96. McCulloch, M. T.; Mortimer, G.; Esat, T.; Xianhua, L.; 115. Nier, A. O.; Schlutter, D. J. ‘‘Extraction of helium from
Pillans, B.; Chappell, J. ‘‘High resolution windows into individual interplanetary dust particles by step-heating.’’
early Holocene climate: Sr/Ca coral records from the Huon Meteorites 1992, 27, 166–173.
Peninsula.’’ Earth Planet Sci. Lett. 1996, 138, 169–178. 116. Nishiizumi, K.; Lal, D.; Klein, J.; Middleton, R.; Arnold,
97. Merrihue, C.; Turner, G. ‘‘Potassium argon dating by acti- J. R. ‘‘Production of Be-10 and Al-26 by cosmic rays in
vation from fast neutrons.’’ J. Geophys. Res. 1966, 71, terrestrial quartz in situ and implications for erosion rates.’’
2852–2857. Nature 1986, 319, 134–136.
98. Middleton, R.; Klein, J. ‘‘26Al: measurement and applica- 117. Nittler, L. R.; Alexander, C. M. O.; Cao, X.; Walker,
tion.’’ Phil. Trans. Roy. Soc. London 1987, A323, 121– R. M.; Zinner, E. ‘‘Stellar sapphires: the properties and
143. origins of presolar Al2O3 in meteorites.’’ Astrophys. J.
1997, 483, 475.
99. Mohler, F. L. Science 1955, 122, 334–335.
118. Noble, S. R.; Lightfoot, P. C.; Scharer, U. ‘‘A method for
100. Muller, R. A. ‘‘Radioisotope dating with a cyclotron.’’ Sci- single-filament isotopic analysis of Nd using in-situ reduc-
ence 1977, 196, 489–494. tion.’’ Chem. Geol. 1989, 79, 15–19.
101. Nelson, D. A.; McCulloch, M. T. ‘‘Petrogenetic applica- 119. Patchett, P. J.; Tatsumoto, M. ‘‘Hafnium isotope variations
tions of the 40K-40Ca radiogenic decay scheme—a recon- in oceanic basalts.’’ Geophys. Res. Lett. 1980, 7, 1077–
naissance study.’’ Chem. Geol. 1989, 79, 275–293. 1080.
102. Nicolaysen, L. O. ‘‘Graphic interpretation of discordant 120. Patterson, C. C. ‘‘Age of meteorites and the Earth.’’ Geo-
age measurements on metamorphic rocks.’’ Ann. New York chim. Cosmochim. Acta 1956, 10, 230–237.
Acad. Sci. 1961, 91, 198–206. 121. Pavich, M. J.; Brown, L.; Valette–Silver, J. N.; Klein, J.;
103. Nier, A. O. ‘‘Evidence for the existence of an isotope of Middleton, R. ‘‘10Be analysis of a Quaternary weathering
potassium of mass 40.’’ Phys. Rev. 1935, 48, 283–284. profile in the Virginia Piedmont.’’ Geology 1985, 13, 39–
104. Nier, A. O. ‘‘Variations in the relative abundances of the 41.
isotopes of common lead from various sources.’’ J. Amer. 122. Pillinger, C. T. ‘‘Light element stable isotopes in meteor-
Chem. Soc. 1938, 60, 1571–1576. ites—from grams to picograms.’’ Geochim. Cosmochim.
105. Nier, A. O. ‘‘A mass spectrometer for routine isotope abun- Acta. 1984, 48, 2739–2766.
dance measurements.’’ Rev. Sci. Instrum. 1940, 11, 212– 123. Pillinger, C. T. ‘‘New technologies for small sample stable
216. isotope measurement: static vacuum gas source mass spec-
106. Nier, A. O.; Thompson, R. W.; Murphy, B. F. ‘‘The isoto- trometry, laser probes, ion probes and gas chromatography
pic constitution of lead and the measurement of geological isotope ratio mass spectrometry.’’ Int. J. Spectrom. Ion.
time.’’ Phys. Rev. 1941, 60, 112–116. Proc. 1992, 118, 477–501.
107. Nier, A. O. ‘‘A mass spectrometer for isotope and gas 124. Raisbeck, G. M.; Yiou, F.; Fruneau, M.; Loiseaux, J. M.;
analysis.’’ Rev. Sci. Instrum. 1947, 18, 398–411. Lieuvin, M.; Ray Lorius, C. ‘‘Cosmogenic 10Be concentra-
tions in Antarctic ice during the past 30,000 years.’’ Nature
108. Nier, A. O. ‘‘Some reminiscences of mass spectrometry 1981, 292, 825–826.
and the Manhattan Project.’’ J. Chem. Ed. 1989, 66, 385–
125. Raisbeck, G. M.; Yiou, F. ‘‘Dating by cosmogenic isotopes
388. 10
Be, 26Al and 41Ca’’, in Nuclear Methods of Dating, Roth,
109. Nier, A. O.; Ney, E. P.; Inghram, M. G. ‘‘A null method E.; Poty, B., Eds.; Kluwer Academic: London, 1989, pp.
for the comparison of two ion currents in a mass spectrom- 353–378.
eter.’’ Rev. Sci. Instrum. 1947, 18, 294–297.
126. Reynolds, J. H. ‘‘High sensitivity mass spectrometer for
110. Nier, A. O. ‘‘A redetermination of the relative abundances noble gas analysis.’’ Rev. Sci. Instrum. 1956, 27, 928–934.
of the isotopes of carbon, nitrogen, oxygen, argon and 127. Reynolds, J. H. ‘‘Comparative study of argon content and
potassium.’’ Phys. Rev. 1950, 77, 789–793. argon diffusion and feldspar.’’ Geochim. Cosmochim. Acta.
111. Nier, A. O.; Booth, E. T.; Dunning, J. R.; Grosse, A. V. 1957, 12, 177–184.
‘‘Nuclear fission of separated uranium isotopes.’’ Phys. 128. Rosman, K. J. R.; Loss, R. D.; De Laeter, J. R. ‘‘The isoto-
Rev. 1940, 57, 546. pic composition of tin.’’ Int. J. Mass. Spectrom. Ion Pro-
112. Nier, A. O.; Thompson, R. W.; Murphy, B. F. ‘‘The isoto- cesses 1984, 56, 281–291.

124

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337

 MASS SPECTROMETRY AND GEOCHRONOLOGY n

129. Rosman, K. J. R.; Lycke, W.; Damen, R.; Werz, R.; Hen- of a set of synthetic uranium isotope mixtures.’’ J. Analyt.
driekx, F.; Traas, L.; De Bièvre, P. ‘‘The preparation and At. Spect. 1995, 10, 395–398.
use of synthetic isotope mixtures for testing mass spec- 143. Thirlwell, M. F. ‘‘High precision multi-collector isotopic
trometers.’’ Int. J. Mass Spectrom. Ion Proc. 1987, 79, 61– analysis of low levels of Nd as oxide.’’ Chem. Geol. 1991,
71. 94, 13–22.
130. Russ, G. P. I.; Bazan, J. M. ‘‘Isotope ratio measurements 144. Tilton, G. R.; Patterson, C.; Brown, H.; Inghram, M.; Hay-
with an inductively coupled plasma source mass spectrom- den, R.; Hess, D.; Larsen, E. ‘‘Isotopic composition and
eter.’’ Spectrochim. Acta. Part B 1987, 42, 49–62. distribution of lead, uranium and thorium in a Precambrian
131. Russell, R. D. ‘‘The systematics of double spiking.’’ J. granite.’’ Geol. Soc. Amer. Bull. 1955, 66, 1131–1134;
Geophys. Res. 1971, 76, 4949–4955. 1139–1148.
132. Russell, W. A.; Papanastassiou, D. A.; Tombrello, T. A. 145. Tyayi, A. K.; Dash, S.; Rajagopalan, S. Secondary Ion
‘‘Ca fractionation on the Earth and other solar system mate- Mass Spectrometry: Application in Materials Science, Ag-
rials.’’ Geochim. Cosmochim. Acta. 1978, 42, 1075–1090. garwal, S. K.; Jain, H. C., Eds.; ‘‘Introduction to Mass
133. Rutherford, E. Radioactive Transformations, Yale Univer- Spectrometry.’’ Perfect Prints, 1997, 303–323.
sity Press: New Haven, 1906. 146. Walder, A. J.; Freedman, P. A. ‘‘Isotopic ratio measure-
ments using a double focusing magnetic sector mass ana-
134. Rutherford, E. ‘‘Origin of actinium and the age of the
lyser with an inductively coupled plasma as an ion source.’’
earth.’’ Nature 1929, 123, 313–314.
J. Anal. Atomic. Spect. 1992, 7, 571–575.
135. Scharer, U.; Allegre, C. J. ‘‘Uranium lead system in frag-
147. Walker, R. L.; Carter, J. H.; Smith, D. H. ‘‘A bulk resin
ments of a single zircon grain.’’ Nature 1982, 295, 585–
bead procedure to obtain uranium and plutonium from ra-
587.
dioactive solutions for mass spectrometric analyses.’’ Anal.
136. Schoeller, D. A.; Hayes, J. M. ‘‘Computer controlled ion Lett. 1981, 14, 1603–1612.
counting isotope ratio mass spectrometer.’’ Anal. Chem.
148. Wendt, J. L.; Wendt, I.; Tuttas, D. ‘‘Determination of U-
1975, 47, 408–414.
Pb ages of zircons by direct measurement of the 210Pb/206Pb
137. Schreiner, G. D. L. ‘‘Comparison of the 87Rb/88Sr age of ratio.’’ Chem. Geol. 1993, 106, 467–474.
the red granite of the Bushveld complex from measure- 149. Wasserburg, G. J.; Papanastassiou, D. A.; Nenow, E. V.;
ments on the total rock and separated mineral fractions.’’ Baum, C. A. ‘‘A programmable magnetic field mass spec-
Proc. Royal Soc. Ser. A 1958, 245, 112–117. trometer with on-line data processing.’’ Rev. Sci. Instrum.
138. Simons, D. S. ‘‘Isotopic analysis with the laser microprobe 1969, 40, 288–295.
mass analyser.’’ Int. J. Mass Spectrom. Ion Proc. 1983, 150. Webster, R. K. In ‘‘Methods in Geochemistry,’’ Smales,
55, 15–30. A. A.; Wager, L. R., Eds.; Interscience: New York, 1960,
139. Smith, D. H.; Donahue, D. L.; Young, J. P. ‘‘Determina- pp. 202–246.
tion of isotopic ratio of neodymium by resonance ioniza- 151. Wetherill, G. W. ‘‘An interpretation of the Rhodesia and
tion mass spectrometry.’’ Int. J. Mass Spectrom. Ion Proc. Witwatersrand age patterns.’’ Geochim. Cosmochim. Acta.
1985, 65, 289–297. 1955, 9, 290–292.
140. Strassman, F.; Walling, E. ‘‘The possibility of determining 152. Wetherill, G. W.; Aldrich, L. T.; Davis, G. L. ‘‘40Ar/40K
the geological age of rubidium-containing minerals and ratios of feldspars and micas from the same rock.’’ Geo-
rocks.’’ Z. Anorg. Allgem. Chem. 1938, 236, 78–82. chim. Cosmochim. Acta. 1955, 8, 171–172.
141. Svec, H. J. ‘‘Mass spectroscopy—ways and means. A his- 153. York, D.; Hall, C. M.; Yanase, Y.; Hanes, J. A.; Kenyon,
torical prospectus.’’ Intern. J. Mass. Spectrom. Ion Proc. W. J. ‘‘40Ar/39Ar dating of terrestrial minerals with a con-
1985, 66, 3–29. tinuous laser.’’ Geophys. Res. Lett. 1981, 8, 1136–1138.
142. Taylor, P. D.; De Bièvre, P.; Walder, A. J.; Entwistle, A. 154. Zeitler, P. K.; Sutter, J. F.; Williams, I. S.; Zartman, R.;
‘‘Validation of the analytical linearity and mass discrimina- Tahirkheli, R. A. ‘‘Geochronology and temperature history
tion correction model exhibited by a multiple collector of the Nanga Parbat—Haramost Massif, Pakistan.’’ Geol.
inductively coupled plasma mass spectrometer by means Soc. of America 1989, 23, 1–22.

125

/ 821A$$0337 01-23-99 11:01:18 msra W: MSR 337