1 The Noble Gases as Geochemical Tracers: history and background Pete Burnard 1 , Laurent Zimmermann 1 and Yuji Sano 2 1. Centre de Recherches Pétrographiques et Géochimiques, BP20, Vandoeuvre- lès-Nancy Cedex, France 2. Center for Advanced Marine Research, Ocean Research Institute, The University of Tokyo, Nakano, Tokyo 164–8639, Japan Abstract This chapter describes the discovery of the noble gases and the development of the first instrumentation used for noble gas isotopic analysis before outlining in very general terms how noble gases are analysed in most modern laboratories. Most modern mass spectrometers use electron impact sources and magnetic sector mass filters with detection by faraday cups and electron multipliers. Some of the performance characteristics typical of these instruments are described (sensitivity, mass discrimination). Extraction of noble gases from geological samples is for the most part achieved by phase separation, by thermal extraction (furnace) or by crushing in vacuo. The extracted gases need to be purified and separated by a combination of chemical and physical methods. The principles behind different approaches to calibrating the mass spectrometers are discussed.
29
Embed
The Noble Gases as Geochemical Tracers: history and backgroundrecherche.crpg.cnrs-nancy.fr/IMG/pdf/Chapter_1... · Clarke-type instrument. Figure 1 shows the technical progress of
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
The Noble Gases as Geochemical Tracers: history and background
Pete Burnard1, Laurent Zimmermann1 and Yuji Sano2
1. Centre de Recherches Pétrographiques et Géochimiques, BP20, Vandoeuvre-lès-Nancy Cedex, France 2. Center for Advanced Marine Research, Ocean Research Institute, The University of Tokyo, Nakano, Tokyo 164–8639, Japan
Abstract
This chapter describes the discovery of the noble gases and the development of the first
instrumentation used for noble gas isotopic analysis before outlining in very general terms how
noble gases are analysed in most modern laboratories. Most modern mass spectrometers use
electron impact sources and magnetic sector mass filters with detection by faraday cups and
electron multipliers. Some of the performance characteristics typical of these instruments are
described (sensitivity, mass discrimination). Extraction of noble gases from geological samples is
for the most part achieved by phase separation, by thermal extraction (furnace) or by crushing in
vacuo. The extracted gases need to be purified and separated by a combination of chemical and
physical methods. The principles behind different approaches to calibrating the mass
spectrometers are discussed.
2
Introduction
Occupying the last column of the Periodic Table and therefore characterized by a
filled outer valence shell, the noble gases do not form chemical compounds at
conditions relevant to natural processes on Earth: while Kr and Xe do have an
extensive chemistry (Grochala 2007), this is typically only with highly
electronegative elements such as F and O, which in natural terrestrial systems are
already bound to more reactive elements than the noble gases. The noble gases
and their isotopes are able to provide unique constraints on certain geological
processes owing to their inert behavior and because numerous nuclear reactions
are recorded in the noble gas isotopic compositions. While many of these
applications are related to constraining the duration of geologic events
(thermochronology, absolute dating, surface exposure dating etc), this volume
aims to describe sampling strategies, analytical methodologies and data
interpretation for the use of noble gases as tracers in the Earth Sciences: tracers of
geological fluids, of mantle circulation, of seawater circulation, of dust falling
from space, of hydrocarbon reserves, of past climate change and of potential
future CO2 storage in crustal reservoirs.
History
The discovery of the noble gases was initially by inference rather than isolation.
Helium was first identified as an element simply from its absorption spectrum in
the solar chromosphere (by Lockyer in 1868; it would be another 27 years before
it was actually isolated as an element). Similarly, the presence of an unknown gas
of low density (which later turned out to be Ar) in the atmosphere was proposed
by Rayleigh in 1895 because nitrogen from distilled air had a different density to
chemically produced nitrogen. This observation lead to more refined distillation
experiments by Ramsey (1898) which positively identified Ar as an element and
also led to the isolation of Kr, Ne and Xe (for which Ramsey and Rayleigh would
receive Noble Prizes).
The discovery that an alpha particle produced by decay of U was in fact a helium
nucleus (by accumulating α particles in an evacuated tube by recoil through a thin
window) allowed Rutherford and Strutt to make the first radiometric age
measurements in 1905: by heating various U-bearing minerals and weighing the
3
He that was produced, they were able to constrain the age of the earth to be
greater than 400 Ma. Rutherford and Strutt were well aware of the problem of
potential He loss from minerals on geological timescales and so correctly
considered the 400 Ma estimate to be a minimum estimate; but even so, this was a
major advance as previous estimates considered that the maximum age of the
Earth was of the order 20 Ma. However, as a result of the He loss problem the
U+Th/He radiochronometer was largely ignored until the beginning of the 1990s
when a better understanding of the kinetics of He loss in certain U bearing
minerals ushered in the possibility of quantitative U+Th/He thermochronometry
(Farley 2002).
Noble gases were central to understanding the structure of the atom, and the first
isotopic separation – using the first mass spectrometer – was made by JJ Thomson
and F.W. Aston in 1913 to separate 20Ne from 22Ne. This development opened the
floodgates for research into isotopes, and ten years later multiple isotopes of all
the noble gases except for 3He had been identified (despite a hiatus forced by the
First World War), for the most part by Dempster at the University of Chicago:
Dempster's 1918 mass spectrometer (Dempster 1918) laid down the blueprint for
the magnetic field filters used in mass spectrometers to this day. Helium-3 was
only identified in 1939 by cyclotron mass spectrometry at Lawrence Berkely Nat.
Labs, CA, by Alvarez and Cornog (Alvarez and Cornog 1939). who used the 60-
inch cyclotron as a mass spectrograph to show that 3He is a stable isotopic
constituent of natural helium. The 3He2+ ions were detected using the nuclear
reaction 28Si(3He,p)30P(β)30Si and the radio-activity of 30P quantified by a Geiger
counter. The activity of pure atmospheric helium was twelve times higher than
that of tank helium derived from a natural gas well. They concluded that the 3He/4He ratio of air helium is about ten times greater than that of natural gas
helium.
The advances in mass spectrometry resulting from the Manhattan Project (1942-
1946) opened up opportunities in geo- and cosmochemistry in the post war years.
At the forefront of mass spec development during the Manhattan project, Nier,
now in Minnesota, returned to the geochronology and geochemistry he started in
the pre-war years as a grad student at Harvard. Seminal papers on K-Ar dating
(Aldrich and Nier 1948a) and on the He isotopic composition of the atmosphere
(Daunt, et al. 1947) resulted from the impressive 15cm radius, 60° mass
4
spectrometers and electron impact ion sources that Nier built at that time. These
machines were capable of separating 3He+ from HD+ and HHH+ interferences.
Nier and his co-workers discovered large variations in 3He/4He ratios of natural
materials: Aldrich and Nier presciently stated in their 1948 paper (Aldrich and
Nier 1948b) that “The present study can hardly be regarded as more than a
preliminary exploration of a new and fascinating field of investigation. It is
apparent that a far more comprehensive and systematic study will be required to
definitely establish the natural source of 3He and 4He.”
Noble gas mass spectrometers owe significantly to Nier’s legacy; indeed, many
modern mass specs still employ a Nier-type ion source (see Section Noble Gas
Mass Spectroscopy, below). Further advances in noble gas mass spectroscopy in
the 65 years or so since Nier’s inventions have been gradual rather than leaps in
performance. Two exceptions are Reynold’s (at Berkeley) development of static
(i.e. isolated from pumps) mass spectrometry in 1956 (Reynolds 1956) thereby
significantly increasing sensitivity, and the development of a new electron impact
source by Baur and Signer in the late 1970s (Baur 1980) which is characterized by
better pressure linearity.
Clarke et al. (Clarke, et al. 1969) measured the 3He/4He ratios of deep Pacific
water by using a 10-inch radius, low volume, and static vacuum mass
spectrometer, with an electron multiplier for ion detection, reducing samples sizes
by three orders of magnitude compared to Aldrich and Nier (Aldrich and Nier
1948a). Two examples of Clarke mass spectrometers remain operational to this
day (Scripps Institution of Oceanography and Woods Hole Institution of
Oceanography (e.g., Jenkins (Jenkins 1987) , Lupton (Lupton, et al. 1990)). In
1976, Takaoka and co-workers developed a noble gas mass spectrometer to
measure all noble gas isotopes, which was commercialized in 1982 as the VG
MM-3000 (Hooker, et al. 1985). At the same time Sano et al. (Sano, et al. 1982)
used a 6-inch static mass spectrometer (6-60-SGA, Nuclide Co.) for terrestrial He
isotope measurements, reporting an atmospheric 3He/4He ratio of (1.43±0.03)x10-
6, which agreed well with the critical value of (1.399±0.013)x10-6 determined by
Mamyrin et al. (Mamyrin, et al. 1970) (see below).
5
In the ex-Soviet Union, Mamyrin et al. (Mamyrin, et al. 1969)
independently developed the “Reflectron” time of flight mass spectrometer with
m/ΔM ~ 2000 and even more sensitive than the latest magnetic sector mass
spectrometers. Until the mid- 1980s, there were only seven laboratories in the
world routinely measuring terrestrial He isotopes: Ontario, St. Petersburg, Boston,
San Diego, Tokyo, Osaka and Cambridge.
In 1986, the VG-5400, a successor of the MM-3000 was introduced and
soon after the MAP 215-50 became available. These machines have since
produced the majority of terrestrial noble gas analyses in many laboratories
around the world (e.g., Sano and Wakita (Sano and Wakita 1985)). However,
further development in noble gas mass spectrometry ceased for almost twenty
years following the success of the VG5400 and MAP215 instruments.
In 2004 two brand-new multicollector noble gas mass spectrometers
(HELIX-SFT, GV instruments and Noblesse, Nu Instruments) were announced,
with the possibility of precisely measuring the atmospheric 3He/4He ratio on
samples size of 1x10-7 ccSTP, three order of magnitude smaller than the original
Clarke-type instrument. Figure 1 shows the technical progress of helium isotope
measurements with time. Taking into account only mass spectrometry
(MS1~MS6) developments, the sample size required for helium isotope analysis
as well as the precision of the analyses have improved on an approximately
logarithmic scale with time since their invention. Recently, high sensitivity
(sample sizes of 10-7 ccSTP) and high precision (estimated error of 0.1% at 1σ)
helium isotope measurements have been reported by Sano et al. (Sano, et al.
2008).
With the advent of high resolution, multicollector noble gas mass
spectrometry, the precision and sensitivity for the remaining noble gases – which
all have more than two isotopes – have also increased. It is important to note that
all noble gases have at least one (and usually more) non-abundant isotopes
requiring multicollection using secondary electron multipliers (SEM) and
6
sometimes requiring multiple SEMs. Multiple SEM machines are considerably
more complex to construct and operate than faraday-only multicollector mass
spectrometers.
One of the major issues for Ne isotope analyses are the isobaric
interferences on 20Ne+ by 40Ar+ and H218O+ and on 22Ne+ by CO2
++ (see Table 1),
requiring high resolving power mass filters in order to cleanly separate the
signals. This is now possible for 20Ne+ on both Nu and GV (now Thermofisher)
mass spectrometers, producing data with 20Ne/22Ne and 21Ne/22Ne uncertainties that
have decreased by factor 2 – 5 compared to single collector data (Colin, et al. 2011;
Marrocchi, et al. 2009; Parai, et al. 2009).
Multicollection also provides major advantages for the analysis of Kr and
Xe isotopes, as both noble gases have numerous isotopes (6 and 9 stable isotopes,
respectively) and their atoms have short half lives in the mass spectrometer due to
their high ionization cross-sections compared to the lighter noble gases. These
multicollector (>2 collector) machines have been routinely producing data for two
or three years at the time of writing (Holland and Ballentine 2006; Holland, et al.
2009); from this early work it is clear that the precision of noble gas analyses that
is now possible will open up new opportunities and directions in the isotope
geochemistry of the noble gases.
Figure 1. Technical advances in helium isotope measurements with time. Top
and bottom show the decrease in sample size required to measure He isotopes and
the improvement in analytical precision, respectively. MS1: Aldrich and Nier
(1948); MS2: Clarke et al. (1969); MS3: Mamyrin et al. (1969); MS4: Nagao et
al. (1981); MS5: Sano and Wakita (1988); MS6: Lupton (1990); MS7: Sano et al.
(2008); NAA: Coon (1949); LR: Wang et al. (2003).
Obviously, improvements in materials, vacuum techniques, multiple
collectors and ion optics have contributed to the high standards expected of
commercial mass spectrometers today, but the for the most part modern mass
spectrometers closely resemble those of Nier and Reynolds (indeed, some are still
operational with excellent data still produced by glass ‘Reynolds’ machines).
Some individual labs have made technical advances that have very specific
applications. Wang et al. (Wang, et al. 2003) reported that a sensitive laser
spectroscopic method could be been applied to the quantitative determination of 3He/4He for ratios in the range 10-7 ~ 10-5. Resonant absorption of 1083 nm laser
light photons by metastable 3He atoms in a discharge cell was measured by
frequency modulation saturation spectroscopy while the abundance of 4He was
7
measured by a direct absorption technique. Even though the results on three
different samples extracted from the atmosphere and commercial helium gas were
in good agreement with values obtained by mass spectrometry, a large amount of
sample helium, about 0.2 ccSTP was required for analysis. This is far beyond the
terrestrial helium sample size for practical use. Resonance Ionization Mass
Spectrometry has also been developed in some noble gas laboratories which
increases precision and reduces detection limits for Xe and Kr isotope
determination (Crowther, et al. 2008; Gilmour, et al. 1994; Iwata, et al. 2010;
Mamyrin, Tolstikh.In, Anufriye.Gs and Kamenski.Il 1969; Strashnov, et al. 2011;
Thonnard 1995). The turbo compressor source developed by H. Baur of ETH
increases the partial pressure of (noble) gases in the source by using a molecular
drag pump within the mass spectrometer volume, thereby increasing sensitivity
for He and Ne by a factor ~100 (Baur 1999).
Conventions
Quantities of noble gases are reported either in moles of gas or as cm3 STP
(Standard Temperature and Pressure); it should be noted that although STP is
usually defined as 273.15 K and 101.325 kPa (= 1 bar), some definitions use
100.000 kPa (McNaught and Wilkinson 1997). Use of moles should be
encouraged. As a consequence, noble gas concentrations are usually expressed as
moles (or cm3 STP) per gram of rock or fluid. Sometimes, although rarely, these
are converted to percent or ppm of the noble gas by weight, as is common for
other elements in many branches of geochemistry. However, performing the
simple conversion demonstrates how ‘rare’ the noble gases are: a typical olivine
crystal might contain ~ 10-13 moles 4He g-1, or about 0.5 ppt 4He. The abundances
of 3He or the rare isotopes of Xe are typically >106 times lower.
There is no single convention for reporting isotope ratios. Most commonly, the
actual ratio is reported; this is practical where isotopic variations are relatively
large. Where small isotopic variations are involved, the ‘delta’ notation is
commonly used where:
( )( ) 10001
tan
•⎥⎥⎥
⎦
⎤
⎢⎢⎢
⎣
⎡−=
dardsr
isample
ri
X
XX
XX
iδ
8
Where iX is the ‘anomalous’ isotope and rX is the reference isotope. For all
systems with the exception of He, it is usual to place a non-radiogenic isotope as
the denominator. Thus δ40Ar would refer to the permil variations in the 40Ar/36Ar
ratio.
Because the non-radiogenic He isotope (3He) is the least abundant isotope,
historically He isotope ratios have been reported as 3He/4He ratios (i.e. radiogenic
isotope as the denominator). Furthermore, it is common to report 3He/4He ratios
normalized to the atmospheric 3He/4He ratio, Ra. Thus a sample with a 3He/4He
ratio of 1 Ra would have real ratio of 1.39 x10-6 (see chapter 2, this volume).
However, when plotting He isotopes against other isotope ratios (which are
ubiquitously quoted with the more traditional non-radiogenic/radiogenic notation),
this often leads to unnecessarily complex, inverse relationships. In these instances,
it is preferable to use 4He/3He (not normalized to Ra).
Relative noble gas abundances are frequently expressed normalized to 36Ar and
the ratio normalized to that of air, commonly known as the F-value notation:
( ) ( )( )air
sample
AriAri
iF 36
36
//
=
Thus F(80Kr) = 2 corresponds to a 80Kr/36Ar ratio twice the ratio of air.
Noble Gas Mass spectrometry
Principles
As with all mass spectrometers, noble gas mass spectrometers consist of an ion
source, a mass filter and collector(s). The “standard” configuration which
accounts for >95% of operational mass spectrometers at the time of writing
consists of an electron impact source, a magnetic sector filter and detection using
a combination of Faraday cups and secondary electron multipliers (SEMs). The
magnetic field is measured by one or more Hall probes. Magnet control and data
collection are usually done via proprietary software; on the modern generation of
mass spectrometers (post 2002), control of source and multiplier voltages are also
performed via a software interface.
The most common “non-standard” mass spectrometer solution involves ionisation
using resonant lasers and a time of flight mass filter with detection by
microchannel plates (Gilmour, Lyon, Johnston and Turner 1994) (Lavielle, et al.
9
2006). Alternatively, atom trap or penning trap noble gas mass spectrometers also
exist (Lu and Mueller 2010), (Neidherr, et al. 2009). These configurations are
adapted to specific analytical problems and are not “off the shelf” solutions,
therefore will not be dealt with here: the reader is instead referred to the
publications above.
Most, but not all, noble gas mass spectrometers are operated in “static” mode, that
is to say all the gas available is admitted into the mass spectrometer with the
pumps isolated (Reynolds 1956) (by contrast, “dynamic” mode is where the
sample gas is slowly bled into the mass spectrometer while simultaneously
pumping the spectrometer). “Static” mode considerably increases mass
spectrometer sensitivity with the disadvantage that the signal changes over time as
the gas is gradually consumed by ionisation. Usually this is accounted for by
performing an extrapolation to the instant the gas was diluted into the mass
spectrometer (Figure 2). An additional consequence of operating in “static” mode
is that very low mass spectrometer backgrounds are required because, as soon as
the pumps are isolated, background levels in the spectrometer envelope will begin
to rise. This is countered by including one or more getters in the isolated volume
and by rigorous vacuum techniques: noble gas mass spectrometers are exposed to
air very infrequently (usually several years between ventings) and are thoroughly
baked at temperatures up to 350°C after being at atmospheric pressure.
FIGURE 2: Signal of 40Ar vs. time in a VG5400 mass spectrometer (200 µA trap current). ≈ 200
seconds are necessary to dilute the gas into the mass spectrometer and to determine field positions
for each argon isotope. The signals were extrapolated to the instant the gas was diluted into the
mass spectrometer (t=0) in order to account for consumption by ionisation of the gas during
measurement.
10
Electron bombardment sources
While magnet and collector configurations are relatively straightforward
and have much in common with non-noble gas mass spectrometers, good mass
spectroscopic analyses require a thorough understanding of how a noble gas
source functions. All commercial noble gas mass spectrometers past and present
use either modified Nier-type sources (Figure 3) or Baur-Signer sources, both of
which depend on a filament to create an electron beam which then ionize gas
molecules in an ionization chamber by electron bombardment (where an energetic
electron colliding with a neutral gas atom or molecule dislodges an electron from
the neutral species, thereby creating a positive ion). An ion repeller behind the
electron beam and a series of lenses extract the ions from the source and focus
these into a well-collimated ion beam.
FIGURE 3: Schematic of a Nier-type source. A tungsten filament operating at a current of around
2.5 A emits a beam of electrons which passes through the ionisation volume and is measured by
the trap. The electron current received by the trap, H, is usually of the order 10 – 800 μA; the
filament current is regulated in order that H is kept constant. A magnetic field (B) in the source
increases the electron path length and thus increases sensitivity. The ions are extracted by
adjusting the voltages on the repeller and drawout plates.
The ionisation chamber of Nier-type sources are placed within a magnetic
field (usually created by small external permanent magnets) which increases the
path length of the electrons in the ionisation region, thereby increasing the 11
sensitivity of the source but with the disadvantage that this induces mass
discrimination effects in the source itself. Baur-Signer sources have no magnetic
field within the source relying instead on a circular filament to provide a focussed,
conical electron beam. These sources therefore have considerably less mass
discrimination but also typically have lower sensitivity than Nier-type sources.
The efficiency of the source increases with mass (Figure 4) for a given source
setting.
FIGURE 4: Sensitivity vs. mass of gas analysed (4He, 20Ne, 36Ar, 84Kr & 130Xe). For the same
source settings, sensitivity increases with the mass of the gas analysed and depends only on the
effective ionisation cross section of the target atom (Botter and Bouchoux 1995)
The intensity of the electron beam is monitored within the source (the
“trap current”, usually between 10 and 800 μA) and the filament current
controlled so as to keep electron beam intensity constant. Intensity of the electron
beam has a strong effect on both sensitivity and mass discrimination (see section
Pressure dependent effects below) of the mass spectrometer (Figure 5), therefore
it is important to ensure the electron beam intensity does not fluctuate.
12
FIGURE 5: 40Ar signal as a function of trap current, VG5400. A constant pressure of argon was
introduced in the mass spectrometer for each measurement. For all other source settings kept
constant, the signal increases approximately linearly with the trap current.
Pressure dependent effects
The response of a noble gas ion source depends on the pressure within the source
itself (Burnard and Farley 2000) {Mabry, 2012 #1738}; as a result, the sensitivity
of the mass spectrometer is not linear with pressure, and mass discrimination also
depends on the quantity of gas being analyzed (i.e. the pressure in the source).
These effects are considerably less important for Bauer-Signer type sources than
for modified Nier-type sources (see Figures 6 and 7). As a result, it is important
that pure noble gases are admitted into the mass spectrometer preferably one
noble gas at a time, and that samples and standards are similar in purity, pressure
and composition (see section Purification). It is thought that these pressure
dependent effects result from the distribution of space charges within the
ionization volume (Burnard and Farley 2000), (Baur 1980). Space charges result
from passing an electron beam through the insulated ionization volume. At small
ion currents (low source pressures), the ion chamber is essentially well isolated
and large space charge potentials can exist. However, increasing the density of
ions in the source (high source pressures), these space charges can be redistributed
by the ions themselves. Hence the pressure-dependent effects are more significant
at high electron beams (high trap currents; Figure 6). Pressure dependent effects
are most significant for He and can be > 5% for typical sample sizes for Nier-type
13
14
sources (Sano, Tokutake and Takahata 2008): mass discrimination will be greater
for He than for the remaining noble gases due to the large mass difference
between the two He isotopes, and, in addition, He requires higher electron beam
currents (higher trap currents) than the heavier noble gases in order to compensate
for the lower He ionization efficiency (Figure 4), and increasing trap current is a
convenient way of increasing He sensitivity.
FIGURE 6: Non-linearity of a Nier-type source. a) He isotope discrimination and He sensitivity as
a function of pressure (≡4He mol) for constant source settings (trap current = 400 μA). b) 40Ar/36Ar
discrimination as a function of trap current and pressure (≡40Ar signal). Coulié (Coulie 2001)
similarly observes an increase in mass discrimination for Ar, up to 8%, as a function of the trap
current.
15
FIGURE 7: Non-linearity of Ar sensitivity of a Baur-Signer source (from (Baur 1980)). Measured
(shaded areas) and simulated (data points) relative Ar-sensitivity of a GSVII (GS98) Baur-Signer
ion source versus equivalent Ar pressure (equivalent Ar pressures were obtained by measuring the
ion current of 40Ar during stepwise admission of known amounts of He to the static mass
spectrometer; the measured 4He ion current was multiplied by √(4He/40Ar) to get an equivalent 40Ar ion current and a corresponding equivalent Ar pressure that would cause the same space
charge). The measurements were performed with two different α-slit settings, 0.063 rad for the
upper curve and circular symbols, 0.038 rad for the lower curve and square symbols. The
maximum sensitivities (=100%) for the two slits are 1.1 and 1.0 mA torr-1 respectively. The
measurements were made at 2.5 x10-4 A total filament emission and 100eV electron energy.
Isobaric interferences
Despite thorough baking and the presence of getters in the mass spectrometer,
there are frequently background atoms or molecules in the region of the noble gas
peaks. Common interferences that are known in noble gas mass spectrometers are
given in Table 1 (an exhaustive list is not possible, particularly at high masses
where many different molecules may exist).
Species Mass/Charge Noble gas
Resolution required1
comment
1H2D+ 3.0215 3He 509 Present in all mass spectrometers
3H+ 3.0161 3He 150 000 1H1H1H+ 3.0238 3He 384 Present in all mass