Construction of a Newly Designed Small-Size Mass Spectrometer for Helium Isotope Analysis: Toward the Continuous Monitoring of ³He/⁴He Ratios in Natural Fluids

Abstract

The construction of a small-size, magnetic sector, single focusing mass spectrometer (He-MS) for the continuous, on-site monitoring of He isotope ratios (³He/⁴He) is described. The instrument is capable of measuring ⁴He/²⁰Ne ratios dissolved in several different types of natural fluids of geochemical interest, such as groundwater and gas from hot springs, volcanoes and gas well fields. The ion optics of He-MS was designed using an ion trajectory simulation program “TRIO,” which permits the simultaneous measurement of ³He and ⁴He with a double collector system under a mass resolution power (M/ΔM) of >700. The presently attained specifications of He-MS are; (1) a mass resolving power of ca. 430, sufficient to separate ³He⁺ from interfering ions, HD⁺ and H₃⁺, (2) ultra-high vacuum conditions down to 3×10⁻⁸ Pa, and (3) a sufficiently high sensitivity to permit amounts of ³He to be detected at levels as small as 10⁻¹³ cm³ STP (3×10⁶ atoms). Long term stability for ³He/⁴He analysis was examined by measuring the ³He/⁴He standard gas (HESJ) and atmospheric He, resulting in ∼3% reproducibility and ≤5% experimental error for various amounts of atmospheric He from 0.3 to 2.3×10⁻⁶ cm³ STP introduced into the instrument. A dynamic range of measurable ³He/⁴He ratios with He-MS is greater than 10³ which was determined by measuring various types of natural fluid samples from continental gas (with a low ³He/⁴He ratio down to 2×10⁻⁸) to volcanic gas (with a high ³He/⁴He ratio up to 3×10⁻⁵). The accuracy and precision of ³He/⁴He and ⁴He/²⁰Ne ratios were evaluated by comparing the values with those measured using well established noble gas mass spectrometers (modified VG5400/MS-III and -IV) in our laboratory, and were found to be in good agreement within analytical errors. Usefulness of the selective extraction of He from water/gas using a high permeability of He through a silica glass wall at high temperature (700°C) is demonstrated.

Introduction

Helium in several geochemical reservoirs of the Earth has characteristic isotopic ratios, i.e., ³He/⁴He=1.4×10⁻⁶ in air, 1.1×10⁻⁵ in the upper mantle, and <10⁻⁷ in the crust, resulting in a wide range of ³He/⁴He ratios of more than three orders of magnitude in geochemical/geological samples.^1,2) The diffusivity of helium is very high compared to other geochemical tracers because of its low molecular weight and chemical inertness. Moreover He can escape from the Earth’s atmosphere to outer space because of its high velocity in the upper atmosphere, resulting in low concentrations of He in the atmosphere (5.2 ppm). These features make the ³He/⁴He ratio a very sensitive tracer for the transportation of volatiles and identification of their source reservoirs. Therefore, if we continuously monitor the isotopic ratio of He in fluids derived from the interior of the Earth, it would be possible to detect a change of ³He/⁴He ratio resulting from He release with various ³He/⁴He ratios from different reservoirs as the result of seismic or volcanic activities occurring in the crust and/or mantle. For example, variations in ³He/⁴He ratios in groundwater samples have been observed after earthquakes, which were considered to be the result of changes in groundwater flow systems due to the destruction of brittle, crustal rocks associated with earthquakes.^3–5) Sano et al.⁶⁾ reported post-eruptive increase in ³He/⁴He ratios in steam wells from volcanic islands, possibly resulting from additional magmatic He discharges from activated magmas. Predictive signals for earthquakes/volcanic eruptions have been observed as geophysical and geochemical signals such as ground deformation, seismic swarm, changes in the water level of observation wells, and radon emissions, during continuous monitoring with a time resolution of less than an hour or a day.^7–9) On the contrary, sampling intervals in previous He studies were more than several days^4,6) or even years.^3,5) This is due to the complicated analytical procedures for determining He using an ultra-high vacuum system after transporting the samples to the laboratory. Such procedures include the purification of noble gases, separation of He from other noble gases (Ne, Ar, Kr, and Xe), followed by time consuming work associated with mass spectrometry. This makes it difficult to detect an abrupt change in ³He/⁴He ratios associated with seismic or volcanic activities. Therefore, the continuous monitoring of ³He/⁴He ratios of natural fluids with a time resolution of less than one day would be highly desirable.

In addition to ³He/⁴He monitoring, the determination of the ⁴He/²⁰Ne ratio could be a candidate because it can vary by more than 5 orders of magnitude from 0.3 to >40,000,¹⁰⁾ as has been reported for gases and rocks. In such cases, ²⁰Ne abundances are regarded as an indicator of the degree of contamination from the atmosphere, because ⁴He/²⁰Ne ratios in the atmosphere can be as low as 0.32, while those in samples derived from the Earth’s deep interior often reach values of more than 100.^1,2,10) A compact mass spectrometer with ion optics of a single-focusing magnetic-sector with large incident and exit angles was first designed and examined by Wu and Matsuda.¹¹⁾ Another mass spectrometer with the same ion optics was constructed and attained a mass resolving power of ∼600 at a 5% peak height, succeeded in measuring the ³He/⁴He ratio of atmospheric He.¹²⁾ The mass spectrometer had the potential to be practically used for ³He/⁴He monitoring, but no applications have been reported to date.

We designed and constructed a mass spectrometer (He-MS) based on a new ion optics for the purpose of He isotope monitoring. In this paper, we report on the design, manufacture, tuning and evaluation of He-MS. The main objective was the continuous ³He/⁴He measurement to monitor ³He/⁴He variations in naturally occurring gases/groundwaters that are associated with seismic or volcanic activities deep underground.

Design and Manufacture of the He-MS

Required specifications and ion optics of He-MS

The He-MS was originally intended for use at an observation station such as a seismographic station, where groundwater or gas is continuously supplied from a monitoring well. For such purpose, a compact size is preferable, and there are additional requirements, 1) a simple system; 2) maintenance-free for a long period; 3) capable of the simultaneous measurement of ³He and ⁴He by a double collector system; 4) measurable ²⁰Ne; 5) high vacuum condition; 6) large dispersion of ion trajectories to achieve high resolving power; and 7) high transmission of the ion beam inside the flight tube for achieving a sufficient high sensitivity to permit the minor isotope ³He to be detected.

To simultaneously satisfy three of the requirements, i.e., compact, large dispersion and high conductance, suitable ion optics was investigated with the computer programs, “TRIO” (Matsuo et al.)¹³⁾ and “TRIO-DRAW” (Toyoda and Matsuo),¹⁴⁾ for calculating ion trajectory considering third order aberration.

The designed ion optical system is basically a 90° deflection, single focusing and with 10 cm of the ion curvature, combined with an electrostatic quadrupole lens (Q-lens). A schematic diagram of the ion optics is illustrated in Fig. 1. Simulated ion trajectories from an exit slit of the ion source to the focal point in the x-axis (parallel to the curvature of the ion orbit) and y-axis (parallel to the direction of the sector magnetic field) using “TRIO-DRAW”¹⁴⁾ are shown in Figs. 2(a) and 2(b), respectively. The Q-lens prevents the ion orbit from diverging in the y-direction and enables ions to pass through a narrow gap between the pole pieces of the main magnet. The combination of large incident and exit angles (29°20′) and the Q-lens allows high transmission for the ion beam and high mass dispersion. The large angles can improve the convergence in the small geometry by the lens effect at the fringing magnetic field. As a result, the ion optics achieves a high resolving power despite having a small ion orbit radius. Simulated mass spectra of ³He⁺ and HD⁺ by “TRIO-DRAW”¹⁴⁾ are shown in Fig. 3, where an ion source exit slit size of 2 mm×0.1 mm is adopted. Based on the calculation, the theoretical mass resolving power is >700 in ideal conditions.

Fig. 1. Schematic drawing of the ion optical system of the mass spectrometer, He-MS. The shapes of the ion beams of ³He and ⁴He emitted from the ion source exit slit are regulated by the quadrupole lens (Q-lens), dispersed by the magnetic field with a dispersion angle of ca. 90° (incident and exit angles of 29°20′), gap width of 14 mm and radius of 100 mm and 115 mm for ³He and ⁴He, respectively, and then focused on each focal points where collector slits are placed. Two shunts regulate the edge of the magnetic field. Units for lengths are in mm.

Fig. 2. Ion trajectories simulated by “TRIO-DRAW” (Toyoda and Matsuo)¹⁴⁾ in (a) x-direction (deviation from the main axis of ion orbital in horizontal direction to the plane of Fig. 1) and (b) y-direction (likewise in perpendicular direction to the Fig. 1 plane).

Fig. 3. Result of a simulation of a mass spectra (left: ³He; right: HD). Horizontal and vertical axes correspond to m/z and ion beam intensity, respectively, both with arbitrary units.

Plan of He-MS based on the ion optics

Figure 4 is a schematic showing the overall configuration of He-MS, which accommodates the design of ion optics obtained above. Most parts are made of SUS 304 with flanges of ConFlat-type with Cu-gaskets and with Au-gaskets. A conventional Nier type, electron impact ion source is installed. The ion source is the same as those used in modified-VG5400 mass spectrometers in our laboratory, which simplifies the evaluation and/or investigation of the performance as well as trouble shooting. A split-type flight tube was designed for the simultaneous detection of both ³He and ⁴He. Because of the low ³He/⁴He ratio (≤10⁻⁵) of naturally occurring He, the beam intensity of ³He is much weaker than the ⁴He ion beam, ³He is measured with a secondary electron multiplier operated in the ion-counting mode, while measuring ⁴He involves the use of a Faraday cup (“High Faraday” in Fig. 4). In the flight tube, the ³He ion beam passes a narrow collector slit (300 μm in width) so as to separate ³He ions from interfering HD and H₃ ions with a resolving power of ≥500, while ⁴He is introduced into the outer flight tube, which passes through the wider collector slit (900 μm in width). In order to simultaneously obtain ³He and ⁴He signals, the position of the collector slit for ⁴He can be adjusted by means of a manipulator and the focal point of the ⁴He ion beam can be moved slightly by applying positive voltages to a pair of deflecting plates that are located in front of the collector slit. Another Faraday cup (“Axial Faraday”) can be inserted into the ³He beam line to measure a strong ion beam such as ²⁰Ne⁺ and ⁴⁰Ar⁺ (Fig. 4).

Fig. 4. Design drawing of the He-MS instrument, consisting of the ion source, Q-lens, flight-tube, and collectors. High and axial Faraday mean the Faraday cup for ⁴He (high m/z) and coaxially-arranged one. The axial Faraday can be inserted on the main axial beam line in front of the SEM when an intense ion beam is required to be detected.

The split-type flight tube also can reduce background noise near the ³He spectrum caused by the scattering of ⁴He ions inside the flight tube. Additional advantages for the double collection system are: 1) easy and stable operation for measuring two ion beams of largely different intensities by use of suitable ion detection methods; and 2) easier setting of the main electromagnetic field intensity which should be stable most of the time when ion beams are being measured. When a mass spectrometer is equipped with a single-collector system, the isotope ratio is measured with a peak-jumping mode, in which each isotope peak and baselines are measured by changing the main magnet field intensity corresponding to each m/z. Such peak jumping thus requires repeated changes in the magnetic field during an isotope ratio analysis, making precise and stable control of the magnetic field difficult. In contrast, a double collector system is capable of reducing magnetic field jumping, since only two magnetic field settings (one for He peak center and the other for baseline) are required.

Ultra high vacuum pumping system for He-MS is composed of a turbo-molecular pump (150 L/s), an ion pump (20 L/s) and a rotary pump. The whole He-MS can be heated to ca. 300°C in a baking box, because the components are made of metal (SUS304, copper and gold) and ceramics. After a baking operation at 250–280°C for a few days, a vacuum of <3×10⁻⁸ Pa can be attained. Two getter pumps (NP-10: SAES Getters Co.) placed in the flight tube effectively reduce H₂ background. After achieving ultra high vacuum conditions, an ion pump is used to maintain the vacuum instead of a turbo-molecular pump.

Noble gas preparation line

He-MS is equipped with a noble gas preparation line similar to those described in Nagao et al.¹⁵⁾ and Sumino et al.,¹⁶⁾ in which noble gases are purified by chemically removing reactive gases with titanium–zirconium getters heated at about 800°C, He and Ne are then separated from Ar, Kr and Xe by adsorbing the latter on charcoal traps cooled at liquid nitrogen temperature, and finally the He is separated from Ne by adsorbing the Ne on a sintered stainless steel trap (Cryogenic trap) cooled at about 20 K. The purified He in the line is introduced into He-MS for determining the ³He/⁴He ratio and the amount of ⁴He. Next, Ne is desorbed from the Cryogenic trap at approximately 50 K, and then introduced into He-MS for measuring amount of ²⁰Ne. Standard gases such as known amounts of the atmosphere and the ³He/⁴He standard gas with ³He/⁴He=(28.88±0.14)×10⁻⁶ (HESJ)¹⁷⁾ stored in reservoirs are used for tuning the ion source, the collector system and the alignment of the flight tube, or can be used to evaluate the performance of He-MS. When an unknown gas sample is analyzed, a small amount of gas (∼1 cm³ STP) is separated from a sampling bottle (30–100 cm³) using a gas pipette with a known volume (∼2 cm³) and its pressure and temperature are measured, and the sample is then introduced into the preparation line. Most of the preparation line is also composed of stainless-steel and thus can be baked at 250°C to maintain ultrahigh vacuum conditions.

Controller for the ion source

The Nier-type ion source is composed of an ionization chamber, a filament, two electrodes to repel produced-ions from the chamber (repellers), an electrode to collect the current of the electron beam from the filament (trap), an electrode to push out the produced-electron from the filament toward an entrance slit of the chamber (grid), two lenses to draw out ions from the chamber and focus the ion-beam (half plate), and an ion exit slit at the ground potential (earth slit). A magnetic field is applied to the ion source by a pair of permanent magnets placed on the outside of the ion source housing, to improve ionization efficiency by collimating the electron beam. The controller of the ion source is basically composed of six circuits: 1) A constant current supply to the filament; 2) A voltage setting circuit to apply suitable voltages to each electrode; 3) Circuits to detect and control the electric current through the trap (hereafter trap current) to the desired values (400 μA for the usual operation); 4) A resistive divider to apply an electric potential, which is adjustable between the potentials of the chamber and the earth slit (ground), to the half plates. A high voltage of 3.5 kV from HER-5N6 power supply (Matsusada Precision Inc.) is applied to the chamber to accelerate the ions that are produced in the chamber. Each voltage relative to that for the chamber is adjustable within the ranges from −50 to −90 V for the filament, from −24 to +24 V for the repellers, and from +0 to +50 V for the trap. The grid voltage is adjustable from 0 to −24 V relative to the filament voltage. The stability of the each voltage is better than ±0.1% relative to the adjusted one within a temperature range of ±2°C.

Computer-controlled DC current supply for the main magnet

The DC current supply for the main electromagnet consists of a Hall probe (BHT-910: F.W. BELL) inserted between the pole pieces of the main magnet and a constant DC current supply (NL035-30: Takasago Ltd.) controlled by a GP-IB programmer (AP-1628A: Takasago Ltd.). The Hall probe has a gain of 0.75 mV/kG under a control current of 100.00 mA. The Hall-voltage is further amplified by an instrumental-amplifier with a gain of 130. The constant DC current supply applies electric current to the magnet under the control of a reference voltage produced by GP-IB programmer with a resolution of 16 bits. By using an electrical circuit and the computer program written in HP-Basic on a Microsoft Windows PC, which is based on the programs reported by Nagao et al.¹⁵⁾ and Sumino et al.,¹⁶⁾ an excitation current for the main magnet can be set and maintained at a given value by the program to obtain a mass spectrum by scanning the magnetic field.

Voltage supply for the Q-lens

Based on the ion optics calculation, the required voltages for the Q-lens are about ±25 V when the ion acceleration voltage is 3.5 kV. To compensate for a possible slight disagreement between the calculated ion optics and actual ion-orbit, the power supply can adjust voltages of the four electrodes of the Q-lens independently by changing connections among segmented resistances using rotary switches. Two DC power supplies (PL-120-0.6: Matsusada Precision Inc.) provide source positive and negative voltages.

Collector assemblage for ion detection

A schematic diagram of the measuring instruments assemblage for ion detection is shown in Fig. 5. A weak ion current of ³He is measured using an ion counting system composed of a secondary electron multiplier (ETP AF133H: SGE Analytical Science) and a pulse counting unit composed of amplifier, discriminator, and counter (ORTEC Co.). The counting rate of ions output from the counter (ORTEC 996) is acquired by a computer via GP-IB, and then converted to “imaginal” voltage for convenience by the following formula,

Fig. 5. Schematic diagram of ion detectors and signal processing. Output signals of high and axial Faraday and ion counting are acquired by a computer through a GP-IB interface.

where e is an elementary charge (1.6×10⁻¹⁹ C), N counting rate (cps), and R imaginal electric register (1×10¹⁵ Ω).

The intense ⁴He beam is collected in the Faraday cup and measured as a current mode by converting to voltage through an operational amplifier (OPA104CM: Burr-Brown Co.) with a 1.0×10¹⁰ Ω feedback register (RHA2B: Hydragin Co.). The output voltage is measured using a digital multimeter (HP 34401: Agilent Technologies) and read by the computer via GP-IB. Data acquisition is carried out using the HP-Basic program reported by Nagao et al.¹⁵⁾ and Sumino et al.¹⁶⁾ A pair of electrodes (deflectors in Figs. 4 and 5) is installed in front of the collector slit for ⁴He, which functions as a zoom lens as well as a deflector of the ⁴He ion beam to focus the ion beam at the collector slit and to produce a flat top mass spectrum.

Results and Evaluation of the He-MS

Mass resolution, sensitivity, and detection limit

The basic specification of He-MS are summarized in Table 1, and the current instrument is compared with the miniature and conventional mass spectrometers for noble gas measurements developed by Sano et al.¹²⁾ and the Modified-VG5400 installed at the University of Tokyo. Vacuum conditions, 5% valleys of the mass resolutions, sensitivities for ⁴He, background peak heights of HD, detection limits for ³He, and dynamic ranges of He detection, respectively, are listed. The mass resolution power is 430 for a 5% peak height (upper panel in Fig. 6), which is lower than the simulated value of 700 (Fig. 3). The low value may be due to inaccurate alignment of the source and collector slits as well as the flight tube, and the vertical length of ion beam emitted from the earth slit (5 mm×0.1 mm) of the ion source which is larger than that assumed in the simulation (2 mm×0.1 mm), resulting in the scattering of ions and broadening of the observed ³He peak. Considering only sensitivity, the estimated detection limit for He ions corresponds to <10⁻¹⁵ cm³ STP of He in the MS. However, the tailing effect of HD due to the relatively low mass resolution power and background noise due to scattered ⁴He ions both of which interfere with the ³He signal, would make the practical detection limit worse.

Table 1. Basic specification of the He-MS instrument and other noble gas mass spectrometers.

	Vacuum	Mass resolution	Sensitivity	Background HD†	Detection limit	Dynamic range
	Pa	5% valley	cm3 STP/cps	cm3 STP	cm3 STP 3He
This study	2.7×10−8	430	3.0×10−14*	1.5×10−12	1×10−13	4×107
Sano et al.12)	4.0×10−8	600	1.6×10−15**	1.3×10−11	1×10−12‡	<1×107
MS-III#	1.3×10−8	550	6.4×10−15	3.6×10−13	1×10−15	1×109

^# Modified-VG5400 installed at the University of Tokyo. * ⁴He at ion counting mode. ** Converted 1.6×10⁻¹⁹ A to 1 cps. ^† HD ion current is converted to cm³ STP by comparing with ³He ion current and sensitivity. ^‡ Estimated from S/N ratio of mass spectra.

Fig. 6. Mass spectra of ³He⁺ and HD⁺ (+H₃⁺). The upper panel is the spectrum when ³He of the standard He gas (HESJ)¹⁷⁾ is introduced into He-MS, and lower one is the background spectrum when the MS is operated in a static mode, i.e., isolated from the pumping system.

As can be seen in the background mass spectra around ³He⁺ in Fig. 6 (lower panel), a residual signal which would be a tailing of the HD⁺ spectrum is still observed at the position of ³He⁺ even in evacuated conditions. The mean sensitivity for ⁴He by the ion counting mode determined by seven measurements of the standard air is (3.0±0.1)×10⁻¹⁴ cm³ STP/cps (Table 1). Although the background signals of HD and H₃ are small and equivalent to 1.5×10⁻¹² cm³ STP of ³He (Table 1), its tailing seems to appear at the center of the ³He peak. The tailing effect at m/z=3.0160 (corresponding to ³He center) can be virtually converted to an amount of ³He of 1.6×10⁻¹⁴ cm³ STP.

The scattering of ⁴He ions also affects the baseline at the ³He spectrum. Figure 7 shows the correlation between the baseline signal intensity adjacent to ³He and the amount of ⁴He in the MS. The good linear correlation suggests the presence of interfering ions equivalent to 1 cps per 6.4×10⁻⁷ cm³ STP ⁴He in the MS. If the interfering ions are equally detected at both the peak and baseline of ³He, they would not seriously affect ³He detection by a baseline correction. However, in the case of low counting rates for ³He ions, statistical variations in the counting rates for both ³He and the baseline would strongly affect the determination of the ³He peak intensity, occasionally resulting in a negative apparent ³He output. Given that the introduced ⁴He is 10⁻⁶ cm³ STP, the noise from the scattered ion would correspond to 4.3×10⁻¹⁴ cm³ STP ³He on the positions of the ³He peak and the baseline. Based on the counting rate and the integration time for the ³He count accumulation in practical measurements, the statistical variation in this noise is estimated to be 1.3×10⁻¹⁴ cm³ STP ³He equivalent.

Fig. 7. Correlation between signals at m/z=2.895, assumed to be baseline signal for ³He in the routine isotopic measurement, and amount of ⁴He introduced into He-MS. A good linear correlation shows that scattered ⁴He ion is detected at the baseline, resulting in an increased signal at baseline for ³He.

Therefore the overall detection limit for ³He is estimated to be 10⁻¹³ cm³ STP (Table 1) defined as a signal-to-noise ratio of ∼3 with the noise as the sum of those derived from noise level of the detector (corresponding to <10⁻¹⁵ cm³ STP), scattering effect of ⁴He⁺ (1.3×10⁻¹⁴ cm³ STP), and the tailing of HD (1.6×10⁻¹⁴ cm³ STP). The detection limit is much better than that of the in-house fabricated MS (10⁻¹² cm³ STP),¹²⁾ and comparable to those of commercially available instruments, ranging from 3×10⁻¹⁴ (Sano and Wakita)¹⁸⁾ to 6×10⁻¹⁶ (Sumino et al.)¹⁶⁾ cm³ STP. Although the remaining problems posed above make the performance of He-MS worse to some extent, a slight difference in the He isotopic ratio in natural samples can still be measured, as shown below.

Reproducibility in He isotopic ratio measurements of the He standard gas (HESJ)

Reproducibility of He isotope measurements is the most important factor for a long-period and continuous monitoring of ³He/⁴He in natural samples. To investigate the reproducibility of the ³He/⁴He ratio measurements with He-MS, we measured the He standard gas, HESJ with a ³He/⁴He ratio of (28.88±0.14)×10⁻⁶ (Matsuda et al.),¹⁷⁾ a total of 35 times using nearly constant amount of He (⁴He ∼1.1×10⁻⁶ cm³ STP). Results of the replicate measurements of ³He/⁴He ratios are shown in Fig. 8, for which different sensitivities between the ion counting and the Faraday cup used for the ³He and ⁴He measurements, respectively, are uncorrected. Since HESJ has a high ³He/⁴He ratio, 20 times higher than that of terrestrial air: (1.399±0.013)×10⁻⁶ (Mamyrin et al.),¹⁹⁾ and is composed of almost pure He, an isotope measurement is relatively easy, since the He purification is not necessary and the effect of interfering ions such as the HD on ³He is smaller than in the case of natural samples with lower ³He/⁴He (<10×10⁻⁶).

Fig. 8. Result of replicate measurements of HESJ. Open and solid symbols are obtained using different ion source conditions before and after the replacement of a filament in the ion source. Error bars are one standard deviation of repeated data acquisitions for each sample analysis.

An average of measured ³He/⁴He ratios is 1.734±0.062 (2σ) in the first period (24 November 2007–5 January 2008) and 1.770±0.050 (2σ) in the second period (13 October 2008–30 November 2008). Between the two periods, the filament of the ion source was replaced and the conditions for the ion source were changed slightly. The ³He/⁴He stability was 3–4% for more than a month, which is sufficient to permit the detection of changes in the ³He/⁴He ratio during the continuous monitoring of natural He. For example, changes in He isotopic ratio reported for seismic/volcanic activities are more than 10% (Bräuer et al.)⁴⁾ or a factor of two or more (Sano et al.)^3,6) compared to those observed during periods of steady-state. He-MS therefore has the capabilities for distinguishing between such isotopic changes in nature.

Measurements of natural gas samples

In case of analyses of gas samples in nature, the procedure for measuring He differs from that for HESJ. Volcanic, spring and natural gases are mainly treated as samples in noble gas geochemistry. In the natural samples, noble gases are trace components, and thus most of the sample gases, e.g., carbon dioxide, nitrogen, oxygen, methane etc., must be removed by a purification procedure before a noble gas analysis can be performed. We measured two natural gases, four volcanic gases, six hot-spring gases, and one soil gas (Table 2), which were collected in 50 cm³ lead-glass bottles with two vacuum stopcocks on both ends. The amounts of the natural gases introduced into the purification line were limited to <1 cm³ STP because of the capacity of the purification system. The introduced noble gases were purified and the He and Ne were separated from other noble gases and then analyzed following the procedure described above.

Table 2. Helium isotope ratios, ⁴He and ²⁰Ne concentrations of gas samples measured by He-MS and other mass spectrometers.

Sample	Total gas volume	C(4He)	3He/4He	C(20Ne)	4He/20Ne	3He/4He	4He/20Ne	Ref.
	cm3 STP	ppm	10−6	ppm		10−6
	He-MS					MS-III or MS-IV
Methane rich continental gas	0.279	278	0.084±0.013	0.0706	3930	0.1006±0.0037	5780	1
	0.247	163	0.025±0.015	0.122	1340	0.0170±0.0022	1040
Volcanic gas in Galápagos	0.314	4.36	17.25±0.31	3.97	1.10	17.79±0.20	1.56	2
	0.700	12.6	24.14±0.46	0.0054	2320	25.03±0.35	2360
Volcanic gas in China	0.700	2.31	7.59±0.14	0.0249	92.7	7.501±0.053	82.1
	0.505	10.9	8.17±0.15	0.0485	224	8.573±0.061	162
Hot-spring gas in China	0.719	167	3.27±0.10	14.1	11.8	3.498±0.056	11.1
Hot-spring gas in Japan	1.190	12.2	9.12±0.15	1.33	9.18	9.067±0.080	11.0
Carbonate-rich hot spring gas	0.667	5580	0.880±0.041	12.6	442	0.854±0.014	541
	0.085	6560	0.685±0.057	10.8	609	0.698±0.030	856
Carbonate-rich hot spring gas	0.783	93.2	7.10±0.30	0.483	193	7.140±0.080	588
	0.958	1.19	3.01±0.11	0.0077	154	2.890±0.030	181
Soil gas in Canary Islands	0.707	6.59	1.45±0.09	21.0	0.31	1.610±0.064	0.281	3
		4He	3He/4He	20Ne
		10−10 cm3 STP	10−6	10−10 cm3 STP
Blank		2.6	60±41	5.7

1: Kotarba and Nagao¹⁰⁾; 2: Padrón et al.²⁵⁾; 3: Padrón et al.²⁶⁾Errors for ⁴He and ²⁰Ne concentrations are ca. 10%. Errors associated with ³He/⁴He ratios are one standard deviation.

Results for the natural samples are summarized in Table 2 with data measured by two modified-VG5400 (MS-III and MS-IV) noble gas MS instruments for comparison. Details of the analysis of noble gas samples using MS-III and MS-IV have been described by Aka et al.²⁰⁾ and Sumino et al.,²¹⁾ respectively. Table 2 also shows the procedural blank levels which are sufficiently low compared to the amounts of He and Ne from the samples introduced into He-MS. As shown in Fig. 9, ³He/⁴He ratios measured using He-MS are in good agreement with those obtained by MS-III and -IV over a ³He/⁴He ratio range from 10⁻⁵ to 10⁻⁸. Because the amounts of He introduced into He-MS were adjusted to be almost constant (corresponding to ∼3×10⁻⁸ Torr) to maintain the ion source conditions nearly constant, the amount of ³He in the MS decreased with decreasing ³He/⁴He ratio and, as a result, it was difficult to precisely determine the ratio. Figure 10 compares ⁴He/²⁰Ne ratios measured by He-MS with those by MS-III/IV. They show good agreement in general, and when differ to an extent beyond their analytical errors, the discrepancy is acceptable unless the ⁴He/²⁰Ne ratio of a sample is close to the atmospheric value (0.32), because it is used as an index of atmospheric contamination.

Fig. 9. Comparison of ³He/⁴He ratios obtained with He-MS and those with MS-III or MS-IV. The solid line indicates 1 : 1 correlation. Error bars are 1σ.

Fig. 10. Comparison of ⁴He/²⁰Ne ratios obtained with He-MS and those with MS-III or MS-IV. The solid line indicates 1 : 1 correlation. Error bars are 1σ.

Measurements of the atmospheric He

In order to measure He (5.2 ppm) in the atmosphere, the more abundant Ar (0.9%) and Ne (18 ppm) need to be separated from the He fraction. Moreover, since the ³He/⁴He ratio of the atmospheric He is 1.4×10⁻⁶, consequently the volume fraction of ³He in the atmosphere is 7.3 ppt, measurement of the He isotopic ratio requires a high sensitivity for ³He detection and a technique for isolating He from Ne. It is possible to isolate He from other noble gas species using cold traps inserted in the purification line, and He-MS has a high sensitivity for ³He and mass resolving power which can distinguish HD⁺ and H₃⁺ from ³He⁺. The results of repeated measurements of atmospheric He are shown in Fig. 11. The correction factor for raw ³He/⁴He ratios, involving the sensitivity ratio between ion counting and Faraday cup collectors used for ³He and ⁴He, respectively, and the mass discriminating effect of the MS, was calculated as a ratio of the measured and absolute ³He/⁴He ratio of HESJ.¹⁷⁾ Measured ³He/⁴He ratios using various amounts of He, (0.31–2.3)×10⁻⁶ cm³ STP ⁴He, introduced into He-MS are plotted in Fig. 11. The observed atmospheric ³He/⁴He ratios are identical to the reference value of (1.399±0.013)×10⁻⁶ (Mamyrin et al.)¹⁹⁾ within analytical errors.

Fig. 11. ³He/⁴He ratios of atmospheric He as a function of the amount of He introduced into He-MS. The amounts of He correspond to 0.06−0.44 cm³ STP of atmosphere. The broken line and hatched area show the average and standard deviation of the measured ³He/⁴He ratios. Error bars are 1σ.

Helium Extraction System and Its Application

A helium extraction system was developed for practical application to routine measurements of natural gas/water samples based on exploiting the high permeability of He through a silica glass tube. As a result, it can simplify the procedure for purifying He in the gas phase compared with the conventional method described above. The continuous He sampling method from a gas–water mixture is schematically presented in Fig. 12. The gas–water separator module²²⁾ contains a bunch of silicone hollow tubes, and the insides of the tubes are isolated from the outsides. The module selectively extracts and isolates the gas phase from water running through the insides of the silicone hollow tubes, by decompression of the outside volume of the tubes using a diaphragm pump. Gases from the exchange module are introduced into the outside volume of the heated silica tube, and nearly pure He passes through the silica glass wall, and it can then finally be introduced into He-MS. When the inner volume of the silica tube is under a vacuum, the pressure difference for He between the outer and inner volumes of the tube enables He to penetrate through the silica glass into the inner volume connected to the He analysis system. Because the permeability of He through silica glass is still as low as 10⁻¹⁰ cm³ STP/s/cm²(area)/mm(thickness)/Torr(pressure difference) at room temperature,²³⁾ the silica tube needs to be heated to attain high permeability to collect a sufficient amount of He into the inner volume for the isotopic measurement. It is known that the permeability of He increases exponentially with an increase in the temperature of the silica glass. The much lower permeability of other gases than He permits the measurement of He without the need for the purification/separation procedure described above. For example, at 700°C, the permeability of hydrogen, Ne, Ar and nitrogen are respectively 15, 76, >1×10⁷, and >1×10⁷ times smaller than that of He.^23,24)

Fig. 12. Schematic drawing of a system for extracting He from water. Gas species in water running through silicone hollow tubes permeate into the outside of the tubes and transferred to the outer volume of the silica tube by the diaphragm pump. The diaphragm pump maintains gas pressure outside the silicone hollow tubes low, keeping continuous gas extraction from the water. The silica tube can be heated up to 700°C by a sheath heater. When the gas–water separator module is removed and a gas sample is directly introduced into the outer volume of the silica tube, the system can selectively introduce He into He-MS from the gas sample through the silica tube.

By using this extraction system, the time required for the He isotope measurement can be greatly reduced due to time savings associated with the purification and separation of He, resulting in a high time resolution and simple procedure for an automated continuous monitoring system to be constructed in the future. Tentative data are summarized in Fig. 13. In this study, the data were obtained from air in the laboratory without the gas–water separator module and diaphragm pump. Type 1 silica tubes consist of silica tube connected to a ConFlat flange with an O-ring made of Viton. On the other hand, type 2 silica tubing, as shown in Fig. 12, is connected to the metal flange with a Kovar seal instead of a Viton O-ring.

Fig. 13. Plot of ³He/⁴He vs. ⁴He/²⁰Ne obtained by using the silica-tube extraction system. Filled circle represents the data with the type 1 tube and open-square with the type 2. In the type 1, silica tube is connected to a metal flange with a Viton O-ring. On the other hand, silica tube is connected to a metal flange through a glass–metal joint (Kovar seal) in the type 2. Two dotted lines in the figure are mixing lines represent addition of commercially available He to the mixture of atmospheric He and Ne permeated through the two types of silica tubes. For the commercial He ³He/⁴He ratio of 10⁻⁸ is assumed. Note that ⁴He/²⁰Ne ratios of all samples shifted to higher values due to the higher permeability of He than that of Ne in silica tube. Permeability ratios of ⁴He to ²⁰Ne (⁴He/²⁰Ne) through silica tubes of type 1 and 2 are estimated as 63 and 130 based on the lowest ⁴He/²⁰Ne ratios.

The low ⁴He/²⁰Ne ratios, ∼25, observed for Type 1 tubing compared with those for Type 2 tubing may be attributed to the infiltration of atmospheric Ne through the Viton O-ring. The infiltration of Ne is reduced when Type 2 tubing is used as inferred from the higher ⁴He/²⁰Ne ratios, >50. In addition to the different ⁴He/²⁰Ne ratios among the two different types of silica tubes, the decreased ³He/⁴He ratios down to 0.45×10⁻⁶ accompanied by an increase in ⁴He/²⁰Ne ratios up to 200 were observed in both cases, as shown in Fig. 13. When these data were observed, the amounts of He infiltrated through the silica glass increased.

This can be explained as an increase of He with low ³He/⁴He in the laboratory air, because on one occasion we checked the vacuum leakage of the purification line before the silica tube experiment using a commercially available sample of He stored in a metal cylinder. The commercially available He was mainly collected in well gas field in the old continental area,¹⁰⁾ and thus was enriched in ⁴He produced from α-decay of U and Th and accumulated during gas storage in the crust over the geological timescale. Since such He is frequently used in laboratory experiments, the observed He increase in the laboratory atmosphere could be due to the release and/or leakage of He from other laboratories in our building. If this is the case, we eventually demonstrated the ability of He-MS for the continuous monitoring of ³He/⁴He and/or ⁴He/²⁰Ne variations in nature.

The ⁴He/²⁰Ne ratio in air is 0.32 and the observed lowest value is 43 when type 2 tubing was used, thus the transmission factor for He/Ne ratio through the silica tube at ca. 700°C in this system becomes 130. The value generally agrees with 76, measured at a temperature range of 25−500°C by Altemose²³⁾ and Norton²⁴⁾ if we consider the uncertainty associated with the extrapolation up to 700°C.

Conclusion

We designed and constructed a new mass spectrometer, He-MS, for the continuous monitoring of He isotopic ratios in naturally occurring gases and waters. Although the mass resolution still remains insufficient, He-MS performed satisfactorily in the measurement of small-amounts of He in natural gas samples under the present conditions. Detection limits for ³He as low as 10⁻¹³ cm³ STP were achieved by improving the quality of the vacuum in the MS. An acceptable reproducibility of 5% or less for ³He/⁴He measurements was demonstrated by the repeated isotopic ratio measurements of a standard He sample. The dynamic range in ³He/⁴He ratio measurement was determined to be >1,000 by measuring different types of natural gas samples. There is no systematic difference in exceeding the analytical errors between the ³He/⁴He ratios obtained with He-MS and those by the VG5400 mass spectrometers in our laboratory.

An extraction system composed of a gas–water separation module and hot silica tubes was designed to extract He dissolved in water. The hot silica-tube showed high selectivity for He extraction from gas samples compared with other noble gases and reactive gas species. Since this He extraction system is maintenance-free and does not require a cryogen for separating He from other gas species, it will be very useful in the continuous monitoring of ³He/⁴He ratio at a field observation station.