Superconducting Tunnel Junction Detectors for ... - IEEE Xplore

IEEE TRANSACTIONS ON APPLIED SUPERCONDUCTIVITY, VOL. 24, NO. 4, AUGUST 2014

2400208

Superconducting Tunnel Junction Detectors for Analytical Sciences Masataka Ohkubo, Member, IEEE, Shiki Shigetomo, Masahiro Ukibe, Go Fujii, and Nobuyuki Matsubayashi

(Invited Paper)

Abstract—Superconducting tunnel junction (STJ) detectors exhibit superior detection performance for photons and particles at a high spectroscopic resolution of ∼10 eV, a short dead-time (decay time) of ∼μs, a high quantum efficiency of ∼100%, and a low detection threshold energy of less than 1 eV, which cannot be achieved by conventional detectors. The outstanding detection performance originates from a small superconducting energy gap of ∼meV, which is three orders of magnitude smaller than ∼eV in semiconductors. This paper reports our recent progress in two applications of STJ detectors to fluorescence-yield X-ray absorption fine structure (XAFS) spectrometry for trace light elements in matrices and mass spectrometry (MS) for ions with the same mass/charge-number ratio (m/z) but different charge states and neutral fragments. Index Terms—Mass spectroscopy (MS), superconducting devices, superconducting tunnel junction (STJ), X-ray absorption fine structure (XAFS), X-ray detection.

I. I NTRODUCTION

S

UPERCONDUCTING tunnel junction (STJ) detectors have a sandwich structure of superconductor-insulatorsuperconductor with an insulator thickness of ∼1 nm. The device structure is the same as that of Josephson junctions. For detecting photons or particles, the STJ detectors are operated in the so-called Giaever mode [1], in which the dc Josephson effect is suppressed by applying a small magnetic field of 10–100 G so that the change of quasiparticle tunnel current can be measured. The junctions are biased within the sub-gap region to observe the change of quasiparticle tunnel current.

Manuscript received January 30, 2014; revised March 26, 2014; accepted April 7, 2014. Date of publication April 21, 2014; date of current version May 12, 2014. This work was supported in part by the Nuclear Research program of the Ministry of Education, Culture, Sports, Science, and Technology (MEXT), the JST-SENTAN Program of Japan Science and Technology Agency (JST), the Grant-in-Aid for Scientific Research (A) of the Japan Society for the Promotion of Science (No. 22246056), and by the Nanotechnology Platform Program of MEXT (12024046). This paper was recommended by Associate Editor A. Kleinsasser. M. Ohkubo is with Tsukuba Innovation Arena Headquarter (TIA), National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba 3058568, Japan (e-mail: [email protected]). S. Shigetomo, M. Ukibe, G. Fujii, and N. Matsubayashi are with the Research Institute of Instrumentation Frontier (RIIF), AIST, Tsukuba 305-8568, Japan (e-mail: [email protected]; [email protected]; [email protected]; [email protected]). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TASC.2014.2318316

The thermally excited quasiparticles must be suppressed by cooling to a temperature well below superconducting transition temperatures (Tc ). When X-ray photons or such particles as atoms or molecules strike a superconducting electrode, and their energies are above the superconducting energy gap (2Δ), Cooper pairs are broken and quasiparticles are created at near to the final stage of the excitation and relaxation processes. The excess quasiparticles with energies just above 2Δ result in an increase of the quasiparticle tunnel current. Since the quasiparticles have an effective lifetime of ∼ μs for recombination into Cooper pairs, current pulses with a decay time of ∼ μs are produced for individual photons or particles. This detection mechanism is called direct detection, in contrast to coherent detection (or heterodyne detection) of electromagnetic waves of photon energies of less than 2Δ. One of the remarkable differences between Josephson-effect applications and detector applications is junction size. For superconducting quantum interference devices (SQUIDs) or single flux quantum (SFQ) logic circuits using Josephson effects, the junction size is as small as 1 μm square or less [2]. STJ detectors, on the other hand, should have a large junction size of 100–200 μm for X-ray spectroscopy or mass spectroscopy in order to obtain a reasonable number of photon or particle events. Furthermore, because one junction is still small even though it is a large junction, array detectors must be fabricated for high-throughput practical analyses. Current technologies for STJ array fabrication, cryogenics, and signal processing allow us to achieve a sensitive area of ∼ mm2 , which is still smaller than ∼ cm2 in the latest silicon drift detectors (SDDs). Nevertheless, a mean energy resolution of 14 eV [3] in 100 STJ pixels (1 mm2 ) outperforms 48 eV [4] in an SDD (10 mm2 ) for the oxygen K line (525 eV photons). Further progress of STJ technologies is required for analytical sciences. In this paper, we discuss two applications of STJ detectors: X-ray absorption fine structure (XAFS) spectrometry and mass spectrometry (MS). XAFS spectrometry is widely used in synchrotron radiation (SR) facilities in order to analyze the atomicscale local structure around the atoms that absorb SR photons or the electronic states of the atoms [5]. STJ detectors play an important role in selecting trace elements in bulk samples [6]. In particular, STJ detectors enable the analysis of light elements whose characteristic X-rays are in the soft X-ray region of less than 2 keV at a high energy resolution beyond the limit of semiconductor detectors [3], [7].

1051-8223 © 2014 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

2400208


Fig. 1. Top view of a part of the superconducting-tunnel-junction (STJ) array detector with 100 pixels for soft X-rays and low energy ions. The junction size is 200 × 200 μm2 . The sensitive area is 4 mm2 in total. The wiring leads are connected to the top and bottom superconducting electrodes.

In MS, STJ detectors are used to detect analyte atoms or molecules that are ionized and accelerated at a static voltage of 3–30 kV [8], [9]. STJ detectors enable charge state discrimination of ions or direct separation of different neutral fragments without ionization, which cannot be realized by MS instruments with a conventional ion detector [10], [11]. In this paper, we report the operational basis and our latest results of the applications of STJ detectors to analytical sciences. II. O PERATING P RINCIPLE OF STJ D ETECTORS An example of an STJ array detector with 100 pixels is shown in Fig. 1. Each pixel is a tunnel junction of 200 × 200 μm2 . For detecting photons or particles, a reasonably large sensitive area is necessary. However, the larger the junction size is, the lower the spectroscopic resolution becomes because of spatial nonuniformity, i.e. the amplitude of the signal that depends on the absorption position of quanta [12], [13]. Therefore, we select a junction size between 100 and 200 μm, depending on the application: for example, 100 μm for high resolution in XAFS and 200 μm for a large sensitive area in MS. The spatial nonuniformity can be reduced by carefully designing a junction layer structure of Nb/Al/AlOx /Al/Nb [13]. The spatial nonuniformity has been studied experimentally and theoretically [14]–[16]. Although a quasiparticle diffusion and tunneling model predicts experimental spatial profile shapes, some of the physics remains unsolved. For example, the spatial profile for 5-keV photons can change with bias [17]. Since it is unlikely that the quasiparticle diffusion constant depends on the bias point, some unusual non-linear effects may be responsible [18]. Nevertheless, in a low energy range of less than 2 keV, our STJ detectors exhibit a linear relation between deposited energy and output pulse height, and therefore are applicable to soft X-ray spectrometry or low energy particle detection. For practical use, the surface of the top Nb electrode is bare so that soft X-ray photons, atoms, and molecules impinge on the Nb surface directly, although the Nb surface exposed in air may be contaminated by oxidization and the detector response is slightly degraded [19].

Fig. 2. Tunneling processes in a superconducting tunnel junction biased in the sub-gap region (a) and cross-sectional TEM image (b). “S1” and “S2” denote two superconducting electrodes separated by the 1 nm-thick AlOx insulating layer indicated by “I.” Absorption of an X-ray photon or particle surface impact results in Cooper-pair breaking and creates quasiparticles. The two major tunneling processes are shown. The superconducting electrodes are Nb/Al bilayers in actual STJ detectors in order to obtain a better tunneling barrier quality, a better spatial uniformity, and signal amplification due to quasiparticle trapping and multiple tunneling.

STJ detectors use 2Δ as a scale for measuring the energy deposited upon a single-photon absorption event or a singleparticle surface impact event. Energy deposition of more than 2Δ leads to Cooper-pair breaking and creates excess quasiparticles. The mean number of quasiparticles, N , for an energy deposition of E is expressed by N = E/ε ∼ = 2E/2Δ

(1)

where ε is the mean energy necessary for creating one quasiparticle. The factor 2 corresponds to the fact that two quasiparticles are created by breaking a Cooper pair. The ε value deviates from Δ because of energy loss processes in superconductors. Numerical Monte Carlo calculations taking into account the quasiparticle and phonon excitations were performed in order to estimate the ε value [20], [21]. The calculated ε was 1.7Δ (2.6 meV in Nb). Thus, the mean lowest detectable energy is 3.4Δ (5.2 meV). Actually, it is impossible to detect a 5.2-meV photon, since there is currently no way to measure a few quasiparticles. The experimental lowest detectable energy is about 1 eV [22], [23], which is equivalent to about 400 quasiparticles. Tunneling probability and the noise level of semiconductor read-out circuits may limit the lowest detectable energy. Output pulse signals of STJs biased in the sub-gap region are created by an increase of tunneling current due to excess

OHKUBO et al.: SUPERCONDUCTING TUNNEL JUNCTION DETECTORS FOR ANALYTICAL SCIENCES

quasiparticles. Two tunneling processes mainly contribute to the current pulse generation, as shown in Fig. 2. One is the direct tunneling of a quasielectron from S1 to S2. The other is one electron of a Cooper pair in S1 tunnels to S2, producing a quasielectron in S1, and forms a Cooper pair with a quasielectron in S2. The current direction is the same as that of the first tunneling process, but the quasielectron flow direction is opposite. This process is equivalent to tunneling of a quasihole. These two tunneling processes allow multiple tunneling of quasielectrons within a quasiparticle effective lifetime (∼ μs), which amplifies the signal, but also produces additional statistical noise [24]– [26]. Nevertheless, multiple tunneling effectively improves the energy resolution eventually. In actual STJ detectors, the superconducting electrodes are composed of Nb/Al bilayers, as shown in Fig. 2(b). The Al layers with a 2Δ value smaller than that of Nb are effective for quasiparticle trapping near the tunneling barrier, which improves the tunneling probability. Moreover, the thick Al layers result in a long quasiparticle lifetime and a long quasiparticle diffusion length. Therefore, the Al layers enhance the signal amplification and improve the spatial uniformity. Apart from the detailed tunneling processes in superconductivity, the energy resolution (δE) divided by E of general quantum detectors that have an energy gap and follow the ideal Poisson statistics, which provide that individual carrier generation processes are independent, can be expressed by σN δE = 2.355 = 2.355 ε/E E N

(2)

where σN is the standard deviation of the number of created carriers. In Poisson statistics, the standard deviation is equal to the square root of N . The δE value in (2) is expressed in the unit of full width at half maximum (FWHM) of the pulse height distribution of the detector output pulses for individual quanta. The factor of 2.355 is to convert the standard deviation into FWHM. ε is larger than Δ and equals 1.7Δ in superconductors, as mentioned before. In addition, individual quasiparticle generation processes are not completely independent. The deviation from the ideal Poisson statistics is represented by the Fano factor (F ). The above calculations predict F = 0.195. As a result, the theoretical limit of energy resolution is expressed by √ √ (3) δE = 2.355 F E = 1.36 Δ × E. For Nb, the theoretical δE value for the characteristic X-ray photons of the oxygen K-line (O-K) at 525 eV is 1.2 eV, provided that all of the quasiparticles are collected and are counted exactly once to produce each current pulse. In our actual STJ detectors, the experimentally achieved δE value, i.e. the mean energy resolution for 100 STJ pixels with a junction size of 100 × 100 μm2 , is 14 eV including read-out circuit noise. The distribution of pixel number with δE is shown in Fig. 3 [3]. The mean intrinsic δE value after subtracting the read-out circuit noise (10 eV) is 10 eV. The degradation from the theoretical value is caused by spatial inhomogeneity, incomplete quasiparticle collection, and tunneling statistical noise, as mentioned above. Nevertheless, the STJ array detector outperforms the latest SDD with a δE value of 48 eV [4].

2400208

Fig. 3. Pixel number histogram of energy resolution (δE) for an STJ array detector [3]. The mean δE value of the 100 STJ pixels with 100 × 100 μm2 is 14 eV at the oxygen K-line of 525 eV. The total sensitive area is 1 mm2 . The intrinsic mean δE value after subtracting the read-out noise is 10 eV.

It was also pointed out that superconducting detectors would be useful for detecting high-energy heavy atoms that produce slow recoil atoms that create no carriers in semiconductors [20]. The slow atoms, on the other hand, create phonons that superconductors are sensitive to. The maximum energy of the phonons roughly corresponds to the Debye energy, which is 24 meV in Nb, and any phonons with energy > 2Δ can break Cooper pairs [27]. Therefore, the detection of low-energy atoms would be feasible through the phonon excitation. The initial idea was to detect atomic displacement cascades in superconductors, but phonons can be produced even in soft impact events of large molecules at a surface without penetration into a superconductor [28]. After the phonon excitation on the superconductor surface, the following detection processes are identical to those of the photon detection. III. A PPLICATION OF STJ D ETECTORS TO A NALYTICAL I NSTRUMENTS One of the advantages of STJ detectors over thermal detectors such as transition edge sensor (TES) microcalorimeters and metallic magnetic microcalorimeters (MMCs), which use the rise in temperature due to the absorption of each quantum, is a fast recovery time after photon or ion impact events, leading to a high counting rate which is especially important for XAFS and MS applications. The fast recovery time originates from the fact that the lifetime of quasiparticles produced by the events is about one-thousandth of the thermal recovery time. Another advantage of STJ detectors is insensitivity to base temperature drift or variation, as long as they are kept at a temperature below ∼Tc /10. This advantage originates from the electronic energy gap of the absorbers, which are the superconducting electrodes of tunnel junctions. Thermal detectors, on the other hand, usually have a gap-less absorber, so the amplitude of temperature rise is affected by base temperature drift because of the temperature dependence of heat capacity. In exchange, the calorimeters have an extremely high energy

2400208


resolution of better than 1 eV [29]. The energy gap promises long-term stability of detector operation without peak position drift and a high tolerance to heat input from the environment. Especially, heat input is unavoidable in ion detection, since an open space is necessary between the STJ detectors at 0.3 K and the ion source at room temperature to enable ions to strike the detector surface. Our Nb/Al-based STJ detectors operate at temperatures of less than about 0.5 K. We use a cryostat equipped with a closed-cycle 3 He cooler and a pulse tube cooler with a base temperature of 0.3 K. Cooling from room temperature to 0.3 K and recycling of the 3 He cooler can be performed automatically without supplying liquid helium. Therefore, this analytical instrument can be used without a knowledge of cryogenics. The detector biasing, signal amplification, and digital signal processing are performed by semiconductor circuits at room temperature. Details of the pulse signal processing using field programmable gate array (FPGA) logic circuits for the 100 STJ pixels are reported elsewhere [30]. A. X-ray Absorption Fine Structure (XAFS) Spectrometry in Synchrotron Radiation Facilities Synchrotron radiation (SR) facilities provide intense monochromatic photon beams from < 1 eV to > 1 MeV, which are versatile for material analyses, for example, crystal structures, atomic-scale local structures, and electronic states. STJ detectors are well suited to analyses at SR beam lines, measuring the photon energies of characteristic X-rays from elements at a high energy resolution that cannot be achieved by conventional semiconductor technology, and a high photon counting rate [6]. Our XAFS instrument with an STJ array detector having 100 pixels (SC-XAFS) covers a soft X-ray region of less than ∼2 keV, in which the characteristic X-rays of the light elementK-lines and the heavy element L-lines exist. The STJ array detector has an asymmetric layer structure so that a 300 nm-thick top Nb electrode efficiently absorbs photons of less than 2 keV at detection efficiencies between 20 and ∼100% [31]. Semiconductor detectors also operate in this soft X-ray region, but there are difficulties with an insufficient energy resolution and a surface dead layer. The most difficult case for semiconductor detectors is a trace light element embedded in the matrix containing a major element on the immediate left in the periodic table. The characteristic X-rays of the trace light element are buried in the skirt of the strong peak signal of the major element. STJ array detectors can be used to analyze such challenging cases. As shown in Fig. 3, the average energy resolution of the 100 STJ pixels is 14 eV for O-K 525 keV, which is five times better than that of the latest Si drift detector, SDD. At this energy resolution, the STJ detector can clearly separate the characteristic X-rays from B, C, N, and O, while in a semiconductor detector with a δE value of 48 eV there is peak overlap (Fig. 4). The spectrum measured by the asymmetric STJ array detector is ideal with no double peaks originating from the absorption events in the top and bottom Nb electrodes and almost no artifact events due to wiring and contact hole structure [31], [32].

Fig. 4. Comparison of the X-ray fluorescence spectra for a boron nitride sample with carbon and oxygen contamination, measured by an STJ detector (solid blue line) and calculated for an Si drift detector (dotted red line) with an energy resolution of 48 eV.

The STJ array detector was used to measure fluorescenceyield X-ray absorption fine structure (FY-XAFS) spectra of nitrogen dopant (300 ppm) in SiC at Beamline 11 of the High Energy Accelerator Research Organization, Photon Factory (KEK PF) [33]. FY-XAFS spectra are recorded by monitoring the yield of the K-line of nitrogen (N-K), as the SR beam energy is scanned in steps of 1–2 eV. Since the peak of Si-K (1.74 keV) is far from those of C-K (277 eV) and N-K (392 eV), no X-rays from Si are produced at SR energies near the C or N edges. However, it is very difficult to analyze an elemental combination of C and N, because the weak N-K peak of the trace N atoms is masked by the skirt of the strong C-K peak. There has been no report on XAFS measurement of the nitrogen dopant in SiC before. The energy resolution of the STJ array detector has made it possible to select the weak N-K peak. Fig. 5 shows the first successful FY-XAFS analysis of the N dopant in SiC. Comparison between the experimental XAFS spectra and the ab initio calculation by FEFF [34] reveals that the N dopant substitutes the C site; substitution of the Si site, which has been reported in the literature [35], is ruled out in our samples. We also analyzed nitrogen dopant in ZnO, and revealed the change of N sites with rapid thermal annealing [36]. The SC-XAFS instrument may play an important role in analyzing other materials containing trace light elements such as wide-gap semiconductors. B. Mass spectrometry (MS) for Different Charge States and Neutral Fragments MS is a method to analyze ions according to mass/chargenumber ratio (m/z is used in the MS community). MS instruments have three components: 1) an ion source that ionizes analytes (atoms or molecules of mass m) at a charge state z, which are accelerated by a voltage V (normally between 3 and 30 kV), 2) a temporal or spatial ion separation section according to m/z, and 3) an ion detector that converts the ion arrival to an electric signal. Each MS instrument is called by a specific name depending on the type of ionization and the type of ion separation. A traditional MS instrument is a


2400208

Fig. 5. Fluorescence yield X-ray absorption fine structure (FY-XAFS) spectra for nitrogen-ion-implanted SiC in the as-implanted state and after rapid thermal annealings [3]. Comparison between the experiment (a) and the ab initio calculation (b) reveals that the N dopant occupies the C site.

Fig. 6. Schematic diagram of matrix-assisted laser desorption ionization time of flight mass spectrometer (MALDI TOF MS).

combination of electron ionization and bending magnet separation: magnetic-sector MS. Another common type of MS instrument is the combination of matrix-assisted laser desorption ionization (MALDI) and time of flight (TOF) ion separation: MALDI TOF MS (Fig. 6). In MALDI TOF MS, analytes in a mixture with matrix molecules on a sample plate are desorbed from the surface and ionized by UV laser pulses, and then accelerated to an energy zV . The section for the ion separation in TOF is just a vacuum tube, in which the ions fly a length of l toward the ion detector. The TOF of each ion, which is the interval between a laser pulse shot and a detector output pulse, is expressed by m 1 l. (4) TOF = z 2 eV By using (4) and experimentally obtained TOF values, the m/z values of ions are determined. MALDI TOF MS instruments are widely used in life sciences for identifying biomolecules such as peptides, proteins, and protein complexes in bottom-up and top-down proteomics, and metabolomics. One of the latest techniques using MALDI TOF MS is molecular imaging [37]. One disadvantage of STJ detectors for the TOF MS application is a long dead time of ∼ μs. In modern high-

mass-resolution TOF MS, 1 ns pulse width is desirable. Superconducting strip ion detectors (SSIDs) are promising, because they can generate narrow output pulses of faster than 1 ns and valuable discrimination levels for different charge states [38], [39]. Nevertheless, STJ detectors have a time response of < 100 ns in pulse leading edge discrimination, so that they are acceptable for some TOF MS applications. In addition, the STJ performance fits well to magnetic-sector type MS instruments and ion storage rings. Interestingly, MS is unable to measure masses (m values) directly, because the response of ions to an electromagnetic force is determined by m/z. Uncertainty about the charge states is always present; the components of conventional MS instruments are unable to distinguish different charge states precisely. In addition to this uncertainty, analytes must have a charge, which means that MS is unable to analyze neutral atoms or molecules. Consequently, there are two inherent limitations in MS. The first one is that the peaks of ions with the same m/z but different charge states completely overlap in MS spectra. For example, it is impossible to distinguish between 14 N+ and 14 2+ N 2 (m/z 14). The second limitation is that neutral atoms or molecules produced in gas phase fragmentation cannot be analyzed, which is called “neutral loss.” STJ detectors can overcome both these limitations. Conventional ion detectors such as microchannel plates (MCPs) rely on the emission of one secondary electron upon ion impact from the detector surface with a low work function [9]. The number of electrons is multiplied from one secondary electron by a factor of 106 with an electron multiplier. The MCPs are unable to distinguish different charge states. Moreover, the detection efficiency of the MCPs decreases as the mass increases. The quantum efficiency is roughly proportional to the ion velocity, v = (2 zeV /m)1/2 , so the ion detection efficiency of conventional detectors is inversely proportional to m1/2 . On the other hand, the ion detection mechanism of STJ detectors is Cooper-pair breaking and sub-gap current increase,

2400208


Fig. 7. Mass spectrum of the air. Atoms and molecules in the air are identified 2+ have the same m/z of 14, and they from the m/z values. 14 N+ and 14 N2 completely overlap. The same overlap occurs for oxygen at m/z 16.

which is induced by the phonon excitation upon ion surface impact. STJ detectors measure the energy deposited on the top superconducting electrode, which ensures 100% detection efficiency independent of m, because the kinetic energy of ions is independent of m and equal to zV . Furthermore, it is possible to determine the charge states from the kinetic energy measurement. After the pioneering work on biomolecule detection in 1996 [40], STJ detectors were extensively applied to MS. The works before 1999 were reviewed in [8]. However, they focused on detection of large molecules, except one report on charge state discrimination using a normal-insulator-superconductor junction [41]. In this paper, we report on overcoming the two inherent limitations of MS. The typical MS spectrum of the air ionized by electron impact is shown in Fig. 7. Most of the peaks are identified from m/z values, but some ions are masked by other ions with the same m/z. In the ionization of the air, the 14 N+ 2 ions 2+ 14 (m/z 28) are naturally dominant. The N2 ions (m/z 14) may also be produced at a very small percentage, but there 2+ have been no reports on the presence of 14 N2 . It is expected that N2 molecules dissociate into atoms and a large amount of N+ ions (m/z 14) should be produced. Therefore, it is 2+ impossible to distinguish 14 N2 and 14 N+ with conventional MS spectrometers. A similar peak overlap occurs in oxygen at an m/z value of 16. The first direct analysis of N2 2+ was realized by the STJ detector as shown in Fig. 8 [10]. The scatter plot of m/z vs. kinetic energy (pulse height) in Fig. 8(a) shows the event groups at different charge states along the y axis. The separation between the singly-charged ions and the doubly-charged ions can be seen clearly in Fig. 8(b), which is a pulse height spectrum for the events at m/z 14. Two peaks are assigned to N+ and N2 2+ from the peak positions on the y axis. Doubly-charged diatomic ions are investigated in planetary science, and may play an important role in the escape of atmosphere from planets [42]. Therefore, the detection and ionization cross-section measurement of doubly-charged diatomic ions on the ground is the forerunner of space instrumentation. Similar m/z peak

overlapping is found in oligomerization of such biomolecules as lysozyme and immunoglobulin [38]. The charge state discrimination also simplifies mass spectra of multiply-charged protein ions produced by electrospray ionization: for example, fragment ions of non-covalent complexes such as hemoglobin [43]. The ability of STJ detectors to measure kinetic energy also enables different neutral molecules to be analyzed. In tandem mass spectrometry (so-called MS/MS), m/z-selected ions are broken into fragments by gas phase atomic collision or electron transfer, which produces metastable ions. Fragmentation patterns are used to identify the precursor ions or analyze their molecular structures. The fragments often include both of ions and neutral molecules. The neutral fragments cannot be analyzed in conventional MS, and are lost from an ion trajectory. This is called “neutral loss.” In order to solve the neutral loss problem, we installed an STJ detector into a magnetic sector type MS instrument [11]. In this setup, m/z-selected singly-charged ions interact with a target atom of Xe, and are excited by molecular collision or electron transfer. The metastable singly-charged ions or neutral molecules dissociate into neutral fragments as well as ionic fragments through collisionally-activated dissociation (CAD) or electron transfer dissociation (ETD). Especially, the analysis of reaction branches in ETD of singly-charged precursor ions is very tough, since all fragments are neutral. It is possible to use re-ionization of neutral fragments and then analyze them with conventional MS, but there is a possibility of additional fragmentation on re-ionization. STJ detectors allow direct mass analysis of the neutral fragments without re-ionization, taking into account kinetic energy conservation and partition of the kinetic energy of the precursor ions into fragments. The neutral or ionic fragments have kinetic energies depending on the m values. The sum of the kinetic energies is equal to that of the precursor ions. Therefore, STJ detectors can distinguish different neutral fragments. In this experiment, the branching ratios of dissociation of neutral acetone are identified from kinetic energy spectra: for example, (CH3 )2 CO → CH3 CO + CH3 and (CH3 )2 CO → CH3 + CH3 + CO. These reactions produce the neutral radicals, which are thought to be sources of pre-biotic organic molecules in interstellar space through dissociative recombination [44]. IV. C ONCLUSION STJ detectors are versatile devices for analytical sciences and outperform conventional detectors based on semiconductor technology. The recent progress in two STJ applications was discussed: XAFS and MS. In XAFS, the analysis of atomicscale local structure of trace light element impurities in matrices composed of light elements has been realized. In addition to the soft X-ray region of less than 2 keV, it is necessary to cover X-rays up to 10 keV for future analyses in synchrotron radiation facilities. In MS, STJ detectors can overcome two inherent limitations of conventional MS instruments. Ions with the same m/z but different charge states can be separated by using the ability of STJ detectors to measure kinetic energy, which also solves


2400208

Fig. 8. Confirmation of the existence of N2 2+ ions by mass spectrometry with the STJ detector. The simultaneous measurement of m/z and kinetic energy of 2+ each ion produced by ionizing the air is shown in (a). The events at m/z 14 only are plotted in (b), confirming the first successful detection of 14 N2 without 14 + disturbance of the N ions. The figure is based on [10].

the problem of neutral loss. In addition, gas phase reaction branches that create neutral products only can be revealed by STJ detectors, which may thus yield new insights in gas phase ion reaction experiments using electrostatic storage rings, for investigating dissociative electron recombination. One weakness of current STJ array detectors is the small sensitive area compared with that of semiconductor detectors. Since our microfabrication technology, cryogenics, and signal processing can be used to produce instruments with five times more STJ pixels, we aim to develop a sensitive area of ∼cm order. ACKNOWLEDGMENT The authors thank all members of the Superconducting Spectroscopy Group of AIST, the staff of the superconducting device fabrication facility called CRAVITY, and Dr. Y. Kitajima and the staff of KEK PF. Special thanks are due to Prof. S. Hayakawa and Dr. S. Tomita for fruitful discussion and MS experiments. R EFERENCES [1] P. Lerch and A. Zehnder, “Quantum Giaever detectors: STJ’s,” in Cryogenic Particle Detection, C. Enss, Ed. Berlin, Germany: SpringerVerlag, 2005, pp. 217–266. [2] M. Hidaka, S. Nagasawa, K. Hinode, and T. Satoh, “Improvements in fabrication process for Nb-based single flux quantum circuits in Japan,” IEICE Trans. Electron., vol. E91C, no. 3, pp. 318–324, Mar. 2008. [3] M. Ohkubo et al., “X-ray absorption near edge spectroscopy with a superconducting detector for nitrogen dopants in SiC,” Sci. Rep., vol. 2, pp. 831-1–831-5, Nov. 2012. [4] D. Schlosser, “Expanding the detection efficiency of silicon drift detectors,” Nucl. Inst. Methods in Phys. Res. A, vol. 624, no. 2, pp. 270–276, Apr. 2010. [5] J. J. Rehr and R. C. Albers, “Theoretical approaches to X-ray absorption fine structure,” Rev. Mod. Phys., vol. 72, pp. 621–654, Sep. 2000, For example. [6] S. Friedrich, “Cryogenic X-ray detectors for synchrotron science,” J. Synchroton Radiat., vol. 13, pp. 159–171, Feb. 2006. [7] O. B. Drury and S. Friedrich, “Sensitivity and S/N-ratio of superconducting high-resolution X-ray spectrometers,” IEEE Trans. Appl. Supercond., vol. 15, no. 2, pp. 613–617, Jun. 2005. [8] M. Frank, S. Labov, G. Westmacott, and W. Benner, “Energy-sensitive cryogenic detectors for high-mass biomolecule mass spectrometry,” Mass Spectrom Rev., vol. 18, no. 3/4, pp. 155–186, Aug. 1999.

[9] D. W. Koppenaal et al., “MS detectors,” Anal. Chem., vol. 77, no. 21, pp. 418A–427A, Nov. 2005. [10] S. Shiki et al., “Kinetic-energy-sensitive mass spectrometry for separation of different ions with the same m/z value,” J. Mass Spectrom., vol. 43, no. 12, pp. 1686–1691, Dec. 2008. [11] M. Ohkubo, S. Shiki, M. Ukibe, S. Tomita, and S. Hayakawa, “Direct mass analysis of neutral molecules by superconductivity,” Int. J. Mass Spectrom., vol. 299, no. 2/3, pp. 94–101, Jan. 2011. [12] H. Pressler, M. Ohkubo, M. Koike, M. Ukibe, and T. Zama, “Experimental imaging diagnosis of superconducting tunnel junction X-ray detectors by low-temperature scanning synchrotron microscope,” Appl. Phys. Lett., vol. 77, no. 4, pp. 4055–4057, Dec. 2000. [13] M. Ukibe, T. Ikeuchi, T. Zama, and M. Ohkubo, “Aluminum thickness dependence of spatial profile in niobium-based superconducting tunnel junctions,” Nucl. Instrum. Methods Phys. Res. A, vol. 520, no. 1–3, pp. 260–262, Mar. 2004. [14] R. Cristiano et al., “Quasiparticle diffusion edge losses back-tunneling in superconducting tunnel junctions under X-ray irradiation,” J. Appl. Phys., vol. 86, no. 8, pp. 4580–4587, Oct. 1999. [15] A. Kozorezov et al., “Quasiparticle diffusion and the energy resolution of superconducting tunneling junctions as photon detectors. I. Theory,” Phys. Rev. B, vol. 66, pp. 094510-1–094510-14, Sep. 2002. [16] S. Bechstein, B. Beckhoff, R. Fliegauf, J. Weser, and G. Ulm, “Characterization of an Nb/Al/AlOx /Al/Nb superconducting tunnel junction detector with a very high spatial resolution in the soft X-ray range,” Spectrochimica Acta B, vol. 59, no. 2, pp. 215–221, Feb. 2004. [17] H. Pressler et al., “Spatially resolved study of superconducting tunnel junctions X-ray detectors by low temperature scanning synchrotron microscopy,” IEEE Trans. Appl. Super., vol. 11, no. 1, pp. 696–699, Mar. 2001. [18] A. Poelaert, A. Kozorezov, A. Peacock, and J. Wigmore, “Strong nonlinear response of superconducting tunnel junctions due to localized traps,” Phys. Rev. Lett., vol. 82, pp. 1257–1260, Feb. 1999. [19] M. Ohkubo et al., “Asymmetric response of superconducting niobiumtunnel-junction X-ray detectors,” Phys. Rev. B, vol. 54, pp. 9484–9490, Oct. 1996. [20] M. Kurakado, “Possibility of high resolution detectors using superconducting tunnel junctions,” Nucl. Instrum. Methods, vol. 196, pp. 275–277, May 1982. [21] N. Rando et al., “The properties of niobium superconducting tunneling junctions as X-ray detectors,” Nucl. Instrum. Meth. in Phys. Res. A, vol. 313, no. 1/2, pp. 173–195, Mar. 1992. [22] A. Peacock et al., “Single optical photon detection with a superconducting tunnel junction,” Nature, vol. 381, pp. 135–137, May 1996. [23] N. Rando et al., “S-Cam: A spectrophotometer for optical astronomy: Performance and latest results,” Rev. Sci. Instrum., vol. 71, no. 12, pp. 4582– 4591, Dec. 2000. [24] C. A. Mears, S. E. Labov, and A. T. Barfknecht, “Energy-resolving superconducting X-ray detectors with charge amplification due to multiple quasiparticle tunneling,” Appl. Phys. Lett., vol. 63, no. 21, pp. 2961–2963, Nov. 1993. [25] D. J. Goldie, P. L. Brink, C. Patel, N. E. Booth, and G. L. Salmon, “Statistical noise due to tunneling in superconducting tunnel junction detectors,” Appl. Phys. Lett., vol. 64, no. 23, pp. 3169–3171, Jun. 1994.

2400208


[26] K. Segall et al., “Noise mechanisms in superconducting tunnel-junction detectors,” Appl. Phys. Lett., vol. 76, no. 26, pp. 3998–4000, Jun. 2000. [27] M. Ohkubo, “Superconductivity for mass spectroscopy,” IEICE Trans. Electron., vol. E90-C, no. 3, pp. 550–555, Mar. 2007. [28] M. Ohkubo, “Superconducting detectors for particles from atoms to proteins,” Physica C, vol. 468, no. 15–20, pp. 1987–1991, Sep. 2008. [29] S. Bandler et al., “Advances in small pixel TES-based X-ray microcalorimeter arrays for solar physics and astrophysics,” IEEE Trans. Appl. Supercond., vol. 23, no. 3, p. 2100705, Jan. 2013. [30] S. Shiki, M. Ukibe, Y. Kitajima, and M. Ohkubo, “Fluorescence yield X-ray absorption spectroscopy for O-K α with elimination of N-K α background using superconducting tunnel junction detectors,” Physica C, vol. 471, no. 21, pp. 1263–1266, Nov. 2011. [31] M. Ukibe, S. Shiki, Y. Kitajima, and M. Ohkubo, “Soft X-ray Detection performance of superconducting tunnel junction arrays with asymmetric tunnel junction layer structure,” Jpn. J. Appl. Phys., vol. 51, no. 1R, pp. 010115-1–010115-5, Jan. 2012. [32] M. Ohkubo et al., “Double peak phenomenon in superconducting tunnel junction X-ray detectors,” J. Appl. Phys., vol. 85, no. 1, pp. 595–599, Jan. 1999. [33] Photon Factory, Jan. 30, 2014. [Online]. Available: http://pfwww.kek.jp/ index.html [34] A. Ankudinov, A. Nesvizhskii, and J. Rehr, “Dynamic screening effects in X-ray absorption spectra,” Phys. Rev. B, vol. 67, pp. 115120-1–115120-6, Mar. 2003. [35] J. Hartman et al., “NMR studies of nitrogen doping in the 4H polytype of silicon carbide: Site assignments and spin-lattice relaxation,” J. Phys. Chem. C, vol. 113, no. 33, pp. 15024–15036, Aug. 2009. [36] P. Fons et al., “Direct observation of nitrogen location in molecular beam epitaxy grown nitrogen-doped ZnO,” Phys. Rev. Lett., vol. 96, pp. 0455041–045504-4, Feb. 2006. [37] J. M. Perkel, “Life science technologies: Mass spec imaging: From bench to bedside,” Science, vol. 340, no. 6136, pp. 1119–1121, May 2013. [38] K. Suzuki et al., “Charge-state-derivation ion detection using a superconducting nanostructure device for mass spectrometry,” Rapid Commun. Mass Spectrom., vol. 24, no. 22, pp. 3290–3296, Oct. 2010. [39] N. Zen, S. Shiki, M. Ukibe, M. Koike, and M. Ohkubo, “Ion-induced dynamical change of supercurrrent flow in superconducting strip ion detectors with parallel configuration,” Appl. Phys. Lett., vol. 104, no. 1, pp. 012601-1–012601-5, Jan. 2014. [40] D. Twerenbold et al., “Detection of single macromolecules using a cryogenic particle detector coupled to a biopolymer mass spectrometer,” Appl. Phys. Lett., vol. 68, no. 24, pp. 3503–3505, Apr. 1996. [41] G. Hilton et al., “Impact energy measurement in time-of-flight mass spectrometry with cryogenic microcalorimeters,” Nature, vol. 391, pp. 672– 675, Feb. 1998. [42] R. Thissen et al., “Doubly-charged ions in the planetary ionospheres: A review,” Phys. Chem. Chem. Phys., vol. 13, no. 41, pp. 18264–18287, Sep. 2011. [43] M. Ohkubo et al., “Superconducting tunnel junction detectors for mass spectrometry,” Progr. Supercond., vol. 14, pp. 77–81, Dec. 2012. [44] L. Snyder, “Interferometric observations of large biologically interesting inter-stellar and cometary molecules,” Proc. Nat. Acad. Sci. USA, vol. 103, no. 33, pp. 12243–12248, Aug. 2006.

Masataka Ohkubo (M’13) received the M.S. degree in 1983 and the Ph.D. degree in 1991 in electrical and electronic engineering from Toyohashi University of Technology, Toyohashi, Japan. He worked with Toyota Central R&D Labs., Inc., Nagoya, Japan, from 1983 to 1993, and joined Electrotechnical Laboratory (ETL) in 1993. From 1994 to 1996, he was a Guest Researcher at Karlsruhe Research Center, Germany. His research interests include the epitaxial growth of high-temperature superconductors, ion beam analysis, and analytical sciences with superconductivity. He was appointed as the Director of Research Institute of Instrumentation Frontier (RIIF), National Institute of Advanced Industrial Science and Technology (AIST) in 2011. His current position is the Supervisory Innovation Coordinator in Tsukuba Innovation Arena (TIA). His other interests are superconducting strip detectors (SSDs) and single-flux quantum (SFQ) read-out circuits.

Shiki Shigetomo received the M.S. degree in astronomy and the Ph.D. degree in nuclear engineering and management from the University of Tokyo, Tokyo, Japan, in 1997 and 2009, respectively. He is currently a Researcher in National Institute of Advanced Industrial Science and Technology (AIST). He has been pursuing instrumentation of XAFS and MS with superconducting detectors since 2001.

Masahiro Ukibe received the M.S. and Ph.D. degrees in nuclear engineering from the University of Tokyo, Tokyo, Japan, in 1995 and 1998, respectively. He joined Electrotechnical Laboratory (ETL) in 1998. From 2004 to 2005, he was a Guest Researcher at le Centre de Spectrométrie Nucléaire et de Spectrométrie de Masse, France, devoted to the development of bolometer array detectors for cosmic microwave background (CMB). Since 2007, he has been senior researcher at the Research Institute of Instrumentation Frontier (RIIF), AIST. His research interests include developing fabrication processes for superconducting detectors.

Go Fujii received the M.S. degree and the Ph.D. degree in quantum information from Nihon University, Tokyo, Japan, in 2009 and 2012, respectively. He joined National Institute of Advanced Industrial Science and Technology (AIST) in 2012 as a Postdoctoral Staff, and was appointed as a Researcher in 2013. His research interests include the fabrication of three-dimensional advanced superconducting devices.

Nobuyuki Matsubayashi received the Ph.D. degree in inorganic and physical chemistry from Osaka University, Osaka, Japan, in 1986. He is currently a Principal Research Manager with National Institute of Advanced Industrial Science and Technology (AIST). His research interests include analyses of materials by XAFS and XPS.