© MICROSCOPY SOCIETYOF AMERICA 2015 Atom-Probe …€¦ · Atom-Probe Tomographic Analyses of Hydrogen Interstitial Atoms in Ultrahigh Purity Niobium Yoon-Jun Kim1 and David N. Seidman1,2,*

Microsc. Microanal. 21, 535–543, 2015doi:10.1017/S143192761500032X

© MICROSCOPY SOCIETYOF AMERICA 2015

Atom-Probe Tomographic Analyses of HydrogenInterstitial Atoms in Ultrahigh Purity NiobiumYoon-Jun Kim1 and David N. Seidman1,2,*

1Department of Materials Science and Engineering, Northwestern University, 2220 Campus Dr., Evanston, IL 60208, USA2Northwestern University Center for Atom-Probe Tomography (NUCAPT), 2220 Campus Dr., Evanston, IL 60208, USA

Abstract: Atomic-scale characterization of hydrogen and formation of niobium hydrides, using ultraviolet(wavelength = 355 nm) picosecond laser-assisted local-electrode atom-probe tomography, was performed forultrahigh purity niobium utilizing different laser pulse energies, 10 or 50 pJ/pulse or voltage pulsing. At 50 pJ/pulse, hydrogen atoms migrate onto the 110 and 111 poles as a result of stimulated surface diffusion, whereas theyare immobile for <10 pJ/pulse or for voltage pulsing. Accordingly, the highest concentrations of H and NbH wereobtained at 50 pJ/pulse. This is attributed to the thermal energy of the laser pulses being transferred to pureniobium specimens. Therefore, we examined the effects of the laser pulse energy being increased systematicallyfrom 1 to 20 pJ/pulse and then decreasing it from 20 to 1 pJ/pulse. The concentrations of H, H2, and NbH and theatomic concentration ratios H2/H, NbH/Nb, and Nb3+/Nb2+ were calculated with respect to the systematicallychanging laser pulse energies. The atomic concentration ratios H2/H and NbH/Nb are greater when decreasingthe laser pulse energy than when increasing it, because the higher residual thermal energy after decreasing thelaser pulse energy increases the mobility of H atoms by supplying sufficient thermal energy to form H2 or NbH.

Key words: atom-probe tomography, hydrogen interstitial atoms, hydrides, niobium, metallic superconductors

INTRODUCTION

Niobium is a type-II superconducting metal possessing a cri-tical temperature (Tc) of 9.8 K. The high-energy physicscommunity is developing superconducting radio-frequency(SRF) cavities using ultrahigh purity (ASTM B 393, Type 5Grade) niobium to enable the exploration of energy andintensity frontiers employing linear accelerators (Phinneyet al., 2007). The central issue concerning niobium SRF cav-ities is their lack of reproducible performance for large-scaleproduction, which means their yield is low and hence theircost for using them for linear accelerators becomes extremelyexpensive. One of the symptoms of poor performance at thehigh electric-field gradients, 35MV/m, specified by the Inter-national Linear Collider project and a next-generation accel-erator project, Project X, is the appearance of large pits at theedges of the heat-affected zones near the equator welds (Geet al., 2011). Typical pit dimensions are 0.1–1.0 mmwidth and10–100mm depth (Cooley et al., 2011). In particular, badcavities exhibit a strong decrease in the quality factor (Q),which is the ratio of stored energy to dissipated power that isdiminished by the existence of chemical impurities and/or arough surface finish (Padamsee, 2009). Therefore, under-standing and preventing weld pits have received a great deal ofjustifiable attention (Champion et al., 2009; Cooley et al., 2011;Singer et al., 2011). Although the exact formation mechanismof weld pits is not fully understood, it is believed that niobiumhydrides play a significant role.

Several studies were performed on the sub-surfacechemistry of SRF cavities, especially on the behavior of Hand the formation of NbH in Nb coupons, on an atomic scale,using an ultraviolet (UV) laser-assisted local-electrode atom-probe (LEAP) tomograph in conjunction with aberration-corrected high-resolution transmission electron microscopy(Kim et al., 2013). In addition, secondary ion mass spectro-metry (Ciovati et al., 2010; Maheshwari et al., 2011) and elasticrecoil detection (Romanenko & Goncharova, 2011) techni-ques were utilized to obtain quantitative depth profiles ofH and NbH in Nb-SRF cavity specimens. Oxygen and oxideformation on pure Nb surfaces were studied utilizing X-raydiffraction techniques (Ma & Rosenberg, 2003; Delheusy et al.,2008). Earlier, LEAP tomographic studies were successfullyemployed for studying the formation of niobium oxides onNb surfaces (Sebastian et al., 2006; Yoon et al., 2007, 2008)

The behavior of hydrogen interstitial atoms is of para-mount importance because niobium has a body-centeredcubic (b.c.c.) structure containing octahedral and tetrahedralinterstitial sites, and hydrogen atoms can migrate randomlyon them and form different hydrides (NbHx) (Schober,1975a, 1975b). According to the Nb–H phase diagram(Fig. 1; Ricker & Myneni, 2010), hydrogen dissolves exother-mically in niobium producing solid solutions of hydrogenin Nb denoted as α- or α'-phases. It is emphasized that theα'-phase contains H in pure Nb up to ~70 at% above ~180°Cand the α-phase co-exists with the β-NbH phase and ε-NbHphase at a H concentration of up to ~43 at% at room tem-perature and 30 K, respectively. The interactions amongNb and H atoms involve short-range ordering, therebycreating a hydrogen superlattice (Makenas & Birnbaum, 1982).*Corresponding author. [email protected]

Received May 15, 2014; accepted February 16, 2015

Above room temperature, the β-NbH phase (face-centeredorthorhombic, f.c.o.) can be formed from α– and/or α'-solidsolutions by the ordering of H atoms in the tetrahedralpositions in an f.c.o. structure. Employing conventionaltransmission electron microscopy and selected-area diffrac-tion patterns indicate that hydrogen ordering involves thetetrahedral sites in β-NbH (f.c.o.), which had precipitatedfrom the α'-solid solution (Schober et al., 1973, 1974;Birnbaum et al., 1976; Schober & Wenzl, 1976; Grossbeck &Birnbaum, 1977; Gabriel et al., 2011). The β-NbH phase isthe most common one observed in Nb, and its H/Nb atomicratio (ratio of the atomic concentrations of hydrogen andniobium) ranges from 0.7 to 1.1 (Gabriel et al., 2011). Attemperatures <230 K, different ordered hydrides, ε (orthor-hombic) ζ (orthorhombic), γ (pseudo cubic), and λ (uncon-firmed), can be formed from the β-phase (Makenas &Birnbaum, 1982; Ricker & Myneni, 2010). The formation ofordered β-NbH is accelerated by plastic deformation. Theinteractions among hydrogen atoms and dislocationsincrease the binding enthalpy of an H–H dimer from 1,265(0.05485 eV) to 7,300 cal/mol (0.3166 eV) (Baker & Birnbaum,1972). The net result is that the diffusivity of an H–H dimerdecreases from 10− 14 cm2/s in Nb (Baker & Birnbaum, 1973)to 10− 32 cm2/s at 60 K. Thereby, effectively immobilizinghydrogen atoms via their interactions with dislocations resultsin β-NbH precipitates at 60 K.

In this article, we present hydrogen and niobium hydrideformation behavior using a picosecond UV laser-assistedLEAP tomograph for different UV laser pulse energies rangingfrom 1 to 50 pJ/pulse. Hydrogen concentration profiles andthree-dimensional (3D) mapping of H and NbH in pure Nbare determined for different laser pulse energies and voltagepulsing. In addition, the following quantities were determinedand discussed: (1) mass-resolving power (m/Δm); (2) the ratioof charge states of Nb ions (Nb3+/Nb2+ ); and (3) the ratiosof different hydrogen compounds (H2/H and NbH/Nb) inmole percent.

EXPERIMENTAL TECHNIQUES

Analyses were performed utilizing laser energies rangingfrom 1 to 50 pJ/pulse. The remaining parameters were heldconstant: UV wavelength λ = 355 nm; evaporation rate(ions/pulse) = 0.50%; a pulse repetition rate of 250 kHz; anda niobium nanotip temperature of 30 K. In addition, anexperiment was performed using the voltage pulsing modewith a pulse fraction (pulse voltage to standing DC voltage)of 0.20, and a pulse repetition rate of 100 kHz.

A sample of 1 cm diameter, cut from an SRF single-cellcavity (Romanenko & Padamsee, 2010), was supplied by theFermi National Accelerator Laboratory (Batavia, IL, USA).LEAP tomographic specimens were prepared using an FEIHelios dual-beam focused ion beam (FIB) microscope (HeliosNanolab, FEI Co., Hillsboro, Oregon, USA). Specimens with ananotip radius of ~20 nm were fabricated using the FIB-basedlift-out method and attached to Si microposts on a coupon(Miller et al., 2007; Seidman, 2007a, 2007b; Seidman & Stiller,2009). The coupon was then inserted into the LEAP tomo-graph’s ultrahigh vacuum (UHV) chamber and cooled to 30 Kbefore the pulsed laser-assisted evaporation analyses. The LEAPtomograph’s UHV chamber was maintained at ~2× 10−11

Torr, and the partial pressures of residual gases were monitoredcontinuously during the course of each experiment using aresidual gas analyzer (RGA) (Inficon Transpector 2, InficonHolding AG, Bad Ragaz, Switzerland). The dominant partialpressures were PH2O¼ 5:5 ´ 10 - 13 and PH2¼ 2:7 ´ 10 - 13 Torr,which are approximately two orders of magnitude less than thetotal pressure, ~2× 10−11 Torr in the UHV chamber.

RESULTS AND DISCUSSION

Mass spectra obtained utilizing different laser pulse energies,50, 10 pJ/pulse, or voltage pulsing are displayed in Figure 2.Eight different ionic species, 1H+, 1H2

+, 16O+, 93Nb3+ , 93Nb2+ ,93Nb1H2+ , 93Nb16O2+ , and 69Ga+ (the latter is due to theutilization of an FEI Helios dual-beam FIB microscope formilling with gallium ions), were typically obtained, inde-pendent of the laser pulse energies. We note that the ionicspecies 1H+, 1H2

+, and 93Nb1H2+ are from exothermicallydissolved H in the Nb lattice. We ignored possible interac-tions between the residual H2 gas in the UHV chamber and apure Nb sample for the following reasons. (1) As a nanotip isbiased at a positive potential during atom-probe tomo-graphic analyses and as all ions are evaporated as positivelycharged ions all field-evaporated ions including positivelycharged residual H2 molecules should be repelled and notattracted to a nanotip. (2) The only possible way to attractthe residual H2 gas to the nanotip is a result of their polar-izability in the high electric fields involved. The latter doesnot occur in our APT analyses because the partial pressure ofhydrogen needs to be significantly greater than the partialpressure of hydrogen in an UHV chamber, 2.7 × 10− 13 Torrin our UHV chamber, as reported above (Ast & Seidman,1968, 1971; Averback & Seidman, 1973). The flux of hydrogen(F) to a Nb nanotip is given by the classic gas kinetic factor

Figure 1. Nb–H binary phase diagram: temperature versusatomic fraction hydrogen (%).

536 Yoon-Jun Kim and David N. Seidman

times an enhancement factor, owing to the polarizability ofH2, that determines the total flux of hydrogen atoms to thesurface of a nanotip (Brandon, 1968; Southon, 1968), asexpressed by the following equation:

F ¼ F0 ´32

αE2

2kBT

� �; (1)

where α is the polarizability of H2, E the electric field, kB theBoltzmann’s constant, T the temperature (in Kelvin), and F0the number of hydrogen molecules/m2/s impinging on ananotip in the absence of E. The classic gas kinetic factor forthe flux of H2 molecules that arrive at the surface of thenanotip is given by:

F0 ¼ pηffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2πmkBT

p ; (2)

where p is the partial pressure of H2, η the sticking coefficientdescribing the fraction of incident gas molecules that adsorbon the surface of a nanotip, and m the molecular mass of H2.The quantity inside of parenthesis in equation (1) is calledan enhancement factor owing to the polarizability of specificgas species. The following values apply to our experiments:aH2 ¼ 5:314 a:u: (Olney et al., 1997), E = 35 V/nm for Nb(Tsong, 1978, 1990), PH2 ¼ 2:7 ´ 10 - 13 Torr, T = 30 K, andη = 1 [this yields an upper bound value for equation (2)].These values yield F = 2.09 × 1014 molecules/m2/s for the H2

molecule flux to a surface for η = 1. Consequently, the timefor monolayer coverage is ~47,816 s (~13.28 h), which issignificantly longer than our APT runs, which are typically<4–5 h. The above calculation demonstrates that the effectswe report in this article cannot be because of the residualH2 gas in the UHV chamber. In addition, when hydrogenpromotion is utilized the nanotip is continuously evaporating(Ast & Seidman, 1968, 1971; Averback & Seidman, 1973),

which is not the case for our studies of Nb, where the nanotipsare stable when they are not being pulsed with UV light or avoltage.

Quantitative concentration profile of each element wasobtained from a 3D reconstruction (Fig. 3). Nb, H, and NbHwere the predominant species detected, but oxygen andniobium oxides (NbO and/or Nb2O5) were also detectedat a maximum concentration of ~45 mol%Nb2O5 for the UVlaser pulsing mode using 50 (Fig. 3a) and 10 pJ/pulse(Fig. 3b) and ~5 mol% NbO for the voltage pulsing mode(Fig. 3c) within 5 nm from the surface. The behavior of Hand NbH depended on the UV laser pulse energies used. At adepth of ~5 nm in the oxide layer, the concentration of Hincreased continuously with increasing UV laser pulse energy.For 10 pJ/pulse (Fig. 3b) and voltage pulsing (Fig. 3c), the

Figure 3. Concentration profiles of H, Nb, O, NbO, Nb2O5, NbH,and Ga obtained by analyzing Nb with: (a) 50 pJ/pulse; (b) 10 pJ/pulse; and (c) voltage pulsing at a pulse repetition rate of 100 kHz.

Figure 2. Mass spectra obtained from atom-probe tomographicanalyses from: (a) 50 pJ/pulse with a picosecond ultraviolet laserpulse repetition rate of 250 kHz; (b) 10 pJ/pulse with a picosecondUV laser pulse repetition rate of 250 kHz; and (c) voltage pulsingwith a pulse repetition rate of 100 kHz.

Hydrogen interstitials in pure Nb 537

concentration profiles of Nb and H stabilized at averageconcentrations of 44.6 mol% Nb, 26.4 mol% H, and 36.5 mol% Nb, 61.4 mol% H, respectively. In addition, the amount ofNbH detected using 50 pJ/pulse was ~43 mol%, but it isreduced significantly to ~3 mol% for voltage pulsing and to~25 mol% for 10 pJ/pulse.

The behavior of Nb, H, and NbH are best understood byconstructing 2D cross-sectional views of the 3D reconstruc-tions. The field-desorption image of Nb atoms contains crys-tallographic information, poles, and zone lines (Fig. 4a). Thiscontrast is attributed to local electric-field variations on thesurface of a nanotip (Waugh et al., 1976; Gault et al., 2009).Assuming a stereographic projection, the poles are identifiedfrom the symmetry of the image and are accordingly indexed(Fig. 4j; Drechsler & Wolf, 1960; Brandon, 1964).

When the UV laser pulse energy is increased to 50 pJ/pulse, the poles disappear in the Nb-desorption image (Fig. 4c)and concomitantly H decorates the zone lines of Nb (Fig. 4f).This implies that there are high binding energy sites and fastsurface diffusion pathways, which are dependent on thecrystallographic facets along these zone lines at the highestUV laser pulse energy; this was similar for results obtainedusing laser-assisted LEAP tomography to analyze pure Al

(Gault et al., 2010) and field ion microscopy (FIM) ofW usingvoltage pulsing (Brandon, 1962; Moore & Spink, 1974). TheH-enriched zone lines displayed in Figure 4f almost disappearat 10 pJ/pulse (Fig. 4e), and when using voltage pulsing(Fig. 4d), owing to a reduction of the number density of Hatoms. Figure 4f displays clearly the threefold zone linedecoration with H at the 1 11 pole toward the 110-type poles,such as the 101 and 110 poles, and sixfold symmetry at the 110pole toward 111-type poles, such as the 1 11 and 111 poles.The migration of H atoms implies an interaction with itsformer nearest-neighbor (NN) Nb atoms at the edge of a planeas a result of different lattice binding energies, and thereforethemigration energy of H atoms is a function of the number ofNN atoms: see ball model (Waugh et al., 1976). This ballmodel suggests that the majority of atoms (atoms on {110}planes, except for kink-site atoms) can be evaporated from ananotip by migrating toward {110}-type planes, while main-taining two first NN atoms. However, kink-site atoms on<111> zone lines have only one NN atom, which are resistantto field evaporation of W atoms, as determined by FIMobservations. The larger the picosecond laser pulse energythe higher is the nanotip’s temperature, which results inatomic rearrangements on different crystallographic facets, as

Figure 4. Top views of Nb, H, and NbH ions in nanotips: (a) Nb (voltage pulsing); (b) Nb (10 pJ/pulse); (c) Nb(50 pJ/pulse); (d) H (voltage pulsing); (e) H (10 pJ/pulse); (f) H (50 pJ/pulse); (g) NbH (voltage pulsing); (h) NbH(10 pJ/pulse); and (i) NbH (50 pJ/pulse). Nb = brown, H = yellow, and NbH = red. Each individual dot correspondsto an individual atom. j: Stereogram with the relevant poles indicated in (a–i).

538 Yoon-Jun Kim and David N. Seidman

determined by FIM (Brandon, 1962). A field-emissionmicroscope (Gomer, 1961) study of H on W nanotipsdemonstrated that H atoms migrate from the closest-packedplane, {110}, to an atomically rough surrounding surfaceconsisting of terraces, for example, on the 110–211 zones. Thisindicates that low-impedance pathways, with small activationenergies for the diffusion of H atoms, provide the maximummobility of H atoms and the smallest binding energies.

When voltage pulsing is utilized, hydrogen atoms do notdiffuse between poles and are desorbed essentially directlyfrom crystallographic poles. Figure 5 displays a 50 at%H isoconcentration surface obtained using voltage pulsing,which demonstrates that the 100, 001, 110, 101, 111, 1 11, and121 crystallographic poles of Nb are enriched with hydrogen.

Figures 4g to 4i display the distributions of NbH mole-cules obtained using the voltage pulse mode and differentpicosecond UV laser pulse energies, 10 and 50 pJ/pulse,respectively. First, using the voltage pulsing mode (Fig. 4g)NbH is distributed around the 110 pole. Second, picosecondUV laser pulsing produces more NbH molecules throughoutthe Nb nanotips, as displayed in Figures 4h and 4i for 10 and50 pJ/pulse, respectively; the desorption images exhibit zonelines from 001 to 1 1 1 poles and from 001 to 100 poles.The differences in the NbH concentration profiles (Fig. 3)depend on the picosecond UV laser pulse energy, whichdetermines the nanotip temperature, and electric-fieldpulsing (voltage pulsing), which clearly does not affect thenanotip temperature.

Several studies indicate that hydride formation is causedby external forces (the so-called Maxwell stresses); specifi-cally, the mechanical stresses generated by the electric field,E, exerted on a nanotip (Panofsky & Phillips, 2005):

σ¼ ε02E2; (3)

where σ is the negative hydrostatic stress caused by E, and ε0the vacuum permittivity of free space. The evaporation field ofniobium is 35 V/nm (Tsong, 1978, 1990); therefore, the Max-well stress, σ, is −5.42 GPa; the transverse elastic constant(C’ = 0.5(C11–C22)) of pure Nb at 30 K is ~59 GPa (Simmons& Wang, 1971). Akaiwa & Wada (1990) demonstrated, usingFIM, that the formation of NbH is temperature- and E-fielddependent, thereby showing that hydrogen atoms on 111 polestend to rearrange themselves, subsequently producing anexpansion of the nanotip for a stress >1 GPa. Similar findingsexist for the 111 and 001 poles of GaP (Gaussmann et al., 1992)and 111 poles of Si (Sakata & Block, 1982). The presence ofhydrogen is, thus, a precursor to the E-field-induced formationof surface hydrides on specific hkl poles.

Besides the fact that the distribution of H atoms andformation of NbH molecules are related to the hkl poles ofNb, the concentrations of H and NbH are highly correlatedwith the picosecond UV laser pulse energy (Fig. 3). As fieldevaporation is a thermally activated process governed bythe Boltzmann factor, with an activation energy that is afunction of the E-field (Brandon, 1963, 1965; Kellogg, 1983),hydrogen is thermally desorbed with each UV laser pulse.

The higher the picosecond pulse energy, the more copious isthe thermally activated hydrogen desorption. The desorbed1H+ and 1H2

+ ions interact with the surface Nb atoms toform NbH. In addition, hydrogen field desorption in

Figure 5. Isoconcentration surface of 50 at% H:(a) top view and (b) side view.

Hydrogen interstitials in pure Nb 539

combination with field-induced hydride formation havebeen observed using pulsed laser APT studies of Si (Kellogg,1983) and Mo (Kellogg, 1981).

More systematic APT experiments were performed toverify the effects of picosecond UV laser pulse energy(thermal energy) on H interstitial atoms, utilizing differentpicosecond UV laser pulse energy steps: 20→ 10→ 5→ 3→1→ 3→ 5→ 10→ 20 pJ/pulse to analyze the behavior of H inNb. The mass spectra collected are displayed in Figure 6 forthe mass-to-charge state ratio ranges of 0–3.0 and 30–50amu: approximately two million ions were collected for eachUV laser pulse energy. Note that the range of pulse energiesis from 1 to 20 pJ and that the 1H+ peak is essentially unaf-fected by the pulse energy. Although the 1H2

+ peak dependson the pulse energy and pulse energies of 10–20 pJ yield thebest results, the positions of the peaks of 93Nb3+ , 93Nb2+ , and93Nb1H2+ are unaffected strongly by the UV laser pulseenergy in the range 1–20 pJ. The atomic ratios of H2/H,NbH/Nb, and Nb3+ /Nb2+ at each UV laser pulse energy stepand the mass-resolving power, m/Δm, of H are displayed inFigure 7. A symmetrical behavior between increasing UVlaser pulse energy and decreasing UV laser pulse energy isobserved (compare Fig. 7a with Fig. 7d and Fig. 7c withFig. 7f). These figures indicate that the higherm/Δm and thehigher concentration of Nb2+ rather than Nb3+ are obtained

at higher picosecond UV laser pulse energies. An exceptionto this symmetric behavior is displayed for H2/H as a func-tion of picosecond UV laser pulse energy (compare Fig. 7bwith Fig. 7e). In general, the H2 molecules are detected morefrequently for increasing UV laser pulse energies owing tothe thermal energy absorbed from the UV laser pulses, withthe assumption being that thermal energy absorbed in theapex of the nanotip is proportional to the applied UV laserpulse energy. The higher H2/H ratio is achieved whendecreasing the UV laser pulse energy from 20 to 1 pJ/pulseas opposed to increasing the UV laser pulse energy from1 to 20 pJ/pulse. For example, H2/H is ~0.011± 0.00015 at10 pJ/pulse after decreasing it from 20 pJ/pulse; however,alternatively H2/H is ~0.0017± 0.000084 at 10 pJ/pulse afterincreasing it from 5 pJ/pulse. This difference in H2/H valuesat the same UV laser pulse energy between increasing anddecreasing the UV laser pulse energy steps is mainly due tothe residual thermal energy in the apex of the nanotip from aprevious higher UV laser pulse energy step, when decreasingthe laser pulse energy. The higher residual thermal energyafter decreasing the laser pulse energy increases the mobilityof H atoms by supplying sufficient thermal energy to formH2 molecules. This is also true for the NbH/Nb ratio becauseit is obtained when decreasing the UV laser pulse energies asopposed to increasing the UV laser pulse energies.

Figure 6. Mass spectra for different picosecond ultraviolet laser pulse energy steps: 20→ 10→ 5→ 3→ 1→3→ 5→ 10→ 20 pJ/pulse, with a mass-to-charge-state ratio range of: (a) 1.0–2.0 and (b) 30–50.

540 Yoon-Jun Kim and David N. Seidman

In summary, the behavior of hydrogen in Nb (b.c.c.)employed for SRF cavities has been investigated utilizingLEAP tomography, employing different UV laser pulseenergies utilizing a picosecond UV (wavelength = 355 nm)laser, and additional voltage pulsing. It is strongly empha-sized that hydrogen dissolves exothermically in niobium andthat the partial pressure of hydrogen in the UHV chamberwas 2.7 × 10 − 13 Torr during all the analyses as determined byan RGA. We present evidence for accelerated hydrogensurface diffusion resulting from the thermal energy input ofUV laser pulses at 50 pJ/pulse. For a UV laser pulse energy of<10 pJ/pulse this effect almost completely disappears. Theformation of NbH depends on the surface diffusivity ofhydrogen, as NbH is observed around the 110 pole. Ouranalyses include the sub-surface region, where the magneticfield penetrates ~40 nm from an Nb cavity’s surface duringthe operation of a linear particle accelerator. Therefore,residual hydrogen interstitial atoms or ordered niobiumhydrides formed during a cavity’s processing results in radio-frequency losses.

ACKNOWLEDGMENTS

This research was supported by the Fermi Research AllianceLLC under contract number DE-AC02-07CH11359 with theUS DOE. Atom-probe tomography measurements were per-formed at the Northwestern University Center for Atom-ProbeTomography (NUCAPT) and the LEAP tomograph waspurchased and upgraded with funding from the NSF-MRI

(DMR 0420532) andONR-DURIP (N00014-0400798, N00014-0610539, N00014-0910781) programs. NUCAPT is a SharedFacility of the Materials Research Center of the NorthwesternUniversity, supported by the National Science Foundation’sMRSEC program (DMR-1121262). We are also grateful to theInitiative for Sustainability and Energy at Northwestern (ISEN)for grants to upgrade NUCAPT’s capabilities.

REFERENCESAKAIWA, N. & WADA, M. (1990). A stress-induced hydride phase in

niobium observed with a field-ion microscope. J Less-CommonMet 160(2), 283–294.

AST, D.G. & SEIDMAN, D.N. (1968). Field ion microscopy of gold.Appl Phys Lett 13(10), 348.

AST, D.G. & SEIDMAN, D.N. (1971). Noble gas imaging of gold in fieldion microscope. Surf Sci 28(1), 19–31.

AVERBACK, R.S. & SEIDMAN, D.N. (1973). Neon gas imaging of gold infield-ion microscope. Surf Sci 40(2), 249–263.

BAKER, C. & BIRNBAUM, H.K. (1972). Hydrogen-dislocationinteraction in niobium. Scripta Metall 6(9), 851–854.

BAKER, C. & BIRNBAUM, H.K. (1973). Anelastic studies of hydrogendiffusion in niobium. Acta Metall Mater 21(7), 865–872.

BIRNBAUM, H.K., GROSSBECK, M.L. & AMANO, M. (1976). Hydrideprecipitation in Nb and some properties of NbH. J Less-Common Met 49(1–2), 357–370.

BRANDON, D.G. (1962). Image formation in field ion microscope.Philos Mag 7(78), 1003–1011.

BRANDON, D.G. (1963). Resolution of atomic structure—recentadvances in theory and development of field ion microscope.Br J Appl Phys 14(8), 474–484.

Figure 7. Calculated mass-resolving power, m/Δm of H (a,d), ratios of H2/H and NbH/Nb (b,e), and Nb3+ /Nb2+ (c,f).Note well that (a–c) are calculated for decreasing ultraviolet (UV) laser pulse energies 20→ 10→ 5→ 3→ 1 pJ/pulse,while (d–f) are for increasing UV laser pulse energies 1→ 3→ 5→ 10→ 20 pJ/pulse.

Hydrogen interstitials in pure Nb 541

BRANDON, D.G. (1964). Accurate determination of crystalorientation from field ion micrographs. J Sci Instrum 41(6),373–375.

BRANDON, D.G. (1965). Analysis of field evaporation data from fieldion microscope experiments. Br J Appl Phys 16(5), 683–688.

BRANDON, D.G. (1968). Gas impact, field etching and fielddeformation. In Field-Ion Microscopy, Hren, J.J. & Ranganathan, S.(Eds.), pp. 53–68. New York, NY: Plenum Press.

CHAMPION, M.S., COOLEY, L.D., GINSBURG, C.M., SERGATSKOV, D.A.,GENG, R.L., HAYANO, H., IWASHITA, Y. & TAJIMA, Y. (2009).Quench-limited SRF cavities: Failure at the heat-affected zone.IEEE Trans Appl Supercond 19(3), 1384–1386.

CIOVATI, G., MYNENI, G., STEVIE, F., MAHESHWARI, P. & GRIFFIS, D.(2010). High field Q slope and the baking effect: Review of recentexperimental results and new data on Nb heat treatments. PhysRev ST Accel Beams 13(2), 022002.

COOLEY, L.D., BURK, D., COOPER, C., DHANARAJ, N., FOLEY, M., FORD, D.,GOULD, K., HICKS, D., NOVITSKI, R., ROMANENKO, A., SCHUESSLER, R.,THOMPSON, C. & WU, G. (2011). Impact of forming, welding, andelectropolishing on pitting and the surface finish of SRF cavityniobium. IEEE Trans Appl Supercond 21(3), 2609–2614.

DELHEUSY, M., STIERLE, A., KASPER, N., KURTA, R.P., VLAD, A., DOSCH, H.,ANTOINE, C., RESTA, A., LUNDGREN, E. & ANDERSEN, J. (2008). X-rayinvestigation of subsurface interstitial oxygen at Nb/oxideinterfaces. Appl Phys Lett 92(10), 101911.

DRECHSLER, M. & WOLF, P. (1960). Zur Analyse vonFeldionenmikroskop-Aufnahmen mit atomarer Auflosung. In 4thInternational Conference on Electron Microscopy, Bargmann, W.,Mollenstedt, G., Niehrs, H., Peters, D., Ruska, E. & Wolpers, C.(Eds.), pp. 835–847. Berlin, Germany: Springer-Verlag.

GABRIEL, S.B., SILVA, G., CANDIOTO, K.C.G., SANTOS, I.D., SUZUKI, P.A.& NUNES, C.A. (2011). Niobium hydrogenation process effect oftemperature and cooling rate from the hydrogenationtemperature. Int J Refract Met H 29(1), 134–137.

GAULT, B., LA FONTAINE, A., MOODY, M.P., RINGER, S.P. &MARQUIS, E.A.(2010). Impact of laser pulsing on the reconstruction in an atomprobe tomography. Ultramicroscopy 110(9), 1215–1222.

GAULT, B., MOODY, M.P., DE GEUSER, F., TSAFNAT, G., LA FONTAINE, A.,STEPHENSON, L.T., HALEY, D. & RINGER, S.P. (2009). Advances inthe calibration of atom probe tomographic reconstruction.J Appl Phys 105(3), 034913.

GAUSSMANN, A., DRACHSEL, W. & BLOCK, J.H. (1992). Kinetics of fieldevaporation during hydride formation on gap surfaces—a FIMand atom-probe study. Langmuir 8(1), 125–129.

GE, M., WU, G., BURK, D., OZELIS, J., HARMS, E., SERGATSKOV, D.,HICKS, D. & COOLEY, L.D. (2011). Routine characterization of 3Dprofiles of SRF cavity defects using replica techniques.Supercond Sci Technol 24(3), 035002.

GOMER, R. (1961). Field Emission and Field Ionization. Cambridge:Harvard University Press.

GROSSBECK, M.L. & BIRNBAUM, H.K. (1977). Low-temperaturehydrogen embrittlement of niobium. 2. Microscopicobservations. Acta Metall Mater 25(2), 135–147.

KELLOGG, G.L. (1981). Pulsed laser stimulated field desorption ofhydrogen from molybdenum. J Chem Phys 74(2), 1479–1487.

KELLOGG, G.L. (1983). Field evaporation of silicon and fielddesorption of hydrogen from silicon surfaces. Phys Rev B28(4), 1957–1964.

KIM, Y.-J., TAO, R., KLIE, R.F. & SEIDMAN, D.N. (2013). Direct atomic-scale imaging of hydrogen and oxygen interstitials in pure niobiumusing atom-probe tomography and aberration-corrected scanningtransmission electron microscopy. ACS Nano 7(1), 732–739.

MA, Q. & ROSENBERG, R.A. (2003). Angle-resolved X-rayphotoelectron spectroscopy study of the oxides on Nb surfacesfor superconducting r.f. cavity applications. Appl Surf Sci206(1–4), 209–217.

MAHESHWARI, P., TIAN, H., REECE, C.E., KELLEY, M.J., MYNENI, G.R.,STEVIE, F.A., RIGSBEE, J.M., BATCHELOR, A.D. & GRIFFIS, D.P.(2011). Surface analysis of Nb materials for SRF cavities.Surf Interface Anal 43(1–2), 151–153.

MAKENAS, B.J. & BIRNBAUM, H.K. (1982). Phase-changes in theniobium hydrogen system 2. Low-temperature hydride phase-transitions. Acta Metall Mater 30(2), 469–481.

MILLER, M.K., RUSSELL, K.F., THOMPSON, K., ALVIS, R. & LARSON, D.J.(2007). Review of atom probe FIB-based specimen preparationmethods. Microsc Microanal 13(6), 428–436.

MOORE, A.J.W. & SPINK, J.A. (1974). Influence of surfacecoordination on field evaporation processes in tungsten.Surf Sci 44(1), 198–212.

OLNEY, T.N., CANN, N.M., COOPER, G. & BRION, C.E. (1997). Absolutescale determination for photoabsorption spectra and thecalculation of molecular properties using dipole sum rules.Chem Phys 223(1), 59–98.

PADAMSEE, H. (2009). RF Superconductivity: Science, Technology,and Applications. Weinheim: Wiley-VCH.

PANOFSKY, W.K.H. & PHILLIPS, M. (2005). Classical Electricity andMagnetism. Mineola, NY: Dover Publications.

PHINNEY, N., TOGE, N. & WALKER, N. (2007). International LinearCollider Reference Design Report. Vol III-Accelerator. http://www.linearcollider.org/ILC/Publications/Reference-Design-Report

RICKER, R.E. & MYNENI, G.R. (2010). Evaluation of the propensityof niobium to absorb hydrogen during fabrication ofsuperconducting radio frequency cavities for particleaccelerators. J Res Natl Inst Stan 115(5), 353–371.

ROMANENKO, A. & GONCHAROVA, L.V. (2011). Elastic recoil detectionstudies of near-surface hydrogen in cavity-grade niobium.Supercond Sci Technol 24(10), 105017.

ROMANENKO, A. & PADAMSEE, H. (2010). The role of near-surfacedislocations in the high magnetic field performance ofsuperconducting niobium cavities. Supercond Sci Technol23(4), 045008.

SAKATA, T. & BLOCK, J.H. (1982). Field evaporation of silicon-(III)surfaces in the presence of hydrogen. Surf Sci 116(1), L183–L189.

SCHOBER, T. (1975a). Niobium-hydrogen system-electron-microscope study 1. Room-temperature results. Phys StatusSolidi A 29(2), 395–406.

SCHOBER, T. (1975b). Niobium-hydrogen system—Electron-microscope study 2. Low-temperature structures. Phys StatusSolidi A 30(1), 107–116.

SCHOBER, T., LINKE, U. & WENZL, H. (1974). Metallography ofniobium hydrogen system. Scripta Metall 8(7), 805–812.

SCHOBER, T., PICK, M.A. &WENZL, H. (1973). Electron-microscopy ofbeta-hydride in niobium. Phys Status Solidi A 18(1), 175–182.

SCHOBER, T. & WENZL, H. (1976). Beta-phase melting andsolidification phenomena in niobium-hydrogen system. PhysStatus Solidi A 33(2), 673–681.

SEBASTIAN, J.T., SEIDMAN, D.N., YOON, K.E., BAUER, P., REID, T.,BOFFO, C. & NOREM, J. (2006). Atom-probe tomography analysesof niobium superconducting RF cavity materials. Physica C441(1–2), 70–74.

SEIDMAN, D.N. (2007a). Perspective: From field-ion microscopy ofsingle atoms to atom-probe tomography: A journey: “Atom-probe tomography” [Rev. Sci. Instrum. 78, 031101, (2007)].Rev Sci Instrum 78(3), 030901.

542 Yoon-Jun Kim and David N. Seidman

SEIDMAN, D.N. (2007b). Three-dimensional atom-probe tomography:Advances and applications. Annu Rev Mater Res 37, 127–158.

SEIDMAN, D.N. & STILLER, K. (2009). An atom-probe tomographyprimer. MRS Bull 34(10), 717–724.

SIMMONS, G. &WANG, H. (1971). Single Crystal Elastic Constants andCalculated Aggregate Properties: A Handbook. Cambridge, MA:MIT Press.

SINGER, W., SINGER, X., ADERHOLD, S., ERMAKOV, A., TWAROWSKI, K.,CROOKS, R., HOSS, M., SCHOLZ, F. & SPANIOL, B. (2011). Surfaceinvestigation on prototype cavities for the European X-ray freeelectron laser. Phys Rev ST Accel Beams 14(5), 050702.

SOUTHON, M.J. (1968). Field emission and field ionization. InField-IonMicroscopy, Hren, J.J. & Ranganathan, S. (Eds.), pp. 6–27.New York, NY: Plenum Press.

TSONG, T.T. (1978). Field ion image formation. Surf Sci 70(1), 211–233.TSONG, T.T. (1990). Atom-Probe Field Ion Microscopy: Field Ion

Emission and Surfaces and Interfaces at Atomic Resolution.Cambridge and New York, NY: Cambridge University Press.

WAUGH, A.R., BOYES, E.D. & SOUTHON, M.J. (1976). Investigations offield evaporation with a field-desorption microscope. Surf Sci61(1), 109–142.

YOON, K.E., SEIDMAN, D.N., ANTOINE, C. & BAUER, P. (2008). Atomic-scale chemical analyses of niobium oxide/niobium interfaces viaatom-probe tomography. Appl Phys Lett 93(13), 132502.

YOON, K.E., SEIDMAN, D.N., BAUER, P., BOFFO, C. & ANTOINE, C.(2007). Atomic-scale chemical-analyses of niobium forsuperconducting radio-frequency cavities. IEEE Trans ApplSupercond 17(2), 1314–1317.

Hydrogen interstitials in pure Nb 543