Introduction

The concept of liquid metal plasma-facing materials for future fusion reactors has been studied theoretically and experimentally. Liquid metals have the advantage of a quick self-healing capability because of their liquid state, among others. Several devices, such as the Lithium Tokamak eXperiment-β (LTX-β) [1], National Spherical Torus eXperiment (NSTX) [2], EAST [3], ADITYA-Upgrade [4], HIDRA [5], and STOR-M [6] have employed lithium (Li) as a plasma-facing material and investigated its effects on plasma performance. Li is a low-Z element, and, hence, is expected to induce much less Bremsstrahlung radiation loss in the core plasma. Improved plasma performance with a flat electron temperature profile was observed in LTX [7] at the Princeton Plasma Physics Laboratory (PPPL) because of the high retention rate of H species in Li, which significantly suppressed hydrogen recycling.

Over the last decade, PPPL has expanded experimental facilities to study plasma-material interactions (PMIs) for Li plasma-facing materials. The Surface Science and Technology Laboratory (SSTL) at PPPL has several ultrahigh vacuum chambers equipped for surface analysis and materials characterization utilizing X-ray photoelectron spectroscopy (XPS), Auger electron spectroscopy (AES), temperature programmed desorption (TPD), low energy electron diffraction (LEED), secondary electron emission, scanning electron microscopy (SEM), and scanning Auger microscopy (SAM). To bridge tokamak experiments and laboratory surface analysis techniques, we have utilized the Sample Exposure Probe (SEP), with its vacuum suitcase, to transfer a sample after exposures in LTX-β plasmas to the surface analysis chambers in the SSTL [8]. Using those PMI facilities at PPPL, we have investigated erosion, hydrogen retention, chemical state, thermal stability, and spreading of lithium films, and also characterized the surface evolution of lithium-coated surfaces in the LTX-β environment [9,10,11,12,13,14,15].

Sputtering yields and hydrogen retention at Li plasma-facing components are affected by chemical contaminants (e.g., H or D, O, C, and B), film temperature, and incident H or D ion energy and angle [9,10,11, 13, 16,17,18,19,20]. Allain and Ruzic [17] reported that the chemical state of the deuterium-treated (oxygen-removed) lithium surface plays a major role in the decrease of Li sputtering. They also reported that the fraction of sputtered species in an ionic state ranges from 55 to 65% for incident D+, He+, and Li+ ions with energies between 100 and 1000 eV. The charge state of sputtered species is important since, in a fusion device, the motion of sputtered ions is significantly affected by the sheath potential. Positive and negative hydrogen ion production for a hydrogen ion beam incident on graphite, HOPG, diamond, Mo, Al, Ag, and Si [21,22,23,24,25,26,27,28,29,30,31,32] has also been studied in terms of developing a negative hydrogen ion source and other applications. Measurements on sputtered contaminant species and their positive or negative charge are important for understanding surface treatment mechanisms and also their effect on plasmas. Glow discharge cleaning (GDC) using H2, D2, He, Ne, or Ar plasmas have been investigated for surface conditioning in current fusion devices [4, 15, 33,34,35,36,37,38,39,40] and ITER [41]. Incident ion energies of more than 200 eV, accelerated by the applied anode potential for discharge, are expected for GDC [42]. Ar, Ne, and N2 have also been introduced to fusion plasmas by gas puffing in several fusion reactors such as DIII-D [43, 44], ASDEX Upgrade [45], and JET [46] to induce plasma detachment on divertor surfaces.

We report herein measurements made possible by additional upgrades of the surface analysis facilities for Li PMI investigations at PPPL. Using these new capabilities, we measured positive and negative ions sputtered/ejected from oxidized lithium surfaces at 300 or 540 K due to H, D, He+, Ne+, or Ar+ ion irradiations for incident energies from 30 to 2000 eV. The measurements were performed using a differentially pumped quadrupole mass spectrometer with capabilities for detecting positive and negative ions, and an energy filter for determining the energies of the detected ions. Sputtering/ejection yields of both positive and negative ion species were quantified, and the mean kinetic energy of the sputtered/ejected ions was also determined. These measurements will aid in understanding and predicting scenarios in fusion reactors employing lithium as a plasma-facing material.

Upgrades to Facilities for Studies of Plasma-Material Interactions at PPPL

An important development for PMI studies a few years ago, the Sample Exposure Probe (SEP) is a mobile device featuring a vacuum suitcase with ultrahigh vacuum (UHV) pumping and a base pressure < 5 × 10–9 Torr, and has been described [8]. Active pumping under high vacuum enables the transfer of a lithium-coated sample, which is highly chemically reactive, between the Lithium Tokamak eXperiment-β (LTX-β), without air exposure, to an UHV surface analysis chamber for X-ray photoelectron spectroscopy (XPS) in the Surface Science and Technology Laboratory (SSTL) located just down the hall from LTX-β. Experimental results using the SEP for XPS analysis of samples exposed to Li evaporations and H-plasma exposures in LTX-β have been reported previously [15]. Recently, a new, SEP-compatible facility was added, the Sample Exposure Station (SES), which enables ion exposures on a SEP target using an electron cyclotron resonance (ECR) plasma source, and capabilities for secondary ion mass spectroscopy (SIMS) for positive and negative ions, temperature programmed desorption (TPD), and physical vapor deposition of Li films using a Li thermal evaporator. This Li source can produce a relatively fast Li deposition rate (~ 100 nm/min) and, hence, Li deposition, plasma exposure, and surface analysis in the same chamber can be carried out within a short time period. In a separate upgrade, since the SEP was designed for sample heating up to ~ 900 K, the recent installation of a new residual gas analyzer (RGA) on the LTX-β vessel now enables TPD measurements for the target immediately (~ 1 min) after plasma exposure in LTX-β for hydrogen retention and chemical composition studies.

Experimental Methods

We performed experiments in the Sample Exposure Station (SES), which is an ultrahigh vacuum chamber for ion irradiation experiments using an electron cyclotron resonance (ECR) plasma source (GenII Atom/Ion Hybrid Source, tectra GmbH). The SEP was attached to the SES to introduce a target/substrate that could be coated with a Li film (Fig. 1). This target was a stainless steel (SS) substrate (25-mm diameter, 5-mm thickness) that was attached at the end of the SEP sample manipulator rod. The SS substrate surface was hand polished (Rrms ~ 100 nm) using 1-μm alumina powder in water and then rinsed with acetone and ethanol. The Li film for the experiments reported herein was deposited on the SS substrate surface using the in-vessel LTX-β Li evaporators [15]. The Li film thickness deposited (~ 250 nm) was monitored by a quartz crystal microbalance (QCM) in the LTX-β vacuum vessel during the Li evaporation. This single Li film was used for all ion exposures reported in this article. The Li-coated SS target in the SEP was kept at a pressure of below 5 × 10–9 Torr, which mainly consisted of H2O, H2, and CO, during transportation from LTX-β to the XPS surface analysis chamber and the SES. This Li sample was analyzed 48 h after the Li deposition to determine the surface composition by XPS before introducing the sample to the SES. The XPS system has been described elsewhere [8]. XPS broadscan spectra, along with the Li 1s, C 1s, and O 1s regions, were obtained and quantification of the surface composition utilized the sensitivity factors from the ref. [47]. The surface composition was determined to be Li:O:C = 60:38:2 (at.%). This composition was identical to that of another Li surface that was kept in the SEP chamber for several weeks at a background pressure of 5 × 10–8 Torr. Hence, the oxidation level of the target surface appears to be saturated. We note that the XPS technique cannot detect elemental H directly, but these surfaces also contained H due to hydroxyl (OH) groups in LiOH, which can be seen by a chemical-shifted peak in the XPS O 1s spectra.

Fig. 1
figure 1

Schematic of the Sample Exposure Station (SES) viewed from the top

Full size image

Ions were generated from the ECR plasma source [48,49,50,51] in the SES chamber (Fig. 1). The SES chamber was pumped by two turbomolecular pumps (170 and 240 L/s) and had a base pressure below 5 × 10–9 Torr. Gas partial pressures in the chamber were monitored by an RGA (HALO 301-RC, Hiden Analytical Inc.), including during ion irradiations. The distance between the ion source exit aperture and the target was 95 mm. Background pressures during ion irradiations were 5 × 10–6–2 × 10–5 Torr, dominated by the gas fed to the ECR plasma source, while the partial pressure of impurity species (primarily H2 and H2O) were below 5 × 10–9 Torr. The ion current from the plasma source was measured by a Faraday cup at the same location as the target before and after an ion exposure. The extraction grid at the beam exit of the ECR plasma source was negatively biased to prevent electrons from exiting with the ion beam [50, 51], and, hence, only ions and neutrals irradiated the target. We measured the ion species concentration ratios for ion beams produced by several different feed gases previously [50]. The atomic ion flux (\({j}_{{\mathrm{H}}^{+}/{\mathrm{D}}^{+}}\)), ions/cm2/s, and atomic ion kinetic energies (\({E}_{{\mathrm{H}}^{+}/{\mathrm{D}}^{+}}\)) for H and D ion beam irradiations were calculated herein by utilizing our previous finding that trihydrogen ions, H3+ and D3+, were the dominant (~ 80%) species from H2 and D2 plasmas, respectively. Thus, the atomic ion flux and energy of incident H and D atomic ions were calculated by \({j}_{{\mathrm{H}}^{+}/{\mathrm{D}}^{+}} = {j}_{{\mathrm{H}}_{3}^{+}/{\mathrm{D}}_{3}^{+} }\times 3\) and \({E}_{{\mathrm{H}}^{+}/{\mathrm{D}}^{+}} = {E}_{{\mathrm{H}}_{3}^{+}/{\mathrm{D}}_{3}^{+} }/3\), where \({j}_{{\mathrm{H}}_{3}^{+}/{\mathrm{D}}_{3}^{+}}\) was measured as the ion current using the Faraday cup. A negligible amount (< 1%) of doubly charged ions were found from ion concentration measurements, even for Ar plasmas [50]. The sequence of the exposures was He+, Ar+, Ne+, H, and D ion irradiations on the 300 K target followed by D ion irradiation on the 540 K target.

Here, we use “ejection” for species whose origin can be sputtering and/or reflection, e.g., H2+ and H− species observed under H ion irradiation. Sputtered ions species were monitored in the SES by using a differentially pumped quadrupole mass spectrometer (IDP, Hiden Analytical Inc.) with capabilities for detecting positive and negative ions, and an energy filter for determining the energies of the detected ions. The IDP chamber is separated from the main SES chamber by a beam skimmer with an orifice of 0.5-mm diameter and is differentially pumped by a 170 L/s turbomolecular pump so that the base pressure was below 5 × 10–8 Torr at all times. The IDP ion optics consists of two electrodes to focus ion species into the quadrupole mass spectrometer component, and an extractor electrode located just after the IDP sampling orifice. Charged ions, positive or negative, were selected by tuning the electrode potential of the ion optics, quadrupole mass spectrometer, and detector components. The electrodes of the ion optics were tuned by optimizing the applied voltage of each electrode to achieve the highest intensity of sampled positive or negative ions. The reference potential for all IDP electrodes, including the quadrupole rods, can be biased so that the IDP can also be used as a retarding field energy analyzer.

Calibration of IDP Signal to Ion Count

Ions detected by the IDP were obtained as signal intensities Ik (counts/s) for species k, which is given by

$$ I_{k} = J_{k} AP_{k} , $$
(1)

where Jk is the flux of incoming ions at the IDP entrance after ejection from the target, A is the area of the sampling orifice (constant), and Pk is the transmission factor of the IDP [52, 54]. Jk is given by

$$ J_{k} = Y_{k,l} \left( {E_{l} ,\theta_{l} } \right)F_{k} \left( {E_{l} ,\theta_{k} } \right)j_{l} C, $$
(2)

where Yk,l is the sputtering yield of an ejected species k as a function of the incident energy of ion species l bombarding the target, Fk is the angular distribution of the sputtered ion species k, jl is the flux of ion species l bombarding the target, and C is a scaling constant. The scaling constant C represents the portion of total ejected ion flux from the target/sample arriving at the IDP sampling orifice and is defined only by the geometrical configuration. Therefore, C is the same for all species. By substituting eqs. (2) to (1),

$$ I_{k} = Y_{k,l} \left( {E_{l} ,\theta_{l} } \right)F_{k} \left( {E_{l} ,\theta_{k} } \right)j_{l} AP_{k} C , $$
(3)

where θl = 45° and θk = 0° (measured from the target surface normal) in this experimental setup. If we assume that the angular dependences of Fk are the same for different species and incident energies,

$$ F_{k} \left( {E_{l} ,\theta_{k} } \right) = F\left( {0^\circ } \right) . $$
(4)

Justification for this assumption cannot be currently provided due to a lack of available information, but we proceed with this assumption, recognizing this limitation of the IDP calibration method used below. Such incident-energy and species dependencies of Fk can be investigated by angular scanning experiments using an azimuthally rotatable detector [21] and/or computational methods using SRIM [53], VFTRIM-3D [17], and molecular dynamics (MD) based [19] calculations but are beyond our scope in this article. Then, using Eq. (3), Ik normalized by \({I}_{{k}{\prime}}\) for species k’ sputtered by ion species l’ represented as

$$ \frac{{I_{k} }}{{I_{{k^{\prime } }} }} = \frac{{Y_{k,l} \left( {E_{l} } \right)}}{{Y_{{k^{\prime } ,l^{\prime } }} \left( {E_{{{\text{l}}^{\prime } }} } \right)}}\frac{{j_{l} }}{{j_{{l{\prime} }} }}\frac{{P_{k} }}{{P_{{k^{\prime } }} }}{ } $$
(5)

Therefore,

$$ I_{k} = Y_{k,l} \left( {E_{l} } \right)B_{{k^{\prime } }} R_{{k,k^{\prime } }} j_{l} $$
(6)

where

$$ B_{{k^{\prime } }} = \frac{{I_{{k^{\prime } }} }}{{Y_{{k^{\prime } ,l^{\prime } }} \left( {E_{{{\text{l}}^{\prime } }} } \right)j_{{l^{\prime } }} }} $$
(7)

and

$$ R_{{k,k^{\prime } }} = \frac{{P_{k} }}{{P_{{k^{\prime } }} }} . $$
(8)

The sputtering yields of Li+ from a Li2O surface due to He ion beam irradiations at an incident ion angle of 45°, were measured for \({E}_{H{e}^{+}}\) = 500, 700, and 1000 eV [17] previously. We determined the value,

$$ B_{{Li^{ + } }} = \frac{{I_{{{\text{Li}}^{ + } }} }}{{{ }Y_{{Li^{ + } ,He^{ + } }} \left( {E_{{He^{ + } }} , 45^\circ } \right)j_{{He}^{+}}}}$$
(9)

by measuring \({I}_{L{i}^{+}}\) for \({j}_{{He}^{+}}\) at an incident ion angle of 45° and energy of 500 eV of He+ for the Li2O target we utilized herein (described above). Then, having Li+ as k’, Eq. (6) becomes

$$ I_{k} = Y_{k,l} \left( {E_{l} } \right)B_{{Li^{ + } }} R_{{k,Li^{ + } }} j_{l} $$
(10)

and, hence, the sputtering yield for species k,

$$ Y_{k,l} \left( {E_{l} } \right) = \frac{{I_{k} }}{{B_{{Li^{ + } }} }}\frac{1}{{R_{{k,Li^{ + } }} j_{l} }} . $$
(11)

For negative species quantification, the value of \({R}_{{O}^{-},L{i}^{+}}\) was determined by measuring relative values of \({P}_{{O}^{-}}\) and \({P}_{{He}^{+}}\) using the method in ref. [54]. For O− measurement, we utilized the dissociative attachment of O2 at the electron-impact energy of 6.5 eV [55]. For He+ formation, we utilized the electron-impact ionization of He atoms at the electron-impact energy of 70 eV [56]. We assume \({P}_{{He}^{+}}\) ~ \({P}_{L{i}^{+}}\) because of similar mass, so that \({R}_{{O}^{-},L{i}^{+}}\) ~ \({R}_{{O}^{-},{He}^{+}}\). We measured that relative Pk for the positive species in the mass range 4–40 amu (D2, He, Ne, O2, and Ar) varies within 25% from its average, which is smaller than the dominant error mentioned below. For H2 (2 amu), \({{R}_{{H}_{2}^{+},L{i}^{+}}\sim P}_{{H}_{2}^{+}}/{P}_{{He}^{+}}\) = 1.7 was measured, so this value was applied for H2+, H+, and D+ quantification. Due to difficulties to form light negative ions, e.g. H− or D−, using the ionizer filament, we assumed the similar mass dependency with the positive ion species for the negative ion species, \({P}_{{H}^{-},}/{P}_{{O}^{-}}\)~ \({P}_{{H}_{2}^{+}}/{P}_{{He}^{+}}\), to quantify H− or D− species. \({R}_{L{i}^{+},L{i}^{+}}\) is just unity. Then, using Eq. (11), the sputtering yield is determined by measuring the sputtered ion signal intensity Ik and incident ion current jl. The maximum uncertainty estimated in the measurement of \({Y}_{{Li}^{+},H{e}^{+}}\left({E}_{H{e}^{+}}, 45^\circ \right)\), which we used for the sputtering yield calibration, was about 40% [17], which comes from the uncertainty of the angular sputtered distribution. Therefore, we assigned ± 40% for the error of the determined sputtering yields. Additional uncertainty in the experiments reported herein arises from the ion beam current fluctuations, which varied ± 10%, but smaller than the error made by the sputtering yield reference.

Results

Mass Spectra for Ions Sputtered/Ejected from Oxidized Li During Ar/D Ion Irradiation (SIMS)

Figure 2 shows typical mass scans of positive and negative ions during Ar ion irradiation at an ion incident energy of EAr = 250 eV. Sputtered Li+ ions at 6 and 7 amu were seen in Fig. 2(a) along with a small signal for H+ (1 amu). We note that the peak height ratio of 6Li+:7Li+ was 1:5 although the natural abundance ratio of Li-6 and Li-7 is 1:12, and this is reasonable given the higher sputtering yield expected for the light Li isotope. H− (1 amu), O− (16 amu), and OH− (17 amu) peaks were detected in the negative ion scans (Fig. 2(b, c)). The existence of hydrogen species in these scans confirms that the surface contains H, likely as LiOH compounds. No additional multiatomic ions, e.g., Li2, LiH, LiO, and LiOH, were detected.

Fig. 2
figure 2

Mass spectra of (a) positive and (b, c) negative species for an Ar ion bombardment at E = 250 eV

Full size image

Figure 3 shows mass scans for positive and negative ions during D ion irradiation with ion incident energies of ED = 50 and 175 eV. Sputtered Li+ ions at 6 and 7 amu are shown in Fig. 3(a). D2+ (4 amu) species, likely dissociated from D3+ and reflected, are clearly observed for ED = 50 eV. Figure 3(b, c) shows D− (2 amu) but no detectable O− (16 amu) or OD− (18 amu) peaks for D ion irradiation at ED = 50 eV. For ED = 175 eV, H− (1 amu), O− (16 amu), OH− (17 amu), and OD− (18 amu) peaks were observed in the negative ion scan. H ion species likely arise from dissociative chemisorption of water (H2O) present in the background gas. The background intensity is clearly seen for the H− and D− peaks on the ED = 175 eV profile in Fig. 3(b). Such background intensity was possibly the result of the filtering of relatively high energy H− (~ few eV) being unsuccessful. We assumed a linear background profile as shown in Fig. 3(b) and subtract this profile from the mass spectra to determine the D− signal intensity. We empirically estimate that the error of this background subtraction is similar to the error of ± 40% due to the sputtering yield calibration uncertainty.

Fig. 3
figure 3

Mass spectra of (a) positive and (b, c) negative species for D ion bombardments at ED = 50 (dashed) and 175 eV (solid). The linear profile for the background substruction of the D− intensity is shown as a red line in (b)

Full size image

Sputtering Yields for Li+ Ions

Figure 4 shows the measured Li+ (7 amu) sputtering yield for various incident ions and energies. The Li+ ion sputtering yield for He+ ion irradiation at 500 eV is the same as that previously reported [17] because this data point was used for the calibration in this work (see section "Calibration of IDP signal to ion count"). The good agreement seen between this experiment and previously reported data [17] at 1000 eV benchmarks the experimental method we applied here. Li+ sputtering yields for He+ ion irradiation show a relatively flat profile within a range of 0.2–0.4 Li+ ions/He+ ion even at a low incident ion energy of EHe = 90 eV. Those yields are also higher than for D or H ion irradiation at any incident energies, which also show relatively flat profiles for E = 100–700 eV. The Li+ sputtering yields drop by more than a factor of two at E = 50 eV compared to that at E = 100 eV for H and D ion irradiation, but they are still within an order of magnitude of those at higher energy bombardment of up to E = 700 eV. More significant decay of the sputtering yield at E = 50 eV was expected for H and D bombardment based on previously reported TRIM calculations [1], which only consider physical sputtering processes. However, Krstic et al. recently reported higher sputtering yields than TRIM results for H and D bombardment for energies of E < 100 eV using molecular dynamics simulations [19, 20]. Figure 4 shows that both Ne and Ar irradiations cause relatively higher Li+ sputtering yields than for H or D irradiation for E > 300 eV. Such an incident ion energy range is expected to occur during glow discharge conditioning (GDC) or high energy flux events, such as an edge-localized mode (ELM). Li− sputtering was observed for Ar+ and Ne+ ion irradiations but those yields were less than 1% of the Li+ sputtering yields, so the results are not shown in this article.

Fig. 4
figure 4

Sputtering yields of Li+ ions (7 amu) from the Li-coated target at 300 K for Ar, Ne, He, H, and D ion bombardment. For D ion irradiation, a target temperature of 540 K (dashed line) was also measured. Error bars near ± 40% are for every data point but are shown only for the He incident ion profile

Full size image

Sputtering Yields for O− Ions

Sputtering yields for O− (16 amu) and OH− (17 amu) ions during Ar+, Ne+, and He+ ion irradiation are shown in Fig. 5(a). Energy thresholds for detectable O− and OH− intensities were at an incident ion energy of E = 150 eV, which are at higher energies than those for Li+ sputtering at E > 50 eV. He+ ion irradiation showed relatively flat profiles for the sputtering yields for both O− and OH− for incident He+ ions over the energy range of E = 150–2000 eV. Ar+ ion bombardment showed higher O− sputtering yields of 0.6–1.1 O− ions/Ar+ ion for energies of E > 500 eV, which are incident ion energies that can be expected in glow discharge conditioning. Ne+ ion bombardment showed sputtering yields of 0.2–0.4 O− ions/Ne+ ion for energies of E > 500 eV. Figure 5(b) shows O− and OH− sputtering/ejection yields for H ion irradiation, along with O− sputtering yields for D ion irradiation. The OH− ejection yield is much larger than that of O− sputtering for H ion irradiation. D ion bombardment produces much larger O− sputtering yields than H ion bombardment, as was seen for Li+ sputtering yields.

Fig. 5
figure 5

Sputtering yields of (a) O− (solid) and OH− (dashed) species sputtered by Ar+, Ne+, and He+ ion bombardments, and (b) O− (solid) and OH− (solid) species sputtered/ejected by D and H ion bombardment. The target temperature was 300 K for all ion irradiation experiments except for D ion bombardment (shown in (b)) with a target temperature of 540 K

Full size image

Figure 6(a) shows sputtering/ejection yields of O−, OH−, and OD− ions during D ion irradiation on a Li-coated target at 300 K. For incident D ion energies of E > 180 eV, similar sputtering/ejection yields were seen for those three negative ion species. The significantly higher yields of OH− and OD− ejections over that for O− sputtering suggests that the surface is extensively hydroxylated with OH species initially, and likely that such OD species are formed during D ion irradiation, and it is quite possible that there is also a mechanism in which D-attachment to an oxygen atom at the surface occurs (in a “pickup” reaction) to cause OD− ejection with increasing probability at lower incident D ion energies. Ejection of OD− and OH− is a mechanism of removing O species that is dominant over O− sputtering for lower energy D ion bombardments at E = 90 eV. Sputtered/ejected species from a heated target at 540 K were different from those obtained with the target at 300 K, as seen in Fig. 6(b). Ejected OH− and OD− signals were no longer observed, but OHD− (19 amu) was ejected instead. D-attachment to an OH species at the surface is likely to cause OHD− ejection, but further investigations are necessary to understand the process and the role of changes in the surface composition and/or the ejection mechanism at elevated temperatures.

Fig. 6
figure 6

Sputtering/ejection yields of O− (16 amu), OH− (17 amu), and OD− (18 amu) species during D ion bombardment of Li-coated targets at (a) 300 K and (b) 540 K. For the target at 540 K (b) no ejection of OH− (17 amu) or OD– (18 amu) ions were observed, but a signal for OHD− (19 amu) was observed

Full size image

Sputtering and Ejection Yields for H and D Ions

H− and H+ sputtering yields are shown in Fig. 7 for Ar+, Ne+, He+, and D ion irradiation. Ar+ and Ne+ ion bombardment show the highest both H− and H+ ion sputtering yields for incident ion energies of E > 200–300 eV, and, hence, this is consistent with the common knowledge that GDC using Ar or Ne is efficient at removing both H and O contaminants in Li films. He+ ion irradiation causes an H− and H+ ion sputtering yields of ~ 0.05 ions/He+ ion even at E = 100 eV, indicating that He+ ion irradiation may contribute to removing hydrogen species trapped in lithium during fusion plasma operations.

Fig. 7
figure 7

Sputtering yields of (a) H− and (b) H+ ions due to Ar+ , Ne+, He+, D, and H ion irradiation

Full size image

H− and D− ejection yields are shown in Fig. 8(a) for H and D ion irradiations, respectively. Both H− and D− ejection yields show similar profile at E < 100 eV. H−/D− ejection yields increase for H/D ion bombardments at E < 100 eV because of the reflection of incident H/D ions in addition to sputtering processes. It is also seen that H− is ejected more than H+ (Fig. 7(b)). H2+, D2+, and He+ ejection yields are shown in Fig. 8(b) for H, D, and He ion irradiations, respectively. It is seen that D2+/H2+ is ejected less than D−/H− (Fig. 8(a)). Relatively high He+ ejection yields were measured for He ion irradiations while the ejection yields of Ar+ and Ne+ were less than 0.01 (not shown) for the incident ion energy range of E = 100–2,000 eV.

Fig. 8
figure 8

(a) Ejection yields of H− and D− ions due to H and D ion irradiations, respectively. (b) Ejection yields of H2+, D2+, and He+ ions due to H, D, and He+ ion irradiations, respectively

Full size image

Energy Distributions

The kinetic energy of sputtered/ejected ions and reflected ions during ion irradiation was investigated by biasing the mass spectrometer electrodes to work as a retarding field analyzer. Figure 9 shows intensity profiles of Li+ and He+ ions ejected by He+ ion irradiation at E = 125 eV, as a function of the bias voltage away from the reference potential of the IDP. A retarding field analyzer acts as a high pass filter, and so the detected intensity profile is an integration of the ion energy distribution from the bias voltage to the highest kinetic energy of ions. Thus, a broader profile corresponds to higher ion energies. He+ ions, which we can assume to be mainly reflected, will have much lower kinetic energies than Li+ species, which must be sputtered, and this will be indicated by the width difference of the He+ and Li+ ion profiles. Assuming a Maxwell–Boltzmann distribution for the ion kinetic energy, a bias potential that gives 40% of the maximum intensity at 0 V corresponds to the mean ion kinetic energy < Ek > of a sputtered species k.

Fig. 9
figure 9

Energy scan profiles (smoothed) of Li+ (7 amu) and He+ (6 amu) ions due to He+ ion irradiation, as a function of the bias reference potential of the IDP

Full size image

Figure 10 shows the mean kinetic energy < \({E}_{{\mathrm{Li}}^{+}}\)> of sputtered Li+ for H, D, He+, Ne+, and Ar+ ion bombardment, which were determined by using the measured potential scan profiles. The mean kinetic energies of sputtered Li+ ions are in the range of < \({E}_{{\mathrm{Li}}^{+}}\)>  = 2–25 eV, which is one order of magnitude lower than the incident ion energy. The measured mean kinetic energies of the sputtered Li+ ions show similar profiles among the several incident ions. Simulations with VFTRIM-3D calculated an energy of 20 eV for sputtered Li particles from a 700 eV D+ ion beam at 45° incidence [17], and there is a good agreement (within 10%) of our experimental result with this simulation.

Fig. 10
figure 10

The mean kinetic energy < \({E}_{{\mathrm{Li}}^{+}}\)> of sputtered Li+ ions for He+, Ne+, Ar+, D, and H ion bombardment

Full size image

Summary, and Implication for Fusion Applications

We measured Li+ as the dominant positive ion and H−, O−, and OHx− as the dominant negative ions sputtered/ejected from an oxidized Li-coating (Li–O–H) on a stainless steel substrate under irradiation by H, D, He+, Ne+, and Ar+ ions at 45° incidence in the energy range of 30–2000 eV. The absolute sputtering yields and the mean kinetic energies were determined for those sputtered/ejected ions. The measured mean kinetic energies of sputtered Li+ ions during irradiations by H, D, He+, Ne+, and Ar+ ions were one order of magnitude lower than the incident ion energy. The incident ion energy on the plasma-facing walls in reactors is typically in the same order of magnitude as the plasma potential in divertor plasmas [57]. This balance of sputtered ion kinetic energy and sheath potential indicates that the sputtered Li+ ions are pushed back toward the divertor surface by the positive sheath potential, and, therefore, the erosion rate of Li can be greatly suppressed. The negative ions O−, OH−, and H−, are, on the contrary, drawn toward the bulk plasma due to the positive sheath potential. These findings give us a picture, as shown in Fig. 11, that ion bombardment works as a surface cleaning mechanism, i.e., removing oxygen and hydrogen, and suppressing Li erosion. This mechanism would apply during D-T operations because Li+, O−, OH−, and H− ion sputtering/ejection were observed for D and He+ ion irradiations at incident energies of E > 100 eV, which is in a fusion operational regime. Our experimental results also give a prediction for efficient Ar or Ne glow discharge cleaning on Li surfaces, since sputtering rates of O− and H− ions were found higher than 0.01 for Ar+ and Ne+ ion bombardments at incident energies of E > 100 eV. These results also suggest that Ar or Ne gas, which can be injected to induce plasma detachment during plasma operation, may also enhance Li surface cleaning.

Fig. 11
figure 11

A schematic of the expected behavior of sputtered/ejected positive (Li+) and negative (H−, O−, OH−) ions due to ion (X+) bombardment at a Li–O–H surface in a fusion reactor under the sheath potential

Full size image