Hydrogen retention in lithium and lithium oxide films

Journal of Nuclear Materials 502 (2018) 161e168 Contents lists available at ScienceDirect Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat Hydrogen retention in lithium and lithium oxide films
rrez b, A.O. Nelson c, M. Hofman d, P.S. Krsti
L. Buzi a, *, Y. Yang a, F.J. Domínguez-Gutie c b, c a R. Kaita , B.E. Koel a Department of Chemical and Biological Engineering, Princeton University, Princeton, NJ 08544-5263, United States Institute for Advanced Computational Science, Stony Brook University, NY 11794-5250, United States c Princeton Plasma Physics Laboratory, Princeton, NJ 08543-0451, United States d Department of Chemistry, Princeton University, Princeton, NJ 08544, United States b a r t i c l e i n f o a b s t r a c t Article history: Received 2 October 2017 Received in revised form 2 February 2018 Accepted 6 February 2018 Available online 9 February 2018 Pure lithium (Li) surfaces are difficult to maintain in fusion devices due to rapid oxide formation, therefore, parameterizing and understanding the mechanisms of hydrogen (H, D) retention in lithium oxide (Li2O) in addition to pure Li is crucial for Li plasma-facing material applications. To compare H retention in Li and Li2O films, measurements were made as a function of surface temperature (90e520 K) under ultrahigh vacuum (UHV) conditions using temperature programmed desorption (TPD). In both cases, the total retention dropped with surface temperature, from 95% at 90 K to 35% at 520 K Li2O films retained H in similar amounts as pure Li. Molecular Dynamics (MD) modeling was used to elucidate the mechanisms of H retention, and results were consistent with experiments in terms of both retention fraction and the drop of retention with temperature. © 2018 Elsevier B.V. All rights reserved. Keywords: Hydrogen retention Lithium Lithium oxide Nickel single crystal Molecular dynamics 1. Introduction Li conditioning of plasma facing components (PFCs) has improved plasma performance with energy confinement and lowered H recycling in magnetic fusion devices [1e4], and suppressed edge-localized modes (ELMs) in NSTX [5,6]. Li conditioning of the NSTX divertor resulted in significant reduction (50%) of the heat load due to enhanced radiation [7,8]. These effects may be due in part to lithium's efficiency in binding H isotopes, thereby increasing the H retention and lowering the recycling of these species. Accordingly, an understanding of H retention mechanisms and parameterization of H uptake in Li is needed for future applications of Li in high heat-flux and long-pulse duration machines, as well as for H storage applications [9e11]. Such information is also needed for Li2O because at typical base pressures of 1  10 8 Torr, which is not uncommon in tokamaks, Li2O can form rapidly (2Li þ H2O/Li2O þ H2 [12]). For instance, in NSTX the walls were exposed to 100e600 L (1 L ¼ 1  10 6 Torr  sec) of water vapor inbetween the plasma shots, during which the Li oxidation occurred. * Corresponding author. E-mail address: [email protected] (L. Buzi). https://doi.org/10.1016/j.jnucmat.2018.02.010 0022-3115/© 2018 Elsevier B.V. All rights reserved. According to the residual gas analysis, the gas consisted of 77% of hydrogenic species (mass 2, 3, 4) and 18% of water vapor (mass 17, 18, 19, 20) [13]. Li oxidation has also been observed in in-vacuo measurements of Li-coated samples of PFCs in LTX using the Materials Analysis Particle Probe (MAPP) [14e16]. The problem will continue to be important for the high-Z PFC phase of operation in NSTX-U in which Li2O is also quickly formed, during Li evaporation, under typical water partial pressure conditions in the 1  10 9 Torr range. Several studies have addressed the mechanisms of H retention in Li. Baldwin et al., [17] measured deuterium (D) retention in Li as a function of ion fluence and reported full uptake of D until volumetric conversion to LiD. Ion fluences beyond saturation led to a switch from the low to the high recycling regime (i.e. high retention to low retention), independent of the Li temperature (523e673 K). Taylor et al., [18,19], after analyzing NSTX tiles and performing exsitu experiments on ATJ graphite samples, reported D bonding with Li after Li interacted with carbon and oxygen. Krsti
c et al. demonstrated that O concentrations could increase up to 40 at% with significant D fluence, which was later reconfirmed by experiments by Taylor et al. [20]. Using quantum classical molecular dynamics calculations (QCMD) on lithiated graphite surfaces, Krsti
c et al. [21,22] showed that D is bound to O containing complexes rather 162 L. Buzi et al. / Journal of Nuclear Materials 502 (2018) 161e168 than to Li, thus promoting oxygen as having a main role in D retention. MAPP results on LTX indicated that it is not crucial to have just elemental Li to bind H, and Li oxide could also act as a binding agent [16]. However, in order to gain a better understanding and evaluate Li oxide's ability to reduce recycling, the efficiency of Li oxide compared to elemental Li in retaining H needs further investigation. This work is built upon previous experiments where an ultrathin (3 monolayer, ML) Li film was deposited on a polycrystalline TZM sample and irradiated with D ions [23]. Release of oxygen from TZM and subsequent oxidation of Li films during TPD, as well as the complex Li-TZM interface, introduced uncertainties into interpreting those experiments. In order to address these issues, we have conducted a systematic study in which five to seven times thicker Li films were deposited on a nickel (Ni) single crystal. The purity of the Li film was checked with auger electron spectroscopy (AES). Modeling of the results is done by Molecular Dynamics (MD) using REAXFF bond-order potentials [24,25], with correction for dynamical polarization effects due to difference in Li, H and O electronegativities by Electronegativity Equalization Method (EEM) [26,27]. 2. Experimental setup All experiments were performed in a stainless steel ultrahigh vacuum (UHV) chamber with a 2  10 10 Torr base pressure. Low energy electron diffraction (LEED) was performed by using PHI 15e120 LEED optics. AES was done using a PHI 15-255G doublepass cylindrical mirror analyzer (CMA). TPD experiments were performed with the sample in line-of-sight of the ionizer of a shielded UTI 100C quadrupole mass spectrometer (QMS), with the shield nozzle located 1 mm from the sample, using a heating rate of 4 K/s. A K-type thermocouple (chromel-alumel) was spot-welded directly on the sample to monitor the temperature. The Ni(110) single crystal (8 mm square, ±0.5 orientation) sample was spot-welded onto tantalum (Ta) wires used for resistive heating. The crystal was cleaned using cycles of 1.5-keV Arþ ion sputtering and annealing in vacuum at 1100 K. Oxidation in 4  10 8 Torr O2 with the sample at 1000 K was used to eliminate residual carbon, and reduction in 4  10 8 Torr H2 with the sample at 1000 K was used to eliminate residual oxygen. Good surface order was confirmed with LEED. Surface purity was checked with AES to ensure carbon and oxygen concentrations were below 1%. The quality of the Ni(110) surface was also confirmed to be good using the position and shape of the H2 desorption peaks in TPD [28]. In these experiments, we used a Ni (110) single crystal as a substrate to avoid effects due to grain boundaries, intrinsic defects, and impurities diffusing to the surface. Moreover, due to the low solubility of alkali metals in Ni, Li and Ni are immiscible, and thus do not form either bulk alloys or two-dimensional surface alloys [9,29]. Li dosing was performed with a commercial Li metal dispenser (Li/NF/7.3/17/FT, SAES Group) by thermal evaporation onto the Ni substrate. Hþ 2 ions were produced in a differentially pumped ion gun (PHI 04e303 A) with adjustable ion energy from 0 to 5 keV, and a liquid nitrogen trap was used in the H2 gas line to mitigate H2O contamination. H2 gas (Praxair, 99.999%) and O2 gas (Praxair, 99.995%) was introduced into the chamber following a liquid nitrogen trap and using a high precision variable leak valve. Hydrogen, rather than deuterium, was used in these experiments for convenience and H has the same chemistry as D. After surface preparation, Li films were exposed to a 500 eV Hþ 2 ion beam, which þ þ is nominally composed of 90% Hþ 2 and 10% H [30]. The H2 flux was defocused over the surface and the current measured on the sample was 1.74 mA. The H2 pressure in the chamber during Hþ 2 irradiation was 4  10 8 Torr and the total exposure time was 120 s. These conditions provided a total fluence of 4  1015 Hþcm areal density was ~3  1015 Li cm 2, as discussed later. 2 . The Li 3. Computational approach We simulated these experiments by using Molecular Dynamics. Amorphous target surfaces of pure Li and Li2O were prepared for a set of temperature values T (90, 300, 400, 500, and 600 K), following the procedure in Refs. [21,22] for each temperature. Computational cells of about 2000 atoms were used. These amorphous cells were created initially at 300 K, one with random distribution of lithium atoms, and another one with a predefined random distribution of 33% O and 67% of Li atoms. These cells were energy optimized in a succession of heating (1000 K max) and annealing processes, and finally thermalized to a desired temperature using a Langevin thermostat with time constant of 100 fs. The final numbers of atoms in the prepared cells at various temperatures are shown in Table 1. Periodic boundary conditions were applied in the x-y directions to simulate an infinite surface slab, with D impact in the z-direction. The lateral dimensions of the cells were 3.6 nm in z direction and about 3.4 nm in x and y directions for both surfaces (Fig. 1). The cell depth of 3.6 nm is sufficient to prevent penetration of the D atoms to the cell bottom boundary, thus avoiding artificial reflections. The atomistic simulations were performed by MD, using Large Scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) [31] with Reactive Force Field (ReaxFF) Bond Order (BO) potential [24,25,32], and corrections for dynamic atom charge effects by semi-empirical EEM [26,27]. The classical ReaxFF potentials used for Li, O, D were verified in our previous computations [33] of retention and sputtering of Li-C-O-D surfaces by Quantum-Classical Molecular Dynamics (QCMD), using Self Consistent Charge Tight Binding Density Functional Theory (SCC-DFTB) [34]. The ReaxFF potentials implemented in LAMMPS are able to model the dynamics of breaking and forming of chemical bonds [24,32], as well as to calculate the dynamic changes in charges of the atoms in the system with the change of atomic coordinates using the Electronegativity Equalization Method [26,27]. The latter is important in the presence of mutually polarizable materials with very different electronegativities, such as Li (0.94) and O (3.4), while the H electronegativity (2.2) is in the middle. The prepared computational cells, with various atomic contents and at various temperatures, were bombarded by N ¼ 5040 independent 10 eV D atoms, with trajectories starting 1 nm at random location above the surface, in the direction orthogonal to the surface. This large number of trajectories led to the adequate statistics of H retention probability and was done at supercomputing facilities using “embarrassing parallelization”. If the number of D atoms retained in the surface is ND, then the retention probability per D is calculated as ND/N. The retention chemistry of D evolves at the end of the collision cascade when the impact particle is thermalized allowing comparison with the experimental results at higher impact energies [21]. We carry out the analysis of the resulting chemistry after the final rest location of each D impact, by performing the nearest-neighbor (NN) estimation for each retained D (H) atom in the surface, defining the most-probable bonds [21]. While role of Li in bonding hydrogen is not challenged in the purelithium surface, it is surprising that O-D and Li-D NNs are similarly represented. Although there is two times more Li than oxygen atoms in the surface, oxygen has coordination number 2, two times larger than coordination number of Li (1). This indicates approximate similar efficiency of H (D) retention in Li and Li2O surfaces. We note that the O-D bond percentage is slightly bigger than Li-D one, and that effectiveness of O to bond D slightly increases with temperature. 163 L. Buzi et al. / Journal of Nuclear Materials 502 (2018) 161e168 Table 1 Atomic content for Li and Li2O target surfaces at various temperatures after energy optimization and thermalization processes. Temperature (K) Pure Li Li2O Li atoms Li atoms O atoms 90 305 400 500 600 2000 1340 660 2000 1340 660 1998 1335 658 1995 1329 656 1991 1320 658 Fig. 1. a) Li system (2000 lithium atoms), b) Li2O system: 33% of O (660 atoms) and 67% of Li (1340 atoms). Green and red symbols represent Li and O, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) 4. Results and discussion Auger electron spectroscopy (AES) analysis was done on the sample surface to determine the elemental composition. Electrons with energy 3 keV were used to irradiate the sample and the generated Auger electrons were analyzed with a double-pass cylindrical mirror analyzer (CMA). Fast AES scans were taken on the pure Li film to avoid oxidation and/or other impurity deposition. AES spectra of the Li metal films on Ni (110) showed a Li peak at 51 eV, which is characteristic of pure metallic Li, as shown in Fig. 2. Carbon and oxygen impurity concentrations in the film, as detected by AES and also carbon monoxide (CO) TPD, were less than 1%. When the Li film was oxidized by exposure to 10 L O2 at fixed exposure temperatures in the range 90e520 K, the Li AES transitions split into two negative-going peaks at 36 and 42 eV, indicating the formation of Li2O [9,35,36]. No metallic Li signal in AES remained. The estimated Li film thickness was 5 nm; slightly below the AES probing depth (~10 nm) therefore the Ni peaks are still visible. Fig. 2. AES of (top) Li metal and (bottom) Li2O films on Ni(110). A total fluence of 4  1015 Hþ cm 2 of 500 eV Hþ 2 ions was used to irradiate the Li and Li2O films after deposition. TPD was performed within 2e5 min after ion irradiation, time needed to move the manipulator and place the sample in front of the mass spectrometer. Li and H2 TPD profiles are shown in Fig. 3. The initial Li coverage before irradiation was checked with TPD at room temperature and its reproducibility was within 1e2%. During the Li and Hþ 2 exposures, the Ni substrate was kept at a fixed exposure temperature, ranging from 90 to 520 K. Li and H2 desorption curves are given in Fig. 3 for each exposure temperature in order to illustrate this influence, however we will focus our discussion below on the two highlighted curves at 90 and 520 K. As seen in Fig. 3(a), the amount of Li desorbing in the multilayer peak near 600 K starts to decrease when exposure temperatures above 450 K are used, since Li starts evaporating significantly at temperatures above 450 K [37]. The higher temperature Li desorption (700e1000 K), as shown more clearly on an expanded y scale in Fig. 3(a1), corresponds to desorption from the Li monolayer [37]. Hþ 2 irradiation caused the formation of LiH and a new Li desorption peak near 650 K due to LiH decomposition [9]. This Li peak is coincident with the H2 desorption peak, shown in Fig. 3(b), demonstrating that Li and H2 evolution in TPD is rate-limited by LiH decomposition in the films, as previously reported [9]. The H2 TPD curves show that the amount of H retained in the Li film at 90 K is higher than that at 520 K and only a slight decrease was observed from 300 K to 520 K. The amount is quantitatively analyzed in the next section. No LiH (8 amu) desorption was detected in these experiments. To understand the H retention measurements, we need a determination of the absolute coverage (atoms/cm2) for both Li and H in the films. The Li coverage was calibrated by determining the Li TPD area from a Li film that formed a saturation monolayer coverage (just prior to formation of a Li multilayer desorption feature) and assigning the integrated area under this Li TPD curve, after correction for the mass spectrometer sensitivity to the translational energy of the desorbed species [38], to a Li coverage corresponding to a hexagonal close packed structure of Li adatoms with their metallic atomic radii (3.02  1015 atoms/cm2) [37]. The surface coverage (q) is given in ML, where 1 ML corresponds to the 164 L. Buzi et al. / Journal of Nuclear Materials 502 (2018) 161e168 LiOH (m/z ¼ 24) desorption was detected in TPD. TPD curves for these desorbed products are provided in Fig. 4(aec). No desorption of LiO, H2O, or LiH was observed. Fig. 4(a) shows that Li is retained at the Ni surface to higher temperatures after oxidation, due to the increased thermal stability of Li2O. In this case, Li is only desorbed when Li2O decomposes, which produces a Li TPD peak at 900 K. The temperature was increased up to 1200 K during TPD due to the limitations with the thermocouple operation therefore the highest temperature peak was not completely captured. For 300 K exposure case (Fig. 4(a)), 18% of the Li layer was not oxidized since a Li metallic peak at 570 K was still present (2.2 ML) and about 10% of the Li layer (1.2 ML) formed the LiH (Li peak at Fig. 3. (a) Li and (b) H2 TPD curves after pure Li films at 90e520 K were irradiated with 500 eV Hþ 2 . Panel (a1) shows data from panel (a) on an expended y scale. surface Ni atom density of Ni(110) of 1.14  1015 atoms/cm2. This gives an absolute Li coverage, qLi, of 2.56 ML referenced to the Ni(110) surface atom density. The H coverage was determined in a similar method by determining the integrated area under the H2 TPD profiles. We assign a surface saturation of qH ¼ 1.5 ML (1.71  1015 atoms/cm2) when produced from dissociative adsorption of H2 on Ni(110) at 300 K [28,39]. Similar experiments to those described above were done for Li2O films. These films were oxidized by the same exposure of 10 L O2 at fixed exposure temperatures in the range 90e520 K. After Hþ 2 irradiation of Li2O films with Hþ 2 ions, H2 (m/z ¼ 2), Li (m/z ¼ 7) and Fig. 4. (a) Li and (b) H2 TPD curves after Li2O films at 90e520 K were irradiated with 500 eV Hþ 2. L. Buzi et al. / Journal of Nuclear Materials 502 (2018) 161e168 630 K). The exposure to 400 K produced a Li peak (LiH decomposition temperature) at 630 K (Fig. 4(a)), indicating that about 27% of the Li film formed LiH (4.7 ML). H2 TPD showed a low temperature peak at 630 K only for 90 K and 300 K exposure temperature (Fig. 4(b)). When the temperature increased beyond 300 K, only the high temperature peak remained at 900 K. It is not understood why the H2 peak is missing for the 400 K exposure, and presumably this is related to another channel for the consumption of H2 during or prior to TPD, but this requires further studies. LiOH (m/z ¼ 24), Li2O (m/z ¼ 30) and O2 (m/z ¼ 32) were also monitored and desorbed at 900 K, although the calibration for these species is not provided in this work (see LiOH decomposition in Ref. [40]). It is important to point out that no H was measured from TPD experiments after irradiation with 500 eV Hþ 2 ions on Ni at room temperature, therefore the hydrogen retention in these experiments can be attributed solely to the Li/Li2O layer. Energetic Hþ 2 ions (500 eV) get implanted in the first few nm of the Ni substrate, but since Ni does not retain H at room temperature and above, H immediately diffuses out of Ni at these temperatures and reacts with the Li/Li2O layer. In this case Ni serves as a reservoir (virtual source) of thermal H, which interacts with Li/Li2O and forms other compounds such as LiH, LiOH etc. From this point of view, the 10 eV impact energy used in the MD calculation (in which case the impact particles do not reach the bottom of the slab) is applicable to the experimental results. This is supported by the fact that the retention chemistry develops mainly when impact particles thermalize [21], irrespectively of their impact energy. AES spectra were obtained from all of these surfaces at fixed exposure temperatures in the range 90e520 K after Hþ 2 irradiation and these are given in Fig. 5. As a reference, AES spectra from the clean Ni(110) surface and pure Li on Ni(110) at 90 K is provided in Fig. 5(a) and (a1). Also, AES spectra from the oxidized Li film surface at 90 K prior to Hþ 2 irradiation is provided in Fig. 5(b) and (b1). The top panels of Fig. 5 show AES spectra after Hþ 2 irradiation of pure Li for the Li (a) and O (a1) regions. In Fig. 5(a), at 90 K, Hþ 2 irradiation causes no shift in the metallic Li peak at 51 eV; 500 eV Hþ 2 irradiation of the 5 nm Li film mostly results in H implanted in 165 Ni, and at 90 K this H is not mobile, and remains primarily in the Ni substrate, and so the metallic Li signal in AES is maintained. The penetration depth in Ni for 500 eV Hþ 2 ions, calculated by the SRIM code [41] is 6 nm and it scales up to 10 nm for the Li/Ni system. Li normally attenuates the Ni peak at 61 eV, except on the curve labeled “a”, where Li exposure occurred at 300 K. In the 300 K case, the Li layer may have been slightly thinner. Alternatively, the “as dosed Li” may have not formed a uniform layer at 300 K, exposing some Ni, whereas at higher temperatures Li diffuses to form a more uniform layer covering the Ni surface. The O AES spectra shown in Fig. 5(a1) show a small oxygen peak at 512 eV, which is still less than 2% in the film. The bottom panels of Fig. 5 provide AES spectra taken after Hþ 2 irradiation of Li2O films for (b) Li and (b1) O AES regions. Exposure of a 5 nm thick Li film to 10 L O2 oxidizes the Li to create a Li2O film. As shown in Fig. 5(b), this causes the metallic Li AES peak at 51 eV to be eliminated and two peaks near 36 and 42 eV, characteristic of Li2O, to appear. However, also LiH has peaks in the range 36e42 eV [9,35]. In order to understand whether the peak represents LiH or Li2O (peak at 40þ-1 eV), we look at the shift in the O peak [35,42]. At temperatures up to 470 K, O peaks appear at 508e509 eV (LiOH), and at higher temperatures it shifts to 512 eV (Li2O). In the literature [36] it is shown that slow conversion of LiOH to Li2O and H2O occurs on the order of minutes. In the present experiments, TPD was performed 2e5 min after Hþ 2 irradiation, therefore, at temperatures higher than 470 K, it is believed that LiOH decomposed to form Li2O and H2O according to the reaction: 2LiOH> Li2O þ H2O (g). H2O formation between Hþ 2 exposure and TPD may be another cause of the H retention drop at elevated temperatures. In Fig. 6, the total amounts of Li and H2 are plotted for both cases, i.e., pure and oxidized Li. The data points are connected with a spline fit to guide the eye and the dashed and dotted lines in the graph show the initial Li coverage in Li and Li2O experiments before Hþ 2 ion irradiation. For the oxidized Li film, the relative amount of Li desorbed from the irradiated Li2O film is lower than for the Li film indicating that the Hþ 2 sputtering of pure Li is higher than þ Fig. 5. a) Li and a1) O AES signal after Hþ 2 exposure of pure Li and b) Li and b1) O AES signal after H2 exposure of Li2O. 166 L. Buzi et al. / Journal of Nuclear Materials 502 (2018) 161e168 Fig. 6. Concentrations of H and Li from TPD measurements after irradiation of Li films. The dashed and dotted lines show initial Li coverage in Li and Li2O experiments. sputtering of Li2O. We note no large effect of exposure temperatures on H retention exceeding the Li melting point. This was to be expected since the H retaining compounds, LiH and LiOH, decompose at much higher temperatures than pure Li (see Figs. 3 and 4). These measurements were carried out below the saturation H uptake level (Li:H ¼ 1) to more clearly identify changes in retention in the films due to exposure temperature. Results from both experiments and MD simulations are compared in Fig. 7(a) and (b). In Fig. 7(a), the retention fraction of H is plotted for pure and oxidized Li and compared with computed MD results for D. The total incident H was calculated by the TPD profile of H retained in Ni (110) at 90 K (8.2 ML/Ni). The retention for Li and Li2O is rather similar. We have to keep in mind that retention in the case of oxidized Li may be higher than shown due to the release of LiOH, which is not calibrated and not accounted for in this graph. Noise/scattering in the data is attributed to the statistical error due to the integration of the TPD profiles, and the 10% error bar is associated with the inherent uncertainty of the measuring instruments and the reproducibility of Hþ 2 dosing. At Fig. 7. H retention fraction (H retention/H incident) in Li and Li2O as a function of exposure temperature and comparison with D and H retention from MD calculations. temperatures higher than 450 K (Fig. 6(b)), Li coverage starts to drop rapidly due to evaporation, which may also contribute to a larger scatter in the data for H retention. For each computed data point, we report a standard error defined as sS ¼ sN 1/2, where s is the standard deviation of our sample of N cases. In summary, the MD calculations demonstrated that the oxidation of the Li slab results in the formation of various Li oxides and the retained D atoms are located in the interstitial positions of Li or bonded to oxygen. We also calculate the retention of the same samples but with impact of H. The values obtained for the retention are 89 ± 2.5% at 90 K, 72 ± 3% at 300 K and 62 ± 3% at 500 K. A comparison between the computed values of H and D retention and experimental results for H retention is provided in Fig. 7(b). Experiments and calculations consistently show retention by both pure Li and Li2O, indicating that both pure and oxidized Li can bind H with similar efficiency. The same drop of retention with Texp is followed for both Li and Li2O. The exponential decay of retention with temperature may originate from diffusion, since it is a thermally activated process (i.e., higher Texp leads to higher H desorption from the Li surface.) An interesting observation was the comparison between H and D in terms of total retention in pure Li (Fig. 7(b)). From the calculations, one can see a small drop in H retention compared to D. Although H and D are chemically similar, their different mass could lead to different diffusion rates in Li. As a final comment, a similar chemistry is expected for tritium and since tritium inventory is recognized as an important challenge for a fusion reactor, possible solutions are being considered (see Ref. [44]). 5. Conclusions H retention in pure Li and Li2O films was calculated by molecular dynamics and measured experimentally as a function of surface temperature under UHV conditions. It was experimentally shown that upon oxidation, Li thermal stability increased. It was also shown that both pure Li and Li2O are able to retain H in almost same amounts. In addition, it was shown that H and D retention drops with surface temperature in the range 90e520 K from 95% to 35% due to outwards diffusion of H at high temperatures. Results of the MD modeling with EEM corrections were qualitatively consistent with experimental results in terms of both retention fraction and the drop of retention with temperature, when using either H or D as impact particles. Similar trends and agreements were seen between the experimental results for Li2O and the MD results for a mixture of Li and oxygen with atomic concentration ratio Li:O ¼ 2:1. In TPD measurements, the Li and H desorption peaks were observed at the same temperature (650 K) and this was interpreted as, corresponding to the decomposition of LiH. In addition to TPD analysis for Li2O films irradiated with H ions, AES measurements were consistent with the formation of LiOH, which decomposed to Li2O and H2O at exposure temperatures higher than 470 K. The results from experiment and MD modeling do not preclude the possibility that the hydrogen is retained through the formation of large LinHn molecules. Earlier DFT studies, for example, predicted the formation of a rock-salt structure for LiD under deuterium bombardment of lithium films [43]. The Li and H could then bind to form LiH at the surface, which would subsequently dissociate at the same temperature as observed experimentally. The details of the chemistry underlying H retention and Li and Li2O films may thus be complex. Nevertheless, both experiment and modeling support the possibility that low H recycling can be achieved if Li2O is formed under fusion reactor conditions, as both Li and Li2O have comparable efficiency for trapping H. L. Buzi et al. / Journal of Nuclear Materials 502 (2018) 161e168 Note that the digital data for the figures in this paper can be found at http://arks.princeton.edu/ark:/88435/dsp01x920g025r. [14] Acknowledgements [15] This material is based upon work supported by the U.S. Department of Energy, Office of Science/Fusion Energy Sciences under Award DE-SC0012890 (BEK) and Award DE-SC0013752 (PSK). FJDG acknowledges support by a Postdoctoral Scholarship Number 267898 of the National Council of Science and Technology, Mexico. [16] [17] References [18] [1] D.K. Mansfield, K.W. Hill, J.D. Strachan, M.G. Bell, S.D. Scott, R. Budny, E.S. Marmar, J.A. Snipes, J.L. Terry, S. Batha, R.E. Bell, M. Bitter, C.E. Bush, Z. Chang, D.S. Darrow, D. Ernst, E. Fredrickson, B. Grek, H.W. Herrmann, A. Janos, D.L. Jassby, F.C. Jobes, D.W. Johnson, L.C. Johnson, F.M. Levinton, D.R. Mikkelsen, D. Mueller, D.K. Owens, H. Park, A.T. Ramsey, A.L. Roquemore, C.H. Skinner, T. Stevenson, B.C. Stratton, E. Synakowski, G. Taylor, A. von Halle, S. von Goeler, K.L. Wong, S.J. Zweben, Enhancement of tokamak fusion test reactor performance by lithium conditioning, Phys. Plasmas 3 (1996) 1892, https://doi.org/10.1063/1.871984. [2] M.G. Bell, H.W. Kugel, R. Kaita, L.E. Zakharov, H. Schneider, B.P. LeBlanc, D. Mansfield, R.E. Bell, R. Maingi, S. Ding, S.M. Kaye, S.F. Paul, S.P. Gerhardt, J.M. Canik, J.C. Hosea, G. Taylor, Plasma response to lithium-coated plasmafacing components in the National Spherical torus experiment, Plasma Phys. Contr. Fusion 51 (2009) 124054, https://doi.org/10.1088/0741-3335/51/12/ 124054. [3] R. Majeski, R. Doerner, T. Gray, R. Kaita, R. Maingi, D. Mansfield, J. Spaleta, V. Soukhanovskii, J. Timberlake, L. Zakharov, Enhanced energy confinement and performance in a low-recycling tokamak, Phys. Rev. Lett. 97 (2006) 1e4, https://doi.org/10.1103/PhysRevLett.97.075002. [4] J.C. Schmitt, R.E. Bell, D.P. Boyle, B. Esposti, R. Kaita, T. Kozub, B.P. LeBlanc, M. Lucia, R. Maingi, R. Majeski, E. Merino, S. Punjabi-Vinoth, G. Tchilingurian, A. Capece, B. Koel, J. Roszell, T.M. Biewer, T.K. Gray, S. Kubota, P. Beiersdorfer, K. Widmann, K. Tritz, High performance discharges in the Lithium Tokamak eXperiment with liquid lithium walls, Phys. Plasmas 22 (2015), https:// doi.org/10.1063/1.4921153. [5] R. Maingi, S.M. Kaye, C.H. Skinner, D.P. Boyle, J.M. Canik, M.G. Bell, R.E. Bell, T.K. Gray, M.A. Jaworski, R. Kaita, H.W. Kugel, B.P. Leblanc, D.K. Mansfield, T.H. Osborne, S.A. Sabbagh, V.A. Soukhanovskii, Continuous improvement of H-mode discharge performance with progressively increasing lithium coatings in the National Spherical Torus Experiment, Phys. Rev. Lett. 107 (2011), https://doi.org/10.1103/PhysRevLett.107.145004. [6] R. Maingi, D.P. Boyle, J.M. Canik, S.M. Kaye, C.H. Skinner, J.P. Allain, M.G. Bell, R.E. Bell, S.P. Gerhardt, T.K. Gray, M. a. Jaworski, R. Kaita, H.W. Kugel, B.P. LeBlanc, J. Manickam, D.K. Mansfield, J.E. Menard, T.H. Osborne, R. Raman, A.L. Roquemore, S. a. Sabbagh, P.B. Snyder, V. a. Soukhanovskii, The effect of progressively increasing lithium coatings on plasma discharge characteristics, transport, edge profiles and ELM stability in the National Spherical Torus Experiment, Nucl. Fusion 52 (2012) 83001, https://doi.org/10.1088/00295515/52/8/083001. [7] M. Ono, M.A. Jaworski, R. Kaita, H.W. Kugel, J.-W. Ahn, J.P. Allain, M.G. Bell, R.E. Bell, D.J. Clayton, J.M. Canik, S. Ding, S. Gerhardt, T.K. Gray, W. Guttenfelder, Y. Hirooka, J. Kallman, S. Kaye, D. Kumar, B.P. Leblanc, R. Maingi, D.K. Mansfield, A. McLean, J. Menard, D. Mueller, R. Nygren, S. Paul, M. Podesta, R. Raman, Y. Ren, S. Sabbagh, F. Scotti, C.H. Skinner, V. Soukhanovskii, V. Surla, C.N. Taylor, J. Timberlake, L.E. Zakharov, Recent progress in the NSTX/NSTX-U lithium programme and prospects for reactorrelevant liquid-lithium based divertor development, Nucl. Fusion 53 (2013), https://doi.org/10.1088/0029-5515/53/11/113030. [8] T.K. Gray, J.M. Canik, R. Maingi, a. G. McLean, J.-W. Ahn, M. a. Jaworkski, R. Kaita, M. Ono, S.F. Paul, The effects of increasing lithium deposition on the power exhaust channel in NSTX, Nucl. Fusion 54 (2014) 23001, https:// doi.org/10.1088/0029-5515/54/2/023001. [9] J. Engbæk, G. Nielsen, J.H. Nielsen, I. Chorkendorff, Growth and decomposition of lithium and lithium hydride on nickel, Surf. Sci. 600 (2006) 1468e1474, https://doi.org/10.1016/j.susc.2006.01.040. [10] L. Wang, M.Z. Quadir, K.F. Aguey-Zinsou, Direct and reversible hydrogen storage of lithium hydride (LiH) nanoconfined in high surface area graphite, Int. J. Hydrogen Energy 41 (2016) 18088e18094, https://doi.org/10.1016/ j.ijhydene.2016.07.073. [11] L. Wang, M.Z. Quadir, K.F. Aguey-Zinsou, Ni coated LiH nanoparticles for reversible hydrogen storage, Int. J. Hydrogen Energy 41 (2016) 6376e6386, https://doi.org/10.1016/j.ijhydene.2016.01.173. [12] C. Haertling, R.J. Hanrahan, R. Smith, A literature review of reactions and kinetics of lithium hydride hydrolysis, J. Nucl. Mater. 349 (2006) 195e233, https://doi.org/10.1016/j.jnucmat.2005.10.005. [13] C.H. Skinner, R. Sullenberger, B.E. Koel, M.A. Jaworski, H.W. Kugel, Plasma [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] 167 facing surface composition during NSTX Li experiments, J. Nucl. Mater. 438 (2013), https://doi.org/10.1016/j.jnucmat.2013.01.136. C.N. Taylor, B. Heim, S. Gonderman, J.P. Allain, Z. Yang, R. Kaita, A.L. Roquemore, C.H. Skinner, R.A. Ellis, Materials analysis and particle probe: a compact diagnostic system for in situ analysis of plasma-facing components (invited), Rev. Sci. Instrum. (2012), https://doi.org/10.1063/1.4729262. M. Lucia, R. Kaita, R. Majeski, F. Bedoya, J.P. Allain, D.P. Boyle, J.C. Schmitt, D.A.S. Onge, Development progress of the materials analysis and particle probe, Rev. Sci. Instrum. 85 (2014), https://doi.org/10.1063/1.4890257. R. Kaita, M. Lucia, J.P. Allain, F. Bedoya, R. Bell, D. Boyle, A. Capece, M. Jaworski, B.E. Koel, R. Majeski, J. Roszell, J. Schmitt, F. Scotti, C.H. Skinner, V. Soukhanovskii, Hydrogen retention in lithium on metallic walls from invacuo analysis in LTX and implications for high-Z plasma-facing components in NSTX-U, Fusion Eng. Des. 117 (2017) 135e139, https://doi.org/ 10.1016/j.fusengdes.2016.06.056. M.J. Baldwin, R.P. Doerner, S.C. Luckhardt, R.W. Conn, Deuterium retention in liquid lithium, Nucl. Fusion 42 (2002) 1318e1323, https://doi.org/10.1088/ 0029-5515/42/11/305. C.N. Taylor, B. Heim, J.P. Allain, Chemical response of lithiated graphite with deuterium irradiation, J. Appl. Phys. 109 (2011), https://doi.org/10.1063/ 1.3555097. C.N. Taylor, J.P. Allain, B. Heim, P.S. Krsti
c, C.H. Skinner, H.W. Kugel, Surface chemistry and physics of deuterium retention in lithiated graphite, J. Nucl. Mater. (2011), https://doi.org/10.1016/j.jnucmat.2010.09.049. C.N. Taylor, J.P. Allain, K.E. Luitjohan, P.S. Krsti
c, J. Dadras, C.H. Skinner, Differentiating the role of lithium and oxygen in retaining deuterium on lithiated graphite plasma-facing components, Phys. Plasmas 21 (2014), https://doi.org/10.1063/1.4874340. P.S. Krsti
c, J.P. Allain, C.N. Taylor, J. Dadras, S. Maeda, K. Morokuma, J. Jakowski, A. Allouche, C.H. Skinner, Deuterium uptake in magnetic-fusion devices with lithium-conditioned carbon walls, Phys. Rev. Lett. 110 (2013), https://doi.org/ 10.1103/PhysRevLett.110.105001. P.S. Krsti
c, J.P. Allain, A. Allouche, J. Jakowski, J. Dadras, C.N. Taylor, Z. Yang, K. Morokuma, S. Maeda, Dynamics of deuterium retention and sputtering of Li-C-O surfaces, Fusion Eng. Des. 87 (2012) 1732e1736, https://doi.org/ 10.1016/j.fusengdes.2011.07.009. A.M. Capece, J.P. Roszell, C.H. Skinner, B.E. Koel, Effects of temperature and surface contamination on D retention in ultrathin Li films on TZM, J. Nucl. https://doi.org/10.1016/ Mater. 463 (2015) 1177e1180, j.jnucmat.2014.10.048. A.C.T. Van Duin, S. Dasgupta, F. Lorant, W.A. Goddard, ReaxFF: a reactive force field for hydrocarbons, J. Phys. Chem. A 105 (2001) 9396e9409, https:// doi.org/10.1021/jp004368u. D. Bedrov, G.D. Smith, A.C.T. Van Duin, Reactions of singly-reduced ethylene carbonate in lithium battery electrolytes: a molecular dynamics simulation study using the ReaxFF, J. Phys. Chem. A (2012) 2978e2985, https://doi.org/ 10.1021/jp210345b. W.J. Mortier, S.K. Ghosh, S. Shankar, Electronegativity-equalization method for the calculation of atomic charges in molecules, J. Am. Chem. Soc. 108 (1986) 4315e4320, https://doi.org/10.1021/ja00275a013. Y. Cong, Z.Z. Yang, General atom-bond electronegativity equalization method and its application in prediction of charge distributions in polypeptide, Chem. Phys. Lett. 316 (2000) 324e329, https://doi.org/10.1016/S0009-2614(99) 01289-0. T.E. Jackman, K. Griffiths, W.N. Unertl, J.A. Davies, K.H. Gurtler, D.A. Harrington, P.R. Norton, Surface phases of Ni(110) induced by adsorption of deuterium, Surf. Sci. 179 (1987) 297e321, https://doi.org/10.1016/00396028(87)90060-4. K. Wang, P. Ross, Reversible Li deposition on Ni in ultrahigh-vacuum, J. Electrochem. Soc. 142 (1995) L95eL97, https://doi.org/10.1149/1.2044257. H. Tawara, Y. Itikawa, H. Nishimura, M. Yoshino, Cross sections and related data for electron collisions with hydrogen molecules and molecular ions, J. Phys. Chem. Ref. Data 19 (1990) 617e636, https://doi.org/10.1063/1.555856. S. Plimpton, Fast Parallel algorithms for short-range molecular dynamics, J. Comput. Phys. 117 (1995) 1e19, https://doi.org/10.1006/jcph.1995.1039. S.S. Han, A.C.T. van Duin, W.A. Goddard III, H.M. Lee, Optimization and application of lithium parameters for the reactive force field, ReaxFF, J. Phys. Chem. A 109 (2005) 4575e4582, https://doi.org/10.1021/jp051450m.
rrez, P.S. Krsti
F.J. Domínguez-Gutie c, Sputtering of lithiated and oxidated carbon surfaces by low-energy deuterium irradiation, J. Nucl. Mater. 492 (2017) 56e61, https://doi.org/10.1016/j.jnucmat.2017.05.014. M. Elstner, D. Porezag, G. Jungnickel, J. Elsner, M. Haugk, T. Frauenheim, S. Suhai, G. Seifert, Self-consistent-charge density-functional tight-binding method for simulations of complex materials properties, Phys. Rev. B 58 (1998) 7260e7268, https://doi.org/10.1103/PhysRevB.58.7260. R.E. Clausing, D.S. Easton, G.L. Powell, Auger spectra of lithium metal and lithium oxide, Surf. Sci. 36 (1973) 377e379, https://doi.org/10.1016/00396028(73)90268-9. G.L. Powell, O. Ridge, G.E. Mcguire, D.S. Easton, R.E. Clausing, M.C. Division, O. Ridge, O. Ridge, N. Division, U.C. Corporation, L.A.S. Laboratories, L. Alamos, Auger spectra of lithi~ hydride* 46 (1974) 345e357. V. Saltas, C.A. Papageorgopoulos, Adsorption of Li on Ni(110) surfaces at low and room temperature, Surf. Sci. 461 (2000) 219e230, https://doi.org/ 10.1016/S0039-6028(00)00588-4. Y.-L. Tsai, C. Xu, B.E. Koel, Chemisorption of ethylene, propylene and 168 L. Buzi et al. / Journal of Nuclear Materials 502 (2018) 161e168 isobutylene on ordered Sn/Pt(111) surface alloys, Surf. Sci. 385 (1997) 37e59. http://ac.els-cdn.com/S0039602897001143/1-s2.0-S0039602897001143main.pdf?_tid¼f06a722e-7933-11e7-867a00000aab0f02&acdnat¼1501865144_fe246f0f393ad353fe1e76f3379fb1d6. [39] B. Voigtl€ ander, S. Lehwald, H. Ibach, Hydrogen adsorption and the adsorbateinduced Ni(110) reconstruction- an EELS study, Surf. Sci. 208 (1989) 113e135, https://doi.org/10.1016/0039-6028(89)90039-3. [40] L.N. Dinh, C.M. Cecala, J.H. Leckey, M. Balooch, The effects of moisture on LiD single crystals studied by temperature-programmed decomposition, J. Nucl. Mater. 295 (2001) 193e204, https://doi.org/10.1016/S0022-3115(01)005487. [41] J.F. Ziegler, M.D. Ziegler, J.P. Biersack, SRIM - the stopping and range of ions in matter (2010), Nucl. Instrum. Methods Phys. Res. Sect. B Beam Interact. Mater. Atoms 268 (2010) 1818e1823, https://doi.org/10.1016/j.nimb.2010.02.091. [42] J.P. Tonks, M.O. King, E.C. Galloway, J.F. Watts, Corrosion studies of LiH thin films, J. Nucl. Mater. 484 (2017) 228e235, https://doi.org/10.1016/ j.jnucmat.2016.12.008. [43] M. Chen, T. Abrams, M.A. Jaworski, E.A. Carter, Rock-salt structure lithium deuteride formation in liquid lithium with high-concentrations of deuterium : a first-principles molecular dynamics study, Nucl. Fusion 56 (2016) 16020, https://doi.org/10.1088/0029-5515/56/1/016020. [44] C.A. Gentile, L.E. Zakharov, R.D. Woolley, C.J. Huber, Flowing Liquid Lithium for the Purpose of Reducing Tritium Inventory Levels in Fusion Energy Reactors, 2013. http://ieeexplore.ieee.org/stamp/stamp.jsp?tp¼&arnumber¼6635397.