Hydrogen and oxygen in-depth evolution during electrochemical hydrogenation/dehydrogenation of Y–Pd thin films analyzed by Glow Discharge Optical Emission Spectroscopy

Hydrogen and oxygen in-depth evolution during electrochemical hydrogenation/dehydrogenation of Y–Pd thin films analyzed by Glow Discharge Optical Emission Spectroscopy E. Rayón Encinas a, ⁎, R. Escobar Galindo b , J.M. Albella Martín b , E. Matveeva c a Instituto de Tecnología de Materiales, Technical University of Valencia, Camí de Vera s/n, E-46022, Valencia, Spain b Instituto de Ciencia de Materiales de Madrid, Cantoblanco, E-28049 Madrid, Spain c Institute for Nanophotonics Technology, Technical University of Valencia, Camí de Vera s/n, E-46022, Valencia, Spain Abstract Elemental in-depth analysis of the Y–Pd thin film under hydrogenation/dehydrogenation has been performed by Glow Discharge Optical Emission Spectroscopy (GDOES). This technique was able to detect qualitatively the changes in the composition of the layer, particularly the indepth H and O distribution during hydrogen uptake and release. The charge of the film with hydrogen was done under electrochemically controlled conditions, while the discharge was effectuated in different modes: electrochemically, in alkali electrolyte (forced anodic discharge) or spontaneously, in air ambient or in the electrolyte (open potential circuit discharge). In-depth GDOES profiles of hydrogen at different times during the charge/discharge process confirm the processes of capture/release of hydrogen into the film corresponding to a reversible transition between the two hydrogenation states of yttrium: di- and tri-hydrides. The study of the oxygen profile reveals the presence of this element in the as prepared films and demonstrates its further accumulation under cyclic hydrogenation/dehydrogenation. Anodic discharge of the electrode rapidly increases oxygen content and probably leads to the formation of hydroxides along the whole yttrium depth with the consequent deterioration of the optical switching properties. . Keywords: Yttrium hydrides; Switchable mirrors; Palladium hydrides; Electrochemical hydrogenation; Optical switching; Radio-frequency glow discharge optical emission spectroscopy; Depth profiling; Smart optical windows 1. Introduction Since 1996, when reversible optical transition was discovered in yttrium thin films (100–300 nm) [1,2] under hydrogenation–dehydrogenation processes, many other materials have been proposed to develop and improve optical switching devices based on this phenomenon. Among them, the rare earth metals allowed with Mg were first proposed (second generation optical switching materials) [3] and then alloys of Mg with Ni, Co, Mn, Ti [4–6] were developed (third generation). The hydrogenation process can be performed by a controllable change of hydrogen pressure in the gaseous atmosphere [7,8], or under application of cathodic polarization in the electrochemical setup [9,10]. In both cases a thin film cap of palladium (5–15 nm) is deposited onto the active layer to protect against oxidation and catalyze H 2 dissociation. In the electrochemical mode, the hydrogen obtained by water electrolysis (is negative polarization of the thin film cathode) penetrates by diffusion into the alloy layer and converts it in a hydride. Alkaline solutions (typically 1–6 M KOH) are generally used as electrolytes. This is necessary as materials possessing hydrogenation ability are only stable at high pH (pHN13). Using neutral or acidic electrolytes provokes fast corrosion (oxidation) of the active metal [11]. The advantage of the electrochemical hydrogenation is an easier control of the hydrogen concentration in the material that is directly related to ⁎ Corresponding author. Universidad Politécnica de Valencia. Instituto de Tecnología de Materiales. Edif 5E. Camí de Vera s/n, E-46022. Valencia, Spain. Tel./fax: +34 667 238511. E-mail address: [email protected] (E.R. Encinas). the level of transparency–reflectance [12].However,this method implies using of aggressive electrolytes that may oxidize the active layer even protected by thin Pd film, leading to a rapid deterioration. In devices that use any of both methods of control, it is desirable to achieve longer lifetime and stability under cyclic optical switching. However, in practice, the switchable mirrors upon working typically demonstrated degeneration features such as lost of reversibility in optical switching [13]. The analysis of the evolution of lateral hydrogen diffusion in rare earth metals was possible by monitoring the optical changes in thin films under gas loading [14]. Direct monitoring of the amount of hydrogen absorbed in Y–Pd thin film under gas loading has been obtained by quartz crystal microbalance techniques and by electrochemical experiments, but in-depth analysis has never been performed in these experiments. A method for the measurement of the concentration of hydrogen versus depth in solids was reported by Hjövarsson et al. using the 1 H( 15 N,αβ) 12 C resonant nuclear reaction in pure Y [15] and Mg–Y alloys [16], but the samples were loaded under gas pressure and this technique loses resolution with increasing depth due to energy straggling [17]. The distribution of hydrogen and possible oxygen uptake under different modes of electrochemical hydrogenation/dehydrogenation steps (most real mode in devices application) still remains not clear. The purpose of the paper was to in situ study the hydrogen and oxygen evolution in depth under hydrogenation/dehydrogenation processes in order to get a better understanding of the internal composition and behaviour of the electrodes. We have applied a chemical analysis method based on Glow Discharge Optical Emission Spectroscopy (GDOES). This technique permits studying, in a very fast and accurate manner, the indepth distribution and concentration of the elements in the electrode. In this work, a thin film electrode of palladium covered by yttrium (optical switching material) was studied and results discussed. The hydrogen and oxygen distribution profiles have been analyzed by means of GDOES in the as-deposited thin films and throughout their optical transitions induced by hydrogen uptake/release. The effect of oxygen is of particular interest and importance due to its influence on hydrogen diffusion and reversibility of optical transition. Additionally, other changes in physical properties, i.e. reversible swelling of the deposited structures, were observed and monitored during this transition. Similar studies performed on Mg alloyed with Ni (third generation material) will be presented in an incoming paper. 2. Experimental details 2.1. Electrode preparation The samples used in this work were prepared in a Pfeiffer Vacuum Classic 500 deposition system equipped with electron gun evaporator. The substrates were previously cleaned in ultrasonic bath and lately treated by ion bombardment in the deposition chamber during 30 min in a 2.2 Pa vacuum level. The pressure in the chamber during evaporation was of 2.0×10 −4 Pa. Y–Pd electrodes (consisting on 200 nm of Y+ 5 nm of Pd) were deposited by electron beam evaporation onto 100×100 mm borosilicate glass substrates. The evaporation rate of yttrium and palladium was 8 Å/s and 7 Å/s rate, respectively, monitored by a quartz microbalance. To improve the homogeneity of the deposition, a continuous rotation of the substrate was performed. After deposition, the glasses with the deposited metal were cut in 3×4 cm 2 pieces using a ruby glass cutter and kept in a dry vessel during several days before treatment and analysis. The Y–Pd deposited films show a brown metallic colour that should be an indication of its partial oxidation during deposition procedure, further confirmed by GDOES analysis. Improving (lowering) the pressure level during deposition allows fabrication of less oxidized thin films. Nevertheless, we studied the as prepared Y/Pd films with some presumed initial oxidation level, mainly to verify the viability of GDOES analysis in resolving the oxygen in-depth distribution. Moreover, the electrode does demonstrate optical switching, getting very transparent for a maximal level of hydrogenation. 2.2. Electrode hydrogenation Electrochemical hydrogenation was performed in a two electrode cell. The 3×4 cm 2 Y–Pd sample was partially immersed in a 1 M KOH alkali solution resulting in a working electrode area of 3×3 cm 2 . The non-submerged area was used to make the electrical contact to a copper wire, covered with epoxy resin. A platinum bar of 20 cm 2 was used as counter electrode. Under galvanostatic conditions, the potential of a counter electrode during experiment (oxygen evolution on a Pt electrode) is constant. Thus, the charging curve of the electrochemical kinetics reflects only the different processes developed on the working electrode. For this reason, we decided not to use a potentio-galvanostat but a power supply, IPS-603 Isotech, used in constant current mode. The voltage kinetics was monitored by a PC controller. Current densities of −0.3 mA/ cm 2 and +0.1 mA/cm 2 were applied for cathodic and anodic modes, respectively. The loss of electrode area after each GDOES analysis (4 mm diameter) was taken into account in order to maintain the same current density in every electrochemical experiment. 2.3. Electrode analysis In-depth profile composition analyses were obtained by GDOES using a Jobin Yvon RF GD Profiler and operating at a typical radio-frequency discharge pressure of 650 Pa and power of 40 W. This method allows analyzing the chemical composition of dielectric, semiconductor and metallic materials in very short time (erosion rateN1 µm min −1 ) with good sensitivity (1 ppm) and in-depth resolution (few nanometres) [18,19]. During GDOES experiments, the samples are sputtered by Ar + ions and accelerated neutral species of very low energies (b50 eV). The sputtered atoms are then excited by the plasma and de-excited by photon emission with a characteristic wavelength, enabling element distinction. In all the cases, the intensity of the signal corresponding to each element was normalized to variations of the luminous source signal recorded throughout the experiment. The diameter of the analyzed area was 4 mm, so the 9 cm 2 Y–Pd electrodes permitted many experiments on the same sample. The GDOES analyses were performed after each electrochemical treatment once the samples were rinsed with ethanol. The time elapsed between the sample removal from the electrolyte and the GDOES experiment was less than 10 s. This time is short enough to assume that there is not a substantial spontaneous desorption of hydrogen from the film in air (see Section 3.3 below). The experimental GDOES time was of 2–3 s and the sample was re-immersed in the solution to perform the next electrochemical step. 3. Experimental results and discussion 3.1. Composition of as-deposited films Fig. 1 shows the in-depth distribution of elements in the asdeposited Y–Pd films. The signals for each element have been scaled to be adapted to the same graph. However, the yttrium signal could not be measured because no detector was available for the corresponding wavelength emission in the polychromator analyzer. For this reason, the discussion will be focused on the qualitative hydrogen/oxygen evolution (both in intensity and in depth) at different stages of the hydrogenation/ dehydrogenation processes. The thin palladium layer (5 nm approximately) is easily noticeable at the surface. Signals from components of the borosilicate glass (Si, B, etc.) were definitely recognized when the etching process reached the substrate. Within the yttrium film, the profile reveals a noticeable quantity of oxygen and hydrogen just below the Pd top layer in the as-deposited sample. It can be due to a spontaneous oxidation of the film either during deposition or after its preparation (lack of Pd protection at the surface). It is known that a 5 nm layer of Pd grows more as islands than as a homogeneous film. Therefore, the Pd layer do not entirely covers the yttrium film, providing a path for oxidation. It is worth mentioning that the signals of both H and O show a very similar in-depth profiles, which might be related to the formation of a hydroxide, rather than oxide, layer in the sub-surface portion of the film (approximately half of its entire thickness). 3.2. Composition during hydrogenation Electrochemical hydrogenation of the Y–Pd electrode was performed in 1 M KOH under constant current density of −0.3 mA/cm 2 . During hydrogenation the polarization process was interrupted, the electrode removed, washed, dried and subsequently measured by GDOES. The handling of the sample before analysis did not last longer than 10 s. Nevertheless, during this short period of time some discharge could occur. In Fig. 2A the total curve is reconstructed from those partial processes. Arrows indicate the time when analyses were performed. Three hydride phases are formed during yttrium hydrogenation [20] and can be observed in the electrochemical kinetics of Fig. 2A: i) a solid state hydrogen solution giving rise to an αhydride composition, produced during the first fast voltage drop, from −0.7 to −1.6 V; ii) the di-hydride or β-phase corresponding to the voltage plateau at −1.72 V; and iii) the trihydride or γ-phase, formed as the cell voltage decreases to −1.9 V. Below that voltage, the electrode becomes completely transparent and H 2 gas evolution is visible. Namely, this optical transition is of interest for developing of switchable devices with variable transparency. We have observed that the changes of composition during hydrogen uptake lead to an increase of time required to reach the glass interface during the analysis, as monitored by the appearance of the B or Si signal (Fig. 2B). The time needed for erosion of the yttrium layer in the as-deposited material (curve 1 in Fig. 2B) is 3.2 s versus 5.6 s for the fully hydrogenated Fig. 1. GDOES depth profile in the as-deposited 200 nm yttrium film covered with 5 nm palladium deposited on borosilicate glass. Fig. 2. (A) Voltage on the cell with a Y–Pd electrode immersed in 1 M KOH charged with current density of 0.3 mA cm −2 . The numbers indicate the instance when the GDOES analysis was taken. (B) In-depth boron signal during different hydrogenation states. electrode (curve 5). This 75% increase of the time can be attributed to the three phenomena: 1) Non-homogeneity of the thickness of the film along with a surface. As we have noted in the Experimental section, we checked the thickness distribution at different locations by profilometry and found it fairly constant. Only 1% of deviation was observed. In order to definitively withdraw this option, we have also taken into consideration the reversibility of this phenomenon (see below). Under dehydrogenation the observed time variation was developed in the opposite direction (less time for more discharged electrodes). 2) A slower erosion rate of the hydrogenated material under the glow discharge. This possibility could be correct if the hardness of the material increases, after hydrogenation. Therefore, this second alternative could explain the observed dependences of the full erosion time on the levels of hydrogenation. 3) An increase of the film thickness due to hydrogenation. Expansion of metals and alloys under hydrogenation and even phase transition is a well-documented phenomenon for all the three generations of the optical switchable materials. As seen in Fig. 2B, the 50% of the total time increase occurs between stages 4 and 5 of the hydrogenation process. This is just the instance when the yttrium hydride undergoes a phase (and optical, and metal/insulator) transformation from β- YH 2+ε to γ-YH 2.8 , and the crystalline structure of material changes from cubic to hexagonal [21]. However, only 15% expansion of material volume due to the yttrium hydrogenation was reported [22] and not 75% as appreciated in the figure. In conclusion, the time needed to reach the glass substrate interface acts as an indicator of changes in the physical properties of the film. These changes are reversible, as was mentioned above, so that during hydrogen release, the curves are shifted in the opposite direction and the analysis time is gradually reduced. We assume that two out of these three alternatives (hardness and volume expansion) are contributing to the overall phenomenon. In order to avoid these hardness/expansion effects associated with the H uptake in the films, the depth profiles of the samples have been normalized to the time where the substrate signals start to rise. Therefore, we have taken the same level of GDOES signal for the fixed element (B, in all the experiments) as the true position of the interface between the deposited film and a glass. Hence, we have normalized the thickness of the entire thin film whatever the hydrogenation conditions and compared the in-depth distribution of different elements throughout the thickness. Thus, in the graphs the depth 0 corresponds to the sample surface while the depth 1 corresponds to the interface between the film and the glass substrate. In Fig. 3 the H-profiles are presented. The results indicate that the hydrogen concentration increases gradually with time of cathodic polarization. In the first stages of H-loading, including state 1 before treatment, a gradient in H-distribution exists: the H signal is stronger at the outer side of the film. The profile becomes flatter when hydrogenation advances, in particular for state 4, corresponding to the end of di-hydride formation. Curve 5 was obtained at the final hydrogenation transition, where the film is homogeneously hydrogenated and thus, di-to-tri-hydride phase transition occurs simultaneously over the entire thickness, converting the metal into a transparent dielectric. It is worth to note, that we have not been able to observe an accumulation of hydrogen inside the Pd protective layer. This film is a catalyst permitting hydrogen trespassing in/out. From Fig. 3 it is clear that hydrogenation of the thin film does not start just at the surface but 20–30% deeper into the sample. Oxides/ hydroxides occupied approximately the same position (Fig. 1) clearly influence the H-uptake and explain our findings. Nevertheless, more detailed experiments with full quantitative standardization of atomic content of the studied element are needed. Large fluctuations of the signal over the depth in the different experiments were observed. The sampling of the GDOES experiment was of 1000 points per second. Therefore, one could expect that the resolution of the experiment would be high enough to minimize the statistical noise. However, the very short experimental time (approximately 2 s) does not allow to completely stabilize the discharge so we cannot assure that the oscillations observed are real fluctuations of the hydrogen content. 3.3. Electrode composition during spontaneous discharge in the ambient atmosphere As we mentioned above the hydrogenated material is not stable with respect to its hydrogen content. The discharge processes start immediately after decreasing the (effective) hydrogen pressure around the sample until the equilibrium is reached. The composition of ambient atmosphere or the presence of reagents able to capture hydrogen at the sample surface (i.e. molecular oxygen together with water) can accelerate the discharge/release process. Fig. 3. GDOES in-depth distribution of hydrogen in the hydrogenated Y–Pd electrode immersed in 1 M KOH, at the instance indicated on the kinetics curve of Fig. 2A. In a first attempt, we analyzed spontaneous discharge of the Y–Pd layer in air atmosphere. GDOES analyses started with the fully hydrogenated electrode (state 5 above), and were then consecutively made after different exposure times (minutes). In this set of experiments, the same sample was shifted on the 4 mm diameter window of the GDOES detector, because this area is eroded after each analysis. The discharge time was accounted beginning from the first measurement in a series. Fig. 4 shows the evolution of the hydrogen profile under dehydrogenation. One can see that a fast spontaneous discharge process in air and then a steady equilibrium is achieved in few minutes. Nevertheless, the “residual”H-content in the film is higher than in the as prepared one. This surely can be attributed to the fact that a reversible transition in Y hydrides is produced between the di- and tri-hydride states and, once loaded with hydrogen, the yttrium film may hardly return to be purely metal. The depth scale in Fig. 4 was normalized as described above due to a clearly observed contraction phenomenon, opposite to the expansion previously described. Real erosion time for the electrode discharged in air to the steady equilibrium (t 3 on Fig. 4) was equal to the time obtained for the film with YH 2 composition (state 4 in Fig. 2A, B). This finding demonstrates that during reversible transition between the two hydrogenated states, other physical properties (i.e. resistance to erosion due to changes in the crystallography structure, density, etc.) are also reversibly affected. It is worth mentioning that only small changes in oxygen content and its distribution in the sample was appreciated during air discharge, see also below. 3.4. Electrode composition during spontaneous and anodic discharge in electrolyte In a throughout electrical controlled switchable device, an anodic discharge could be a way to adjust the transparency. For this purpose, we have studied the hydrogen profiles in the Y–Pd electrode discharged in 1 M NaOH electrolyte both under an open circuit potential (OCP) mode and while the discharge was anodically performed. Fig. 5 depicts the entire electrochemical scheme used in this case. Starting with the as-deposited film (point 1 in Fig. 5), hydrogen was subsequently loaded into the Y–Pd thin film entirely in a single step (point 2). Note that, in this case, the total charge time needed to convert a film onto transparent trihydride was shorter (100 s vs. 150 s under same cathodic current) than for the step-loading showed in Fig. 2A. This evidently was a result of the partial discharge of the sample during preparation for GDOES analysis and the subsequent measurement. After step 2 the electrode was removed from the electrolyte, washed, dried and measured. Following GDOES analysis and partial discharge, the sample was subjected to a second loading (point 2′) and, subsequently, to an OCP discharge. Both treatments were done consecutively until the equilibrium was established in the electrolyte solution (point 3). After the GDOES analysis was done, a third loading 3′was completed. In this case the charging was followed by an anodic forced discharge (point 4). The electrode was analyzed by GDOES after anodizing. Thus, using the three GDOES results we were able to compare the composition of the electrode after different types of discharge processes and assess the feasibility of a full electrical control in the switchable devices based on hydrogenation in alkali electrolyte. The voltage on the electrochemical cell during successive experimental stages is also displayed in Fig. 5. The characteristic voltage plateaux were practically the same as shown in Fig. 2A. In a first stage, with the cell filled with 1 M KOH, the asdeposited electrode yields a voltage of −0.75 V (point 1); subsequently, under cathodic charging, the voltage drops to −2.25 V (point 2). In a second stage, a short charging reveals a remnant capacity of the electrode to uptake hydrogen (2′); subsequent spontaneous OCP discharge shows a short plateau at −0.75 V and then stabilizes at 0.0 V (point 3). The total durability of the −0.75 V plateau is dependent on purging of Fig. 4. GDOES hydrogen profiles during spontaneous discharge in air for: after full cathodic charging (t 0 ), after 1 min (t 1 ), after 2 min (t 2 ), and 5 min (t 3 ) in air. Fig. 5. (A) Galvanic program applied to 200 nm Y/5 nm Pd electrode immersed in 1 M KOH. (B) Voltage kinetics recorded for each step. The numbers indicate the instance when profile analysis was done. the system with nitrogen. Under careful protection of the cell from oxygen, this plateau could last for many hours. Thus, chemical spontaneous hydrogen discharge under OCP conditions is due to interaction with dissolved oxygen [23]. Finally, in the third stage: previous charging shows the same behaviour as in stage 2 (point 3′) and under anodic discharging the voltage reaches +1.7 V (point 4) immediately. The in-depth H- and O-profiles are shown in Figs. 6 and 7 as measured after these experiments. As seen in Fig. 6, curves 3 and 4, the high hydrogen content is preserved in the Y–Pd electrode in two conditions: relaxation in electrolyte under OCP and under anodic polarization. The erosion time indicates that YH 2 composition has not been reached in the OCP experiment. Therefore, and despite the well stabilized voltage (point 3 in Fig. 5), the equilibrium discharge is still not fully complete. Under relaxation in electrolyte, the discharge process is spontaneous and mainly due to the presence of reducible components [24] (dissolved oxygen); it supposed to occur through uncontrollable and slow chemical reaction: YHxþO2→YHx−yþH2OðþH2O2Þð1Þ with xthe hydrogen molar fraction in the compound and ythe electrons consumed by the oxygen reduction reaction delivered by the YH x hydride, which itself is oxidized. We have included in Eq. (1) hydrogen peroxide (besides water) as a possible final product to generalize an interaction and reflect possible changes of the electrolyte composition during repetitive cyclic switching of the device. Curve 4 in Fig. 6 corresponds to an anodic (electrically activated) discharge: the hydrogen concentration in the metal was surprisingly higher than after OCP experiment. The H- profile is flat along with the entire thickness of the film. Thin film electrode has acquired a dark brown colour after this experiment, evidently due to the concurrent oxidation of the Y metal. This was confirmed by the oxygen profiles obtained from the GDOES analysis. The in-depth O-profiles, Fig. 7, demonstrate that under applied anodic discharge (curve 4) the entire film contains a high concentration of oxygen. O-content is of similar level (and even higher) that was found for the sub-surface layer in the asdeposited film (compare with curve 1). Moreover, the erosion time was extended to more than 5 s for this sample that is definitely out of range for the metal-hydride films used in this work. These are indications that an unexpected (oxidized or hydroxide) material was formed. Comparing together the in-depth H- and O-profiles, we can conclude that these two elements became equally distributed within the film after anodic discharge. Moreover, their contents are alike to the ones observed in the sub-surface area of the asdeposited film. This means that during anodic discharge yttrium oxide or hydroxide is produced. As can be seen in Fig. 7 there is small uptake of oxygen even in the conditions when the electrode was not polarized anodically: after cathodic polarization, curve 2, or under OCP discharges, curve 3. In these two cases, the O-profiles present only a small growth but a clear in-depth expansion. This spontaneous chemical oxidation of Y (and, perhaps, other metals/alloys) could be a reason for many degeneration features developed with time and/or under cycling of the optical devices basedonreversiblehydrogenationinwaterelectrolytes. Consequently, a full electrical control of hydrogenation and even possibility for stable function of a device in alkali-based electrolytes is problematic. From the very beginning, the protection of the external surface was considered of primary importance in such devices. For this purpose, a thin Pd film was first chosen that permitted development of the three generations of switchable mirrors. On the other hand, the high optical opacity of Pd makes thicknesses of 10 nm its natural limit. At the same time, the island-like morphology of the very thin Pd layer (5–10 nm) cannot be avoided, resulting in penetration of oxygen through the uncovered area. Recent materials developed for switchable devices, such as Mg–Ni alloys, have shown higher stability versus oxidation due Fig. 6. GDOES hydrogen profiles after the treatments presented in Fig. 5 for: (1) as-deposited electrode, and after (2) full cathodic charging, (3) OCP discharge, (4) anodic forced discharge. Fig. 7. GDOES oxygen profiles after the treatment presented in Fig. 5 for: (1) asdeposited electrode, and after (2) full cathodic charging, (3) OCP discharge, (4) anodic forced discharge. to their more noble nature in comparison with Y. These have demonstrated better reversibility and higher contrast in optical transition. We expect that an electrically controlled discharge (under anodic polarization), would be achievable in those systems due to reduced spontaneous surface oxidation. Other materials, e.g. Ni, might be used as cap layers with good optical and good hydrogen permeability. In a subsequent paper dedicated to the GDOES analysis of Mg–Ni thin film alloys we will present and discuss both issues. 4. Conclusions GDOES analysis was employed for the first time for studying the in-depth distribution of chemical elements in Y– Pd thin films deposited on glass substrates. The technique is sensitive enough to routinely resolve the 5 nm palladium protective film from the underlying yttrium layer and variations in the chemical (H and O) composition during hydrogenation/ dehydrogenation. The erosion time of the metal (hydride) film during GDOES analysis is sensitive to properties and composition of the material. The time increased under hydrogenation revealing not only expansion of the studied Y–Pd thin film volume/thickness but also changes of other properties such as hardness and/or crystal (phase) structure of material. The erosion time was correspondingly reduced when hydrogen was released from the film. Normalization of thickness of the studied film was done based on the (time) position of the glass substrate signal (boron). The in-depth H- and O-profiles were then compared for different stages of the Y–Pd film hydrogenation/dehydrogenation. A reversible development of the H-profile under hydrogenation/dehydrogenation was observed when composition of the Y film was alternated between the di- and tri-hydride states. Spontaneous H-release in air was very fast and the di-hydride composition was reached within few minutes after removing the material from the electrolyte. Under OCP conditions (in the electrolyte), the stability of the fully hydrided state depends on the presence of dissolved oxygen. Analysis of the O-profile revealed spontaneous oxidation (or formation of hydroxides) of the films deposited in the relatively low pressure conditions employed in this work. Oxides/ hydroxides are located at the yttrium/palladium sub-surface. The same process of yttrium oxidation is enhanced in the electrolyte due to its reaction with dissolved oxygen and water. In anodic discharge process, the entire yttrium film could be converted into an oxide/hydroxide. Discussion of perspectives for a full electrically controlled switchable device in alkali electrolytes allows concluding that for the Y–Pd system this is a very unfavourable situation due to high chemical activity of Y. For third generation materials, Mg– Ni alloys, an electrically controlled device is more feasible yet the type of surface protection is of the highest importance. Acknowledgment The help of Mr. Yuri Makhushok in the development of a portable power current supply controlled by computer is fully acknowledged. References [1] J.N. Huiberts, R. Griessen, J.H. Rector, R.J. Wijngaarden, J.P. Dekker, D.G. de Groot, N.J. Koeman, Nature 380 (1996) 231. [2] R. Griessen, J.N. Huiberts, M. Kremers, A.T.M. van Gogh, N.J. Koeman, J.P. Dekker, P.H.L. Notten, J. Alloys Compd. 253–254 (1997) 44. [3] D.G. Nagengast, A.T.M. van Gogh, E.S. Kooij, B. Dam, R. Griessen, Appl. Phys. Lett. 75 (1999) 2050. [4] T.J. Richardson, J.L. Slack, R.D. Kostecki, B. Farangis, M.D. Rubin, Appl. Phys.Lett. 78 (2001) 3047. [5] T.J. Richardson, J.L. Slack, B. Farangis, M.D. Rubin, Appl. Phys Lett. 80 (2002) 1349. [6] W. Lohstroh, R.J. Westerwaal, A.C. Lokhorst, J.L.M. van Mechelen, B. Dam, R. Griessen, J. Alloys Compd. 404–406 (2005) 490. [7] R.J. Wijngaarden, J.N. Huiberts, D. Nagengast, J.H. Rector, R. Griessen, M. Hanfland, F. Zontone, J. Alloys Comp. 308 (2000) 44. [8] S.J. Van der Molen, J.W.J. Kerssemaker, J.H. Rector, N.J. Koeman, B. Dam, R. Griessen, J. Appl. Phys. 86 (1999) 6107. [9] E.S. Kooij, A.T.M. van Gogh, R. Griessen, J. Electrochem. Soc. 146 (1999) 2990. [10] P.H.L. Notten, M. Kremers, R. Griessen, J. Electrochem. Soc. 143 (1996) 3348. [11] M. Pourbaix, in: Pergamon (Ed.), Atlas of Electrochemical Equilibria in Aqueous Solutions, New York, NY, 1966, p. 181. [12] V. Parkhutik, E. Matveeva, Y. Makushok, E. Rayon, T.J. Richardson, J. Electrochem. Soc. 152/12 (2005) 209. [13] A.M. Janner, P. van der Sluis, V. Mercier, Electrochim. Acta 46 (2001) 2173. [14] F.J.A. den Broeder, S.J. van der Molen, M. Kremers, J.N. Huiberts, D.G. Nagengast, A.T.M. van Gogh, W.H. Huisman, N.J. Koeman, B. Dam, J.H. Rector, S. Plota, M. Haaksma, R.M.N. Hanzen, R.M. Jungblut, P.A. Duine, R. Griessen, Nature 394 (1998) 656. [15] C. Chacon, E. Johansson, B. Hjörvarsson, C. Zlotea and Y. Andersson. J. Appl. Phys. 97 (2005) 104903. [16] J.N. Huiberts, J.H. Rector, R.J. Wijngaarden, S. Jetten, D. de Groot, B. Dam, N.J. Koeman, R. Griessen, B. Hjörvarsson, S. Olafsson, Y.S. Cho, J. Alloys Compd. 239 (1996) 158. [17] W.A. Landford, H.P. Trautvetter, J.F. Ziegler, J. Keller. Appl. Phys. Lett. 28 (1976) 566. [18] M.R. Winchester, R. Payling, Spectrochim. Acta, Part B 59 (2004) 607. [19] R. Escobar Galindo, E. Forniés, J.M. Albella. J. Anal. At. Spectrom. 20 (2005) 1108. [20] E.S. Matveeva, R.J. Ortega, A. Sanchez, C. Ferrer, V. Parkhutik, Sens. Actuators, B 84 (2002) 83. [21] A. Remhof, G. Song, C. Sutter, D. Labergerie, M. Hübener, H. Zabel, J. Härtwig, Phys. Rev. B. 62 (2000) 2164. [22] D.G. Nagengast, J.W.J. Kerssemakers, A.T.M. van Gogh, B. Dam, R. Griessen, Appl. Phys. Lett. 75/12 (1999) 1724. [23] R.A.H. Niessen, P.H.L. Notten. Electrochim. Acta. 50 (2005) 2959. [24] E. Matveeva, V. Parkhutik, J. Electrochem. Soc. 149/10 (2002) 148.