Segregation and Preferential Sputtering of Cr in WCrY Smart Alloy

The temperature driven segregation of Cr to the surface of the tungsten-based WCrY alloy is analysed with low energy ion scattering of He+ ions with an energy of 1 keV in the temperature range from room temperature to 1000 K. Due to the high surface sensitivity, these measurements probe only the composition of the outermost monolayer. The surface concentration of Cr increases slightly when the temperature of the sample is increased up to 700 K and exhibits a much stronger increase when the sample temperature is further raised. The segregation enthalpy for Cr is obtained from the Langmuir-McLean relation and amounts to 0.7 eV. The surface concentration of Y shows a similar behaviour to the Cr concentration. The temperature thresholds between slow and accelerated surface density increases for Cr and Y are nearly the same. At a temperature of 1000 K the low energy ion scattering detects almost no W on the surface. The modified surface composition due to the segregated species, i.e. the mixed Cr/Y layer, stays stable during cool-down of the sample. Preferential sputtering is investigated using ion bombardment of 250 eV D atoms, resulting in an increase of the W surface density at room temperature. This effect is counteracted at elevated temperatures where segregation replenishes the lighter elements on the surface and prevents the formation of an all-W surface layer. The flux of segregating Cr atoms towards the surface is evaluated from the equilibrium between sputter erosion and segregation.


Introduction
The first wall in a future fusion power plant needs to ensure long reliable operational periods in the presence of high particle and heat loads [1]. Tungsten is the candidate material for the first wall of a fusion reactor [2]. During long-term operation of the power plant the first wall material will become activated due to neutron irradiation. This poses a potential danger in the unlikely case of a loss-of-coolant event. The nuclear decay heat could raise the temperature of the first wall up to 1200 • C for several weeks or months [3]. A simultaneous air ingress would allow tungsten to form volatile WO 3 which is released into the reactor vessel. This problem can be alleviated by the utilisation of self-passivating ternary tungsten alloys which were developed and investigated during recent years [4]. These alloys show under plasma exposure preferential sputtering of the lighter constituents, leaving behind an enriched tungsten layer which has superior sputter resistance on the surface. In case of water or air ingress a protective layer containing stable oxides of the lighter alloying elements develops on the surface which suppresses further oxidation. Because of this situation dependent behaviour of the alloy, these materials are commonly labelled as smart alloys [5]. Such alloys with various compositions are presently under development.
WCrY alloys were recently developed and tested in laboratory experiments (oxidation resistance) and in a plasma environment [5]. Plasma exposure of WCrY samples in the linear plasma device PSI-2 and subsequent analysis by secondary ion mass spectrometry confirm surface enrichment of W caused by preferential sputtering of Cr [6]. The plasma exposure in these experiments uses a pure D plasma with sample biasing between 120 V to 220 V and is performed at temperatures between 900 K and 1000 K, which is a typical temperature range estimated for the first wall during operation of a fusion power plant and temperatures of the divertor tiles are probably even higher. However, a reverse mechanism to the Cr depletion by preferential sputtering could be the thermal segregation of Cr to the surface of the sample at elevated temperatures. This process cannot be assessed easily in the plasma exposure experiments.
In this work, both processes, the thermally activated segregation as well as the preferential sputtering by deuterium of the lighter alloy components, are investigated in an ion beam setup. Low energy ion scattering (LEIS) is utilised to analyse the surface composition of a WCrY sample at various temperatures. Surface composition changes due to preferential sputtering by D ions are investigated at different temperatures in order to resolve the influence of segregation at higher temperatures.

Experimental
LEIS is an extremely surface sensitive method to analyse the elemental composition of the first monolayers of a sample [7]. In LEIS a beam of low energy (few keV) singly charged ions (usually noble gas ions) is directed toward the surface and ions scattered under a certain angle pass an energy analyser and are detected and counted [8]. The collision process can be described in binary collision approximation which allows straightforward to infer the mass of the atom from which the detected projectile atom has been scattered [9].
The used LEIS apparatus is described in detail in [10]. In brief, ions are produced in a Bayard-Alpert type ion source, extracted from the source volume with a voltage of about 100 V, and accelerated by a second electrode to their final energy of q × 1 kV, where q denotes the charge of the ion. A 90 • magnetic sector field is used to select the mass and charge state of the probing ion beam. Two sets of ion optical elements, each one consisting of an einzel lens and steering plates for lateral and vertical beam deflection, are mounted before and after the magnetic sector field and guide and focus the ion beam onto the sample. The sample is mounted on a Prevac PTS 1000 RES/C-K sample holder which can be resistively heated up to 1200 K and allows to measure the temperature of the front plate with a K-type thermocouple. The scattered ions pass a 90 • spherical energy analyser and are counted with a channeltron electron multiplier. The sample holder and the analyser can be independently turned around the scattering centre in order to adjust incident angle and scattering angle of the ions. The setting and scanning of the voltages at the analyser electrodes as well as the ion counting is performed with a National Instruments PXI-8106 data acquisition board which is controlled by a personal computer running LabVIEW software. Sample cleaning and sputter erosion is done with a Perkin-Elmer 04-161 sputter ion gun.
The WCrY alloy is produced by a field-assisted sintering technology [5]. The composition of the produced WCrY alloy is given in table 1. The WCrY ingot obtained after sintering is first cut by wire erosion to get a sample which measures 10 mm x 5 mm x 1 mm. The sample is manually ground to remove residual impurities from cutting. For grinding silicon-carbide (SiC) papers with different particle sizes are used. Afterwards a cloth with a diamond paste of particle size 1 µm is applied. Finally, the sample is polished to a mirror finish using an Active Oxide Polishing Suspension (OPS) with 0.04 µm SiO 2 particles.
Prior to the LEIS measurements the sample is in-situ cleaned by sputtering with 500 eV Kr + ions under normal incidence with a total fluence in the order of 1 × 10 21 m −2 . It is almost unavoidable that the cleaning procedure by sputtering will already modify the surface because the sputtering yields for the various constituents of the WCrY sample are different. As a result the surface concentrations of the cleaned sample deviate from the bulk concentrations. The deviation of the surface concentrations from the respective bulk concentrations can be derived from the equilibrium state which depends the sputter yields and is for a two component alloy given by [11] c s,1 where c denotes the surface/bulk concentration and Y is the sputter yield. Indices s and b refer to surface and bulk, respectively, and numbers label the alloy constituents. Kr is deliberately chosen as sputter gas because the sputter yields for W and Cr at the low ion energy of 500 eV do not differ very much. The elemental sputtering yields of 500 eV Kr + ions for pure W and Cr are 0.6 and 1.0 [12], respectively. Since the difference in sputtering yields is less than a factor of two, the resultant deviation in surface densities is rather small, too. In any case, it is anticipated that surface changes due to thermal segregation and preferential sputtering by low energy deuterons will be much larger and can easily be distinguished from the artefacts of surface cleaning.
The mass resolution of LEIS depends on the mass ratio between the probing ion and the surface atoms [13]. In that respect it would be beneficial to use incident ions with a higher mass, e.g. Ne + or Ar + . However, due to the larger masses of these ions they would induce stronger sputtering of the surface which might lead to unintended changes of the surface composition. He + is chosen as probing ions in order to keep those effects reasonably small.
Although the mass resolution is not optimum, the amplitude of the various peaks in the scattering spectra can be well determined by the peak fitting procedure outlined below. The magnitude A i of an individual peak originating from scattering at atoms of species i is given as the area under the fitted peak which is proportional to the product of peak amplitude and full width of half maximum, both obtained from the peak fitting procedure.
The uncalibrated surface concentration c i for species i on the surface is defined as where the summation goes over all elements j on the surface. The magnitude of an individual peak depends on the differential scattering cross section multiplied with the probability that the projectile ion is not neutralised during the collision. This quantity might be obtained from calibration measurements using pure elemental samples of the various alloy species. Unfortunately, such measurements are not yet available. However, since c i (A 1 , .., A n ) defined in equation 2 is asymptotically correct for zero (c = 0) and full (c = 1) coverage, and strictly monotonous in between, it fulfils the requirements to characterise the surface coverage.

Results and Discussion
Two representative LEIS spectra for the investigation of the thermally activated segre- The segregation enthalpy, ∆H, can be determined from the data shown in figure 3. For the analysis of the Cr segregation we follow the procedure described in [18]. The surface concentration of the segregating species, here Cr, is described by a Langmuir-McLean where c s Cr denotes the fractional surface coverage of Cr, c b Cr the fractional Cr bulk concentration, and ∆G is the Gibbs free energy of segregation. Using ∆G = ∆H − T ∆S, where S is the entropy, and taking the logarithm of equation 3 gives the following equation [20] ln( c s Figure 4 shows the corresponding Arrhenius plot of the Cr surface concentration data.
The data points are clearly separated in two distinct regions. At temperatures below For the low temperature region we refrain from evaluation of a segregation enthalpy because it is not clear that the slight increase of Cr surface concentrations with rising temperature is indeed a thermal segregation as in the high temperature region. As mentioned in the experimental section, the sputter cleaning with Kr + causes some preferential erosion which leads according to equation 1 to a surface depletion of the species with the larger sputter yields, which is Cr. The small increase when raising the sample tempera-ture to 600 K to 700 K, which is less than a factor of 2, is likely due to recovery of the unperturbed surface densities due to Cr diffusion from the bulk.
The same analysis for the Y surface concentration is shown in figure 5. Again, the data is separated in two regions with slow and fast segregation, respectively. The segregation enthalpy for Y is ∆H = 0.3 eV. The critical temperature where both straight lines cross is T c,Y = 696 K. This temperature is not significantly different from the critical temperature obtained for Cr segregation.
It is obvious that for both lighter alloy constituents, Cr and Y, the segregation behaviour looks very similar. In both cases segregation proceeds rather slowly below a critical temperature and accelerates as soon as this temperature is exceeded. However can be investigated with LEIS, too, but some attention has to be paid to the unavoidable sputtering by the He + ions used for the analysis. Sputtering by the probing ion beam is of no concern for the previous analysis of the thermally driven segregation behaviour, but here it needs to be discriminated against a possible de-segregation. In order to keep the This behaviour could result in a chromium oxide layer which might explain the enhanced deuterium retention found in W-10Cr-0.5Y samples which were outgassed at even higher temperatures prior to a study of the hydrogen isotope retention [22].