Long-term evolution of the impurity composition and impurity events with the ITER-like wall at JET

Local impurity line emission of eroded or recycled neutral atoms and subsequently singly charged ions was measured spectroscopically in the visible spectral range. In the case of the monitoring discharges beryllium was detected observing the Be I emission at 457 nm in the divertor while for carbon C II at 515 nm was used. For lines of sight (figure 1(a)) perpendicular to the viewed divertor surface area, the intensity of these lines is approximately proportional to the gross erosion flux of the respective species. For species like C and Be constant deposition and re-erosion are the cause for the visible line emission in the divertor. As the plasma parameters are identical during the monitoring discharges the variation in the photon flux is a direct measure of changes to the impurity (re-)erosion fluxes and yields on the divertor surfaces. Figure 1(a) shows a subset of the available lines of sight of the JET survey high-resolution spectrometers providing integral views of the outer and inner divertor regions, respectively [5].

The VUV spectral range applied for approaches (ii) and (iii) covers spectral lines from a wide range of elements, such as W, Ni, Cr and C, and their various ionization states, making this range particularly valuable from the point of view of core plasma diagnosis. Spectroscopic measurements in the VUV were carried out with the aid of a set of survey SPRED spectrometers by PI [10, 11]. Using routinely either a 450 g mm⁻¹ holographic grating in the 10–110 nm wavelength range employing a horizontal line of sight into the plasma (figure 1(a)) or a spectrometer with a 2105 g mm⁻¹ grating recording spectra in the wavelength range below 40 nm and looking nearly vertically down into the JET divertor.

For both the VIS and VUV spectra usually the integral area defined by a given number of pixels on either side of the line centre is used as the intensity signal, these pixels also defining the background to be subtracted. However, to clearly identify W in the confined plasma a fit of the spectra is performed based on the method described in [12, 13].

2.2.1. Monitoring discharges.

Monitoring discharges were introduced to the routine plasma operation of JET-ILW [6]. The monitoring scenario had to be compatible with operation at the beginning of the JET-ILW experiment. The plasma shape (figure 1(a)) was developed for the JET-ILW operation during the previous JET-C experimental campaign [14]. Figure 2 shows the temporal evolution of plasma parameters for a discharge in the later phase of the campaign. In the initial phase (JPN #80170-#80300) no auxilary power was available and only the ohmic flat-top phase of the plasma was used. With progressive use of auxilary power, the monitoring discharge was modified adding L- and H-mode phases in subsequent plasma operation. This allows the operational evolution to be followed through all stages of exploitation. For the purpose of this paper the ohmic phase has been used as it is available throughout the complete campaign. Further analysis on the L- and H-mode is outside of the scope of this contribution, but envisioned in conjunction with future experimental operation.

Zoom In Zoom Out Reset image size

2.2.2. Overall trends.

2.2.3. Transient impurity events.

Following the impurity evolution in the divertor it is clear that after the installation of the JET-ILW about 300 ohmic discharges (2000 s) of the divertor operation were required to establish a steady-state divertor composition [5].

Figure 3 displays the main low-Z impurities involved in the evolution of the divertor plasmas, Be and C, by means of their photon fluxes normalized to D_alpha, representing the impinging ion flux. The grey areas depict the phase of constant plasmas in the start-up of JET-ILW operation. In the course of the campaign, a further evolution of the Be and C emission is, however, apparent.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Figures 3(c) and (a) display the long-term C and Be evolution measured during the constant ohmic phases of the monitoring discharges. A wide spread with a maximum increase of about 20% in carbon is visible though still more than an order of magnitude below the levels of the JET-C area. In addition, the deuterium emission is depicted (figure 3(b)) as it is used to normalize the C and B emission against subtle plasma variations.

A possible source for additional carbon might be the W coatings utilized at JET as they still contain up to 8% (at.) of carbon [6]. This provides a potential reservoir of C in addition to the mainly present Be and W. Other potential sources of carbon are discussed in [6].

For both the inner and outer divertor an initial level of beryllium is established after 300–500 discharges. While for the outer divertor, the composition between C and Be remains similar, the inner divertor displays a clear drop in beryllium emissions during the initial phase (figure 3, grey area). This observation indicates that the balance between C and Be is changing. Both C and Be are recycling in front of the W, Be mixed surface as W emissions are always present [15].

While the monitoring discharges have constant plasma conditions they do reflect the impact of plasma operation in between. As the experimental conditions evolve during the course of the JET-ILW operation towards higher power and higher performance operation, one can quite clearly observe the daily variations within the monitoring plasmas (for details see [6]). However, repetition of discharges reveals the base level of conditions.

One particular case to highlight this effect is represented by the red line. Here, for scientific purposes a Be evaporation was performed to study its effect on main chamber impurities such as tungsten and nickel as well as the divertor response. The Be evaporation was not required from the conditioning point of view, but was used to locate potential sources of W and Ni from the main chamber and recessed areas by suppressing potential sources with a thin Be layer.

Apart from the main chamber especially the outer divertor is covered by a thin layer of Be [16], hence for a limited set of discharges the Be emission increases as the C emission is reduced until the transient beryllium coating from the outer divertor is eroded by subsequent plasma operation. From migration experiments during JET-C experiments it was established that the outer divertor is typically erosion dominated, hence Be is measured as a recycling impurity, thus as part of the plasma flux impinging on the target [16, 17].

We can observe that after the initial strong evolution (figure 3, grey area), especially in the inner divertor, a base level composition of the divertor is reached. Further variations can mostly be attributed to changes in plasma operation in between monitoring discharges [6] (e.g. Be evaporation). Evolution e.g. towards higher power operation or longer phases without operation, where the O leak causes higher levels of O and C, is still slightly changing the material mix or layer composition especially in the deposition-dominated inner divertor. The main Be source is the main chamber wall whereas new C sources are not apparent. Initial post-mortem results regarding the divertor beryllium deposition were presented in [18].

One indicator of the plasma impurity composition is the effective charge of the confined plasma observed from spectroscopy.

$$\begin{equation} c_{\rm imp}=\frac{Z_{\rm eff}-1}{Z_{\rm imp}\cdot(Z_{\rm imp}-1)}. \end{equation} \tag{ 1 }$$

The evolution of main chamber impurities is deduced from VUV line emission as depicted above. Those line emissions—normalized to the edge density—can be interpreted as effective concentration of the observed element. CXRS analysis is only available for a limited set of discharges with neutral beam heating, thus relative changes are best determined via passive VUV signals. Effects of increased plasma heating can, however, slightly change this relation (figure 5(b)) as discussed below.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The interplay of the different plasma conditions present during the JET-ILW operation influences the impurity composition and production. Higher auxiliary power and operation at low density causes a larger Be influx due to increased main chamber erosion, while lower density also means less plasma flux to the target hence less residual background flux of carbon.

In equation (1) we assume only one impurity is dominating Z_eff and hence we can deduce the change in impurity concentration from JET-C to JET-ILW, assuming that C and Be are the dominant impurities, respectively. Taking into account figure 4, we can also follow up the evolution of the impurity emission, here C, O, in the confined region as an indicator for the relative changes. Emissions from the main plasma are depicted here with O VI originating from deeper in the confined region. C III is chosen here as a signal from the core (C VI) has low emissivity and is close to the detection limit for the JET-ILW. Carbon drops by a factor of 20, while oxygen is, when comparing the restart phase (#71000 versus #80000), at least a factor of 10 lower compared with JET-C. The drop in oxygen is especially impressive as the start-up of the JET-ILW operation was performed with a vacuum leak present. For JET-C an average Z_eff = 1.96 is observed while this drops drastically for the JET-ILW to Z_eff = 1.21. Considering equation (1) we conclude a concentration of carbon of 3.2% for the JET-C and beryllium of 1.75% for the ILW. Residual carbon is still present and originates from the JET-C to JET-ILW changeover and potential contributions from W-coated CFC. This is consistent with previous CXRS observations of the carbon dominated JET [6], the JET-ILW is showing 2–3% Be and 0.1–0.2% carbon. This means that two effects of the change over to the JET-ILW are obvious: drastic reduction of the carbon content and effective oxygen gettering by the Be first wall. The Be first wall is gettering the O much more effectively than in previous campaigns with regular Be evaporations and PFCs made of CFC. This is in line with the fact that no wall conditioning is required during ILW plasma operation [19, 20]. In figure 4(b) a trend towards increased normalized C III emissions in the later phases of the first year of JET-ILW operation can be seen, which can be related to the increase in auxiliary power in the course of the JET-ILW operation. A potential increase of a carbon source needs to be separated from the simple change in core plasma temperature, confinement and hence increased emissions.

Figure 5(a) shows the auxiliary power P_IN as well as the radiated power P_Rad in the course of JET operations since JPN #70000. The wide range of applied power is apparent as is the increase in applied heating since installation of the ILW. The overall input power is steadily increasing from ohmic plasmas to ELMy H-Mode discharges. Dedicated effects relating to RF heating are observed with respect to high-Z core accumulation [9].

Taking the C III emission (figure 4(b)) and comparing this to the trend displayed (figure 5(b)), we can conclude that the main change in the C III emission is caused by the increase in input power, simply due to changed plasma conditions. For both JET-C and JET-ILW one can observe the same slope hence the same dependence on power increase. The drop from JET-C to JET-ILW is still about 15 as seen from the overall magnitude of the two slopes. For comparison the Be evolution is shown as well (figure 5(c)), a similar trend with heating power is visible reinforcing the statement regarding C.

We hence can deduce that despite the small increase of 20% in the divertor C emission (figure 3) the overall change to the core C emission is mostly originating from a change in plasma conditions. In conclusion, no apparent strong carbon sources are evolving so far from potentially damaged W-coated surfaces.

In addition to the general long-term impurity evolution also spurious impurity influxes were detected as irregular occurrences of additional radiation during all available plasma scenarios. The statistics presented here are with respect to events occurring during the flat-top main heating phase of the plasma. Additional events occurring during the current ramp-up are not considered here. These events are probably small particles or dust. The radiation events connected to them are occurring with a typical rise time of ms and sometimes lead to a long lasting increase in radiated power, potentially adding several 100 kW up to a few MW to the steady-state bulk radiated power, typically 10–50% of the applied heating power. Data from the visible divertor spectroscopy show only a few of the events, so far eluding definition of composition and origin of the source. These events can cause a significant impact on plasma operation even though only a few have induced a disruption during the main heating phase. The frequency of these events is quite steady throughout the JET-ILW campaign. Conditioning effects are only appearing during constant plasma operation with repetitive plasmas. Disruptions can, however, increase the number of events occurring in the subsequent 2–3 discharges apparently stirring up additional particles [20]. Performing different plasma scenarios, with strike-point position and input power variations has not led to an overall cleaning. The composition of these particles is determined from main chamber VUV measurements of high-Z elements. Major contributors are W and Ni/Fe/Cr, the latter grouped together since they might come from Inconel or steel originating from the JET in-vessel components. The composition of more than 15% of the detected events could not be determined through VUV measurements and could be caused by low-Z impurities e.g. droplets from beryllium melt damage. The origin of the particles is thought to be either left over cuttings from in-vessel machining, mainly in the first phase where almost all events are identified, and abrasions from W-coatings due to installation. The fact that only a few events are detected in the divertor, tends to the conclusion that main chamber W-coatings are responsible for most W related events. Such coatings are e.g. applied as re-ionization plates or protection tiles. Damage on the re-ionization plate was observed by initial inspection. In addition, melt damage on the beryllium limiters as well as the target Langmuir probe-tips from initial in-vessel inspection was identified.