Influence of AR injection on shielding layer properties and surface protection from transient high heat loads under the QSPA plasma exposures

The paper presents experimental studies of a shielding plasma layer formation in front of a tungsten surface exposed with hydrogen plasma in the QSPA-M test-bed facility under the conditions of additional seeding of argon (Ar) along the target surface into the zone of plasma-surface interaction. A pulsed gas injector on the base of a fast electromagnetic valve has been developed for the local injection of Ar. The injector is capable of generating a homogeneous argon gas flow with a maximum concentration above n Ar = 6 × 1023 m−3 and a pulse duration of 0.5 ms. It is shown that the increase in the argon gas density in front of the surface leads to an essential decrease (in 1.5–2 times) in the energy load delivered to the target surface. In the presence of a strong magnetic field (up to 1 T), both the thickness of the shielding layer and the fraction of energy dissipated by the shield increase further. Even for moderate energy densities of the QSPA plasma streams in the experiments with Ar gas injection, less than 40% of the impacting plasma load is absorbed by the tungsten surface. The results demonstrate that this additional shielding attributed to the formation of a dense Ar plasma layer in front of the exposed W surface would be favourable for the divertor armour performance, causing the decreasing erosion of plasma-facing components in the course of transient events in a fusion reactor.

Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

Introduction
Tokamak divertor and its targets surfaces are expected to be the most loaded areas in future fusion reactors, such as ITER and DEMO, that should be capable of taking on the highest heat and particle fluxes to the surface during the disruptions and other transient events [1][2][3].
The vapour shield attributed to the dense plasma layer formation in front of the divertor surface is an intrinsic effect of extreme plasma heat loads, e.g.unmitigated major disruptions and giant ELMs.A shielding layer consisting of evaporated material and plasma species prevents the direct contact of powerful plasma with a material surface and absorbs most of the impacting plasma energy [4][5][6][7].Thus, it favourably influences the overall erosion of tungsten (as designated divertor material), including not only the sputtering yield, melt, and evaporation losses, but also suppressing W surface cracking and dust production rate [8][9][10].Investigations of plasma energy transfer through the vapour shield significantly contribute to our understanding of physical processes occurring during the plasma-material interaction under extreme conditions and could be used for benchmarking predictive numerical codes [6,7,11,12].
Injection of heavy noble gas in front of exposed surfaces might be one of the additional options for the mitigation of material damage in the course of severe transients and for efficient protection of divertor components in fusion devices.The injection aims to dissipate the plasma energy through increasing radiation from a shielding layer and, as a result, decreasing the heat loads to the plasma-facing components.It is foreseen that there will be a need to seed noble gases or nitrogen into the divertor plasma of ITER to substantially reduce power loads on the tungsten divertor targets [13].
In existing tokamaks, the seeding of nitrogen and argon has already been successfully demonstrated in a number of dedicated experiments.In particular, the impact of gas seeding on the confinement and power load control has been studied in radiative divertor experiments in JET, ASDEX-UG, DIII-D, JT-60U, and other tokamaks [14][15][16][17][18]. Numerical simulation of shielding layer properties is also performed to improve prediction capabilities for ITER and beyond.
However, in presently operating tokamaks and stellarators, the target surface damage is typically rather moderate because of the relatively low energy content in the core plasma as compared to future fusion reactors.Thus, a direct extrapolation of the transient heat flux parameters to ITER and DEMO experimental conditions is not always possible.Therefore, other powerful test-bed facilities (QSPAs, linear devices, electron beams, etc) are involved in plasma-surface interaction studies and analysis of the energy shielding mechanisms under the high heat transient loads typical for a fusion reactor.The efficiency of energy shielding in ITER-and DEMO-relevant conditions is among critical issues and needs to be comprehensively analysed in simulation experiments with plasma test-bed facilities and with numerical modelling [6,11,19,20].This paper presents the experimental studies of the plasma energy transfer to the tungsten surfaces under the transient plasma loads generated by the quasi-stationary plasma accelerator QSPA-M and the influence of the additional injection of Ar cloud in front of the target on the shielding layer parameters and its efficiency.

Experimental setup
The experiments have been performed using the QSPA-M testbed facility described elsewhere [19,21,22].The experimental setup is schematically presented in figure 1. Hydrogen was used as the working gas of the QSPA's plasma stream.The formation of the transient plasma layer in front of the surface was studied for the normal impact of the plasma stream on the tungsten target surface.The size of the target was 5 × 5 cm 2 .The target was introduced along the axis of the system and was able to move from 1.62 m to 1.85 m (diagnostic chamber) from the accelerator output.
A distinctive feature of the quasi-stationary plasma accelerator QSPA-M is the presence of an external magnetic field of up to 1 T, which makes it possible to study the influence of a strong B-field on the plasma shield dynamics and shielding efficiency.Such plasma transportation in a strong external magnetic field in the QSPA-M facility mimics the fusion reactor conditions in the vicinity of the divertor, where plasma propagates along the separatrix.The magnetic system of the QSPA-M device consists of 21 base coils of longitudinal magnetic field and 4 Helmholtz coils in the regions of the plasma accelerator and target chambers [21].The external longitudinal magnetic field increases linearly along the plasma propagation axis from 0.06 T at the discharge area up to 1 T at the distance of 1.85 m from the accelerator.The plasma stream energy density was varied (up to 1 MJ m −2 ) to study the evolution of the shielding layer properties.
A special gas injector was designed for an additional pulsed injection of noble gas cloud (argon) along the exposed target surface into the zone of plasma-surface interaction.The gas injector was installed in a vertical pipe at a distance of 1.62 m from the accelerator.The general view and scheme of the pulsed gas injector on the base of the fast electromagnetic gas valve is presented in figure 2. The power supply of the electromagnetic coil of the pulsed gas valve is provided by a capacitor battery with C = 300 µF.The injector provided a local seeding of Ar gas along the irradiated W surface, as shown in figure 1.The injector is capable of generating a homogeneous gas flow with a maximum concentration over n Ar = 6 × 10 23 m −3 within a 0.5 ms pulse (figure 3).Detailed measurements of the gas dynamic characteristics of the pulsed gas valve have been performed for different values of the applied voltage and current in the winding of its control electromagnetic coil and varied initial pressure in its gas chamber.The valve-opening pulse was synchronized in such a way that at the moment of the QSPA plasma arrival at the surface, the concentration of argon was maximal, and it practically did not change during the plasma-surface interaction.
Plasma parameters in the vicinity of the exposed target (electron density and temperature, impurities behaviour in the plasma shield) have been analysed using optical emission spectroscopy.Plasma pressure distributions were measured by  piezoelectric sensors.The interaction of plasma with the W target recorded with a high-speed digital camera PCO AG (10bit CMOS pco.1200 s) [23] was analysed.
Small movable thermocouple calorimeters (a sensitive area of 5 mm in diameter) were used to measure the energy balance during the plasma-surface interaction.Radial distributions of plasma energy density at different distances from the plasma accelerator without the target were measured previously [19].The measured value corresponds to the impacting plasma load.In the present study, a movable calorimeter (axial) was inserted through a hole in the exposed target.The calorimeter could be moved from pulse to pulse along the axis to perform the local measurements at different distances from the surface.These measurements ranged from flush with the surface to 7-10 cm ahead.This allowed the measurement of energy density delivered to the surface through the shield and the study of the spatial shielding characteristics [19].The position where the measured value reaches the energy of impacting plasma corresponds to the actual size of the shielding layer.
The plasma stream velocity was measured by the time of flight of the plasma flow between two magnetic probes located at different distances from the accelerator.The power density in the plasma stream was calculated on the basis of spectroscopic measurements of the temporal distributions of the plasma stream density and velocity.

Results of experiments
In the present experiments, the discharge voltage in the QSPA-M accelerating channel was 10 kV, and the maximal discharge current reached 400 kA.The duration of the plasma pulse exceeded 100 µs.The maximum value of the hydrogen plasma pressure measured by a movable piezoelectric sensor was ∼0.3 MPa.The diameter of the plasma stream in the presence of an external magnetic field was ∼5 cm and it increased to 15 cm when the external magnetic field was turned off.The energy density in the axial region of the plasma stream could be varied within 0.1-1 MJ m −2 [19,21,22].The average value of the plasma density estimated using the Hβ spectral line of hydrogen was around Ne = (2-3) × 10 21 m −3 in a free plasma flow without an external magnetic field.In the presence of the external B-field, it reached Ne = (2-3) × 10 22 m −3 due to the magnetic compression of the plasma column [22].
The results of the performed calorimetry measurements of the energy load distributions are presented in figure 4. As it can be seen in figure 4(a), the local injection of argon leads to an essential decrease in the value of the energy absorbed by the target surface.The zero concentration of argon in this plot corresponds to the conditions of exposure with a pure hydrogen plasma without Ar injection.It should be noted that for relatively small concentrations of seeded argon, up to n Ar ∼ 2.6 × 10 23 m −3 , the energy absorbed by the target decreases by only 5%-7%, regardless of the magnitude of the magnetic field.In the absence of the external B-field, increasing the Ar gas concentration in front of the surface (up to n Ar = 6 × 10 23 m −3 ) results in a steady decline in the energy absorbed by the target.The fraction of the energy dissipated in the shielding layer grows even more significantly in the presence of the external magnetic field, which confines the shielding layer in front of the surface, thereby restricting its flow around the target.The shield in the Bfield expands mainly towards the impacting plasma stream.In  this case, the measured energy load to the tungsten surface is only about 0.3 MJ m −2 , i.e.∼37% of the energy density in the impacting plasma stream.Moreover, this load is almost 1.5 times less compared to the target exposure by the QSPA hydrogen plasma stream without additional Ar seeding.As shown in figure 4(a), even for pure hydrogen plasma exposures with an energy density of 0.8 MJ m −2 , some dynamical screening of the surface by the stopped head part of impacting hydrogen plasma occurs: the surface load is ∼0.45 MJ m −2 and 0.62 MJ m −2 for B = 0.8 T and B = 0, respectively.
Measurements of the energy density distributions in the plasma layer in front of the target under the local injection of Ar have been performed both in the magnetic field and when it was turned off.A sufficiently moderate argon concentration of approximately n Ar = 2.6 × 10 23 m −3 was chosen as the initial condition.Under this condition, the presence of argon impurities in the transient plasma layer does not significantly affect the fraction of energy absorbed by the target surface.Nevertheless, the addition of even such a relatively small amount of argon causes an increase in the thickness of the protective layer in front of the target in the presence of the magnetic field.
If no external magnetic field is applied, the thickness of the shielding layer in front of the target is only about 1 cm since the impacting plasma stream pressure partly removes the plasma shield.A pronounced plasma flow around the target is observed with high-speed imaging.In the external Bfield, the plasma shield is much better retained by magnetic pressure, and its thickness grows significantly (above 5 cm) (figure 4(b)).The dissipation of the impacting energy load in the shield has a linear dependence due to the additional radiation losses from heavy Ar dopant, so one can expect that a further increase in the shield thickness would result in higher shielding efficiency.
Spectroscopic studies of plasma-surface interactions have been performed in both regimes, with and without an external magnetic field (figure 5).The argon lines are recorded at a distance of 5 cm from the target surface (figure 5(b)), which is in qualitative agreement with the measurement of the energy density distributions in front of the target described above (figure 4(b)).Spectral lines of Ar II were detected approximately 30-40 µs after the start of plasma-surface interaction and were observed for ∼20 µs.The electron density was estimated from the Stark broadening of the spectral line of Ar II (4806 A).The spectroscopic measurements confirm that the injection of Ar leads to the increasing thickness of the shielding layer.Also, as a result of the local gas injection along the exposed surface, the electron density in the near-surface plasma increases significantly, reaching up to (1-3) × 10 24 m −3 .As follows from the spectroscopy studies, the radiation intensity from the shield grows considerably when the external magnetic field is applied.It is important to note that in this case, the hydrogen lines are significantly broadened (figure 5) even at greater distances from the target, while the half-width of the argon lines almost does not change.The reason for that is the magnetic field influence on the pinching of impacting hydrogen plasma stream during its transportation along the B-field lines.On the contrary, Ar is seeded locally in front of the target surface.
The spectroscopy measurements were supplemented by monitoring the radiation intensity from the shielding layer using photodiodes in the visible wavelength range.The diode registered radiation from a thin plasma layer of 10 mm in front of the surface.As follows from the radiation intensity measurements, the Ar-enriched shield effectively absorbs the impacting plasma energy.The absorbed energy is then used for the shield heating and expansion and is also re-irradiated by the shield.As shown in figure 6, the radiation intensity in the case of Ar injection is essentially higher in comparison with pure hydrogen.It starts earlier, and the maximum intensity is achieved already in the first instants of plasma stream interaction with the Ar cloud.
These measurements also confirm that in the presence of the B-field, the shielding layer expands mainly towards the incoming hydrogen plasma flow, and the region in which the energy of impacting plasma is converted into radiation also shifts farther from the surface.Thus, the duration of the radiation peak registered from the near-surface plasma becomes shorter.In addition, as the shielding layer is heated, the emission spectrum should be shifted from the visible region to the vacuum ultraviolet wavelengths.Measurements of radiation in the extreme ultraviolet region of the spectrum with an array of AXUV diodes have been prepared; however, unfortunately, further experiments are interrupted by the Russian invasion of Ukraine.

Summary
Dense plasma shield formation in front of the surface under transient plasma heat loads has been studied during a heavy noble gas injection along the target surface in the QSPA-M test-bed facility.An argon shield in front of the tungsten target surface was created using a designed compact gas injector on the base of a pulsed electromagnetic gas valve, which allowed the generation of a homogeneous argon gas flow with a maximum concentration near the target surface ∼n Ar = 6 × 10 23 m −3 and a pulse duration of 0.5 ms.
It has been shown that the injected Ar cloud, being ionised by the impacting plasma stream, forms a dense transient plasma layer in front of the surface.The thickness of such a shielding layer essentially exceeds the mean free path of plasma ions, which is only fractions of a millimeter under given experimental conditions; thus, this layer protects the material from further contact with the impacting high-energy plasma.The electron density of the plasma layer near the surface also increases significantly, reaching (1-3) × 10 24 m −3 .
The experiments with the QSPA-M have shown that an increase in the argon gas density in front of the surface leads to a decrease (1.5-2 times) in the energy load delivered to the target surface.In the presence of a strong magnetic field (up to 1 T), both the thickness of shielding layer and the fraction of the energy dissipated by the shield grow further.Even at moderate energy densities of the QSPA plasma streams in the experiments with Ar gas injection, less than 40% of the impacting plasma load is absorbed by the tungsten surface.The results demonstrate that this additional shielding attributed to the formation of a dense Ar-H plasma layer in front of the exposed surface should be favourable for the divertor armour performance, being important for decreasing the overall erosion of plasma-facing components and promoting suppression of W impurities transport to the core plasma in the course of transient events in a fusion reactor.
The obtained experimental results on the Ar shield formation and its influence on the plasma energy absorption by the target surface could be used for benchmarking the developed numerical simulation codes (for example, TOKES code [7]).It is worth mentioning that material protection from the impacts of plasma and the resulting high heat and particle loads on the surfaces is of undoubted interest not only for fusion reactor design [1,2], but also for solving a wide range of fundamental and applied problems (space engineering [24,25], plasma technologies [26], etc).For example, with the help of gas seeding, it is possible to vary the thermal load on materials when they are processed with powerful pulsed plasma flows in technologies for modifying surface layers in order to improve the performance characteristics of materials.

Figure 1 .
Figure 1.Schematic view of experiment with an additional pulsed injector for Ar seeding along the target surface.

Figure 3 .
Figure 3.Time distributions of argon pressure near the surface for different initial gas pressures in the gas injector valve (a).Average density of Ar cloud near the target during the plasma surface interaction vs. the valve pressure.(b).

Figure 4 .
Figure 4. Energy density absorbed by the surface vs. the argon density in the shield with and without magnetic field (a).Energy density measurements in shielding layer at different distances from the target surface with additional local injection of argon n Ar = 2.6 × 10 23 m −3 .The energy density of impacting plasma stream is 0.8 MJ m −2 .

Figure 5 .
Figure 5. Optical spectra near the target surface with and without B-field (a) and at various distances from the exposed surface with additional local injection of argon n Ar = 2.6 × 10 23 m −3 , B = 0.8 T (b), B = 0 (c).

Figure 6 .
Figure 6.Photodiode signals of visible radiation from the transient plasma layer of 10 mm in front of the surface.