Experimental investigation of scrape-off layer blob high density transition in L-mode plasmas on EAST

Ping WANG(汪平),Guanghai HU(胡广海),Ning YAN(颜宁),Guosheng XU(徐国盛),Lingyi MENG(孟令义),Zhikang LU(卢智康),Lin YU(余林),Manni JIA(贾曼妮),Yifeng WANG(王一丰),Liang CHEN(陈良),Heng LAN(兰恒),3,4,Xiang LIU(刘祥),Mingfu WU(吴茗甫)and Liang WANG(王亮),5,∗

1 Institute of Plasma Physics,Hefei Institutes of Physical Science,Chinese Academy of Sciences,Hefei 230031,People’s Republic of China

2 University of Science and Technology of China,Hefei 230026,People’s Republic of China

3 College of Physics and Optoelectronic Engineering,Shenzhen University,Shenzhen 518060,People’s Republic of China

4 Advanced Energy Research Center,Shenzhen University,Shenzhen 518060,People’s Republic of China

5 Institute of Energy,Hefei Comprehensive National Science Centre,Hefei 230031,People’s Republic of China

Abstract Lithium Beam Emission Spectroscopy systems in the outer midplane and divertor Langmuir probe arrays embedded in the divertor target plates,are utilized to investigate the scrape-off layer(SOL)blob transition and its relation with divertor detachment on EAST.The blob transition in the near SOL is observed during the density ramp-up phase.When the plasma density,normalized to the Greenwald density limit,exceeds a threshold of fGW～0.5,the blob size and lifetime increases by 2 – 3 times,while the blob detection rate decreases by about 2 times.In addition,a weak density shoulder is observed in the near SOL region at the same density threshold.Further analysis indicates that the divertor detachment is highly correlated with the blob transition,and the density threshold of blob transition is consistent with that of the access to the outer divertor detachment.The potential physical mechanisms are discussed.These results could be useful for the understanding of plasma-wall interaction issues in future devices that will operate under a detached divertor and high density conditions(over the blob transition threshold).

Keywords:blobs,divertor detachment,L-mode,EAST

1.Introduction

Plasma-wall interaction(PWI)is a critical issue for tokamak fusion research,which involves heat and particle exhaust,and the lifetime of the plasma facing components(PFC).The heat and particle distributions between the divertor targets and the first-wall depend on the competition between the perpendicular and parallel transports in the scrape-off layer(SOL).Blobs(or filaments)in the SOL region have been widely recognized as an important channel for cross-field transport in tokamaks.Therefore,the understanding of blob physics is essential for the study of PWI issues.The blobs are coherent structures that extend along the magnetic field line,and their densities are much higher than the background plasmas[1–5].The interchange and ballooning instabilities compete with the parallel currents,which gives an increase to a poloidal polarization.AnE×Bdrift is caused by this polarization and thus propels the radial transport of blobs,which has been observed in Alcator C-Mod[6,7],ASDEX-Upgrade(AUG)[8],DIII-D[9],JET[10],JT-60U[11],and TCV[12].

The high density transition(or modification)is one of the most important experimental features of the SOL blob,which is associated with the enhancement of cross-field transports and the broadening of the SOL density profile[7,8,13–16].The blob transition generally occurs over a density threshold ofis the plasma line-averaged density,andnGW=Ip/πa2is the Greenwald density limit[17],Ipis the plasma current andais the minor radius.It has been observed that the blob sizes increase by several times at densities above a threshold value[18–21].On AUG,the evolution of the Degree of Detachment(DoD)[22]on the outer divertor is seen to be similar to that of the blob size[19],indicating that the divertor detachment and the blob transition seem to be correlated.The simulation made by Schwöreret alshows similar results[23].Further experimental results in COMPASS,AUG and JET suggest that the blob translates from the ‘sheath limited’(SL)to ‘inertial’(IN)regime when the divertor effective collisionality[18,24],which is in line with the condition for the divertor detachment.The simulation results with the two-dimensional model HESEL are qualitatively consistent with experimental observations[25].However,the studies in JET show that the correlation between SOL profiles(which is commonly associated with the modification of blobs)and divertor collisionality does not hold when detachment is achieved with seeded impurities[26].Militelloet alfound that SOL broadening can occur in MAST in the complete absence of detachment[27].A similar result was shown in Alcator C-mod that the SOL density shoulder is observed by increasing the line-averaged density,but not observed through seeding an impurity in the region of divertor[28].These studies imply that the physical relation of the blob transition with the line-averaged density and divertor detachment is not yet clear.

In the EAST tokamak,blob statistical characteristics[29]and blob properties in the lower hybrid wave dominant heating discharges[30]have been reported in previous work.However,the blob transition with the line-averaged density phenomenon has not yet been studied on EAST.In this work,the blob transition on EAST is investigated to identify its density threshold from the observation of SOL blobs and plasma density evolution.In addition,the relation of the blob transition to the divertor detachment is studied.The rest of this paper is organized as follows.The experimental setup,key diagnostics and EAST experiments are presented in section 2.The method for the analysis of blobs is introduced in section 3.The experimental results are presented in section 4.Finally,the discussion and conclusion are given in section 5.

2.Experimental setup and key diagnostics

EAST is a medium-size superconducting tokamak aiming for long-pulse steady-state high-performance operation with the major radiusR～1.88 m and the minor radiusa～0.45 m.An ITER-like tungsten upper divertor,a carbon lower divertor and molybdenum first wall have been equipped in the device since the 2014 experimental campaign.Note that the carbon lower divertor has been upgraded to a new tungsten divertor in the 2021 EAST campaign.The key diagnostics utilized in this work are the divertor Langmuir probe system[31]and Lithium Beam Emission Spectroscopy(Li-BES)[32,33].The poloidal layouts of the two diagnostics are shown in figure 1.Since the discharges analyzed in this work are operated in the USN(upper single null)divertor configuration,the upper outer(UO)divertor Langmuir probes are mainly used.

The Li-BES system measures the Li-I(2s-2p)line emission intensities.According to the collisional radiative model[34],the Li-I line emission is mainly determined byneat the plasma edge region,i.e.δI / I=δne/ ne.The EAST Li-BES system[32]located in the Port D,consists of 4×32(poloidal×radial)channel arrays with a radial observation range ofR=2.03–2.38 m.The detection system of the Li-BES diagnostic consists of two branches,the CMOS branch and the APD branch.The APD branch is a 4×32 pixel APD detector matrix(APDCAM-10G-4×32 from Fusion Instruments)with a micro-lens array.The fill factor of the Hamamatsu S8550 APD matrix is increased from 50% to nearly 100% by focusing light from the microlens surface onto the sensitive detector areas.Its sampling rate is 2 MHz and radial spatial resolution is about 6–10 mm.Since there is a significant impact from photon noise at high frequencies,the Li-BES data is filtered with a low-pass Butterworth filter,and the cut-off frequency of the filter is 20 kHz.The 14 outermost channels in the third column,labeled from CH3 to CH16,are mainly used to measure blob size and velocity in this study.In addition,a divertor Langmuir probe diagnostic has been installed on EAST to measure the plasma parameters in the divertor region[35,36].In this study,the UO divertor Langmuir probe system in the Port D is utilized,which consists of 13 sets of triple Langmuir probes marked UO-LP1 to UO-LP13[31].The triple probes can obtain the floating potentialVf,the positive biased potentialVp,and the ion saturation currentIsreliably[37].The electron temperatureTetand the particle fluxΓionon the divertor target plates can be calculated fromΓion=js/e=Is/(e Apr)andTet=(Vp-Vf)/ ln 2.Here,jsis the ion saturation current density,andApris the effective area of a probe tip andeis the elementary charge.Its sampling rate is 50 kHz and spatial resolution in the poloidal direction is within 18 mm.

Dedicated experiments with the blob transition and divertor detachment were carried out in the 2017 EAST campaign.These discharges are operated in L-mode plasmas with a similar plasma currentIp～400 kA,magnetic fieldBt～2.4 T,edge safety factorq95～7,and 4.6 GHz lower hybrid wave(LHW)heating powerPLHW～1 MW.Three typical shots #75012,#75016 and #75017 are shown here to demonstrate the repeatability of the results.Figures 2(a)–(c)show the time evolutions of,plasma stored energyWMHDand the heating powerPLHWrespectively.In these shots,the plasma density ramps up fromn¯e=2×1019m-3(fGW～0.3)ton¯e=4.3×1019m-3(fGW～0.7).Due to an uncertainty of the electron temperature measured by the divertor Langmuir probes,the roll-over of particle flux is utilized to define the detachment onset instead of the value ofTet[38,39],andΓionnear the strike points is more focused on than that in the far scrape-off layer or the whole targets[40].The time evolutions ofΓionandTetclose to the strike point on the outer divertor targets are shown in figures 2(d)and(e).With the density ramping up,Γionrolls over andTetdrops to about 5 or 10 eV simultaneously,suggesting that the outer divertor accesses the detachment.A more detailed discussion about the divertor detachment and its correlation with the blob transition will be given in section 4.2.Figure 2(f)shows the magnetic configurations of these three shots,the separatrix of shot#75012 shifts inward a little in contrast to shots#75016 and #75017,hence different Li-BES channels and divertor Langmuir probes will be utilized in the later analysis.

Figure 1.The plasma configuration and key diagnostics involved in this work.(a)Li-Beam Emission Spectroscopy(Li-BES),Divertor Langmuir probes(DLPs),(b)the Li-BES channels layouts distinguished by the row and column numbers.

Figure 2.The time evolutions of(a)the line-averaged density ,(b)plasma stored energy WMHD,(c)LHW heating power PLHW,(d)particle fluxΓion and(e)electron temperature Tet near the strike point on the outer divertor targets,(f)magnetic configurations for EAST discharges#75012,#75016,#75017.

Figure 3.(a)The spatial-temporal emission intensity calculated from the conditional averaging technique,(b)the radial profile ofobtained from the Gaussian fitting at τ=0,(c)the conditionally averaged structure ofat the reference channel.

3.Blob detection with Li-BES diagnostic

The BES diagnostic system has been widely used to investigate the blobs in the SOL region[40–42].To calculate the radial size,lifetime(or self-correlation time),radial velocity and detection rate(or frequency)of blobs,the technique of conditional averaging[43]is applied to the low-pass filtered Li-BES data.Due to the fact that Li-BES noise level is low enough to study low level fluctuations for L-mode discharges in EAST and to get a larger sample population,the blob is defined as an event that the zero-mean reference channel line emissionδI exceeds a threshold of 1.5 standard deviation σ[19,44].In this work,the channel near the separatrix is referred to as the reference channel and the averaged time window of the detected blob is～200 μs.The detection rate fbis defined as the number of detected blobs per second.The normalized spatial-temporal emission intensitycalculated from the conditional averaging technique is shown in figure 3(a),where R is the beam radial coordinate andτ is the time shift relative to the peak point of the blob.Figure 3(b)shows the radial profile ofobtained from the Gaussian fitting atτ=0,the blob radial sizeδblobis evaluated as the half-width at half maximum(HWHM).Figure 3(c)shows the conditionally averaged structure ofat the reference channel,and the lifetime of blob τblobis defined as the full temporal width at half maximum(FWHM).In addition,a sensitivity study on the chosen threshold has been performed,showing that the results of the conditional average do not depend strongly on the selected threshold,and the detection rate decreases with the threshold.However,the evolutions of these blob parameters with time or density are nearly the same.

Figure 4.(a)Conditionally averaged emission intensity fluctuation measured by Li-BES channels CH10–CH12,(b)the temporal evolution of thepeak position(asterisks)and linear fitting(solid line)for velocity estimation.

Considering that the blobs can maintain their structure in a propagation path of several centimeters(the minimum is only 1–2 cm)and the error of the linear fitting with two radial channels is relatively large,three adjacent radial channels are used to calculate the blob radial velocityvrin this work.Figure 4(a)shows the conditionally averagedmeasured by Li-BES channels CH10 – CH12,withvrcalculated from the time delay betweenof these three channels[19,21].The detection timeτ of thepeak increases with distance from the separatrix,showing a radial propagation.Figure 4(b)shows the temporal evolution of thepeak position(asterisks)and the linear fitting for velocity estimation(solid line),and the blob velocity is estimated to be about 1.55 km s-1.

Figure 5.(a)Blob sizeδblob,(b)detection rate fb,(c)lifetime τblob and(d)radial velocity vr versus line-averaged density n¯e at the radial position of R-Rsep～5 mm for EAST#75012(red),#75016(blue)and#75017(black).Dashed lines indicate the density threshold of the blob transition.

Figure 6.(a)Blob related particle transportΓblob at different n¯e and(b)upstream SOL density profiles for shots#75012.The vertical dotted line indicates the radial location of last closed flux surface.

4.Experimental results

4.1.SOL blob and density profile modification

To characterize the blob modification,the blob sizeδblob,lifetime τblob,detection rate fband radial velocity vr,and its evolution withn¯ein the density ramp-up phases are studied in the discharges #75012,#75016 and # 75017.The blobs nearest to the last closed flux surface(LCFS)are detected and analyzed,the reference channels are 11 for shots#75017 and#75016 respectively,and 12 for shot #75012.Figure 5 shows these blob parameters in the position ofR-Rsep～5 mm.It is evident that the blob size,lifetime and detection rate remain almost constant beforeexceeds 3.3 ×1019m-3(fGW～0.5).When the density threshold is exceeded,the blob size increases by about 2 to 3 times and the lifetime increases by about 2 times,while the detection rate decreases by about 2 times.Note that the blob size for shot #75012 is larger than those for shots #75016 and#75017.A possible explanation is that the faster ramp rate of density leads to a higher edge electron density and lower electron temperature.It creates the increase in ion-electron collisionality Λ.The high collisionality yields higher Spitzer resistivity for the blob circuit in the parallel direction,thus the growth rate of the filament instability increases as γ ∝Te-3/2[4,19]and the sheath dissipation decreases,which may increase the blob size.In contrast,figure 5(d)shows that there are unnoticeable variations in the blob radial velocity,especially with unclear characteristics at less than the density threshold.In fact,the radial velocity measurements depend heavily on blob size,lifetime and the spatial resolution of diagnostics.The blob size is smaller than 2 cm in low density and the spatial resolution of the Li-BES is around 6–10 mm at the SOL region in EAST.For the smaller blobs,only three adjacent radial channels could be employed to calculate radial velocity.It plays a crucial role on the measurement error.Moreover,the blob lifetime is also smaller in low density,which results in a relatively larger measurement error.Therefore,the more significant deviation of radial velocity may be caused by the measurement error of lifetime,Li-BES system spatial resolution and the small blobs.The radial convective transport associated with blobs has been estimated for shot #75012.It is defined as Γblob=〈nblob〉〈vr〉fbτb,where〈nblob〉 is the conditional averaging blob density,〈vr〉is the blob radial velocity,fbis the blob frequency andτbis the blob lifetime.As shown in figure 6(a),Γblobincreases withne,suggesting an enhancement of the radial particle transport associated with blobs.Figure 6(b)shows the electron density profile measured by fast frequency sweeping reflectometry[45,46],a weak density shoulder forms in the near SOL region when the line-averaged density exceeds the same threshold value of 3.3 × 1019m-3.The observation suggested that the enhanced radial transport carried by blobs may lead to the formation of the density shoulder,which is in agreement with the observation in AUG[19].

Figure 7.Connection length Lc to different divertor target plates as a function of R-Rsep.Black,blue and red symbols represent Lc to the upper inner(UI),low outer(LO)and upper outer(UO)divertor target plates,respectively.

Figure 8.The particle fluxΓion profiles on the upper outer divertor target projected onto the outer midplane at different for(a)shots#75012,(b)#75016 and(c)#75017.

4.2.Relation between blob modification and divertor detachment

The blob behaviours and propagation are associated with the ion polarization current,the sheath current and the ion current caused by neutral friction[47].One hypothesis is that the blob extends along the field line to the target plates in the SOL and the blob current loop can be closed at the sheath[48].Figure 7 shows the connection lengthLcat the outer midplane as a function ofR-Rsepin USN configurations.Here,Lcis calculated by the magnetic field line tracing code.SinceLcto the UO divertor targets is shorter than that to the UI and LO targets,and the UO divertor is the last to detach,it can be considered as the main electrical interface between the blobs and the divertor targets[19].Therefore,the blob behaviors are compared with the evolution of plasma parameters on the UO divertor targets to illustrate their relation.As shown in figure 8,Γionon the UO divertor target plates in different densities are projected onto the outer midplane.The divertor target strike point is located near the Langmuir probe 9 in discharge#75012,and near probe 13 in discharges #75016 and#75017.It is found that the divertor begins to detach at the density threshold of blob transition,i.e..

Figure 9.(a)Blob sizeδb,(b)particle fluxΓion and(c)electron temperature and on the outer divertor targets,(d)DoD versus lineaveraged density at R-Rsep～5 mm.

A more detailed comparison is carried out to clarify the correlation between the blob transition and the UO divertor detachment.The UO-LP 9(strike point,R-Rsep～5 mm)is projected onto the corresponding coordinates within the Li-BES field of view,enabling the direct comparison between the blob properties and divertor conditions for shot #75012.The UO-LP 9 corresponds to Li-BES channel CH12.Figures 9(b)and(c)showΓionandTetmeasured by UO-LP 9 versus,respectively.Γionrolls over andTetreduces to 5 eV,suggesting the beginning of the divertor detachment.The blob sizes remain at about 2 cm before the divertor detachment,and then increase up to～6 cm nearly at the beginning of(or slightly before)the detachment.Taking into account the measurement errors caused by the differences in spatial resolution between the Li-BES and the divertor Langmuir probe system,it is found that the onset of the divertor detachment and the blob transition occur nearly at the same density.The Degree of Detachmentis calculated as shown in figure 9(d),wherecis a normalization constant and obtained experimentally from the low density phase of the density ramp-up discharge.The dramatic increase in DoD seems to act as an indicator of the blob sharp modification.These results suggest that the blob transition is highly correlated with the divertor detachment.

Figure 10.Divertor collisionality Λdiv versus the line-averaged density at R-Rsep～5 mm.

As mentioned in the introduction,the blob behaviors are closely related to the divertor effective collisionality Λdiv.To examine the validity of the mechanisms proposed in[18,24],Λdivas a function offor discharge #75012 is shown in figure 10.Here,Tetandnemeasured by the divertor Langmuir probes are used to calculate Λdiv.It is shown that Λdivin the positions ofR-Rsep～5 mm increases by up to two orders of magnitude whenand stays nearly constant whenwhich is consistent with the blob transition and divertor detachment density threshold value.This result is qualitatively consistent with the model developed by Myraet al[49],i.e.that the blobs become disconnected from the divertor for large collisionalities.However,the threshold of Λdivis different to AUG[18,24],which is nearly 80 instead of 1,and this may be associated with different blob sizes in the two devices according to Myra’s model.

5.Discussion and conclusion

The modification of SOL blob behaviours is studied in L-mode plasmas with density ramp-up on EAST:the blob size and lifetime increase by 2–3 times while the blob detection rate decreases by about 2 times when the density exceeds the threshold value ofThe formation of a weak density shoulder is observed in the near SOL region above the same density threshold.The evolutions of the blob size in the outer midplane and the divertor plasma parameters show that the access to partial outer divertor detachment and the blob transition occur at almost the same density.It is suggested that the blob transition is highly correlated with the divertor detachment.However,the direction of causality is not yet clear.A possible mechanism is that the increase of density in the SOL leads to a rise in the global collisionality Λ(∝neTe-3/2)along field lines,disconnecting the blobs from the divertor.As a result,the sheath dissipation decreases,which increases the blob size and leads to an enhanced convective radial transports caused by blobs.This would increase or spread the particle loads to the divertor targets,which would trigger or at least contribute to the divertor detachment.Another explanation may also exist,i.e.the higher density leads to divertor detachment,which reduces the electron temperature in front of the targets and thus dramatically increases the divertor collisionality Λdiv.The high collisionality Λdivdisconnects the blobs from the divertor target and results in the blob transition[18].However,figure 8 shows that the outer divertor detachment only appeared in the first 5 mm from the strike points,which seems unlikely to affect about 2 – 4 cm wide blobs;the former direction seems more likely.This conclusion should be taken with caution,due to lack of measurements of collisionality upstream.

The modification of the blob detection rate indicates that other mechanisms may also exist,e.g.the poloidalEr×Bflow shear(Er×Bshear can tilt and split the blobs,which would narrow their radial extent and increase their detection rate[50]),or changes of turbulence and instabilities in the blob generation region.The Li-BES data shows ∇VP(which is proportional to the localEr×Bshearing rate)increases withn¯ein the near SOL.It may suppress the turbulence or blob amplitude,suggesting that the blob transition is not correlated with the change of localErin our case.

Acknowledgments

The authors would like to acknowledge the support and contributions of the EAST team.The work is supported by the National Key R&D Program of China(Nos.2017YFE0301300,2017YFA0402500,2019YFE03030000),Institute of Energy,Hefei Comprehensive National Science Center(Nos.GXXT2020004,12105187),National Natural Science Foundation of China(Nos.11922513,U19A20113,11905255,12005004),Anhui Provincial Natural Science Foundation(No.2008085QA38),and China Postdoctoral Science Foundation(No.2021M702245).

ORCID iDs