Role of electric currents for the SOL and divertor target heat fluxes in ASDEX Upgrade

In divertor tokamaks electric currents flowing in the Scrape-Off Layer (SOL) are a frequently observed and well known feature. However, the convective heat flux carried by these currents is often assumed to be small compared to the total parallel heat flux. In this paper it is shown that in ASDEX Upgrade the electric current, which was measured by Langmuir probes at the outer target, can be several times larger than the ion saturation current in low density L- and H-mode discharges. Therefore they have to be taken into account when calculating the target heat flux from the Langmuir probe data. Moreover, a detailed analysis reveals that for low density L- and H-mode discharges the heat flux profile is mainly determined by the current profile. By applying a 1D heat flux transport model it is demonstrated that the electric current can carry a large fraction (>  50 %) of the electron heat flux along the SOL convectively. Thus a Spitzer-Härm approach, where it is assumed that heat is transported only conductively, might not be valid in the investigated regime.


Introduction
In divertor tokamaks, such as the next step fusion device ITER, heat is transported across the separatrix and flows in the Scrape-Off-Layer (SOL) within a narrow channel [1] towards the inner and outer divertor targets. Due to a combination of drifts [2], geometry effects, and poloidally varying anomalous 4  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. transport, the plasma tends have different temperatures at the two target plates. This temperature asymmetry drives a current carried by electrons flowing from the colder towards the hotter target along the SOL [3,4], which will be called thermoelectric current in the following. The flow of electrons can contribute significantly to the parallel heat flux towards the hot plate, and thereby increase the already existing temperature asymmetry further [5,6]. Analytical and numerical SOL models often neglect this contribution to the transport and assume that heat is carried only by conduction. Thus it is of high importance to know under which conditions this is valid.
SOL currents are investigated experimentally in several tokamaks. In Alcator C-Mod [7] and JET [8] electric currents larger than the ion saturation current were measured. By calculating the heat flux from LP measurements the authors showed that the currents contribute significiantly to the target heat flux.
In ASDEX Upgrade (AUG) thermoelectric currents are known to occur [9], and their measurement by shunt resistors are useful to approximate the outer target electron temperature for detachment control [10].
In this study Langmuir Probe (LP) measurements of divertor target currents of AUG L-and H-mode discharges are presented and the implications for LP data analysis and for the heat flux transport are discussed. In section 2 the LP diagnostic setup and analysis are explained, and in section 3 the discharge setup and measurements are presented. The physical mechanism driving these currents and the dependence of the electric current on the density are discussed in section 5. In section 4 the heat flux profiles calculated from the Langmuir probe measurements are compared to infrared data, and the role of the current for the target heat flux is discussed. In section 6 a simplified model for the SOL heat transport with electric currents is solved for typical low density L-and Hmode conditions, and it is shown that a significant contribution of the heat can be carried by the electric current. In section 8 the results are summarized.

Diagnostic setup
The main diagnostics used for this study are the Langmuir probes in the lower divertor of AUG, which are displayed in figure 1 (a). The red dots at the outer target correspond to dome shaped single probes which have a radius of about 2.5 mm and are separated poloidally by about 2 cm. A triangular voltage signal is applied to them, i.e. they are swept to obtain currentvoltage (IV) characteristics. The sweeping frequency is about 200 Hz. As an example in figure 1 (b) an IV characteristic taken in an L-mode discharge is shown. On the horizontal axis the applied voltage with respect to the target potential V tar , and on the vertical axis the measured current is plotted. A model is fitted to the characteristic from which the electron temperature T e and the ion saturation current density j sat are obtained. The electric current density to the target j net is calculated from the current drawn by the probe at V = V tar , i.e. when it is grounded. In this paper the following sign convention for the current densities is chosen: j net is positive for a net electron current and negative for a net ion current towards the target. The ion saturation current density j sat has a different sign, i.e. it is positive. This sign convention allows a more compact representation of j net and j sat profiles in the same plot.
The fitting model from [11] is used to describe the non saturation of the ion flux [12,13] due to the expansion of the Debye sheath with increasing probe voltage. Moreover a double probe characteristic is assumed to take into account the decreased electron saturation current which is typically observed at shallow magnetic field incidence angles [13,14].
The black dots in figure 1 (a) correspond to flush mounted probes, which are toroidally displaced with respect to the dome shaped probes and have poloidal and toroidal extensions of about 5 and 25 mm, respectively. The flush mounted probes are commonly operated as triple probes [15] by which T e and j sat are obtained at a frequency of about 25 kHz. Since in a triple mode operation not the full IV characteristic is obtained, these fitted values are less reliable then the ones obtained by the swept probes. An additional grounded flush mounted probe can be used to obtain j net . The data from the flush mounted probes are only used in section 5 to compare inner and outer target fluxes, whereas the data in the other sections are obtained from the dome shaped probes.
The non saturation of the ion flux depends on the probe geometry and the incident field line angle onto the probe [12,13]. However, the non saturation effect should not play a role when the probe is at the same potential as the target, i.e. when the electric current is measured. Hence, the measurement of the electric current should be independent of the probe shape. In figure 2 the electric current density profiles obtained from swept flush and dome shaped probes in an L-mode during a strike line sweep are compared. The inclination angle between the probes and magnetic field was here about 2.5 • for the flush probe and about 5 • for the dome probe. The good agreement between the two measurements corroborates the assumption that the probe shape does not influence the electric current measurements.
The outer target is also observed by an IR camera to measure the target temperature with a time resolution of about 1 kHz [16]. The heat flux onto the divertor target is calculated from the temperature measurements using the implicit version [17] of the THEODOR code [18].

Discharge setups and current measurement results
In the following, two well diagnosed low collisionality deuterium L-and H-mode discharges and related LP measurements are discussed. We focus on low collisionality discharges because, as will be seen in section 5 the SOL currents have the strongest effect on the target heat flux for an attached inner divertor. Moreover, studies on the heat flux width often focus on low collisionality discharges as well, since the target IR data is more reliable at low density and can be accurately described [19] in the absence of dissipation [20]. Therefore, the results presented here directly contribute to a better understanding of the divertor heat flux estimates in such low density discharges.
In both discharges the magnetic field is B t = -2.5 T, and the ion ∇B drift is directed towards the active X-point, which is the standard direction for tokamak operation. With this setup the outer target tends to have a higher electron temperature than the inner one, i.e. the thermoelectric current is directed towards the inner divertor. For the case with a reversed magnetic field, i.e. with the ion ∇B drift directed away from the active Xpoint, it was shown that the temperature asymmetry between inner and outer divertors becomes smaller [21]. Therefore, it is expected that the standard magnetic field configuration shows larger thermoelectric currents than the reversed field configuration. Figure 3 shows the time traces for the L-mode discharge 36 280 of (a) the plasma current I p , (b) the Electron Cyclotron Resonance Heating (ECRH), Neutral Beam Injection (NBI)  and Ohmic heating powers P ECRH , P NBI and P OH , (c) the density n H5 measured along the interferometry line of sight H5 [22] and (d) the strike point position at the outer target with respect to an arbitrarily chosen offset location. The line of sight H5 passes through the edge, n H5 , and is used as a measure for the edge density.
For the LP analysis the time range between 6.0 and 6.35 s is used, marked red in figure 3. In this time range the outer target strike point is moving by about 2.5 cm to obtain continuous LP profiles. The plasma current I p decreases from 0.8 MA to about 0.75 MA during that time range. The edge density n H5 is about 0.9 ×10 19 m −3 and P ECRH = 0.4 MW. Measurements by the edge Thomson scattering system show that the variation of I p does not affect the upstream density and electron temperature profiles within the measurement uncertainties.
The outer target profiles of j sat and j net are shown in figure 4 (a). Here s is the target coordinate and s sep is the strike line location. The dots are the data obtained from individual IV curves and the solid lines polynomial fits to the data. The j net profile shows a narrow width of about 1 cm and has a peak value of about 5 kA/m 2 . Compared to j sat it is very narrow and about three times higher in peak value. The electron temperature T e is shown in figure 4 (b). The peak value of T e is about 60 eV. This might be an overestimation of the actual value, as the fitted parameters become more uncertain at high electron temperatures. Furthermore, an underestimation of the non saturation of the ion flux can as well lead to an overestimation of T e [23].
The time traces for the H-mode discharge 36 052 are shown in figure 5. The time range used for the LP profiles is between 5.0 and 6.5 s, indicated by the red shaded region. In figure 5 (d) the outer target strike point position is shown. In this time range a strike line sweep of about 4 cm is performed in order to obtain continuous probe profiles. The ELMs, which are clearly observed in P OH and the strike line movement, appear at a low frequency of about 70 Hz . The long inter ELM phases of about 10 ms allow the analysis with swept probes. During the inter ELM phases the strike line position is still fluctuating with an amplitude of about 0.5 cm. This makes the data analysis close to the profile peaks where the decay lengths are about 1 cm challenging. In particular, this leads to a smoothing of the profiles around the peak position. Figure 4 (c) shows the outer j net and j sat profiles and (d) the T e profile for this discharge. In this case j net has a peak value of about 15 kA/m 2 , which is about twice the value of j sat at the same location, and becomes negative in the Private Flux Region (PFR). The dip of j net in the PFR is explained by the diamagnetic current across the divertor leg due to the poloidal pressure gradient [2,24,25]. To ensure that the electric current is divergence free, it must then flow along the field lines towards the target.

Target heat flux calculation from LP measurements and comparison to infrared data
The fact that the outer target currents can be significantly higher than the saturation current is crucial for the interpretation of target heat flux profiles. This will be demonstrated in this section by analysing the heat flux profiles of the two Land H-mode discharges discussed in section 3. From sheath theory [26] it is well known that, assuming equal electron and ion temperatures, about 2/3 of the heat flux is transmitted by the ions to the target for ambipolar conditions, i.e. for j net = 0. This, however, is not valid for strongly non-ambipolar conditions.
The individual heat flux contributions are calculated from LP measurements by decomposing the total heat flux according to the model described in [26], chapter 25.5: with q e = 2T e (j sat + j net ), (2) Here q e and q i are the electron and ion contributions and q rec the energy deposited due to recombination at the target surface. T i is the ion temperature, V sh the potential drop in the sheath forming in front of the target, R E the ion energy reflection coefficient [27], and E rec = 13.6 eV the recombination energy per impinging ion. Secondary electron emission has been neglected due to the prompt redeposition of electrons at small magnetic field incidence angles [28]. The sheath potential drop is calculated according to where the term V pl,fl is the difference between the plasma and the floating potential [26], and the second term V tar,fl is the difference between the floating and the target potential. Since −V tar,fl is approximately the voltage at which an LP draws zero current this term can also be read directly from LP measurements, as it is done in this study. While the contributions q e and q rec are calculated directly from the LP measurements, this is not possible for q i without further assumptions on T i and R E . R E depends both on the ion incidence energy and angle as well as the surface material  [27]. For ions impinging perpendicular to a tungsten surface, and ion temperatures above 20 eV, an almost constant value of R E ≈ 0.5 is reported [27,29]. Due to a complete lack of information about the ion temperature, T i is assumed to be between 0 and 2T e . This assumption should give a range of realistic ion heat fluxes.
It is common to express equation (1) in terms of ion and electron sheath heat transmission factors γ i and γ e , defined by [26] For fixed R E and T i /T e , the total sheath heat transmission factor γ tot ≡ γ i + γ e is a function of the ratio j net /j sat . In figure 6 (a) γ tot , γ i and γ e are plotted for the case T i = 0. Here the convention from [26] was adopted to use eV tar,fl /T e as an independent variable. With increasing eV tar,fl /T e , the sheath potential drop decreases (see equation 3) and more electrons pass the sheath. As a consequence γ e is increasing exponentially. The ion sheath heat transmission factor γ i on the other hand is decreasing corresponding to a decreased energy gain of the ions in the sheath. The total sheath heat transmission factor γ tot is 3.6 for V tar,fl = 0 and increases to about 6.6 at eV tar,fl /T e = 1, which corresponds to the case j net /j sat ≈ 3. The sheath heat transmission coefficients for T i = 2T e are shown in figure 6 (b). Due to the higher ion energies with respect to the case T i = 0, the sheath heat transmission factor increases to 5.8 at V tar,fl = 0 and 8.8 at eV tar,fl /T e = 1. The heat fluxes q e , q rec and q i obtained for the L-mode discharge 36 280 are shown in figure 7 (a). To obtain these profiles, the polynomial fits shown in figure 4 (a) were used for j sat , j net and T e in equation 1. The dashed magenta line shows the ion heat flux for the assumption T i = T e . The lower and upper boundaries of the shaded region correspond to the ion heat fluxes for T i = 0 and T i = 2T e , respectively. The electron heat flux q e is between 2 and 5 times larger than q i , and has a several times shorter decay length. The recombination heat flux q rec is negligible. The total heat fluxes q mod is plotted in figure 7 (b) together with the heat flux q IR obtained from IR measurements. A good agreement with q mod is obtained for s − s sep > 0 for the whole range of assumed ion temperatures. Although q IR tends to be below the q mod obtained for T i = T e it cannot be concluded that T i < T e , since the effect of T i on the total heat flux is relatively small with respect to the uncertainties in the probe measurements and the IR data. In the PFR q IR decays steeply, which is not seen in q mod .
The heat flux profiles for the H-mode discharge 36 052 are shown figure 7 (c) and (d). The obtained peak value of q e is between one and three times that of q i . Close to the peak value, where as discussed in section 3 the uncertainties are high, q IR is underestimated.  The analysis made in this section shows that the heat flux decay length in the discussed cases is strongly determined by the electron heat flux decay length. The electron heat flux decay length, on the other hand, is mainly determined by the j net profile. In section 5 is was furthermore shown that in low density L-mode plasmas the j net profile is determined by the inner target saturation current profile, with additional contributions from the radial diamagnetic drift. Conclusively, the thermoelectric current leads to a coupling between the outer target heat flux and the inner target saturation current in low density discharges.

Change of outer target current through inner target conditions
The examples discussed so far in this paper demonstrate that electric currents several times larger than the ion saturation current can occur at the outer target and that this current can contribute significantly to the target heat flux. It is now shown that the electric current profile is largely explained by a thermoelectric effect. Furthermore, it is analysed how a change of the inner target conditions can affect the thermoelectric current.
For a floating target, i.e. if the current to the target is zero, the sheath potential is approximately proportional to the electron temperature. If the electron temperatures at the inner and outer targets are different thermoelectric currents appear in order to compensate the difference of sheath potentials [3]. For zero parallel electrical resistivity this difference is compensated completely in the sheaths. Assuming a finite resistivity the potential difference is compensated partly in the SOL between the sheaths and the current is reduced compared to the model with zero resistivity. In [4] the thermoelectric current is modeled by assuming a Langmuir type model for the sheaths, finite resistivity and constant plasma pressure in the SOL, resulting in the following equation for the thermoelectric current: Hereĵ ≡ j th /j sat,in is the thermoelectric current j th normalized to the inner target saturation current density j sat,in , T e,out and T e,in are the electron temperatures at the outer and inner targets, the coefficients κ sh = 3.89 and κ presh = 0.85 are the floating and presheath potentials normalized to the target electron temperature, α = 0.7 is the normalized temperature gradient force onto the electrons, L ∥ the parallel connection length between the inner and outer targets and η ∥ the averaged parallel resistivity. For an effective charge number Z eff = 1, η ∥ is given by where lnΛ is the Coulomb logarithm, and T e the average electron temperature. The inner target saturation current density required to solve equation 5 is calculated by j sat,in = n e,in 2eT e,in /m i , where m i is the ion mass. Due to the assumed pressure conservation, n e,in is related to the average density n e by n e,in = n e T e /T e,in . By solving equation 5ĵ is obtained as a function of the ratio T e,out /T e,in . Figure 8 shows the solution of this equation for n e = 0.5 ×10 19 m −3 , T e = 60 eV, T e,in = 20 eV and L ∥ = 60 m, which corresponds approximately to the parameters of the discharge 36 280. The value chosen for T e,in is the peak temperature measured by the triple probes at the inner target.
For T e,out /T e,in < 2, j th is significantly smaller than j sat,in , while for T e,out /T e,in > 2, which is fulfilled in the discharge 36 280, j th approaches to j sat,in .
For the discharge 36 280 no reliable full inner target ion saturation profiles could be obtained. Instead, LP data from the L-mode discharge 34 320, with similar conditions as the discharge 36 280, is used for this comparison. In figure 9 the profiles of j net measured at the outer target and j sat,in f out in with respect to the poloidal flux coordinate ρ p are shown. Here f out in is the flux expansion between inner and outer target. The shape as well as the magnitude of both profiles are in good agreement. The negative dip in j net in the PFR is caused by the diamagnetic currents, as discussed before. The agreement between the two profiles confirms that the electric current observed at the outer target is dominantly driven by the thermoelectric effect.
The discharges discussed in the previous sections correspond to conditions where the outer as well as inner targets are attached. However, as the density is increased the outer target ion saturation current increases and the electric current decreases due to the detachment of the inner divertor. These effects should lead to a decrease of the ratio j net /j sat,out . To quantify this dependence a series of L-mode discharges with varying densities is analysed. The ion saturation current j sat,out in these discharges is obtained from the flush triple Langmuir probes, and the electric current j net from grounded flush probes. Figure 10 (a) shows the peak value of j net and the j sat,out value at the same target location in dependence on the edge density n H5 . j net increases for low densities slightly, while for n H5 > 1.5 ×10 19 m −3 it drops again. Although it should be noted that for higher densities the data is sparse, the drop of j net is consistent with the picture that the thermal current decreases above a certain density due to a detaching inner divertor. j net does however not reach zero, which may be explained by the aforementioned diamagnetic currents. The saturation current j sat,out increases approximately linearly with n H5 . Figure 10 (b) shows the ratio of the j net and j sat,out values from figure 10 (a). For the lowest densities the ratio is around 6 and decreases approximately linearly to values below one with increasing density. As is seen from equation 1 the importance of the electric current increases for an increasing ratio of j net and j sat,out . For increasing densities the contribution of the electric current to the target heat flux becomes therefore less significant.

Model for parallel heat transport with strong electric currents
Thermoelectric currents are carried by a flow of electrons from the colder to the hotter target, and therefore this mechanism leads to a convective heat flux. To estimate the contribution of this convective heat flux to the total electron heat flux, a one dimensional heat transport model is employed. In the model it is assumed that 1) there are no volumetric heat sources and sinks, 2) electron-ion energy exchange is negligible in the SOL. With these assumptions the following heat flux equation for the electrons is obtained from the Braginskii equation for the electron temperature [6,30]: Here q e,∥ is the parallel electron heat flux, κ e ≈ 2400 Wm/eV 2/7 for Z eff = 1 [31], and l the distance from the outer target along the field line. The terms q cond,e and q conv,e on the left hand side of equation 7 are the conductive and convective parts of the electron heat flux, respectively. The conductive term is set according to Spitzer-Härm. Both terms, j net,∥ and q e,∥ , are assumed to be constant along the field line, and the outer target electron temperature T e,out is a free input in the model. The electron heat flux q e,∥ is obtained from the probe measurements by adding the heat flux required to overcome the potential drop of the sheath to the electron target heat flux: The factor 1/sin(α), where α is the angle between the magnetic field and the target plate, is required to convert perpendicular (i.e. towards the target) to parallel heat fluxes. Here α = 2.5 • is used, which is a typical value at the outer target in AUG. The sheath drop V sh , which is calculated from equation 3, is typically between 2 and 3T e /e, i.e. the upstream electron heat flux is around twice that of the target electron heat flux. Equation (7) is solved numerically for both the L-and Hmode cases with the following parameters: • For the L-mode discharge T e,out = 60 eV is used (see. The peak target current j net is in both cases scanned between 0 and 30 kA/m 2 . The solutions for the upstream (25 m parallel upstream away from the target) convective heat fluxes normalized to q ∥,e are shown in figure 11 (a) as functions of j net . For j net = 0 the convective part is zero, and the electron heat flux is described exactly by the Spitzer-Härm term. With increasing j net the normalized convective flux increases until it reaches unity. For the L-mode discharge the peak value of j net is about 4 kA/m 2 , and, therefore, between 40 and 80 % of the total electron heat flux is carried convectively. For the H-mode discharge the peak value of j net is about 15 kA/m 2 , and between 60 and 90 % is carried convectively. However, the maximum target current might be higher than the upstream current, due to the diamagnetic currents. Nevertheless, assuming that the upstream current is only half of the target current, still about 50 % is carried convectively. The convective heat flux from the colder to the hotter target also contributes to the heat flux asymmetry between the targets as discussed in references [5,6].
The upstream electron temperature T u e obtained from the solution of equation 7 is shown in figure 11 (b). For all solutions T u e is decreasing approximately linearly with j net . The slope dT u e /dj net is for all solutions about -1 eV/(kA/m 2 ). Therefore, for j net = 4 kA/m 2 the upstream temperature is about 4 eV lower, and for j net = 15 kA/m 2 about 15 eV lower than what would be obtained by assuming pure Spitzer-Härm heat transport.

Application of results: Estimation of the power decay length from LP data
Attached heat flux profiles can be well described by the model described in in [19,32]. The fitting parameters q 0 , λ q and S used in this model parametrize the upstream peak heat flux, the e-folding length of the profile, and the spreading of the profile towards the PFR. In particular λ q is an important quantity to predict the power exhaust capabilities in future fusion devices. A large λ q is not only beneficial since the power spreads over a larger area, but also since the radiating volume increases [33]. To calculate the target heat flux from LP data it is commonly assumed that the target is floating. In the following it is demonstrated how this assumption influences the estimation of the heat flux profile and the corresponding fitting parameters.
Floating target conditions correspond to V tar,fl = 0 in figure 6 (a) and (b). With respect to V tar,fl > 0 the electron heat flux is underestimated, and the ion heat flux overestimated. Since the underestimation of γ e is stronger than the overestimation of γ i , the total sheath heat transmission factor decreases. Figure 12 shows for the L-mode #36 280 the IR heat flux q IR , the heat flux profiles obtained from the Langmuir probes with the correct sheath heat transmission factor q mod , and with the  Outer target heat flux profiles from IR (q IR ), from LPs calculated according to equation 1 with T i = Te and R E = 0.5 (q mod ) and from LPs calculated with the standard formula for floating conditions (q mod,fl ). The corresponding fits are obtained according to the model described in [19]. assumption of a floating target q mod,fl . The fits of the profiles are shown by the solid lines. For the LP heat fluxes T i = T e and R E = 0.5 is used. The floating target assumption results in a spatially constant sheath heat transmission factor of about 5.
The power decay lengths obtained for the fit functions are about 3 mm for both q IR as well as q mod . For q mod,fl the decay length is 4.1 mm. Thus with a constant sheath heat transmission factor λ q is overestimated by about one third in this case. This is in line with the results from [34], where the power decay length obtained from LP measurements with a constant sheath heat transmission factor is systematically by a factor of about 1.3 larger than the one obtained from IR measurements.

Conclusions
Electric currents measured with shunt resistors in the divertor targets of AUG were previously explained by the thermoelectric effect. The Langmuir probe measurements presented here show that for low collisionality SOL conditions, the electric currents at the outer target can be several times larger than the ion saturation current for both L-as well as H-modes. The outer target electric current and the inner target ion saturation current profiles are compared for an L-mode case, and good agreement between them is found. This shows that the electrons flow from the inner to the outer target, while the inner target flow is depleted of electrons. By increasing the density the electric current decreases, while the outer target ion saturation current increases, leading to a decrease of their ratio.
Due to the strong electric currents the assumption of an ambipolar flux to each of the targets is not fulfilled. This has to be taken into account when calculating the heat flux from the Langmuir probe data. The analysis of the obtained heat flux profiles shows that the ion and electron heat flux shapes differ strongly in the investigated low density discharges. In particular, including the current in the analysis leads to higher peak heat fluxes, and to smaller near SOL heat flux widths derived by the LP method.
If the ratio of electric to ion saturation current is small, the electric current can be neglected in the heat flux calculation.
The assumption of a spatially constant sheath heat transmission factor does therefore not lead to large errors at high densities.
The commonly used assumption of using a pure Spitzer-Härm heat transport is inspected by setting up a simplified model including the heat transport by the electric current. Solutions of this model show that, for the experimentally observed conditions, less than 50 % of the electron heat flux might be carried by conduction. The model predicts, furthermore, that the upstream temperature drops approximately linearly with the current. For the investigated conditions this led to a drop betwen 5 and 15 eV, compared to the case with no current.
High negative target currents require the sheath potential drop V sh to decrease with respect to the commonly used approximation of V sh = 3T e /e, which has not yet been addressed in this work. This has to be taken into account when calculating the electric fields upstream and in the divertor region from the target conditions. In particular the decreasing electric currents through increasing density, might affect the upstream radial electric field and its shear, and thereby have an effect on the L-H transition and blob transport. Analysing the upstream electric field with regards to SOL currents, and its further effects will be the subject of future work.