Multi-pin Langmuir probe measurement for identification of blob propagation characteristics in the Large Helical Device

Accepted Manuscript Multi-pin Langmuir probe measurement for identification of blob propagation characteristics in the Large Helical Device H. Tanaka, S. Masuzaki, N. Ohno, T. Morisaki, Y. TsujiLHD Experiment Group PII: DOI: Reference:

S0022-3115(14)00743-0 http://dx.doi.org/10.1016/j.jnucmat.2014.10.059 NUMA 48556

To appear in:

Journal of Nuclear Materials

Please cite this article as: H. Tanaka, S. Masuzaki, N. Ohno, T. Morisaki, Y. Tsuji, LHD Experiment Group Multipin Langmuir probe measurement for identification of blob propagation characteristics in the Large Helical Device, Journal of Nuclear Materials (2014), doi: http://dx.doi.org/10.1016/j.jnucmat.2014.10.059

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

P2-035 Multi-pin Langmuir probe measurement for identification of blob propagation characteristics in the Large Helical Device H. Tanakaa*, S. Masuzakia, N. Ohnob, T. Morisakia, Y. Tsujib, and LHD Experiment Groupa a b

National Institute for Fusion Science, Toki 509-5292, Japan

Graduate School of Engineering, Nagoya University, Nagoya, Aichi 464-8603, Japan

Abstract In order to investigate the blobby plasma transport, we have measured electrostatic fluctuations around the divertor leg by using a newly-designed multi-pin reciprocating Langmuir probe in the Large Helical Device. Near the low-field side edge of the divertor leg, positive spikes of ion saturation current fluctuation were observed. In addition, the electric field which correlates with the ion saturation current fluctuation was firstly evaluated with the neighboring floating potential measurement. Considering the positional relationship with the magnetic geometry, the identified direction of the electric field inside the blobs is consistent with the theoretically predicted E ×B motion. By applying the conditional averaging method, a quantitative speed of the blobs was preliminary estimated.

PACS: 52.25.Fi, 52.35.Ra, 52.55.Hc, 05.45.Tp PSI-21 Keywords: Non-diffusive transport, Fluctuation, Langmuir probe, Divertor plasma, LHD *Corresponding Author Address: National Institute for Fusion Science, Toki 509-5292, Japan *Corresponding Author e-mail: [email protected] Presenting Author: Hirohiko Tanaka Presenting author e-mail: [email protected]

1

P2-035 1. Introduction In tokamak devices, blobby plasma transport is a well-known non-diffusive convective transport in the scrape-off layer (SOL) [1]. The blobs, which has a field-aligned structure, are mostly detected in the low-field side (LFS) SOL [2, 3], and they greatly affect edge characteristics such as SOL density profile, flow, and other issues [4]. Blobs ejected from the vicinity of the last closed flux surface (LCFS) are theoretically predicted to propagate toward the LFS direction due to the E × B drift with an internal electric field (E) in the blob and the magnetic field (B) [5]. A number of past experiments reported the existence of the radial particle flux, which is consistent with this theoretical preciction by using multi-pin Langmuir probes [6]. In the Large Helical Device (LHD), being the world’s largest heliotron-type device, SOLs essentially locate in front of the helical coil can. Thus, there is no simple LFS SOL unlike as in tokamaks. Previous studies in the LHD showed that blob-like intermittent events were dominantly detected near the divertor leg [7, 8, 9]. Magnetic geometry around the divertor leg – private region has characteristics similar to those at LCFS – LFS SOL in tokamaks. Further, in a recent study [10], the typical velocity and the size of blobs were estimated by using two distantly positioned probes. However, the existence of an internal electric field inside each blob has not been confirmed yet. In this study, we have shown the first measurement of the internal electric field near the 2

P2-035 divertor leg in the LHD. The helical divertor leg is formed in a toroidally and poloidally twisting shape. In addition, the LFS vector near the divertor leg is three-dimensionally distributed. We therefore analyzed detailed three-dimensional magnetic geometry by using the KMAG code [11]. Then, we designed a new multi-pin probe head suited to the magnetic geometry near that region. By reciprocating the multi-pin Langmuir probe, E fluctuation correlating to the blob-like event was confirmed. In the following, experimental setup is described. In Sec. 3, fluctuation measurement and analyses are shown in sequence. Finally, conclusion is given in Sec. 4.

2. Experimental Setup The LHD has a pair of superconducting helical coils with the helical pitch number of l/m = 2/10. We installed a reciprocating Langmuir probe with a triple-pin head. Figure 1(a) shows insertion trajectory of the probe head with a cross-section of the magnetic-field connection length (Lc) distribution, which is roughly perpendicular to the magnetic field and parallel to the LFS (-∇B) vector in this figure. On the insertion trajectory, the magnetic field strength monotonically increases as decreasing the height (z) from the mid-plane. The standby position of the probe head was set in the private region above an upper X-point. During a measurement, the probe head was inserted downward at a speed of approximately 1.5 m/s. The probe head firstly cuts across a divertor leg with two Lc peaks 3

P2-035 from z ~ 1.18 m, and then the probe passes through the edge surface layers and the ergodic region [12]. This insertion trajectory is almost the same as that of the previous studies in Refs. [9, 10], where the blob-like events were observed. To measure the electric field inside the blobs propagating toward the LFS direction, we arranged the probe tips as shown in Fig. 1(b). Three dome-shaped probe tips with a diameter of 2 mm were positioned on a plane perpendicular to the magnetic field gradient around the divertor leg. One probe tip was used to measure an ion saturation current (Isat) fluctuation, which is reflected by an electron density fluctuation. The other two tips positioned roughly perpendicular to the magnetic field acquired floating potentials (Vf1, Vf2) at neighboring different two locations. Distance between the latter two tips (d) was approximately 7 mm. By assuming that electron temperatures (Te) of a passing blob on the two tips were the same, the electric field across the magnetic field on the probe surface can be deduced by

E ⊥ ≡ (Vf1 − Vf2 ) d . If the sign of E ⊥ is positive, the E × B drift direction corresponds to the LFS in this situation. In the vicinity of the divertor leg, which will be analyzed later, the magnetic field strength was approximately 1.3 T. In an experiment, Isat, Vf1, and Vf2 fluctuations were simultaneously sampled at 1 MHz. During the insertion, core plasma parameters were almost constant with the line averaged electron density of approximately 3.8 ×1019 m-3. The magnetic axis position and the toroidal magnetic field strength on the axis were 3.6m and -2.75T, respectively. 4

P2-035

3. Measurement and Analysis Figure 2(a) shows the Isat signal around the LFS edge of the divertor leg. The horizontal axis indicates probe-head-center height (zc) from the mid-plane. Noise components ascribable to a power supply oscillation were removed before the plot in Fig. 2(a) and before other analyses in all subsequent plots. This figure also shows

I sat

m

, where

m

means a simple

moving average with a sliding window of 1,000 points; the width of the sliding window corresponds to the motion with approximately 1.5 mm along zc. The

I sat

m

distribution

peaked at the Lc peaking position, as shown in Fig. 1(a). Figures 2(b) and (c) show the

~ fluctuation level (F) and the skewness (S) of Isat, which are defined as F = I sat2

~ and S = I sat3

~ I sat2

m

~ I sat ≡ I sat − I sat

m

32 m

12 m

I sat

m

, respectively. Then, we defined the Isat fluctuation component by

. At around zc ~ 1.19 m, F ~ 0.15 and S ~ 0.45. The positive skewness is a

characteristic similar to the results in the previous studies [9, 10]. The positive skewness indicates that positive spikes are dominant in the fluctuation; and the passing of blobs makes the skewness positive. At zc > 1.2 m, S is close to zero and F becomes larger with increasing zc. They are mainly caused by the large background noise, as described below. In Fig. 2(d), simultaneously measured Vf1 and Vf2 are shown. They have negative peaks at different zc. These peak positions correspond to the Lc peaks in Fig. 1(a), and the negative Vf reflects large sheath potential with high Te because plasma potential inside the divertor leg 5

P2-035 would be positive [13]. The above statistics do not contain temporal characteristics. Thus, we employed spectral analysis to the positive-skewness region at zc = [1.1875, 1.1925] m. Figure 3 shows power spectra of Isat, Vf1, and Vf2. There is no spectral peak; periodic fluctuations are not dominant. Figure 3(a) also shows the power spectrum of Isat background noise, which was calculated from the Isat signal at the standby position. Noise components over 300 kHz are comparable to the power spectrum at zc ~ 1.19 m. At zc > 1.2 m, the background noise becomes dominant at the high-frequency range. Next, correlation analysis was applied to confirm the existence of the internal electric field inside the blobs in the positive-skewness region. The correlation coefficient of signals x and y is defined by C (x, y ;τ ) = ~ x (t )~ y (t + τ )

~ x2

12

~ y2

12

,

(1)

~ means an average. Figure 4(a) shows auto-correlation coefficient of I sat ,

where

C (I sat , I sat ) . The coefficient value becomes zero at |τ | ~ 10−15 μs. In addition, a sharpened

shape at τ ≤ 1 μs is also observed, which is partly attributed to the noise components around

~ ~ ~ f ~ 300 kHz. Figure 4(b) shows cross-correlation coefficients between I sat and Vf1 , Vf2 , and

(V~

f1

)

~ − Vf2 . It is noted that there are positive and negative correlation peaks of C (I sat ,Vf1 ) and

C (I sat ,Vf2 ) at τ ~ 0, respectively. Further, C (I sat , Vf1 − Vf2 ) has a positive peak, indicating

~ ~ that I sat and E ⊥ had a positive correlation. This result agrees with the theoretical prediction 6

P2-035 of the blob-propagation direction. To estimate E × B drift speed of the blobs quantitatively, the absolute value of the internal electric field is needed. For this aim, we employed the conditional averaging technique in common with similar past research in tokamaks [6]. Figures 5(a) and (b) show

~ ~ ~ auto- and cross-conditional averaged shapes of I sat and Vf1 , Vf2 , and

(V~

f1

)

~ − Vf2 ,

~ respectively. In this calculation, a total of 52 time points were detected when the I sat has a positive peak larger than a standard deviation for more than 3 μs. After that, subsets of each signal around the detected time points were averaged in a same time domain. Central peak of

~ averaged I sat at τ ≤ 1 is mainly caused by the noise components as well as the correlation analysis. Cross-conditional averaged shapes in Fig. 5(b) have similar tendencies to those of cross-correlation coefficients in Fig. 4(b). The analysis shows that the positive-peak

(

~ ~ amplitude of Vf1 − Vf2

)

is approximately 1 V. Because there were a large number of

~ detection errors due to the I sat noise, this value is underestimated. From the conditional averaging result, the underestimated electric field is

E ⊥ ≈ 1 / 7 ≈ 143 V/m; the propagation speed is vb ≈ E ⊥ B ≈ 143 / 1.3 ≈ 110 m/s. This speed is low but of the same order in the previous study [10].

4. Conclusion

We have used the multi-pin Langmuir probe in order to measure the electric field inside 7

P2-035 the blobs witch appear in the vicinity of the divertor leg in the LHD. Near the LFS edge of the

~ divertor leg where the positive spikes of I sat appeared, cross-correlation coefficients ~ between the I sat and floating potentials were firstly discussed and the internal electric field of the blob was identified. This electric field is consistent with the prediction that the blobs propagate toward the LFS direction due to the E × B drift in common with tokamaks. Moreover, the propagation speed of the blobs was preliminary estimated with conditional averaging analysis.

~ In this study, several analyses were influenced from the large background noise of I sat . Reduction of the background noise is required for more reliable characterization of blobs. Further, two-dimensional measurement perpendicular to the magnetic field with high-spatial resolution, such as gas-puff imaging, should be performed to capture the accurate spatial behavior.

Acknowledgements

This work was supported by KAKENHI (23860064 & 25820440). This research was also performed under the auspices of NIFS budget (NIFS12ULPP029 & NIFS13KLPP028) and NIFS/NINS under the project of Formation of International Network for Scientific Collaborations.

8

P2-035 References

[1] D.A. D’Ippolite, J.R. Myra, and S.J. Zweben, Phys. Plasmas 18 (2011) 060501. [2] G.S. Kirnev V.P. Budaev, S.A. Grashin, L.N. Khimchenko and D.V. Sarychev, Nucl. Fusion 45 (2005) 459−467. [3] H. Tanaka, N. Ohno, N. Asakura, Y. Tsuji, H. Kawashima, S. Takamura, Y. Uesugi and the JT-60 Team, Nucl. Fusion 49 (2009) 065017. [4] A.Yu. Pigarov, S.I. Krasheninnikov, B. LaBombard, and T.D. Rognlien, Contrib. Plasma Phys. 48 (2008) 82−88. [5] S.I. Krasheninnikov, Phys. Lett. A 283 (2001) 368−370. [6] J.A. Boedo, D. Rudakov, R. Moyer, S. Krasheninnikov, D. Whyte, G. McKee, G. Tynan, M. Schaffer, P. Stangeby, P. West, S. Allen, T. Evans, R. Fonck, E. Hollmann, A. Leonard, A. Mahdave, G. Porter, M. Tillack, and G. Anter, Phys. Plasmas 8 (2001) 4826−4833. [7] N. Ohno, S. Masuzaki, H. Miyoshi, S. Takamura, V.P. Budaev, T. Morisaki, N. Ohyabu and A. Komori, Contrib. Plasma Phys. 46 (2006) 692−697. [8] H. Tsuchiya, T. Morisaki, V.P. Budaev, A. Komori, H. Yamada, and LHD Experimental Group, Plasma Fusion Res. 5 (2010) S2078. [9] H. Tanaka, N. Ohno, Y. Tsuji, S. Kajita, S. Masuzaki, M. Kobayashi, T. Morisaki, A. Komori and the LHD Experimental Group, Plasma Fusion Res. 7 (2012) 1402152. [10] H. Tanaka, S. Masuzaki, N. Ohno, T. Morisaki, Y. Tsuji, and the LHD Experiment Group, J. Nucl. Mater. 438 (2013) S563−S566. [11] Y. Nakamura, M. Wakatani and K. Ichiguchi, J. Plasma Fusion Res. 69 (1993) 41. [12] N. Ohyabu, T. Watanabe, H. Ji, H. Akao, T. Ohno, T. Kawamura, K. Yamazaki, K. Akaishi, N. Inoue, A. Komori, Y. Kubota, N. Noda, A. Sagara, H. Suzuki, O. Motojima, M. Fujiwara, and A. Iiyoshi, Nucl. Fusion 34 (1994) 387−399. [13] N. Ezumi, S. Masuzaki, N. Ohno, Y. Uesugi, S. Takamura, and LHD Experimental Group, 9

P2-035 J. Nucl. Mater. 313−316 (2003) 696−700.

10

P2-035 Figure captions

Fig. 1. (a) Insertion trajectories of probe tips and a cross-section of Lc which is roughly perpendicular to the magnetic field around the divertor leg. (b) Photographs of the probe head.

I sat

Fig. 2. (a) Distributions of Isat (solid line),

m

(dashed line), (b) fluctuation level, (c)

skewness, (d) Vf1 (solid line), and Vf2 (dashed line) along zc. The hatched region corresponds to zc = [1.1875, 1.1925].

Fig. 3. (a) Power spectra of Isat (solid line), background of Isat (dashed line), (b) Vf1 (solid line) and Vf2 (dashed line).

~ ~ Fig. 4 (a) Auto-correlation coefficient of I sat . (b) Cross-correlation coefficients between I sat

(

~ ~ ~ ~ and Vf1 (solid line), Vf2 (dashed line), and Vf1 − Vf2

)

(dotted line).

~ Fig. 5 (a) Auto-conditional averaged shape of I sat . (b) Cross-conditional averaged shapes of

(

~ ~ ~ ~ Vf1 (solid line), Vf2 (dashed line), and Vf1 − Vf2

11

)

(dotted line).

P2-035 Fig. 1. single-column

12

P2-035 Fig. 2. single-column

13

P2-035 Fig. 3. single-column

14

P2-035 Fig. 4. Single-column

15

P2-035 Fig. 5. Single-column

16