1.4841915.pdf

Extreme ultra-violet movie camera for imaging microsecond time scale magnetic

reconnection

Kil-Byoung Chai and Paul M. Bellan

Citation: Review of Scientific Instruments

84

, 123504 (2013); doi: 10.1063/1.4841915

View online: http://dx.doi.org/10.1063/1.4841915

View Table of Contents: http://scitation.aip.org/content/aip/journal/rsi/84/12?ver=pdfcov

Published by the AIP Publishing

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

131.215.71.79 On: Thu, 20 Feb 2014 15:57:09

REVIEW OF SCIENTIFIC INSTRUMENTS

84

, 123504 (2013)

Extreme ultra-violet movie camera for imaging microsecond time scale

magnetic reconnection

Kil-Byoung Chai and Paul M. Bellan

Applied Physics, Caltech, 1200 E. California Boulevard, Pasadena, California 91125, USA

(Received 24 October 2013; accepted 18 November 2013; published online 12 December 2013)

An ultra-fast extreme ultra-violet (EUV) movie camera has been developed for imaging magnetic

reconnection in the Caltech spheromak/astrophysical jet experiment. The camera consists of a broad-

band Mo:Si multilayer mirror, a fast decaying YAG:Ce scintillator, a visible light block, and a high-

speed visible light CCD camera. The camera can capture EUV images as fast as 3.3

×

10

6

frames

per second with 0.5 cm spatial resolution. The spectral range is from 20 eV to 60 eV. EUV images

reveal strong, transient, highly localized bursts of EUV radiation when magnetic reconnection occurs.

.[

http://dx.doi.org/10.1063/1.4841915

]

I. INTRODUCTION

Time-resolved fast imaging of extreme ultra-violet

(EUV) and soft x-ray radiation is useful to understand mag-

netic reconnection, magnetohydrodynamic (MHD) activity,

particle and energy transport, plasma stability, and turbulence

in laboratory plasmas such as tokamaks

1

–

4

spheromaks,

5

reversed field pinches,

6

and stellarators.

7

Since visible light

technology does not work at EUV wavelengths, conventional

optical schemes based on mirrors and lenses cannot be used

to collect radiation and form images. Various EUV imaging

methods have been used previously, most particularly pinhole

cameras

3

,

4

,

6

,

7

and diode arrays.

1

,

2

,

5

Pinhole cameras consist

of a small pinhole to form an image, a micro-channel plate

intensifier shutter that converts EUV into fast electrons, a

scintillator/phosphor to convert the fast electrons into a visible

light image, and a visible light camera to record the phosphor

image. Pinhole cameras have the advantage of simplicity,

but have extremely low sensitivity because of the necessarily

small pinhole photon collection efficiency. Diode arrays have

the advantage of providing good temporal resolution and

three-dimensional images can be obtained from tomographic

reconstruction of line-integrated diode data. However, diode

arrays have low spatial resolution because cost consider-

ations dictate that only a small number of channels are

feasible.

Photon-efficient, high spatial resolution imaging at EUV

wavelengths has recently become possible using multilayer

mirrors. Unlike conventional silver-coated visible light mir-

rors, multilayer mirrors have high EUV/soft x-ray reflectivity

at normal incidence to the mirror surface. Multilayer mirrors

consist of alternating stacks of two different materials with

half-wavelength periodicity so as to satisfy the Bragg con-

structive interference condition. At EUV wavelengths, the pe-

riodicity is a few nm and so multilayer mirrors are difficult

to fabricate and are expensive. Multilayer mirrors are now

widely used in the EUV lithography industry

8

and in space-

craft telescopes.

9

–

11

For example, the Atmospheric Imaging

Assembly on the Solar Dynamics Observatory uses multilayer

mirrors coupled to back-illuminated CCD cameras

11

to take

EUV images of the sun every 12 s at several different spectral

energies.

We report here an ultra-fast, multilayer-mirror-based,

EUV imaging movie camera. This camera has been designed

to image pulsed plasmas in a lab experiment investigating

complex, dynamic plasma behavior relevant to spheromak

formation, astrophysical jets, and solar corona loops.

12

,

13

Re-

cent observations using filtered EUV-sensitive PIN diode de-

tectors shown in Fig.

1

revealed two distinct strong, transient

bursts of EUV radiation, ranging from 20 eV to 60 eV, that

occur in association with two respective distinct fast, transient

magnetic reconnection events. The characteristic duration of

these EUV radiation bursts is a few

μ

s. The highly transient

nature of the EUV emission and the complex, dynamic, nature

of the visible light images suggest that the EUV burst should

have a well-defined localized, morphology which, if captured,

would provide useful information on the reconnection pro-

cess. In order to capture this image, we have developed the

camera described here.

II. OPTICAL DESIGN

Figure

2(a)

shows the layout of the ultra-fast EUV imag-

ing camera system which consists of a high-speed visible

camera (DRS Imacon 200), a YAG:Ce scintillator (Crytur), a

Mo:Si multilayer mirror (NTT-AT), and two visible-light flat

mirrors. Because EUV cannot penetrate glass and is highly

attenuated in air, the entire EUV optical system is installed

inside the vacuum chamber as shown in Fig.

2(b)

.Thescin-

tillator converts the EUV image into visible light which can

then be observed from outside the vacuum chamber.

The Imacon 200 visible-light camera, capable of 2

×

10

8

fps, is located outside the vacuum chamber and views

the image on the scintillator through a window on the vacuum

chamber. The scintillator must be sufficiently fast to provide

temporal resolution of the reconnection activity which has a

1

μ

s time scale. A YAG:Ce scintillator was selected because

it has a 70 ns decay time. This scintillator converts EUV/x-ray

photons into 550 nm visible photons which are then detected

by the high speed visible light camera. The scintillator is in

the form of crystal to provide high spatial resolution. The

scintillator diameter and thickness are 25 mm and 100

μ

m,

respectively, so that it can stand free without external support.

0034-6748/2013/84(12)/123504/5/$30.00

84

, 123504-1

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

131.215.71.79 On: Thu, 20 Feb 2014 15:57:09

123504-2

K.-B. Chai and P. M. Bellan

Rev. Sci. Instrum.

84

, 123504 (2013)

0

10

20

30

40

50

60

0

10

20

30

40

50

60

70

8

0

90

100

Time

(

s

)

Photodiode signal (mV)

FIG. 1. Al filtered EUV-sensitive PIN diode signal. The first peak corresponds to the spider leg merging time and the second peak corresponds to the

kink/Rayleigh-Taylor reconnection time.

The plasma-facing side of the scintillator is coated with a 200

nm Al film to block visible light while not severely attenu-

ating the desired EUV photons. The transmission for 20–60

eV EUV photons is

>

0.6 and the visible light transmission is

<

10

−

5

. The conversion efficiency of the crystalline YAG:Ce

scintillator for 20–60 eV EUV photons into visible photons

is not known, but is presumed to be similar to the reported

>

1% efficiency of a YAG:Ce powder scintillator (see Fig. 4

in Ref.

14

).

A broadband concave Mo:Si multilayer mirror is used

to focus the EUV photons onto the YAG:Ce scintilla-

tor. This mirror was custom-fabricated by NTT-AT

15

and

has alternating layers of Mo and Si periodically stacked

(

∼

18 nm) to satisfy the Bragg constructive interference con-

dition (

n

λ

=

2

d

sin

θ

where

n

is integer,

λ

is the target wave-

length,

d

is periodicity, and

θ

is the incident angle; here tar-

get wavelength

=

36 nm and

θ

∼

80

o

). The mirror diameter

and focal length are 50.8 mm and 50 mm, respectively. The

M

u

ltilayer

mirror

Mirror

EUV

YAG:Ce Scintillator

(Al coated at front side)

Visi

b

le

Fast

v

isi

b

le camera

Cham

b

er window

(a)

Fast

v

isi

b

le camera

(

b

)

Plasma

FIG. 2. (a) Sketch of ultra-fast EUV imaging diagnostic. (b) Top view of experimental layout showing EUV imaging diagnostic mounted on the first port.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

131.215.71.79 On: Thu, 20 Feb 2014 15:57:09

123504-3

K.-B. Chai and P. M. Bellan

Rev. Sci. Instrum.

84

, 123504 (2013)

FIG. 3. (a) Reflectivity of Mo:Si multilayer mirror.

15

(b) Conversion efficiency defined as number of incoming EUV photons divided by output visible photons.

multilayer mirror reflectivity, shown in Fig.

3(a)

, is provided

by NTT-AT

15

and has a maximum of 13% at 34 eV (36 nm)

and the FWHM is approximately 5 eV. The distance between

plasma and mirror is 60 cm while the distance between mir-

ror and scintillator is 5.45 cm so the optical system provides

a demagnification of 11.

The net photon conversion efficiency of the combined

multilayer mirror, light blocking Al filter, and scintillator, de-

fined as incoming EUV photons divided by output visible-

light photons is shown in Fig.

3(b)

and is between 5

×

10

−

4

and 9

×

10

−

4

for 30–50 eV. Filtered EUV-sensitive PIN diode

detectors in the vacuum chamber indicate that the plasma

emits a transient isotropic 50 kW EUV burst when magnetic

reconnection occurs; this corresponds to 1.0

×

10

22

photons

s

−

1

assuming 30 eV energy for all EUV photons. Taking into

account the solid angle subtended by the multilayer mirror

and the fact that the multilayer mirror is blocked by scintil-

lator support structure by 50%, 2.2

×

10

18

photons s

−

1

will

reach the mirror. Using the EUV to visible photon conversion

efficiency, the number of visible light photons generated by

the scintillator is 2.0

×

10

15

photons s

−

1

. A substantial frac-

tion of these visible light photons are lost because of the small

solid angle subtended by the lens on the visible light cam-

era (lens diameter: 3.57 cm, distance from scintillator: 25 cm,

collection efficiency

=

1.3

×

10

−

3

) and because of the inter-

nal camera optics (efficiency: 1.3

×

10

−

1

). These conversion

and collection efficiencies show that there will be 3.2

×

10

4

visible light photons entering the camera sensor during 100

ns exposure time. Because these photons will cover roughly

50

×

50

=

2500 pixels (obtained from Fig.

7(d)

), one pixel

can get 13 photons for 100 ns, which is sufficient to obtain

meaningful images.

III. RESULTS

A. Visible light test

Before attempting to capture EUV images from an

actual plasma, the spatial resolution and field of view of the

EUV imaging system were tested using visible light. The

scintillator was temporally replaced by a ground glass diffuser

(120 grit) to enable visible photons to travel through the en-

tire system. The remainder of the system was unchanged from

the EUV setup. A set of computer-generated phantom test ob-

jects was displayed on an LCD monitor. These phantom test

objects, shown in Figs.

4(a)

–

4(c)

, were located at the same

60 cm distance from the focusing mirror as a plasma would

be in an actual experiment.

Figures

4(d)

–

4(f)

show images of these phantom objects

produced by the temporary visible-light optical system. The

test object in Fig.

4(d)

is a grid with 25.4 mm

×

25.4 mm spa-

tial periodicity; Fig.

4(d)

shows that the optical system field

of view is from 0 mm to 200 mm in the horizontal direction

and from

−

25 mm to 175 mm in the vertical direction. The

vertical extent of the field of view has been intentionally ar-

ranged to cover the upper part of vacuum chamber because the

plasma jet usually kinks upward. The test object in Fig.

4(e)

consists of 1.27 cm diameter dots, spaced 1.27 cm from each

other; Fig.

4(e)

reveals that the spatial resolution of the op-

tics is about 0.5 cm because FWHM of dots ranges between

0.4 and 0.6 cm. The test object in Fig.

4(f)

is a previously

(a)

(

b

) (c)

(d) (e) (f)

FIG. 4. (a)–(c) Phantom objects used in spatial resolution and field of view

tests. These were displayed on an LCD monitor at the position where plasma

would be present. (d)–(f) Images of phantom objects obtained by replacing

scintillator with ground glass diffuser to achieve visible light measurement.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

131.215.71.79 On: Thu, 20 Feb 2014 15:57:09

123504-4

K.-B. Chai and P. M. Bellan

Rev. Sci. Instrum.

84

, 123504 (2013)

(a) (

b

) (c)

(d) (e) (f)

5.2

s

4.9

s 5.5

s

5.

8

s 6.1

s 6.4

s

FIG. 5. EUV images from actual plasma jet. Images were captured starting

at 4.9

μ

s with 300 ns interframe time and 100 ns exposure time. The source

electrodes are located on the right side of each image as indicated by ellipse.

The circular rings are visible light leakage from the perimeter of the scintil-

lator. The jet speed is estimated to be 21 km/s.

made visible light photo of the plasma; Fig.

4(f)

shows that

the optical system should be able to resolve the structure of a

kinked plasma jet. The spatial variation of brightness and clar-

ity is a consequence of coma aberration caused by the light

source (plasma) being substantially off the axis of the concave

mirror.

B. EUV results

1. First EUV burst (spider leg merging)

Figures

5(a)

–

5(f)

show the EUV images taken from

the actual plasma jet (scintillator now back in place and

EUV optics in vacuum as in Fig.

2(b)

). These images were

captured starting at 4.9

μ

s and had a 300 ns interframe time

and a 100 ns exposure time. The electrode from which the

jet originates is the ellipse located on the extreme lower

right of each image. The large crescent on the left of each of

Figs.

5(a)

–

5(f)

is visible light leaking around the perimeter of

the scintillator. The EUV image originating from the center

of the electrode in each of Figs.

5(a)

–

5(f)

brightens from

4.9

μ

sto5.5

μ

s and then dims; during this time, this bright

region moves leftwards, away from the source electrode. The

size of the brightest EUV segment is about 5 cm

×

3cm.

The estimated velocity of the moving EUV front is

∼

21 km/s

(2.54 cm in 1.2

μ

s) in good agreement with the plasma jet

velocity measured in visible light.

12

Visible light imaging

shows that during this 4.9

μ

sto5.5

μ

s interval, the inner

portions of 8 initial plasma arches having morphology

of a spider, merge to become a single jet.

16

This merging

requires the inner parts of the “spider legs” to undergo

magnetic reconnection. The EUV images are bright where

this merging occurs while regions outside the merging

region are dim. This EUV burst is transient, lasting only

about 2

μ

s. The time at which this bright EUV burst

occurs is in good agreement with the first transient burst

observed by the separately mounted EUV diode. The time

of the first peak in Fig.

1

is different from that of EUV

images in Fig.

5

because they are not from the same plasma

shot.

(a) (c)

(d) (f)

(

b

)

(e)

29.0

s

2

8

.5

s 29.5

s

30.0

s 30.5

s 31.0

s

FIG. 6. Visible light images taken by Imacon camera without EUV optics at

the second port of chamber (See Fig.

2(b)

) from 28.5

μ

s with 500 ns inter-

frame time and 20 ns exposure time. Plasma undergoes kink and Rayleigh-

Taylor instabilities during this period. As a result of this instability cascade,

magnetic reconnection takes place and the plasma breaks off from its source

electrode. Apex of kinked structure dims as the plasma breaks off from its

source electrode.

2. Second EUV burst (kink/Rayleigh-Taylor

reconnection)

The merged jet propagates away from the source

electrode, accelerated by MHD forces.

16

When the jet length

exceeds a critical value given by the Kruskal-Shafranov cri-

terion, the jet suddenly develops a strong kink instability

15

as shown in Fig.

6

(visible light). The lateral acceleration of

the jet caused by the kink instability provides an effective

gravity in the frame of the jet. This effective gravity pro-

vides the environment for a much faster growing secondary

instability, namely, a Rayleigh-Taylor instability.

12

The fast-

growing ripples of the Rayleigh-Taylor instability choke the

jet cross-section down to a scale smaller than the ion skin

depth (Figs.

6(a)

–

6(d)

). This leads to a second and different

magnetic reconnection

12

event which severs the jet from the

source electrode (Figs.

6(e)

–

6(f)

).

Figures

7(a)

–

7(f)

are EUV images, taken during this

time period but from a different shot than the visible images

(from 29.4

μ

s to 30.9

μ

s with 300 ns interframe time and

100 ns exposure time). The lateral velocity of the kink insta-

bility increases from 28 km/s to 85 km/s during 29.7–30.3

μ

s

(a) (c)

(d) (f)

(

b

)

(e)

29.7

s

29.4

s 30.0

s

30.3

s 30.6

s 30.9

s

FIG. 7. EUV images starting at 29.4

μ

s with 300 ns interframe time and

100 ns exposure time. These images were taken from a different plasma shot

than the visible images. As magnetic reconnection occurs, a strong, transient

highly localized EUV burst is observed at 30.3

μ

s.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

131.215.71.79 On: Thu, 20 Feb 2014 15:57:09

123504-5

K.-B. Chai and P. M. Bellan

Rev. Sci. Instrum.

84

, 123504 (2013)

indicating a very large lateral acceleration (

∼

10

11

ms

−

2

). The

EUV signal brightness peaks at 30.3

μ

s coincident with this

large acceleration and appears at the apex of the kinked struc-

ture. The spatial extent of this EUV burst is extremely local-

ized, having an area of only 2.5 cm

×

2.5 cm. After the time

of peak intensity, the bright EUV segment on the apex elon-

gates vertically and then gradually disappears. The duration

of this transient Rayleigh-Taylor-caused reconnection event

is much less than the earlier EUV burst associated with spider

leg merging.

The EUV images during this second magnetic reconnec-

tion event differ considerably from the visible light images.

In the visible light images, the apex of the kinked jet dims

(Fig.

6(e)

) whereas the apex becomes very bright in the EUV

images (Fig.

7(d)

). This presumably indicates that the region

where bright EUV is observed is locally hot because plasma

particles are heated by magnetic reconnection. Also, the later

elongation of the localized “hot spot” suggests that the heated

particles move in both upper and lower directions away from

the original heating location; this hot particle dynamics is not

captured by the visible light images.

IV. CONCLUSION

An ultra-fast EUV imaging movie camera system has

been developed which can capture EUV images as fast as

3.3

×

10

6

fps with 0.5 cm spatial resolution. The images

provided by this camera differ substantially from visible

light images when magnetic reconnection associated with the

kink-driven Rayleigh-Taylor instability occurs. The extreme

spatial and temporal localization of the EUV images suggests

the EUV radiation results from localized heating of plasma

particles by magnetic reconnection. While the temporal and

spatial resolution of the camera system is adequate to resolve

these EUV bursts, the camera is limited by low photon

collection efficiency and by coma aberration of the multilayer

mirror. The low collection efficiency could be improved by

using a multilayer mirror with higher reflectivity, a higher

sensitivity scintillator (LuAG:Ce scintillator improves effi-

ciency 50%

17

), and a thinner Al light-blocking filter (100 nm

film improves efficiency 33%). Each of these improvements

could increase sensitivity by few tens of percent so if the

camera were optimized by these means, the overall camera

performance would increase by a factor of the order 2–5. On

the other hand, if the camera lens system were to be replaced

by an optical fiber imaging bundle system with one face at

the scintillator, sensitivity could be improved by two orders

of magnitude. Eliminating or reducing coma aberration could

be accomplished by either a computer reconstruction method

or by adopting coma-free optics such as Schwarzschild

optics.

18

Moving the scintillator to the axis of the multilayer

mirror would decrease light efficiency by blocking more of

the mirror, but would reduce coma aberration significantly.

We are now investigating upgrading of the camera using a

combination of these methods.

The camera design could be modified to measure higher

energy photons (soft x-ray) by using a multilayer mirror

designed for higher energies. Since scintillator efficiency

increases with incident photon energy,

14

it is expected that

better images would be obtained for soft x-rays than for EUV.

ACKNOWLEDGMENTS

This work was supported by USDOE.

1

G. A. Bobrovskij, D. A. Kislov, E. S. Lyadina, and P. V. Savrukhin,

Rev.

Sci. Instrum.

62

, 886 (1991).

2

B. Alper

et al.

,

Rev. Sci. Instrum.

68

, 778 (1997).

3

S. von Goeler

et al.

,

Rev. Sci. Instrum.

70

, 599 (1999).

4

B. C. Stratton

et al.

,

Rev. Sci. Instrum.

75

, 3959 (2004).

5

A. E. Askren

et al.

,

46th Annual meeting of the Division of Plasma Physics,

Savannah, GA, USA, 15–19 November 2004

(American Physics Society,

2004).

6

T. Onchi

et al.

,

Rev. Sci. Instrum.

81

, 073502 (2010).

7

S. Ohdachi, K. Toi, G. Fuchs, S. von Goeler, and S. Yamamoto,

Rev. Sci.

Instrum.

74

, 2136 (2003).

8

S. Bajt

et al.

,

Opt. Eng.

41

, 1797 (2002).

9

J.-P. Delaboudiniere

et al.

,

Solar Phys.

162

, 291 (1995).

10

J. K. Culhane

et al.

,

Sol. Phys.

243

, 19 (2007).

11

J. R. Lemen

et al.

,

Sol. Phys.

275

, 17 (2012).

12

A. L. Moser and P. M. Bellan,

Nature (London)

482

, 379 (2012).

13

E. V. Stenson and P. M. Bellan,

Phys. Rev. Lett.

109

, 075001 (2012).

14

A. Baciero

et al.

,

J. Synchrotron Radiat.

7

, 215 (2000).

15

Multilayer mirror custom fabricated by NTT-AT, Shinjuku Mitsui Building

2-1-1, Nishi-shinjuku, Shinjuku-ku, Tokyo 163-0431, Japan.

16

D. Kumar and P. M. Bellan,

Phys. Rev. Lett.

103

, 105003 (2009).

17

J. Tous, M. Horvath, L. Pina, K. Blazek, and B. Sopko,

Nucl. Instrum.

Methods Phys. Res. A

591

, 264 (2008).

18

K. Schwarzschild, Abh. Wiss. Gottingen

10

, 4 (1905).

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

131.215.71.79 On: Thu, 20 Feb 2014 15:57:09