optical lever for kagra - gwdoc.icrr.u-tokyo.ac.jp

Optical lever for KAGRA

Kazuhiro Agatsuma 2014/May/16

2014/May/16 GW monthly seminar at Tokyo 1

Contents

Optical lever (OpLev) development for KAGRA • What is the optical lever? • Review of OpLev in TAMA-SAS • Requirements of KAGRA • Selection of components • Layout • Performance test • Toward bKAGRA (cryo-condition)


What is the optical lever?


OpLev is an angular sensor used for • Local angular control of each mirror => to help lock acquisition of the interferometer • Monitoring drift motion


mirror

Position sensor

Light source

Principle


Archimedes "Give me a place to stand and I will move the earth!“ (http://www.buzzle.com/editorials/7-30-2004-57259.asp)

c.f. Lever

Optical lever

The mirror angle is increased by the distance of detection

mirror

Position sensor

Light source

Effort Load

Fulcrum The effort force is reduced by the distance of effort arm

L

OpLev looks easy! It looks a kind of LEGO. Does it need to research? => Yes.

Operation is easy but “suitable” setup is not obvious • to meet requirements (angular spectrum, drift, range) • easy maintenance (easily available) • reasonable price (10-20 OpLevs will be used) => suitable operation is achieved by accumulation of

experiences


mirror

Position sensor

Light source

♦To decide detail design of OpLev ♦Engineering test (prototype test)

Objectives

Review of OpLev in TAMA-SAS (Actual examples)


OpLev in TAMA-SAS


(2002, TAMA project report)

TAMA-SAS Position sensor: QPD (quadrant photo diode)

Sensitivity


There were some exercises to reach the left sensitivity ♦windshield ♦laser angle ♦cabling ♦etc…

Angular control by OpLev


(master thesis of K.A.)

Mirror angular motion was suppressed by OpLev control

KAGRA requirements


Requirement (spectrum)

(※) In this calculation, following parameters are assumed • Intensity noise of light source: RIN ~ 1e-7 /rtHz • Shot noise or dark noise： ~1e-11 m/rtHz (as the spot fluctuation)


by Michimura and Sekiguchi, JGW document (2012/4/27)

To lock IFO RMS: 0.1 urad (for cavity) RMS: 1 urad (for BS mirror)

Intensity noise, shot noise

Requirement

Requirement (drift, range)


To hold fringe for a day Drift: 10 urad/day

3 km

3 cm / 3km = 10 urad 10 urad/week? 10 urad/day? 10 urad/h? =>

3 cm

To be locked by referring to optical lever

remaining fringe

If commissioning phase is assumed

Drift

Range

To follow unstable condition Range: 1~10 mrad (according to VIRGO empirically)

If the local control is failure (turned off), the suspension system is fluctuated

Selection of components


Selection of components • Light source

– SLD (Super Luminescent Diode) – Center wave length: 670 nm – Thermal control is included – Power: 1 mW

• Collimator lens – Beam-spot size： it affects the linear range and sensitivity ⇒ 0.5 - 2 mm for PSD (c.f. 2mm in case of aLIGO)

• Photo receiver – PSD (Position Sensitive Detector) – Φ: 9 mm x 9 mm


Light source • Selected parameters

– Fiber coupled SLD： to avoid interference with a reflection from cladding – Center of wave length: 650-680 nm

• Visible is useful at the installation and commissioning phase • Tradeoff with lifetime (switchable to IR SLD, which has longer

lifetime) – Power: 1 mW (mirror reflection of 20-50% is assumed) – Thermal control (to avoid the mode hop) – PM fiber (to keep polarization because of high incident angle) – FC/APC (to avoid returning light)


Measured RIN and lifetime


This is much lower than the requirement

from LIGO document

Position sensor

PSD (Position Sensitive Detector) • Large linear dynamic range • Low speed response • Spot-size-independent sensitivity

QPD (Quadrant Photo Detector) • Small linear dynamic range • High speed response • Spot-size-dependent sensitivity


Performance has been checked for below candidates

(※ Spot size affects the linear dynamic range)

Calibration factors


Φ 2mm: linear range: 7mm Φ 0.5mm: linear range: 8mm

Sensitivity is independent on the spot size => easy adjustment Linear range of PSD (7mm) is much larger than that of QPD (0.5mm)

PSD: 7 mm (Φ 2 mm) QPD: 0.5-1.5 mm (Φ 2-6 mm)

X-Y stage PSD (Position Sensitive Detector) Φ: 9 mm x 9 mm

Launcher (collimator lens & mirror)

Selected launcher Spot size is 1-2 mm with the

working distance of 4m • Adjustable-focus lens (good match with PSD)


Initial steering mirror

Layout


Optical layout


mirror

Position sensor

Light source

mirror

mirror

Turning mirror

Length coupling

θ

Large θ causes length coupling

Layout design


BS

Option 1 (Mixed signal)

Option 2 (Broad range & length)

BS

Default

Input port

Turning port

Range: 8mm/(2*3m) = 1.3 mrad

lens

Range: 8mm/(2*1m) = 4 mrad

PSD

Chamber for Type-B SAS

Turning port


BS

Option 1 (Mixed signal or Length information)

Option 2 (Broad range & Length)

BS

lens

PSD PSDs have to be placed on the focal plane and image plane using sliders with about 10-um accuracy

Merit: Clear separation between longitudinal and angular motion Demerit: Extra optics (PSD, lens and their sliders) increase total cost.

Subtraction of angular signal (input port) from the mixed PSD output produces the longitudinal signal

Merit: Cheap and simple Demerit: Direct coupling between longitudinal and angular motion

Option 1.5 lens (Broad range or Length information) To adjust the focal or image plane by adding a lens

Option 1 Option 1.5 Option 2

Cost Low Middle High

Information Mix or Length Broad or Length Broad and Length

Layout design for Type B (preliminary)


Drown by Gianni Gennaro

Breadboard

Performance test


Comprehensive drift (NAOJ)


~1m

PSD

Thermometer Paper box was used as windshield

Comprehensive drift (NAOJ)


Fitting: Y = α + β*T + γ*(x – t) T: temperature, β: temperature response x: time, γ: constant drift t: time delay

X direction: β = 8 um/K, γ = 14 um/day => 2 urad/day Y direction: β = 50 um/K, γ = 170 um/day => 30 urad/day (3m OpLev length)

To hold fringe for a day Drift: 10 urad/day

Origin of this drift is in investigation


Comprehensive drift (Nikhef)

~1m

Windshield

PSD

SLD

Thermometer (Pt100)

Comprehensive drift (Nikhef)


Fitting function: Y = α + β×T + γ×(x – t) T: temperature, β: temperature response x: time, γ: constant drift t: time delay

X direction: β = 44 um/K, γ = 6.4 um/day => 0.7 urad (0.1K/day in Kamioka) + 1 urad(drift) = 1.7 urad/day (OpLev length of 3m is assumed)

Y direction: β = 600 um/K, γ = 13um/day => 10 urad (0.1K/day) + 2 urad(drift) = 12 urad/day (OpLev length of 3m is assumed)

These results almost achieved the requirement of KAGRA (10 urad/day)

Temperature control in lab.

25 min.

Comparison with NAOJ result


NAOJ (talk at 2nd ELiTES meeting)

Condition is better than NAOJ ♦all mirror use thread-pitch locks ♦temperature is stable in the clean room ♦heavy optical table ♦metal windshield

X

Y

Nikhef

Spectrum of spot fluctuation


limited by digitization noise Thermal drift is seen below 0.1 Hz

worse than X direction => mechanical vibrations

Optical lever length of 3 m is assumed => divided by factor of 6 => ~ 1 nrad/rtHz at 1 Hz

Intensity noise by offset


Consistent with RIN measurement within difference factor of 2 (The left response has two times smaller RIN than the right graph shows) => an individual difference of SLD

(Blue line)

Investigation of large thermal response


?

Short length setup ⇒ Scaling factor only? ⇒ Scaling factor + bending?

The temperature response of Y direction is relatively worse: (X: β = 44 um/K, Y: β = 600 um/K)

Table bending?

Investigation of large thermal response


(2.5 + 98) + 104 + 104 + 111 + 29 cm = (100.5) + 348 cm, d1 = 111 cm Optical path length: 449 cm OpLev length: 348 cm Return mirror OpLev: 244 cm

(2.5 + 12) + 29 + 29 + 27 + 29 cm = (14.5) + 114 cm, d2 = 29 cm Optical path length: 129 cm OpLev length: 114 cm Return mirror OpLev: 85 cm

Lspot = 2x(4.5xθ1 + 3.5xθ2 + 2.4xθ3) Sspot = 2x(1.3xθ1 + 1.1xθ2 + 0.9xθ3)

d1 d2

Lspot /Sspot = ~3: Scaling factor (θ1=θ2=θ3)

d1/d2 = 3.8 : Table deformation factor

Which is the real factor? 3 (dominated by local mirrors) or 3x3.8=11 (dominated by table deformation)

θ1 θ2

θ3

Short length measurement


Short: β = 44 um/K, γ = 0.6 um/day (Long: β = 44 um/K, γ = 6.4 um/day)

Fitting function: Y = α + β×T + γ×(x – t) T: temperature, β: temperature response x: time, γ: constant drift t: time delay

Short: β = 45 um/K, γ = 1.9 um/day (Long: β = 600 um/K, γ = 13 um/day)

x13 x7 x10

Y temp: Table bending dominant Y drift: Table bending dominant (almost)

x1

X Y

X temp: Translational motion of SLD or PSD X drift: Table bending dominant

Expected drift (scaling factor: x3 for 3m length, 0.1 K/day) X: β = 44 um/K, γ = 2 um/day => ~1 urad/day Y: β = 150 um/K, γ = 6 um/day => ~3 urad/day

Next step


Pylon

Toward bKAGRA (perspective for cryogenic system)


Layout


test mass chamber for iKAGRA

test mass chamber for bKAGRA

~15 m

Layout for iKAGRA


~ 1 m

OpLev length: 3m Optical path length: 4m

Pylon

Range: 8mm/(2*3m) = 1.3 mrad

Input port

Turning port

Layout for bKAGRA


~ 15 m

OpLev length: 15 m Optical path length: 30 m

aLIGO

Pylon

Range: 8mm/(2*15m) ~ 0.3 mrad !

Detection port

Design of collimator lens will be needed

Observable range becomes small (1.3 mrad => 0.3 mrad)

OpLev inside cryostat


Merit inside cryostat ♦Observing Intermediate mass ♦Short length OpLev (broad range)

Technical misc. ・No optical window ・Fiber coupled SLD is useful ・Cryo-compatible PD What should be the reference? - Cryostat => deformation by changing temperature - Making special pylon (super invar?) => difficult…

Summary Selection and prototype test for the OpLev of iKAGRA Selection Laser source: Superlum SLD (670 nm) with PM fiber, Position sensor: PSD (9mm x 9mm), Collimator lens: Adjustable-focus lens is selected for PSD (Φ2 mm at 2-m WD,

Φ1 mm at 4-m WD) Performance tests • The comprehensive drifts are X:1 urad/day, Y: 3 urad/day • The angular spectrum is 1 nrad/rtHz at 1 Hz on an optical bench => Optical components are expected to achieve requirements Layout • Current range (PSD, 8mm range, and input port) is 1.3 mrad for the angular motion

of mirrors. It is useful to detect the signal at not only input port but also the turning port. (It is said that 10 mrad range is sufficient for SA in VIRGO)

Perspectives • Long-length OpLev have a high sensitivity but small range • Additional OpLev inside cryostat has some merits (broad range and seeing

intermediate mass)


Acknowledgements

• Riccardo DeSalvo • Ettore Majorana • Ryutaro Takahashi • Tomotada Akutsu • Takanori Sekiguchi all member of VIS, AOS, and QND group.


optical lever for kagra - gwdoc.icrr.u-tokyo.ac.jp

Documents