Stefan Hild, Andreas Freise, Simon Chelkowski University of Birmingham GWADW, ELBA, May 2008 Virtual Interferometry for future GW detectors.

Stefan Hild, Andreas Freise, Simon Chelkowski

University of Birmingham

GWADW, ELBA, May 2008

Virtual Interferometry for future GW detectors

Stefan Hild ELBA GWADW, May 2008 Slide 2

Virtual interferometry (the Idea) Inspiration from time-delay interferometry

3rd generation detectors are likely to consist of several individual instruments (Triangle …)

Optical and/or electronical combination of the several outputs of the individual instruments might allow to do nice things: Null streams ? Displacement noise free interferometry ? Frequency noise rejection ? … many more … (hopefully ?)

This is a huge multi-dimensional playground.


Virtual interferometry (the Idea) Inspiration from time-delay interferometry

3rd generation detectors are like to consists of several individual instruments (Triangle…)

Optical and/or electronical combination of the several outputs of the individual instruments might allow to do nice things: Null streams ? Displacement noise free interferometry ? Frequency noise rejection ? … many more … (hopefully ?)

This is a huge multi-dimensional playground.


The 3rd Generation Holy Grail:Displacement noise free interferometry (DNFI)

If you get DNFI to work: you can reduce many limiting noise sources.

Example: 2nd Generation noise limits


Starting point: DNFI a la Tarabrin:LIGO-P070109-00Z arXiv:0804.3955v1

Seems to be an interesting concept since it uses a standard Fabry-Perot cavity.

Modified version of the paper on the arXiv (end of April 2008)


DNFI ala Tarabrin:

Using a double pumped (2 Laser of different polarisation) detuned Fabry Perot cavity

Read out at both ends of the cavity by 8 photo diodes (4 homodyne detectors)

Assumption: All auxiliary optics sit on isolated platforms (no relativ movement of the components on the platforms)

Setup shown in the paper:

arX

iv:0

804

.3955v1


DNFI ala Tarabrin:

Experimentalapproch of the setup:

Setup shown in the paper:

arX

iv:0

80

4.3

95

5v1


Intuitive understanding of DNFI ala Tarabrin (1)

Using simple Finesse simulations

GW and mirror displacement can be distinguished in the signals

Disp. M_a Disp. M_b GW

S1 - scaled(S2) gives a GW channel ‘free’ of M_a displacement

In phase

Out of phase


What can we learn?Simple picture of a detuned cavity:

From the input side GW and end mirror displacement look identical

However, displacement of the input mirror looks different than GW (due to direct reflection)


Intuitive understanding of DNFI ala Tarabrin (2)

Disp. M_a Disp. M_b GW

S3 - scaled(S4) gives a GW channel ‘free’ of M_b displacement In phase

Out of phase

Now using the signals from LASER-B.


Creation DNFI channel for Tarabrin setup

Building a linear combination it is possible to create DNFI channel. We can suppress mirror displacement for frequencies below the

detuning of the cavity.

DNFI-channel = k1*S1 + k2*S2 + k3*S3 + k4*S4


Problems of the Tarabrin Concept

Detuning reduces power buildup inside the cavity

Frequency noise (no common mode rejection)

Displacement noise of optics on the Platforms (Homodyne detectors)

arX

iv:0

804

.3955v1


Going forward with the Tarabrin Concept ??

Detuning reduces power buildup inside the cavity

Frequency noise (no common mode rejection)

Displacement noise of optics on the Platforms (Homodyne detectors)

Perhaps one can use more powerful lasers or larger power recycling factors ?

Combine 2 cavities to form a Michelson interferometer

Replace the homodyne detectors by the conventional beam splitter of the Michelson interferometer


Our Playground … Single Michelson IFO with

douple pumped arm cavities.

3 Laser: all slightly different frequency (few GHz).

4 Photo detectors. PD1 only sees Laser1 PD2 only sees Laser2 PD3 only sees Laser3 PD123 sees an optical mix of all

lasers

2 Output mode cleaners.

Arm length of 3km.

Cavity detuning of a few kHz.


Signal Transfer functions

Using a linear combination of PD1, PD2 and PD3 we can remove displacement of IX and IY. (as expected)

But we have no chance to remove EY and EX. (as expected)

To remove EX and EY we probably need sensing at the end of the arms (in reflection of EY and EX).


Extending our Playground … A triangle of 3 Michelson IFOs

with arm cavities.

Each test mass is part of two Michelson IFOs.

3 Laser: all slightly different frequency (few GHz).

12 Photo detectors.

6 Output mode cleaners.

Arm length of 3km.

Cavity detuning of a few kHz


Signal TFs of single Michelson

Now we can subtract IX, IY, EX and EY. One Michelson (made of X and Y arm) ‘displacement

noise free’ !!


Displacement noise free Michelson interferometer

One Michelson (made of X and Y arm) ‘displacement noise free’ !!

DNFI-channel = k1*PD1 + k2*PD2 + k3*PD3 + k4*PDb2 + k5*PDc3

BUT this result is cheating… (for two reasons)

Displacement noise suppressioninside the detection (audio) band


How we were cheating…

If any displacement or GW signal is present in the Z-arm this will couple into PDb2 and PDc3.

PD2 and PD3 will be contaminated by laser frequency noise (no common mode rejection)

DNFI-channel = k1*PD1 + k2*PD2 + k3*PD3 + k4*PDb2 + k5*PDc3


ET design study is a good frame to do this …


Summary Virtual interferometry might be highly beneficial for 3rd

Generation GW detectors

Holy Grail = Realize Displacement noise free interferometry that combines: Noise suppression at (sub)audio band frequencies Keep the common mode rejection of frequency noise Using feasible geometries

The Tarabrin-Concept might be useful.

We presented some first and very preliminary analyses, but this needs to be continued.


E N D

Stefan Hild, Andreas Freise, Simon Chelkowski University of Birmingham GWADW, ELBA, May 2008 Virtual Interferometry for future GW detectors.

Documents

phase slide

3955v1 slide

stefan hild elba gwadw

generation noise limits

direct reflection slide

gw channel free

virtual interferometry

b displacement