Tilt-to-length coupling metrology in the LISA mission

This paper describes a setup aimed at measuring the so-called Tilt-To-Length (TTL) coupling in the optical benches of the LISA mission. The TTL is the coupling of the angular jitter of any optical setup into the optical path length between its input and output pupils. This might be deleterious in laser ranging experiments and must be evaluated for further compensation. The setup is made of two laser beams, one features an angular jitter that mimics the input beam as seen from the jittering bench under test (BUT), the other is aligned to the optical axis of the BUT and provides a phase reference for the jittering beam. The induced phase variations between both beams detected at the BUT’s output pupil gives access to the TTL coupling. The ‘‘TTL probe’’ must feature a negligible residual TTL coupling which implies a micrometric accuracy in the centering of the setup pupil, the beams and the angular jitter associated pivot point. The setup integrates optical masks as a link between the setup optical reference frame to its mechanical reference frame, together with position memories and servo-loops for the beam’s alignment. We show that the stability, the accuracy, and the noise floor of the setup is compliant with the LISA specifications for the TTL mitigation, although it makes use of off-the-shelf components and is operated in a standard environment laboratory.

💡 Research Summary

The paper presents a dedicated laboratory demonstrator, named the “FOGOB” (Tilt‑to‑Length Probe), designed to quantify the Tilt‑to‑Length (TTL) coupling that occurs in the optical benches of the Laser Interferometer Space Antenna (LISA) mission. TTL coupling is the conversion of angular jitter of a bench (or the spacecraft) into an apparent change of the optical path length (OPL) between the input and output pupils. In LISA, where interferometric measurements must reach sensitivities better than 10 pm/√Hz over the 0.1 mHz–1 Hz band, even a tiny TTL effect can mask the gravitational‑wave signal, so it must be measured and compensated.

The FOGOB concept uses two laser beams at 1064 nm: an “Rx” beam that mimics the incoming inter‑spacecraft link and is deliberately given a sinusoidal angular jitter (typically a few tens of hertz), and a “LO” beam that is aligned with the optical axis of the bench under test (BUT) and serves as a phase reference. The two beams are phase‑locked with a 1 MHz heterodyne offset, producing a beat signal that is detected on a four‑quadrant photodiode (QPD). After demodulation at 1 MHz, the phase φ of the beat gives the OPL change ΔL = φ λ/(2π). By summing the phase contributions of the four quadrants and normalising by the known Rx tilt, the TTL coupling coefficient (ΔL/θ) is obtained. The differential wavefront signal (DWS), i.e., the phase difference between opposite quadrants, provides a direct measurement of the instantaneous Rx‑LO tilt.

A central requirement is that the probe itself introduces negligible TTL. This is achieved through a three‑step mask‑based alignment procedure. Three quartz masks (QPD‑like, Pinhole, and Pupil) are sequentially placed at the same mechanical location. The QPD‑like mask contains a four‑hole pattern that is used to centre both beams on a common optical reference. The Pinhole mask (150 µm aperture) defines the phase‑lock region, allowing the Rx and LO beams to be locked at the same point. Finally, the Pupil mask (2.55 mm diameter) defines the actual FOGOB pupil that will be conjugated to the LISA optical bench pupil. Each mask carries four fiducial apertures; their positions are monitored by a set of auxiliary QPDs that serve as a “position memory”. The fiducial‑to‑mask alignment is measured with a resolution of ~30 nm, and long‑term drift stays below 0.4 µm over five days, ensuring sub‑micron repeatability when swapping masks.

Beam alignment is stabilised by closed‑loop servos. The LO beam is clamped to the position‑memory QPD via a piezo‑driven steering mirror, achieving sub‑micron positional stability. The Rx beam is controlled in four degrees of freedom (horizontal/vertical shift and tilt) using a combination of the position‑memory QPD and the DWS error signal. The control matrix is diagonalised, allowing independent actuation of each degree of freedom. The servo bandwidth (≈0.1 Hz) is sufficient to maintain alignment over long periods while also permitting the injection of a pure sinusoidal tilt by adding a calibrated signal to the DWS error channel.

Performance measurements were carried out by imposing a 15 Hz sinusoidal angular jitter on the Rx beam. The TTL signal was extracted from the beat‑phase at the QPD and expressed in meters, then normalised by the known jitter amplitude to obtain the coupling coefficient in µm/rad. Over a 130‑hour run the measured TTL drift was <0.7 µm/rad, with a statistical uncertainty of <0.1 µm/rad for 60‑second averaging. When the QPD was perfectly centred, the residual TTL was ≈4 µm/rad. These figures satisfy the LISA requirements of ≤15 µm/rad accuracy and ≤1 µm/rad precision (3σ) without any recalibration over several days.

In summary, the authors demonstrate that a relatively low‑cost, off‑the‑shelf laboratory setup can meet the stringent TTL metrology specifications required for LISA. The key innovations are the mask‑based alignment strategy that guarantees micrometric centring of beams, pivot point and pupil; the high‑resolution fiducial position‑memory system; and the dual‑loop servo architecture that stabilises both beam position and angular jitter. The reported long‑term stability, sub‑micron positioning accuracy, and residual TTL coupling well below the mission budget indicate that the FOGOB is a viable platform for characterising and eventually compensating TTL effects in the actual LISA optical benches. Future work will focus on integrating the demonstrator with a flight‑representative optical bench, testing under thermal‑vacuum and vibration conditions, and automating the calibration sequence for on‑orbit use.