Construction techniques and commissioning of the Three-Backlink Experiment for the LISA mission

TL;DR

The paper presents the design, construction, and initial commissioning of the Three-Backlink Experiment (3BL) to study non-reciprocal phase noise in LISA-like laser interferometry. It compares direct fiber, frequency-separated fiber, and free-beam Backlinks using ultra-stable quasi-monolithic optical benches and a Calibrated Quadrant Photodiode Singleton (CQS) for micron- and microradian-scale beam alignment. The study demonstrates through initial measurements an upper limit of about $15\ \mathrm{pm}/\sqrt{\mathrm{Hz}}$ on non-reciprocal Backlink noise in the LISA band, identifies dominant noise sources (e.g., backscatter, laser frequency noise, and environmental temperature), and validates the construction and commissioning of the benches as a high-precision metrology testbed. The work provides a foundation for future, more sensitive Backlink tests and informs the design of space-based laser interferometry testbeds for gravitational-wave and geodesy missions, with ongoing efforts to reach the $1\ \mathrm{pm}/\sqrt{\mathrm{Hz}}$ level (Ho-Zhang2025).

Abstract

Designed to detect gravitational waves in the lower-frequency band, the space mission LISA will open a new window to astronomy after its launch in the 2030s. Each LISA spacecraft houses two optical benches that require the exchange of a phase reference between them via an optical connection, called a Backlink. Here we present the construction and commissioning of an ultra-stable quasi-monolithic optical testbed to investigate different Backlink implementations: a direct fiber, a frequency-separated fiber, and a free-beam link, compared in the Three-Backlink Experiment. Dedicated alignment techniques crucial for the construction of these optical benches are presented together with the development of a high-precision beam alignment and measurement tool - a Calibrated Quadrant Photodiode Singleton. An upper limit for the performance of all three investigated Backlink schemes, as determined by initial experiments, can be set at a $15\text{pm}/\sqrt{\text{Hz}}$-equivalent level within the LISA band, spanning 0.1mHz to 1Hz. Our measurements were able to verify the successful construction and commissioning of this very complex interferometer as an interferometric laboratory testbed for LISA. We find no limitations due to the construction on the here reported performance levels. Our results can support the construction of high-precision metrology testbeds for space-based laser interferometry for future gravitational wave or geodesy missions.

Figures (7)

• Figure 1: The 3BL consists of two optical benches with a mirrored layout of the optical components, except for one additional half-wave plate on the right bench. On the left side, the bench is shown with the focus on the beam paths, while the right side presents the optical design with labeled components: bs: beamsplitter, m: mirror, pbs: polarizing beamsplitter, hwp: half-waveplate, att: attenuator, FIOS: fiber injector optical subassembly. Some optics are not labeled for simplicity. This figure is based on IsleifPhD, originally created from IfoCAD simulations Kochkina2013IfoCAD and with a swap of the left and right bench.
• Figure 2: A photograph of a CQS is shown in a). The housing, measured by the CMM, is made of brass and mounted on two micrometer-controlled translation stages. The following two pictures illustrate the working principle of the CQS. b) A calibration procedure provides the distances between the housing to the center of the QPD. The third, z-direction, is not shown for simplicity. c) The beam is centered by using differential power sensing on a QPD, and by measuring the outside of the mount, here highlighted in blue, the position of the beam is revealed.
• Figure 3: Step-by-step illustration of the construction example. Note that the optical bench layout has been significantly simplified compared to the actual design in figure \ref{['fig:3BLoverview']}. a) The first alignment beam is created by commercial components and aligned according to CQS measurements. b) A second alignment beam is created by back-coupling into the fiber coupler. c) The two bs are placed and aligned in order to optimize the contrast. d) The mirror is removed, and by CQS measurements, the height at the FIOS position is estimated. e) A pair of CQS measurements is performed on each alignment beam. f) One of the FIOS is placed and aligned according to the CQS measurements. This figure is based on a similar figure in BischofPhD.
• Figure 4: Pictures of the finalized optical benches that form the 3BL experiment. All components were placed according to the design, with only minor deviations in their absolute positions. The pictures were first printed in BischofPhD
• Figure 5: The 3BL optical benches are installed inside the vacuum chamber. Photodiodes are installed at the front and the back, and the FBBL-steering mirrors are aligned at the back. Parts of the thermal shield have been removed for visibility.