Periodic Reporting for period 1 - QNOIWA (Quantum NOIse evading measurement for gravitational WAve detectors)
Okres sprawozdawczy: 2023-03-01 do 2025-02-28
The detection of gravitation waves (GW) produced by astrophysical sources was predicted by Albert Einstein in 1916 in his theory of general relativity. Such measurement was long thought impossible because of the high strain sensitivity it requires in the typical 10Hz–10kHz spectral range. The typical required relative sensitivity is some 10^(-20)/√Hz, as calculated for the detection of stellar explosion using earth-based detectors. Ever since the pioneer work of Joseph Weber in the late 1950’s, who claimed to have detected GW using high Q-factor resonant bar detectors, there has been an increasing number of similar projects such as the ALLEGRO at Baton Rouge, USA, AURIGA at Legnaro, Italy or EXPLORER at CERN in Switzerland, which are running examples of this technology. Those detectors nevertheless are too narrowband because of their typical sub-100Hz and sharp resonance frequency. Highly sensitive and broadband detectors have emerged with the construction of optical-based interferometric detectors as first proposed by Gerstenshtein and Putovoit in 1963. Since the 70’s, multiple small-scale prototype benches such as the 30m-long Garching interferometer, the 10m one in Glasgow or the 40m one at Caltech have flourished, but GW detection requires interferometric arm length on the order of a few km long. To this end, the 4km-long LIGO US project was founded in 1992 and set up in Hanford and Livingston, the 3km-long VIRGO French-Italian project started in 1993 and set up in Pisa, and the KAGRA project started in 1995 in Tokyo. The GEO project was founded in 1989 through a collaboration between the Glasgow and Garching groups. With its 600m long arm, the GEO 600 interferometer serves as a test bench to improved detection technologies and its sensitivity compares to that of the larger LIGO generation of detectors. Since 2008, improvement of both VIRGO and LIGO strain sensitivity throughout a large bandwidth has skyrocketed. The new generation of Advanced VIRGO (aVIRGO) and Advanced LIGO (aLIGO) have allowed to reach an unprecedented strain level of detection on the order of 10^(-23)/√Hz. Those major upgrades over the original interferometers allowed, for the first time, the simultaneous detection of gravitational waves for a collision of black holes on September 14, 2015.
After all external sources of technical noise are reduced, the key to a further improved sensitivity is the reduction of the remaining and dominating quantum noise of the probe laser. For instance, aLIGO, aVIRGO and GEO600 are today limited by quantum noise throughout most of their detection bandwidth. It is this specific noise source that we address in this QNOIWA project. Quantum shot noise prevails at high frequencies (>100Hz), while an interplay between shot noise and radiation pressure noise has to be accounted for at lower frequencies (30-100Hz). The latter effect first pointed out by Braginsky in the 1960s [13] is known as the Quantum Back Action (QBA). It can be understood as follows: an ideal detection implies that the interferometer end-mirror is only put in motion by a gravitational wave. In such case, the detected optical signal is shot-noise limited, and the signal-to-noise ratio (SNR) increases as the square root of laser intensity. Following this logic, we could in principle reach unlimited detection sensitivity simply by increasing the laser power. In practice however, light acts on the end mirror through the radiation pressure force, a random force due to the Poissonian statistics of photons which causes unwanted random motion. This QBA effect leads to an increase in detection noise. Because of the very low resonance frequency of the end mirror (typically around 1Hz), the effect is mostly visible bellow~100Hz, and decreases quadratically with the GW detection frequency. The objective of this QNOIWA project was to experimentally beat quantum noise, i.e. demonstrate the simultaneous reduction of QBA in the lower part of the detection bandwidth and the reduction of shot noise in the upper part in the kHz frequency range. Our experiment have been successful, and will certainly pave the road to the pending objective of a future integration in large scale GW detectors, in lower frequency range.
After all external sources of technical noise are reduced, the key to a further improved sensitivity is the reduction of the remaining and dominating quantum noise of the probe laser. For instance, aLIGO, aVIRGO and GEO600 are today limited by quantum noise throughout most of their detection bandwidth. It is this specific noise source that we address in this QNOIWA project. Quantum shot noise prevails at high frequencies (>100Hz), while an interplay between shot noise and radiation pressure noise has to be accounted for at lower frequencies (30-100Hz). The latter effect first pointed out by Braginsky in the 1960s [13] is known as the Quantum Back Action (QBA). It can be understood as follows: an ideal detection implies that the interferometer end-mirror is only put in motion by a gravitational wave. In such case, the detected optical signal is shot-noise limited, and the signal-to-noise ratio (SNR) increases as the square root of laser intensity. Following this logic, we could in principle reach unlimited detection sensitivity simply by increasing the laser power. In practice however, light acts on the end mirror through the radiation pressure force, a random force due to the Poissonian statistics of photons which causes unwanted random motion. This QBA effect leads to an increase in detection noise. Because of the very low resonance frequency of the end mirror (typically around 1Hz), the effect is mostly visible bellow~100Hz, and decreases quadratically with the GW detection frequency. The objective of this QNOIWA project was to experimentally beat quantum noise, i.e. demonstrate the simultaneous reduction of QBA in the lower part of the detection bandwidth and the reduction of shot noise in the upper part in the kHz frequency range. Our experiment have been successful, and will certainly pave the road to the pending objective of a future integration in large scale GW detectors, in lower frequency range.
The solution proposed in this QNOIWA project utilizes another physical system, an atomic spin ensemble, to obtain the broadband quantum noise reduction. Crucially, it does not require modifications of current GW detectors, which lowers the cost and alleviates the risk of practical implementation. This ensemble behaves as an oscillator with an “effective negative mass", which is the basis for QBA reduction. One laser source addresses the GW detector at λ_I=1064nm, a typical wavelength used today in all GW detectors, while simultaneously, another laser source centered at λ_S=852nm addresses an atomic spin ensemble placed in a magnetic field. An entangled two-mode squeezed (Einstein-Podolsky-Rosen) state at these two respective wavelengths is generated using the sum frequency generation (SFG) and parametric down conversion (PDC). The resulting quantum light modes are injected into the respective dark ports of the two oscillators, leading to macroscopic entanglement of the GWD and the spin ensemble response. One of the first technical challenge of this experiment was to generate the EPR source. A version of it had already been demonstrated before the beginning of this project, but EPR entanglement was only >10kHz. The first 6 months of this project, I provided support to the laser team dedicated to extending the EPR frequency bandwith, while training on how the operate the atomic spin ensemble. In greater details this worked involved :
- Characterization of the spin preparation state using the MORS technique
- Characterization of the cell containing the atom transmission, and optimization of the top-hat beam
- Generation of ponderomotive squeezing and optimization of the interaction
- Replacement of the pump/probe laser beams and reorganization of the optical setup to improve the stability, and reduce the alignment time
- Motorization of the "balancing" of the photodetector and scanning of the homodyne phase
- Noise hunting to reduce coupling into the atomic spin ensemble ( laser noise, electronic noise and inducting coupling )
When the EPR source was finally ready, we proceed in the injection of the non-classical probe, this work involved:
- replacement of the homodyne detector to ensure the detection was compatible with the simultaneous detection of the beat-note used to lock the EPR phase
- Characterization and optimization of squeezing in absence of atomic response ( Atoms are moved to high frequency, out of the detection bandwidth)
The objective was then to "switch on" the atoms, and aim to an active response in the [10 kHz-250 kHz]. Since we do not have a GW detector in the laboratory for obvious reasons, we will simply emulate the effect of a GW detector using different values for the homodyne detection phase. This work required much resilience, and an upgrade of some of our control electronics. Our first observation of what seems to be frequency dependent stearing of the quantum state by the atoms was performed at 50kHz ( larmor frequency ). We realized the importance of the probe classical noise, so implemented an active "noise eater" before the injection to minimize its influence. Finally, we could clearly observe the frequency dependent strearing of our squeezed state, and even reproduce our result as low as 10kHz.
One of the final goal of the project was the demonstration of virtual rigidity. Indeed, we recall that our final agenda is to improve detection in the [10Hz – 100Hz] most interesting and challenging for GW detection. In order to lower the Larmor frequency in order to match χ_S with that of a GW free mass, one possibility suggested by the group several years ago is to change the detection phase of the state measured by the homodyne detector.
To do so, we tried rotation the quarter waveplate position after the spin ensemble, and indeed observe a down-shift of the effective larmor frequency. The rotation induced rotation feature is also affect by the angle of the waveplate, making the atomic ensemble a tunable noise eater for GW detectors, or any broadband sensor limited by backaction.
- Characterization of the spin preparation state using the MORS technique
- Characterization of the cell containing the atom transmission, and optimization of the top-hat beam
- Generation of ponderomotive squeezing and optimization of the interaction
- Replacement of the pump/probe laser beams and reorganization of the optical setup to improve the stability, and reduce the alignment time
- Motorization of the "balancing" of the photodetector and scanning of the homodyne phase
- Noise hunting to reduce coupling into the atomic spin ensemble ( laser noise, electronic noise and inducting coupling )
When the EPR source was finally ready, we proceed in the injection of the non-classical probe, this work involved:
- replacement of the homodyne detector to ensure the detection was compatible with the simultaneous detection of the beat-note used to lock the EPR phase
- Characterization and optimization of squeezing in absence of atomic response ( Atoms are moved to high frequency, out of the detection bandwidth)
The objective was then to "switch on" the atoms, and aim to an active response in the [10 kHz-250 kHz]. Since we do not have a GW detector in the laboratory for obvious reasons, we will simply emulate the effect of a GW detector using different values for the homodyne detection phase. This work required much resilience, and an upgrade of some of our control electronics. Our first observation of what seems to be frequency dependent stearing of the quantum state by the atoms was performed at 50kHz ( larmor frequency ). We realized the importance of the probe classical noise, so implemented an active "noise eater" before the injection to minimize its influence. Finally, we could clearly observe the frequency dependent strearing of our squeezed state, and even reproduce our result as low as 10kHz.
One of the final goal of the project was the demonstration of virtual rigidity. Indeed, we recall that our final agenda is to improve detection in the [10Hz – 100Hz] most interesting and challenging for GW detection. In order to lower the Larmor frequency in order to match χ_S with that of a GW free mass, one possibility suggested by the group several years ago is to change the detection phase of the state measured by the homodyne detector.
To do so, we tried rotation the quarter waveplate position after the spin ensemble, and indeed observe a down-shift of the effective larmor frequency. The rotation induced rotation feature is also affect by the angle of the waveplate, making the atomic ensemble a tunable noise eater for GW detectors, or any broadband sensor limited by backaction.
Our work represents a significant advancement in the field due to several innovative aspects. Firstly, we successfully "activated" atomic responses in the [10 kHz-250 kHz] range, a crucial frequency band for gravitational wave (GW) detection. Although we lack a GW detector in our laboratory, we ingeniously emulated its effects by varying the homodyne detection phase, demonstrating our system's adaptability. This is the first demonstration of such achievement ever report using "negative mass" auxiliary system, when will certainly be published the higher impact factor journal there is. Work still to be conducted to lower the frequency down to 10-100Hz, but most of the challenges are technical. We have provide the first effective "proof-of-principle" that the idea, publish theoretically more than 10 years ago by E. Polzik and F. Ya. Khalili works.
One of our key breakthroughs in that respect was the first observation of frequency-dependent steering of the quantum state by the atoms at the Larmor frequency of 50 kHz. This required not only resilience but also the upgrade of our control electronics. Recognizing the impact of probe classical noise, we implemented an active "noise eater," significantly reducing noise influence and allowing us to observe and reproduce this steering effect even at frequencies as low as 10 kHz.
Moreover, our project aims to demonstrate virtual rigidity, a concept critical for improving GW detection in the challenging [10Hz – 100Hz] range. By adjusting the detection phase using rotation of the quarter waveplate, we successfully downshifted the effective Larmor frequency, thereby aligning the atomic response with that of a GW free mass. This tunable noise-eating capability makes our atomic ensemble an invaluable tool for GW detectors or any broadband sensors that would require a tailored manipulation of the quantum state used for noise evasion.
In summary, our work not only introduces novel methods for frequency-dependent quantum state manipulation but also enhances the potential for more sensitive GW detection, making it state-of-the-art in its contribution to the field.
One of our key breakthroughs in that respect was the first observation of frequency-dependent steering of the quantum state by the atoms at the Larmor frequency of 50 kHz. This required not only resilience but also the upgrade of our control electronics. Recognizing the impact of probe classical noise, we implemented an active "noise eater," significantly reducing noise influence and allowing us to observe and reproduce this steering effect even at frequencies as low as 10 kHz.
Moreover, our project aims to demonstrate virtual rigidity, a concept critical for improving GW detection in the challenging [10Hz – 100Hz] range. By adjusting the detection phase using rotation of the quarter waveplate, we successfully downshifted the effective Larmor frequency, thereby aligning the atomic response with that of a GW free mass. This tunable noise-eating capability makes our atomic ensemble an invaluable tool for GW detectors or any broadband sensors that would require a tailored manipulation of the quantum state used for noise evasion.
In summary, our work not only introduces novel methods for frequency-dependent quantum state manipulation but also enhances the potential for more sensitive GW detection, making it state-of-the-art in its contribution to the field.