We report the development of a space-compatible packaging system for an integrated monolithic ultra-stable optical reference toward China’s next-generation geodesy mission with low orbit satellite-to-satellite tracking. Building on our previous work, we optimized the mounting structure and thermal insulation mechanism using the finite element method. The comprehensive simulation results demonstrated the robustness of the entire packaging system with enough margins to withstand severe launch loads and maintain an ultra-high geometric cavity length stability. A long-term prediction of the vacuum maintenance around the cavity during in-orbit operation was conducted. An engineering prototype, within which an integrated monolithic optical reference has been mounted, was built based on our optimized design, and it has successfully passed typical aerospace environmental tests, including sinusoidal vibration (∼10 g, 10–100 Hz), random vibration (∼0.045 g2/Hz, 10–2000 Hz), and thermal cycling (0–45, 3 °C/min, lasting for 90 h). The experimental thermal time constant of the prototype exceeded 9.5 × 104 s, enabling a temperature stability of 1.1 × 10−6 K/Hz1/2 at 10 mHz on the optical cavity, with external active temperature control. The design is also suitable and useful for laboratory and terrestrial applications.

The demand for ultra-stable lasers is increasing in the context of space missions, such as the Laser Interferometer Space Antenna (LISA),1,2 Gravity Recovery and Climate Experiment Follow-On (GRACE-FO),3 Next Generation Gravity Mission (NGGM),4 TianQin,5 and Taiji.6 The next generation geodesy mission in China uses an inter-satellite laser ranging interferometer (LRI) for low orbit satellite-to-satellite tracking (SST) over distances exceeding 100 km. The precision of the displacement measurement is 20 nm/Hz1/2 at Fourier frequencies ranging from 10 to 100 mHz. Laser frequency instability is a significant noise source of the LRI, necessitating stringent requirements expressed as7 
(1)
where y = Δν/ν is the fractional frequency instability and f denotes the Fourier frequency. Laser frequency stabilization of the LRI can be achieved through the Pound–Drever–Hall (PDH) technique,8 which utilizes a high-finesse Fabry–Pérot (F–P) cavity as the frequency reference. The instability of the laser frequency is linked to the fractional optical path length instability of the cavity, which can be characterized by two components, as shown in the following equation:
(2)
where ∆n/n is the variation in the refractive index of the gas between the cavity mirrors and ∆L/L is the variation in the geometric cavity length. Here, factors such as the thermal expansion of the mirror coatings and variations in the coating index due to temperature can be considered negligible. To satisfy the requirements of Eq. (1), the fractional changes in the refractive index and geometric cavity length must sum to less than 10−13/Hz1/2 at Fourier frequencies ranging from 10 to 100 mHz.

Over the past few decades, extensive research has been conducted on ground-based ultra-stable optical cavities.9–18 However, despite their good performance on the ground, these cavities are not suitable for space applications due to their inability to withstand severe launch loads. A number of ultra-stable optical reference systems have been designed for current and future space applications,19–23 including GRACE-FO, NGGM, BOOST, and SOC2. These represent a significant advancement in the field. Typically, these prototypes are characterized by separate designs for optical cavities and the associated mode-matching optics. This may lead to an increased susceptibility of the optical coupling to space environmental factors and a decrease in the overall compactness of the packaging system.

In our previous work,7 an integrated monolithic optical reference designed for potential SST mission applications was introduced. This reference includes a 50 mm-long ULE (ultra-low expansion) optical cavity and the associated mode-matching optics, which have been integrated into a ULE optical bench (OB) with a size of 80 × 80 × 65 mm3. Locked to the cavity, an Nd:YAG laser with a wavelength of 1064 nm demonstrated a frequency noise of 30 Hz/Hz1/2 at Fourier frequencies from 10 mHz to 0.1 Hz in the laboratory, meeting the SST mission requirement. However, further research is required to make this promising integrated monolithic optical reference more compatible for space applications.

We developed a space-compatible packaging system based on our previous studies. This paper is organized as follows: Sec. II outlines the design principles of the packaging system, focusing on its robustness against severe launch loads and the maintenance of the ultra-high optical path length stability of the cavity during operation. Section III describes the optimization design, including a detailed examination of the support structure of the OB and the thermal insulation mechanism within the packaging system. Section IV presents a comprehensive performance analysis of the entire packaging system. Section V presents the results of aerospace environmental tests and a thermal time constant experiment conducted on an engineering prototype.

Space instruments must be structurally verified in order to confirm their ability to withstand severe launch loads. In this study, the packaging system was subjected to a series of Aerospace Environmental Tests (AETs) with the following typical conditions: sinusoidal vibration (10–100 Hz) with acceleration up to 10 g, random vibration (10–2000 Hz) with power spectral acceleration up to 0.045 g2/Hz and total rms acceleration up to 5.54 g, shock (100–4000 Hz) with acceleration up to 400 g, and thermal cycling ranging from 0 to 45 °C (at a rate of 3 °C/min and a dwell time of 4 h). These conditions of AETs are referenced to the aerospace environmental test conditions of TianQin-1 satellite,24,25 a low earth orbit (with a height of 700 km) technology demonstration satellite for TianQin project.5 

It is crucial that the delicate monolithic optical reference and other components of the packaging system survive the AETs.

A number of basic principles must be considered when designing the mechanical reliability of a packaging system for space-borne optical reference. First, structural damping should be considered to mitigate mechanical resonance and to avoid excitation of the system under sinusoidal vibrations with significant acceleration amplitudes within the 10–100 Hz frequency range. Furthermore, in order to minimize the thermal-stress accumulation in materials during the thermal cycling, it would be advantageous to minimize the coefficient of thermal expansion (CTE) mismatch between the contact materials and to use materials with a low elastic modulus. The mechanical safety margin of a material is commonly evaluated using the safety factor ratio η=[σ]cσmax, where σmax and [σ]c represent the maximum and ultimate stresses of the material, respectively. In accordance with the ISO space standard,26 the core payload components should have a conservative safety factor under typical space conditions. In this study, if the ULE components and other materials demonstrate a safety factor exceeding 3 and 1.3, respectively, it can be concluded that the entire packaging system is mechanically reliable.

In the context of the SST mission, the satellites will be guided in a drag-free orbit, and the residual acceleration on board will be at a negligible level of 10−10 g/Hz1/2.27 Consequently, the temperature becomes a dominant perturbation affecting the geometric cavity length. The laser frequency noise budget due to temperature was determined to be 20 Hz/Hz1/2 at 10 mHz. The fractional frequency instability of the laser induced by the temperature fluctuation on the cavity is expressed as follows:
(3)
where δνT denotes the laser frequency fluctuation, ΔlT represents the corresponding geometric cavity length fluctuation, and δTcavity is the temperature fluctuation on the optical cavity. The coefficient of thermal expansion of the ULE material at the working temperature T, denoted by CTEuleT, can be expressed as follows:
(4)
where βule is the first-order temperature coefficient of CTEule and values ∼2.0 × 10−9/K2 within the temperature range of 5–35 °C, while Tzero,ule represents the zero-crossing temperature of the ULE material.28 Assuming that the discrepancy between T and Tzero,ule does not exceed 15 K. It is critical to maintain δTcavity below 2.2 × 10−6 K/Hz1/2 at 10 mHz so that δνT can be reduced to less than 20 Hz/Hz1/2 at 10 mHz. To achieve this temperature stability, a passive thermal shield design is our best solution with minimal power consumption, basing on a condition that the active temperature control outside the packaging system on satellite can reach a stability level of 1 mK.24 
The external packaging structures of the monolithic optical reference can act as an efficient temperature filtering system. The equivalent transient heat transfer function, Ht, is expressed as follows:
(5)
where δTcavityt is the temperature variation of the optical reference, δTout(t) represents the ambient temperature perturbation, and τ denotes the equivalent thermal time constant. The heat transfer function in the frequency domain can be approximated as H(f)1/1+f2/fc2, where fc represents the low-pass cutoff frequency and is equal to 1/τ. At high frequencies (ffc), the thermal attenuation coefficient from the vacuum chamber to the optical cavity can be approximated as follows:
(6)
where δTvacuumf denotes the temperature fluctuation at the outer surface of the vacuum chamber. Assuming δTvacuumf to be 1 mK/Hz1/2 at 10 mHz with active temperature control, in order to meet the requirement of δTcavity, the value of 1fτ must be less than 2.2 × 10−3 at 10 mHz, which requires τ to exceed 4.5 × 104 s.
The relationship between the refractive index and gas pressure has been elucidated by Edlén’s equations,29 which has been modified by experiments over the years.30 An experiment conducted by the Space Optical Clocks (SOC2) project team has demonstrated that the frequency instability induced by residual gas pressure in an optical cavity at a vacuum level of 10−4 Pa can be approximately calculated using the following equation:21,
(7)
where δνn is the frequency noise induced by the refractive index fluctuation δn and δP is the pressure fluctuation between the two mirrors at the end of the cavity. Here, the influence of gas composition on the refractive index has been neglected. In our study, we assigned δνn/ν as 3.5 × 10−15/Hz1/2 at 10 mHz. Assuming that Eq. (7) remains valid at a vacuum level of 10−5 Pa, it can be deduced that δP should not exceed 1.3 × 10−6 Pa/Hz1/2 at 10 mHz. Conversely, if the absolute residual gas pressure around the cavity exceeds 10−3 Pa, the heat transfer of the rarefied gases cannot be ignored. At spacecraft altitudes of ∼300–500 km, the gas pressure is typically 10−4–10−6 Pa,31 and the outgassing of other payloads in the spacecraft is likely to give rise to a higher local gas density. Consequently, it is imperative that the monolithic optical reference be enclosed within an ultra-high vacuum chamber. Under typical conditions, when the absolute pressure in the vacuum chamber Pchamber is maintained at the level of 10−5 Pa, the amplitude of long-term pressure variation δP is expected not to exceed 0.1 × Pchamber ≈ 10−6 Pa.21 Consequently, the residual gas pressure requirement inside our vacuum chamber will be regarded as lower than 10−5 Pa in this study. Materials of components within the vacuum chamber should ideally exhibit minimal outgassing rates.

Space radiation is one of the distinctive features of the space environment, which may affect the performance of optics components, particularly the cavity mirrors and optical fibers, as well as the mechanical properties of other materials. Despite the fact that satellite shells are usually designed to withstand space radiation, the packaging system of the optical reference should be able to furtherly shield these components from radiation damage. Generally, multi-layer metal structures with large total thickness will be advantageous for mitigating radiation exposure.

The design of a packaging system for space-borne optical reference involves a number of complex trade-offs. These include minimizing the thermal stress in materials, damping external vibration and shock, isolating ambient temperature noise, and maintaining a long-term ultra-high vacuum around the cavity during in-orbit operation.

Figure 1 shows a simplified diagram of the packaging system, which comprises a monolithic optical reference, support structure of the bench, thermal shield, insulation structure, vacuum chamber, and packaging box. The support structure and thermal insulation mechanism (comprising the thermal shield and insulation structure) play a pivotal role in maintaining the mechanical reliability and thermal stability of the optical reference. This is the primary focus of the following optimization.

FIG. 1.

Simplified diagram of the packaging system.

FIG. 1.

Simplified diagram of the packaging system.

Close modal

In our previous work,7 the monolithic optical bench was supported by a four-point mounting method with inserted Invar and Viton supporting structures in turns. This mounting method can sufficiently withstand vibration and shocks. However, further analysis shows that when this support structure is assembled to the outer vacuum packaging system, it becomes not sufficient enough to withstand thermal cycling and is also more sensitive to temperature fluctuations than we expected.

In this study, we modified the geometry of the mounting structure and designed four cylindrical cup-shaped flexible structures.

Figures 2(a) and 2(b), respectively, show an assembly model and an exploded model of the monolithic optical reference mounted by this support structure with a flexible design. In the assembly model, four identical flexible structures were inserted into the support holes located on the opposite side walls of the OB. The opening of each flexible structure was aligned with the protruding cylindrical head of the Invar-metal structure. The threaded holes located at the center of the cylindrical heads permitted the application of screw preloads to secure the flexible structures to the OB. The four metal structures were affixed to a square Invar plate measuring 131 mm in length and 5 mm in thickness.

FIG. 2.

(a) An assembly model of the integrated monolithic optical reference and the support structure with a flexible structure. (b) An exploded view of model (a). (c) A sectional view of the flexible structure.

FIG. 2.

(a) An assembly model of the integrated monolithic optical reference and the support structure with a flexible structure. (b) An exploded view of model (a). (c) A sectional view of the flexible structure.

Close modal

The flexible structure, shown in Fig. 2(c) with a cross-sectional view, was designed to have an outer diameter (the cylindrical part) of 10 mm and a height of 6 mm. The wall thickness was 1.5 mm. Due to their advantageous properties, including a low CTE, low Young’s modulus, low thermal conductivity, and a relatively low outgassing rate compared to other polymer materials, fluoropolymers such as Viton-A, Teflon, and PCTFE are preferred. The mechanical and thermal properties of the ULE and potential support structure materials are presented in Table I. Although the ULE optical bench features a mirror-polished top surface for bonding the optical cavity and coupling optics, along with supporting holes (each with a diameter of 10 mm and a depth of 5 mm) processed into its side walls, we anticipate that the strength limit of the ULE optical bench will remain consistent with its initial value of 49.8 MPa as provided by Corning Inc.28 

TABLE I.

Thermal and mechanical properties of ULE and the materials of the support structure.

ComponentMaterialCTE (ppm/K)Thermal conductivity (W/m K)Young’s modulusUltimate strength (MPa)
Optical reference ULE 0.03 1.31 67.6 GPa 49.8 
Metal support Invar 0.63 13.4 144 GPa 400 
Alternative material of the flexible structure Teflon 145 0.25 1.2 GPa 23 
Viton-A 230 0.25 7.8 MPa 8.8 
PCTFE 70 0.21 1.2 GPa 38 
ComponentMaterialCTE (ppm/K)Thermal conductivity (W/m K)Young’s modulusUltimate strength (MPa)
Optical reference ULE 0.03 1.31 67.6 GPa 49.8 
Metal support Invar 0.63 13.4 144 GPa 400 
Alternative material of the flexible structure Teflon 145 0.25 1.2 GPa 23 
Viton-A 230 0.25 7.8 MPa 8.8 
PCTFE 70 0.21 1.2 GPa 38 

Finite element analysis (FEA) simulations were performed on an assembly model consisting of a monolithic optical reference and a support structure encompassing the previous and current designs. The simulations included mechanical and thermo-mechanical analyses under typical space conditions, as well as heat transfer simulations (without considering thermal contact resistance). The results of the simulations are presented in Table II.

TABLE II.

Simulation result of models comprising a monolithic optical reference and a support structure.

Maximum von Mises stress of ULE
bench directly contacted with different
Conditionsmaterials of support structure (MPa)
InvarTeflonPCTFEViton-A
Sin vibration 1.9 0.4 0.4 0.6 
Random vibration 6.6 0.6 0.6 39.4 
Shock 47.9 20.5 20.5 8.4 
Thermal cycling 6.5 102.0 59.0 4.8 
Safety factor of ULE 1.03 0.49 0.84 1.26 
Temperature attenuation at 10 mHz 0.015 0.008 0.008 0.008 
Maximum von Mises stress of ULE
bench directly contacted with different
Conditionsmaterials of support structure (MPa)
InvarTeflonPCTFEViton-A
Sin vibration 1.9 0.4 0.4 0.6 
Random vibration 6.6 0.6 0.6 39.4 
Shock 47.9 20.5 20.5 8.4 
Thermal cycling 6.5 102.0 59.0 4.8 
Safety factor of ULE 1.03 0.49 0.84 1.26 
Temperature attenuation at 10 mHz 0.015 0.008 0.008 0.008 

The numerical simulations demonstrated that the monolithic optical reference, when mounted by the current support structure with a flexible design (regardless of Teflon, PCTFE, or Viton-A material), exhibited a lower maximum von Mises stress (see columns 3–5 of Table II) under both sinusoidal vibration and shock loads than the previous design (see column 2 of Table II). As shown in the last row of Table II, the flexible structures significantly improved the thermal isolation between the metal structure and optical cavity. However, Teflon and PCTFE generated excessive thermal stress in the OB during the thermal cycling simulations due to their large Young’s modulus. Conversely, Viton-A, with a Young’s modulus of 7.8 MPa, mitigated the thermal stress in the OB despite its high CTE.

It is worth noting that the mechanical resonance of Viton-A led to a high von Mises stress of ∼39.4 MPa on the OB in the random vibration simulation, as shown in Fig. 3(b). A modal shape map of the monolithic optical reference and support structure at the first-order natural frequency (289 Hz) contributed by Viton-A is shown in Fig. 3(a). Decreasing the wall thickness of Viton-A can increase its natural frequencies; however, this may compromise the thermal stability and thermo-mechanical reliability of the monolithic optical reference.

FIG. 3.

(a) Modal shape of the monolithic optical reference and support structure at 289 Hz. (b) Maximum von Mises stress of the monolithic optical reference in random vibration.

FIG. 3.

(a) Modal shape of the monolithic optical reference and support structure at 289 Hz. (b) Maximum von Mises stress of the monolithic optical reference in random vibration.

Close modal
The constraints imposed by the support structure may potentially alter the thermal properties of the ULE optical cavity, including the coefficient of thermal expansion and zero-crossing temperature. The effective coefficient of thermal expansion (CTEeff) of the optical cavity can be calculated by deriving the relative variation in the cavity length with respect to temperature, according to the following formula:
(8)
where LT and LT0,ule represent the geometric cavity length at the operating temperature (T) and the zero-crossing temperature (T0,ule) of ULE, respectively.
Through the FEA simulation of thermo-mechanical coupling, the variation in the geometric cavity length with temperature can be calculated. The theoretical CTEeff of the optical cavity can then be determined using Eq. (7). As shown in Fig. 4, the calculated CTEeff can be approximated as a linear function of temperature. This function is expressed as follows:
(9)
FIG. 4.

Effective CTE of the optical cavity varies with temperature.

FIG. 4.

Effective CTE of the optical cavity varies with temperature.

Close modal

This equation can be represented as CTEeffTβeffTT0,eff, where βeff is the linear coefficient of CTEeffT and remains the same value as the initial ULE, while T0,eff represents the effective zero-crossing temperature of the cavity and exhibits a slight shift of −0.55 K from T0,ule.

According to Fig. 5, the observed shift in T0,eff from T0,ule indicates a lack of sensitivity to the preload on the OB generated by Viton-A squeezed by the screws. The analysis of CTEeff suggests that the monolithic optical reference mounted by the support structure with a flexible structure entirely made of Viton-A maintains a CTE similar to that of the original ULE material. This represents a significant improvement over our previous mounting method, which resulted in a notable shift in the zero-crossing temperature (∼−4 K) of the ULE cavity.

FIG. 5.

Shift of zero-crossing temperature varies with the preload of screw.

FIG. 5.

Shift of zero-crossing temperature varies with the preload of screw.

Close modal

In conclusion, the support structure with a flexible Viton-A structure showed clear advantages in withstanding severe launch loads, attenuating thermal fluctuations, and inducing less interference to the CTE of the ULE cavity.

Thermal shields can function as heat sinks and reduce thermal radiation. Theoretically, increasing the number of layers and thickness of the thermal shields is beneficial for minimizing the effect of external temperature fluctuations on the optical cavity.32–34 However, this approach is not conducive to compact and lightweight packaging systems, and it may compromise the mechanical reliability of the components.

A single-layer thermal shield with an additional insulation structure was designed using FEA. As shown in Fig. 6(a), the designed thermal shield consists of a rectangular cover with a square opening at the bottom and a square metal plate [corresponding to the square Invar plate of the support structure shown in Fig. (2)]. The cover was made of beryllium bronze (QBe-1.9) owing to its high specific heat capacity, strength, and excellent surface machinability. It measured 130 mm in length and 80 mm in height and had a thickness of 4.5 mm on the top wall and 3.5 mm on the four side walls. To minimize the radiative heat transfer, the thermal shield and all other metal components within the vacuum chamber were mirror-polished and gold-coated.

FIG. 6.

(a) A cutaway view of the model with thermal insulation mechanism, monolithic optical reference, and support structure. (b) 3D geometric model of the insulation structure.

FIG. 6.

(a) A cutaway view of the model with thermal insulation mechanism, monolithic optical reference, and support structure. (b) 3D geometric model of the insulation structure.

Close modal

To support the thermal shield, an additional insulation structure is required to simultaneously fulfill multiple functions such as rigid connections, thermal isolation, and vibration damping. Furthermore, the mechanical reliability of the insulation structure itself should be considered.

We designed an insulation structure composed of a polymer material shaped as a square plate with symmetrically distributed holes, as shown in Fig. 6(b). The design of the holes aims to reduce thermal-stress accumulation during thermal cycling and minimize the cross-sectional area of heat conduction within the structure. Through FEA simulations, the optimal hole size (35 × 35 mm2 for the large holes and 10 × 10 mm2 for the small holes) and a suitable plate thickness (8 mm) were determined. In addition, small round platforms were reserved at the top of the plate to minimize the thermal contact area with the metal thermal shield. PCTFE was the preferred material for the insulation structure due to its favorable characteristics, including low thermal conductivity, low CTE, low outgassing rate, and higher ultimate strength.

FEA simulations similar to those presented in Table II were conducted on the entire model (non-sectional view), as shown in Fig. 6(a). The simulation results are outlined in Table III and show a notable enhancement in the safety factor for the ULE components compared to those documented in Table II. The temperature attenuation from the insulation structure to the optical reference was 0.002 at 10 mHz, which satisfied the thermal isolation requirement.

TABLE III.

Simulation results of the complete model in Fig. 6(a).

Maximum von Mises stress
Conditionsof fragile material/MPa
ULE benchPCTFE
Sin vibration 0.5 5.8 
Random vibration 2.4 6.1 
Shock 10.0 27.3 
Thermal cycling 2.4 15.4 
Safety factor 4.98 1.4 
Temperature attenuation at 10 mHz 0.002  
Maximum von Mises stress
Conditionsof fragile material/MPa
ULE benchPCTFE
Sin vibration 0.5 5.8 
Random vibration 2.4 6.1 
Shock 10.0 27.3 
Thermal cycling 2.4 15.4 
Safety factor 4.98 1.4 
Temperature attenuation at 10 mHz 0.002  

Figure 7 shows a sectional view model of the entire packaging system at an angle of 45° to the XOZ plane. The monolithic optical reference, support structure, and thermal insulation mechanism are enclosed in a vacuum chamber that is subsequently mounted in a packaging box.

FIG. 7.

An incident-cutaway view of the assembly of the packaging system.

FIG. 7.

An incident-cutaway view of the assembly of the packaging system.

Close modal

The vacuum chamber and packaging box were constructed using the Ti–6Al–4V (TC4) alloy due to its advantageous properties, including low density, high ultimate tensile strength, and relatively low CTE as well as low outgassing rate. The two metal structures were measuring a thinnest thickness of 4 and 3 mm, respectively.

Several groups of reticulated cylindrical metal shock absorbers were positioned between the packaging box and the vacuum chamber to provide primary vibration damping. The dimensions of the entire packaging system were 200 mm on each side and a height of 214 mm. In addition, a degassing unit comprising outlet pipes, an ion pump, and a getter assembly was installed at the top of the vacuum chamber. The total mass of the system was 13 kg (the mass of ion and getter pumps was included), with the majority of this attributed to the vacuum chamber and packaging box.

The performances of the entire packaging system were closely interconnected and mutually constrained in terms of mechanical reliability, thermal stability, and vacuum maintenance. In order to thoroughly evaluate the performance of the entire packaging system, a series of FEA simulations were conducted, including solid mechanics, thermo-mechanical coupling, and heat transfer. In addition, long-term vacuum maintenance was predicted based on the outgassing rate measurements.

The maximum von Mises stresses of the major components of the packaging system obtained from the simulations are summarized in Table IV along with their safety factors. The core material ULE exhibited a safety factor of 3.4, with no material having a safety factor of less than 1.5, thereby satisfying the mechanical reliability requirement. Furthermore, a modal analysis was performed, which indicated low-order natural frequencies of 189, 191, 209, and 222 Hz attributed to the shock absorber.

TABLE IV.

Results of FEA simulation for the entire system under aerospace environmental test conditions.

MaterialMaximum von Mises stress (MPa) in FEA
Sin vibrationRandom vibrationShockThermal cyclingSafety factor
ULE 1.1 14.6 7.6 1.9 3.4 
PCTFE 0.8 3.8 3.1 15.6 2.4 
Viton-A 0.4 5.3 3.0 0.3 1.5 
Invar 25.3 149.5 99.1 89.4 2.6 
QBe-1.9 3.0 20.5 23.4 125.0 8.0 
TC4 18.5 384.8 129.6 99.8 2.3 
MaterialMaximum von Mises stress (MPa) in FEA
Sin vibrationRandom vibrationShockThermal cyclingSafety factor
ULE 1.1 14.6 7.6 1.9 3.4 
PCTFE 0.8 3.8 3.1 15.6 2.4 
Viton-A 0.4 5.3 3.0 0.3 1.5 
Invar 25.3 149.5 99.1 89.4 2.6 
QBe-1.9 3.0 20.5 23.4 125.0 8.0 
TC4 18.5 384.8 129.6 99.8 2.3 

In the heat transfer simulation, which considered the thermal contact resistance, a step temperature excitation of 1 mK was applied to the bottom surface of the vacuum chamber. The simulated temperature variations on the cavity are shown in Fig. 8. The heat transfer function from the vacuum chamber to the optical cavity was obtained through curve fitting and can be expressed as δTcavitytδTout(t)=(1et/91164). This gives a thermal time constant of 9.1 × 104 s, which is twice the required value.

FIG. 8.

Heat transfer simulation under a 1 mK step temperature excitation.

FIG. 8.

Heat transfer simulation under a 1 mK step temperature excitation.

Close modal
Before launch, the ion pump must be shut down; consequently, vacuum maintenance is dependent upon the getters during the operation phase in orbit. In an ultra-high vacuum environment, the outgassing rate of a material varies with temperature and decays with time. At room temperature, the equation35,36 can be expressed as follows:
(10)
where q is the outgassing rate (measured in [Torr L/(cm2 s)], 1 Torr ≈133 Pa) at time t. The initial outgassing rate, q1, is defined as the outgassing rate at time t = 1 h. α is the power-law decay exponent of the outgassing rate of the material. The total outgassing of each gas species released in the vacuum chamber at room temperature can be calculated as follows:33 
(11)
where Q̇ ([Torr L/s]) represents the total outgassing rate of materials within the vacuum chamber, N is the total number of materials, qj,1 is the initial outgassing rate of material j at room temperature, and Aj is the total surface area of material j.

Typically, metals exhibit significantly lower outgassing rates than those of polymer materials. The outgassing rates of Viton-A and PCTFE polymer samples were accurately measured using the equipment shown in Fig. 9(a). The results of the measurement are presented in Fig. 9(b). The horizontal axis represents time in hours, while the left and right vertical axes correspond to the measured outgassing rate and temperature, respectively.

FIG. 9.

(a) Equipment for measuring the outgassing rate. (b) Measurement results of the samples. During the second 30 °C stage, the samples were exposed to clean atmosphere for 30 min. qVitonA(t11) and qPCTFE(t11) represent the outgassing rate of the two samples at the 11th measuring moment, respectively.

FIG. 9.

(a) Equipment for measuring the outgassing rate. (b) Measurement results of the samples. During the second 30 °C stage, the samples were exposed to clean atmosphere for 30 min. qVitonA(t11) and qPCTFE(t11) represent the outgassing rate of the two samples at the 11th measuring moment, respectively.

Close modal

The measurement conditions were designed to simulate the temperature and atmosphere-exposure conditions that the polymer materials used in the vacuum chamber may experience during assembly. This process comprises four temperature-phases, with temperatures set sequentially at 303.15 K (30 °C), 373.15 K (100 °C), 303.15 K (30 °C), and 333.15 K (60 °C). During the second phase at 303.15 K (30 °C), the sample was exposed to a clean atmosphere for 30 min to simulate the potential atmospheric exposure of the polymer components during the assembly. This phase involved mounting the integrated optical reference in the ultra-high vacuum chamber where all other components had already been positioned.

The results obtained from the 11th measurement were utilized as the initial outgassing rate for each polymer sample, as they closely approximate the outgassing rate of the polymer components at the moment of packaging system assembly.

The value of α for these polymer samples was estimated within the range of 0.5–2 based on the limited available measurement data. A value of α = 0.5 was determined to be consistent with the commonly reported range for polymer materials at room temperature, which typically ranges from 0.4 to 0.8.35,36 The primary outgassing species detected in both samples were H2O, H2, and CO/N2 (the device could not differentiate between CO and N2 due to their similar molecular masses), as shown in Table V.

TABLE V.

Outgassing rate of the main materials.

Initial outgassing rate [Torr L/(cm2 s)]
TotalThe main outgassing species
MaterialSurface area (cm2)q1H2OH2CO/N2
Viton-A 21.0 1.07 × 10−7 4.5 × 10−8 1.9 × 10−8 1.4 × 10−8 
PCTFE 316.3 1.01 × 10−8 3.2 × 10−9 1.2 × 10−9 1.7 × 10−9 
Teflon 0.8 4.89 × 10−8 1.2 × 10−8 5.8 × 10−9 3.2 × 10−8 
TC4 758.3 2.5 × 10−14 ⋯ 2.0 × 10−14 5 × 10−13 
QBe-1.9 1061.0 4.2 × 10−12 ⋯ 3.4 × 10−12 8 × 10−13 
Invar 665.8 3.0 × 10−12 ⋯ 2.4 × 10−12 6 × 10−13 
ULEa 393.0 1.9 × 10−12 ⋯ ⋯ ⋯ 
Initial outgassing rate [Torr L/(cm2 s)]
TotalThe main outgassing species
MaterialSurface area (cm2)q1H2OH2CO/N2
Viton-A 21.0 1.07 × 10−7 4.5 × 10−8 1.9 × 10−8 1.4 × 10−8 
PCTFE 316.3 1.01 × 10−8 3.2 × 10−9 1.2 × 10−9 1.7 × 10−9 
Teflon 0.8 4.89 × 10−8 1.2 × 10−8 5.8 × 10−9 3.2 × 10−8 
TC4 758.3 2.5 × 10−14 ⋯ 2.0 × 10−14 5 × 10−13 
QBe-1.9 1061.0 4.2 × 10−12 ⋯ 3.4 × 10−12 8 × 10−13 
Invar 665.8 3.0 × 10−12 ⋯ 2.4 × 10−12 6 × 10−13 
ULEa 393.0 1.9 × 10−12 ⋯ ⋯ ⋯ 
a

The outgassing rate of the ULE material is substituted by the measured value of quartz glass, given by the Beijing Orient Institute of Measurement and Test.

The optical fiber, employed for transmitting the laser to the monolithic optical bench, is coated with Teflon. The outgassing rate of a Teflon sample was also measured. The outgassing rates of the metal materials used within the vacuum chamber were obtained from public reports,34,35 with a power-law decay exponent of 1. In ultra-high vacuum environments, the primary outgassing species of metals are H2 gas (about 80%) and CO gas (about 20%). Nevertheless, the initial outgassing rates of the materials listed in Table V, particularly the polymer materials, could be further reduced by one or more orders of magnitude through extensive degassing processes such as thorough surface cleaning, long-time baking (at a safe temperature), and long-time degassing by an ion pump during the ground phase.

To address outgassing, the SAES St172/HI/16-10/300C non-evaporable getter was selected due to its high saturation capacity for these outgassing species. The getter assembly consisted of three parallel getters. Because the CO saturation capacity of this type of getter is significantly lower than those of other major gas species, the safety factor of the getter assembly is defined as the ratio of its maximum CO absorption capacity to the target CO absorption amount. The integration of Eq. (11) over time estimates that the amount of CO released over five years would be less than 1.3 Torr L (even if all outgassing of ULE is considered to be CO). The typical CO sorption curve of SAES St172/HI/16-10/300C (activated at 900 °C for 10 min) at room temperature shows an approximately linear relationship between the sorbed quantity (represented by xc in units of “Torr cm3”) and pumping speed (represented by ys in units of “cm3/s”).37 This relationship can be expressed as lgys=3.70.82lg(xc). It was assumed that when ys falls below 20 cm3/s, the sorbed quantity of CO for one getter reaches saturation (xc ≈ 842Torr cm3), suggesting a CO saturation capacity of ∼2.5 Torr L for the getter assembly.

Considering the significant overestimation of the initial outgassing rate of the materials, the safety factor of the getter assembly was predicted to be an order of magnitude greater than 1.9. It should be noted that this prediction was a rough estimate. It would be beneficial to conduct a more accurate experiment on long-term vacuum maintenance using getters in the future.

A full engineering prototype of a packaging system based on this design was constructed, incorporating the integrated monolithic optical reference, input optics, support structure, thermal shield, and vacuum chamber housed within. A series of typical AETs, including sinusoidal vibration, random vibration, thermal cycling, and thermal vacuum tests, were conducted on the prototype, with the conditions introduced in Sec. II. In addition, a thermal time-constant experiment was performed.

The vibration test setup is illustrated in Fig. 10(a). During the vibration tests, the acceleration responses of the prototype to the vibration excitations were measured using acceleration sensors. Prior to and following each sinusoidal and random vibration, low-amplitude (0.5 g) sinusoidal vibration resonance scans (10–2000 Hz) were conducted.

FIG. 10.

(a) Vibration test setup. (b) Thermal cycling test setup.

FIG. 10.

(a) Vibration test setup. (b) Thermal cycling test setup.

Close modal

The resonance frequencies observed in the scans were recorded and are detailed in Table VI. These data showed no significant changes before and after the tests. This indicates that there was no structural damage during the vibration testing. The measured resonant frequencies were consistent with the calculated low-order natural frequencies of the packaging system, supporting the reliability of the solid-mechanical FEA simulations and vibration reduction design.

TABLE VI.

Low-order natural frequencies in simulation and resonance frequencies in 0.5 g sinusoidal vibration sweep.

Natural/peak frequencies
Statusof the entire system (Hz)
Modal analysis   189 191 209 222 247 263 281 295 
0.5 g sinusoidal Along X axis Before tests 177 211 268 298 
vibration resonance scan After tests 173 211 266 299 
 Along Y axis Before tests 188 222 268 309 
 After tests 185 214 254 306 
 Along Z axis Before tests ⋯ 222 272 316 
 After tests ⋯ 209 273 321 
Natural/peak frequencies
Statusof the entire system (Hz)
Modal analysis   189 191 209 222 247 263 281 295 
0.5 g sinusoidal Along X axis Before tests 177 211 268 298 
vibration resonance scan After tests 173 211 266 299 
 Along Y axis Before tests 188 222 268 309 
 After tests 185 214 254 306 
 Along Z axis Before tests ⋯ 222 272 316 
 After tests ⋯ 209 273 321 

Thermal cycling tests in both atmospheric pressure and vacuum of 10−3 Pa were conducted in accordance with the conditions specified in Sec. II, using a test chamber as shown in Fig. 10(b). A thermal vacuum test in the same temperature range was also conducted in a thermal vacuum test chamber with a vacuum level of 10−3 Pa. Following ten cycles (∼90 h) of atmospheric pressure temperature cycling and ∼39.5 h of thermal vacuum cycling, no material failure was observed.

According to the characterization results of the optical reference conducted in lab before and after all the aforementioned environmental tests, the performance of the optical reference has not significantly changed, the measured fineness of the cavity remains 280 000, and the detected PDH error signals remain normal.

As shown in Fig. 11(a), in the thermal time constant experiment, an engineering prototype with the outer packaging box removed was placed in a thermostat. Pt100 platinum resistance temperature sensors were affixed to the exterior surface of the vacuum chamber and top surface of the OB (where the optical cavity was located). The temperature signals from the OB were transmitted to the exterior through a vacuum feedthrough. During the experiment, the gas pressure in the vacuum chamber was maintained at 10−6 Pa using an ion pump.

FIG. 11.

(a) Thermal time constant experiment. (b) Temperature response curve. (c) Temperature noise spectrum.

FIG. 11.

(a) Thermal time constant experiment. (b) Temperature response curve. (c) Temperature noise spectrum.

Close modal

A step temperature excitation of 2 °C was applied to the exterior surface of the vacuum chamber. In Fig. 11(b), the measured temperatures of the vacuum chamber and bench are represented by the black and blue curves, respectively. The heat transfer function from the vacuum chamber to the optical bench was derived through curve fitting (red curve). This function is mathematically expressed as δTOBtδTV act(1et95 590), where δTOBt and δTV act represent the temperature variations on the surfaces of the optical bench and vacuum chamber, respectively. The equivalent thermal time was determined as 95 590 s in the experiment, which closely matched the simulated value.

The temperature noise spectra of both the vacuum chamber and optical bench in a steady state were analyzed, and the results are shown in Fig. 11(c). It was observed that above 3 mHz, the amplitude of the temperature noise was limited by the sensitivity floor of the signal acquisition devices. At 1 mHz, the temperature attenuation from the vacuum chamber to the optical reference was calculated as δTOBf|f=1mHzδTVacf|f=1mHz1/100. Consequently, using Eq. (6), it was determined that the thermal time constant from the vacuum chamber to the optical bench was 1 × 105 s, in good alignment with the value derived through curve fitting of the time-dependent heat transfer function. With this thermal shielding factor, active temperature control with a noise floor of 1 mK/Hz1/2 at 10 mHz is sufficient to achieve a temperature stability of 1.1 × 10−6 K/Hz1/2 at 10 mHz on the optical cavity.

Regarding radiation testing, a beta radiation exposure assessment has been conducted on a sample of optical fiber, which is the same as that used in our system. The radiation dose applied was 5 krad, as recommended by the satellite environmental testing facility. Post-assessment results indicate that there is no significant change in the laser power transmission performance of the optical fiber, based on calibration data obtained in our laboratory. However, comprehensive radiation testing of the entire packaging system, which includes the integrated monolithic optical reference, has yet to be performed.

In addition to radiation tests, subsequent research will entail the execution of particular experiments and tests on the assembled prototype, to achieve the full space qualification of the packaging system. These tasks include shock tests, calibration of the effective zero-crossing temperature of the cavity, and precise monitoring of long-term vacuum maintenance using getters.

Basing on our previous work, we developed a packaging system for an integrated monolithic optical reference toward space application. The modification of the support structure is described in detail, along with the thermal shielding mechanism. Preliminary studies were conducted to ascertain the feasibility of maintaining a vacuum around the cavity during in-orbit operation.

The comprehensive performance analysis showed that the packaging system possesses sufficient safety factors to withstand typical launch loads, filter out external temperature variations, and maintain a long-term ultra-high vacuum level around the cavity. An engineering prototype built based on the aforementioned design successfully passed a series of typical environmental tests including sinusoidal vibration (10 g, 10–100 Hz), random vibration (0.045 g2/Hz, 10–2000 Hz), and temperature cycling (0–45, 3 °C/min, for 90 h). In addition, the experimental thermal time constant from the vacuum chamber to the optical bench exceeded 9.5 × 104 s, which was consistent with the simulated value and twice the required value. The optical performance of the integrated monolithic optical reference has already been demonstrated, and the results have been published;7 we have not repeated it in this paper.

This paper provides an important step process toward a completed space-qualified packaging system for an integrated optical reference.

It is noteworthy that parallel activities in space qualification of the laser required are being undertaken by our colleagues;24,25,38 however, the latest processes have yet to be published. Once both the optical reference system and the laser achieve full space qualification, a pre-launch experiment of spaceborne laser frequency stabilization will be conducted. The results of all the anticipated tests and experiments discussed above will be published, in alignment with the project’s timeline.

The design presented in this paper can also be used to study future mission concepts and is suitable for laboratory and terrestrial applications.

This research was funded by the National Key Research and Development Program of China (Grant Nos. 2022YFC2204603 and 2020YFC2200200).

The authors have no conflicts to disclose.

Zhenhai Zhan: Data curation (lead); Formal analysis (lead); Investigation (equal); Methodology (lead); Validation (equal); Visualization (equal); Writing – original draft (lead); Writing – review & editing (equal). Yingxin Luo: Conceptualization (lead); Funding acquisition (equal); Investigation (equal); Methodology (lead); Supervision (lead); Validation (equal); Writing – review & editing (equal). Hsien-Chi Yeh: Funding acquisition (equal); Project administration (equal); Supervision (equal); Writing – review & editing (equal). Hongyin Li: Investigation (equal); Resources (equal). Weilu Chen: Formal analysis (equal); Validation (equal). Chongzhi Ren: Methodology (equal); Validation (equal). Bingcheng Zeng: Methodology (equal); Validation (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.

The combination model of the integrated monolithic optical reference and the support structure with a flexible Viton-A structure can be represented as a simplified block–spring model. In Figs. 12(a) and 12(b), the simplified B–S models are shown in the horizontal and vertical planes, respectively. The flexible structure, constructed of Viton-A, functions as a spring with an elastic coefficient denoted by k1 and damping coefficient c1 (∼0.1). The masses of the metal part of the support structure and monolithic optical reference are represented by M1 (∼0.908 kg) and M2 (∼0.67 kg), respectively.

FIG. 12.

Block–spring model of the monolithic optical reference and support structure in view of (a) XOY and (b) X(Y)OZ.

FIG. 12.

Block–spring model of the monolithic optical reference and support structure in view of (a) XOY and (b) X(Y)OZ.

Close modal
The acceleration ai(i = 1, 2) and displacement xj(j = 1, 2) represent the block motions. The equation for the state of motion is as follows:
(A1)
where K = 4k1 and C = 4c1 denote the total elasticity coefficient and damping of the four springs, respectively. The vibration transfer function is obtained as follows:
(A2)
where ωd [with a unit of (rad/s)] represents the natural frequency of Viton-A and is equal to K/M2, while ζ denotes the damping coefficient defined as C/M2. As the frequency ω approaches ωd, the acceleration gain from the metal support structure to the OB is proportional to ωd and inversely proportional to ζ. Conversely, at frequencies significantly higher than ωd, the external vibration acceleration is effectively reduced. Therefore, when designing a flexible structure, it is preferable to set ωd in a low-frequency range with low acceleration to minimize the external vibration effects.

The elastic coefficient of Viton-A can be calculated as k1 = EA/d, where E represents Young’s modulus (∼7.8 MPa), A denotes the cross-sectional area perpendicular to the spring direction, and d is the spring length equivalent to the wall thickness. For an X-axis aligned spring structure, the bottom surface area (π × 52 ≈ 78.54 mm2) can be used as an approximation for A, while d approximates at ∼1.25 mm (resulting from compressing the bottom wall of Viton-A by 0.25 mm with a preload of 100 N). Conversely, in the case of a spring structure aligned along the Y or Z axis, A is equivalent to the lateral surface area (∼2π × 5 × 5 ≈ 157.08 mm2), whereas d is 1.5 mm. The theoretical low-order natural frequencies of the model shown in Fig. 12 were calculated as fd=12π×4k1/M2. As shown in Table VII, the calculated value of fd closely matches that of fd, which represents the simulated natural frequency of the model shown in Fig. 2(a). However, small discrepancies existed between fd and fd, which could be attributed to the simplification of the geometric model in Fig. 12 compared to the model shown in Fig. 2(a).

TABLE VII.

Geometrical parameters of Viton-A and the calculated natural frequencies.

DirectionA (mm2)D (mm)K/kN m−1fd (Hz)fd (Hz)
x axis 78.54 1.25 4 × 490 272 289 
y axis 157.08 1.50 4 × 817 351 336 
z axis 157.08 1.50 4 × 817 351 340 
DirectionA (mm2)D (mm)K/kN m−1fd (Hz)fd (Hz)
x axis 78.54 1.25 4 × 490 272 289 
y axis 157.08 1.50 4 × 817 351 336 
z axis 157.08 1.50 4 × 817 351 340 

As shown in Fig. 13, the amplitude–frequency response curve of the vibration obtained using Eq. (A1) is close to that of the FEA simulation. The second peak in the simulation curve represents a higher order natural frequency. However, this cannot be resolved using Eq. (A1).

FIG. 13.

Amplitude–frequency response curve of the vibration transfer function from the metal plate of the support structure to the OB.

FIG. 13.

Amplitude–frequency response curve of the vibration transfer function from the metal plate of the support structure to the OB.

Close modal

For a complex system with multilayer structures, the vibration transfer function from the outermost layer to the secondary outer layer can be described by Eq. (A2). M1 and M2 represent the mass of the outermost layer and the total mass of the components within the outermost layer, respectively. The  Appendix presents a convenient method for estimating vibration transmission between adjacent layers in a packaging system.

1.
M.
Peterseim
,
O. S.
Brozek
,
K.
Danzmann
,
I.
Freitag
,
P.
Rottengatter
,
A.
Tünnermann
, and
H.
Welling
,
AIP Conf. Proc.
456
(
1
),
148
155
(
1998
).
2.
J.
Stacey
,
G.
Barwood
,
A.
Spampinato
,
P.
Tsoulos
,
C.
Robinson
,
P.
Gaynor
et al, in
International Conference on Space Optics
(
ICSO
,
2022
).
3.
R.
Thompson
,
W. M.
Folkner
,
G.
de Vine
,
W.
Klipstein
,
K.
McKenzie
,
R. E.
Spero
,
N.
Yu
,
M.
Stephens
,
J.
Leitch
,
R.
Pierce
et al, in
European Frequency and Time Forum (FCS) Proceedings
(
IEEE
,
2011
), pp.
1
3
.
4.
K.
Nicklaus
,
K.
Voss
,
A.
Feiri
,
M.
Kaufer
,
C.
Dahl
,
M.
Herding
,
B. A.
Curzadd
,
A.
Baatzsch
,
J.
Flock
,
M.
Weller
et al,
Remote Sens.
14
(
16
),
4089
4106
(
2022
).
5.
J.
Luo
,
L. S.
Chen
,
H. Z.
Duan
,
Y. G.
Gong
,
S.
Hu
,
J.
Ji
et al, “
TianQin: A space-borne gravitational wave detector
,”
Classical Quantum Gravity
33
(
3
),
035010
035029
(
2016
).
6.
W. R.
Hu
and
Y.-L.
Wu
,
Natl. Sci. Rev.
4
(
5
),
685
686
(
2017
).
7.
Y.
Luo
,
H.
Li
,
Y. Q.
Li
,
Z. H.
Zhan
,
C. Y.
Huang
,
J. T.
Lai
, and
H. C.
Yeh
,
Appl. Opt.
60
(
10
),
2877
2885
(
2021
).
8.
R.
Drever
,
J.
Hall
,
F.
Kowalski
,
J.
Hough
,
G. M.
Ford
,
A. J.
Munley
, and
H.
Ward
,
Appl. Phys. B
31
,
97
105
(
1983
).
9.
S.
Webster
and
P.
Gill
, in
2010 IEEE International Frequency Control Symposium
(
IEEE
,
2010
), pp.
470
473
.
10.
T.
Kessler
,
T.
Legero
, and
U.
Sterr
,
J. Opt. Soc. Am. B
29
,
178
184
(
2012
).
11.
Q. F.
Chen
,
A.
Nevsky
, and
S.
Schiller
,
Appl. Phys. B
107
(
3
),
679
683
(
2012
).
12.
B. K. Q. F. K.
TaoChenTao
and
Q. F.
Chen
,
Appl. Phys. B
124
,
228
234
(
2018
).
13.
J. M.
Robinson
,
E.
Oelker
,
W. R.
Milner
,
W.
Zhang
,
T.
Legero
et al,
Optica
6
(
2
),
240
243
(
2019
).
14.
J.
Shang
,
J.
Cao
,
K. S. P.
CuiWangZhang
,
J. S. H.
YuanChaoShu
,
P.
Zhang
,
J.
Yuan
, and
X.
Huang
,
Opt. Commun.
382
,
410
414
(
2017
).
15.
J.
Keller
,
S.
Ignatovich
,
S. A.
Webster
,
T. E.
Mehlstäubler
, and
E.
Tanja
,
Appl. Phys. B
116
,
203
210
(
2014
).
16.
S.
Häfner
,
S.
Falke
,
C.
Grebing
,
S.
Vogt
,
T.
Legero
,
M.
Merimaa
,
C.
Lisdat
, and
U.
Sterr
,
Opt. Lett.
40
(
9
),
2112
2115
(
2015
).
17.
S.
Herbers
,
S.
Häfner
,
S.
Dörscher
,
T.
Lücke
,
U.
Sterr
, and
C.
Lisdat
,
Opt. Lett.
47
(
20
),
5441
5444
(
2022
).
18.
S.
Häfner
,
S.
Herbers
,
S.
Vogt
,
C.
Lisdat
, and
U.
Sterr
,
Opt. Express
28
(
11
),
16407
16416
(
2020
).
19.
K.
Abich
,
A.
Abramovici
,
B.
Amparan
et al, “
In-orbit performance of the GRACE follow-on laser ranging interferometer
,”
Phys. Rev. Lett.
123
(
3
),
31101
31108
(
2019
).
20.
J.
Sanjuan
,
K.
Abich
,
M.
Gohlke
et al,
Opt. Express
27
(
25
),
36206
36220
(
2019
).
21.
D.
Świerad
,
S.
Häfner
,
S.
Vogt
et al, “
Ultra-stable clock laser system development towards space applications
,”
Sci. Rep.
6
(
1
),
33973
33981
(
2016
).
22.
B.
Argence
,
E.
Prevost
,
T.
Lévèque
,
R.
Le Goff
,
S.
Bize
,
P.
Lemonde
, and
G.
Santarelli
,
Opt. Express
20
,
25409
25420
(
2012
).
23.
G. D.
Cole
et al,
Opt. Express
32
(
4
),
5380
5396
(
2024
).
24.
Z.
Li
,
H.
Duan
,
X.
Huang
,
M.
Ming
et al,
Opt Laser. Technol.
141
,
107155
(
2021
).
25.
M. B. F.
MingCaoLiao
et al,
Classical Quantum Gravity
41
,
165014
(
2024
).
27.
J.
Mei
,
Y.-Z.
Bai
,
J.
Bao
,
E.
Barausse
,
L.
Cai
,
E.
Canuto
et al, “
The TianQin project: Current progress on science and technology
,”
Prog. Theor. Exp. Phys.
2021
(
5
),
05A107
(
2020
).
28.
29.
B.
Edlén
,
J. Opt. Soc. Am.
43
,
339
344
(
1953
).
30.
G.
Bönsch
and
E.
Potulski
,
Metrologia
35
,
133
139
(
1998
).
32.
J.
Sanjuan
,
N.
Gürlebeck
, and
C.
Braxmaier
,
Opt. Express
23
(
14
),
17892
17908
(
2015
).
33.
X.
Dai
,
Y.
Jiang
et al,
Opt. Express
23
(
4
),
5134
5146
(
2015
).
34.
J.
Zhang
,
W.
Wu
,
X. H.
Shi
,
X. Y.
Zeng
,
K.
Deng
, and
Z. H.
Lu
,
Rev. Sci. Instrum.
87
(
2
),
023104
(
2016
).
35.
R.
Grinham
and
A.
Chew
,
Appl. Sci. Convergence Technol.
26
(
5
),
95
109
(
2017
).
36.
P.
Chiggiato
, arXiv 2006.07124 (
2020
).
37.
Saes getters, technical specification specifiche techniche Doc.M.K300. 0030.21 Rev.1. ST 172/HI/16-10/300 C (SAES Code 5K0405) (
1992
).
38.
W.
Fan
,
C.
Ma
,
D.
Liu
,
R.
Zhu
,
G.
Zhou
,
X.
Gong
,
S.
Zhou
,
J.
Xu
,
W.
Yuan
,
C.
Guo
, and
H. C. C.
YehYeh
,
Opt. Express
31
,
13402
13413
(
2023
).