PaperlayerView exact revision on arXiv

Source-linked AI summary

The Feasibility of a Fully Miniaturized Magneto-Optical Trap for Portable Ultracold Quantum Technology

Joseph Rushton, Matthew Aldous, Matt Himsworth

arXiv:1405.3148v1physics.ins-detcond-mat.mtrl-sciquant-ph

TL;DR

Ultracold-atom experiments have promising applications, but their surrounding infrastructure remains a barrier to rugged, practical devices. This paper evaluates a sealed, microfabricated MOT integrating the vacuum system, atom source, and optics. It concludes that such a device could maintain UHV for at least 1000 days using passive pumping, subject to stringent material and fabrication requirements.

  • Problem

    Ultracold atom-chip components remain part of laboratory-scale systems, and practical devices require miniaturized, integrated infrastructure.

    Source: §I · p. 1

  • Method

    The study assesses a microfabricated, sealed MicroMOT design using selected materials, pumping elements, and structural approaches to control leaks, permeation, and outgassing.

    Source: E. Vacuum discussion · p. 16

  • Results

    The study concludes that microfabrication techniques and materials can support sealed microlitre chambers maintaining UHV for at least 1000 days with passive pumping.

    Source: §VI · p. 18

  • Takeaways & Limitations

    The proposed MicroMOT combines a miniaturized MOT geometry and atom source with a sealed chamber, while requiring extremely low leakage, low outgassing, and negligible noble-gas outgassing from chamber walls and sputtered films.

    Source: §VI · p. 18

    Outgassing cannot be reliably predicted for real materials, and the diffusion-limited model is accurate only to about an order of magnitude, especially for reactive species such as hydrogen.

    Source: §VI · p. 18

Abstract

from arXiv · show

Experiments using laser cooled atoms and ions show real promise for practical applications in quantum- enhanced metrology, timing, navigation, and sensing as well as exotic roles in quantum computing, networking and simulation. The heart of many of these experiments has been translated to microfabricated platforms known as atom chips whose construction readily lend themselves to integration with larger systems and future mass production. To truly make the jump from laboratory demonstrations to practical, rugged devices, the complex surrounding infrastructure (including vacuum systems, optics, and lasers) also needs to be miniatur- ized and integrated. In this paper we explore the feasibility of applying this approach to the Magneto-Optical Trap; incorporating the vacuum system, atom source and optical geometry into a permanently sealed micro- litre system capable of maintaining $10^{-10}$ mbar for more than 1000 days of operation with passive pumping alone. We demonstrate such an engineering challenge is achievable using recent advances in semiconductor microfabrication techniques and materials

I. ULTRACOLD QUANTUM TECHNOLOGY

Ultracold-atom experiments have broad potential, but their laboratory-scale infrastructure remains too large and specialized for rugged, mass-produced use. The paper proposes a microfabricated, sealed MicroMOT and examines the design requirements and constraints of miniaturizing the trap and its vacuum system.

I. ULTRACOLD QUANTUM TECHNOLOGY

  • Existing ultracold systems are typically bespoke, occupy at least a cubic metre, and require expert operators, while atom chips remain components within a laboratory-scale system.The paper compares the miniaturization goal to chip-scale atomic clocks and notes that ultracold-system miniaturization remains at an earlier stage.Source: §I · p. 1
  • The proposed MicroMOT integrates the vacuum system, atom source, and MOT optics in a centimetre-scale device targeting 10^-10 mbar for 1000 days with passive pumping.The authors frame this as an engineering challenge aimed at mass-producible technology rather than a bespoke proof-of-concept system.Source: §I · p. 2

II. THE MAGNETO OPTICAL TRAP SYSTEM

  • Standard MOTs use a litre-scale UHV chamber with optical ports, atom sources, gauges, and pumps, and their pumps and gauges make compact, mass-produced systems difficult.The paper describes typical systems as assembled and evacuated through staged pumping and baking to reach about 10^-10 mbar.Source: §II · p. 2
  • The standard MOT geometry is poorly suited to integration because it requires many optical ports, precise alignment, large volumes, and fragile optical elements.It also makes bringing atoms close to an atom-chip surface difficult.Source: §II · p. 2
  • The G-MOT uses a planar grating reflector and a single incident beam, and is presented as the most suitable geometry for microfabricated devices.It can capture many atoms and has lower backscatter, although the grating affects beam wavefronts and polarizations and complicates use of widely spaced wavelengths.Source: §II · p. 3, Figure 2 · p. 3
  • A practical G-MOT capture volume is constrained by atom number and trap geometry, while a 3 mm chamber can limit free-fall interaction time to 5 ms under the stated rubidium assumptions.The cited minimum practical beam-overlap volume is 0.045 cm3; sub-Doppler cooling may increase the example interaction time by a factor of 3 to 4.Source: §II · p. 3

III. ATOM SOURCE AND CONTROL

The MicroMOT must control its atom source to load a trap without sustaining vapor pressures that compromise vacuum or atom coherence. The paper considers vapor loading and source, pumping, and temperature-control approaches, with rubidium as the challenging example.

III. ATOM SOURCE AND CONTROL

  • The study focuses on loading a single MOT from room-temperature vapor and controlling it on short timescales, rather than producing a degenerate gas.A 2D-MOT/3D-MOT arrangement can separate high-pressure loading from the higher-vacuum trap chamber, but multichamber systems are likely needed for BECs.Source: §III · p. 3
  • Rubidium vapor pressure must be regulated because room-temperature vapor saturates microlitre volumes, while MOT loading needs about 10^-8 mbar and decoherence-sensitive measurements require an order-of-magnitude reduction.At room temperature rubidium has a vapor pressure of 5 × 10^-7 mbar, causing collisions, excess fluorescence, and rapid vapor saturation in the small chamber.Source: §III · p. 4
  • Alkali sources include pure metals, compounds, wax pellets, azides, and enriched glass, but many are unsuitable for UHV or have poorly controlled or limited lifetimes.Pure rubidium must be sealed away during fabrication; commercial dispensers can be activated by Joule heating or a focused laser.Source: §III · p. 4
  • Surface pumping alone may not sustain the target lifetime: the stated loading scenario would consume about 1 mg of rubidium over 1000 days, exceeding a typical monolayer capacity.The passage suggests adding high-surface-area materials, provided they can be sufficiently degassed before encapsulation.Source: §III · p. 4
  • Cooling rubidium to −30°C is described as necessary to reach 10^-10 mbar, and an integrated cold finger can avoid condensation on windows and reflectors.Other discussed approaches include pumping excess rubidium with heated gold or rubidium/bismuth alloying; getter-film pumping of alkali metals may be negligible, with little data available.Source: §III · p. 4
  • A narrow conductance channel can separate a high-pressure source chamber from the UHV MOT chamber, while local source heating can increase vapor pressure for loading.A 1 mm-long channel with a 100 × 100 µm cross-section is described as maintaining UHV at room temperature given a pumping rate above 0.1 l s−1 in the laser-cooling chamber.Source: §III · pp. 4–5

IV. UHV IN A CHIP

Maintaining sealed UHV in a chip requires controlling leaks, permeation, and outgassing while retaining a means to assess internal pressure. The paper reviews encapsulation options and material choices for a microfabricated chamber.

IV. UHV IN A CHIP

  • The study identifies leak, permeation, and outgassing rates as central challenges for sealed passive UHV chips and compares the requirements with typical UHV systems.It proposes addressing demanding values through material selection, fabrication processes, and structural features.Source: §IV · p. 5
  • The paper describes MOT loading rate as a pressure gauge, but its stated sensitivity range of 10^-7 to below 10^-9 mbar is slightly above the target pressure.It may still indicate the internal environment, and improved sensitivity to 10^-12 mbar may be possible if background rubidium is quickly reduced after loading.Source: §IV · p. 5
  • Wafer-to-wafer bonding under vacuum enables thorough degassing and evacuation and can incorporate a structural scheme to reduce leaks, while post-bond evacuation-tube sealing is simpler to construct.The latter approach faces difficulty evacuating through a small aperture and challenges in wafer-level sealing complexity, uniformity, and time.Source: §IV · pp. 5–6, Figure 3 · p. 5
  • The device needs optical access for cooling beams and fluorescence detection, but glass-metal thermal-expansion mismatch can constrain baking, bonding, and operating temperatures.The passage notes that shaped sealing edges can alleviate some mismatch, while many metals do not meet the required low outgassing rates or are impractical for microfabrication.Source: §IV · p. 6
  • Silicon is proposed as the principal substrate because it has low room-temperature permeation and outgassing, high purity, thermal robustness, and compatibility with optical materials.Its brittleness limits it to small structures, which the authors consider acceptable for this application; silicon is also non-magnetic and poorly conducting.Source: §IV · p. 6

A. Pumping

The MicroMOT uses passive non-evaporable getters (NEGs) to pump residual gases, but cannot rely on perfect seals or impermeable materials. Getter capacity, gas-specific pumping, and hydrogen equilibrium constrain the design.

A. Pumping

  • NEGs chemisorb common residual gases and absorb reacted surface material into their bulk during activation, continuing to pump at room temperature until their surfaces saturate.A passivating surface layer diffuses into the bulk during heating, exposing a fresh metallic pumping surface.Source: A. Pumping · p. 6
  • NEGs do not pump noble gases or some hydrocarbons such as methane at room temperature, so the MicroMOT must be sealed at UHV to evacuate most noble gases beforehand.A lithium getter could replace or complement an NEG if NEGs alone cannot achieve the required vacuum; lithium has greater capacity and can pump methane.Source: A. Pumping · p. 7
  • Sputtered TiZrV NEG films can activate at 180°C, and coatings have demonstrated a room-temperature vacuum of 10^-14 mbar.Pumping rates and capacities depend on film surface area and deposition conditions; CO saturation can reduce pumping rates for other gases.Source: A. Pumping · p. 7, Table III · p. 7
  • At room temperature, a TiZrV NEG’s hydrogen disassociation pressure remains negligible at 10^-19 mbar even when the film is very saturated (xH = 0.01).The pressure only endangers the vacuum above 150°C, when helium permeation through glass becomes equally problematic.Source: A. Pumping · p. 7
  • Lithium films have a carbon-monoxide pumping capacity over 10^4 times greater than TiZrV NEGs, although their use in the MicroMOT remains a possible alternative or complement.Rubidium may also getter some gases, but the paper expects less gettering than lithium because rubidium has higher vapour pressure.Source: A. Pumping · p. 7
  • The pumping analysis assumes a 0.5 cm3 internal volume, 5 cm2 surface area, and a 1 cm2 NEG area with 1 µm thickness.Hydrogen diffuses through the NEG bulk, unlike most gases that stop chemisorbing effectively after a monolayer forms.Source: A. Pumping · pp. 7–8

B. Permeation

Permeation through chamber walls depends on gas, material, and temperature, with helium ingress through optical glass posing a key vacuum risk. Alumino-silicate glass and silicon offer material-specific ways to limit permeation.

B. Permeation

  • Permeation involves adsorption, dissociation, absorption and diffusion through the wall, followed by surface escape into the evacuated volume.For thoroughly degassed materials, gas takes time to break through; this delay can be extremely long when diffusion rates are very low.Source: B. Permeation · p. 8
  • Helium permeation through Pyrex is high enough that the system would be expected to lose UHV within days after bonding.The amorphous glass network provides channels for helium diffusion, while modifier molecules can plug these channels and reduce permeation.Source: B. Permeation · p. 8
  • Corning 1720 alumino-silicate glass has helium permeation rates five orders of magnitude lower than Pyrex and can maintain the target vacuum duration.Figure 4 compares 1 mm optical wafers exposed to atmospheric helium after evacuation to 10^-10 mbar; the Corning 1720 data define the AS-glass properties used.Source: B. Permeation · pp. 8–9, Figure 4 · p. 8
  • Hydrogen permeation through glass can be neglected because its rate is far below helium’s, atmospheric hydrogen partial pressure is lower, and NEGs can pump it.The hydrogen permeability data for AS glass are limited, so the study estimates diffusivity using a scaling law, glass composition, and Pyrex solubility.Source: B. Permeation · p. 9
  • Silicon is practically impermeable to noble gases, while hydrogen permeation through silicon at room temperature is negligible.Some recent studies indicate room-temperature helium permeation through silicon may be more significant than expected, but more work is required to confirm this against other research.Source: B. Permeation · p. 9

C. Leaking

Sealed, long-lived microdevices require exceptionally low leakage because internal pumping may be impractical. The section’s comparison is bounded by material-data assumptions, including limited helium-solubility measurements.

C. Leaking

  • Wafer-level vacuum seals must achieve extremely low leakage because small-volume, long-lived chips may preclude internal pumping.Standard commercial seals quote leak rates below 10^-11 mbar l s^-1, but these are not suitable for wafer-level vacuum encapsulation.Source: C. Leaking · p. 9
  • The study considers bonding methods with demonstrated leak rates below 10^-13 mbar l s^-1 in air and temperatures no higher than 400°C.Low-temperature indium bonding is excluded because noble-gas outgassing may occur and its leak rate is insufficiently low.Source: C. Leaking · pp. 9–10
  • The tabulated silicon helium-solubility estimate relies on a limited-range measurement and a theoretical solution energy.The associated diffusivity values are theoretical but agree with experimental results; the AS-glass solubility is assumed equal to Pyrex’s.Source: C. Leaking · p. 10

Glass Frit

The section compares bonding approaches for sealing microfabricated vacuum devices. Glass frit is less sensitive to surface quality, gold-silicon eutectic bonding has the lowest reported leak rate, and anodic bonding requires getter films to address residual oxygen.

Glass Frit

  • Glass frit bonding deposits a low-melting-point glass compound between materials, then heats it to melt and seal the surfaces upon cooling.Its leak rate is estimated below 10^-15 mbar l s^-1, but the supporting pressure measurement is near the gauge limit and the immediate post-bonding pressure is unknown.Source: Glass Frit · p. 10

Eutectic Bonding

  • Gold-silicon eutectic bonding has demonstrated a leak rate below 10^-15 mbar l s^-1, the lowest the authors found.The bond forms at 363°C; disadvantages include multilayer-film requirements, inert-gas storage before bonding, and restrictions on subsequent heating.Source: Eutectic Bonding · p. 10

Anodic Bonding

  • Anodic bonding joins oxide-forming silicon and alkali-containing glass using heat and high voltage, with leak rates below 10^-14 mbar l s^-1.Bonding produces oxygen at the inner seam, which can be significant in small evacuated volumes, so getter films are mandatory.Source: Anodic Bonding · pp. 10–11

Direct Bonding

Bond leakage and getter capacity set the constraints for maintaining ultra-high vacuum in a sealed MicroMOT. The analysis finds that a buffer moat can substantially reduce leakage, while bonding methods and getter performance determine the remaining limits.

Direct Bonding

  • Several bonding methods achieve reliable leak rates below 10^-14 mbar l s^-1, and some measurements fall below 10^-15 mbar l s^-1; eutectic bonding has the lowest measured rates and low outgassing.Anodic and direct bonds may be hermetic but were beyond measurement sensitivity; glass frit may match low leak rates, though its residual gas pressure is unknown.Source: Direct Bonding · p. 11
  • Without pumping, a 0.5 cm^3 cavity permits only 3×10^-22 mbar l s^-1 for a 50% pressure rise over 1000 days; the helium example permits 1.5×10^-17 mbar l s^-1.The higher example rate reflects helium's atmospheric partial pressure of 5×10^-3 mbar.Source: Direct Bonding · p. 11
  • The NEG model does not dynamically account for one gas reducing the pumping speed of others; instead, it assumes the lowest pumping speed as a worst case.The lifetime approximation assumes pumping remains constant until getter capacity is reached, when pumping speed drops sharply.Source: Direct Bonding · p. 11
  • For nitrogen, a 0.1 l s^-1 getter with a capacity of 10^14 molecules permits a maximum leak rate of 5×10^-15 mbar l s^-1 for a 1000-day lifetime.This estimate uses a base pressure of 5×10^-14 mbar and assumes pumping remains constant until the NEG reaches capacity, then drops to zero.Source: Direct Bonding · p. 11
  • Argon is the most significant noble-gas leak in the model, with a maximum permissible rate of 1×10^-19 mbar l s^-1 and a nearly 10^4 reduction required from the lowest measured bonding leak rate.A buffer moat is proposed to reduce leakage by several orders of magnitude.Source: Direct Bonding · p. 12
  • A 0.05 cm^3 moat around a 0.5 cm^3 cavity reduces leakage by 2×10^5, allowing bonds with leak rates in the 10^-14 mbar l s^-1 range.The moat need not be bonded at UHV; the model shows little variation below an initial moat pressure of 10^-6 mbar.Source: Direct Bonding · p. 12

D. Outgassing

Outgassing from surfaces and material bulk is a major limit on sealed UHV, and its rate depends on gas species, material, and processing history. The analysis uses surface-desorption and bulk-diffusion models to set stringent outgassing requirements, especially during high-temperature bonding.

D. Outgassing

  • Outgassing sets the lowest base pressure for a given pumping rate, and hydrogen is the dominant gas at UHV.With NEG pumping around 0.1 l s^-1cm^-2, reaching 10^-10 mbar requires outgassing below 10^-11 mbar l s^-1.Source: D. Outgassing · p. 13
  • Surface outgassing is fastest for weakly bound gases, while species with desorption energies from 0.7 to 1.1 eV are difficult to remove without high-temperature baking.The paper names this difficult range the outgassing energies range (OER).Source: D. Outgassing · p. 13
  • Bulk outgassing is modeled as diffusion from a degassed slab, with gas transport and trapping affecting the rate and surface recombination sometimes limiting desorption.For non-dissociative gases, initial bulk concentration is x0 = SP; for dissociative gases it scales with P^0.5.Source: D. Outgassing · p. 13
  • For xH = 0.01, a 3600 s bond at 400°C requires a bonding outgassing rate below 10^-7 mbar l s^-1cm^-2 to keep η = 0.1.Bonding and NEG reactivation can raise outgassing; the temperature-scaling estimate assumes instantaneous temperature changes and a bonding period short enough not to alter operating outgassing.Source: D. Outgassing · pp. 13–14
  • For ED = 0.9 eV, room-temperature hydrogen outgassing must be below 10^-16 mbar l s^-1cm^-2 to meet the stated vacuum target.Carbon monoxide, with only single-monolayer NEG capacity, requires rates over a hundred times lower than hydrogen without continuous reactivation.Source: D. Outgassing · p. 14
  • Hydrogen outgassing rates are extrapolated using diffusion energy, while carbon monoxide requires rates over a hundred times lower because the NEG has only single-monolayer capacity.The analysis assumes clean surfaces and standard wafer cleaning; it also warns that detergents can remove vanadium from NEG films and should be avoided.Source: D. Outgassing · pp. 13–14

Hydrogen

Hydrogen outgassing from silicon is difficult to predict from high-temperature measurements, and processing can leave substantial hydrogen near the surface. The paper compares silicon and aluminosilicate glass behavior and describes how baking and surface barriers affect release.

Hydrogen

  • Hydrogen diffusion measurements in silicon vary widely, and the commonly cited vWW values are consistently higher than subsequent measurements and serve as an upper limit in the analysis.Room-temperature migration is predominantly atomic with an activation barrier near 0.5 eV; doping, impurities, defects, growth, orientation, and quantum effects affect exact values.Source: Hydrogen · p. 14
  • Measured room-temperature hydrogen solubilities in silicon are very low, and using different reported values changes calculated permeation by no more than a factor of two or three.The vWW solubility extrapolates to 10^-10 molecules per cubic centimetre under 1 bar H2; other measurements report at most a few hundred atoms per cubic centimetre.Source: Hydrogen · p. 14
  • Room-temperature hydrogen near silicon surfaces can reach 10^18 molecules per cubic centimetre after standard processing, then fall below 10^-30 mbar l s^-1cm^-2 within an hour under vacuum.The initial outgassing rate can be 10^-7 mbar l s^-1cm^-2, with later release limited by surface recombination.Source: Hydrogen · p. 14
  • After high-temperature baking, AS glass has a theoretical hydrogen outgassing rate just over 10^-17 mbar l s^-1cm^-2, close to an experimental value just below 10^-16.The paper attributes the low glass outgassing to strong Si-H and Si-OH bonds.Source: Hydrogen · p. 14

Noble gases

Noble gases must be removed before sealing because they cannot be pumped from the finished MicroMOT. The paper considers diffusion from glass and gas incorporated during metal-film deposition as sources to control.

Noble gases

  • Noble gases cannot be removed after sealing, so components must be completely degassed before bonding.Atmosphere-exposed glass may contain 10 ppb noble gases that diffuse out of the bulk; a thin AS-glass slab can be degassed, with wafer-level degassing scaling proportionally to area.Source: Noble gases · pp. 14–15
  • Figure 7 depicts substantial scatter in reported hydrogen diffusion through silicon, with the commonly quoted vWW rate shown as a solid black line.The figure includes hydrogen data without isotopes and indicates silicon type where known.Source: Figure 7 · p. 15
  • Sputtered metal films can trap argon that later permeates to the surface, while vacuum-arc deposition or e-beam evaporation avoids additional process gases.Replacing argon with krypton can reduce incorporation by a factor of 10^3 or more, but measured krypton outgassing from TiZrV NEG films remains as high as 10^-19 mbar l s^-1cm^-2 after activations.Source: Noble gases · p. 15

Carbon monoxide

Carbon monoxide is treated primarily as a surface gas, while other residual gases and bonding can threaten the sealed vacuum. The discussion identifies material, cleaning, getter, and bonding choices that constrain outgassing and residual-gas risks.

Carbon monoxide

  • Carbon monoxide is mainly a surface outgassing species because its size makes bulk diffusion negligible, and reported glass concentrations vary with glass type, processing, and measurement conditions.Carbon monoxide desorption from silicon has energy below 0.5 eV and should be readily degassed; the figure models glass and silicon rates across baking temperatures.Source: Carbon monoxide · p. 15, Figure 8 · p. 16

Other gases

  • Methane may arise from gauge reactions or reactions between carbon monoxide and hydrogen, but TiZrV NEG films do not pump it.Methane therefore needs evacuation before sealing or reactive getters; other organics in encapsulated devices are associated with insufficient cleaning or residual gases before sealing.Source: Other gases · p. 16
  • Anodic bonding can release oxygen at inner bonding edges, with an estimated 10^13 to 10^14 molecules per millimetre of inner circumference and possible NEG saturation.High-temperature treatment after bonding is important to absorb oxidized NEG layers into the bulk; rubidium can react with remaining oxygen to form Rb2O.Source: Other gases · p. 16

E. Vacuum discussion

  • The vacuum discussion favors eutectic and direct bonding for low outgassing and hermeticity, with anodic bonding suitable if released oxygen can be pumped away.A bonding-seam moat and NEG coating on inner edges, plus a barrier coating on outer edges, are proposed to reduce leakage.Source: E. Vacuum discussion · p. 16
  • NEGs are straightforward to deposit and can coat surfaces, while reactive getters offer higher capacity and methane pumping; combining them is presented as a compromise.The proposed sequence activates NEGs during bonding and reactive getters afterward.Source: E. Vacuum discussion · p. 17
  • After degassing, chosen materials are expected to meet room-temperature outgassing of 10^-13 mbar l s^-1cm^-2, but bonding can raise rates by eight orders of magnitude.The stated room-temperature targets during bonding are below 10^-16 for bulk-gettered hydrogen, 10^-18 for surface-gettered carbon monoxide, and 10^-21 mbar l s^-1cm^-2 for non-gettered noble gases; AS glass is expected to dominate hydrogen outgassing.Source: E. Vacuum discussion · p. 17

V. PROTOTYPE MICROMOT

The prototype combines a microfabricated trap geometry, atom-source and vacuum chambers, and sealed-chip pumping. Its estimated lifetime exceeds the target, while atom-source vapor and pumping determine practical performance.

V. PROTOTYPE MICROMOT

  • The generic design assumes a G-MOT with a 10 mm grating and a 15×15×3 mm3 cavity, avoiding commitment to one application.About 90% of the beam-overlap volume lies within 2.5 mm of the grating surface.Source: §V · p. 17
  • The simple generic design is intended to demonstrate feasibility, with adaptation expected for specific applications or manipulation techniques.Source: §V · p. 17
  • Four chambers separate science, atom sourcing, alkali getter/LIAD/peltier functions, and reactive-getter pumping; a 1×0.1×0.1 mm channel restricts vapor flow.The channel dimensions still require experimental optimization.Source: §V · p. 17
  • The 20×24×5 mm3 prototype uses bonded glass and silicon wafers, an internal NEG film, and a reflector eutectically bonded under UHV after baking.The design includes a moat at the bonding region to reduce argon leakage.Source: §V · p. 17, Figure 10 · p. 18
  • Nearly an order of magnitude above target, the estimated lifetime is limited by argon leakage.Source: §V · p. 17
  • After bonding, AS glass outgassing is negligible, yielding XHV base pressure and essentially unlimited lifetime; practical pressure and lifetime depend on atom-source vapor and pumping.Source: §V · p. 18

VI. DISCUSSION AND CONCLUSIONS

The study concludes that microfabrication can produce sealed microlitre chambers that maintain UHV for at least 1000 days with passive pumping. It identifies bonding, outgassing, and temperature-sensitive components as important design constraints.

VI. DISCUSSION AND CONCLUSIONS

  • Degassing bulk materials up to 600°C may be unsuitable for heat-sensitive integrated components, requiring lower-temperature cleaning and bonding methods.Source: §VI · p. 18
  • The leakage estimate assumes joined materials are intact and homogeneous, although fabrication defects can add leak paths; barrier films, NEG coatings, and stress-relief annealing can reduce these effects.Source: §VI · p. 18
  • The silicon outgassing model is reliable only to about an order of magnitude at best, and material-specific experimental studies are needed.Source: §VI · p. 18
  • Microfabrication can yield sealed microlitre chambers maintaining UHV for at least 1000 days with passive pumping elements.Source: §VI · p. 18
  • The design uses a single laser beam, patterned reflectors, and an atom source; its vacuum depends on very low-leakage bonds, low-outgassing materials, and negligible noble-gas release.Source: §VI · p. 18
Loading 1405.3148v1…