Atomic Clock Technology

From the cesium beam clocks that defined the SI second to optical lattice clocks that won't lose a second in the age of the universe — a complete technical guide to every major type of atomic clock.

Quick Comparison: All Clock Types

TechnologyStabilityPowerSizeCostBest For
Cesium-133 Atomic Clock10⁻¹⁴ per day30 W typicalRack unit (bench top)$50,000–$100,000Gold Standard
Rubidium Clock10⁻¹¹ per day (cell-based)5–20 WModule / rack card$500–$5,000Compact & Affordable
Optical Lattice Clock10⁻¹⁸ per day500 W+ (laser systems)Optical table (2×3 m)$2M–$10MMost Accurate Ever Built
Cesium Fountain Clock10⁻¹⁶ per day200 W (incl. lasers)2-meter stainless steel tower$3M–$10MPrimary Standards
Chip-Scale Atomic Clock10⁻¹⁰ per day120 mW17 × 35 × 11 mm (Microsemi SA.45s)$1,500–$3,000Miniaturized
Hydrogen Maser10⁻¹⁵ per hour (unmatched short-term)60 WRefrigerator-sized cabinet$200,000–$300,000Best Short-Term Stability
Nuclear Clock10⁻¹⁹ to 10⁻²⁰ (predicted)TBDLab bench (early prototype)Multi-million (research)Next Frontier

Cesium-133 Atomic Clock

Gold Standard
Microwave hyperfine transition of Cs-133 ground state

Overview

The cesium-133 atomic clock is the foundation of modern timekeeping. The SI second is defined by exactly 9,192,631,770 oscillations of the cesium-133 hyperfine transition between two ground-state energy levels. Cesium beam clocks achieve excellent long-term accuracy and are the workhorse of national timekeeping laboratories.

How It Works

Cesium atoms are heated in an oven and collimated into a beam. A magnetic state selector separates atoms in the desired hyperfine state. The beam passes through a microwave cavity tuned to the resonance frequency. A second state selector and detector measure how many atoms changed state — maximum transitions indicate exact resonance. A feedback loop locks the local oscillator to this frequency.

Applications

National metrology institutesPrimary UTC contributorsTelecommunications master clocksScientific research

Known Implementations

Symmetricom 5071A, PTB CS1/CS2, USNO cesium ensemble

Frequency
9,192,631,770 Hz (9.19 GHz)
Stability
10⁻¹⁴ per day
Accuracy
10⁻¹⁴ relative
Size
Rack unit (bench top)
Power
30 W typical
Cost
$50,000–$100,000

Rubidium Clock

Compact & Affordable
Optical pumping + microwave of Rb-87 at 6.835 GHz

Overview

Rubidium clocks use a vapor cell filled with Rb-87 gas. A laser or lamp optically pumps atoms into a specific hyperfine state, and a microwave field induces transitions. They are far more compact and affordable than cesium clocks, making them the dominant choice for GPS satellites (with ~90 on orbit), cellular base stations, and portable timing systems.

How It Works

An Rb-87 vapor cell is illuminated by an Rb lamp filtered to excite only atoms in the ground state. A microwave cavity at 6.835 GHz drives hyperfine transitions. The transmitted light intensity dips at resonance (absorbed by more atoms). A VCXO is phase-locked to this dip. The cell must be temperature-stabilized for long-term accuracy.

Applications

GPS satellitesTelecom base stationsMilitary portable timingTest and measurement

Known Implementations

GPS Block IIR satellites, Microsemi 8040, Stanford Research FS725

Frequency
6,834,682,610.9 Hz (6.83 GHz)
Stability
10⁻¹¹ per day (cell-based)
Accuracy
10⁻¹¹ relative
Size
Module / rack card
Power
5–20 W
Cost
$500–$5,000

Optical Lattice Clock

Most Accurate Ever Built
Laser-cooled Sr-87 or Yb-171 atoms at optical frequencies

Overview

Optical lattice clocks are the most accurate timekeeping instruments ever built. They trap thousands of atoms (strontium-87 or ytterbium-171) in a laser lattice and interrogate them at optical frequencies (~430 THz) — 100,000× higher than cesium. The higher frequency means each "tick" is much shorter, enabling far finer time resolution. These clocks would not lose a second in 15 billion years — longer than the age of the universe.

How It Works

Atoms are laser-cooled to near absolute zero (1 μK), trapped in a 1D optical lattice at the "magic wavelength" where lattice-induced light shifts cancel. A clock laser interrogates the ultra-narrow Sr clock transition. An optical frequency comb (Nobel Prize 2005) converts the optical frequency to a usable microwave output. Systematic effects (gravity, blackbody radiation, collisions) are meticulously characterized.

Applications

Frontier metrology researchTests of relativity and fundamental constantsGeodesy (measuring gravity via clock height difference)Future redefinition of the SI second

Known Implementations

NIST Sr clock, SYRTE Paris lattice, PTB Yb+ ion clock

Frequency
429,228,004,229,873 Hz (429 THz, Sr)
Stability
10⁻¹⁸ per day
Accuracy
2 × 10⁻¹⁸
Size
Optical table (2×3 m)
Power
500 W+ (laser systems)
Cost
$2M–$10M

Cesium Fountain Clock

Primary Standards
Laser-cooled Cs-133 atoms tossed upward in vacuum

Overview

Cesium fountain clocks are the global primary standards for UTC. Unlike beam clocks, fountain clocks laser-cool a ball of ~10⁷ cesium atoms near absolute zero (1 μK), then toss them 1 meter upward in a vacuum. The slow-moving atoms spend ~1 second in the microwave cavity — far longer than beam clocks — allowing much sharper frequency discrimination via the Ramsey interrogation method.

How It Works

Six laser beams slow cesium atoms from 100 m/s to under 1 cm/s. The atom ball is tossed upward through a microwave cavity. Atoms travel through the cavity twice (up and down). The Ramsey method uses these two interactions for interference-based frequency discrimination 1000× sharper than a single-cavity beam clock. Gravity and thermal environment are carefully controlled.

Applications

NIST-F2 (US primary standard)PTB-CSF2 (Germany)SYRTE-FO2 (France)NICT-CsF2 (Japan)TAI computation

Known Implementations

NIST-F2, PTB CSF1/CSF2, SYRTE FO1/FO2, NICT-CsF2

Frequency
9,192,631,770 Hz (defined SI second)
Stability
10⁻¹⁶ per day
Accuracy
8 × 10⁻¹⁶ (NIST-F2)
Size
2-meter stainless steel tower
Power
200 W (incl. lasers)
Cost
$3M–$10M

Chip-Scale Atomic Clock (CSAC)

Miniaturized
Miniaturized Rb/Cs cell with MEMS + VCXO loop

Overview

Chip-Scale Atomic Clocks (CSACs) pack atomic clock physics into a matchbook-sized package consuming under 150 mW. They use MEMS-fabricated miniaturized vapor cells and VCSEL lasers instead of conventional bulky components. While their accuracy is 10,000× worse than cesium beam clocks, they are dramatically better than quartz oscillators and can operate for days on battery power — critical for GPS-denied navigation.

How It Works

A VCSEL (vertical cavity surface-emitting laser) is locked to the Rb-87 or Cs-133 hyperfine transition using coherent population trapping (CPT) — a technique where two coherent light fields simultaneously drive the microwave transition, producing a narrow dark resonance. A MEMS atomic vapor cell contains the reference atoms. A low-power VCXO is phase-locked to this resonance.

Applications

Military navigation (GPS-denied)Submarine systemsPortable test equipmentSmart grid timingDrone swarm coordination

Known Implementations

Microsemi SA.45s CSAC, Teledyne CSAC, AccuBeat AR125

Frequency
~6.8 GHz (Rb) or ~9.2 GHz (Cs)
Stability
10⁻¹⁰ per day
Accuracy
10⁻¹⁰ relative
Size
17 × 35 × 11 mm (Microsemi SA.45s)
Power
120 mW
Cost
$1,500–$3,000

Hydrogen Maser

Best Short-Term Stability
Stimulated emission of H atoms at 1420.405 MHz

Overview

Active hydrogen masers are unrivaled for short-term frequency stability — outperforming even cesium fountains over timescales of seconds to hours. A hydrogen maser amplifies microwave radiation at 1420 MHz by stimulated emission from spin-flipping hydrogen atoms. They are the preferred timing reference for VLBI radio telescope networks (EVN, VLBA, ALMA) where picosecond synchronization between widely separated antennas is required.

How It Works

Hydrogen molecules are dissociated into atoms by a RF discharge. A magnetic state selector passes only high-energy (spin-up) atoms into a Teflon-coated quartz storage bulb. Atoms spontaneously emit 1420 MHz photons as they relax, building up microwave power in a resonant cavity. A low-noise amplifier extracts this signal and phase-locks a quartz oscillator to it.

Applications

VLBI radio telescopesDeep space navigation (NASA/ESA)Gravitational wave detector timingNational metrology labs

Known Implementations

Symmetricom MHM-2010, PTB H-masers, Effelsberg 100m telescope, LISA Pathfinder

Frequency
1,420,405,751.768 Hz (1.42 GHz)
Stability
10⁻¹⁵ per hour (unmatched short-term)
Accuracy
10⁻¹³ long-term
Size
Refrigerator-sized cabinet
Power
60 W
Cost
$200,000–$300,000

Nuclear Clock (Thorium-229)

Next Frontier
Nuclear isomeric transition in Th-229 at 8.4 eV (UV)

Overview

Nuclear clocks use transitions in atomic nuclei rather than electron shells. The thorium-229 nucleus has an extraordinarily low-energy nuclear isomeric transition at ~8.4 eV (vacuum UV light) — 100 million times less energetic than a typical nuclear transition. This makes it the only known nuclear transition accessible to lasers. Nuclear clock transitions are 1,000× less sensitive to electromagnetic perturbations than electronic transitions, promising 10⁻¹⁹ accuracy. The first thorium nuclear clock prototype was demonstrated in 2024.

How It Works

A VUV laser at ~148 nm excites the Th-229 nuclear isomeric transition. The nucleus is embedded in a crystal lattice (CaF₂ or ThF₄) that shifts the transition to accessible UV wavelengths. The clock transition is insensitive to stray fields because the nuclear magnetic moment is ~1000× smaller than atomic magnetic moments. An optical frequency comb connects the UV frequency to usable microwave output.

Applications

Fundamental physics testsDark matter detectionVariation of fundamental constantsNext-generation GPS (2030s+)

Known Implementations

JILA thorium prototype (2024), PTB Th-229 research, CU Boulder / MIT experiments

Frequency
~2.0 × 10¹⁵ Hz (ultraviolet)
Stability
10⁻¹⁹ to 10⁻²⁰ (predicted)
Accuracy
Unknown — prototype only (2025)
Size
Lab bench (early prototype)
Power
TBD
Cost
Multi-million (research)