Terahertz Communication

Terahertz (THz) frequencies — 100 GHz to 10 THz — represent the next frontier in wireless communication.

Period2020s–2030s

The THz Gap

The electromagnetic spectrum between microwave (roughly 30 GHz) and infrared (roughly 10 THz) has been largely unexplored for communications. This region — spanning frequencies of 100 GHz to 10 THz — is called the THz gap because neither conventional microwave electronics nor optical photonics work well within it.

Below the gap, transistors and RF circuits generate and detect microwave signals efficiently. Above the gap, lasers and photodetectors handle infrared and optical frequencies. In between, electronic devices are too slow (transistor cutoff frequencies drop off rapidly above 300 GHz) and photonic devices are too energetic (photon energies at THz frequencies correspond to temperatures of 5–50 K, requiring cryogenic cooling for efficient generation).

The gap is closing. Quantum cascade lasers (QCLs),resonant tunneling diodes (RTDs), and graphene transistors are pushing electronic generation into the sub-THz range. Photomixing (beating two laser frequencies) generates THz radiation optically. These technologies are enabling the first practical THz communication systems.

Why THz Matters for 6G

Current 5G mmWave operates at 24–100 GHz, with channels up to 400 MHz wide. THz frequencies could support channels of 10–100 GHz, enabling data rates exceeding 1 Tbps — orders of magnitude beyond current 5G capabilities. The potential applications:

  • Tbps wireless backhaul: Replacing fiber-optic links for 5G/6G base station interconnection — especially valuable where trenching fiber is impractical (historic districts, water crossings, temporary venues)
  • Data center wireless cables: Replacing copper and fiber cables between server racks — eliminating cable management, enabling flexible rack layouts, and reducing latency
  • Ultra-high-resolution radar: THz wavelengths (0.3–3 mm) enable millimeter-scale imaging for security screening, industrial inspection, and autonomous vehicles
  • Device-to-device links: Ultra-short-range (1–10 m) Tbps connections between laptops, phones, and peripherals — replacing USB cables and Thunderbolt
  • Wireless projection displays: THz links could carry uncompressed 8K video (48 Gbps) wirelessly from a source to a display

Atmospheric Absorption: The Water Vapor Problem

THz propagation faces severe atmospheric absorption, primarily fromwater vapor (H₂O) rotational transitions. Specific frequencies experience strong absorption:

FrequencyAbsorption LineAttenuationUseable Band
118.75 GHzO₂ resonanceHighNarrow — avoided
183.31 GHzH₂O resonanceVery highOnly for short-range links
325 GHzH₂O resonanceExtremeAvoided for comms
220 GHzWindow between H₂O linesModerateUsable for 100+ m links
280–320 GHzWindowModerateUsable — promising for 6G
340–360 GHzWindowModerateUsable for indoor links
450 GHzH₂O resonanceHighShort-range only
670 GHzWindowModeratePotentially usable

Window frequenciesbetween strong absorption lines offer the best propagation for communication. The 220 GHz, 280–320 GHz, and 340–360 GHz windows are the most promising for 6G. At sea level, typical attenuation is 5–15 dB/km in these windows (compared to <1 dB/km for 5G mmWave). At altitude (aircraft, high-rise buildings), attenuation drops significantly because water vapor density decreases exponentially with altitude.

Semiconductor Technologies for THz

Silicon-based transistors cannot operate above ~300 GHz. THz electronics require specialized semiconductor technologies:

  • InP HEMT (Indium Phosphide High Electron Mobility Transistor): Highest frequency operation — fmax > 600 GHz. Used for sub-THz receivers and low-noise amplifiers. Expensive, low volume.
  • SiGe BiCMOS (Silicon-Germanium Bipolar CMOS): fmax ~300–500 GHz. Most practical for mass production — integrates RF and digital on the same chip. Leading 6G candidate.
  • GaN HEMT (Gallium Nitride): High power output at sub-THz frequencies — suitable for transmitters. Used in automotive radar (77 GHz) and being pushed toward 300 GHz.
  • Graphene FET: Theoretical fmax > 1 THz — graphene's electron mobility (200,000 cm²/V·s) far exceeds silicon. Experimental stage — difficult to fabricate reliably.
  • RTD (Resonant Tunneling Diode): Quantum-effect device that generates THz radiation through resonant tunneling. Demonstrated at 1.98 THz — the highest-frequency electronic oscillator. Low output power (~μW).
  • Quantum Cascade Laser (QCL): Intersubband semiconductor laser — generates THz optically. Requires cryogenic cooling for continuous-wave operation. Used in spectroscopy and imaging.

THz Modulation Approaches

Two fundamentally different approaches exist for THz modulation:

  • Electronics-based (direct modulation): The THz carrier is generated by an electronic oscillator (RTD, multiplier chain, or QCL) and directly modulated with data. Limited by transistor switching speed — achievable data rates scale with bandwidth.
  • Photonics-based (heterodyne): Two optical lasers with a frequency difference in the THz range are beat together on a photomixer. The difference frequency is the THz carrier, and one laser is modulated with data. Higher output power and wider tunability than electronics-based approaches.

For 6G communication, electronics-based approaches are preferred for integration with mass-produced semiconductor processes. Photonics-based approaches are preferred for laboratory demonstrations and long-range links requiring higher output power.

THz Link Budget: A Practical Example

Frequency: 300 GHz (λ = 1 mm)
TX power: 10 dBm (10 mW)
TX antenna gain: 30 dBi (15 cm dish, 60% efficiency)
Path loss (100 m): FSPL = 20·log₁₀(4π×100/0.001) = 132 dB
Atmospheric attenuation: 10 dB/km × 0.1 km = 1 dB
RX antenna gain: 30 dBi
─────────────────────────
Received power: 10 + 30 − 132 − 1 + 30 = −63 dBm
Noise floor (B = 10 GHz, NF = 10 dB): −94 + 10 = −84 dBm
SNR = −63 − (−84) = 21 dB
Shannon capacity: C = 10 × log₂(1 + 126) = 69 Gbps

At 100 m range with a 300 GHz link, 10 GHz bandwidth, and modest antenna gains, Shannon-limited capacity reaches ~70 Gbps. With 100 GHz bandwidth (available in the 300 GHz window), Tbps rates become feasible. In practice, achievable rates are 50–70% of the Shannon limit due to hardware imperfections, channel estimation overhead, and coding limitations.

Technical Challenges

  • Path loss: Free-space path loss increases with frequency² — a 300 GHz link has 20 dB more path loss than a 30 GHz link at the same range. Combined with atmospheric absorption, THz links are fundamentally short-range.
  • Blockage: THz waves cannot penetrate walls, foliage, or even paper. Human bodies cause 20–40 dB attenuation. Links require clear line-of-sight, limiting deployment to controlled environments.
  • Antenna size: At 300 GHz, a half-wavelength antenna is 0.5 mm — enabling massive arrays on small chips. A 1024-element phased array fits on a 4×4 mm die, enabling electronic beam steering with narrow beams (2–3°) to compensate for path loss.
  • Power generation: Current THz sources produce microwatts to milliwatts. A 300 GHz SiGe transmitter might produce 0 dBm (1 mW), while a 30 GHz 5G transmitter produces 20 dBm (100 mW). This 20 dB gap must be compensated with antenna gain.
  • PAPR (Peak-to-Average Power Ratio): High-order modulation (64-QAM, 256-QAM) has high PAPR, requiring highly linear power amplifiers — challenging at THz frequencies where amplifier linearity is poor.

Potential Applications

  • Data center interconnect: Wireless replacement for cables between server racks — eliminating cable management, enabling flexible rack layouts. Links of 1–5 m at 100+ Gbps.
  • Wireless backhaul: Urban point-to-point links at 100–500 m — replacing fiber in dense cities where trenching is impractical. E-band (70–80 GHz) already serves this role; THz extends it to higher capacities.
  • Device-to-device: Tbps links between laptops, phones, and peripherals — replacing USB cables and Thunderbolt for docking stations and external displays.
  • Security screening: THz imaging can see through clothing, packaging, and materials — detecting concealed weapons, drugs, and explosives. Already deployed in airports (L3 ProVision).
  • Spectroscopy: Many molecules have rotational absorption lines in the THz range — enabling chemical identification for pharmaceutical quality control, food safety, and environmental monitoring.
  • Autonomous vehicles: High-resolution THz radar could supplement lidar and camera sensors, providing 1 mm range resolution at 100 m.

Timeline

1896Oliver Lodge demonstrates wireless at 1.2 m wavelengthFirst laboratory demonstration of sub-millimeter wave transmission
1923Kannie & others observe sub-mm emissionsEarly experimental observations of radiation beyond 100 GHz
1964First maser at 1.4 mmBell Labs demonstrates coherent radiation at 214 GHz
1970First far-infrared laserHCN laser at 337 μm (890 GHz)
1991Quantum cascade laser proposedFaist, Capasso et al. propose QCL — enabling practical THz sources
1994Quantum cascade laser demonstratedFirst QCL at Bell Labs — mid-infrared, later extended to THz
2002First THz wireless link at >100 GHzDemonstrated wireless data transfer at sub-THz frequencies
2013100 Gbps wireless at 300 GHzNTT and other labs demonstrate sub-THz wireless links exceeding 100 Gbps
2017FCC opens 95 GHz – 3 THzFCC allocates new spectrum above 95 GHz for experimental use
2019University of Oulu 6G FlagshipFinland leads global 6G research targeting THz communications
2020Samsung 6G white paperSamsung outlines THz-based 6G vision — 1 Tbps peak, < 1 ms latency
2021140 GHz wireless link demonstratedBrunner et al. — 240 Gbps at 140 GHz over 100 m
2023WRC-23 identifies 6G bandsWorld Radiocommunication Conference identifies candidate THz bands for IMT-2030
20246G test networks in Japan, ChinaNTT DOCOMO and Chinese labs demonstrate THz prototypes
2030Expected 6G commercializationIndustry targets commercial THz wireless systems by 2030