Physical Layer Deep Dive

Waveforms · Modulation Order · Channel Bandwidth

Giga-MIMO · Spectrum Bands · Channel Coding · Numerology

IMT-2030 / 3GPP Release 21 | June 2026

1. Executive Summary

6G, formally designated IMT-2030 by the ITU, represents the next epoch of mobile communications targeting commercial availability around 2030. The 3GPP has initiated study items under Release 20 (started June 2025) with normative specifications expected in Release 21. By the end of 2026, the physical layer foundations—including waveform, coding, and modulation—should be substantially defined.

6G Key Performance Indicators (IMT-2030 Targets) Peak Data Rate: Up to 1 Tbps (sub-THz bands) | 50–200 Gbps (wide-area) User Experience Rate: 1 Gbps everywhere Latency (User Plane): < 0.1 ms (URLLC scenarios) Spectral Efficiency: 3× improvement over 5G NR Connection Density: Up to 10,000,000 devices/km² Energy Efficiency: 100× better than 5G Sensing Accuracy: 20× improvement over 5G

2. Waveform Candidates for 6G

Waveform selection is the cornerstone of any new generation’s physical layer. For 6G, OFDM is expected to remain the dominant foundational structure for a third consecutive generation, but several enhanced and alternative candidates are being actively studied.

2.1 OFDM — The Continuing Foundation

Orthogonal Frequency Division Multiplexing has served two generations (4G LTE and 5G NR) reliably. For 6G, it is expected to persist as the baseline waveform, with key innovations layered on top for higher efficiency, improved energy performance, and support for high-Doppler environments.

2.2 OTFS — Leading Challenger

Orthogonal Time Frequency Space (OTFS) modulates data in the delay-Doppler domain rather than the traditional time-frequency domain used by OFDM. This gives it fundamental advantages for high-mobility scenarios (V2X, high-speed rail, aerial communications):

• Resilient to high Doppler shifts — ideal for 6G mobility use cases
• Full channel diversity exploitation without requiring adaptive transmitters
• Excellent performance for short packets and large antenna arrays
• Simplifies channel estimation under doubly-dispersive (time & frequency varying) channels
  

2.3 AFDM — Affine Frequency Division Multiplexing

AFDM is a newer contender that achieves full diversity by spreading data symbols across the entire time-frequency domain under doubly dispersive channels. It is being seriously studied alongside OTFS for 6G, particularly for ISAC (Integrated Sensing and Communication) waveform design.

2.4 Other Candidate Waveforms

Waveform	Key Strength	Best Use Case	Status
OFDM (Enhanced)	Backward compatibility, mature ecosystem	Broadband, FR1 & FR2 continuation	Baseline — certain
DFT-S-OFDM	Lower PAPR, uplink efficiency	UE uplink (power-limited devices)	Evolved from 5G NR
OTFS	Doppler resilience, delay-Doppler domain	High-mobility V2X, aerial, satellite	Strong candidate
AFDM	Full diversity in doubly dispersive channels	ISAC, high-speed sensing+comms	Active research
Index Modulation	Spectral & energy efficiency	IoT, low-power devices	Research stage
AI-native Waveforms	AI/ML-optimized physical layer	Adaptive, self-optimizing networks	Early exploration

3. Modulation Order in 6G

5G NR peaked at 256-QAM (8 bits/symbol) in Release 15/16. For 6G, the modulation order is expected to push significantly higher, with the target dependent on the frequency band and channel conditions:

3.1 Expected QAM Orders by Band

Band / Range	Expected Max QAM	Bits/Symbol	Notes
Sub-6 GHz (FR1 continuation)	1024-QAM	10	High SNR indoor/dense urban scenarios
FR3 Upper Mid-Band (7–24 GHz)	1024-QAM	10	Qualcomm/Ericsson prototype validated
mmWave (24–100 GHz)	256-QAM – 1024-QAM	8–10	SNR constrained by path loss
Sub-THz (100–300 GHz)	64-QAM – 256-QAM	6–8	Hardware nonlinearity limits higher orders
THz (>300 GHz)	QPSK – 64-QAM	2–6	Extremely SNR-sensitive at range

Research has demonstrated 4096-QAM (12 bits/symbol) in terahertz photonic-assisted systems under ideal conditions. While 4096-QAM and even 8192-QAM have been demonstrated in lab THz systems, practical deployments for wide-area coverage are expected to rely on 1024-QAM as the practical ceiling for 6G commercial systems.

3.2 Constellation Shaping

Beyond raw QAM order, 6G is expected to introduce probabilistic constellation shaping (PCS), which allows non-uniform distribution of QAM points closer to a Gaussian distribution, providing up to 1.5 dB SNR gain without changing modulation order. This is aligned with Qualcomm’s work on ‘constellation shaping’ as part of 6G PHY evolution.

4. Channel Bandwidth in 6G

Channel bandwidth is being dramatically expanded in 6G. The Ericsson–Qualcomm MWC 2026 prototype demonstrated a 400 MHz component carrier at 30 kHz subcarrier spacing as an agreed study item for 3GPP Release 20. Sub-THz bands will enable channel bandwidths of 1–30 GHz per carrier.

Frequency Range	Expected Bandwidth per Carrier	Typical Aggregated BW	Standard Reference
Sub-6 GHz (FR1)	100 MHz (inherited)	Up to 400–800 MHz CA	5G-Advanced evolution
FR3: 7–15 GHz	200–400 MHz	Up to 1–2 GHz CA	3GPP Rel.20 study item (400 MHz @ 30 kHz SCS)
FR3: 15–24 GHz	400 MHz – 1 GHz	Multi-GHz aggregation	WRC-27 study bands
mmWave (24–52 GHz)	400 MHz – 2 GHz	Up to 4 GHz CA	Evolved from 5G NR FR2
Sub-THz (90–300 GHz)	2 GHz – 30 GHz	Up to 50+ GHz	IEEE 802.15.3d up to 69.12 GHz

The sub-THz range between 100–300 GHz contains approximately 97 GHz of allocated spectrum across non-contiguous windows, making it the primary enabler for >100 Gbps peak throughputs in 6G.

5. MIMO Evolution: Massive → Giga → XL-MIMO

MIMO technology evolves dramatically in 6G. While 5G NR used Massive MIMO (typically 64–256 antenna elements), 6G introduces Gigantic MIMO (gMIMO) and Extra-Large MIMO (XL-MIMO) with hundreds to thousands of antennas, redefining how spatial resources are exploited.

5.1 Evolution Roadmap

Generation	MIMO Type	Antenna Count	Key Characteristics
4G LTE	MIMO / MU-MIMO	2–8 elements	Spatial multiplexing, beamforming basics
5G NR	Massive MIMO	32–256 elements	3D beamforming, FDD/TDD, CSI-RS feedback
5G Advanced	Enhanced mMIMO	256–512 elements	AI-assisted beamforming, improved CSI
6G (FR3/mid-band)	Giga-MIMO	512–1024+ elements	Near-field regime, spatial multiplexing ×10
6G (sub-THz)	XL-MIMO / Holographic MIMO	1000–10,000+ elements	Holographic beamforming, near-field communications

5.2 Near-Field vs Far-Field — The Critical Distinction

As antenna arrays grow to XL-MIMO scales, the Rayleigh distance (the boundary between near-field and far-field) moves into the hundreds of meters. This fundamentally changes the propagation model: devices inside the near-field experience spherical wavefronts rather than planar ones, enabling new capabilities like near-field beamforming that focuses energy at a specific distance and angle simultaneously.

5.3 Giga-MIMO: Qualcomm’s Vision for FR3

Qualcomm has specifically highlighted Giga-MIMO as a key 6G innovation for the 7–8 GHz upper mid-band (FR3), with key design elements including:

• 400 MHz component carrier bandwidth with 30 kHz subcarrier spacing
• Advanced DFT-S with MIMO for uplink gain over 5G single-layer DFT-S
• Evolved LDPC + constellation shaping + MIMO mapping
• AI-ML-enhanced CSI feedback to support extreme beamforming precision
• Improved cell-edge coverage with 4-Tx/Rx antenna devices
  

6. Spectrum Strategy: The 6G ‘Golden Band’

6G will use a multi-layered spectrum strategy spanning from sub-6 GHz to sub-THz. However, the industry is converging on a clear hierarchy of importance, with the upper mid-band (FR3) at 7–15 GHz emerging as the ‘golden band’ for 6G wide-area deployments.

6.1 The Multi-Layer Spectrum Architecture

Layer	Frequency Range	Role	Key Characteristics
Coverage Layer	Sub-3 GHz (reused)	Ubiquitous coverage, IoT, LPWA	Inherited from 5G; backward compatible
Golden Band (FR3)	7–15 GHz (cmWave)	Primary 6G capacity + coverage	Balance of coverage, capacity, and BW
Capacity Layer	15–52 GHz (mmWave)	Urban hotspots, dense areas	Wide BW, moderate coverage
Extreme Capacity	90–300 GHz (sub-THz)	Fixed wireless, indoor Tbps links	Vast spectrum, very short range

6.2 Why 7–15 GHz Is the Golden Band

The cmWave range at 7–15 GHz has attracted massive industry and regulatory attention for the following reasons:

• WRC-23 designated 6.425–7.125 GHz for IMT globally; WRC-27 will study 7.125–8.4 GHz and 14.8–15.35 GHz
• 3GPP Release 19 initiated FR3 (7–24 GHz) study items in December 2023
• Coverage advantage over mmWave: propagation is far better, enabling macro-cell deployments
• Bandwidth availability: 200–400 MHz per carrier — far more than sub-6 GHz
• Ericsson analysis: 6G will need ~3 GHz of additional wide-area spectrum, best served by cmWave
• Qualcomm’s Giga-MIMO demonstrations centered on the 6–8 GHz range
  

6.3 Sub-THz: The Speed Record Band

For indoor and short-range deployments, the sub-THz band (90–300 GHz) offers the path to >100 Gbps and eventually 1 Tbps per user. Key sub-bands under study:

• D-band: 110–170 GHz — favoured for backhaul and fixed wireless
  
• H-band: 220–330 GHz — demonstrated 100 Gbps with 30 GHz occupied BW in Keysight testbeds
  
• IEEE 802.15.3d already defines PHY modes up to 100 Gbps at 252–325 GHz
  

7. Channel Coding in 6G

5G NR used LDPC codes for data channels and Polar codes for control channels. For 6G, the industry is evaluating whether to evolve these codes, unify them, or introduce AI/ML-based approaches.

7.1 Evolved LDPC — Most Likely Baseline

LDPC codes remain the frontrunner for 6G data channel coding. Qualcomm explicitly mentions evolved 6G LDPC in their 6G PHY roadmap. Key improvements expected:

• Enhanced belief-propagation decoding for lower latency
  
• Improved performance in short block-length regimes (critical for 0.1 ms latency targets)
  
• LDPC combined with constellation shaping for joint modulation-coding gain
  
• Better support for HARQ with incremental redundancy
  

7.2 Polar Codes — Continued Role in Control

Polar codes are expected to continue for control channels, potentially with improvements in list decoding and CRC-aided polar decoding. Some research proposes unifying LDPC and Polar into hybrid schemes.

7.3 GLDPC-PC: Generalized LDPC with Polar Component Codes

Generalized LDPC with Polar component codes (GLDPC-PC) is an actively researched candidate for 6G. It combines the high-throughput, low-latency advantages of LDPC with the near-capacity performance of polar codes, and supports parallel decoding. It also shows promise for MIMO iterative detection and decoding joint optimization.

7.4 AI/ML-Based Channel Coding

Research is underway on AI/ML-assisted channel coding for 6G. This includes:

• Neural network decoders that approach theoretical capacity limits
  
• AI-optimized code design for specific channel conditions
  
• Semantic coding: transmitting meaning/intent rather than raw bits
  
• Joint source-channel coding enabled by deep learning
  

7.5 Summary Comparison

Coding Scheme	Use in 5G NR	Expected Role in 6G	Key Benefit
LDPC	Data channels (eMBB)	Primary data channel code (evolved)	High throughput, parallel decoding
Polar Codes	Control channels, PBCH	Control channels (improved)	Near-capacity for short blocks
GLDPC-PC	Not in 5G	Research candidate, potential adoption	Best of LDPC + Polar combined
Turbo Codes	Data channels (4G LTE)	Retired in 6G	Outperformed by LDPC/Polar
AI-Native Codes	Not in 5G	Likely for semantic comms, ISAC	Adaptive, environment-aware coding

8. Numerology in 6G: Evolution of the 5G Framework

5G NR introduced a flexible numerology framework (μ = 0–4) where subcarrier spacing (SCS) scales as 15 × 2^μ kHz, enabling the same waveform to operate across sub-1 GHz to mmWave bands. For 6G, this framework is being evolved—not abandoned.

8.1 5G NR Numerology — Baseline Reference

μ	SCS (kHz)	Slot Duration (ms)	Typical Band Use
0	15	1.0	Sub-3 GHz, eMBB
1	30	0.5	3–6 GHz, FR1
2	60	0.25	6 GHz, unlicensed
3	120	0.125	mmWave FR2
4	240	0.0625	mmWave FR2 reference signals

8.2 6G Numerology: What Changes

Ericsson–Qualcomm’s MWC 2026 prototype validated 400 MHz component carrier at 30 kHz SCS as a concrete 6G study item. Nokia’s white paper on 6G PHY proposes:

• Retaining 5G-defined subcarrier spacings to ensure seamless coexistence in Multi-RAT Spectrum Sharing (MRSS)
  
• Reducing number of SCS options per band/range to simplify system complexity
  
• Adding new numerologies for sub-THz bands (very wide SCS for large channel bandwidths and hardware phase noise tolerance)
  
• Setting 3 MHz as the minimum UE bandwidth for native LPWA support
  
• Forward-compatible numerology design with min/max bandwidth flexibility across releases
  

8.3 New Sub-THz Numerology Requirements

At sub-THz frequencies, phase noise from local oscillators becomes a dominant impairment. This drives the need for wider subcarrier spacings (potentially 480 kHz or 960 kHz) at sub-THz bands to maintain inter-carrier orthogonality. The expected 6G numerology extensions:

New μ (proposed)	SCS (kHz)	Target Band	Rationale
5 (proposed)	480	100–200 GHz sub-THz	Phase noise compensation
6 (proposed)	960	200–300 GHz sub-THz	Very large BW, phase noise
FR3 aligned	30–120	7–24 GHz FR3	Validated at 30 kHz for 400 MHz CC

9. Other Key 6G Physical Layer Changes

9.1 AI-Native Physical Layer

Unlike 5G where AI/ML was bolt-on, 6G embeds AI natively at the PHY layer. This includes AI-driven channel estimation, CSI compression (continuing from 3GPP Release 19), beam prediction, waveform optimization, and even end-to-end learned communication systems.

9.2 Integrated Sensing and Communication (ISAC)

6G unifies sensing and communication in a single waveform and hardware platform. ISAC is one of the six official IMT-2030 usage scenarios. This changes waveform requirements (must serve dual purposes), introduces new reference signals for sensing, and has implications for sub-THz waveform design and MIMO beam management.

9.3 Reconfigurable Intelligent Surfaces (RIS)

RIS (also called Intelligent Reflective Surfaces) are passive/semi-active meta-surfaces that reshape the radio propagation environment. In 6G, RIS is expected to be standardized for channel manipulation, coverage extension at mmWave/sub-THz, and energy efficiency improvements. RIS interacts deeply with waveform and MIMO design.

9.4 Non-Terrestrial Networks (NTN)

6G expands NTN support to cover LEO, MEO, GEO satellites, HAPS (High Altitude Platform Stations), and UAVs as first-class citizens in the standard. This requires numerology and waveform robustness to high Doppler (LEO moves at ~7.8 km/s), long propagation delays, and hybrid terrestrial/non-terrestrial handover.

9.5 Full Duplex and New Duplexing

6G is expected to introduce in-band full duplex (IBFD) — simultaneous transmission and reception on the same frequency — made feasible by advances in self-interference cancellation. This could double spectral efficiency in suitable scenarios and fundamentally alter the FDD/TDD duplexing paradigm.

10. Standardization Timeline

Milestone	Date	Body	Details
IMT-2030 Vision Published	June 2023	ITU-R WP5D	6 usage scenarios, KPI framework defined
3GPP Release 20 Study Start	June 2025	3GPP RAN	6G Study Items started; FR3, sub-THz, AI-native
Ericsson–Qualcomm Prototype	Feb 2026 (MWC)	Industry	400 MHz CC @ 30 kHz SCS validated in lab
IMT-2030 Requirements Final	2026	ITU-R WP5D	Technical performance requirements locked
3GPP Release 21 Start	~2027	3GPP	Normative 6G specifications begin
First 6G Specs Available	~End of 2028	3GPP	Release 21 complete; ecosystem builds
Commercial 6G Launch	~2030	Global	Initial markets; Olympics LA 2028 demos

11. Quick Reference Summary

Parameter	5G NR	6G (Expected)
Primary Waveform	CP-OFDM / DFT-S-OFDM	CP-OFDM + OTFS / AFDM (band-dependent)
Peak Modulation Order	256-QAM (8 bits/sym)	1024-QAM practical; 4096-QAM research
Max Channel Bandwidth	400 MHz (mmWave)	400 MHz (FR3) → 30 GHz (sub-THz)
MIMO Evolution	Massive MIMO (up to 256)	Giga-MIMO / XL-MIMO (512 → 10,000+)
Channel Coding — Data	LDPC	Evolved LDPC + constellation shaping
Channel Coding — Control	Polar Codes	Evolved Polar / GLDPC-PC hybrid
Key New Coding	—	AI/ML semantic coding, GLDPC-PC
Numerology (SCS)	15/30/60/120/240 kHz	Retain + add 480/960 kHz for sub-THz
Golden Band	3.5 GHz (sub-6) / 26 GHz (mmWave)	7–15 GHz cmWave (FR3)
Sub-THz Band	Not standardized	90–300 GHz (key extreme-rate band)
AI in PHY	Add-on (Rel.17/18 CSI)	Native — waveform, coding, beamforming
Sensing Integration	Radar study items (Rel.18)	ISAC as core IMT-2030 use case
First Commercial Use	2019	~2030