Three material classes driving low-loss dielectric innovation

The low-loss dielectric materials landscape for 5G and 6G is dominated by three principal classes: polymer-based substrates (polyimide, liquid crystal polyester, and cyclo-olefin polymer), inorganic ceramic composites, and nano-structured porous composite materials. Each class addresses the same core challenge — simultaneously minimising the dielectric constant (Dk) and the dissipation factor (Df, or loss tangent) while preserving mechanical integrity, thermal stability, and moisture resistance — but via fundamentally different material strategies and across frequency regimes extending from sub-6 GHz through millimeter-wave D-band (110–170 GHz) and into the terahertz ranges targeted by 6G development.

The dataset examined for this analysis comprises more than a dozen directly relevant technical literature entries and patents spanning 2018–2023, with additional foundational references dating to the early 2000s. Key institutional contributors span Chinese academic institutions, European research groups, and industry players including Unimicron Technology Corporation and Kyocera. According to WIPO, patent filings in wireless communication materials have intensified significantly since 2018, consistent with the commercial 5G rollout timeline. The technical concerns documented across this literature converge on a single engineering imperative: achieving near-air-like permittivity in a substrate that can survive the thermal and mechanical demands of high-volume manufacturing.

Polymer substrates: from polyimide to cyclo-olefin

Polymer dielectrics have emerged as one of the most actively investigated material families for 5G and 6G substrates, driven by the need for flexible, processable materials with ultra-low loss characteristics. The central achievement in this space is the demonstration that non-fluorinated polyimide chemistry can reach record-low dielectric performance through molecular engineering alone — without the environmental concerns of fluorinated polymers and without the mechanical fragility of porous architectures.

Sun Yat-sen University researchers reported the design and synthesis of a novel amorphous polyimide designated TmBPPA, achieving a dielectric constant (k) of 2.23 and a loss tangent below 3.94 × 10⁻³ at 10⁴ Hz — claimed as the lowest among all non-fluorinated, non-porous polymers reported at that time. The design strategy exploited the secondary relaxation behaviors of polymer chains, specifically suppressing dipole reorientation mechanisms that contribute to dielectric loss. This approach, documented in 2019, represents a significant departure from the conventional assumption that fluorination is a prerequisite for ultra-low-k polymer performance.

“Conventional polyimide exhibits dielectric constants in the range of 3.0–4.0 and dielectric loss of approximately 0.02 — inadequate for the signal integrity demands of high-frequency 5G circuits.”

Polyimide more broadly remains a critical interlayer dielectric in 5G-era advanced packaging, particularly wafer-level fan-out packaging. As reviewed by Shenyang University of Chemical Technology (2023), conventional polyimide’s Dk of 3.0–4.0 and Df of approximately 0.02 are insufficient for next-generation signal integrity, and further reductions require molecular engineering, porosity introduction, or composite formulation. The gap between where conventional PI sits and where 5G mmWave demands it to be defines much of the active research agenda in this field.

Liquid crystal polyesters (LCP) address a different part of the frequency spectrum. Researchers at the University of Electronic Science and Technology of China documented LCP’s outstanding combination of good thermal stability, low water absorption, and stable dielectric constant and loss tangent across millimeter-wave frequency ranges (2020). These properties position LCP as an ideal high-performance substrate and packaging material for both sub-6 GHz and mmWave 5G antenna modules. The stability of LCP’s dielectric properties across frequency — a property many polymer systems lack — is particularly valued for antenna designs that must maintain consistent radiation characteristics.

Cyclo-olefin polymers (COP) are gaining ground for millimeter-wave filter applications. Researchers at UBO/Lab-STICC CNRS demonstrated narrow-band substrate integrated waveguide (SIW) bandpass filters using COP materials with a loss tangent as low as 7.5 × 10⁻⁴ at 40 GHz for Zeon’s RS420-LDS grade, fabricated using a laser direct structuring (LDS) process (2018). This loss tangent value — more than an order of magnitude lower than conventional polyimide — makes COP a compelling candidate for high-Q mmWave front-end components. The LDS fabrication route also enables three-dimensional metallization of complex cavity geometries, pointing toward a miniaturization pathway for mmWave modules. Standards bodies including IEEE have documented the increasing importance of low-loss substrate selection in mmWave circuit design.

For prototyping and additive manufacturing contexts, Czech University of Life Sciences Prague evaluated common 3D-printable plastics (PLA, PET-G, ABS, ASA) over the 1–100 MHz frequency range, finding relative permittivities between 2.88 and 3.48 and loss factors from 0.03 to 4.31%. PLA (colorless) and ABS were identified as more suitable for RF applications based on lower loss factor values and frequency dependences (2022). While these values do not approach the performance of LCP or COP at mmWave frequencies, 3D-printable RF substrates represent an accessible prototyping pathway for early-stage hardware development.

Nano-composites and ceramics: engineering air into the substrate

Porous nano-composite structures represent the frontier of ultra-low permittivity substrate development specifically targeted at future 6G frequency bands, where the design philosophy of maximising air content — since air possesses the ideal permittivity of 1 — is a recurring theme across both polymer and ceramic composite systems.

Researchers at the University of Oulu fabricated a porous composite of hollow silica nanoshells and cellulose nanofibers using a template-assisted Stöber process, in which polystyrene nanosphere cores were coated with amorphous SiO₂ shells and subsequently burned off to yield hollow spheres. The resulting composite, with a mass density of just 0.19 ± 0.02 g/cm³, was demonstrated as a viable candidate for 6G dielectric substrates (2020). The University of Oulu’s work, published in 2020, represents one of the first published material studies explicitly framed around 6G frequency requirements, predating most comparable work globally. Research on next-generation wireless materials is also tracked by bodies such as ITU, which oversees 6G spectrum and standardisation discussions.

Low-density composite materials combining microwave-transmissive polymer matrices with particulate fillers have also been patented for microwave antenna and lens applications. A 3M Innovative Properties Company patent (2012, JP) disclosed a composite with a dielectric constant tunable between approximately 1.2 and 100 and a microwave loss tangent no greater than 0.10 at 1 GHz, achieved by controlling the volume fraction and coating thickness of electrically conductive filler particles with major dimensions below 0.5 mm. The system allows filler volume fraction to increase while decreasing composite density — an unusual design inversion enabled by the hollow or low-density nature of the filler particles.

Inorganic ceramic composite systems continue to find application in microwave resonator and filter components. Cold-sintered (Bi₀.₉₅Li₀.₀₅)(V₀.₉Mo₀.₁)O₄–Na₂Mo₂O₇ composites were shown to achieve near-zero temperature coefficient of resonant frequency (TCF) with a relative permittivity around 40 and a Microwave Quality Factor (Qf) of approximately 4,000 GHz, as reported by Xi’an Jiaotong University (2019). Critically, cold sintering operates at approximately 150°C — far below the conventional ceramic sintering temperatures of 800–1,200°C — offering a low-energy manufacturing pathway for temperature-stable filter and antenna components in 5G base station infrastructure. Research published through Nature journals has documented cold sintering as one of the most significant ceramic processing innovations of the past decade.

Measurement under pressure: characterizing dielectrics at mmWave and THz

Accurate measurement of Dk and Df at high frequencies is both technically challenging and commercially critical — discrepancies between vendor datasheets and actual board-level performance can compromise signal integrity at mmWave frequencies, with direct consequences for link budget and system qualification. The characterization methods landscape has evolved substantially to meet the demands of 5G and 6G material qualification.

Unimicron Technology Corporation employed the Fabry–Perot open resonator (FPOR) technique to characterize three commercially sourced low-loss dielectric materials, cross-validated against a fabricated coplanar waveguide with ground (CPWG) test vehicle measured by time-domain reflectometry (TDR) and vector network analysis (VNA) (2022). The study highlighted significant differences between vendor-provided Dk and Df values and those measured on actual PCB structures using real cross-section geometries. This finding has direct implications for high-frequency PCB design and qualification: engineers cannot rely on datasheet values alone when designing for mmWave performance.

“Vendor-provided Dk and Df values diverge meaningfully from PCB-level measurements — underscoring the necessity of Fabry–Perot open resonator and TDR-based qualification at the substrate fabrication stage.”

For anisotropic materials — including many polymer laminates used in multilayer PCB constructions — resonant cavity methods provide high accuracy but require careful interference mode suppression. Researchers at Northwestern Polytechnical University designed a multi-mode split rectangular cavity operating from 300 to 1,000 MHz to characterize low-loss anisotropic dielectric materials, validating the approach on polytetrafluoroethylene (PTFE) (2022). The method addresses the directional dependence of permittivity in laminated structures, which is often overlooked in single-axis characterization approaches.

At terahertz frequencies — directly relevant to 6G systems envisioned to operate above 100 GHz — conventional analytical and iterative numerical methods for complex refractive index extraction from THz time-domain spectroscopy (THz-TDS) data suffer from sensitivity to initial conditions and convergence problems. Researchers at Huazhong University of Science and Technology proposed a 4-layer neural network model validated on three materials spanning a wide range of dispersions and thicknesses (TPX, z-cut crystal quartz, and 6H SiC), demonstrating smaller extraction errors than traditional iterative algorithms (2022). This represents a methodological innovation of direct relevance to the material qualification pipeline for 6G hardware development, and reflects a broader trend of machine learning integration into materials characterization documented by bodies such as NIST.

At D-band frequencies (110–170 GHz) specifically planned for B5G and 6G outdoor-to-indoor coverage scenarios, penetration loss through common building and environmental materials becomes a critical system-level constraint. Researchers at Fudan University investigated penetration losses through vegetation, planks, glass, and slate at D-band and found that thick materials present significant blocking challenges, complicating macro-station indoor coverage strategies for future B5G mobile communication (2022). This result establishes a system-level constraint that dielectric substrate designers must account for when engineering low-loss propagation paths and antenna window materials.

Institutional landscape and the five innovation trends reshaping the field

Analysis of the sourced literature and patent data reveals a concentration of innovation activity in Chinese academic institutions, European research groups, and a small number of industry players — with each cluster contributing distinct capabilities to the overall materials pipeline.

Institutional contributors

Chinese academic institutions dominate the literature output across every sub-domain: Huazhong University of Science and Technology (THz characterization methods), Fudan University (D-band propagation loss), Sun Yat-sen University (non-fluorinated polyimides), University of Electronic Science and Technology of China (LCP materials for 5G), Northwestern Polytechnical University (cavity measurement of anisotropic dielectrics), Xi’an Jiaotong University (cold-sintered ceramic composites), and Shenyang University of Chemical Technology (polyimide synthesis). The breadth of Chinese institutional contributions reflects coordinated national investment in materials for 5G infrastructure and the pre-competitive 6G research phase.

Unimicron Technology Corporation, a major PCB manufacturer based in Taiwan, stands out as the only PCB industry contributor with a materials characterization study directly tied to production-level circuit board fabrication (2022). This positions Unimicron at the interface between raw material development and manufacturing qualification — a strategically important position as 5G mmWave volume production scales. University of Oulu in Finland is notable as an early mover in 6G-specific dielectric substrate development, with its 2020 porous silica-cellulose nanocomposite study representing one of the first published material studies explicitly framed around 6G frequency requirements. UBO/Lab-STICC CNRS (France) has focused on translating low-loss thermoplastics into practical mmWave filter structures using LDS manufacturing. Kyocera Corporation holds an earlier patent on low dielectric loss ceramic material for high-frequency applications (filed 2000, DE jurisdiction), reflecting the company’s long-standing position in ceramic and multilayer dielectric components for wireless communication.

Five innovation trends

Five key innovation trends are visible across the dataset:

• From fluorinated to non-fluorinated polymers: driven by environmental and regulatory pressure, the move toward non-fluorinated low-loss polymers is accelerating, with Sun Yat-sen University’s TmBPPA demonstrating that performance parity is achievable.
• Porous architectures approaching air-like permittivity: both polymer and ceramic composite systems are adopting porous designs to drive Dk toward 1, with the University of Oulu’s 0.19 g/cm³ silica-cellulose composite as the leading example explicitly targeting 6G.
• Machine learning in dielectric characterization: Huazhong University of Science and Technology’s 4-layer neural network for THz refractive index extraction exemplifies the integration of ML into the material qualification pipeline, particularly at frequencies where traditional algorithms fail.
• Cold sintering as an energy-efficient ceramic processing route: Xi’an Jiaotong University’s ~150°C cold-sintered composites with Qf ~4,000 GHz demonstrate a low-energy pathway to temperature-stable microwave resonators for base station infrastructure.
• D-band and THz characterization methods: as 6G system architectures are defined around D-band and terahertz operation, the measurement and qualification infrastructure for materials at these frequencies is becoming a competitive differentiator in its own right.