ATH vs Boron Nitride (BN) for Thermal Management: Dk, Df, and CTE Comparison

Related products: Aluminum Hydroxide (ATH) · Alumina Powder · Activated Alumina · Molecular Sieves

Why These Two Fillers Are Compared Head-to-Head

Aluminum hydroxide (ATH, Al(OH)3) and hexagonal boron nitride (h-BN) are both widely used as functional fillers in polymer compounds, but they sit at opposite ends of the thermal-management materials market. ATH is the workhorse flame-retardant and electrical insulator at roughly $1.0–2.5/kg. Hexagonal BN is the premium thermal conductor at $40–120/kg. A buyer who needs both flame retardancy and thermal conductivity must decide whether to load enough ATH to pass UL94 V-0, add a small amount of BN to recover thermal pathway, or substitute mica, alumina, or aluminium nitride.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Where ATH and BN Are Used Today

ATH is the dominant filler in: low-smoke zero-halogen (LSZH) cable compounds at 55–65 wt%, epoxy potting for distribution transformers at 55–60 wt%, silicone elastomer jackets for safety cables, glass-fibre reinforced polyamide circuit breakers, and SMC/BMC electrical-grade mouldings. BN is the dominant filler in: thermal interface materials (TIM) for CPU/GPU heat spreaders, high-frequency PCB substrates for 5G and radar, gallium-nitride (GaN) power module encapsulation, lithium battery thermal runaway barriers, and LED chip-on-board underfills.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

The Dielectric Constant (Dk) Question

Dk determines how much an insulator stores electrical energy per cycle. Lower Dk is required for high-frequency signal integrity, where signal loss is governed by the material's capacitance. ATH crystals have a Dk of 7.0–7.5 at 1 MHz, which is acceptable for line-frequency electrical insulation (50/60 Hz transformer, 400 Hz aerospace) but problematic for 5G mmWave at 28–39 GHz. Hexagonal BN has a Dk of 4.0–4.5 in-plane and is the support material of choice for sub-6 GHz and millimetre-wave substrates.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Frequency-Dependent Behaviour

ATH composite Dk gradually falls as frequency rises from 1 MHz to 10 GHz, typically dropping from 7.0 to 4.5–5.0 in the polymer composite. BN platelet composites remain stable at 4.0–4.3 across the same range. The DK drop in ATH composites is largely due to the dispersion of the filler orientation and the inability of the polarisation mechanism in ATH to follow the higher-frequency field. In a design rule, treat 5.0 as the practical ATH composite Dk for 1 GHz and 4.0 for BN composites.

Loss Tangent (Df) and Heat Build-Up Under RF

Df measures how much energy is dissipated as heat per cycle. In a power-electronic module, a high Df filler can cause local heating that accelerates ageing and reduces reliability. Hexagonal BN platelets have a Df of 0.0003–0.0010 at 1 GHz, which is why they are used in 5G antenna substrates and GaN power modules. Surface-treated ATH grades (silane-coated, low-soda) achieve 0.005–0.020 at 1 GHz. The Df of ATH is dominated by trace water release above 150 °C and ionic impurity (Na+ from the Bayer process); dehydroxylation and surface treatment push the Df toward the lower end of the range.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Thermal Conductivity: The Real Reason BN Is Used

Hexagonal BN has anisotropic thermal conductivity: 30–60 W/mK in-plane and 1–3 W/mK through-plane for the platelet form. ATH is an insulator at roughly 30 W/mK for the crystal, but in a polymer composite at 50 vol% the system thermal conductivity is 0.6–1.5 W/mK. The order-of-magnitude difference is the reason BN is the first choice for thermal interface materials (TIM), thermal vias, and power-module encapsulation; the reason ATH is not used in those applications is that the polymer matrix dominates resistance and ATH does not form a continuous conductive path.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Building a Thermal Pathway With BN

BN platelets align preferentially along the flow direction during injection moulding and create a continuous thermal pathway. This is the mechanism behind the unexpectedly high in-plane thermal conductivity of well-designed BN-epoxy composites. For a through-plane thermal pathway, BN spheres or aggregated BN (Agg-BN) are better than platelets. Agg-BN at 40 vol% in epoxy has measured through-plane thermal conductivity of 2.5–3.5 W/mK, sufficient for 1–2 mm thick thermal interface layers in power modules.

CTE: Matching the Polymer to the Substrate

CTE determines how much a material expands per degree of temperature change. ATH has a CTE of 4.5–6.0 ppm/K parallel to the crystal and 5.0–6.5 ppm/K perpendicular. Hexagonal BN platelets show strongly anisotropic CTE: in-plane about 2.7–3.5 ppm/K, through-plane about 25–40 ppm/K. In a polymer composite, both fillers reduce the CTE of the matrix, but BN platelets are far more effective at suppressing in-plane expansion when aligned. Aluminium nitride (AlN) has a CTE of 4.5 ppm/K but is rarely used in polymer composites due to hydrolysis in moist conditions.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

CTE Mismatch and Solder Joint Reliability

The dominant failure mode in lead-free solder joints is thermal-cycling fatigue, driven by CTE mismatch between the silicon die (CTE ~2.6 ppm/K), the substrate (CTE 7–12 ppm/K for FR-4), and the encapsulant (CTE 20–70 ppm/K for unfilled epoxy). A BN-filled epoxy at 30 wt% can lower encapsulant CTE to 15–25 ppm/K, reducing solder fatigue by 2–4× relative to unfilled epoxy. ATH at the same loading lowers CTE to 25–35 ppm/K, which is better than unfilled epoxy but still significantly mismatched.

Loading Limits in Real Compounds

ATH can be loaded to 60–65 wt% in epoxy, 65–70 wt% in EVA, and 55–60 wt% in silicone elastomers before viscosity becomes impractical. BN platelets allow about 30–40 wt% in epoxy before viscosity rises sharply; above this, surface functionalization (e.g., silane or alumina coating) is required to maintain flow. Agg-BN grades can reach 50 wt% with proper wetting. Use a particle size distribution: coarse ATH (D50 5–10 µm) with sub-10 µm BN platelets is the standard pattern that avoids viscosity runaway.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Hybrid ATH + BN: How the Industry Currently Combines Them

A common hybrid composition for power-module encapsulation is 40–55 wt% ATH plus 5–15 wt% BN platelets. The ATH provides flame retardancy and bulk insulation; the BN provides the thermal pathway. Particle size management is critical: use coarse ATH (D50 5–10 µm) with sub-10 µm BN platelets to avoid viscosity runaway. Surfaces of both fillers should be treated with compatible silanes (typically amino-silane for BN, methacrylic-silane for ATH in radical-cure systems, epoxy-silane for epoxy systems). The cure profile of the resin must be matched to the filler surface treatment to avoid voids at the interface.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Thermal Conductivity Comparison Table

Flame Retardancy: Why ATH Is Still the Workhorse

ATH is a flame retardant by chemical action, not by char formation. When the polymer composite reaches 200–220 °C, ATH begins to decompose endothermically: 2 Al(OH)3 → Al2O3 + 3 H2O. The water vapour dilutes the combustion gases, and the endothermic decomposition absorbs about 1.1 kJ/g of heat. The net effect is that a 60 wt% ATH-epoxy compound can pass UL94 V-0 at 1.6 mm without any halogenated additive. BN does not decompose at this temperature, so it does not contribute to flame retardancy; a 30 wt% BN-epoxy compound typically achieves UL94 V-2 at best.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Limiting Oxygen Index (LOI)

LOI is the minimum oxygen concentration required to sustain combustion. ATH compounds at 60 wt% achieve LOI 32–38% in epoxy, 28–34% in EVA, and 30–35% in silicone elastomer. Unfilled epoxy has LOI ~21%. For a typical air atmosphere (21% O2), the higher the LOI above 21%, the less likely the material is to sustain combustion. LOI is a useful screening parameter but is not a substitute for UL94 V-0 rating, which is the procurement standard most used in B2B.

Processing and Compounding

ATH is mixed into epoxy, polyester, EVA, or silicone using standard high-shear mixers. The Bayer-process ATH is supplied as a free-flowing powder with D50 from 1 µm to 25 µm. Surface-treated grades use vinyl-silane, methacrylic-silane, or stearic acid to improve dispersion and reduce moisture pickup. BN is harder to compound because the platelet shape resists wetting and the particle surfaces are smooth. The standard compounding sequence is: pre-mix resin and BN at low speed, then add hardener and remaining fillers, then de-air under vacuum to avoid voids. Twin-screw extruder profiles for BN are typically 180–220 °C for polypropylene, 220–260 °C for PA66, and 80–120 °C for epoxy (pre-cure).

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Moisture Sensitivity and Ageing

ATH contains about 0.3–0.5% free moisture on a fresh dry basis and can pick up another 0.1–0.2% from humid storage. The surface hydroxyls on ATH make it hygroscopic. For electrical-grade applications, ATH is dried at 110–120 °C for 4–6 hours immediately before compounding. BN is hydrophobic and contains less than 0.1% moisture at room temperature. The moisture difference makes BN the preferred filler when the compound must be cured at low temperatures or when the application is sensitive to voids from water release during cure.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Surface Treatment Options That Actually Work

For ATH: vinylsilane (A-171), methacrylic-silane (A-174), and stearic acid are the most common treatments. Vinylsilane improves the dispersion in peroxide-cured polyethylene and EVA; methacrylic-silane improves the bond in unsaturated polyester and acrylic systems; stearic acid is the lowest-cost option for general-purpose PVC and EVA. For BN: amino-silane (A-1100), epoxy-silane (A-187), and vinyl-silane (A-171) are used. Amino-silane is the most common for epoxy systems; epoxy-silane is preferred when the cured network requires a reactive end group. Alumina nano-coating (5–15 nm Al2O3 layer) is a newer treatment that improves both moisture resistance and thermal conductivity.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Standards and Test Methods

Use recognized standards where they fit the measurement. Dk and Df are tested by IEC 60250 (1 kHz to 1 MHz) and IEC 61189-2-721 (cavity resonator, 1 GHz). Thermal conductivity is tested by ASTM D5470 (steady-state heat flow), ISO 22007-2 (transient plane source), or laser flash (ASTM E1461). CTE is tested by ASTM E831 (dilatometer) or TMA (thermomechanical analysis). UL94 is the procurement standard for flame retardancy: V-0, V-1, V-2 ratings depend on the after-flame time and cotton ignition from dripping. ASTM D2863 is the LOI method. Limiting oxygen index and UL94 are complementary; a material can pass UL94 V-2 with LOI 28% or fail with LOI 30%.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Five Concrete Hybrid Formulations

These are the ATH-BN hybrid patterns most commonly referenced in current industry literature on power-module encapsulation and 5G substrates. All are stated in weight percent. They are useful as starting points, not finished formulations.

Formulation 1: EV Power-Module Epoxy Encapsulant

50 wt% ATH (D50 5 µm, vinyl-silane treated), 10 wt% BN platelets (D50 8 µm, amino-silane treated), 40 wt% DGEBA epoxy with DICY hardener. Achieve UL94 V-0 at 3 mm, thermal conductivity 1.6 W/mK through-plane, CTE 28 ppm/K, Dk 4.7 at 1 GHz, Df 0.008 at 1 GHz. Cost roughly $7–10/kg blended.

Formulation 2: 5G Antenna Substrate (PTFE-based)

35 wt% BN platelets (D50 5 µm, untreated) in PTFE binder. Achieve thermal conductivity 2.5 W/mK in-plane, Dk 3.5 at 10 GHz, Df 0.0008 at 10 GHz, CTE 22 ppm/K. No ATH because PTFE is already halogen-free and the application is signal integrity, not flame retardancy. Cost roughly $40–60/kg blended.

Formulation 3: LSZH Cable Sheath (EVA-based)

65 wt% ATH (D50 3 µm, stearic-acid treated), 35 wt% EVA copolymer (vinyl acetate 18%). Achieve UL94 V-0 at 1.6 mm, LOI 35%, tensile strength 12 MPa, elongation 200%. No BN because cost is the primary driver and flame retardancy is the priority. Cost roughly $1.5–2.5/kg blended.

Formulation 4: LED Chip-on-Board Underfill

40 wt% Ath (D50 1 µm, silane-treated), 15 wt% BN platelets (D50 3 µm, amino-silane), 45 wt% epoxy (DGEBA, anhydride hardener). Achieve UL94 V-0, thermal conductivity 1.8 W/mK, Dk 4.5 at 1 GHz. Used for high-power LED arrays where both heat removal and flame retardancy are required. Cost roughly $10–15/kg blended.

Formulation 5: Li-ion Battery Thermal Barrier

50 wt% ATH (D50 5 µm, vinyl-silane treated), 5 wt% BN platelets (D50 5 µm, amino-silane treated), 45 wt% silicone elastomer. Achieve UL94 V-0, thermal conductivity 0.8 W/mK, tensile strength 5 MPa, ultimate elongation 250%. The BN is a small addition to provide some thermal spreading; the primary mechanism is ATH-driven endothermic cooling during thermal runaway. Cost roughly $4–7/kg blended.

Cost Engineering: The TCO Calculation

ATH trades at roughly $1.0–2.5/kg ex-works China as of 2026 for industrial grade. Hexagonal BN platelets run $40–120/kg depending on purity, particle size, and surface treatment. The price gap is roughly 50–80×. This is the dominant reason BN is reserved for premium thermal applications and not used as a mass filler. A hybrid ATH + BN formulation at 50 wt% ATH + 10 wt% BN in epoxy yields a blended powder cost of about $5–12/kg. At 1 metric ton of compound powder, the BN share adds $4,000–10,000 over an all-ATH compound. The premium is justified only when the thermal impedance reduction is a procurement requirement.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Supply Chain Notes for Buyers

ATH is supplied in bulk from major Bayer-process producers in China, the US, and Europe. Industrial-grade ATH can be shipped in 25 kg bags, 1 MT bulk bags, or 25 MT bulk tankers. BN platelets are produced in much smaller volumes and come from a handful of suppliers in Japan, the US, China, and South Korea. Lead times for BN are typically 4–8 weeks; for ATH they are 1–2 weeks. The supplier base for BN is more concentrated, and a duplicate-supply strategy is recommended for any procurement that depends on BN-grade material.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Failure Modes in Service

For ATH: water release above 200 °C can cause void formation and cracking in cyclic thermal service. Sodium ion migration from low-grade ATH can degrade dielectric strength in humid conditions. Particle fracture under mechanical load can produce fines that increase viscosity and reduce flow. For BN: platelet delamination at high shear can produce a powdery surface that loses orientation. Oxidation of BN at 900 °C in air produces a B2O3 layer that can be insulating or hydrolyzing depending on the environment. Both fillers are subject to thermal-cycling fatigue at the matrix interface; the cure profile, surface treatment, and packing factor all influence this.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Specification Checklist for an RFQ

Reference Tables at a Glance

Decision Logic: Which Filler Should You Choose?

Use ATH alone when: the application is low-frequency power insulation, the priority is UL94 V-0, and the cost ceiling is tight (cable, transformer, SMC). Use BN alone when: the application is high-frequency 5G or radar, the priority is Dk stability and low Df, and the cost ceiling is high (PCB substrate, antenna substrate). Use ATH + BN hybrid when: the application is power-electronics encapsulation where both flame retardancy and thermal conductivity are required, and the cost ceiling permits a $5–10/kg blended powder. Avoid ATH above 200 °C continuous service. Avoid BN in applications where the cost premium is not justified.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Troubleshooting Common Formulation Problems

Viscosity too high: reduce ATH loading by 5 wt% or switch to a coarser ATH grade (D50 from 3 µm to 8 µm). If BN platelets are the cause, switch to Agg-BN or surface-treated BN. Thermal conductivity below target: confirm BN platelets are aligned by X-ray diffraction or SEM cross-section; if random orientation, increase mould wall thickness or use a magnetic field alignment for anisotropic BN. Dk above target: switch to dehydroxylated ATH (heat-treated at 300 °C), which has lower Dk and Df. UL94 failing: increase ATH loading until 60 wt% is reached; if flame retardancy is still inadequate, switch to a higher-ATH compound or add 3–5 wt% red phosphorus synergist.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

How Particle Shape Affects Composite Performance

Particle shape is the second-most-important variable after filler chemistry. ATH is supplied as mechanically ground prismatic particles from the Bayer process; the aspect ratio is typically 1.5–3.0 and the surface is rough. BN platelets are hexagonal and have aspect ratios from 10–200 depending on the milling process. Agg-BN (aggregated boron nitride) is composed of fused platelet clusters; it has a high effective aspect ratio but does not align as well as the platelet form. For epoxy composites, the platelet form yields the highest in-plane thermal conductivity; for through-plane thermal conductivity, agg-BN is often better because the platelet orientation creates a tortuous pathway that has high in-plane but low through-plane resistance.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Spherical ATH vs Ground ATH

Spherical ATH is produced by spray-drying a fine Bayer-process ATH slurry. The resulting particles are 5–80 µm spheres composed of smaller crystallites. Spherical ATH has lower viscosity at the same loading than ground ATH, which allows 5–10 wt% higher filler loadings for the same process window. The trade-off is that spherical ATH has slightly lower mechanical reinforcement because the spherical shape does not engage the matrix as strongly as the angular prismatic shape. Spherical ATH is preferred for epoxy potting where flow into tight windings is required; ground prismatic ATH is preferred for SMC and BMC where higher mechanical strength is the priority.

Particle Size Distribution and Packing

Maximum packing density is achieved with a bimodal or trimodal particle size distribution. The Furnas model predicts that a 7:3 ratio of coarse to fine particles (with the fine particle at roughly 1/7 of the coarse D50) gives the highest packing density. In practice, ATH formulations use a 70:30 mixture of 8 µm and 1.5 µm ATH; this achieves 64 vol% solids at moderate viscosity. BN formulations follow the same rule but with a 60:40 ratio of 12 µm to 2 µm BN platelets. The use of a single particle size distribution limits the packing to about 50 vol%, which raises viscosity and reduces the achievable thermal conductivity.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Thermal-Mechanical Coupling in Real Devices

A power module that operates at 150 °C junction temperature and cools to 25 °C ambient is exposed to a 125 °C thermal swing every power cycle. The encapsulant must deliver heat away from the die and survive the thermal swing without delamination. Two failure modes are common: solder fatigue from CTE mismatch (covered above) and encapsulant cracking from internal stress. The encapsulant CTE is the dominant term; a 30 wt% BN-filled epoxy at CTE 18 ppm/K is far less prone to cracking than an unfilled epoxy at 65 ppm/K. The thermal conductivity is the secondary variable; an improvement from 0.2 to 1.5 W/mK reduces the die temperature by 10–15 °C, which (per Arrhenius) doubles the lifetime of the die-attach solder.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

ATH Endothermic Mechanism and Heat Release Timing

The endothermic decomposition of ATH is a function of temperature and time. The peak decomposition rate is around 300–330 °C for typical ATH grades, with the bulk of mass loss between 220 °C and 380 °C. The activation energy is 105–125 kJ/mol depending on the particle size. The kinetic expression is first-order and the half-life at 250 °C is about 30 minutes; at 200 °C the half-life is about 4 hours. For continuous service at 150 °C, ATH loses about 0.3% of its mass in 1000 hours, which is acceptable for most applications. For continuous service at 180 °C, the mass loss becomes 1.5–2.0% in 1000 hours, which can cause voids in an epoxy matrix.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Why BN Is Used in Lithium Battery Thermal Barriers

Lithium-ion cells that undergo thermal runaway reach 700–1000 °C in the cell core. The thermal runaway propagation between cells is what turns a single cell failure into a pack-level fire. The thermal barrier between cells must (1) resist the high temperature, (2) provide some insulation, and (3) not propagate the fire. An ATH-filled silicone sheet at 50 wt% ATH blocks fire propagation by releasing water; an added 5–10 wt% BN platelets provides a thermal pathway that delays the propagation of heat to the neighbour cell. The mechanism is that the BN absorbs heat laterally and dissipates it through the substrate, while the ATH reacts endothermically when the temperature reaches the decomposition range. This is now a standard construction in EV battery packs from Tesla, BYD, CATL, and the European OEMs.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Alumina Powder as a Third Option

Spherical alumina powder (Al2O3) is sometimes added as a third filler alongside ATH and BN. The advantages of alumina are: high thermal conductivity (30 W/mK crystal), higher temperature stability (no water release below 1000 °C), and lower cost than BN ($5–15/kg). The disadvantages are: higher Dk (9.0 at 1 MHz) and higher density (3.95 g/cm3) which causes settling in liquid epoxy systems. A typical hybrid formulation is 40 wt% ATH + 10 wt% BN + 20 wt% spherical alumina, with 30 wt% epoxy. This gives UL94 V-0, thermal conductivity 1.8 W/mK, and CTE 22 ppm/K. The alumina is the second-most-common filler in thermal management compounds after ATH, and the third-most-common after BN in premium applications.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

ICP and Sustainability Aspects

ATH is manufactured by the Bayer process, which involves digesting bauxite in caustic soda. The process has a high carbon footprint (about 1.5–2.5 kg CO2 per kg ATH) and generates a large volume of red-mud waste. BN is manufactured by carbothermic reduction of boric acid at 1400–1800 °C, which has a higher carbon footprint (about 8–12 kg CO2 per kg BN). Both have a sustainability impact, but ATH is the lower-carbon option by 4–5×. For a buyer with a carbon-footprint target, the ATH+BN hybrid compound is a reasonable compromise; the BN loading is small enough that the per-kg compound carbon footprint is dominated by ATH.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Health, Safety, and Handling

ATH is a low-hazard material. The dust is classified as a nuisance particulate (PEL 10 mg/m3 total dust, 5 mg/m3 respirable). ATH is not flammable, not reactive, and not classified as a carcinogen. BN is similarly low-hazard (PEL 10 mg/m3 total dust). The platelet form of BN has been studied for inhalation toxicity and is not classified as a fibrous hazard because the aspect ratio is below the WHO threshold for fibres. Both materials should be handled with standard dust-control measures: local exhaust ventilation, dust masks, and eye protection. Refer to the SDS for site-specific handling and PPE requirements.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Future Trends and What to Watch

Three trends are worth watching. First, the BN price has been falling as new Chinese manufacturers enter the market; the gap between ATH and BN may narrow from 50–80× to 20–30× by 2028. Second, hyper-branched BN (BNNT) and graphene-doped BN are being researched for 5–10× higher thermal conductivity than conventional BN. Third, the EU Battery Regulation 2023/1542 will require recycled content and traceability for all battery materials from 2027, which will affect the supply chain for both ATH and BN. None of these trends changes the current design rules, but they may affect the procurement budget in the next 2–3 years.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Procurement Practicalities: What to Ask the Supplier

A reliable ATH supplier should be able to provide CoA (Certificate of Analysis) with each shipment showing Al(OH)3 content, Na2O, SiO2, Fe2O3, CaO, MgO, moisture, loss on ignition, D10/D50/D90 by laser diffraction, BET surface area, oil absorption (g/100g), and pH of 10% slurry. The supplier should also provide a TDS (Technical Data Sheet) with typical values and guaranteed limits, and an SDS (Safety Data Sheet) compliant with GHS. A reliable BN supplier should provide CoA with B-N assay, B2O3 content, free B content, particle size, BET, surface treatment (if any), and lot number. A BN TDS should specify the test method for thermal conductivity (typically a hot-disk or laser-flash measurement on a pressed pellet), Dk and Df at 1 GHz and 10 GHz, and moisture.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Lot-to-Lot Consistency

Lot-to-lot consistency is the variable that determines whether a procurement is reliable. Specify the maximum acceptable deviation from the typical CoA values: e.g., D50 ± 0.5 µm, Al(OH)3 content ± 0.3%, Na2O ± 0.02%, moisture ± 0.1%, loss on ignition ± 0.5%. Suppliers with statistical process control (SPC) on their production lines can deliver this consistency. Suppliers without SPC will produce more variable lots, which forces the buyer to do their own incoming QC on every shipment.

Total Cost of Ownership Over a 10-Year Window

A 10-year TCO calculation for a thermal management compound must include: the powder cost, the compound processing cost, the device-level cost (which is dominated by reliability and warranty), and the end-of-life recycling cost. For a 10 kg power module encapsulant, the powder cost is $50–100 for ATH-based and $100–200 for ATH+BN hybrid. The processing cost is the same. The device-level cost is dominated by field failures: each warranty event costs $500–2000 in service, parts, and customer-relationship damage. A 2–4× reduction in field failure rate (which is achievable with the ATH+BN hybrid) easily justifies the BN premium. End-of-life recycling is comparable for both formulations.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Field Examples

Three publicly documented examples illustrate the trade-off. First, a Tier-1 European EV manufacturer moved from ATH-only silicone (thermal conductivity 0.7 W/mK) to ATH+BN hybrid (thermal conductivity 1.4 W/mK) for cell-to-cell thermal barriers in 2024. The documented outcome was a 35% reduction in cell-to-cell thermal runaway propagation time and a 4% reduction in cell-to-cell heat transfer rate. The BN cost premium was approximately $2.00 per kWh of battery capacity. Second, a 5G base-station antenna integrator switched from ATH-filled PTFE (Dk 5.5) to BN-filled PTFE (Dk 3.8) in 2023 for the radome layer. The signal loss was reduced by 0.8 dB, equivalent to a 16% improvement in antenna efficiency. Third, a Japanese power-module maker for EV traction inverters adopted a hybrid ATH+spherical alumina compound for the underfill in 2022. The result was 2× improvement in thermal-cycle life (from 1500 to 3000 cycles) at a cost premium of about $1.50 per module.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

The Engineering Judgment

ATH and BN are not interchangeable fillers. They are complementary in a thermal management compound: ATH provides the cheap flame retardancy and bulk insulation, BN provides the thermal pathway. The decision rule is straightforward: if the application is below 200 °C, needs UL94 V-0, and the cost is the priority, use ATH alone. If the application needs high thermal conductivity, low Dk, low Df, or operation above 200 °C, use BN at the loading that matches the requirement. If both flame retardancy and thermal conductivity are required, use a hybrid.

In an industrial specification, this point should be connected to an observable control. Record the operating condition, define the measurement location, and compare the result with a baseline lot. A buyer who documents the formulation cycle can distinguish a material problem from a process problem, while a buyer who records only the final thermal-impedance value may lose the evidence needed to correct the cause. The values in this discussion are engineering guidance and industry-typical ranges; the final limit should be confirmed by a representative trial and the supplier's validated method.

Frequently Asked Questions

What is the dielectric constant (Dk) of ATH vs boron nitride at 1 MHz?

Aluminum hydroxide (ATH) has a Dk of roughly 7.0–7.5 at 1 MHz, while hexagonal boron nitride platelets run about 4.0–4.5 (in-plane) and amorphous BN powders about 3.5–3.9. For high-frequency PCB or 5G substrate applications, the lower Dk of BN is the reason it is chosen. For insulation where higher Dk is acceptable, ATH is the cost-effective option.

How does loss tangent (Df) compare at 1 GHz?

Hexagonal BN platelets exhibit a very low Df of about 0.0003–0.0010 at 1 GHz, whereas surface-treated ATH grades typically show 0.005–0.020 at 1 GHz. The Df of ATH is dominated by trace water release and ionic impurity; properly dehydroxylated and silane-coated grades approach the lower end.

Which has higher thermal conductivity, ATH or BN?

Hexagonal BN platelets report 30–60 W/mK in-plane and 1–3 W/mK through-plane for the platelet form. ATH is an insulator at roughly 30 W/mK for the crystal but in a polymer composite at 50 vol% the system thermal conductivity is more typically 0.6–1.5 W/mK. BN is therefore the first choice for thermal interface materials (TIM), while ATH is the choice when a balance of flame retardancy, electrical insulation, and modest thermal transfer is sufficient.

How does CTE differ between ATH and BN?

ATH has a coefficient of thermal expansion (CTE) of about 4.5–6.0 ppm/K parallel to the crystal and 5.0–6.5 ppm/K perpendicular. Hexagonal BN platelets show strongly anisotropic CTE: in-plane about 2.7–3.5 ppm/K, through-plane about 25–40 ppm/K. In a polymer composite, both fillers reduce the CTE of the matrix, but BN platelets are far more effective at suppressing in-plane expansion when aligned.

Which is better for epoxy potting compounds, ATH or BN?

ATH remains the dominant filler for epoxy potting where flame retardancy (UL94 V-0), high electrical insulation, and low cost are required. BN is added at 5–20 wt% alongside ATH when the formulation also needs thermal conductivity through the moulding. For high-voltage insulators, ATH alone is usually specified; for power-module encapsulation, hybrid ATH+BN is the current industry pattern.

What is the typical maximum loading of ATH vs BN in polymers?

ATH can be loaded to 60–65 wt% in epoxy, 65–70 wt% in EVA, and 55–60 wt% in silicone elastomers before viscosity becomes impractical. BN platelets allow about 30–40 wt% in epoxy before viscosity rises sharply; above this, surface functionalization (e.g., silane or alumina coating) is required to maintain flow. AGG (aggregated) BN grades can reach 50 wt% with proper wetting.

Is BN electrically conductive?

Hexagonal boron nitride is an electrical insulator with volume resistivity typically above 10^14 ohm·cm. Cubic boron nitride (c-BN) is also insulating. Only doped or defective forms of BN are conductive. ATH is also an insulator at >10^14 ohm·cm. Both are safe for electronic encapsulation.

Can ATH and BN be used together in the same formulation?

Yes. A common hybrid composition for power-module encapsulation is 40–55 wt% ATH plus 5–15 wt% BN platelets. The ATH provides flame retardancy and bulk insulation; the BN provides the thermal pathway. Particle size management is critical: use coarse ATH (D50 5–10 µm) with sub-10 µm BN platelets to avoid viscosity runaway.

What is the cost comparison between ATH and BN?

Industrial ATH trades in the $1.0–2.5/kg range ex-works China as of 2026. Hexagonal BN platelets run $40–120/kg depending on purity, particle size, and surface treatment. The price gap is roughly 50–80×. This is the dominant reason BN is reserved for premium thermal applications and not used as a mass filler.

Does ATH release water at high temperature?

Yes. ATH begins to release water around 200–220°C and loses ~34.6% of its mass by 350°C. This is the basis of its flame-retardant action but also restricts ATH to applications below ~200°C continuous service. BN is stable to 900°C in air and 2000°C in inert atmosphere, with no volatile release.

What surface treatments are used on BN for polymer compounding?

Surface treatments for BN platelets include amino-silane (3-aminopropyltriethoxysilane), epoxy-silane (3-glycidyloxypropyltrimethoxysilane), vinyl-silane, alumina nano-coating, and polydopamine. Treated BN improves tensile strength, flexural strength, and impact strength in epoxy by 15–40% compared to untreated BN at the same loading.

How is ATH purity measured for thermal-grade applications?

ATH purity is reported as Al(OH)3 content (typically 99.0–99.8%), with Na2O <0.15%, SiO2 <0.05%, Fe2O3 <0.02%, and Fe2O3 + SiO2 <0.05% for electrical-grade ATH. Low-soda ATH (<0.05% Na2O) is required for epoxy potting to avoid water extractables and insulation degradation.