FR-4 vs G-10 Fiberglass Polishing Templates: Material Properties & Selection Guide
Two materials. Nearly identical names. Genuinely different performance envelopes. This guide explains exactly when each is the right choice — and when neither is sufficient.
What FR-4 and G-10 Actually Are
FR-4 and G-10 are both members of the NEMA (National Electrical Manufacturers Association) LI-series of industrial laminate materials — composite sheets built from woven E-glass fabric impregnated with an epoxy resin system and cured under heat and pressure into rigid, dimensionally stable laminates. They have been manufactured to standardized specifications since the 1950s, originally for printed circuit board substrates, and their consistent dimensional and mechanical properties have made them the default carrier plate material in semiconductor polishing templates worldwide.
The naming convention is straightforward in principle: the letters describe the flame retardancy class, and the number describes the base fabric and resin system. G-10 is the base specification: woven E-glass / general-purpose epoxy, with no flame retardant requirement. FR-4 is the flame-retardant version of G-10, manufactured with halogenated (brominated) epoxy to achieve a UL 94 V-0 flammability rating. In all mechanical and dimensional respects, they are essentially identical. The difference is in the resin chemistry — specifically, what has been added to the epoxy to make it flame retardant.
Understanding this distinction matters for polishing template material selection because the flame retardant additive — tetrabromobisphenol A (TBBPA) in most FR-4 formulations — affects the epoxy matrix’s response to acidic chemical environments in ways that are meaningful for slurry-contact applications, even though they are irrelevant for the original circuit board application context for which both materials were designed.
The One Real Difference Between FR-4 and G-10
With all the technical language around laminate grades, the practical difference between FR-4 and G-10 in polishing template applications reduces to a single sentence: G-10 tolerates mildly acidic polishing environments (pH 5–7) somewhat better than FR-4, because its epoxy matrix does not contain the brominated flame retardant that makes FR-4 more susceptible to acid-induced swelling.
The mechanism is as follows. In an acidic aqueous environment, the ester linkages in epoxy resins are susceptible to hydrolytic degradation — the acid catalyzes ester bond cleavage, causing gradual absorption of water into the resin matrix and progressive dimensional swelling. In FR-4, the TBBPA flame retardant is chemically bonded into the epoxy backbone; the presence of the halogen substituents makes the resin’s ester groups slightly more electrophilic and therefore more susceptible to acid-catalyzed hydrolysis. G-10’s epoxy, without the halogenated additive, is marginally more resistant to this mechanism.
In practice, this difference manifests as a longer service life for G-10 templates in pH 5–7 slurry environments — typically 20–40% more polishing cycles before dimensional drift of the carrier plate exceeds the replacement threshold. For pH 8–12 (standard alkaline silicon polishing), both materials perform equivalently and the cost advantage of FR-4 makes it the correct default choice.
Full Material Property Comparison
Beyond chemical resistance, FR-4 and G-10 share nearly identical mechanical, thermal, and dimensional properties — which is precisely why the chemical resistance difference is the only meaningful selection criterion between them for polishing template applications. The following table presents the full property comparison relevant to template engineering.
| Property | FR-4 | G-10 | CXT Grade | Relevance to Template Performance |
|---|---|---|---|---|
| Tensile strength | 270–310 MPa | 270–310 MPa | Equivalent | Determines resistance to carrier head clamping forces |
| Flexural modulus | 18–22 GPa | 18–22 GPa | Similar | Higher modulus → better carrier plate bow resistance under polishing load |
| CTE (in-plane) | 14–16 × 10⁻⁶/°C | 14–16 × 10⁻⁶/°C | Similar | Must be compatible with carrier head material to prevent bow at process temperature |
| Water absorption (24h) | 0.10–0.20% | 0.10–0.20% | Lower | Lower absorption → less dimensional change in wet polishing environment |
| Dichte | 1.80–1.90 g/cm³ | 1.80–1.90 g/cm³ | Similar | Affects template weight; relevant for carrier head balance on large multi-pocket templates |
| Surface hardness (Rockwell M) | M-110 typical | M-110 typical | Equivalent | Hardness determines machinability and edge quality after CNC milling |
| Dielectric constant (@ 1 MHz) | 4.5–5.0 | 4.5–5.0 | N/A | Proxy for material homogeneity; tight Dk range indicates consistent fiber/resin distribution |
| Acid resistance (pH 3–5) | Moderate | Gut | Ausgezeichnet | Primary selection criterion for acidic slurry applications |
| Alkaline resistance (pH 8–12) | Ausgezeichnet | Ausgezeichnet | Ausgezeichnet | Both grades perform equivalently for standard alkaline Si polishing |
| Oxidant resistance (KMnO₄, H₂O₂) | Poor | Poor | Ausgezeichnet | Critical for SiC CMP; both laminate grades fail in KMnO₄ environments |
| Laminate delamination risk | Present (layer interface) | Present (layer interface) | None (seamless) | Delamination introduces dimensional instability and contamination |
| Halogen content | ~18–21% Br (TBBPA) | None | None | Halogen-free preferred for some fab chemical management programs |
pH & Chemical Compatibility: The Decisive Factor
For semiconductor polishing template material selection, chemical compatibility with the process slurry is the primary criterion, and pH range is the most practical way to characterize it. The following visualization shows the effective operating range for each material against the pH scale.
Reading this chart against your actual slurry chemistry leads directly to the correct material selection:
- Colloidal silica slurry for Si SSP, pH 9–11: FR-4 is entirely adequate. No G-10 premium justified.
- Oxide CMP slurry with NH₄OH additive, pH 10–11: FR-4. Standard alkaline environment.
- Citric acid-buffered silica slurry for glass polishing, pH 5–6: G-10 preferred. FR-4 may show swelling over 50+ cycle lifetimes.
- HNO₃-buffered diamond slurry for sapphire, pH 4–6: G-10 minimum; CXT preferred for high cycle count production.
- KMnO₄-based slurry for SiC CMP, pH 2–4: CXT mandatory. Neither FR-4 nor G-10 is viable. See our detailed SiC polishing template guide.
- H₂SO₄/H₂O₂ (piranha-type) slurry, pH < 2: CXT mandatory. Extreme acid conditions.
- KOH-based slurry for compound semiconductors, pH 12–13: CXT preferred. Strong alkali at pH above 12 degrades both FR-4 and G-10 over time.
How Each Material Fails in Service
Understanding the failure progression of FR-4 and G-10 in out-of-envelope chemical environments helps predict template replacement timing and identify early warning signs before a process excursion occurs. The failure sequence for laminate materials in acidic slurry is consistent and observable.
Template performs within dimensional specification. Slurry contact at the work-hole surfaces and carrier plate periphery begins gradual epoxy resin attack, but the rate is slow enough that no measurable dimensional change occurs.
Visible yellowing or darkening of the epoxy resin at the machined work-hole sidewall surfaces. This is the first observable sign of acid attack. Dimensional tolerance is still within spec; template can continue in service but replacement should be planned.
Swelling of the epoxy matrix at the work-hole sidewall creates a measurable reduction in work-hole diameter — typically 5–15 µm. This tightens the wafer-to-hole clearance, increasing lateral retention force beyond design intent and creating stress concentrations at the wafer edge. TTV begins to show a systematic shift associated with template-induced pressure variation.
The acid penetrates to the glass fabric / resin interface and begins attacking the silane coupling agent that bonds the resin to the glass fibers. Micro-delamination initiates, visible as white or translucent blisters between laminate layers at the carrier plate periphery. Once delamination begins, it propagates rapidly.
Delaminated material and liberated glass fibers shed into the polishing slurry. These particles cause scratch defects on the wafer surface and contaminate the slurry bath. Carrier plate bow increases as the laminate structure loses coherence. Template is a process hazard and must be removed from service immediately.
The same failure sequence occurs for G-10 in pH-incompatible environments, but the cycle onset of each stage is approximately 20–40% later due to the marginally better acid resistance of the non-halogenated epoxy. For CXT-grade templates, this failure mode does not exist: there is no laminate interface to delaminate and no epoxy matrix susceptible to acid attack.
Edge Treatment: Why It Matters More Than Material Grade
In practice, for silicon polishing with alkaline slurry, the quality of the edge treatment on an FR-4 template is a more important performance variable than whether FR-4 or G-10 was chosen as the base material. This is because the primary contamination risk from both materials in normal service is not chemical degradation of the bulk epoxy — it is mechanical shedding of glass fibers from machined edges.
Both FR-4 and G-10 are woven glass fabric composites. When a router bit or end mill cuts through the material to create the work-hole pocket or the outer carrier plate profile, the cutting action severs individual glass fiber bundles at the cut surface. If these fiber ends are left exposed, they can fray during polishing and release sub-micron glass particles directly into the slurry stream at the wafer surface. A single glass fiber fragment of 0.3–1.0 µm diameter is sufficient to leave a scratch on a 300 mm prime silicon wafer that fails surface inspection.
The solution is edge treatment: a precision finishing operation applied to all machined surfaces before backing pad lamination. At Jizhi, this consists of a three-step sequence applied to every template regardless of whether FR-4 or G-10 is the specified material:
All work-hole and outer profile surfaces are finish-milled to drawing dimensions using carbide end mills with controlled feed rate and cutting speed to minimize heat-induced fiber pull-out and achieve maximum surface quality at the cut edge.
All machined edges are inspected at 20–40× magnification for fiber fraying, delamination, and dimensional conformance. Any template showing visible fiber exposure beyond the specified limit is rejected before proceeding to the next step.
A thin coat of chemically compatible epoxy sealant is applied to all machined edges by precision brush or spray, encapsulating any exposed fiber ends. The sealant is cured under controlled temperature and then inspected for complete coverage and absence of runs or voids that could introduce particles in service.
Machining & Fabrication Considerations
FR-4 and G-10 are both machinable with standard CNC tooling, but their glass fabric reinforcement creates specific tooling and process requirements that distinguish them from pure polymer materials. Understanding these requirements helps evaluate supplier fabrication quality and interpret the dimensional tolerances that are achievable in production.
Tooling and Feed Rates
The woven glass fabric in both laminates is highly abrasive and causes rapid wear of conventional high-speed steel tooling. Carbide or diamond-coated carbide end mills are standard for production template machining. Feed rate and cutting speed must be balanced to minimize heat generation (which causes epoxy softening and fiber pull-out) while maintaining dimensional precision. Typical parameters for work-hole machining are surface speeds of 100–180 m/min with feed rates of 0.05–0.15 mm/tooth, adjusted for cutter diameter and work-hole depth.
Dimensional Tolerance Achievability
With proper tooling and process control, work-hole depth tolerances of ±5 µm and diameter tolerances of ±10 µm are routinely achievable in FR-4 and G-10 on CNC machining centers with temperature-controlled work fixtures. Carrier plate flatness (bow) of ≤10 µm across the working surface requires starting with a flat-lapped raw material panel and managing thermal input during machining to prevent stress-induced warp. For specifications tighter than ±3 µm on work-hole depth, in-process CMM verification and closed-loop CNC compensation are used.
CXT Machining Differences
CXT-grade materials machine similarly to G-10 in terms of tooling and feed parameters, but the seamless construction means that there is no laminate layer interface that could delaminate under cutting forces. This makes CXT somewhat more forgiving of aggressive cutting parameters and allows faster material removal rates without the delamination risk that limits aggressive machining of laminates. Edge sealing is not required for CXT because there is no glass fabric to expose at cut surfaces.
When Neither FR-4 Nor G-10 Is Sufficient: CXT Grade
Both FR-4 and G-10 are laminate materials — stacks of glass fabric layers bonded by resin, with discrete layer interfaces running through the thickness of the plate. This laminate structure is the fundamental source of their chemical vulnerability: once acid or oxidant chemistry penetrates the outer resin surface and reaches the fiber-resin interface, delamination propagates rapidly between layers, and the structural integrity of the carrier plate deteriorates quickly.
CXT-grade templates address this at the structural level by eliminating the laminate construction entirely. CXT is a seamless, monolithic material with a homogeneous cross-section — there are no layer interfaces to delaminate, no fiber bundles to expose at machined edges, and no epoxy matrix susceptible to the specific chemical attack mechanisms that limit FR-4 and G-10. The matrix resin is selected from inert polymer families that maintain dimensional stability across the full pH 2–13 range, including in the presence of strong oxidants.
The manufacturing implications of seamless construction extend beyond chemical resistance. Because CXT templates are not laminate stacks, thickness uniformity across the carrier plate is achieved through precision machining rather than laminate pressing — giving tighter control over carrier plate bow for applications where ≤5 µm flatness is required. The absence of a fiber reinforcement phase also means that there is no differential CTE between fiber and matrix that can cause micro-cracking under thermal cycling.
The trade-off is cost and lead time: CXT templates are custom-fabricated items with longer production cycles than FR-4 or G-10 catalog templates. For applications where they are required — SiC CMP, aggressive oxide CMP, certain compound semiconductor polishing processes — this cost is non-negotiable. For applications where FR-4 or G-10 is chemically adequate, specifying CXT adds cost without process benefit. The full engineering case for SiC-specific template requirements is covered in our SiC wafer polishing templates guide.
Material Selection Matrix by Application
The following matrix consolidates the selection guidance from all previous sections into a quick-reference format organized by semiconductor polishing application. For applications not listed here, follow the pH and oxidant selection logic from Section 4, or contact our engineering team for an application-specific recommendation. For a broader understanding of how material selection fits into the complete specification process, see our 6-parameter template specification guide.
Common Material Selection Mistakes
Mistake 1: Defaulting to FR-4 for Every Application Without Checking Slurry pH
FR-4 is the lowest-cost option and the correct default for alkaline silicon polishing. But it is also the most commonly mis-specified material for non-alkaline applications. Engineers who specify templates primarily based on dimensional requirements and leave material selection to “standard FR-4” without verifying slurry chemistry compatibility create template failure timelines of 40–60 cycles rather than the 100–200+ cycles achievable with the correct material. The template replacement cost and process disruption are typically far higher than the cost difference between FR-4 and G-10 or CXT.
Mistake 2: Using G-10 as a Conservative “Upgrade” When CXT Is Required
G-10 is meaningfully better than FR-4 in mildly acidic environments. It is not meaningfully better than FR-4 in strongly acidic or oxidant-containing environments. For SiC CMP with KMnO₄ slurry at pH 2–4, G-10 fails at approximately the same cycle count as FR-4 — perhaps 15–20% later, but still catastrophically early compared to CXT. Specifying G-10 as a conservative upgrade for SiC applications is a false economy; only CXT provides genuine chemical resistance in that environment.
Mistake 3: Ignoring the Slurry Oxidant Component When Selecting Material
pH is a good primary filter for material selection, but oxidant chemistry is an independent variable that overrides pH-based decisions. A slurry at pH 7 (neutral) that contains 2% H₂O₂ is more aggressive toward FR-4 and G-10 epoxy matrices than a pH 5 slurry with no oxidant. Engineers who select material based on pH alone without checking oxidant components will find that templates fail far earlier than the pH-based prediction suggests. Always provide the complete slurry chemistry — pH, oxidant type, oxidant concentration, any chelating or surfactant additives — when requesting a material selection recommendation.
Mistake 4: Accepting Templates Without Specifying or Verifying Edge Treatment
The most common cause of glass fiber contamination in polishing operations is not material grade — it is inadequate edge sealing on otherwise acceptable FR-4 or G-10 templates. A G-10 template with poor edge treatment will shed more contamination in service than an FR-4 template with excellent edge sealing. When qualifying a new template supplier or a new template design, always include a wafer-level particle count test in the first qualification lot — this is the only reliable way to verify that edge treatment quality meets production requirements.
