How to Read a PCB Laminate Datasheet: Dk, Df, Tg, CTE, and What Actually Matters for Your Design

Every PCB starts as a material selection decision, and that decision starts with a laminate datasheet. Yet the document that determines whether your board will perform reliably at operating frequency, survive lead-free assembly without delamination, and meet impedance targets within tolerance is routinely misread, partially understood, or ignored entirely until something goes wrong in production.

In our fabrication facility, we estimate that roughly three out of every ten engineering reviews reveal a material specification mismatch — an engineer who selected a laminate based on one attractive number without verifying the parameter that actually mattered for their design. The most common example is specifying high-Tg FR-4 for a low-frequency consumer board because “higher is better,” adding 15 percent to the material cost while the real risk factor — moisture absorption for a product shipping to tropical climates — was never checked on the datasheet at all.

This guide walks through every section of a PCB laminate datasheet from a manufacturer’s perspective, explains which parameters matter for which applications, and identifies the traps that cause costly mistakes in procurement and production.

Anatomy of a Laminate Datasheet

A laminate datasheet is a technical document published by the material manufacturer — companies like Shengyi, ITEQ, Isola, Panasonic, and Rogers — that reports the measured or guaranteed properties of a specific product grade. Understanding the document’s structure is the first step to reading it correctly.

The header section identifies the product family and grade designation. This is where you find the product name (such as Shengyi S1000-2M or ITEQ IT-180A), the UL designation number, and the applicable standards compliance (typically IPC-4101 slash sheet numbers). Pay close attention to the exact grade designation — a single product family often has multiple variants with different performance characteristics. Shengyi’s S1000 family, for example, includes S1000, S1000-2, S1000-2M, and S1000H, each with meaningfully different Tg, Dk, and processability characteristics.

The property tables form the core of the document and are organized into electrical, thermal, and mechanical sections. Each row reports a specific parameter alongside its test method (typically an IPC-TM-650 reference or ASTM standard), test conditions (frequency, temperature, humidity), and the reported value. The test conditions column is arguably more important than the value itself, because a Dk of 4.2 measured at 1 MHz tells you something fundamentally different from a Dk of 4.2 measured at 10 GHz.

Footnotes and qualifications at the bottom of the datasheet contain critical information that is easy to overlook. These notes often specify the resin content of the test specimen, the copper foil type used for peel strength testing, conditioning requirements before measurement, and tolerance ranges that may not appear in the main table. A manufacturer reporting Dk “at 50 percent resin content per IPC-TM-650 2.5.5.9” is following standard practice, but your actual board may use glass styles that produce 42 percent or 62 percent resin content, shifting the effective Dk by as much as 0.3 units.

Electrical Parameters: The Numbers That Determine Signal Performance

Dielectric Constant (Dk)

Dielectric constant, written as Dk or the Greek letter epsilon-r, quantifies the material’s ability to store electrical energy relative to a vacuum. In practical PCB terms, Dk controls signal propagation velocity and the trace geometry required for a target characteristic impedance. A higher Dk means slower signals and narrower traces for a given impedance target.

Standard FR-4 typically reports Dk between 4.2 and 4.8, depending on the specific grade and measurement frequency. This range is wide enough to cause real problems. An engineer who assumes Dk = 4.3 for impedance calculations on a production board may find that the fabricator’s actual material — a different FR-4 grade with Dk of 4.6 at the relevant frequency — produces traces that miss the impedance target by 7-8 percent. That is nearly the entire tolerance budget for a controlled-impedance design.

What matters more than the absolute Dk value is Dk stability across frequency, temperature, and resin content variation. Premium laminates like Rogers RO4350B specify Dk tolerance of plus or minus 0.05, while standard FR-4 grades may vary by plus or minus 0.3 or more. When reading the datasheet, always note the test frequency — Dk reported at 1 MHz is essentially useless for any design operating above 1 GHz, because FR-4’s Dk drops measurably as frequency increases. Request frequency-swept data from the material supplier if the datasheet only provides a single frequency point.

The resin content issue deserves special emphasis. Dk varies with the ratio of glass fiber to resin in the laminate, and prepreg Dk differs from core Dk for the same product family because they have different glass styles and resin percentages. When your fabricator builds a stackup simulation, they use “press-out” Dk values specific to each prepreg and core layer in your construction — not the single headline number from the datasheet. This is why specifying a Dk value in your fab notes to two decimal places (such as “Dk = 4.25”) without referencing a specific material grade and glass style creates confusion rather than clarity.

Loss Tangent (Df)

Loss tangent, written as Df or tan-delta, measures the fraction of electromagnetic energy converted to heat within the dielectric per signal cycle. In engineering terms, higher Df means more insertion loss in your traces, which directly limits achievable channel length at a given data rate.

Below approximately 500 MHz, Df is rarely the performance bottleneck even for standard FR-4 with Df around 0.018 to 0.025. Above 1 GHz, dielectric loss becomes a first-order design constraint. Above 10 GHz, the difference between standard FR-4 (Df approximately 0.020) and an ultra-low-loss material like Isola Astra MT77 (Df approximately 0.0017) determines whether a millimeter-wave design functions or fails completely.

When comparing Df values across datasheets, verify that the measurements use the same test method. Df measured by stripline resonator may differ from Df measured by split-post cavity on the same physical material. Mixing test methods across suppliers in a comparison spreadsheet produces conclusions that look data-driven but are actually misleading. Also note that Df, like Dk, is frequency-dependent and typically increases at higher frequencies — a material reporting Df of 0.003 at 1 GHz may measure 0.005 at 10 GHz.

For a detailed comparison of Dk and Df across high-speed material families, including Megtron, I-Speed, and TU-872 series laminates, our material comparison guide provides side-by-side data at multiple frequency points.

Thermal Parameters: What Determines Assembly Survival and Long-Term Reliability

Glass Transition Temperature (Tg)

Tg is the temperature at which the epoxy resin matrix transitions from a rigid, glassy state to a softer, more flexible condition. This transition is reversible — the material returns to its original mechanical properties when it cools below Tg — but the consequences of operating above Tg during processing are significant. Above Tg, the Z-axis coefficient of thermal expansion increases by a factor of three to five, placing substantial stress on plated through-hole barrels and via interconnections.

Standard FR-4 has Tg typically around 130 to 140 degrees Celsius. Lead-free solder reflow peaks reach 245 to 260 degrees Celsius. Every board going through lead-free assembly spends time above Tg, which means higher-Tg materials reduce but do not eliminate the thermal excursion above the transition point. High-Tg FR-4 (Tg above 170 degrees Celsius) is the standard specification for lead-free compatible boards, but the widespread assumption that higher Tg always equals better reliability is an oversimplification.

In our production experience, specifying Tg 170 when the application only requires Tg 150 (such as a board that will only see wave soldering at 260 degrees Celsius peak for a few seconds) adds approximately 10 to 20 percent to material cost with no measurable improvement in field reliability for that specific use case. The more important question is often the Td value.

Decomposition Temperature (Td)

Td measures the temperature at which the resin begins to decompose irreversibly — losing 5 percent of its mass through chemical breakdown. Unlike Tg, Td represents a permanent damage threshold. A material with Tg of 150 degrees Celsius but Td of 350 degrees Celsius may actually be more reliable through multiple reflow cycles than a material with Tg of 180 degrees Celsius but Td of only 310 degrees Celsius, because the higher-Td material has more thermal margin before irreversible damage begins.

The related parameters T260 and T288 measure how many minutes the material can withstand sustained exposure at 260 and 288 degrees Celsius before delamination occurs. For boards requiring multiple reflow passes — double-sided SMT assembly or rework scenarios — T260 time greater than 30 minutes is a practical minimum, and many high-reliability specifications require T288 greater than 15 minutes. These parameters are more directly predictive of assembly survival than Tg alone.

Coefficient of Thermal Expansion (CTE)

CTE values on a laminate datasheet are reported separately for the X/Y plane (in-plane) and the Z-axis (through-thickness), and often separately for temperatures below and above Tg. The Z-axis CTE above Tg is the critical number for via reliability — it determines how much the board thickness changes during thermal excursions through the reflow oven, directly stressing plated via barrels.

Standard FR-4 Z-axis CTE below Tg is typically 45 to 65 ppm per degree Celsius, increasing to 250 to 300 ppm per degree Celsius above Tg. High-performance laminates with filled resin systems can reduce the above-Tg Z-axis CTE to 150 to 200 ppm per degree Celsius, significantly improving via reliability for thick boards with high aspect ratio holes. If your design has a board thickness above 2.0 mm with drilled holes smaller than 0.3 mm, Z-axis CTE should be a primary material selection criterion.

Mechanical and Environmental Parameters

Peel Strength

Peel strength measures the adhesion between the copper foil and the laminate surface, reported in pounds per linear inch or Newtons per millimeter. This parameter matters for any design that will experience thermal cycling, mechanical flexing, or chemical processing (such as multiple plating cycles for HDI builds). Datasheets typically report peel strength after thermal stress (such as solder float at 288 degrees Celsius for 10 seconds), which simulates the adhesion you can expect after assembly rather than the initial bond strength before processing.

Low-loss laminates historically suffered from reduced peel strength because the smoother resin surface that reduces dielectric loss also provides less mechanical bonding area for copper adhesion. Modern materials have largely addressed this through surface treatment advances, but it remains important to verify peel strength after thermal stress when evaluating any unfamiliar low-Df material.

Moisture Absorption

Moisture absorption, reported as a weight percentage after 24-hour water immersion, indicates how much water the laminate absorbs under standardized conditions. Absorbed moisture degrades dielectric properties — increasing Dk by as much as 5-10 percent and Df even more significantly — and creates delamination risk during reflow when trapped moisture vaporizes rapidly.

For boards destined for humid operating environments or long warehouse storage before assembly, moisture absorption below 0.10 percent is preferable. Standard FR-4 typically reports 0.10 to 0.15 percent, while PTFE-based materials like Rogers RO4350B achieve 0.06 percent or lower. This is one of the most commonly overlooked datasheet parameters, yet it directly impacts field reliability for any product shipping to tropical or high-humidity markets.

CAF Resistance

Conductive Anodic Filament (CAF) growth is an electrochemical failure mechanism where copper migration along glass fiber interfaces creates conductive paths between adjacent features — vias, traces, or planes — eventually causing short circuits. CAF resistance testing evaluates a material’s susceptibility to this failure mode under accelerated conditions (high temperature, high humidity, applied voltage).

For dense HDI designs with via-to-via spacing below 0.5 mm or fine-pitch BGA fanout patterns, CAF resistance is a serious reliability consideration. Not all datasheets report CAF test results, and the absence of data should not be interpreted as adequate performance. If your design places vias within 0.3 mm of each other in a high-humidity application, request CAF test data from the material supplier.

The Decision Matrix: Which Parameters Matter for Your Application

Not every datasheet parameter is equally important for every design. The most expensive mistake in material selection is treating every specification as critical and choosing the highest-performing laminate across all categories. The second most expensive mistake is ignoring a parameter that turns out to be the binding constraint for your specific application.

For standard digital boards operating below 1 GHz — consumer electronics, industrial controls, IoT devices — Tg and Td are the primary concerns. Dk matters only if you have controlled impedance requirements, and even then, the fabricator’s stackup simulation accounts for Dk variation. Df is essentially irrelevant at these frequencies. Focus your datasheet review on thermal parameters, available thickness options, and moisture absorption if the product operates in harsh environments.

For high-speed digital designs from 1 to 28 Gbps — server interconnects, storage interfaces, networking equipment — Df becomes the dominant parameter because dielectric loss drives insertion loss budgets at these data rates. Dk stability (not absolute value) matters for impedance consistency. Z-axis CTE becomes important if the board is thick enough to create via reliability concerns. Review the material’s Dk and Df data at your actual operating frequency, not the 1 MHz headline number.

For RF and microwave designs above 5 GHz — 5G infrastructure, automotive radar, satellite communications — Dk tolerance, Dk stability versus frequency and temperature (TCDk), and Df are all critical. You need frequency-swept data, not single-point measurements. Moisture absorption matters because water’s high Dk (approximately 80) can shift your antenna tuning and filter characteristics. This is where the cost premium for Rogers and similar RF materials is justified.

For automotive applications requiring AEC-Q200 or similar qualification, focus on T260/T288 times (multiple reflow survival), CTE (thermal cycling reliability over the vehicle lifetime), and CAF resistance (humidity exposure in underhood environments). Tg is important but is often over-emphasized relative to Td and CTE in automotive procurement specifications.

For aerospace and military Class 3/3A applications, nearly every parameter on the datasheet is relevant, and you will likely need to supplement datasheet data with qualification testing specific to your program requirements. Peel strength after thermal stress, CAF resistance, moisture absorption, and long-term thermal aging data become as important as electrical and thermal parameters.

Common Datasheet Traps

Trap 1: Dk at the wrong frequency. A datasheet showing Dk = 4.4 at 1 MHz tells you almost nothing useful if your design operates at 10 GHz. FR-4 Dk typically drops 0.2 to 0.5 units between 1 MHz and 10 GHz. Always match the datasheet test frequency to your operating frequency, or request supplementary frequency-swept data from the supplier.

Trap 2: Confusing core Dk with prepreg Dk. Many engineers assume that the core and prepreg versions of the same product have identical Dk. They do not. A 1080 glass style prepreg has significantly higher resin content than a 7628 glass style core of the same product family, and the Dk difference can exceed 0.3 units. Your fabricator’s stackup must account for this layer-by-layer — confirm that your impedance model uses per-layer Dk values, not a single generic number.

Trap 3: “Equivalent” materials that are not equivalent. When a fabricator proposes a material substitution — your specified Isola 370HR replaced with their stocked ITEQ IT-180A — both may meet a generic “high-Tg FR-4” description, but their Dk, Df, CTE, and moisture absorption characteristics differ enough to affect controlled impedance results. Always review the substitute material’s full datasheet rather than accepting brand-name equivalence. In our production, we have seen impedance shifts of 3-5 percent between nominally equivalent materials from different manufacturers, which can push results outside specification for tight-tolerance designs.

Trap 4: Same brand name, different UL designations. Material manufacturers update their product formulations over time, sometimes changing the UL designation number without changing the product trade name. If your qualification testing was performed with a specific UL-designated version of a material, verify that the current production version carries the same UL number. A changed formulation may have measurably different Dk, Df, or thermal performance even though the marketing name is identical.

Trap 5: Ignoring the test conditions column. A CTE value reported “below Tg” versus “above Tg” can differ by a factor of five. A peel strength reported “as received” versus “after thermal stress” can differ by 30 percent or more. The number without its test condition is meaningless — always read the full row including method and conditions.

Cost Impact: When Over-Specifying Materials Costs You Money

Material cost typically represents 15 to 25 percent of total PCB fabrication cost, but the material choice can influence overall board cost by a much larger margin because it affects processing requirements, yield rates, and available fabricator options. Understanding the real cost multipliers helps engineers make informed trade-offs rather than defaulting to the most conservative specification.

From our production data, the approximate material cost multipliers relative to standard mid-Tg FR-4 are as follows. Standard FR-4 with Tg around 135 to 150 degrees Celsius represents the baseline at 1.0x. High-Tg FR-4 (Tg 170 and above, such as Shengyi S1000-2M or ITEQ IT-180A) runs approximately 1.10 to 1.20x. Mid-loss laminates for high-speed digital applications (Panasonic Megtron 4, ITEQ IT-968SE) cost approximately 1.5 to 2.0x. Low-loss laminates (Megtron 6, Isola I-Speed) run 2.5 to 3.5x. PTFE-based RF laminates (Rogers RO4350B, RO4003C) cost 3.0 to 5.0x depending on thickness. Ultra-low-loss ceramics and exotic materials can exceed 8.0x.

The cost impact extends beyond the raw material price. Premium materials often require longer procurement lead times — Rogers materials may add 5 to 10 business days if not stocked by the fabricator, while standard FR-4 is always available. Specialized materials may limit your fabricator options to shops with experience processing that specific material family, reducing competitive pricing leverage. And some materials require modified fabrication processes (different drill speeds, adjusted lamination cycles, special surface preparation) that add processing cost on top of the material premium.

The practical lesson is straightforward: specify the material performance your design actually requires, not the highest grade available. A consumer IoT board running at 100 MHz with two controlled impedance layers does not benefit from Megtron 6 — standard high-Tg FR-4 delivers identical functional performance at one-third the material cost.

What to Put in Your Fabrication Notes

The fabrication drawing is where your material selection becomes a manufacturing instruction. Getting this section right prevents substitution surprises, impedance failures, and unnecessary cost additions. Based on thousands of orders we have processed, here is what works well and what causes problems.

Specify performance requirements, not just brand names. Instead of writing “Material: Isola 370HR,” write “Material: Isola 370HR or equivalent meeting Tg greater than 170C (DSC), Dk 4.04 plus or minus 0.10 at 1 GHz, Df less than 0.012 at 1 GHz, Td greater than 340C.” This gives the fabricator flexibility to use stocked equivalent materials while ensuring the parameters that matter for your design are protected.

Call out the parameters that bind your design. If your impedance budget depends on Dk consistency, specify the Dk tolerance explicitly. If your thermal cycle reliability depends on Z-axis CTE, specify a maximum value. If your product ships to high-humidity environments, specify a maximum moisture absorption percentage. Do not list every datasheet parameter — that signals to the fabricator that you do not understand which constraints actually matter, and it eliminates flexibility that could reduce your cost and lead time.

Include an approved materials list when possible. For production orders, providing two or three pre-approved material options (such as “Shengyi S1000-2M, ITEQ IT-180A, or Isola 370HR”) gives the fabricator procurement flexibility while keeping your qualification boundaries clear. This approach typically saves 5 to 15 percent versus locking to a single material brand and reduces lead time by avoiding material-specific procurement delays.

Leave fabrication-specific parameters to the manufacturer. Drill speed, lamination pressure and temperature profiles, plasma desmear parameters, and similar process details should not appear in your fabrication drawing. These are manufacturing process controls that vary by equipment, lot, and facility — specifying them constrains the fabricator without improving your board quality and may actually prevent them from using their optimized process.

Putting It All Together

A PCB laminate datasheet is a dense technical document, but you do not need to understand every number on it to make sound material decisions. The key is knowing which parameters bind your specific design, reading those values at the correct test conditions, and communicating your requirements to the fabricator in a way that protects performance without unnecessarily restricting flexibility or inflating cost.

Start with your application requirements — operating frequency, thermal environment, reliability class, and cost target. Match those requirements to the relevant datasheet parameters. Verify the test conditions match your operating conditions. Then write fabrication notes that specify the parameters that matter and allow flexibility on the parameters that do not.

Material selection is not about finding the best laminate — it is about finding the right laminate for your specific design at the lowest total cost. The datasheet gives you the data to make that determination. Reading it correctly is the first step.

Reviewed by AtlasPCB Engineering Team