Aluminaworld is a Zibo, Shandong-based Chinese manufacturer (ISO 9001 certified) producing calcined alumina (Al2O3 99.5-99.8%, alpha-phase 95%+) for low-temperature co-fired ceramic (LTCC) and high-temperature co-fired ceramic (HTCC) multilayer ceramic substrates used in 5G RF modules, automotive ECU power modules, and LED heat-spreader layers, with 15+ years export experience serving 60+ countries.
This guide targets procurement engineers, tape process engineers, and quality managers who must hit ±0.2% co-fired shrinkage reproducibility on 50-200 mm LTCC and HTCC substrates and need to qualify a calcined alumina lot. We cover the four alpha-phase / PSD / surface-area / binder-burnout levers that govern shrinkage, the cost-per-m2 trade-off between LTCC tape and HTCC tape, and eleven procurement-and-process FAQs drawn from production audits across 60+ countries.
A multilayer ceramic substrate is not a single material but a system: it is a stack of 4 to 40 ceramic green tapes, each screen-printed with conductive vias and traces, laminated under isostatic pressure, and co-fired so that the ceramic and the metal (silver, silver-palladium, or tungsten-molybdenum) densify in a single thermal cycle. The substrate has to deliver three simultaneous properties: dielectric constant matched to the RF or power circuit design, thermal conductivity high enough to remove the IGBT or LED junction heat, and dimensional tolerance tight enough for downstream pick-and-place, wire bonding, or solder-paste printing. Each of these three properties is sensitive to the alumina filler, but in different ways, and the dominant failure mode in production is not "the alumina is dirty" but "the alumina varies lot to lot by enough to push the fired substrate out of tolerance". The first conversations with a new supplier almost always focus on price and lead time; the second conversation, six months into production, almost always focuses on lot-to-lot consistency. Plan the procurement conversation accordingly.
The four levers that govern lot-to-lot consistency are alpha-phase fraction (XRD internal-standard method), particle size distribution (laser diffraction, d10, d50, d95), BET surface area (ISO 9277:2022 nitrogen adsorption), and tap density (ASTM B527). Each lever contributes a known, measurable, lot-to-lot variation that maps linearly (or near-linearly) to post-sintering shrinkage. A supplier qualification that measures all four and locks them under statistical process control will deliver ±0.15% shrinkage reproducibility with no further effort. A supplier qualification that measures only one or two will deliver ±0.35% or worse and force the substrate manufacturer into weekly fabrication, more frequent SPC re-baselining, and a higher scrap rate that erodes the supposed cost advantage of the cheaper alumina.
The four levers and their relative impact on shrinkage are described in section 2 below in the recipe format used by most substrate manufacturers' process-engineering teams. The recipe is a six-step loop; each step catches a class of variation that the previous step would let through. Run all six on every new supplier qualification and at every quarterly re-qualification of an incumbent supplier. Sections 3 through 6 cover the cost-per-m2 trade-off between LTCC and HTCC tape, the relevant industry standards (IPC-6012D, IEC 60268-7, ASTM F1471, SEMI G96, AEC-Q100/Q104), the seven CoA numbers that matter, and the procurement specs a qualified alumina shipment should meet.
A multilayer ceramic substrate is a stack of 4 to 40 ceramic green tapes, each screen-printed with conductive vias and traces, laminated under isostatic pressure, and co-fired so that the ceramic and the metal (silver, silver-palladium, or tungsten-molybdenum) densify in a single thermal cycle. The substrate has to deliver three simultaneous properties: dielectric constant matched to the RF or power circuit design, thermal conductivity high enough to remove the IGBT or LED junction heat, and dimensional tolerance tight enough for downstream pick-and-place, wire bonding, or solder-paste printing. Each of these three properties is sensitive to the alumina filler, but in different ways, and the dominant failure mode in production is not "the alumina is dirty" but "the alumina varies lot to lot by enough to push the fired substrate out of tolerance".
LTCC (low-temperature co-fired ceramic, sinter 850-900 degC) uses alumina as a filler and CTE anchor inside a glass-ceramic matrix. The glass-ceramic flows first, encapsulates the alumina grains, and carries most of the shrinkage load. Alumina controls the coefficient of thermal expansion (typically 5-7 ppm/K, matched to silicon and to the copper or aluminium heat-spreader the substrate is bonded to) and pins the post-sintering geometry. HTCC (high-temperature co-fired ceramic, sinter 1500-1650 degC) uses alumina as the primary load-bearing phase with a small glassy-grain-boundary phase to assist densification. Thermal conductivity of HTCC (24-30 W/mK for 95% alumina) is 4-8x that of LTCC (3-5 W/mK), which is why power-electronics and LED heat-spreader applications choose HTCC despite the higher furnace energy cost.
In both systems the calcined alumina grade is the single largest variable the procurement engineer can control to stabilise shrinkage. Four calcined-alumina properties matter, in order of impact on shrinkage reproducibility:
All four properties are measured routinely in qualified-supplier audits. What is not routine is the linkage between incoming-alumina variation and post-sintering shrinkage variation. The recipe below closes that linkage.
The recipe is a six-step loop. Each step catches a class of variation that the previous step would let through. Run all six on every new supplier qualification and at every quarterly re-qualification of an incumbent supplier.
The alpha-phase fraction is the single highest-leverage property. HTCC tape needs alpha fraction above 95% to avoid the additional 13-18% volumetric shrinkage that transition alumina dehydroxylation would produce mid-firing. LTCC tape can tolerate 70-90% alpha because most of the shrinkage is carried by the glass matrix, but the residual alpha still anchors post-sintering geometry and binder-burnout blistering risk. Insist on an XRD report (Cu-K-alpha, scan step 0.02 deg, internal standard such as 10 wt% silicon or 20 wt% corundum) with each lot, not just an XRF total-Al2O3 number. A CoA showing 99.8% total Al2O3 by XRF but only 60% alpha fraction indicates the calciner temperature was below the alpha-conversion threshold, and the batch will produce non-reproducible shrinkage in your process. Reject the lot or negotiate a discount.
Top-cut control is the second-highest leverage property. Top-cut d95 above 10 micron in any LTCC or HTCC tape produces pinhole defects in 10-50 micron via metallisation after screen printing. Run a wet-dispersion laser-diffraction PSD (e.g., Malvern Mastersizer, Horiba LA-960, or Beckman Coulter LS) with the supplier-reported dispersant, sonicate for 60 sec before measurement to break soft agglomerates, and report d10, d50, d95. Reject the lot if d95 is above your spec, regardless of how good d10 and d50 look.
Run a multi-point BET (ISO 9277:2022) on every incoming lot. Single-point BET is faster but tends to under-report the surface area of high-BET alumina by 5-10% because the single-point method assumes a known cross-sectional area of the adsorbate. The multi-point value is the one you correlate to your shrinkage control chart. If the supplier CoA shows BET 1.5 m2/g and your incoming measurement shows 3.0 m2/g, the lot will drive your shrinkage outside ±0.2% and you should reject it. We have seen suppliers ship a 1.5 m2/g spec lot and then on the next shipment deliver a 2.8 m2/g lot without changing the CoA — multi-point BET on your incoming dock catches this.
After the first three verifications pass, cast six 50 mm x 50 mm green tiles from your standard slurry formula using the incoming lot, fire them in your production furnace with the standard profile, and measure pre-fire and post-fire dimensions with a calibrated toolmaker's microscope (or optical comparator at 10x) to ±2 micron. Plot the six shrinkage values on an X-bar-R control chart against your historical control limits. If the mean shrinkage moves above ±0.15% from your centreline or any single tile is above ±0.30%, hold the lot pending investigation. The six-tile monitor is the only test that integrates all the variation sources (alumina, slurry, casting, drying, lamination, burnout, sintering) and gives you a single yes/no signal.
A furnace that drifts 10 degC at the dwell setpoint will shift your average shrinkage by about 0.2% absolute, and a furnace that has hot and cold zones will produce bimodal shrinkage distribution (some tiles shrink 14%, others 16%) that no incoming inspection can rescue. Run a nine-point thermocouple mapping (top, middle, bottom x front, centre, back) at your standard firing profile every quarter. If uniformity drifts above ±5 degC at the dwell setpoint, schedule a furnace refurbishment before the next qualification run.
The two production-environment variables that most often get overlooked are humidity equilibration in the green tape storage room (above 60% RH the binder plasticises and green density drops 1-2%, driving shrinkage up about 0.3%) and lamination pressure decay on the isostatic laminator (below 20 MPa leaves 1-3% residual porosity that compresses during sintering and produces a wider shrinkage distribution). Log humidity and lamination pressure per shift and review the control charts weekly. A ±5% RH drift in the storage room is detectable in the shrinkage monitor within 24 hours; a 5 MPa decay in lamination pressure shows up within a week.
The table below summarises the typical alumina-filled formulation and the post-sintering properties across the three competing multilayer ceramic substrate systems. Numbers are industry-typical ranges for 2025-2026 production-grade tape; actual values vary by manufacturer and by lot.
| Property | LTCC tape (850-900 degC) | HTCC tape (1500-1650 degC) | AlN tape (1600-1800 degC) |
|---|---|---|---|
| Alumina filler content (vol%) | 30-50% | 85-95% | N/A (AlN primary) |
| Glass-ceramic binder (vol%) | 50-70% (Ca-B-Si or Ba-Al-Si glass-ceramic) | 5-15% (silicate binder) | Y2O3 or CaO sintering aid |
| Required alpha-Al2O3 fraction | 70-90% | >95% | N/A |
| Required BET (m2/g) | 1.5-3 | 0.5-2 | N/A |
| Required d50 (micron) | 0.8-2 | 1.5-3 | N/A |
| Co-fire temperature (degC) | 850-900 | 1500-1650 | 1600-1800 |
| Dielectric constant (1 MHz) | 5.5-8.0 | 9.0-10.0 | 8.5-9.5 |
| Dielectric loss (Df at 1 GHz) | 0.002-0.005 | 0.0005-0.001 | 0.001-0.005 |
| Thermal conductivity (W/mK) | 3-5 | 24-30 | 150-200 |
| CTE (ppm/K, 25-300 degC) | 5-7 | 6-8 | 4-5 |
| Co-fired metal | Ag, Ag-Pd, Au | W, Mo, Mo-Mn | W, Mo (inert atmosphere) |
| Green tape cost (USD/m2) | 80-150 | 40-80 | 200-400 |
| Furnace energy (relative) | 1x | 4-6x | 5-8x (also inert atm) |
| Best-fit application | 5G RF front-end modules, automotive ECU signal layer | IGBT power modules, LED heat spreaders | High-power-density EV inverter substrates |
Source: industry-typical ranges from publicly available datasheets (Dupont 951, Heraeus A200, NTK FC-2, Ferro A6M equivalents) and Aluminaworld production audits 2025-2026. Numbers are representative and may vary ±10% by lot and by manufacturer.
A common question from procurement teams who already buy tabular alumina for castable refractories is whether they can substitute it for calcined alumina in their LTCC or HTCC tape. The short answer is no, and the reason is grain morphology. Calcined alumina (rotary-kiln product at 99.5-99.8% Al2O3) is an angular, porous grain with high surface area (BET 1-3 m2/g) and rough surface texture. The angular shape and surface roughness interlock in the green tape and produce high green density (60-65% by volume). Tabular alumina (sintered alpha-alumina at >99% density) is a rounded, dense grain with low surface area (BET <0.5 m2/g) and smooth surface. The rounded shape and smooth surface do not pack to the same green density, and the low surface area does not anchor the binder as effectively.
The result is a green tape with 3-5% lower packing fraction, which translates to higher porosity after sintering and a 15-25% drop in thermal conductivity in the fired substrate. The substrate is mechanically weaker and the dielectric loss is slightly higher because the residual porosity scatters the RF signal. For LTCC tape in particular the difference is acute because the glass-ceramic matrix needs the alumina filler to provide nucleation sites for crystallisation during the glass-ceramic conversion, and calcined alumina's rough surface provides those nucleation sites more effectively than tabular alumina's smooth surface. Tabular alumina is the right choice in castable refractories, dense ceramic load-bearing parts, and ball-mill grinding media, but the wrong choice for thin multilayer ceramic tape where green packing and post-sintering flatness are the constraints.
A reputable calcined-alumina CoA for LTCC/HTCC tape should report the following seven numbers in this order. Numbers outside the ranges below are not necessarily disqualifying, but they trigger additional incoming inspection or a process-control review.
| Number | Method | LTCC tape typical range | HTCC tape typical range | What happens if outside range |
|---|---|---|---|---|
| Total Al2O3 (wt%) | XRF or ICP | >99.5% | >99.7% | Higher impurity load = lower thermal conductivity, glass-ceramic devitrification risk |
| Alpha-phase fraction (%) | XRD internal-standard | 70-90% | >95% | Each 10% shortfall moves shrinkage 0.3-0.5% absolute; binder-burnout blistering below 60% |
| BET surface area (m2/g) | ISO 9277:2022 multi-point N2 | 1.5-3.0 | 0.5-2.0 | Higher BET requires more dispersant, lowers solids loading, raises drying-shrinkage anisotropy |
| d10 (micron) | Laser diffraction | 0.3-0.6 | 0.4-0.7 | Below 0.2 micron = excess dispersant demand; above 0.8 micron = poor green packing |
| d50 (micron) | Laser diffraction | 0.8-2.0 | 1.5-3.0 | Above spec = slower sintering densification, higher residual porosity |
| d95 (micron) | Laser diffraction | <6 | <10 | Above spec = pinhole defects in 10-50 micron via metallisation |
| Tap density (g/cm3) | ASTM B527 | 0.8-1.0 | 0.9-1.1 | Below 0.7 = high grain porosity, absorbs more binder, non-uniform drying |
Range values are industry-typical for 2025-2026 production-grade LTCC and HTCC tape; exact limits should be set during a six-lot supplier qualification against your process control chart.
The cost-per-substrate decision is application-driven, not purely material-driven. The table below maps the three most common multilayer ceramic substrate use cases to the cost-relevant factors.
| Application | Dominant requirement | Recommended system | Typical alumina filler cost share |
|---|---|---|---|
| 5G RF front-end module | Low Df at 28-39 GHz | LTCC (Ag conductor) | 15-20% of tape cost |
| Automotive ECU signal layer | CTE match to silicon, 150 degC operating | LTCC | 15-20% of tape cost |
| IGBT power module heat spreader | Thermal conductivity >20 W/mK | HTCC (W or Mo-Mn conductor) | 35-45% of tape cost |
| LED COB heat spreader | Thermal conductivity >15 W/mK, <1 W/mK variation across 50 mm | HTCC | 35-45% of tape cost |
| EV inverter substrate | Thermal conductivity >150 W/mK | AlN tape (or AMB silicon-nitride bonded copper) | N/A (AlN is the matrix) |
| Sensor substrate (lambda, NOx) | Chemical durability, 600 degC operating | HTCC | 35-45% of tape cost |
Cost-share percentages are estimates based on industry-typical material splits 2025-2026 and may vary with conductor cost (silver paste in particular has ranged 800-1500 USD/kg over the last 3 years).
The relevant standards for multilayer ceramic substrate dimensional tolerance, microstructural characterisation, and supplier qualification are listed below. None of them specifies a ±0.2% shrinkage number; the ±0.2% is an industry-typical control limit derived from six-lot supplier qualification plus ongoing SPC control charts.
When your SPC chart shows a single tile outside ±0.30% shrinkage, the first instinct is to reject the incoming alumina lot. That is the right call only after you have ruled out the four production-side root causes: (a) humidity excursion in the green-tape storage room (above 60% RH), (b) lamination-pressure decay on the isostatic laminator (below 20 MPa), (c) furnace temperature uniformity drift (above ±5 degC), and (d) slurry-mixing variation (dispersant dose, milling time, casting speed). Each of those four variables moves shrinkage 0.2-0.5% in the same direction as a bad alumina lot. Walk the six-step recipe in reverse on every shrinkage alarm; if the production-side variables all check clean, then the alumina lot is the suspect and a single six-tile monitor is not enough to reject — repeat with twelve tiles and verify reproducibility before holding the supplier responsible.
| Standard | Scope | Where it applies |
|---|---|---|
| ISO 9277:2022 | BET surface area by nitrogen adsorption | Incoming calcined alumina QC |
| ASTM B527 | Tap density of metal powders | Incoming calcined alumina QC |
| ISO 787-2 | Loss on drying | Incoming calcined alumina QC |
| ASTM F1471 | Camber and twist of ceramic substrates | Post-sintering dimensional QC |
| IPC-6012D | Qualification and performance of rigid printed boards | Surface finish and trace integrity |
| IEC 60268-7 | Ceramic dielectric materials | Dielectric property qualification |
| SEMI G96 | High-purity alumina for semiconductor applications | High-end LTCC/HTCC raw material |
| AEC-Q100/Q104 | IC stress test qualification | Automotive power-module substrates |
If you are starting a new LTCC or HTCC tape line, qualify a calcined alumina supplier with the six-step recipe above. If you are already in production and your shrinkage control chart shows drift, walk the six steps in reverse — start with the shrinkage monitor (Step 4), confirm furnace uniformity (Step 5) and green-tape environment (Step 6), and only then investigate incoming alumina lot-to-lot variation (Steps 1-3).
Aluminaworld supplies LTCC-grade and HTCC-grade calcined alumina with the following standard specs and can customise PSD and alpha-fraction within the ranges above for tape-casting and multilayer ceramic substrate applications. Contact us below for a 5 kg free sample, a six-lot qualification batch protocol, or a joint audit at your facility.
For calcined alumina for LTCC / HTCC tape, pseudo-boehmite for catalyst binders, activated alumina for hydrogen and gas drying, or molecular sieves for gas and liquid purification, contact our engineering team:
WhatsApp: +86 133 2522 2240
Email: barry@aluminaworld.com
Related products: Calcined Alumina · Alumina Powder · Pseudo Boehmite · Activated Alumina
Related articles: Pseudo Boehmite for FCC Catalyst · Spherical vs Extrudate Catalyst Carrier · Tabular vs White Fused Alumina · Reactive Alumina for Castables