The density gap between a 3.6 g/cm³ and a 3.9 g/cm³ alumina ceramic part is rarely a measurement artefact; it is a fingerprint of purity, porosity and forming method, with theoretical alpha-Al₂O₃ fixed at 3.987 g/cm³ [S1][S3].
Commercial alumina substrates cluster between 3.6 and 3.95 g/cm³ depending on Al₂O₃ content and sintering practice, while the underlying alpha-alumina crystal sits at 3.987 g/cm³ (about 0.144 lb/in³) [S1][S3]. Any alumina grade sold as 3.6 g/cm³ is telling you about its glassy secondary phase and residual porosity, not about the Al₂O₃ molecule itself. For background on the wider alumina ceramic family and how it sits next to other oxides, the reference page is the natural starting point.
What 3.6 vs 3.9 g/cm³ Actually Tells You
A finished alumina part measured at 3.6 g/cm³ corresponds to a relative density near 90% against the 3.987 g/cm³ theoretical figure, while 3.9 g/cm³ lines up with roughly 98% [S1][S5]. That 8-point swing in relative density is the same 0.3 g/cm³ you see on the bulk scale, and it maps almost one-to-one onto porosity: 7-10% open plus closed porosity in a 3.6 g/cm³ part versus 2-3% in a 3.9 g/cm³ part [S5]. Sintering aids (MgO, SiO₂, CaO) and any secondary spinel or mullite phase pull the lattice average down, so lower purity grades physically cannot reach 3.9 g/cm³ without an unusual hot-press or HIP route [S1][S9]. The 3.6-3.9 g/cm³ gap is therefore not a tolerance band on one material; it is two different microstructures on the same Al₂O₃ backbone.
Density by Purity: A Working Spec Table
Published commercial data line up within tight bands once you separate forming method from chemistry. The table below consolidates the values most often quoted on supplier datasheets [S1][S3][S5][S7][S8]:
85% Al₂O₃, bulk density about 3.40 g/cm³ (relative density ~85%) [S5]. 94% Al₂O₃, about 3.70 g/cm³ [S1]. 95% Al₂O₃, 3.7 g/cm³ (white or pink body, water absorption near 0) [S3][S7]. 96% Al₂O₃, 3.72 g/cm³ (~93% relative) [S1][S5]. 97% Al₂O₃, 3.75 g/cm³ [S3]. 99% Al₂O₃, 3.8-3.9 g/cm³ depending on forming [S3][S7]. 99.5% Al₂O₃, 3.90 g/cm³ (~98% relative) [S1][S3][S5]. 99.7% Al₂O₃, 3.94 g/cm³ [S3]. 99.9% Al₂O₃, 3.95 g/cm³ (~99% relative) [S3][S5]. Sapphire / single-crystal Al₂O₃, 3.97-3.99 g/cm³, near theoretical, and not interchangeable with polycrystalline substrates [S1]. The practical takeaway: 3.6 g/cm³ on a data sheet means you are below 96% Al₂O₃ or the part was deliberately underfired; 3.9 g/cm³ means you are at 99% or above with a dense, low-porosity body.
Selection Criteria: Which Density to Specify
Three decision variables drive the choice between a 3.6 and a 3.9 g/cm³ alumina grade: electrical requirements, mechanical/thermal loading, and chemistry compatibility [S1][S3][S8].
Electrical: thick-film and general PCB substrates centre on 96% alumina at 3.70-3.75 g/cm³, where the small glassy phase is what lets the metallisation bond. Demanding dielectric or high-insulation-resistance work usually jumps to 99.5% (3.90 g/cm³) for lower loss and tighter tolerance. Mechanical and thermal: higher density correlates with higher compressive strength (around 2,000 MPa for a 3.75 g/cm³ grade, climbing past 3,500 MPa near 3.98 g/cm³) and tighter grain structure, which also reduces wear [S4]. Chemical and purity: 99.5% and above run under 0.1% porosity, supporting clean-room, medical and semiconductor uses; 95% (3.7 g/cm³) is the cost-effective workhorse for wear liners, insulating rings and general structural pieces [S3]. For broader industrial ceramic context, alumina's density advantage versus steel (about 7.85 g/cm³) is roughly a 2:1 mass saving at the same volume, which is why the 3.6-3.9 g/cm³ window is so often quoted in aerospace and rotating equipment [S3][S5].
Use-Case Recommendations
For a 3.6 g/cm³ part, the realistic fit is general technical ceramic duty: wear plates, pump seals, insulating bushings, threaded spacers and metal-replacement structural pieces where the 95-96% Al₂O₃ body is enough and cost per kg matters more than the last few percent of porosity [S1][S3][S7].
For a 3.9 g/cm³ part, target it where the dense microstructure pays back: high-purity insulators, protection tubes, semiconductor handling fixtures, medical-device components, and any high-humidity or vacuum application where open porosity would be a contamination source [S3][S8]. A protection tube or insulator datasheet that quotes 3.65-3.9 g/cm³ for a 99-99.5% grade is implicitly telling the buyer to expect forming-method sensitivity: dry-pressed parts sit at the low end, isostatically pressed or extruded bodies at the high end [S8]. Where the question is dielectric behaviour specifically, the 96% vs 99.5% trade-off is laid out in 96% vs 99.5% alumina dielectric strength spec comparison, which complements this density discussion.
Measurement and Definition Pitfalls
Engineers get burned when one number is quoted in three different units of meaning. Three definitions to keep separate on every alumina certificate of analysis [S1][S5]:
Theoretical density 3.987 g/cm³ for alpha-Al₂O₃, fixed by the crystal structure, useful only for relative-density maths. Bulk density is mass divided by the part's external envelope including all open and closed pores; this is what a substrate mass calculation or incoming inspection should use. Apparent solid density excludes open porosity, sits between theoretical and bulk, and is reported under a defined liquid-displacement method such as ASTM C373 or the Archimedes water-absorption test. The Archimedes route, using dry, saturated and suspended mass, is the most common way to back out both bulk density and apparent porosity on a fired part [S5]. Relative density, bulk divided by theoretical as a percentage, is the better predictor of strength and wear than the absolute bulk number, and it is what most quality clauses should reference [S5]. Gamma-alumina and activated alumina are porous, high-surface-area materials whose particle or powder density figures must not be substituted for a dense alpha-alumina ceramic [S1].
Limitations and What the Density Number Cannot Tell You
Density alone is a coarse filter. Two parts both reading 3.90 g/cm³ can have very different grain size, fracture toughness and dielectric loss if the secondary phase distribution is not the same [S1][S3].
Higher density usually means higher compressive and tensile strength, but it also means harder sintering, longer firing cycles and higher piece-price, so 3.95 g/cm³ is not automatically the right target for every drawing [S3]. For 95-97% grades, the small additional glassy phase actually helps thick-film metallisation bond to the substrate, which is why 96% alumina remains the PCB-substrate default even though 99.5% is denser [S1]. Purity above 99.5% keeps shrinking the gains: 99.5% to 99.7% to 99.9% adds only about 0.04-0.05 g/cm³ while raising cost and machining difficulty, so 3.9 vs 3.94 vs 3.95 g/cm³ is a meaningful material decision, not a noise-level spec [S3]. As one supplier table puts it, the 99% and above grades deliver "a nearly fully dense structure (porosity under 0.1%), ultra-high surface finish, suitability for precision machining, and low dielectric loss" [S3].
Sourcing and Standards Anchors
Most published alumina density values are quoted against ISO/ASTM bulk-density methods rather than a single international standard for the Al₂O₃ chemistry itself, so the practical anchor is the supplier datasheet and the test method on the certificate [S1][S5].
For PCB work, the consensus reading across suppliers is about 3.70-3.75 g/cm³ for many 94-96% alumina grades and about 3.85-3.92 g/cm³ for high-purity grades, with CoorsTek and KYOCERA material data showing the same grade-dependent pattern; differences between suppliers are normal because material codes are complete formulations, not just Al₂O₃ content [S1]. For comparison ceramics in the same family, the zirconia ceramic page shows how a different oxide system pushes bulk density into the 5.7-6.1 g/cm³ range, which is the reference point engineers use when alumina at 3.6-3.9 g/cm³ is too light for the inertial or shielding requirement.
Two trackable signals to watch on incoming material: insist on relative density reported as a percentage, not just bulk density, because that single number exposes porosity without you having to back it out from the chemistry [S5]; and require the Archimedes method to be named explicitly on the CoA, since a 3.85 g/cm³ number measured by geometry versus one measured by water displacement can differ by 1-2% on a porous 95% body.