Ceramic testing and quality control

In short

Ceramic testing and quality control is the set of standard measurements - density and porosity, strength, hardness, thermal behaviour, surface and chemical durability - by which a ceramic or glass product is accepted or rejected, plus the statistical system (sampling plans, control charts, capability indices) that keeps a plant inside those limits. It matters because ceramics are brittle: strength is a statistical quantity set by the largest flaw, so a single untested batch can fail in service. In India the test methods are codified by BIS - the IS 13630 series for ceramic tiles, the IS 1528 series for refractories, IS 731 for porcelain insulators, IS 2553 for safety glass - and laboratories are accredited to ISO/IEC 17025 by NABL. It is examined in Ceramic Technology papers and GATE-level questions on properties and testing, and it is the single most common entry role for fresh graduates at Kajaria, Somany, RAK, Hindware, Asian Granito, Orient Bell, TRL Krosaki, Dalmia-OCL, Vesuvius, IFGL, Saint-Gobain, Asahi India Glass, Borosil and Aditya Birla Insulators.

Explained

What is tested, and why ceramics need their own rules

• A ceramic fails from the largest flaw present (pore, crack, agglomerate, inclusion), not from an average property - so strength scatters widely and must be reported with a standard deviation and a Weibull modulus, never as a single number.
• No plastic yielding, so there is no yield point, no elongation, and tensile tests are impractical (grip failure, alignment). Strength is measured in bending (modulus of rupture) or in compression (cold crushing strength).
• Strength is size- and stress-volume dependent: a bigger specimen, or one loaded over a larger volume, breaks at a lower stress. Hence span, cross-section and loading geometry must be exactly as the standard says.
• Testing is split into: raw-material QC (incoming), in-process QC (body, green ware, dried, glazed), fired-product testing (acceptance), and plant/calibration QC (furnace profiles, instruments).
• Destructive vs non-destructive: MOR, CCS, PCE, RUL are destructive and use samples; dye penetration, ultrasonic velocity, impulse excitation (IET/resonance for E), radiography, acoustic emission and visual/dimensional checks can be applied to whole pieces.

Raw material and in-process control

• Moisture: oven-dry at 110 ± 5 °C to constant mass; spray-dried granule moisture is typically held in a narrow band (commonly 5-7% for tile presses) because it controls green density and pressed dimensions.
• Particle size: wet sieve residue on 45 µm / 63 µm for body slip and glaze; laser diffraction for full distribution. Residue on sieve is the single most-used daily QC number in a tile plant.
• Slip properties: density by pyknometer or Baumé, viscosity and flow time by Ford/Marsh cup or Brookfield, thixotropy (flow time at 30 s and 30 min), deflocculant demand titration.
• Plasticity and workability: Atterberg liquid and plastic limits, Pfefferkorn plasticity, water of plasticity; dry strength of test bars for handling losses.
• Chemical and thermal: oxide analysis by XRF or wet chemistry, loss on ignition, carbonate/sulphate content, soluble salts (scumming), DTA/TG for decomposition and dilatometry for thermal expansion and the glaze-body fit.
• Green and dried ware: green bulk density (mercury displacement or dimensional), green/dry MOR, drying shrinkage, drying sensitivity (Bigot curve), lamination and crack checks.
• Firing control: thermocouple/pyrometer profiles, draw trials and pyrometric rings/cones inside the kiln as independent heat-work witnesses, firing shrinkage and loss on firing, warpage.

Physical property tests on fired ceramics

• Apparent porosity, bulk density, water absorption, apparent specific gravity - all four from one Archimedes (boiling or vacuum) test: dry mass, soaked mass in air, soaked mass suspended in water. Boiling 2 h or vacuum saturation; vacuum gives higher, more reliable saturation.
• True density by helium or liquid pyknometer on powdered sample; true porosity = 1 − bulk/true density; closed porosity = true − apparent porosity.
• Water absorption classifies tiles (IS 15622 / ISO 13006): BIa ≤ 0.5%, BIb 0.5-3%, BIIa 3-6%, BIIb 6-10%, BIII > 10% - vitrified tiles are BIa, wall tiles usually BIII.
• Dimensions and surface quality: length, width, thickness, straightness of sides, rectangularity, centre and edge curvature, warpage, and surface defects on at least 1 m² of tiles (IS 13630 part on dimensions, mirroring ISO 10545-2).
• Modulus of rupture and breaking strength: three-point bend; tiles report breaking strength in newtons as well as MOR in N/mm².
• Hardness: Mohs scratch for glazes (quick, shop-floor), Vickers (HV) and Knoop microindentation for engineering ceramics; Rockwell is unsuitable for brittle, hard ceramics.
• Abrasion: deep abrasion (volume removed) for unglazed tiles; surface abrasion by PEI class 0-5 for glazed tiles - class 4/5 for heavy-traffic floors, class 1-2 for walls and light domestic use.
• Slip resistance: coefficient of friction (wet and dry) and ramp-test R-ratings - specified for bathrooms, kitchens and public floors; the exact test referenced varies by project specification.
• Fracture toughness K_IC: single-edge-notched beam (SENB), chevron-notch or SEVNB; the Vickers-indentation estimate is convenient but approximate and must be labelled as such.
• Elastic constants: ultrasonic pulse-echo or impulse excitation of vibration - non-destructive, and far more reliable for ceramics than reading a slope off a load-deflection curve.

Thermal, chemical and surface durability

• Pyrometric cone equivalent (PCE): a test cone of the refractory is fired beside standard Seger/Orton cones; the standard cone that bends to the same degree gives the PCE. It measures refractoriness without load (softening range), not service temperature.
• Refractoriness under load (RUL): specimen heated at a set rate under constant compressive load (commonly 0.2 N/mm²); temperatures at defined subsidence (e.g. 0.5%, 1%, 2%) are reported - a far better service indicator than PCE. Creep in compression gives time-dependent deformation at a held temperature.
• Permanent linear change (PLC) / reheat change: dimensional change after soaking at service temperature - predicts joint opening or spalling in a lining.
• Thermal shock / spalling resistance: repeated quench cycles (water or air) with strength or mass loss measured after each; the ISO 10545-9 style tile test cycles between hot water and cold water and looks for visible defects.
• Thermal expansion by dilatometer and thermal conductivity by hot-wire, calorimeter or laser flash - critical for refractory lining design.
• Chemical resistance: acid/alkali and household-chemical classes for tiles (class A to C by effect on surface); slag attack by crucible, static dip or rotary-slag tests for refractories; CO disintegration and alkali (alumina-silica) resistance tests.
• Stain resistance (classes 1-5) and crazing resistance of glazed ware by autoclave test - typically steam pressure held and the glaze inspected for hairline cracks; crazing is a glaze-in-tension fault from expansion mismatch or moisture expansion of the body.
• Moisture expansion of the body, and frost resistance by freeze-thaw cycling - required for external and cold-region flooring.
• Lead and cadmium release from glazed surfaces by acetic acid extraction - essential for tableware and glazed tiles; for food-contact ceramics the migration limits sit in the FSSAI packaging regulations and the limits/test differ by article, so quote the current regulation rather than a remembered number.

Testing of glass and electrical porcelain

• Glass: residual stress and annealing quality by polariscope / strain disc (optical path difference); thermal shock by hot-cold immersion; hydrolytic (chemical) durability by grain or surface test giving a hydrolytic class; container internal pressure, vertical load and impact tests; thickness, bow and optical distortion for float glass; visual defects - bubbles, seeds, stones, cords, knots, striae.
• Toughened (tempered) glass: fragmentation test (particle count in a 50 mm × 50 mm square), surface compression and bending strength; laminated glass: boil test, bake test and radiation test for delamination, plus impact tests - the safety-glass requirements are in the IS 2553 series, and transparent float glass in IS 14900.
• Electrical porcelain insulators (IS 731 for overhead-line insulators above 1000 V): routine, sample and type tests - visual, mechanical failing load, electromechanical test, dry/wet power-frequency and impulse withstand, puncture test, porosity by dye penetration under pressure (a classic viva question: fuchsine dye at high pressure must not penetrate the fired body), thermal shock cycles, and galvanizing checks on metal fittings.
• Sanitaryware: the vitreous sanitary appliance standards (IS 2556 series) cover dimensions, water absorption, crazing, staining, load and flushing performance.

Statistical quality control and the Indian regulatory frame

• Sampling inspection by attributes (IS 2500 series, aligned with ISO 2859-1): lot size → sample size code letter → AQL → accept/reject numbers; single, double and multiple plans; AQL, LTPD, producer's and consumer's risk, OC curve.
• Control charts: variables charts X̄-R or X̄-s (dimensions, water absorption, slip density) and attribute charts p, np, c, u (defective tiles per lot, pinholes per m²). Control limits come from the process, specification limits from the customer - never mix them on one chart.
• Process capability: Cp compares spread to tolerance, Cpk also penalises off-centring; Cp ≥ 1.33 is the usual industry acceptance, ≥ 1.67 for critical features.
• Weibull analysis for strength: plot ln ln[1/(1−P_f)] against ln σ; slope = Weibull modulus m (scatter: m ≈ 5-10 for traditional ceramics, 10-20+ for good engineering ceramics), intercept gives characteristic strength σ₀.
• Defect analysis tools: Pareto of defect types, fishbone for cause, 7QC tools, SPC on the critical few; first-piece and in-line inspection on tile lines, plus automated optical sorting into first/second/commercial quality.
• Regulatory and system framework: the BIS Act, 2016 and the BIS (Conformity Assessment) Regulations, 2018 govern the ISI mark; several ceramic and glass products are covered by Quality Control Orders issued under the Act that make conformity to the relevant IS mandatory - the notified list changes, so check the current QCO before claiming a product is covered. Test laboratories are accredited to ISO/IEC 17025 by NABL; plant quality systems to ISO 9001:2026 (current edition). Dust-exposure monitoring for respirable crystalline silica in body preparation and glaze areas falls under the OSH&WC Code, 2020, in force with the four Labour Codes from 21 November 2025.

Key points

• Brittle failure - strength is set by the worst flaw, so it is reported as a mean with scatter, not a single value.
• Archimedes test - one weighing set gives apparent porosity, bulk density, water absorption and apparent specific gravity.
• Apparent porosity - open pores only; true porosity needs true density; the difference is closed porosity.
• Water absorption - the classifying property for tiles (BIa ≤ 0.5% to BIII > 10%) and a proxy for vitrification.
• MOR - measured in 3-point or 4-point bending; 4-point stresses a larger volume and gives a lower, more conservative strength.
• Breaking strength (N) - the tile convention F·L/b, distinct from MOR in N/mm².
• CCS - cold crushing strength, the standard handling/strength index for refractory bricks.
• PCE - softening temperature by cone comparison; refractoriness *without* load.
• RUL - refractoriness under load, with subsidence temperatures; closer to real service than PCE.
• PLC / reheat change - permanent dimensional change after soaking; predicts lining joint behaviour.
• Thermal shock resistance rises with strength and falls with E and α - fused silica excels, alumina does not.
• Vickers hardness - microindentation for hard ceramics; Mohs for glaze scratch, Rockwell unsuitable.
• PEI class 0-5 - glazed surface abrasion rating that decides where a tile may be laid.
• Crazing - glaze cracking from glaze being in tension; checked by autoclave test, cured by expansion matching.
• Weibull modulus m - higher m means tighter strength scatter and more usable design stress.
• Cp vs Cpk - Cp ignores centring, Cpk does not; Cpk ≤ Cp always.
• Control limits ≠ specification limits - the first describe the process, the second the customer's need.
• IS 13630 (tile test methods), IS 15622 (pressed tiles), IS 1528 (refractory tests), IS 731 (porcelain insulators), IS 2553 (safety glass) - the numbers to quote in an Indian interview.

Formulas and examples

Example 1 - full Archimedes set on a wall tile body. W_D = 150.0 g, W_W = 160.0 g, W_S = 95.0 g, true density 2.75 g/cm³.

• W_W − W_S = 65.0 g (volume of the whole specimen = 65.0 cm³)
• P_a = 10.0/65.0 × 100 = 15.4%
• ρ_b = 150.0/65.0 = 2.308 g/cm³
• WA = 10.0/150.0 × 100 = 6.67% → group BIIb (6% < E ≤ 10%), a typical glazed wall tile, not a vitrified floor tile
• Apparent sp. gr. = 150.0/(150.0 − 95.0) = 150/55 = 2.727
• P_t = (1 − 2.308/2.75) × 100 = 16.1%; closed porosity = 16.1 − 15.4 = 0.7%

Example 2 - MOR and breaking strength of the same tile. Span L = 100 mm, b = 25 mm, h = 10 mm, breaking load P = 1,250 N.

• σ (3-point) = 3 × 1250 × 100/(2 × 25 × 10²) = 375,000/5,000 = 75 N/mm²
• Breaking strength S = 1250 × 100/25 = 5,000 N
• Had the same load been applied in third-point 4-point bending: σ = 1250 × 100/(25 × 100) = 50 N/mm² - lower, because for a given load the maximum bending moment is PL/6 instead of PL/4. (Separately, in a real test the measured 4-point MOR comes out below the 3-point value because a larger volume is held at peak stress.) Always state the geometry with the number.

Example 3 - thermal shock, alumina vs fused silica.

• Alumina: R = 350×10⁶ × 0.78/(370×10⁹ × 8.0×10⁻⁶) = 273×10⁶/2.96×10⁶ = ≈ 92 K
• Fused silica: R = 70×10⁶ × 0.83/(73×10⁹ × 0.55×10⁻⁶) = 58.1×10⁶/40,150 = ≈ 1,450 K
• Conclusion: the weaker material survives the quench by a factor of about 16, because α is 15 times smaller - strength alone never predicts thermal shock.

Example 4 - Weibull analysis. A porcelain body gives m = 10 and σ₀ = 300 MPa on 10 cm³ bars.

• At σ = σ₀, P_f = 1 − e⁻¹ = 63.2% (the meaning of characteristic strength).
• Design stress for P_f = 10⁻³: σ = σ₀[−ln(1 − 0.001)]^1/10 = 300 × (1.0005×10⁻³)^0.1 = 300 × 0.501 = ≈ 150 MPa - half the characteristic strength for a 1-in-1000 failure rate.
• Scaling to a component of 80 cm³ (8× the volume): σ₂ = 150 × (1/8)^0.1 = 150 × 0.812 = ≈ 122 MPa. Low m is punished hard here; if m were 5, the factor would be (1/8)^0.2 = 0.660 → 99 MPa.

Example 5 - capability and control chart on tile thickness. Spec 9.0 ± 0.5 mm (LSL 8.5, USL 9.5). From 25 subgroups of n = 5: X̄ = 9.10 mm, R̄ = 0.28 mm.

• σ̂ = R̄/d₂ = 0.28/2.326 = 0.1204 mm
• X̄ chart: UCL = 9.10 + 0.577 × 0.28 = 9.262 mm, LCL = 9.10 − 0.162 = 8.938 mm
• Cp = (9.5 − 8.5)/(6 × 0.1204) = 1.0/0.7224 = 1.38
• Cpu = (9.5 − 9.10)/(3 × 0.1204) = 0.40/0.3612 = 1.11; Cpl = (9.10 − 8.5)/0.3612 = 1.66 → Cpk = 1.11
• Reading: spread is adequate (Cp 1.38) but the press is set 0.10 mm high; re-centring to 9.00 mm would lift Cpk to ≈ 1.38 with no change in variation. This is the standard "Cp is fine, Cpk is not - what do you do?" interview question.

Where it is used

• Tile and sanitaryware plants (Kajaria, Somany, Orient Bell, Asian Granito, RAK, Hindware, Cera, Morbi cluster): daily sieve residue, slip density, press moisture, dimension and water absorption checks, PEI and crazing lots, ISI-marked despatch.
• Refractory works (TRL Krosaki, Dalmia-OCL, IFGL, Vesuvius, Calderys, RHI Magnesita India): PCE, RUL, CCS, PLC, slag-attack and thermal-conductivity testing for steel, cement and glass customers, plus post-mortem lining analysis.
• Glass plants (Asahi India Glass, Saint-Gobain, HNG, Borosil, Piramal Glass): annealing-stress polariscopy, fragmentation and boil tests, container pressure and hydrolytic class, defect counts per tonne.
• Electrical porcelain and technical ceramics (Aditya Birla Insulators, BHEL vendors, CGCRI-linked units): routine mechanical and dielectric tests, dye-penetration porosity, HV and K_IC on engineering grades.
• Third-party and government labs: NABL-accredited private labs, CSIR-CGCRI Kolkata, CGCRI Khurja/Naroda centres, BIS laboratories, CBRI and NTH - testing for QCO conformity, tenders and dispute resolution.
• Job roles: QC chemist/engineer, lab in-charge, process QA, production QC supervisor, standards and certification executive (BIS liaison), customer-complaint/failure-analysis engineer.

Interview questions

• Q: Why is a tensile test rarely used on ceramics? A: Ceramics are brittle and notch-sensitive; gripping and misalignment cause premature failure, so bending (MOR) or compression tests with standardised geometry are used instead.
• Q: Difference between apparent and true porosity? A: Apparent porosity counts only pores open to the surface (measured by liquid saturation); true porosity is calculated from bulk and true density and includes closed pores - the difference is closed porosity.
• Q: Why does the same material give a lower MOR in 4-point bending than in 3-point? A: Four-point loading subjects a larger volume to the maximum stress, so the probability of including a critical flaw rises - a direct consequence of Weibull statistics.
• Q: PCE or RUL - which would you quote to a steel plant? A: RUL, because it measures deformation under a service-like compressive load; PCE is a no-load softening indicator and overstates usable temperature.
• Q: What does water absorption tell you about a tile? A: It measures open porosity and therefore degree of vitrification; it classifies the tile (BIa ≤ 0.5% vitrified up to BIII > 10%) and controls frost resistance, staining and strength.
• Q: What causes crazing and how do you test for it? A: Glaze in tension - usually glaze expansion higher than the body's, or moisture expansion of an underfired body; tested by an autoclave (steam pressure) test and cured by lowering glaze expansion or firing the body tighter.
• Q: Why is fused silica better in thermal shock than alumina, although it is far weaker? A: R = σ(1−ν)/(Eα); fused silica's α is roughly 15 times lower and its E one-fifth, which outweighs its lower strength - about 1,450 K versus 92 K.
• Q: A process has Cp = 1.38 but Cpk = 1.11. What is wrong? A: Variation is acceptable but the mean is off-centre; re-centre the setting rather than tighten the process.
• Q: Can control limits be set equal to specification limits? A: No - control limits come from the process's own variation (X̄ ± A₂R̄) and detect instability; specification limits come from the customer and judge conformance.
• Q: What is the Weibull modulus and what is a good value? A: The slope of the Weibull plot, measuring strength scatter; roughly 5-10 for traditional ceramics and 10-20 or more for well-processed engineering ceramics - higher means a higher safe design stress.
• Q: How is porosity checked on a fired porcelain insulator? A: By dye penetration under pressure on sectioned samples per IS 731 - a destructive sample test - fuchsine dye forced into the body must not penetrate; ultrasonic and visual checks support it on whole pieces.
• Q: What governs the ISI mark in India? A: The BIS Act, 2016 with the BIS (Conformity Assessment) Regulations, 2018; for products covered by a Quality Control Order the relevant Indian Standard becomes mandatory - the notified product list is revised periodically and must be checked.

Common mistakes

• Using the soaked-in-air mass where the suspended mass belongs (or vice versa) in the Archimedes formulas, which inverts porosity and density.
• Reporting water absorption as the same thing as apparent porosity - they differ by the bulk density factor.
• Quoting MOR without the span, specimen size and loading geometry, making the number meaningless and non-comparable.
• Treating PCE as the maximum service temperature of a refractory.
• Giving a single strength figure with no standard deviation, sample size or Weibull modulus for a brittle material.
• Confusing control limits with specification limits, or plotting individual readings on an X̄ chart drawn for subgroups of five.
• Judging thermal shock resistance from strength alone and ignoring E and α.
• Citing withdrawn IS numbers from an old lab manual - verify the current number and edition before quoting it in a report or interview.

Related topics

Materials characterisation: XRD, SEM and TEM · Ceramic and glass industry careers in India · Ceramic processing: forming, drying and sintering · Glazes, enamels and glass-ceramic systems · Refractories and high-temperature materials · Mechanical behaviour of brittle solids and fracture mechanics