ISO 14704
Flexural Strength of Monolithic Ceramics

Downloads: ISO 14704 Articulating Bend Fixture for Monolithic Ceramics – Universal Grip Data Sheet (PDF)
Written by Chris Thomas, Universal Grip Company · Reviewed by Frank Bertini · Updated October 27, 2026
ISO 14704 is the international test method for the flexural strength of monolithic fine ceramics at room temperature. A small rectangular bar is broken in three-point or four-point bending, and the force at fracture, the span and the bar's cross-section give the flexural strength in MPa. Because ceramics fail from flaws under tensile stress, this bend strength is the everyday stand-in for tensile strength in ceramic data sheets, design allowables and Weibull reliability work.
The method is used by producers of alumina, zirconia, silicon nitride and silicon carbide, by makers of electronic, wear and dental ceramics, and by labs that qualify ceramic parts for aerospace, semiconductor and medical use. Universal Grip makes an articulating bend fixture built for ISO 14704 (part ref. UGC-ISO14704-BF), with spring-retained free-rolling rollers, sliding supports on a scaled beam and a four-point upper loading member option. This guide walks through the whole standard: specimens, fixtures, procedure, calculations, common errors and how it compares with ASTM C1161.
In this guide: What ISO 14704 Measures · Current Edition and Scope · Test Configurations and Spans · Test Piece Dimensions and Preparation · Fixture, Bearings and Testing Machine · Test Procedure Step by Step · Calculating Flexural Strength · Weibull Statistics and Strength Scaling · Interpreting Results, Fracture Patterns and Common Errors · What the Test Report Must Include · ISO 14704 vs ASTM C1161 and Related Standards · Equipment for ISO 14704 from Universal Grip · FAQ
What ISO 14704 Measures
ISO 14704 defines flexural strength as the maximum nominal stress at fracture of a specified elastic beam loaded in bending. The bar is assumed to be isotropic and linearly elastic right up to fracture, which is a good description of dense technical ceramics. The result is a nominal stress at the tensile surface, calculated with simple beam theory from the fracture force, the fixture geometry and the measured width and thickness.
Ceramic strength is not a single material constant. It depends on the size and distribution of flaws such as pores, inclusions, machining damage and large grains, and on how much material is highly stressed. That is why the standard fixes the bar size, span and machining practice so closely, and why results are usually reported with Weibull statistics rather than a mean alone. Typical uses named in the scope are material development, quality control, characterization and generation of design data.
Note: Flexural strength is sometimes called modulus of rupture (MOR) or bend strength. For a given ceramic, the number you get depends on the loading configuration and bar size, so always quote the configuration with the result.
Current Edition and Scope
The current edition is ISO 14704:2016 (third edition, published April 2016, prepared by ISO/TC 206 Fine ceramics). It cancels and replaces ISO 14704:2008, which itself replaced the original 2000 edition and its 2004 corrigendum. ISO lists the 2016 edition as reviewed and confirmed in 2021, so it remains the version to cite.
-
Materials covered: monolithic fine (advanced technical) ceramics and whisker- or particulate-reinforced ceramic composites
-
Grain size limit: materials with a grain size below 200 µm
-
Temperature: room temperature only (high-temperature bend testing is covered by separate standards such as ISO 17565 and ASTM C1211)
-
Not covered: continuous-fiber ceramic matrix composites, which do not behave as linear-elastic, flaw-controlled beams
-
Thin plates of 0.2 mm to 1.0 mm thickness: tested to ISO 23242, which references ISO 14704
Test Configurations and Spans
ISO 14704 allows three loading configurations. The four-point-1/4 point configuration is the recommended one because it puts a uniform bending moment over the central half of the span, so a larger volume and surface of the bar are tested and the result is less sensitive to exactly where the critical flaw sits.
-
Four-point-1/4 point: support span L = 40 mm ± 0.1 mm, inner loading bearings one quarter of the span in from each support (inner span 20 mm)
-
Four-point-1/3 point: support span L = 30 mm ± 0.1 mm, inner bearings one third of the span in from each support (inner span 10 mm)
-
Three-point: support span L = 30 mm or 40 mm, ± 0.1 mm, with the load applied midway between the supports
-
Bearing positioning: the middle bearing (three-point) or the inner pair (four-point) centered on the outer pair within ± 0.1 mm
Three-point bending stresses only a thin line of material under the center roller, so it gives a higher apparent strength than four-point bending on the same material. Results from different configurations should not be mixed without the size-scaling corrections described in clause 10 and Annex E.
Test Piece Dimensions and Preparation
The standard test piece is a rectangular bar 4 mm wide and 3 mm thick, tested with the 4 mm face in tension and compression. Key dimensional requirements:
-
Width b: 4 mm nominal, cross-section tolerance ± 0.2 mm
-
Thickness d: 3 mm nominal, cross-section tolerance ± 0.2 mm
-
Length LT: at least 35 mm for 30 mm fixtures and at least 45 mm for 40 mm fixtures
-
Parallelism of opposite long faces: within 0.015 mm
-
Edges: at least the two long edges on one 4 mm face (the tensile face) chamfered or rounded; all four long edges is highly recommended
-
Chamfer: 45° × 0.12 mm ± 0.03 mm, or edge radius 0.15 mm ± 0.05 mm; larger chamfers need the correction factors in Annex D
Four preparation routes
ISO 14704 recognizes that strength is controlled by surface condition, so it lets you pick a preparation route and requires you to report it: as-fired (no machining; dimensions may deviate and must be reported), customary machining (your established process), component-matched (machined like the real part, so the test reflects service flaws) and the basic machining procedure. The basic procedure is the reference route when nothing else is specified:
-
Grind parallel to the long axis of the bar, in at least two stages, with filtered coolant
-
Coarse stage: diamond wheel no finer than 120 mesh (D126), depth of cut no more than 0.03 mm per pass
-
Finishing stage: diamond wheel of 320 to 800 mesh (for example D46 or finer), depth of cut no more than 0.002 mm per pass
-
Remove at least 0.06 mm per face in finishing, roughly equally from opposite faces
Handle finished bars so the tensile face is not scratched or chipped, check parallelism, squareness and chamfer size, and measure surface roughness when you can. For an established, proven machining process the standard allows inspection of three to five bars per batch.
Number of test pieces
The standard calls for at least 10 test pieces to estimate a mean flexural strength and at least 30 test pieces when a statistical analysis such as a Weibull analysis is to be made. Thirty is a minimum for a usable Weibull modulus; many qualification programs test more.
Fixture, Bearings and Testing Machine
The fixture is where most ceramic bend tests go wrong, and ISO 14704 is specific about it. Ceramic bars are stiff and brittle, so any friction at the rollers or any twist in the load path changes the stress at the tensile face and shifts the measured strength.
Articulation
-
Semi-articulating fixture: for bars with flat, parallel faces such as machined bars. In four-point, the inner pair and the outer pair are each parallel within 0.015 mm, and one pair articulates as a unit
-
Fully articulating fixture: required for as-fired, heat-treated or oxidized bars whose faces may be uneven, non-parallel or twisted, and usable for machined bars too. In four-point, one bearing may be fixed while the other three articulate independently
-
Three-point fixtures: either the outer pair or the middle bearing articulates; the middle bearing does not roll
-
All bearings must share the load equally
Bearings (rollers)
-
Free to roll: in four-point the inner bearings roll inward and the outer bearings roll outward to relieve friction
-
Diameter: about 1.5 × the bar thickness, with 4.5 mm to 5 mm recommended for the 3 mm bar; diameter uniform within ± 0.015 mm
-
Length: at least 12 mm, with a smooth surface
-
Steel bearings: hardness at least HRC 40 for strengths up to 1,400 MPa, or at least HRC 46 for strengths up to 2,000 MPa
-
Alternative bearings: ceramic or hardmetal with an elastic modulus of 200 GPa to 500 GPa and a flexural strength above 275 MPa
-
Fixture body: hard enough not to be dented by the bearings; steel of HRC 25 or more is adequate when bearings and fixture are at least 12 mm wide
Machine and measuring tools
The testing machine must meet ISO 7500-1 Class 1, with force accurate to 1 % of the indicated load at fracture, and must apply force at a constant loading or displacement rate while capturing the peak load. Width and thickness are measured with a micrometer to ISO 3611 type with 0.002 mm resolution and flat anvils; ball or pointed tips are not allowed. An electromechanical universal testing machine, often a benchtop frame, is the usual choice, because ceramic bars break at forces from a few hundred newtons to roughly 1.5 kN.
Test Procedure Step by Step
-
1. Measure each bar's width and thickness at the middle of the span to 0.002 mm, and record surface condition and chamfer size
-
2. Check the fixture: correct span (40 mm or 30 mm ± 0.1 mm), rollers clean, free to roll and centered, articulation free
-
3. Place the bar on the supports with the chamfered face down (tensile side) and the bar centered on the span and square to the rollers
-
4. Apply a slight preload of no more than 10 % of the expected mean strength to seat the bar and rollers
-
5. Load at a crosshead rate of 0.5 mm/min; the standard expects fracture within 3 s to 30 s, so adjust the rate if your bars break outside that window
-
6. Record the maximum force at fracture
-
7. Recover all fragments; note the fracture location (inside or outside the inner span) and, where possible, identify the fracture origin by fractography
-
8. Repeat for the full set (10 minimum for a mean, 30 minimum for Weibull)
Run tests in a normal laboratory atmosphere and report it. Moisture can promote slow crack growth in some oxide ceramics and glasses, so very slow loading tends to lower strength; the 3 s to 30 s window keeps that effect small and results comparable between labs.
Note: Ceramic bars can launch fragments at fracture. Fit a guard or shield around the fixture and wear eye protection.
Calculating Flexural Strength
Flexural strength σf (MPa) is calculated from the fracture force F (N), the support span L (mm), the width b (mm) and the thickness d (mm):
-
Three-point: σf = 3FL / (2bd²)
-
Four-point-1/4 point (inner span L/2): σf = 3FL / (4bd²)
-
Four-point-1/3 point (inner span L/3): σf = FL / (bd²)
Use each bar's own measured b and d. Because thickness is squared, a 0.02 mm error on a 3 mm bar shifts the result by about 1.3 %. If the chamfers exceed 0.12 mm ± 0.03 mm (or the radius exceeds 0.15 mm ± 0.05 mm), apply the chamfer correction factor from Annex D, since the reduced second moment of area raises the true stress.
Worked example: four-point-1/4 point
-
Material: 96 % alumina, basic machining
-
Support span L = 40 mm, inner span 20 mm
-
Measured width b = 4.00 mm, thickness d = 3.00 mm
-
Fracture force F = 450 N
-
σf = 3 × 450 × 40 / (4 × 4.00 × 3.00²) = 54,000 / 144 = 375 MPa
Worked example: three-point, 30 mm span
-
Same bar size, b = 4.00 mm and d = 3.00 mm, L = 30 mm
-
Fracture force F = 400 N
-
σf = 3 × 400 × 30 / (2 × 4.00 × 3.00²) = 36,000 / 72 = 500 MPa
As a sizing check, a zirconia with 1,200 MPa strength tested four-point-1/4 point on the standard bar fails at F = 1,200 × 4 × 4 × 9 / (3 × 40) = 1,440 N, which tells you the load cell and fixture capacity you need.
Weibull Statistics and Strength Scaling
Ceramic strengths scatter widely because each bar fails at its own worst flaw. The usual way to describe that is the two-parameter Weibull distribution: the Weibull modulus m (the spread; higher is more uniform) and the characteristic strength σθ (the strength at which 63.2 % of bars have failed). Weibull parameters are estimated by maximum likelihood following ISO 20501 (fine ceramics Weibull statistics), which is why ISO 14704 asks for at least 30 bars.
The Weibull model also predicts a size effect: bigger stressed volumes or surfaces contain bigger flaws and are weaker. ISO 14704 clause 10 and Annex E give effective volume and effective surface factors so that strengths measured in three-point, four-point-1/3 point and four-point-1/4 point can be scaled to each other or to a component. Use volume scaling when failures start inside the material (pores, inclusions) and surface scaling when they start at machining damage or chamfers, which is why fractography matters.
Interpreting Results, Fracture Patterns and Common Errors
Annex C of the standard shows typical fracture patterns. A strong bar shatters into several pieces because more stored energy is released; a weak bar often breaks into two. In four-point, a valid break starts in the inner span; a break under or outside a loading roller points to contact damage, wedging or misalignment and should be flagged in the report.
Common errors
-
Fixed (non-rolling) rollers: friction adds a restraining moment and can inflate measured strength by several percent
-
Non-articulating fixture on as-fired or twisted bars: one corner carries the load, giving early, low failures with origins at an edge
-
Wrong roller size or soft rollers: contact damage starts fracture under the roller instead of at the tensile face
-
Unchamfered or chipped edges: edge flaws dominate and strength drops; oversize chamfers raise true stress unless corrected
-
Using nominal instead of measured dimensions: thickness enters squared, so small errors become noticeable
-
Mixing configurations: three-point and four-point results differ by design and need size scaling to compare
-
Too few bars: ten bars give a mean, but a Weibull modulus from fewer than 30 is unreliable
-
Grinding across the bar instead of along it: transverse grinding scratches act as flaws perpendicular to the stress and lower strength
What the Test Report Must Include
-
Reference to ISO 14704:2016 and the configuration (three-point, four-point-1/4 or four-point-1/3) and span
-
Material identification, test piece preparation route, surface condition, chamfer type and size
-
Number of bars tested and dimensions of each
-
Crosshead rate, test atmosphere and temperature
-
Fracture force and flexural strength of each bar, mean and standard deviation
-
Weibull modulus and characteristic strength with confidence bounds when 30 or more bars are tested
-
Fracture location and, where available, fracture origins
-
Any deviation from the standard, including as-fired dimensions outside the nominal size
ISO 14704 vs ASTM C1161 and Related Standards
The closest ASTM equivalent is ASTM C1161, flexural strength of advanced ceramics at ambient temperature (current edition C1161-18, reapproved 2023). The two methods share the same physics and the same standard bar, but there are practical differences:
-
Standard bar: both use 4 mm × 3 mm bars tested on a 40 mm support span; C1161 also offers half-size and double-size bars and spans
-
Four-point geometry: C1161 uses four-point-1/4 point only; ISO 14704 also permits four-point-1/3 point on a 30 mm span
-
Three-point spans: ISO 14704 allows 30 mm or 40 mm
-
Edge finish: both specify chamfered or rounded edges of similar size
-
Machining: both offer as-fired, customary, application-matched and a defined standard grinding procedure
-
Crosshead rate: ISO 14704 sets 0.5 mm/min with a 3 s to 30 s time to fracture; check the C1161 table for the rate tied to each configuration
Related standards worth knowing: ASTM C1211 for flexural strength at elevated temperature, ASTM C1684 for cylindrical rods, ASTM C1273 for direct tensile strength of ceramics, ISO 20501 for Weibull analysis, ISO 23242 for thin ceramic plates, and EN 843-1 for the European ceramic bend test. For polymer and composite bend testing, see ISO 178 and ISO 14125; those use deflection-based methods and much larger specimens and are not interchangeable with ceramic bend testing.
Equipment for ISO 14704 from Universal Grip
The Universal Grip ISO 14704 Bend Fixture (part ref. UGC-ISO14704-BF) is an articulating three-point bend fixture for flexural strength testing of monolithic ceramics at room temperature, with a four-point upper loading member available for ISO 14704 four-point configurations. Per the data sheet:
-
Loading members: spring-retained, free-rolling articulating rollers, so the brittle bar is loaded without frictional or twisting constraint
-
Articulating supports compensate for specimen twist and non-parallel faces
-
Span adjustment: supports slide on a scaled beam for accurate span setting
-
Load capacity: 5 kN standard; higher capacities on request
-
Material: stainless steel fixture with hardened rollers
-
Specimens: rectangular ceramic bars per ISO 14704
-
Machine connection: clevis pin or threaded adapter to suit your load frame
-
Applications: alumina, zirconia and silicon nitride; technical and electronic ceramics; dental and bioceramics; Weibull strength qualification
When ordering, specify three- or four-point configuration, spans, specimen cross-section, roller diameter, capacity and machine connection. Pair the fixture with an electromechanical universal testing machine or benchtop tester, a load cell sized to your expected fracture forces, and test software that captures peak load and exports the strength set for Weibull analysis. See our full range of bend testing fixtures and our bend testing overview, and contact us for a quote.
ISO 14704 FAQ
What is the current edition of ISO 14704?
The current edition is ISO 14704:2016, the third edition, published in April 2016 by ISO/TC 206. It replaced ISO 14704:2008 and was reviewed and confirmed by ISO in 2021.
What specimen size does ISO 14704 use?
The standard bar is 4 mm wide and 3 mm thick, with a cross-section tolerance of ± 0.2 mm. It must be at least 45 mm long for 40 mm span fixtures or at least 35 mm long for 30 mm span fixtures, with opposite faces parallel within 0.015 mm.
Is three-point or four-point bending better for ceramics?
ISO 14704 recommends four-point-1/4 point bending on a 40 mm span. It stresses a larger volume and surface uniformly, so it samples more flaws and gives a more representative strength. Three-point bending gives higher apparent strengths and is best used for screening or comparison within one study.
What crosshead speed does ISO 14704 specify?
The standard sets a crosshead rate of 0.5 mm/min, provided the bar fractures within 3 s to 30 s. If fracture falls outside that window, adjust the rate so it falls inside.
How many specimens are needed for ISO 14704?
At least 10 bars are needed to estimate the mean flexural strength. At least 30 bars are required if a statistical analysis such as a Weibull analysis is to be made.
Why must the rollers be free to roll?
As a ceramic bar bends, its contact points move slightly along the rollers. Fixed rollers resist that motion with friction, adding a moment that changes the stress at the tensile face and biases the strength. ISO 14704 requires rolling bearings, with inner rollers rolling inward and outer rollers rolling outward in four-point bending.
What is the difference between semi-articulating and fully articulating fixtures?
A semi-articulating fixture suits machined bars with flat, parallel faces; in four-point one pair of bearings articulates as a unit. A fully articulating fixture lets three of the four bearings move independently, and is required for as-fired, heat-treated or oxidized bars that may be twisted or non-parallel.
Do ISO 14704 specimens need chamfered edges?
Yes. At least the two long edges on the tensile 4 mm face must be chamfered at 45° × 0.12 mm ± 0.03 mm or rounded to a 0.15 mm ± 0.05 mm radius, and chamfering all four edges is highly recommended. Larger chamfers need the correction factors in Annex D.
How is flexural strength calculated in four-point-1/4 point bending?
Use σf = 3FL / (4bd²), with F the fracture force, L the 40 mm support span, b the width and d the thickness. For example, a 4.00 × 3.00 mm bar breaking at 450 N gives 3 × 450 × 40 / (4 × 4 × 9) = 375 MPa.
Is ISO 14704 equivalent to ASTM C1161?
They are closely aligned: both test 4 × 3 mm bars on 40 mm spans with similar chamfer and machining rules, so results are broadly comparable for the same configuration. ISO 14704 also permits a four-point-1/3 point configuration on a 30 mm span. See our ASTM C1161 guide for the ASTM side.
Can ISO 14704 be used for ceramic matrix composites?
Only for whisker- or particulate-reinforced ceramics that behave as linear-elastic, flaw-controlled materials with grain size below 200 µm. Continuous-fiber ceramic composites are outside its scope because they do not fail in a brittle, linear-elastic way.
What fixture does Universal Grip offer for ISO 14704?
The UGC-ISO14704-BF articulating bend fixture has spring-retained free-rolling rollers, articulating supports, sliding supports on a scaled beam, a 5 kN standard capacity and a four-point upper loading member option. Contact us with your configuration, spans, bar size and frame connection for a quote.
Need a fixture, grips or a complete test system for ISO 14704? Contact Universal Grip with your material, specimen size and machine connection and we will quote the right setup.
