Silicone Hardness Testing: Shore A, IRHD and the Sample Thickness

What Shore A actually measures

Shore A measures the resistance of a material to a spring loaded indentor. A truncated cone is pressed into the surface under a defined force, and the depth of penetration is read on a scale from 0 to 100. A soft gel reads near 0, a hard rubber near 100, and a typical silicone gasket sits between 40 and 70.

It is an empirical test, not a material property. The number depends on the geometry of the indentor, the spring rate, the dwell time and the thickness of the sample, and two instruments from different makers can disagree by several points on the same sample. That is why the standard specifies the instrument, the sample and the procedure together, and why a reading without those details is not a measurement.

The scale is also non-linear in terms of modulus. A change from 40 to 50 Shore A is a much larger change in stiffness than a change from 70 to 80, because the indentor geometry and the material response interact differently across the range. Comparing hardness values as if they were linear is a common source of confusion.

For silicone specifically, hardness is set by the filler loading rather than by the polymer. A 40 Shore A compound might have 20 parts of fumed silica per hundred of polymer; a 70 Shore A compound might have 45. That is why hardness and tear strength move together, and why a hardness specification is partly a specification of the filler system.

The thickness problem

The indentor penetrates a finite depth, and the material under it has to be thick enough to behave as if it were infinite. If the sample is thin, the indentor feels the support underneath and reads harder than the material actually is. The standard requires at least 6 mm of material, either as a single slab or as a stack of thinner ones.

The effect is large. The chart on the left shows the same three compounds measured at thicknesses from 1 to 12 mm. At 1 mm, a nominal 80 Shore A compound reads 91, a nominal 60 reads 78 and a nominal 40 reads 62. That is 11 to 22 points of error, which is larger than the tolerance band on most drawings.

The reading stabilises from about 6 mm. Between 6 and 12 mm the three curves are flat, which is the region the standard is trying to reach. A stack of three 2 mm slabs is acceptable if they are flat and in full contact; a stack of loose pieces with air gaps is not.

The practical consequence is that a part thinner than 6 mm cannot be hardness tested directly. A 2 mm gasket has to be tested as a stack, or the hardness has to be inferred from a plaque moulded from the same compound. Testing the part itself and quoting the number is the most common error in silicone incoming inspection.

Shore A against IRHD

IRHD is the international rubber hardness degree, and it measures the same thing with a different instrument. The IRHD method uses a ball indentor and a different force, and it is designed to be less sensitive to surface condition and more reproducible across laboratories. For most rubbers the two scales agree closely in the middle of the range and diverge at the ends.

Around 50 to 70 Shore A, IRHD and Shore A differ by about one to three points. Below 40 the divergence grows, and above 85 it grows again. A specification that says 60 plus or minus 5 without naming the method is ambiguous, and a supplier and a customer using different methods can disagree on a part that is in fact identical.

IRHD is the better choice for thin sections and for small parts, because the ball indentor is less affected by the support underneath. It is also the more common method in European rubber specifications. Shore A is faster, cheaper and more common in Asia and North America, and for a silicone gasket in the 40 to 70 range the difference rarely matters.

The rule that resolves it is simple: name the method, name the standard, and name the sample thickness in the specification. A drawing that says 60 plus or minus 5 Shore A, ISO 868, on a 6 mm plaque is unambiguous and can be checked. A drawing that says 60 plus or minus 5 hardness cannot.

Sample preparation

The sample has to be flat, parallel and at least 6 mm thick, and it has to be at the test temperature for long enough to be in equilibrium. A sample taken straight from a warm press reads softer than the same sample the next day, because silicone stiffness changes with temperature and because the cure is still completing.

The surface matters too. A moulded surface is smoother than a cut surface, and a cut surface can read one or two points harder because the indentor sits on a rougher plane. For a specification, the sample should be a moulded plaque from the production tool, not a cut piece from a finished part.

The stack, if one is used, must be in full contact with no air gaps and no relative movement. Three pieces of 2 mm sheet stacked loosely will read differently from three pieces bonded together, and neither will match a solid 6 mm slab exactly. The standard accepts a stack but requires the surfaces to be flat and in contact.

Table 1 — What a valid Shore A sample looks like

Reading the instrument correctly

The instrument has to be held vertically and pressed down until the presser foot is fully in contact with the surface. A reading taken before the foot is seated is high, because the indentor has not yet reached its full penetration. This is the second most common error after sample thickness, and it is worth four to five points.

The reading has to be taken within one second of full contact for most silicone compounds, because the material creeps under the load. A reading taken after five seconds is lower, sometimes by three points, and the standard specifies one second for exactly this reason. Instruments with a maximum hold function help, but the operator still has to read at the right moment.

Temperature matters more than most operators expect. A sample at 5 C reads two points harder than the same sample at 23 C; a sample at 35 C reads two points softer. In an unheated inspection room in winter, this is a real source of disagreement between a supplier in a warm climate and a customer in a cold one.

The instrument itself needs checking. A reference block of known hardness, checked at the start of each session, catches a worn indentor or a tired spring. An instrument that has been in service for years without calibration can be five points out, and nobody notices because the readings are consistent with each other.

Hardness and the properties behind it

Hardness is a proxy for several properties at once, and it is worth knowing which. Higher hardness means higher modulus, higher tensile strength, higher tear strength and lower elongation at break. It also means lower compression set resistance in some compounds, because a harder compound has less ability to relax.

That is why a hardness change is never just a hardness change. A customer who asks for 70 Shore A instead of 60 is also asking for a stiffer part, a stronger part and a part that stretches less. If the application needed the flexibility of 60, the change may solve one problem and create another.

The relationship to compression set is the one that catches people out. A gasket at 70 Shore A seals better against high pressure but conforms less well to a rough surface, and a gasket that does not conform leaks regardless of its hardness. The right hardness is the one that matches the sealing surface, not the highest one available.

For design purposes, a hardness specification should be accompanied by a tolerance on the finished part, not just on the compound. A moulded part can vary by two to three points across a batch because of cure variation, and a specification of plus or minus 3 on the compound plus a process variation of 3 gives a part that can be 6 points away from nominal.

A test method you can defend

The method that survives an audit has four parts: a named standard, a defined sample, a defined number of readings and a defined acceptance rule. Everything else is detail. Without those four, two people can measure the same part and disagree, and neither is wrong.

The acceptance rule should be based on the median of five readings rather than the mean, because a single bad reading from a surface defect can pull a mean out of tolerance. The median is robust to one outlier and is easy to explain to an operator.

Table 2 — A Shore A procedure that can be audited

The record is the part that is usually missing. A hardness number without the sample thickness, the temperature and the standard is a number that cannot be defended, and in a dispute about a batch it is worth nothing. A single line on the inspection sheet fixes that permanently.