Glass transition temperature (Tg) is measured by heating a material at a controlled rate while tracking a physical property that shifts abruptly as the material transitions from a rigid, glassy state to a softer, rubbery one. The most common technique is Differential Scanning Calorimetry (DSC), but Dynamic Mechanical Analysis (DMA) and Thermomechanical Analysis (TMA) each detect the transition through different physical signals, and the best choice depends on your material and how precise you need to be.
Differential Scanning Calorimetry (DSC)
DSC is the standard method for measuring Tg and the one referenced by ASTM E1356, the formal test standard for glass transition assignment. The principle is straightforward: you place a small sample in a metal pan on one side of a sensor and an empty reference pan on the other, then heat both at the same rate. As the sample passes through its glass transition, it absorbs slightly more heat than before because the polymer chains gain mobility. The instrument records this as a step change in heat flow, which corresponds to a change in heat capacity.
For polymer samples, prepare a thin film weighing between 5 and 15 mg and place it in an aluminum pan. Five milligrams is a common starting point for materials like polystyrene. Use nonhermetic (unsealed) pans unless your material releases volatiles, and keep the film flat so it contacts the pan bottom evenly. Never heat aluminum pans above 500 °C, though most polymer Tg values fall well below that range. Standard heating rates are typically 10 or 20 °C per minute, though specialized instruments can ramp as fast as 2,000 °C per minute when you need to outrun unwanted crystallization or form changes in unstable samples.
On the resulting curve, the glass transition appears as a gradual step rather than a sharp peak. You can report Tg in three ways from this step: the onset (where the curve first departs from the glassy baseline), the midpoint (halfway through the step), and the inflection point (the steepest part of the step). ASTM E1356 defines how to assign each of these, and reporting which convention you used is essential for anyone comparing your data.
When Standard DSC Falls Short
A common problem with conventional DSC is that the glass transition step can be masked by overlapping thermal events. If a polymer has been stored or aged, it builds up enthalpy relaxation, which shows up as a peak sitting right on top of the Tg step and makes it hard to read. Crystallization, curing reactions, or moisture loss can cause similar interference.
Modulated DSC (MDSC) solves this by applying two heating rates simultaneously: a steady linear ramp plus a small sinusoidal oscillation. This separates the total heat flow into two components. Because the glass transition involves a change in heat capacity, it appears cleanly in the reversing heat flow signal. Kinetic events like crystallization, curing, and evaporation show up only in the non-reversing signal. If your material has a weak or buried Tg, MDSC can pull it out of the noise in a single measurement.
Dynamic Mechanical Analysis (DMA)
DMA measures Tg through mechanical properties rather than heat capacity, and it is significantly more sensitive. The instrument applies a small oscillating force to a sample while ramping the temperature, then tracks how the material responds. It reports three quantities: the storage modulus (a measure of stiffness, or how much energy the material stores elastically), the loss modulus (how much energy is dissipated as internal friction), and tan delta (the ratio of the two, which represents damping).
As a polymer passes through Tg, its stiffness drops dramatically. The storage modulus falls by orders of magnitude, the loss modulus spikes as internal friction peaks, and tan delta reaches a maximum. Each of these gives you a Tg value, but they won’t be the same number. The onset of the storage modulus drop gives the lowest value and is closest to the DSC Tg. The loss modulus peak falls a few degrees higher, and the tan delta peak is higher still. Always specify which signal you’re reporting.
DMA is the better choice when DSC struggles. Temperature transitions are more detectable by DMA because the mechanical changes at Tg are far more dramatic than the changes in heat capacity. DMA can pick up short-range molecular motion before the main glass transition even begins, making it especially useful for highly cross-linked resins, thin coatings, biomaterials like starches and flours, and any sample with low moisture content where DSC transitions become indistinguishable. Typical DMA instruments use liquid nitrogen cooling to start well below Tg and can ramp up to around 250 °C, covering the range for nearly all common polymers.
Thermomechanical Analysis (TMA)
TMA takes a simpler approach: it measures how a sample’s dimensions change as it heats. A probe rests on the sample surface under a light load, and the instrument records expansion or contraction as a function of temperature. Below Tg, the material expands at a relatively low, steady rate. At the glass transition, the coefficient of thermal expansion increases sharply as polymer chains gain mobility, producing a visible change in slope on the dimension-versus-temperature curve.
When using a penetration probe on a coating or soft material, the transition appears as a step change in dimension where the probe begins to sink into the softening surface. Tg is reported as the extrapolated onset of that dimensional change, found by drawing tangent lines along the curve before and after the transition and calculating their intersection. TMA is particularly useful for coatings, films, and samples where you care about dimensional stability as a practical property, not just the transition temperature itself.
Reading the Tg From Your Data
Regardless of technique, extracting Tg from a curve requires choosing a consistent convention. For DSC, the onset is found by drawing a tangent along the pre-transition baseline and another along the steepest part of the step, then finding where they intersect. The midpoint sits halfway between the onset and endset (where the curve returns to a new baseline). The inflection point is the temperature at which the curve’s slope is greatest.
For DMA, the onset Tg is calculated the same way using the storage modulus curve: draw a line along the glassy plateau and another through the steepest part of the modulus drop, then find their intersection. The loss modulus and tan delta Tg values are simply the temperatures at each peak. Tan delta peaks tend to give values 10 to 25 degrees higher than the storage modulus onset, so comparing Tg data across studies requires knowing which definition was used.
Choosing the Right Method
For routine polymer characterization with a well-defined transition, DSC is the standard. It’s fast, requires minimal sample preparation, and aligns with ASTM E1356. If the transition is weak, broad, or buried under other thermal events, switch to MDSC to separate the glass transition from overlapping signals, or to DMA for maximum sensitivity. DMA also gives you mechanical property data alongside Tg, which matters in applications where stiffness and damping are design-critical. TMA is the go-to when dimensional change or thermal expansion behavior is the property you actually care about, such as in coatings or electronic packaging.
In practice, many labs measure Tg by more than one method and report the values together. This provides a more complete picture, since DSC, DMA, and TMA each probe the transition through a different physical lens and will yield slightly different numbers for the same material. The differences are not errors. They reflect the fact that the glass transition is not a single sharp event but a gradual process, and each technique catches a different stage of that process.