Two pans warm inside the same furnace
Even without a transition, the sample adds heat capacity to its pan. Raising that side through the same temperature interval requires more energy than warming an empty reference. Heat reaches both pans through finite conducting paths, so the sample side generally warms a little more slowly. During a steady ramp, the difference can settle to an approximately constant value.
Near melting, some incoming heat changes the sample’s state. An ideal pure substance can absorb heat at nearly constant temperature as it melts, while the reference continues warming. The sample temporarily falls further behind. Once melting finishes, its warming rate catches up again. The enlarged difference between the two temperature curves supplies a melting signal.
A calibrated thermal path converts difference into heat flow
Heat flow is energy transferred each second. A heat-flux instrument connects sample and reference to a common furnace through similar thermal paths and measures their temperature difference. Calibration converts that signal into a difference of heat flows. Here an endothermic process points upwards; another sign convention can draw the same melting process downwards.
A power-compensation instrument instead controls the two sides separately and compares their required electrical powers. Both architectures can measure transition heat, but the signal follows a different physical path.
Sources: [1]
Normal warming lies underneath the peak
The entire area between zero and the measured curve does not belong to melting. Ordinary warming requires heat, and the pans can store different amounts as their temperatures change. Instrument offsets can also add a blank signal. An empty-pan measurement removes the instrument background. The heat flow expected without a transition supplies the sensible-heat baseline. Removing both leaves the extra transition response.
Integrating power over elapsed time gives energy. Let be transition enthalpy per unit mass, the sample mass, the calibrated heat-flow difference, its baseline and time. At approximately constant pressure, a fully collected transition gives the relation below.
For example, a sample weighing ten milligrams that absorbs eighty joules per gram during a transition takes in an extra eight tenths of a joule. Doubling the mass and completing the same transition approximately doubles the heat. Enthalpy per unit mass stays the same. If the horizontal axis is temperature, integration must first convert temperature intervals to elapsed time using the actual heating history.
A faster scan can move the peak
When the furnace warms faster, finite heat transfer lets the sample lag farther behind. Assigning the furnace or programmed temperature to the sample folds this delay into the apparent peak position. Recording the actual temperatures of both pans helps separate that thermal delay.
The material’s transformation can also take time. Melting and molecular rearrangement have finite rates. Even a curve plotted against actual sample temperature can shift or develop a longer tail during faster heating. The total heat of the completed transition can remain nearly unchanged while the timing changes. Peak temperature, a specified onset convention and equilibrium transition temperature describe different features.
Keep the tail and measure a reference
Ending a program as the peak rises collects only part of the transition heat. The sample may continue transforming during a final hold. Waiting for the response to return to a stable baseline helps collect the full area. More complex materials can also have temperature-dependent sensible heat, requiring a correspondingly different baseline.
Reference materials such as indium have known melting temperatures and fusion enthalpies that help check the temperature axis and heat-flow sensitivity. NIST’s 2024 certification of a new indium batch established its temperature and calorimetric values against traceable references. A useful measurement report also identifies heating rate, sample mass and baseline treatment, so its peak area and position can be interpreted under the conditions that produced them.