A differential scanning calorimeter (DSC) measures how much heat a material absorbs or releases as it’s heated or cooled at a controlled rate. It does this by comparing a small sample against an empty reference pan, tracking the difference in heat flow between the two as temperature changes. That difference reveals phase transitions like melting, crystallization, and glass transitions, each producing a distinct signature on the output graph. The technique works across a standard range of about -120°C to 500°C, with some instruments extending further.
The Core Principle: Sample vs. Reference
Every DSC measurement relies on the same basic idea: heat your sample and an empty reference side by side, then measure the difference. A material with no transitions would track closely with the empty pan. But when something interesting happens, like melting, the sample absorbs extra energy and falls behind the reference in temperature. The instrument detects that difference and converts it into a heat flow signal.
There are two main instrument designs that accomplish this differently.
Heat Flux DSC
In a heat flux DSC, both the sample pan and the empty reference pan sit on a single thermoelectric disk inside one furnace. As the furnace heats at a steady, linear rate, heat transfers through the disk to both pans. Because the sample has a heat capacity that the empty pan lacks, a temperature difference develops between the two sides. Thermocouples beneath each pan measure that temperature gap, and the instrument calculates heat flow using the thermal equivalent of Ohm’s law: heat flow equals the temperature difference divided by the thermal resistance of the disk. This is the more common design and tends to be simpler mechanically.
Power Compensated DSC
A power compensated DSC takes a different approach. The sample and reference each sit in their own separate furnace with their own heater. Instead of measuring a temperature difference, the instrument actively keeps both sides at the same temperature. When the sample undergoes a transition that absorbs heat (like melting), its furnace has to work harder to keep up. The instrument records that extra power as the signal. The result is the same kind of data, just measured through power input rather than temperature difference.
Reading a DSC Thermogram
The output of a DSC run is a thermogram: a plot of heat flow on the vertical axis against temperature (or time) on the horizontal axis. Flat regions where the baseline is stable mean nothing dramatic is happening. The interesting information lives in the peaks and shifts.
Peaks pointing in the endothermic direction (heat absorbed) correspond to processes like melting or certain solid-to-solid phase transitions where the material moves from a lower-energy crystal form to a higher-energy one. Peaks in the exothermic direction (heat released) correspond to events like crystallization or curing of a thermoset resin, where the material locks into a more ordered or cross-linked state and gives off energy in the process.
The area under a peak is directly proportional to the energy involved in that transition. For a melting peak, the area gives you the heat of fusion. For crystallization on cooling, the peak area is roughly the same as the melting peak, typically within about 20%, since the heat of fusion changes slightly with temperature depending on how much the material supercools before crystallizing.
The Glass Transition Looks Different
Not every thermal event produces a peak. The glass transition, where an amorphous material shifts from a rigid, glassy state to a softer, rubbery one, shows up as a step change in the baseline rather than a sharp peak. This happens because the material’s heat capacity increases at the glass transition temperature. Below it, the molecular chains are essentially frozen in place and don’t absorb much extra heat. Above it, they gain enough mobility to store more thermal energy, so the heat flow signal shifts upward.
One quirk of glass transitions in DSC is that the signal depends on how the glass was formed. A material that was cooled slowly before being reheated will produce a more prominent overshoot peak near the glass transition compared to one that was cooled quickly. This time-dependent behavior is what distinguishes the glass transition from true phase changes like melting, which happen at a fixed temperature regardless of thermal history.
Sample Preparation and Pan Selection
DSC samples are small, typically just a few milligrams, placed into metal pans (usually aluminum) that are then crimped shut. The type of pan matters more than you might expect.
- Standard pans work for solid, non-volatile materials. They’re crimped but not airtight.
- Hermetic pans are sealed airtight and used for liquids or any sample that loses more than about 0.5% of its weight through evaporation over the temperature range. Without a hermetic seal, evaporation creates misleading endothermic signals that can be confused with real transitions.
- Pin-hole pans have a tiny opening in the lid and are used for boiling point or vapor pressure measurements, where you want controlled release of gas.
For materials that are soft or reactive at room temperature, samples may need to be prepared and sealed in a refrigerated environment to avoid premature changes before the run even starts.
Why Purge Gas Matters
DSC runs don’t happen in open air. A purge gas flows through the cell during the entire experiment, typically at around 50 mL/min for nitrogen or 25 mL/min for helium. Nitrogen is the default because it’s inert, inexpensive, and widely available. It prevents oxidation that would otherwise create exothermic artifacts in the data.
Helium is the better choice for subzero work, especially below -100°C when using liquid nitrogen cooling accessories. Its high thermal conductivity improves the instrument’s response time and cooling performance. The tradeoff is that helium’s different conductivity affects the cell calibration constant, so switching gases means recalibrating.
Calibration With Indium
DSC instruments need regular calibration to ensure their temperature and energy readings are accurate. The most widely used calibration standard is indium, certified by NIST with a melting point of 156.5985°C and a heat of fusion of 28.58 J/g. A calibration run melts a small piece of indium and checks whether the instrument reads the correct temperature and peak area. If the values drift, correction factors are applied. This two-point check (temperature and energy) catches both sensor drift and changes in thermal contact within the cell.
Modulated DSC
Standard DSC heats at a constant rate, which works well when transitions are spaced apart. But when two events overlap in the same temperature range, a standard run blends them into one ambiguous signal. Modulated DSC (MDSC) solves this by overlaying a small sinusoidal temperature oscillation on top of the linear heating ramp. This lets the instrument separate the total heat flow into two components: a “reversing” signal related to heat capacity changes, and a “non-reversing” signal capturing kinetic events.
Glass transitions, which are heat capacity changes, appear cleanly in the reversing signal. Crystallization, enthalpic relaxation, evaporation, and decomposition show up in the non-reversing signal. This separation is especially useful in polymer analysis, where a glass transition and a crystallization event can land at nearly the same temperature and would be impossible to distinguish in a standard run.
Common Applications
DSC is one of the most versatile thermal analysis tools across industries. Polymer scientists use it to identify plastics by their melting points and glass transitions, or to measure how crystalline a material is by comparing its heat of fusion against a fully crystalline reference value. Pharmaceutical researchers rely on DSC to check the purity of drug compounds, since impurities lower and broaden the melting peak in predictable ways. Food scientists track fat melting profiles in chocolate and dairy products. Materials engineers use it to study metal alloys and ceramics.
In quality control, DSC is fast and requires very little material. A single run at 10°C per minute through a relevant temperature range takes under an hour and consumes only a few milligrams of sample, making it practical for routine batch testing alongside more detailed characterization work.