Accurate measurements of specific heat capacity for lipids as a function of temperature, Cp (T), are needed for modeling their crystallization behavior. Differential Scanning Calorimetry, DSC, has been the main technique to determine Cp (T) for numerous materials, including lipids. Key experimental conditions (heating/cooling ramps, sample size, purge gas, and temperature modulation) that affect the measured heat flow, from which Cp (T) is calculated, have been extensively discussed in the literature. Usually, DSC manufacturers provide procedures of what they consider to be the best experimental conditions to measure accurate Cp (T) values with the least uncertainties. The successive nature of these procedures requires the user to perform each step separately, which means that the user needs to take out the empty pan from the DSC furnace to load either the standard material (usually sapphire) or the sample and place it back again into the furnace. Following this method will result in a different pan placement on the sensor each time the DSC furnace is opened, which consequently will influence the heat flow signals. Utilizing the Guide to the Expression of Uncertainty in Measurements (GUM),this paper is intended to quantitatively evaluate the uncertainties in Cp (T) measurements due to the pan position on a heat-flux DSC sensor. Due to this effect, relative expanded uncertainty U values were ∼1.5%, and at least 15–25% as a result of pan placement. The sapphire uncertainty values were much smaller than those from the trimyristin (MMM) sample. With the assistance of FEM simulation, the effect of the different thermal diffusivity of MMM and sapphire on the Cp (T) measurements is elucidated.
Read full abstract