Micro-Thermal Analysis is an emerging technology in which thermal analysis is performed on small specimens as small as 2 x 2 µm. The sensor is a microscopically fine 5 µm diameter, "V" shaped platinum wire. Current is passed through the wire to raise its temperature and the temperature of the specimen according to a linear program. The resistance of the wire is simultaneously measured to yield sample temperature information. Most of the energy in the probe is transferred to the heat capacity of the test specimen. During a transition, however, additional power is needed. When this signal is compared to that of a reference, a Micro Differential Thermal Analysis signal results. Moreover, the probe is loaded with a small force, so that it penetrates into melted specimen resulting in a second analytical signal, that of Micro Thermomechanical Analysis. Due to the small thermal mass of the sensor, heating rates up to 25°C/s are used.

The micro-TA sensor is mounted on the tip of an Atomic Force Microscope (AFM) probe. This permits imaging capabilities of the AFM to be used to map the surface area of interest and to select the appropriate regions for subsequent thermal analysis.

Micro thermal analysis is currently a qualitative tool; that is, it is primarily used to identifying the material being examined. The temperature at which a transition takes place is the primary output signal. Like all thermal analysis techniques, interpretation of results requires detailed temperature calibration procedures and an understanding of the precision of the temperature measurement.

A good reference material has a number of desirable properties including a well-documented value, availability in a suitable form for analysis, homogeneity, stability, low toxicity, and traceability to a national reference laboratory (NRL). In traditional DTA and TMA, metals like indium, tin, and zinc meet these criteria. These metals are not suitable for temperature calibration in micro-TA, however, as they may alloy with the platinum sensor with which they come in to contact.

Organic calibration materials are more suitable for calibration in micro-TA as they do not react with the platinum sensor and the sensor is easily cleaned at the end of an experiment by "burning off" at 500 °C in air. The British Laboratory of the Government Chemist (LGC), a national reference laboratory, conveniently offers eleven organic reference materials with melting temperatures ranging from 41 to 285°C. Six of these standard reference materials, which are used as calibrants in this study, are shown presented in below with their nominal melting temperatures.

Organic chemicals are not ordinarily suitable for use as micro-TA calibration materials in the fine powdered form in which they are received. The individual crystals are usually too small and rough to be used directly. For ease of handling, powdered organic materials need to be formed into larger crystals by recrystallizing from the melt. Suitable crystals are prepared by melting a few milligrams of the organic compound in a standard 6.3 mm aluminum DSC sample pan with an aluminum lid placed lightly (not crimped) on top. The sample is heated in a DSC at 10°C/min to a temperature 10°C about the nominal melting temperature, then quickly cooled. (Alternatively, the DSC sample pan may be placed for a few seconds on a hot plate set to a temperature about 10°C above the nominal melting temperature.) This procedure melts the material without exposing it for long periods (and potential degradation) at temperatures above the melt. The resulting organic crystals are prepared for mounting by cutting down the edges of the DSC sample pan with an Exacto? knife. These crystals may then be fixed to the micro-TA sample mounts using double-sided tape to prevent the individual crystals from moving around during storage and analysis. The mounted samples are stored in small, labeled petri dishes for future use.

The use of the aluminum sample lid during large crystal formation results in a smooth flat surface required for micro-TA work. The aluminum pan and lid provide smooth, inert surfaces upon which the molten organic may crystallize. Crystals created without the use of the lids have an unsuitably ragged surface.

Most thermal analysts prefer a two-point calibration procedure. This is performed using a low temperature calibrant, such as biphenyl at 69°C, and a high temperature one, such as anisic acid at 183°C. A typical thermal curve for micro-TA of a sample of anisic acid is shown in below. The melting endotherm is seen near 170°C. In micro-TA, the sensor is located at the surface (i.e., outside) of the test specimen, as it is in DSC, and so the extrapolated onset is used to identify the melting temperature.

Triplicate measurements of the melting temperature are made and the mean temperature is used to temperature calibrate the apparatus. The pooled standard deviation of the two triplicate measurements gives an estimation of measurement precision.

A comparison of melting temperatures observed by micro-DTA versus the certified melting temperatures provided by LGC are shown below. A straight calibration line (correlation coefficient of 99.9%) is observed for covering a 150°C temperature range. This calibration profile is stored in the instrument for future work. If desired, first, second or third order fits to calibration data may be used.

Measuring the melting temperatures of polymeric samples is the most common use of micro-TA. The micro-DTA response signal for polymers has a somewhat different shape than that for pure organic chemicals. The difference in response is thought to be due to the organic melt shrinking away from the hot thermal probe and loosing contact, creating the melting "peak" response. The less mobile polymer materials remain in contact and simply continue to absorb energy as the molten front grows away from the probe embedded in the viscous melt.

The extrapolated onset for the change in power consumption at the melt may be used to identify the polymer melting transition temperature. This value may be correlated with the melting onset temperature observed by DSC. As is observed with the organic materials, the DTA response with polymers is a straight line with a correlation coefficient of 99.9%. The line has the same slope as that observed with the organic chemicals but is offset to higher temperatures by about 11°C. The "kick-in" transition, which results when the temperature program passes through ambient temperature (see anisic acid curve above), provides a second, low-temperature calibration point in each thermal curve.

A second and somewhat simpler temperature calibration procedure may be used in micro-TA with polymer samples. A single, high melting polymer (the authors prefer Nylon 6, Nylon 6,6 or polyethylene terephthalate) is used to mark the high temperature end of the calibration curve and the extrapolated ambient temperature "kick-in" signal at the low end.

Temperature calibration with polymeric films offers a number of ease-of-use advantages, chief among these is the reduction (by half) in the time and number of experiments needed to calibrate the apparatus. The sample mounting procedure for smooth, clean and flat polymeric films is much easier. Samples are conveniently cut out with a 6.3 mm paper punch and mounted using double sided tape. Additionally, polymers have very high melting temperatures that provide a calibration range of nearly 300°C, a range difficult to achieve with organic chemicals.

Calibration curves for organic and polymer materials may be stored separately within the instrument. The appropriate calibration curve may then be selected to match the type of test specimen being examined. Replicate measurements made on both organic crystals and polymeric films indicates that the standard deviation for melting temperature determine by the micro-TMA signal is about 0.2°C while that for the micro-DTA is about 2.1°C. In addition, the micro-TMA transition is about 3.5°C higher than the extrapolated onset temperature determined by micro-DTA. The precision of temperature measurement appears to be somewhat dependent upon heating rate at the very fast rates typically employed with micro-TA. Heating rates between 10 and 25°C/s (i.e., 600 to 1500°C/min) give consistent results, while those at lower rates between 2 and 8°C/s (120 and 480°C/min) have substantially poorer precision.

The text and figures shown above are adapted from "Micro-Thermal Analysis Calibration, Repeatablity and Reproducibility." by Roger Blaine, Gray Slough and Duncan Price, published in Proc. 27th NATAS Sept 20-22 (1999) Savannah, Georgia, pp. 691-696 [available as PDF (105kB)]