Because thermocouples come with varying wire diameters (see Table 2.34), select the thermocouple wire size best suited to measure your sample.

Be advised that the upper range of temperatures cited (within thermocouple catalogs) for a given thermocouple are related to the larger wire sizes. Thus, a smallwire thermocouple is likely to fail at the upper temperature ranges for that particular thermocouple type.

Environment also influences the selection of thermocouples. The condition of the region where the thermocouple will be placed can be oxidizing, reducing, moist, acidic, or alkaline, or it can present some other condition that could cause premature failure of the thermocouple. Selection of the right type of thermocouple can help avoid premature failure. Fortunately, there are sleeves and covers available for thermocouples that prevent direct contact with their various environments. These covers are made of a variety of materials, from metals to ceramics, making selection of the right material easy. On the other hand, covers add to the heat capacitance of the entire probe and therefore can slow thermocouple response time; and because of a greater heat capacity, they are more likely to affect the temperature of the material being studied.

2.5.12 Resistance Thermometers

In 1821, Sir Humphrey Davy discovered that as temperature changed, the resistance of metals changed as well. By 1887 H.L. Callendar completed studies showing that purified platinum wires exhibited sufficient stability and reproducibility for use as thermometer standards. Further studies brought the Comite International des Poids et Measures in 1927 to accept the Standard Platinum Resistance Thermometer (SPRT) as a calibration tool for the newly adopted practical temperature scale.

Platinum resistance thermometers are currently used by the NIST for calibration verification of other thermometer types for the temperature range 13.8 to 904 K. In addition, they are one of the easiest types of thermometers to interface with a computer for data input. On the other hand, platinum resistance thermometers are very expensive, extremely sensitive to physical changes and shock, have a slow response time, and therefore can take a long time to equilibrate to a given temperature. Thus, resistance thermometers are often used only for calibration purposes in many labs.

Platinum turned out to be an excellent choice of materials because it can withstand high heat and is very resistant to corrosion. In addition, platinum offers a reasonable amount of resistivity (as opposed to gold or silver), yet it is very stable and its resistance is less likely to drift with time. However, because it is a good

conductor of electricity, the SPRT requires a sufficiently long enough piece of platinum wire* to record any resistance.

One of the easiest ways to get a long piece of material in a small (convenient) area is to wrap or wind the material around a mandrel. The typical mandrel used on SPRTs is made of either mica or alumina. The sheaths covering the wrapped thermometer may be borosilicate, silica glass, or a ceramic (see Fig. 2.33). Note that both examples in Fig. 2.33 show four leads where, logically, there should be only two. This is done to reduce any unwanted resistance from the region beyond the thermometer.

An alternative approach to "loose winding" the wires is to form the platinum wire and then flow melted glass around the wire to "lock" the wire in place. This method protects the wire from shock and vibration. These resistance thermometers are, unfortunately, limited to temperatures where the expanding platinum will not crack the containment glass. To limit this occurrence, the expansion of the platinum must closely match the glass around which it is wrapped. Resistance thermometers of this design are more rugged, and therefore, they are more likely to be used in the laboratory.

There are many challenges in the construction and use of resistance thermometers, including:

1. The wire itself must be very pure; any impurities will affect the linearity and reliability of the resistance change. The wire itself must also be uniformly stressed, meaning that after construction, the wire must be temperature-annealed to achieve uniform density. Uniform density provides uniform resistance.

Fig. 2.34 Two different SPRT wrapping designs.

'Currently, about 61 cm of 0.075-mm platinum wire is typically used on resistance thermometers.

2.When the wire is wound on the support device, it must be left in a strainfree condition, as any strain can also affect the resistance of metals. Because the platinum will expand and contract as the temperature changes, it must be wound in such a manner that there is no hold-up or strain. Even this strain could affect the resistance of the metal.

3.Although it is doubtful that a lab will make its own SPRTs, I mention construction demands to impress upon the user the importance of maintaining a strain-free platinum wire. Strain on the wire can be introduced not only during construction, but also in use. The most likely opportunities for wire strain are through vibration and bumping.

4.The SPRT must not receive any sharp motions or vibrations. Such actions can affect the resistance of the metal by creating strain. To place this challenge in the proper perspective, consider a SPRT rapped against a solid surface loud enough to be heard (but not hard enough to fracture the glass sheath). The temperature readings could be affected by as much as 0.001°C. Although this variation may not seem like much, over a year's time, such poor use could cause as

much as 0.1°C error.19

Transportation of SPRTs should be as limited as possible. If you are shipping, special shock-absorbing boxes are recommended. If you have an SPRT shipped to you, keep the box for any future shipping needs and storage. Calibrated SPRTs should be hand-carried whenever possible, to minimize shock or vibration. For instance, you should not lay an SPRT on a lab cart for transportation down the hall. The vibration of the cart may cause changes in subsequent temperature readings.

Other precautions that should be taken with SPRTs include the following:

1.When SPRTs are placed into an apparatus, they should be inserted carefully, to avoid bumps and shocks.

2.Try to avoid rapid dramatic changes in temperature. A cold-to-hot change can cause strains as the wire expands within the thermometer. A hot-to-cold change can cause fracture of the glass envelope encasing the thermometer, or of any glass-to-metal seals (which are structurally weak). A hot-to-cold heat change can also cause a calibration shift of the thermometer.

3.Thermometers with covers of borosilicate glass should not be used in temperatures over 450-500°C without some internal support to prevent deformation.

4.Notable grain growth has been observed in thermometers maintained at 420°C for several hundred hours.20 Such grain growth causes the

AG, printed in Switzerland.

14.Ibid, Ref. 9, pp. 69. 74, 100, and 118.

15.Jacquelyn A. Wise, Liquid-in-Glass Thermometry, U.S. Government Printing Office, Washington, D.C., 1976, p. 23.

16.E.L. Ruh and G.E. Conklin, "Thermal Stability in ASTM Thermometers," ASTM Bulletin, No. 233, p. 35, Oct. 1958.

17.W.I. Martin and S.S. Grossman, "Calibration Drift with Thermometers Repeatedly Cooled to -30° C," ASTM Bulletin, No. 231, p. 62, July, 1958.

18.W.P. White, "Lag Effects and Other Errors in Calorimetry," Physical Review, 31, pp. 562-582 (1910).

19.J.L. Riddle, G.T. Furukawa, and H.H. Plumb, Platinum Resistance Thermometry, National Bureau of Standards Monograph No. 126, U.S. Government Printing Office, April 1972, p. 9.

20.Ibid, Ref. 19, p. 11.