Author: Thomas Krysciak, Engineer at Heraeus Sensor Technology, Kleinostheim, Germany

Today's market presents an ever increasing number of low-temperature measurement applications for resistance temperature devices. Among these are storage of food and medical supplies, laboratory research, and liquification of pure gases. All require repeatable temperature control. The sensors used to measure cryogenic temperatures include germanium semiconductors, carbon sensors, thermistors, thermocouples and platinum resistance elements

Many of these sensors, however, exhibit significant shortcomings:

• Measuring spans of ~ 100C.  For wider spans, more specifically in the range below  -200C to above 150C, it may be necessary to use a combination of sensor types in parallel. This applies in particular germanium semiconductors, carbon sensors and thermocouples.
• Wide variation of the characteristic curve.  This can necessitate changing the analyzing electronics when the sensors are replaced (applies to carbon, germanium sensors and thermocouples).
  
• Deterioration of the characteristic curve. The characteristic curve is greatly influenced by cyclical temperature exposure (applies to carbon sensors and thermocouples).

These negative effects, for all applications over the temperature of 10 kelvin (K), can be avoided by using platinum temperature sensors. In addition to the geometrical dimensions, nominal resistance and price, one must also consider the following when using platinum RTDs in cryogenic applications:

• Sensitivity (R/T) determines the temperature resolution and helps defines which temperatures are to be measured at the lower end.
• Repeatability is the ability to maintain measurement accuracy over the long term and after an extensive number of temperature cycles.
• Response time is the time constant required to react to temperature change in relation to the control instrumentation and corresponding safety circuits.
• Self-heating can also be worth considering.  To minimize the potential effect of self-heating, it is advisable to minimize the measuring current.  For an example, when using a 1000 platinum RTD, the measuring current used should be 0.1 mA.

 

Sensitivity (R/T)

The resistance of metals can be determined according the Matthiessen's rule, an empirical rule stating that the total resistivity of a crystalline metallic specimen is the sum of the resistivity due to thermal agitation of the lattice's metal ions and the resistivity due to imperfections in the crystal.  The temperature-dependent part can thus be seen to change nearly linearly with the temperature.  The temperature coefficient of pure platinum, when  measured between 0C and 100C is 3.929 x 10-3 K-1.

The part that is not  temperature dependent, also known as "rest resistance" is determined by impurities in the crystal. An impurity embedded in the platinum lattice causes an increase in the rest resistance.  Impurities are therefore intentionally introduced in order to manufacture a standard temperature coefficient. Because most temperature measurement applications are above 220 K, the effects of rest resistance can usually be disregarded.  Temperatures below 220 K cause problems with the rest resistance, especially when the temperature falls below 30 K.  Depending on the amount of doping of the platinum lattice, a rest resistance of <0.5  to 2  can be achieved.  The resistance curve flattens strongly below 30 K. (See Figure 1). The unpredictable spread of the rest resistance and the very strong flattening of the temperature characteristics necessitate an individual calibration of each resistance sensor in this measurement range.

A higher output signal facilitates easier analysis for low-temperature measurement applications.  This can be achieved only with thin film platinum temperature sensors.  Thin films allow for nominal values to be increased to 1000 (and higher) at 0C.  When compared to the standard platinum 100 element, the resolution is increased 10X.

Resistance value in function versus temperature

In the transition range of an almost linear resistance temperature curve into the rest resistance, the sensitivity strongly decreases and limits the measuring span.  To carry out measurements below 30 K, only one method is practical - individual calibration of each sensor element.

When summarizing the characteristics of platinum sensors, it is necessary to look at three temperature ranges separately:

• Temperatures above 30 K can be measured easily with platinum temperature sensors.Temperatures of 10 K to 30 K can be measured with high ohmic and individually calibrated platinum temperature sensors.
• Temperatures below 10 K can only be measured with intense technical efforts.

Repeatability

The stability of platinum RTDs is seen most clearly in low-temperature applications. A shift in the calibration point is precipitated at low-temperature exposure. The resistance elements exhibit a hysteresis when subjected to temperatures below -50C and then later used at temperatures above 150C. At 150C or below, the shift in the temperature curve remains stable; at higher temperatures, the curve returns to its original calibration value. The reason for this is the combination of the platinum and the ceramic carrier materials, whose different temperature expansion coefficients cause a plastic deformation of the platinum at low temperatures. This deformation increases with decreasing temperature. Cycling the sensor between the lowest temperature to be measured and room temperature stabilizes the deformation and permits a consistent measuring accuracy. Sensors designed specifically for cryogenic applications are conditioned (e.g. in liquid nitrogen) to anticipate the shift of the characteristic curve. Further cycling during regular use does not change the values, but it should be noted that to avoid a change in the curve, sensors that have been cryogenically conditioned should not be exposed to temperatures above 150°C.

Response Time

The response time of the temperature sensors is influenced by several factors:

• Thermal mass
• Heat transfer coefficient
• Heat dissipation through the connection leads

Platinum RTD response time is determined according to Standard IEC 751 in air and in water at room temperature.  The heat transition coefficient strongly decreases at low temperatures, making the response time significantly greater than it is at room temperature. The corrective factors for the heat transfer coefficient in different media and varying at various median temperatures should also be taken into consideration.

Summary

Platinum temperature detectors are becoming the sensors of choice for low-temperature measurements.A single sensor, properly conditioned and limited to a maximum temperature of 150C, provides long-term stability, and fast response times.

In addition, output signals are repeatable, have the potential for high values, and can be based on industry standards