Problems with Thermocouple Measurements

/Home/Lyophilization Articles & Reports

Problems with Thermocouple Measurements
July 2003
Insight
There are at present two basic temperature transducers used in conjunction with the lyophilization or freeze drying process, i.e., RTD and thermocouple. In the previous INSIGHT (Vol. 6 No. 6) we considered the impact that RTD?s can have on the process, namely the potential for false readings of the product temperature. A few of the readers of INSIGHT expressed their views with respect to this INSIGHT and they appear in Comments 66.
I realize that by showing RTD?s have the potential for indicating false temperature measurements, it may be implied that the thermocouple transducer is the ideal sensor to use. That is hardly the case and as with any measuring device it has its own set of limitations and likewise, if incorrectly used, can give false temperature readings. Therefore, it will be the objective of this INSIGHT to examine the use of the thermocouple in the lyophilization or freeze drying process and in particular the measurement of the product temperature.
A previous issue of INSIGHT (Vol. 4 No. 8) described just how a thermocouple works. If you have not read this INSIGHT it is suggested that you read that INSIGHT before proceeding with this one. An understanding of the previous INSIGHT will prove helpful when reading this INSIGHT. Assuming that you now have a basic understanding of just how a thermocouple works, let us consider some issues concerning its use as a product temperature sensor.
Fundamental Measuring Circuit.
The fundamental (classical) measuring circuit for a type ?T? thermocouple is illustrated by Figure 1. Examination of this figure shows that copper wire of the thermocouple sensor is connected to a terminal. A constantan wire is connected to the copper wire to form a copper-constantan junction as described in INSIGHT Vol. 4 No. 8. This constantan wire also forms a copper constantan junction which is immersed in a ice bath that is maintained at 0 oC. The copper wire from this second junction is connected to a second terminal. A null current indicating instrument is used to determine the difference in potential between the two copper wires when the current ?i? of the circuit is equal to zero. It is important to note that the output of all thermocouple readings are expressed in terms relative to the output of the thermocouple junction of the same composition maintained at 0 oC. To illustrate the latter point, let us examine three examples using the circuit shown in Figure 1.

Example 1. The temperature of the measuring sensor is at a temperature greater than 0 oC. In this case the null indicator will show that the output voltage of the sensor (millivolt) is greater than that of the reference in the ice bath and difference in potential will be positive. As the temperature sensor junction is increased the magnitude of the mV reading will also increase.
Example 2. In this example the temperature of the sensor is equal to 0 oC. At this point the output of the sensor and reference thermocouples are equal but opposite in sign so the resultant mV reading read by the null indicating instrument will be 0 mV. This is an important point for the reader may see a table showing the mV readings at various temperatures; however, the reader should now realize that there is always an output voltage from the sensor thermocouple and the 0 mV merely signifies that the output of the measuring circuit - not the measuring thermocouple - will be zero when the temperature of the sensor is equal to zero.
Example 3. If the temperature of the sensor is now reduced so that it is lower than 0 oC, then the output of the junction in the ice bath will exceed that of the sensor. Hence the output voltage of the circuit will have a negative value. Lowering of the temperature of the sensor to even lower temperatures will increase the magnitude of the mV output from the null indicating instrument but the sign of the voltage will remain negative.
Two important points to keep in mind at all times when making temperature measurements using a thermocouple is that the output voltage is relative to the output of the same type of thermocouple at 0 0C. In addition, it is equally important that the current in the circuit at the time the temperature reading is taken is equal to zero.
Now the reader should be thinking at this point, that is all well and good but I don?t use a reference ice bath to make my temperature measurements. The answer is no, your system does not need an ice bath for today?s thermocouple instrumentation uses an output voltage to simulate the thermocouple output voltage at an ice bath temperature.
Sources of Errors or Effects of the Thermocouple
Now let us consider some potential sources of errors when using a thermocouple.
Drift in the Reference Voltage - This is by no means a common error but it can happen especially in those systems that employ a battery to provide the reference voltage to simulate the ice bath. When using such instruments one must be careful to check to see that the battery supply is within the prescribed limits. It is always a good practice to check the reference voltage output by measuring the temperature of several thermocouples in an ice bath at 0 oC. If there is a significant error in the reference voltage output then one would obtain constantly erroneous temperature readings for all of the thermocouple sensors.
Fabrication of the Thermocouple - Thermocouples are relatively easy devices to fabricate. In fabricating the thermocouple one must avoid making two basic mistakes. The first mistake may be in the selection of the gauge (diameter) of the thermocouple wire. As the diameter of the wire increases, it becomes increasingly difficult to position the thermocouple in the product container. The junction may end up contacting the walls of the container and movement of the shelves, e.g. during stoppering may cause the wire to remove closures from some of the vials. It is recommended that the diameter of the wire be equal or smaller than 0.010 inches (2.54 mm)

Second, one should exercise care in the formation of the thermocouple junction. One can make a junction by simply twisting the two wires together to form a braid. For reasons not entirely clear to me, I have found that there is often the misconception that the greater number of braids, the more sensitive will be the junction of the thermocouple. Figure 2 shows two examples of junction for product temperature thermocouples

In example 1, the number of braids is excessive and the actual mV output of the thermocouple will be when the wires first join at T(1). Since the thermocouple junction extends well above the fill-volume of the product, the sensor may only indicate the correct temperature when the system is at ambient temperature, after the product has reached and maintained some steady state at the end of the freezing process and possibly at the end of the secondary drying depending on how long the product temperature has reached and maintained the shelf temperature. In this example there is little hope of an accurate measurement of degree of supercooling (see INSIGHT Vol. 5 No. 5 and Vol. 2 No. 1) nor will one have a reliable indication of the completion of the primary drying process.
Example 2 shows that there is a minimum number of braids and the junction is located 1 to 2 mm from the bottom of the container. Thus in this example, the thermocouple junction is indicating the temperature at a point in the product. This will prove most useful in determining the degree of supercooling during the freezing process and the completion of the primary drying process. By positioning the thermocouple junction directly in the center of the vial where the final ice sublimation will take place, there will be a sharp increase in temperature as the ice-gas interface passes by the junction. This will be helpful in establishing the completion of the primary drying process. However, one must always keep in mind that there will be a frequency distribution of heat transfer coefficients (Co ) for the vials [1]. Thus the completion of the primary drying in the vial containing the thermocouple is no guarantee that all of the other vials have completed the primary drying process. Those vials not completing the primary, when the secondary drying is commenced, may show signs of partial meltback (see INSIGHT Vol. 1 No. 5)
Clean Contacts - It is important the connection at the copper - copper and constantan - constantan junctions are kept clean and free of any oxide or any other form of contamination. This will be particularity troublesome for those systems which require steam sterilization. The presence of such oxides or contamination can alter the ?electron work function? of the surface (see INSIGHT Vol. 4 No. 8). This can lead to the formation of a second junction at the contact and an error in the thermocouple reading. I have found that contamination of the contacts can lead to an error in the temperature reading as much a 2 oC. The same would be true if the wire used to form the thermocouple sensor junction also becomes corroded or contaminated. It is best that the junction be protected from any direct contact with the formulation where in time corrosion can take place.
Effect on the Cake Structure - It has been reported in the literature by several investigators that the metal surface of the junction or the bare wire can act as a nucleation site for ice formation and thus greatly alter both the degree of supercooling and the ice structure in the cake. Thus these authors found that vials containing thermocouples would have a significantly different drying rate than the other vials in the system [1]. I do not disagree with the finding of these authors but, in spite of performing a rather large number of drying processes, I never observed any significant difference in cake structure between those vials with and those without the thermocouples. Should you find