Can a -20 °C reference point be established in a laboratory?

Speaker / Author: Yvette Volschenk

Company: AFRIQLABS

Centurion

South Africa

Cell: 083658 8373

E-mail:

Abstract:

When measuring temperature, an ice point is used as a zero reference point. Using an ice point, we can monitor our reference standard’s drift in a laboratory. An ice point is stable and can be made up exactly to read the same temperature using the same method every time. An ice point temperature is homogeneous from top to bottom.

By mixing saved ice and salt we can make a -20 °C point. How accurate is a -20 °C ice and salt mixture? Can we use the -20 °C point as a reference point as well or to calibrate other temperature equipment? The question that we must ask ourselves is: How stable is a -20 °C ice salt mixture? How accurate is a -20 °C ice salt mixture? How repeatable is a -20 °C ice salt mixture? Is the -20 °C point also homogeneous?

This paper discussed the research and development conducted by the laboratory and presents some interesting conclusions.

  1. Introduction

There were a few discussions around using a -20 °C reference point to calibrate thermocouples on-site. I was very curious on how this can work. There is Dry Block Calibrators available that can go down to -20 °C. Unfortunately it is not cheap.

In the food industry, especially the fish industry, this is a very important temperature to calibrate. Fish must be kept at -20 °C, on the boats, after it was caught.

I wanted to know if a -20 °C reference temperature point can be accurate and stable.

  1. Experiment

When I consulted for a few laboratories, I told them about the -20 °C reference temperature point, but never used it myself and never analysed a -20 °C reference temperature point. I was glad that we could make one and measured ± -20 °C.

I decided to experiment with the -20 °C reference temperature point and see for myself what the outcome was. I was very surprised of the outcome of the results.

It was not just mixing ice and salt together as I thought it would be. Talking with my fellow temperature metrologist, it sounded like that.

If you google -20 °C reference temperature point, there is no information regarding using it in metrology, how stable it is, and how long this reference point last.

There is a lot of chemical explanations on a salt and ice mixture. Quoted from the internet: ‘The lowest temperature obtainable for a brine from a salt and ice mixture is about - 22°C. (Walker reports - 21° and Bigelow - 22.4°C.) This temperature is called the cryohydric point. If calcium chloride is used with ice the lowest temperature obtainable is - 55 °C.
Read more: ’

I decided to make an -20 °C temperature reference point. The very first one was very watery and definitely not stable. The second one the same. I was very disappointed.

I came to the conclusion that the -20 °C temperature reference point was a waste of time, and not accurate or stable. That is why I decided to present a paper about this topic.

I was very glad that I did not stop after the first two attempts. I bought a polystyrene box to make my -20 °C temperature reference point. Working with ice and salt is very messy and if you do not clean up properly, it can rust all the metal parts in the laboratory that it comes in contact with.

BE CAREFULL – DO NOT TOUCH THE ICE AND SALT MIXTURE WITH YOUR BARE HANDS – IT IS VERY COLD AND CAN FREEZE YOUR HANDS!

I bought a kilo ice cubes from the shop, and shaved it very fine. I threw it in the polystyrene box and filled it to the top with ice. I then mixed it with a kilogram of fine coarse salt. It was not easy to mix, and was impossible to mix the salt and ice properly.I used two reference resistance thermometers to measure the temperature over a period of time. I placed the two resistance thermometers in two different holes in the polystyrene box, more or less at the same immersion depth. This was a much better -20 °C temperature reference point then the first 2 attempts. It was however not stable and did not last very long.

In Table 1 and Graph 1, one can see that over a period of an hour after the two resistance thermometers was immersed into the -20 °C temperature reference point, there was no stability. Probe 1’s temperature differ over the one hour period with 0.513 °C and probe 2 with 0.748. The two reference thermometers temperatures did not change at the same rate either.

Table 1

Time / Probe 1 / Probe 2 / Difference between Probe 1 and Probe 2
10:43 / -20,85 / -20,757 / -0,093
10:49 / -20,779 / -20,747 / -0,032
10:56 / -20,713 / -20,689 / -0,024
11:03 / -20,644 / -20,652 / 0,008
11:10 / -20,548 / -20,283 / -0,265
11:15 / -20,512 / -20,258 / -0,254
11:20 / -20,467 / -20,223 / -0,244
11:24 / -20,488 / -20,154 / -0,334
11:31 / -20,458 / -20,114 / -0,344
11:38 / -20,397 / -20,066 / -0,331
11:41 / -20,373 / -20,049 / -0,324
11:44 / -20,337 / -20,009 / -0,328

Graph 1

When I open the lid of the polystyrene box I saw that the ice was not melted. The ice salt mixture was pulled away from the resistance thermometers that made thermal contact difficult. I compacted the ice and salt mixture and took some readings again.

Table 2

Time / Probe 1 / Probe 2 / Difference between Probe 1 and Probe 2
11:46 / -19,776 / -18,949 / -0,827
11:50 / -20,111 / -18,817 / -1,294
11:54 / -20,197 / -18,847 / -1,35
11:59 / -20,202 / -18,819 / -1,383
12:04 / -20,182 / -18,799 / -1,383

Graph 2

The temperature became colder again, but as one can see in Table 2 and Graph 2, it was not good results.

I decided to make the -20 °C temperature reference point a bit different this time, because of what I observed in the previous attempt. Also because it was difficult to mix a huge pile of ice and salt.

I threw a thin layer of ice, then covered it with salt. Again a thin layer of ice and covered it again with salt. I did this until I filled the polystyrene box to the top with the layers of ice and salt. Again I placed the two resistance thermometers in two different holes into the -20 °C temperature reference point at more or less the same immersion depth. This was a much better attempt. It was however not stable long enough as I wish it to be. I was however very excited, because I got better results.

In Table 3 and Graph 3 below, one can see that over a period of a half an hour, the difference between the two resistances thermometers were smaller. Probe 1’s temperature differ over the ± half an hour period with 0.073 °C and probe 2 with 0.1. This method of making a -20 °C temperature reference point was much better and the difference between the two resistance thermometers was much smaller, however the -20 °C temperature reference point was still not stable and this was still a challenge.

Table 3

Time / Probe 1 / Probe 2 / Difference between Probe 1 and Probe 2
13:31 / -20,802 / -21,074 / 0,272
13:32 / -20,821 / -21,078 / 0,257
13:34 / -20,833 / -21,081 / 0,248
13:35 / -20,834 / -21,083 / 0,249
13:37 / -20,839 / -21,083 / 0,244
13:39 / -20,842 / -21,083 / 0,241
13:42 / -20,846 / -21,078 / 0,232
13:44 / -20,857 / -21,051 / 0,194
13:46 / -20,862 / -21,041 / 0,179
13:47 / -20,866 / -21,034 / 0,168
13"48 / -20,851 / -21,031 / 0,180
13:49 / -20,871 / -21,029 / 0,158
13:52 / -20,877 / -21,014 / 0,137
13:53 / -20,873 / -21,006 / 0,133
13:55 / -20,871 / -20,999 / 0,128
13:57 / -20,873 / -20,994 / 0,121
13:59 / -20,873 / -20,989 / 0,116
14:01 / -20,874 / -20,981 / 0,107
14:03 / -20,875 / -20,974 / 0,099

Graph 3

I made another -20 °C temperature reference point. I wanted to see if I can make a -20 °C temperature reference point that is stable and also stable over a longer period. Again with layers of ice and salt.

This time I threw a thin layer of ice and covered it with salt. Then a thin layer of ice again. This time I compressed the ice and the salt. Then I threw a thin layer of ice salt and ice again.Again I compressed it. The polystyrene box was filled to the top with thin layers of ice and salt and then compressed.The two resistance thermometers was placed into the same hole at the same depth. I got very good results.

The -20 °C temperature reference point was stable for ± one hour. See results in Table 4 and Graph 4. The difference between the two resistance thermometers over one hour was 0.015 °C. After a few minutes the -20 °C temperature reference point was stable and both probes was drifting at more or less at the same rate. After an half an hour, the resistance thermometer Probe 1, was heating up a bit quicker that Probe 2.

Graph 4

Table 4

Time / Probe 1 (°C) / Probe 2 (°C) / Difference between Probe 1 and Probe 2
11:56 / -20,903 / -20,984 / 0,081
12:01 / -20,971 / -21,061 / 0,090
12:03 / -20,977 / -21,063 / 0,086
12:05 / -20,988 / -21,071 / 0,083
12:07 / -20,994 / -21,074 / 0,080
12:08 / -20,996 / -21,076 / 0,080
12:11 / -20,998 / -21,081 / 0,083
12:13 / -20,999 / -21,078 / 0,079
12:14 / -21,000 / -21,081 / 0,081
12:16 / -21,000 / -21,081 / 0,081
12:17 / -21,000 / -21,083 / 0,083
12:18 / -21,001 / -21,083 / 0,082
12:19 / -21,001 / -21,081 / 0,080
12:20 / -21,002 / -21,083 / 0,081
12:22 / -21,001 / -21,081 / 0,080
12:23 / -21,000 / -21,083 / 0,083
12:25 / -20,999 / -21,081 / 0,082
12:27 / -20,998 / -21,081 / 0,083
12:29 / -20,998 / -21,083 / 0,085
12:31 / -20,997 / -21,083 / 0,086
12:34 / -20,996 / -21,083 / 0,087
12:37 / -20,994 / -21,081 / 0,087
12:40 / -20,993 / -21,081 / 0,088
12:43 / -20,990 / -21,081 / 0,091
12:49 / -20,987 / -21,081 / 0,094
12:52 / -20,987 / -21,081 / 0,094
12:53 / -20,987 / -21,078 / 0,091
13:00 / -20,985 / -21,081 / 0,096
13:03 / -20,985 / -21,081 / 0,096
Difference from start to finish / 0,082 / 0,097 / -0,015
  1. Conclusion

When a -20 °C temperature reference point is used for on-site calibration, the following should be considered:

a)Proper analysis of -20 °C temperature reference point.

b)A good descriptive procedure for making a -20 °C temperature reference point.

c)An uncertainty budget taking all the data from the analysis into consideration.

A -20 °C temperature reference point is not for high accuracy work. It is just for a quick verification. When using a -20 °C temperature reference point one has to take readings at the correct time, otherwise the readings obtained will be when the -20 °C temperature reference point is cooling down, or when it is starting to heat up again.

When a -20 °C temperature reference point is not analysed or a proper procedure is used, one will not be sure if the -20 °C temperature reference point’s results can be trusted, especially using only one reference standard and a unit under test. When using two reference standards, the results can be monitored and one can have confidence in the results obtained from a -20 °C temperature reference point.

I am sure that with more experiments one can make a -20 °C temperature reference point that can be used for an on-site reference point. However one cannot claim small uncertainties.One thing that is very critical is how accurate the -20 °C temperature reference point is made up! The other is the diameter of the reference standard and the UUT, and how quick response each have! The response of the reference standard and UUT is very critical, because it will not reach stability at the same time, and the stability of a -20 °C temperature reference point is for a short period only!

Test and Measurement Conference