In DC and resistance metrology temperature is one of the most important variables. Everything in physical and chemical world affected by temperature change and amplitude. This is one of the factors why monitoring temperature in a very precise and controlled manner is important. This is done at a reference temperature points and at this temperature it is possible to calibrate the sensors used. Like all sensors, the ideal one would does not change in value over time. But in real sensors relationship between output signal and temperature may change in value, with ideally value of change at constant rate/slope. Though to get a good calibration points and uncertainty of the measurement it needs to be compared against a very precise and stable temperature reference. One of such references is triple point of water point at 273.16 K. Some other fixed points are also used in world-wide metrology laboratories:

• Argon (83.8058K)
• Mercury (234.3156K)
• Water (273.16K)
• Gallium (302.9146K)
• Indium (249.7458K)
• Tin (505.078K)
• Zinc (692.677K)
• Aluminum (933.473K)
• Silver (1234.93K)

The water triple point (WTP) is a thermometric fixed point central to the definitions used in the International Temperature Scale of 1990 (ITS-90) and served as the definition of the SI unit of temperature, the kelvin, from 1954 to 2019. Today temperature unit kelvin is defined by the physical Boltzmann constant which is in effect since 2019 redefinition.

TPW Cell consists of hermetic glass vessel with carefully calibrated water composition with vacuum area for vapours. Inner stem is provided to accept SPRT and other type temperature sensors for calibration. Once cell is properly conditioned and prepared it will maintain very stable and uniform temperature point at 273.16 K or +0.01 °C over long period of time. A quick weekend visit to a friendly lab with experimental test for realization of +0.01 °C point with TPW standard. Isotech-Jarrett triple point water cell standard used in the experiment, Model A-13-1435. This is NBS/NIST style type A full size cell.

These cells are very fragile and require utmost care to safely operate and transport. They are shipped in triple-packed crates and hardcases to reduce risk of shock or vibration damages.

Cell is maintained with regular water ice bucket or maintenance ovens to prolong the useful fixed point measurement time. But to precool and start the TPW one require liquid nitrogen or dry ice, to form the ice mantle around the center stem. Key point is to have water inside at all three matter states – frozen ice, liquid water and water vapour. Actual composition of the water inside the cell is carefully measured and formulated for best uncertainty and stability. Composition used is called “Vienna Standard Mean Ocean Water” or VSMOW. Unlike the name implies this is not just some ocean water scooped up in Vienna. But a finely controlled mix of some isotopes which the ratios of are listed below. As result “Vienna” is part of the name because the International Atomic Energy Agency, based in Vienna, promulgated the standard formulation for VSMOW.

• H₂ : H₁ = 155.76 ± 0.1 ppm (ratio of 1/6420 parts)
• H₃ : H₁ = 1.85 ± 0.36 – 10-11 ppm (ratio of 1/5496 parts, largely ignored)
• O₁₈ : O₁₆ = 2005.20 ± 0.43 ppm (ratio of 1 part per approximately 498.7 parts)
• O₁₇ : O₁₆ = 379.9 ± 1.6 ppm (ratio of 1 part per approximately 2632 parts)

However VSMOW quantity being very limited, with only a few liters available worldwide. Locally cells are made often with pure distilled water but this makes the triple point of water vary between regions. These is then referenced to a reference VSMOW triple point water cell.

Isopropyl alcohol is used as thermal transfer medium for the thermometers and DUT sensors inside TPW stem.

Next few photos shows beautifully growing ice mantle with dry ice sublimating in the alcohol mixture.

Sometimes ice mantle cracks due to heat from handling the cell. That’s normal and nothing to worry about. As ice anneals over hours later cracks will disappear.

CO₂ dry ice has sublimation temperature at 194.7 K at the normal pressure and can be used to make the mantle in the TPW cell.

Image: Comparison carbon dioxide water phase diagram, CC BY-SA 3.0 referenced from Wikimedia .

For fun I’ve put my recent digital sensor with TI TMP119 into a pile of crushed dry ice with alcohol.

Both TMP117 and TMP119 continued to operate with such cold temperature and reported -83.x °C which is 5 degrees off the expected value. That is not a reason to panic, since such low temperature is way outside of operating temperature range for digital sensor like Texas Instruments TMP11x series. According to TMP119 datasheet it’s rated for 0.3 °C error for lowest temperature -55 °C. We are way below that here and I’m impressed sensor still working and providing a measurement.

Another interesting fact to notice – both TMP117 and TMP119 have about same error with dryice reading, so it’s not a random fluke at one particular sensor chip.

Digital TMP117 and TMP119 sensor was also submerged into TPW to check how far off the reading would be. After few minutes it stabilized at about +0.25 °C for TMP117 and about +0.17 °C for TMP119. Perhaps this small error from expected 0.01 °C was due to self-heating introduced by active circuitry inside TMP117/TMP119 from +3.3V power supply.

Typical sensor that is calibrated with help of fixed point standard such as Isotech/Jarrett is high performance SPRT. We got one of these as well, such as Fluke 5699 that required new terminals:

Platinum thermometers like these are very sensitive and fragile and require great care as well.

Hopefully this one still works and we will continue to explore thermometry metrology in future articles.