Units for Temperature Measurement

[Crafter_Man]

Before 2019:

Triple point of VSMOW = 0.01 °C ± 0 °C (defined)
Boltzmann constant = k = 1.38064903(51) X 10⁻²³ JK⁻¹ ± 3.7 X 10⁻⁷JK⁻¹ (measured)

After 2019:

Triple point of VSMOW = 0.01 °C ± ??? °C (measured)
Boltzmann constant = k = 1.380649 X 10⁻²³ JK⁻¹ ± 0 JK⁻¹ (defined)

So as can be seen, the triple point of VSMOW and the Boltzmann constant recently swapped places in terms of what’s measured and what’s defined.

So what’s the uncertainty for the triple point of VSMOW now? Last I checked, they’re… not sure.

In 2021 I emailed the group leader for NIST’s temperature metrology lab, and asked her if this will affect how they calibrate SPRTs. Her response was interesting. She said the redefinition only applies to the Thermodynamic Temperature Scale, and not to the International Temperature Scale (ITS), the latest of which is the 1990 version. She went on to say,

The vast majority of SPRT and ITS-90 users (including the NIST SPRT Calibration Laboratory) will go on with their lives as if the redefinition never occurred. Since the redefinition only applies to thermodynamic temperature, it doesn’t affect ITS-90 island.

NIST and other labs are developing different capabilities for deployable thermodynamic temperature measurements, but all of those efforts are still years or decades away from widespread commercialization and adoption, and none of these capabilities are currently able to match the ITS-90 in terms of precision and range. For better or for worse, the ITS-90 is here to stay for quite a while longer.

To be honest, the separation between the Thermodynamic Temperature Scale and the ITS is something I’ve never quite been able to wrap my head around. Based on my (limited) understanding, it would seem any modification of the former can and should have a direct and instant affect the latter. At any rate, and reading between the lines, I think what she’s saying is that, for now, users of the ITS will go on assuming the triple point of VSMOW is exactly 0.01 °C, but an uncertainty will probably be assigned to it within the next few decades.

[Chronos]

Surely, the new relative error in the triple point of water would be equal to the old relative error in the Boltzmann constant, since it’s the same experiment either way.

[abcdefghij]

Maybe not. Avogadro’s number is now defined, too, which would impact the uncertainty in the density. Which could have knock-on effects of the uncertainty of the measurement of triple-point.

[Sage_Rat]

@puzzlegal

My European-made freezer defaulted to -4F, which seems to be a fine temperature to store food at.

Do you mean F or C? Of course -4F would be fine since that’s even colder than 0F.

If you meant to say C, it looks like that’s not legal in the EU, it would need to be at least -15C?

https://www.somerset-chamber.co.uk/news/member-news/is-18c-the-best-temperature-to-freeze-or-is-it-just-a-costly-habit/#:~:text=Under%20EU%20and%20UK%20food,18°C%20or%20lower.

[puzzlegal]

I meant -4°F. I was just pointing out that my fridge didn’t default to 0°F.

[DrDeth]

@Ynnad

The Celsius scale arose from originally assigning 0 for the freezing point of water and 100 as the boiling point of water.

Nope. Celsius himself originally made it just the opposite- O boiling and 100 freezing (he changed it fairly quickly). Cecil had a great article on this. Celsius also went to his grave insisting that his system be named Centigrade, not after him.

[quote=“Sage_Rat, post:16, topic:1020711”]
Fahrenheit was determined with an eye towards measuring weather. I’m not sure but I’d expect that most organic chemistry happens within or near the 0-100 range of temperature, F. (E.g. freezers are calibrated to 0F, even in Metric-land, because bacteria continue to multiply at 0C. This activity bottoms out with 0 on the Fahrenheit scale. 100F is, effectively, the midpoint between a healthy human body temperature and a dangerous fever.
[/quote] The early history of F is interesting- basically Fahrenheit didnt want weather commonly below zero, and perhaps used his own body temp as 100.

@puzzlegal

My understanding is that Fahrenheit chose the coldest temperature he could easily create (by mixing ice with salt) to be zero, and his own body temperature that day as 100.

It is obviously an arbitrary, human-centric scale, but it works very well for “temperatures people routinely encounter”. Certainly, anything about 100F is too damn hot and anything below 0F is bitterly cold.

That is correct, but Cecil’s column had some tidbits about the history.

F is better for human life and weather. C or K is better for science.

.

[Zoobi]

Celsius is just fine for human life and weather.

[Earl_Snake-Hips_Tucker]

And it was Centigrade until the 1948 when some yoyo dug up the name Anders Celsius and ascribed it to a temperature scale that Celsius neither invented nor used.