Talk:International Temperature Scale of 1990

ITS–90 is not a “scale”
“International Temperature Scale of 1990” has the word “scale” in its name but this is a misnomer and can be misleading to many people. Only scales like kelvin, Celsius, Fahrenheit, and Rankine are scales. There is no such thing as “99.987 °ITS.”

ITS–90 is an equipment calibration standard for making accurate measurements of temperatures that can then be expressed as values on various temperature scales such as Celsius, kelvin, or Fahrenheit.

The lasting effect of making a temperature measurement using instruments calibrated per ITS–90 is the measured value will likely be very slightly different from those produced using a different calibration standard—regardless of which scale is used to express the value of the measurement. For instance, the boiling point of water can be said to occur at 211.9710 °F when calibrated to the two-point definition of absolute thermodynamic temperature. It is just as valid to say the boiling point of water is 211.953 °F when calibrated per ITS–90. Clearly, the scale on which both temperatures are being expressed is Fahrenheit even though the later was per a kelvin-based calibration standard that has the word “scale” in its title.

The important point is that when documenting high-precision temperature measurements (to a resolution of 10 mK or better), it becomes increasingly important to specify the calibration standard employed. And further, the calibration standard employed places no restrictions as to which scale the temperature may be expressed.

Temperature formating conventions
The name “kelvin” is not capitalized unless it begins a sentence. With the exception of the kelvin, all the scale symbols are immediately preceded by a ° (degree) symbol. Note also that a space is inserted between the value and the degree symbol (or in the case of the kelvin, before the “K”. Only angular measurements such as 180° have the degree symbol placed immediately after the value.  Also, the scale's symbol is the always capitalized; thusly, we have…

The freezing point of water (expressed as kelvins) is 273.15 K The freezing point of water (expressed as degrees Celsius) is 0 °C The freezing point of water (expressed as degrees Fahrenheit) is 32 °F and The freezing point of water (expressed as degrees Rankine) is 459.67 °R

Remember that it is important to capitalize the kelvin symbol (K) because the lowercase version (k) is the SI prefix for 1 x 103

Various reasons can be found for why the name “kelvin” is lowercase and its symbol, “K,” is not preceded with a ° (degree) symbol. For instance, some sources say that kelvin has no degree symbol because it has the distinction of being an absolute scale. However, the definition of “degree” is simply “a unit division or step” so it doesn’t matter where a temperature scale’s null point is. Furthermore, the symbol for Rankine (also an absolute scale) is properly preceded by the degree symbol. There are two simple reasons for why the unit name kelvin and its symbol are written as they are: Notwithstanding the naming convention historically used for the other temperature scales, SI unit names are never capitalized. Also, the 13th General Conference on Weights and Measures (CGPM) in 1967 declared that the kelvin symbol (K) would no longer be preceded by a degree symbol because kelvin would no longer refer to the entire scale; it would thereafter be recognized as referring to an individual unit comprising the scale. Thus, the numeric value preceding the unit symbol would simply specify “how many kelvins there are” (just like 15 km or 4 µg). This is why the kelvin is the only unit measure of temperature that may be written in its plural form (e.g. “…is 12 kelvins greater than…”) in the same fashion as with all the other SI units such as mass (e.g. “…is 12 grams greater than…”). Thus, since kelvin has not referred to the entire scale since 1967, it is not appropriate to refer to a degree-related position on it, nor is it proper to precede the “K” with a ° (degree) symbol.

For more information on how the full word “kelvin” is not capitalized, see this article by the NIST as well as Wikipedia's kelvin article.

For more information on conventions used in technical writing, see the informative SI Unit rules and style conventions by the NIST.

Still more information on SI (International System of Units) conventions from the NIST can be found here.

Issues regarding freezing v.s. melting point
ITS-90 differentiates between “freezing” and “melting". See Table 1 in this link for more information.

ANS: my guess is that the original author made the distinction due the nature of gallium, more specifically its tendency to supper cool. This property would make freezing point unreliable, hence the use of he triple point by NIST. AlphaTeller (talk) 14:14, 15 April 2009 (UTC)

Why do you see the mixed use of the kelvin and Celsius scales?
In science and engineering, the Celsius and kelvin scales are often used simultaneously in the same article. This is because Celsius is used to express a specific temperature (e.g. “The melting point of gallium is 29.7646 °C.”)

The kelvin scale, on the other hand, can be used two ways. It is uniquely used to express differences between temperatures and their uncertainties (e.g. “The error is 10 mK,” and “Agar exhibited a melting point hysteresis of 25 K.”) And of course, the kelvin may also be used to express specific temperatures along its absolute scale (e.g. “The triple point of hydrogen is 13.8033 K.”)

Celsius is not used to express differences between temperatures or their uncertainties. In part, this prohibition eliminates ambiguity. In a chart, the value “25 °C” would be confusing if it was actually intended to represent a difference between two temperatures.

@Greg L : I agree 99.9% ( and don't correct that to 99.9 %, that's another story), however in the main article Kelvin is used extensively as the name of the temperature scale, and kelvin as the units of this scale. With which I concur. Your thoughts? — Preceding unsigned comment added by 2001:4200:7000:1:0:0:0:10 (talk) 13:30, 12 February 2016 (UTC)

And another, more profound, thought

> Celsius is not used to express differences between temperatures or their uncertainties 

I disagree strongly here: IMHO the Celsius scale should ONLY be used to describe temperature differences (one temperature from another, or if not a "conventional temperature difference", then from 273.15 K) - reason being that temperature in almost all thermodynamic applications has to be the absolute temperature (measure of molecular kinetic energy), it appears wrong to me to use a scale that differs from the Kelvin scale by more than a scale factor. So temperature should ALWAYS be expressed in K, while temperature difference could be in K or in °C, whichever suits. (E.g. all heat conduction and convection is fine in °C, but try radiation in °C and you will burn your fingers in the end in coded equations that can go due South).

I know we won't settle this in finite time, there are hundreds of years of legacy behind the Celsius scale. Does any sane person still use the Rankine scale &#12471; — Preceding unsigned comment added by 2001:4200:7000:1:0:0:0:10 (talk) 13:49, 12 February 2016 (UTC)

Typography hints
In any document that has variable-width pages and where word-wrap can occur at almost any point on a line, try using a hard-space (option-space on a Mac) between a temperature value and its degree or scale symbol. Thusly, you will obtain: “A temperature of 99.987 °C at the end of a line instead of 99.987 °C, which is fragmented and is harder to read.”

Similarly, try to use an en-dash (option-hyphen on a Mac) rather than a simple hyphen when writing “ITS–90” to prevent the same thing. Thusly, you will obtain: “A calibration standard per ITS–90, rather than a less controlled hyphenated ITS- 90 which is fragmented and is harder to read.”

“Typographers quotes” should be used rather than simple "typewriter quotes."

For further information on writing style and typography issues, see Wikipedia’s article Manual of style.

"±0" notation

 * "It specifies fourteen calibration points ranging from 0.65±0 K to 1357.77±0 K (-272.50±0 °C to 1084.62±0 °C) and is subdivided into multiple temperature ranges which overlap in some instances."

Please help the reader understand the meaning of the odd "±0" notation. Does it mean that the error bar range is plus/minus one whole unit? (As opposed to maybe "±0.0000" that would specify number of digits of error range?) Or does it mean that this temp is infinitely exact by definition?-73.61.15.243 (talk) 17:40, 12 November 2017 (UTC)


 * It means that the temperature is exact by definition of the ITS-90.--Eheran (talk) 09:44, 8 December 2020 (UTC)

Article cleanup done, minor formatting issues
Hello everyone!

I've corrected a few grammatical errors in the article (mainly ITS-90 to the ITS-90), added a few extra references, removed the weird ±0 notation, and I've also written a new paragraph on the effects of impurities and isotopic composition on the defining fixed points.

I've also added references from the BIPM to the table in Standard interpolating thermometers and their ranges, and added a citation needed tag to the triple point of water, since I couldn't find anything on that in the BIPM's Guide to the Realisation of the ITS-90. Now that I've added sources for the table, can I remove the template?

I am planning to add a specific reference to the pressure correction statement soon. However, I've also accidentally cited the same thing three times - can someone fix that up please?

Cheers, MeasureWell (talk) 03:09, 24 June 2021 (UTC)