Re: SI and nothing else.

[David Bowman <dbowman@tiger.gtc.georgetown.ky.us>]

John Ertel wrote:
> ....
> The SI unit of temperature seems to still be defined the same way as it
> was in my youth.  The SI unit is the Kelvin and it is defined to be
> equivalent to the Celsius degree.  The Celsius (or Centigrade) degree is
> defined to be 1/100th of the temperature difference between the freezing
> point (defined to be zero) and the boiling point (defined to be 100) of
> pure water at 1 atm.

Not quite. This is somewhat backwards.  The SI unit of the kelvin is
defined as a unit of thermodynamic temperature.  This scale of temperature
set by the requirement that the triple point of pure water (with average 
terrestrial isotopic ratios) is defined to have the value of *exactly*
273.16 K.  This definition of the kelvin sets the experimentally measured
value of Boltzmann's constant to 1.3806258(12) x 10^(-23) J/K.   The Celsius
scale is a derived temperature scale which is *defined* such that
t(deg. C) = T(K) - 273.15 exactly.  The old centigrade scale was defined such
that the freezing point of pure water occurred at 0 (deg. C) (under 1 atm of
pressure) and the vapor presure of pure water reached exactly 1 atm of
pressure at exactly 100 (deg. C).

(See NIST spec. pub. #811 'Guide for the Use of the International System of
Units (SI)' and the pub. 'The 1986 CODATA Recommended Values of the
Fundamental Physical Constants' both available from the NIST web site.  Try:
http://physics.nist.gov/Pubs/SP811/sp811.html
http://physcis.nist.gov/Pubs/SP330/sp330.html
http://physics.nist.gov/PhysRefData/codata86/codata86.html
http://physics.nist.gov/Divisions/Div840/FedReg/metric.html)

>                      The thing that varies (but not much over the last 20
> years) is the "position" of the zero-point of the Kelvin system (which
> currently seems to be -273.1475C) relative to the Celsius system.

The things that vary now (as their measurements get better defined) are the
freezing points and boiling points of pure water.  The freezing point of
pure water seems to be about +273.1475 K = - 0.0025 (deg. C).

>                                                                   I have
> looked in over twenty current documents and I find no exceptions other
> than updates on the zero-point of the Celsius system relative to zero in
> the Kelvin system (or as it is most commonly quoted, as the zero-point of
> the Kelvin system relative to zero in the Celsius system).

These are current documents you say?  The zero of the Celsius system is
defined as exactly 273.15 K hotter than absolute zero.  This has nothing
at present to do (directly*) with water's freezing point under 1 atm of
pressure. 

* I said 'directly' above because the triple point of water which is defined
to have the value of 273.16 K is indirectly tied to the freezing point of
water by fixed relative chemical potential functions for the 3 phases of
water which are themselves fixed by the masses and interparticle potentials
for water molecules.

> ...                                                      .  The small or
> thermodynamic calorie still seems to be defined as the heat required to
> raise the temperature of 1 gram of pure water (at 1 atm.) 1 degree Celsius
> when averaged over the range from the freezing point to the boiling point.

The thermodynamic calorie is defined as 4.184 J (exactly).  The
International Table calorie is defined as 4.1868 J (exactly).  The mean
calorie (you define above) has the experimental value of about 4.19002 J.
The (15 deg. C) calorie has the experimental value of about 4.18580 J.  The
(20 deg. C) calorie has the experimental value of about 4.18190 J.  There are
also 'nutritional' (kg) versions of the thermodynamic, International table,
and mean calories, each being 10^3 times larger than the corresponding gram
calorie values.

David Bowman
dbowman@gtc.georgetown.ky.us