Making a Good Ice Bath

[Hawaii596]

Okay, I'll take a break from my less serious postings....

I recently started my new position back home in Texas as lead metrologist and qualilty manager of this lab.  It's pretty well equipped, but has a few areas needing work.  I'm trying to get a triple point (hope to have one within the year).

Meanwhile, I'm sending my best probe out for real cal with freeze points, etc.  And then want to use it for my variety of Burns Engineering B-12001's and similar.  Problem is, getting RTPW (RO) using an ice bath with adequate uncertainty.  I found a NIST tech note and a New Zealand national lab technical guide that both recommend about the same techniques (which I haven't yeet experimented with) to achieve about +/-1 mK and potentially as good as +/-0.5 mK uncertainty.  I was pretty amazed.  If there is any interest, I can either post the info - or maybe all of you out there are already way ahead of me on this.

Have a good day, everyone.

[Bryan]

As I recall from a class many years ago we used dry ice and water, distilled I suppose & stirred.  I imagine there was an ideal place in the vessel at which to measure but it was long ago.  I belive it was referred to as "triple point", the water supposedly exisited as liquid, solid & gas at some location or darn close to it.
Though I haven't done it since I'd be interested just for the sake of general knowledge.

[metrologygeek]

Another good place to start is:
ASTM E563-02e1 Standard Practice for Preparation and Use of an Ice-Point Bath as a Reference Temperature

In my temperature lab, we have a still for producing our own distilled water plumbed into a commercial ice shaver.  Ice is packed into a dewar and backfilled with water.  We can normally achieve < 0. 5mK uncertainty.  We use the zero point as a reference for T/C calibrations.  For resistance thermometry, we use triple point cells.  A triple point is a sealed glass cell with pure water under a vacuum.  Do not try to make a triple point with water and dry ice.  You'll just end up with dry ice and wet ice in a big blob.

[flew-da-coup]

I will just stick with a triple point cell.

[metrologygeek]

Can't go wrong with the triple point.   We usually have two made up at any given time.   The ice point is really only good for t/c work. 

[Hawaii596]

We have a large temperature lab for which I am trying to get a new set of fixed point furnaces.  Its to calibrate high temp noble metal thermocouples, and you need fixed points for the higher temps.  I won't get into the details, as I probably shouldn't here.  But in my proposal package, I'm asking for a triple point of H2O (which I am hoping to get).

And I couldn't disagree at all that a triple point is the way to go.  But for us who don't have one (YET!!), making a good ice bath is the next best thing.  According a a NIST tech note and a New Zealand national lab tech instruction, you can get to within about +/-0.002 Degrees C. with a properly made ice bath (the devil is in the details).  When I use a tall, thin dewar (a good one with glass interior and matching rubber stopper), distilled water that is finely crushed with the right consistency, I tested with my Rosemount PRT and Fluke 8508A and got within about +/-0.001 or so.  For lower end applications, that is actually pretty satisfactory.  Calibrating PRTs, it is a little on the edge, and some ISO17025 auditors don't like it.