Copper Thermometer

7/30/2019 Copper Thermometer

1/29

TEMPMEKO 2010Slide 1

Welcome

A NEW THERMOMETER FOR THE

COPPER POINT

John Tavener : President & Head of Laboratories

Isothermal Technology Ltd, UK(Isotech)


2/29


Demand

There is a significant demand for CopperCells and Apparatus for thermocouplecalibration

However the uncertainties associated withmeasuring the Copper Point are larger thanthose at the Silver Point for exampleThe major source of measurementuncertainty is the sensor

The preferred sensor at the millikelvin levelwould be a quarter ohm PlatinumResistance Thermometer.


3/29


A New Design

1 The Sheath Recrystalised alumina waschosen for its rigidity. Because alumina isslightly porous the inside of the senor was

kept at a slight positive pressure comparedto the outside of the sheath using oxygenrich gas

The theory being that any leakage would beoutward. Alumina has a tempco of about9ppm/C and is thermally fragile and so thesheath is designed to be easily replaceable


4/29


A New Design

2 The Mandrel on which thePlatinum sensor is wound is made of

synthetic sapphire because firstlysapphire is a more nobel insulatorthan either quartz or alumina, andsecondly its expansion and

contraction closely matches that ofthe Platinum


5/29


A New Design

3 The Bias The thermometer isbiassed to +9vDC. The reasons are

two foldFirstly biasing the thermometerincreases its insulating resistance

Secondly a positive voltage acts asan electro magnet repelling themetalic ions.


6/29


A New Design

4 Physically The thermometer is500mm long, 7.8mm diameter with a

100mm long handle and 2 metres oflead


7/29TEMPMEKO 2010Slide 7

The Rest of the Apparatus

A microK 100 was used in conjunction witha Temperature Controlled 1 TinsleyResistor

The Copper Point was realised in anIsotower described elsewhere

It is however significant to mention that theCopper Cell is housed in a metalic

enclosure, including the re-entrant tube inwhich the thermometer was positionedduring measurements.



Testing

Testing was simple, the thermometer wasused repeatedly for the melts and freezesof the Copper Cell in the Isotower

Between each sequence Rtpw and Wgawere measured

Part way through the testing the Copper

Cell was heated to 1100C under vacuumto remove oxygen

since 1ppm of oxygen in Copper can depress itstransition point by 5mK.



Immersion Test

Because, uniquely theImmersionCompensator's

temperature can beadjusted the gradientover the bottom100mm is 2 or 3mK.(less than 1mK overthe bottom 40mm)



History of Measurements

Rwtp & Wga Rcu

.236, 0412,2( ) M 1.092,387( )

116

Hours

F 1.092,362( )

M 1.092,350( )

F 1.092,308( )




.236, 034,0( ) 1.118120 M 1.092,310( ) 47

Hours

F 1.092,302( )

M 1.092,300( )

F 1.092,299( )




Rwtp & Wga Rcu

.236, 029,7( ) 1.118175 M 1.092,298( ) 53

Hour

s

F 1.092,294( )

.236, 040,0( ) 1.118129 M 1.092,293( ) 68

HoursF 1.092,287( )




Rwtp & Wga RcuPrevious

Value

1.092,287.236, 040,8( ) 1.118124 M* 1.092,287( )

25

HoursF* 1.092,283( )

.236, 041,0( ) 1.118133

*After Copper Cell deoxidised

All measurements use 10(mA)

The Copper Point was realised in an Isotower described elsewhere

The resistances marked M and F are the liquidus and solidus resistances




Rwtp & Wga Rcu

5 Cycles 7 x M & F -27(mK) 310 Hours

.All measurements use 10mA

The Copper Point was realised in an Isotower described elsewhere

The resistances marked M and F are the liquidus and solidusresistances



Results

The graphs chart Rwtp and Wga and showsthat over the test period of 5 weeks thesevalues remained stable having no

significant drift or contaminationThis is amazing considering thethermometer was being repeatedly put intoand removed from a metalic re-entranttube at 1084.62C

The outside of the thermometer's sheathwas blackened by the end of the tests



Results - Rwtp



Results Wga



Results

The following graph charts Rcu during thetestsAfter the initial sequence where the

thermometer was stabilising; over a periodof 300 hours at or above 1085C the totaldownward shift is 30mK or around0.1mK/hourInspection of this graph shows that the

majority (25mK) of the drift occured whilethe thermometer was above the liquidustemperature


19/29


Results

In fact it occured during the cold rod processwhere the thermometer exits the cell at1084C and is quenched in air for a minute

while the cold rod nucleates the Copper alongits re-entrant tube the thermometerreaches 400C before it is replaced in theCopper Cell

What is still a puzzle is why, when Rcu is

slowly decreasing Rtpw and Wga remainstable

However at 0.1mK/hour the numbers arerelatively small compared to alternative

sensors


20/29


Results - Changes in Rcu


21/29


Thermometer Drift over aFreeze/Melt

Date F to M T(C) F/M (mK)

3-8 Jan 2010 5 days 1085 +28-9 Jan 2010 1 day 5 -2

9-29 Jan 2010 20 days 1085 -1

1-2 Feb 2010 3 days 1085 -1

3-6 Feb 2010 4 days 1085 0


22/29


Thermometer Drift During a Meltor Freeze

The change in Rcu has been identified as mainlyhappening during the cold rod process.To discover the drift during a melt or freeze it isnecessary to consider the changes in Rcu during afreeze followed by a melt (FM). The previous tableshows 5 such cycles that occured during testing.In 4 of the 5 cycles the thermometer was cooled over3 hours to 480C and then Rwtp and Wga weremeasuredTo summarise the table, independent of the number

of days between freeze and melt or whether thethermometer's temperature excursion is 5C or1085C the solidus/liquidus temperature change wasbetween 0 and 2mK.So the thermometer drift is between 0 and 1 mKfor a melt or freeze


23/29


A Second Copper Cell

A new high purity copper point cellwas manufactured and the new

thermometer was used to measurethe melt and freezes


24/29


Results in Second Cell

The table shows the resistances at liquidusand freeze initiation

Intial Rtpw 0.236,041,0 Wga 1.118133

Slopes RcuPreviously

1.092,283( )

5 Hour Melt 80% 3(mk) 1.092,285()

4 Hour Freeze 50% Flat+ 1.092,280()

Final Rtpw 0.236,041,7 Wga 1.1181343


25/29



This table shows a further melt and freezemade with a second synthetic sapphiremandreled thermometer which had been

stabilised for 100 hours prior to use

Feb 28 2hr F, 50% 0mK 1.099,557

Feb 28 2.5 hr M, 80% 3mK 1.099,553

Mean Melt =3mK Mean Freeze 0mK Drift 4mK


26/29



These results confirm the melt and freezemeasurements of the first thermometer and

this newer thermometer also exhibits averagedrift of less than 0.1mK / hour after 100hours of stabilisation


27/29

The graphs clearly show the unique stabilityof the thermometer during the freezing andmelting of the very pure cell

Copper Isotower Freeze Plateau25th Feb 2010

1.092238

1.09224

1.092242

1.092244

1.092246

1.0922481.09225

1.092252

1.092254

1.092256

1.092258

1.09226

1.092262

1.092264

1.092266

1.092268

1.09227

1.092272

1.092274

1.092276

1.092278

1.09228

1.092282

1.092284

160700 164300 167900 171500 175100

Time/Secs

Res/Ohms

108642/001 Res/Ohms

1 division = 2mK Four Hour Freeze

50% Freeze 2mK


28/29

The graphs clearly show the unique stability of thethermometer during the freezing and melting of thevery pure cell

Copper Isotower Melt Plateau24th Feb 2010

1.09227

1.092272

1.092274

1.092276

1.092278

1.09228

1.092282

1.092284

1.092286

1.092288

1.09229

1.092292

1.092294

1.092296

1.092298

1.0923

85700 89300 92900 96500 100100 103700 107300

Time/Secs

Res/Ohms

108642/001 Res/Ohms1 division = 2mK Five Hour Melt

80% Melt in 3mK


29/29

Conclusion

The new thermometer works up to1085C with little drift (0.1mK/hour)

and no contamination(stable Rwtp and Wga)

It is therefore suitable for theintercomparison and characterisation

of Copper Cells at uncertaintiessmaller than previously achieved

Documents

Copper Thermometer