Progress towards Implementing
the new kelvin
Graham Machin, BSc, DPhil, DSc, CPhys, FInstP, FInstMC
Head of Temperature Standards, NPL Fellow
Cryogenic Cluster Day, CCD7: 20 September 2016
Acknowledge support from European Metrology Research Programme (EMRP) and
European Metrology Research Programme for Innovation and Research (EMPIR)
Introduction




A few words about NPL
The SI and the proposed 2018 redefinition
The mise en pratique for the definition of the kelvin 2019
Towards implementing the new kelvin (InK)
Achievements of the InK 1 project
Prospects for the InK 2 project
 Summary
A few words about NPL
 Founded in 1900 NPL is the UK’s
national standards laboratory for
physical measurement
 A world-leading National
Measurement Institute (one of the
top 3 with NIST, USA and PTB,
Germany)
 More than 800 specialists in
measurement research
 Located in South West London,
Teddington
The SI and the proposed 2018
redefinition
http://www.bipm.org/en/si/
The “old” kelvin
 “The kelvin, unit of thermodynamic temperature, is
the fraction 1/273.16 of the thermodynamic
temperature of the triple point of water”
The new kelvin – from late 2018
 “The kelvin, symbol K, is the SI unit of thermodynamic
temperature; its magnitude is set by fixing the numerical
value of the Boltzmann constant to be equal to exactly
1.380 65X×10−23 when it is expressed in the unit
s−2·m2·kg K−1, which is equal to J·K−1”
The mise en pratique for the
definition of the kelvin
(MeP-K-19)
The MeP-K
 CIPM recommendation 1 CI-2005 “approve in
principle the preparation of new definitions and
mises en pratiques of the kilogram, the ampere
and the kelvin…..”
[Details: http://www.bipm.fr/utils/en/pdf/MeP_K.pdf]
 A mise en pratique is a document that guides the
user from the SI unit definition i.e. (for
temperature) one based on k to a practical,
laboratory realization of the unit
The MeP-K-19
Section 1: The definition of the kelvin
Section 2: Current agreed differences between T and the
defined scales
Section 3: Primary thermometry for measuring
thermodynamic temperature, T
Absolute primary thermometry – no fixed points
Relative primary thermometry – fixed points with explicit T
The MeP-K-19
Section 4: Formal approximations to T: the International
Temperature Scale of 1990 (ITS-90) and the Provisional
Low Temperature Scale from 0.9 mK to 1 K (PLTS-2000)
Methods explicitly defined by an official Temperature Scale
Section 5: Approximation methods that are neither primary
nor defined on a temperature scale, yet capable of
exceptionally low uncertainties or increased reliability.
More details: Fellmuth, B., Fischer, J., Machin, G., Picard, S., Steur, P., Tamura,
O., White, R., Yoon H., “The kelvin redefinition and its mise en pratique”,
Phil. Trans R. Soc. A. 374: 20150037 (2016)
http://dx.doi.org/10.1098/rsta.2015.0037
The MeP-K-19
Section 1: The definition of the kelvin
Section 2: Current agreed differences between T and the
defined scales
Section 3: Primary thermometry for measuring
thermodynamic temperature, T
Absolute primary thermometry – no fixed points
Relative primary thermometry – fixed points with explicit T
These two sections are the
focus of the InK projects
Implementing the new
kelvin (InK) projects
The main aims of the InK projects
 To prepare the international thermometry community for the
kelvin redefinition and ensure a smooth transition to the
new arrangements through the MeP-K-19
 Achieved through:
 Develop primary thermometry methods some of which
could supplant the defined scales (i.e. ITS-90 and PLTS2000) at high (>1300 K) and low (<1 K) temperatures
 Determine new values of T – T90 with the world’s lowest
uncertainties (≤ 1mK) between approximately 1K to 1000 K
 Determine new values for T – T2000 to address the
discrepancy in the background data of the PLTS-2000
Most major institutes in the world are
contributing to the InK activity
V N IIO FI
The InK projects
 Two projects together cover the temperature range
~1mK to 3000 K
 InK 1 Sep 2012 – Oct 2015
Part funded by the European Metrology Research
Programme (EMRP)
 InK 2 Jun 2016 – May 2019
Part funded by the European Metrology Programme
for Innovation and Research (EMPIR)
The InK 1 project had four distinct
research areas
 Assignment of thermodynamic temperature to high
temperature fixed points (HTFPs) >1300 K
 Realisation and dissemination of thermodynamic
temperatures at high temperature >1300 K
 Determination of T – T90 with ultra low uncertainties
 Primary thermometry for low temperatures
…. and novel high temperature fixed points (HTFPs)
HTFP “family”
More details: Machin, G., “Twelve years of high temperature fixed point research: a
review”, AIP Conf. Proc. 1552, 305 (2013); doi: 10.1063/1.4821383
Assignment of thermodynamic
temperature to HTFPs >1300 K
NPL, CNAM, CSIC/CEM, NMIA, NIM, PTB, VNIIOFI [KRISS, NIST, NRC, UC,
NMIJ]
 Determine, for the first time, definitive thermodynamic
temperatures to a set of HTFPs
• Re-C, Pt-C, Co-C [and re-determine Cu]
 Target uncertainty of temperature assignment 0.5 K (k = 2)
at Re-C, scaling to < 0.05 K (k = 2) at the Cu point.
First definitive HTFP
thermodynamic temperatures
Woolliams, E., et al., “Thermodynamic temperature assignment to the point of
inflection of the melting curve of high temperature fixed points”, Phil. Trans R. Soc. A.
374: 20150044 (2016) http://dx.doi.org/10.1098/rsta.2015.0044
Realisation and dissemination of
thermodynamic temperatures at
high temperature
LNE-CNAM, CEM, MIKES, NPL, PTB, TUBITAK-UME
 Assessment of two different MeP-K approved methods for
disseminating thermodynamic temperature above 1300 K
and provide recommendations to CCT on the most
effective method
• Interpolation equation combined with a priori calibrated
HTFPs – MeP-K relative primary thermometry
• Absolute radiometry (via calibrated radiometers),
directly traceable to the electrical watt, the metre and
the second – MeP-K absolute primary thermometry
Recommendations to CCT
WG-NCTh concerning T
dissemination >1300 K
 HTFPs can be used to disseminate ITS-90 scale or T
from NMIs to users with uncertainties comparable with
current approaches.
 Filter radiometers and radiation thermometers, directly
traceable to the radiant Watt, can also be used directly
to disseminate thermodynamic temperature to users
with uncertainties comparable to current methods.
Unknown radiometer drift remains an issue and it is
recommended that if this dissemination approach is adopted a
HTFP be used to periodically assess radiometer stability
More details to be found in

Machin, G., Bloembergen, P, Anhalt, K., Hartmann, J., Sadli, M.,
Saunders, P., Woolliams, E., Yamada, Y. & Yoon H., “Practical
implementation of the mise-en-pratique for the definition of the kelvin
above the silver point”, Int. J. Thermophys., 31, p. 1779-1788, 2010

Sadli, M., Machin, G., Anhalt, K., Bourson, F., Briaudeau, S., del Campo,
D., Diril, A., Lowe, D., Mantilla Amor, J. M., Martin, J.M., McEvoy, H.,
Ojanen, M., Pehlivan, O., Rougié, B., Salim S. G. R.., “Realisation and
dissemination of thermodynamic temperatures above the silver
point”, Phil. Trans R. Soc. A. 374: 20150043 (2016)
http://dx.doi.org/10.1098/rsta.2015.0043
Determination of T-T90 with ultra-low
uncertainties
INRIM, CNAM, CEM, LNE, NIM, NPL, PTB, TUBITAK-UME, UVa, UNa2
 World’s lowest uncertainty for T – T90 in
ranges ~1 K to ~1358 K [InK 1 + 2]
 Achieved by applying different thermometry
techniques to identify systematic sources of
uncertainty
 InK 1 from ~25 K to ~ 300 K by acoustic
(speed of sound) and dielectric constant gas
thermometry
Dielectric Constant Gas thermometer of PTB: 2 K to 100 K
CCT WG4 consensus T-T90 curve with uncertainties
[before the InK projects]
10
90
T- T (mK)
5
0
-5
-10
-250 -200 -150 -100 -50
t90 (°C)
0
50
10
WG4
consensus
curve did not
identify this
artefact - due
to the ITS-90
algorithms
0
T- T
90
(mK)
5
-5
-10
-250 -200 -150 -100 -50
t90 (°C)
0
50
Full InK 1 summary: T-T90: AGT and
DCGT
More details to be found in

Underwood, R., Sutton, G., de Podesta, M., Stanger, L., Rusby, R., Harris,
P., Morantz, P., Machin, G., “Estimates of the difference between
thermodynamic temperature and the ITS-90 in the range 118 K to 303 K”,
Phil. Trans R. Soc. A. 374: 20150048 (2016)
http://dx.doi.org/10.1098/rsta.2015.0048

Gavioso, R.M., Ripa, D.M., Steur, P.P.M., Gaiser, C., Zandt, T., Fellmuth, B.,
de Podesta, M., Underwood, R., Sutton, G., Pitre, L., Sparasci, F., Risegari,
L., Gianfrani, L., Castrillo, A., Machin, G., “Progress towards the
determination of the thermodynamic temperature with ultra-low
uncertainty”, Phil. Trans R. Soc. A. 374: 20150046 (2016)
http://dx.doi.org/10.1098/rsta.2015.0046

Underwood, R., de Podesta, M., Sutton, G., Stanger, L., Rusby, R., Harris,
P., Morantz, P., Machin, G., “Further estimates of T-T90 close to the triple
point of water”, Int. J. Thermophys. 2016 submitted
Primary thermometry
for low temperatures
PTB, CNAM, LNE, VTT, RHUL, Aalto
 Objectives were to:
To undertake low uncertainty primary thermometry to
resolve discrepancies in the background data of the
PLTS-2000 (InK 1 [+2])
Develop primary thermometers for the direct
realisation and dissemination of the kelvin below 1 K
Primary thermometry at low
temperatures
PTB, CNAM, LNE, VTT, RHUL, Aalto
 Development of low noise dc-SQUIDS for JNT [PTB]
 Development of three primary thermometry methods
in ULT range:
 Two based on Johnson noise in metallic conductor
Magnetic field fluctuation thermometry (MFFT) [PTB]
Current sensing noise thermometry (CSNT) [RHUL]
RHUL CSNT
29
Primary thermometry at low
temperatures
PTB, CNAM, LNE, VTT, RHUL, Aalto
 The third based on arrays of tunnel junctions
Coulomb blockade thermometry (CBT) [A!+VTT]
 Measure differential conductance (V1/2 ∝ T)
 Comparison of T <1 K using CSNT, MFFT and CBT,
and T2000 mediated by SRDs
Differential conductance curve from CBT
30
PLTS-2000 – InK 1 results
(2016)
InK 1 Summary of contribution to
world thermometry
 Significant contributions to world thermometry
• First T values for HTFPs
• HTFPs versus absolute radiometry for high temperature
realisation and dissemination
• Recommendations to CCT about possible dissemination
mechanisms at high temperatures
• Lowest uncertainty values for T – T90 ever determined
• Practical sensors for ULT dissemination and new values
for T – T2000 from 0.02 K to 1 K
• Outcomes of InK 1 project published in: Phil. Trans R.
Soc. A. 374: 20150048 (2016)
InK 2: June 2016-May 2019
 Key activities – fill in the gaps remaining from InK 1:
 T-T90 by different primary thermometry methods
 430 K – 1358 K: high temperature acoustic thermometry
and low temperature primary radiometry
 ~5 K – 200 K: acoustic thermometry, dielectric constant gas
thermometry, refractive index thermometry
33
InK 2: June 2016-May 2019
 Novel primary thermometry methods; e.g. Doppler
broadening, double wavelength – to provide robust checks
for previous (InK 1)T evaluations
 T from 0.02 K to 0.0009 K, complete T-T2000 evaluation
Development of practical primary thermometers for direct T dissemination
34
PLTS-2000 – in InK 2
InK 1 project
2012-2015
To be measured
as part of the
InK 2 project
2016-2019
InK 2: June 2016-May 2019
 Formal CCT workshop early 2019, definitive contributions
to MeP-K-19
Including:
Definitive descriptions of allowable primary thermometry
methods; acoustic, radiometric, refractive index, range
of ULT approaches
Full re-evaluation, with ultra-low uncertainty of T-T90 and
T-T2000
36
Summary
 Introduced the SI and the proposed 2018 redefinition
and the MeP-K-19
 Introduced the implementing the new kelvin (InK 1 + 2)
projects
 A brief overview of the achievements of the InK 1
project
 Prospects for the InK 2 project and overall impact into
world thermometry 2019 and beyond
CONTACT DETAILS:
[email protected]
00 44 208 943 6742