Journal of Alloys and Compounds 645 (2015) S51–S55

Contents lists available at ScienceDirect

Journal of Alloys and Compounds

journal homepage: www.elsevier.com/locate/jalcom

Hydrogen isotope separation for fusion power applications

R. Smith a,b,⇑, D.A.J. Whittaker a,b, B. Butler a,b, A. Hollingsworth a,b, R.E. Lawless a,b, X. Lefebvre a,b,
S.A. Medley a,b, A.I. Parracho a,b, B. Wakeling a,b, JET-EFDA Contributors b,1
a
EURATOM/CCFE Fusion Association, Culham Science Centre, Abingdon OX14 3DB, UK
b
JET-EFDA, Culham Science Centre, Abingdon OX14 3DB, UK

a r t i c l e i n f o a b s t r a c t

Article history: The invited talk given at MH2014 in Salford ranged over many issues associated with hydrogen isotope
Available online 9 February 2015 separation, fusion machines and the hydrogen/metal systems found in the Joint European Torus (JET)
machine located near Oxford. As this sort of talk does not lend itself well to a paper below I have
Keywords: attempted to highlight some of the more pertinent information. After a description of the Active Gas Han-
Fusion dling System (AGHS) a brief summary of isotope separation systems is described followed by descriptions
Tritium: isotope of three major projects currently being undertaken by the Tritium Engineering and Science Group (TESG),
the upgrade to the Analytical Systems (AN-GC) at the AGH, the construction of a Water Detritiation Sys-
tem (WDS) and a Material Detritiation Facility (MDF). Finally, a review of some of the challenges facing
fusion with respect to metal/hydrogen systems is presented.
Ó 2015 Elsevier B.V. All rights reserved.

1. Introduction various hydrogen isotopes and helium could have unforeseen

The Joint European Torus (JET) tokomak and the connected
Active Gas Handling System (AGHS), operated by the Tritium Engi-
neering and Science Group (TESG), were built for the study of a
fusion reactor with operating conditions relevant to a potential
power plant. This required the use of deuterium–tritium gas mix-
tures for plasma fuelling. This meant that the design of the JET
tokomak and its associated support systems had to make provision
for processing of radioactive tritium gas mixtures and tritium con-
taminated materials.
JET is the only tritium compatible magnetic confinement fusion
device currently active. Tritium was injected into a large tokomak
for the first time during the Preliminary Tritium Experiment (PTE)
in 1991 and since then 2 follow up T2 experiments in 1997 and
2003. The next tritium campaign, DTE2 (Deuterium Tritium Exper-
iment 2), is being planned for 2017 and lessons learned during this
campaign will provide further valuable information to ITER and
DEMO.
The main motivational factors for DTE2 focus on risk mitigation
for ITER (the next generation tokomak being constructed in Cadar-
ache, France, by an international consortium. The operation with
⇑ Corresponding author at: EURATOM/CCFE Fusion Association, Culham Science
Centre, Abingdon OX14 3DB, UK.
E-mail address: robert.smith@ccfe.ac.uk (R. Smith).
1
See appendix of Romanelli F. at al Proc. 24th IAEA Fusion Energy Conf. (San Diego,
CA, 2012).
influences on the ‘ITER scenarios’ and an integration of DT H-mode
plasma operations with a Be/W first wall is required. Tritium
inventory control is an area which requires improvement, as does
a validation of wall conditioning techniques for tritium, and meth-
ods for tritium removal. Possibly the most valuable area is that of
operational experience. The training of staff in DT operations, and
the development of organisational competence, is an area that is
vital. The final details of DTE2 are to finalised but the planned
throughput of tritium is to rise from 100 g in DTE1 to 1000 g in
DTE2 and the total inventory in the AGHS will increase from 20 g
in DTE1 to 60 g in DTE2.
As the operators of the world’s only fusion related tritium plant
the members of the TESG are uniquely placed to understand the
issues associated with tritium/fusion operations. As part of the
new EuroFusion consortia the TESG has become involved in the
design of DEMO, the European Fusion device currently in the con-
ceptual design stage which is planned to provide >300 MW (elec-
tric) to a national grid by 2050.
2. Overview of the AGHS PLANT
The Joint European Torus (JET) Active Gas Handling System
(AGHS) was designed and built for the purpose of safely storing,
supplying and recycling the hydrogen isotopes of deuterium (D2)
and tritium (T2) used in fusion experiments at JET [1]. The AGHS
consists of an interconnected series of subsystems (Fig. 1). The
Torus contains the deuterium–tritium plasma during short (of

http://dx.doi.org/10.1016/j.jallcom.2015.01.231
0925-8388/Ó 2015 Elsevier B.V. All rights reserved.
S52 R. Smith et al. / Journal of Alloys and Compounds 645 (2015) S51–S55
the order of 10 s) pulsed fusion experiments. The adjacent Neutral
Injection Boxes (NIBs) are used for heating and fuelling the plasma
through highly energetic neutral beams of deuterium and tritium.
The AGHS provides the tritium used in the plasma and subse-
quently extracts the gas (a mixture of deuterium, tritium and
impurities) after each experiment.
Deuterium and tritium are stored in the Product Storage (PS)
subsystem in the form of uranium hydride, using bespoke uranium
beds (UBeds). The PS UBeds contain either 1 kg or 4 kg of depleted
uranium allowing for storage of 7 or 27 mol of hydrogen, respec-
tively [2,3]. The UBed is heated to a temperature of 873 K to release
gaseous D2 and T2. The D2 and T2, is then expanded into the Gas
Introduction and Distribution (GI and GD) transfer lines for supply
to the Torus, either directly via a Gas Introduction Module (GIM) or
from the NIBs.
Hydrogen isotopes and other impurity gases (predominately
helium and traces of nitrogen and hydrocarbons) resulting from
the JET plasma experiments are pumped back to the AGHS via
ML1 and ML2 using the Mechanical Forevacuum (MF) subsystem
or, when tritium is present, the Cryogenic Forevacuum (CF) subsys-
tem, which provides cryogenic pumping at liquid helium tempera-
tures. CF separates the hydrogen isotopes from the other gases
(impurities) using a mixture of activated charcoal trapping and
rough cryogenic distillation. The impurities are then taken to the
Impurity Processing (IP) subsystem. The hydrogen isotopologues
(a mixture of H2, D2, T2, HD, HT and DT) are sent to the Intermedi-
ate Storage (IS) subsystem, where the gas is temporarily stored on
UBeds.
The IP gas processing loop consists of a 2 m3 reservoir, a cold
trap cooled by gaseous nitrogen (to 160 K) for removal of water,
a nickel catalyst module for removal of hydrocarbons, a thin palla-
dium permeator membrane for removal of any remaining impuri-
ties, four UBeds for hydrogen storage and a 150 m 3 h 1 pump
connected in series. The IP subsystem separates any remaining
hydrogen isotopologues from the impurity gases.
The Cryogenic Distillation (CD) subsystem is used to roughly
separate the deuterium and tritium isotopes and remove the hydro-
gen isotope from the gas mixture. The deuterium and tritium can
then be separated using the Gas Chromatography (GC) subsystem
(note, the interconnections between the CD and GC systems are
not shown in Fig. 1). The resulting D2 and T2 products are then sent
to PS for long term storage ready for supply to the Torus.
Additional AGHS subsystems, which are not part of the main D2
and T2 recycling process, are used to aid the gas processing and
ensure safety. The Analytical Laboratory (AN) is capable of analys-
ing gas samples from each subsystem, allowing the composition of
the gas at every processing stage to be determined. The Exhaust
Detritiation (ED) system (EDS) removes the remaining traces of tri-
tium, by converting all free hydrogen and hydrogen containing
compounds to water, from any gases destined for discharge to
the environment.
3. Hydrogen isotope separation for future fusion applications
The balance of deuterium and tritium in a deuterium–tritium
plasma is crucial for fusion applications. Fusion reactions also have
a very low burn-up rate for the tritium. Therefore, large quantities
of tritium must be separated from deuterium (after removal of
other impurities) and re-cycled after extraction from the Torus.
Some of the work currently being undertaken by TESG involves
looking into alternative isotope separation techniques that can be
utilised by future fusion reactors. At present, the only fusion reac-
tor capable of operating with tritium is the JET and the two meth-
ods used are Gas Chromatography (in the GC subsystem) and
Cryogenic Distillation (in the CD subsystem)
It is likely that the methods used at the AGHS will not be fully
suitable for larger reactors such as ITER and DEMO. This is because
these methods are costly, complicated to operate, energy intensive,
operator intensive and time consuming. Because of the require-
ment for more efficient technologies in future reactors, a number
of different techniques (both established and new) will be assessed
by the TESG, in order to determine their viability.
All isotope separation systems work by exploiting the small dif-
ferences in physiochemical behaviour deriving from their differ-
ence in mass (isotope effects, see e.g. [4]). Examples of these
effects include variations in vapour pressure, chemical bond
strength, boiling and freezing points, viscosity, surface tension,
and variations in optical omission spectra. Chemical equilibrium
and reaction rates can also change between isotopes and be
exploited. In the case of hydrogen, the large proportional mass dif-
ference can prove beneficial for isotope separation as it enhances
the isotopic effects.
In order to establish the most promising candidates for an iso-
tope separation system for future reactors, a review of current
and promising future technologies will be produced by TESG. The
use of metal hydrides as a separation tool will be a key area of
investigation, along with the use of nanoporous framework mate-
rials and molecular sieves. Other areas of research will include:
Atomic vapour laser isotope separation (ALVIS), catalytic exchange,
gaseous diffusion, thermal cycling absorption, centrifugation, bio-
logical techniques and continuous ion exchange alongside the
existing technologies used in the AGHS.
4. Outline of the AN-GC upgrade
The primary instrument for determining the hydrogen isotopo-
logue content of samples (including hydrogen in hydrocarbons) in
AN is the Analytical Gas Chromatograph (AN-GC) instrument [5].
GC instruments separate the various molecular components of a
mixed gas by passing a sample through a series of separation col-
umns, which separate the gases in space, followed by an appropri-
ate detector which distinguishes the boundaries between the
gaseous species (see [5] for a detailed description with diagrams).
Hydrogen isotopes are chemically very similar and there are diffi-
culties associated with handling tritium due to its radioactivity.
Therefore, there are no commercially available GC solutions for a
tritium processing plant.
A new customised GC is being constructed for use during
DTE2. In addition to several standard separation columns, a cus-
tomised hydrogen isotope separation column containing palla-
dium will be included. For safety reasons, the entire assembly
is contained within a nitrogen filled glovebox, thus bespoke fur-
naces and liquid nitrogen dewars are required for temperature
control.
The AN-GC assembly includes several detectors; Thermal Con-
ductivity Detectors (TCDs) for the detection of hydrogen isotopes,
a Methaniser and Flame Ionisation Detector (FID) combination
for hydrocarbons, and Ionisation Chambers (ICs) for tritium. The
ICs for this system are custom built low volume units, with 5 cc
chambers and flush gold plated outer chambers and zero dead vol-
ume micrometering valves to ensure the separation within the AN-
GC is not affected. A Quadrupole Mass Spectrometry (QMS) will
also be connected to the system, which allows detailed analysis
of small samples (pressures 10 12 < P < 10 5 hPa) through a Resid-
ual Gas Analyser (RGA) detector. The RGA allows species that the
AN-GC cannot separate to be discriminated, as well as determining
the isotopic content of overlaying non-radioactive species. The
exhaust of the AN-GC will include a bypass containing a custom
built, tritium compatible Laser Raman (LARA) gas cell for further
analysis of the gasses [6].
 R. Smith et al. / Journal of Alloys and Compounds 645 (2015) S51–S55
Fig. 1. Schematic flow diagram of the JET AGHS [1]. This shows the Torus, NIBs and the D2 and T2 supply systems (left) and the AGHS subsystems used for processing and
storing D2 and T2 (right). The AGHS is connected to the Torus via the D2 and T2 supply lines for supply and the matrix lines 1 and 2 (ML1 and ML2, respectively) for exhaust
extraction.
5. Water detritiation system
The EDS removes hydrogen species from the exhaust by con-
verting them to water and collecting the condensate. The average
activity of this condensate tritiated water is 10 GBq/l, with a max-
imum activity of 180 GBq/l with 4–10 tonnes of water generated
each year and stored. A new capability is being developed by TESG
to process this water (Fig. 2). This reduces the waste disposal costs
and demonstrates a more enclosed tritium cycle during the next
fusion tritium campaign.
The purpose of the Water Detritiation System (WDS) is to crack
water into hydrogen isotopologues and oxygen, then concentrate
the tritium ready for processing by the existing AGHS instruments.
The non tritiated gas is released direct to the atmosphere.
Water is first purified using a bespoke series of tritium compat-
ible water filters. Then, the hydrogen isotopologues are separated
from oxygen using an electrolyser unit. A palladium permeation
membrane is used to remove any impurities and the resultant Q2
(where Q is any hydrogen species) is separated in a two column
Fig. 2. The process cycle of the AGHS, including the capability being developed to
process the tritiated water.
S53
cryogenic distillation system. The two columns can be used in par-
allel for low separation but high volume processing or series to
achieve higher separation but low volume. Comparison of the
two column design and an equivalent single column has been con-
ducted through modelling with PROSIM software [7]. The two col-
umn solution reduced the cooling requirement at 20 K by a factor
of three and allowed for a greater throughput by reducing the
reflux ratio by two (the reflux ration being a measure of recycling
of the distillation top product by the column).
6. New materials detritiation facility
There is around 30 tonnes of intermediate level tritiated waste
(specific activity above 12 kBq/g of T) expected to be produced dur-
ing JET operations up to the end of DTE2. The waste consists of var-
ious steels (70%), carbon fibre composite tiles (20%), beryllium
(3%), and other metals (7%). The specific activity of the waste
is up to 100 kBq/g. It is necessary to re-classify this waste as
low level waste (between 0.1 kBq/g and 12 kBq/g tritium) before
the end of the JET operating contract in 2018 for disposal. The
materials detritiation facility (MDF) has been designed to meet this
requirement through processing two batches of 35 kg of waste per
day during 400 days of operation between early 2015 and the end
of 2018.
Depth profiling studies [8–12] show that between 18% and
62% of the tritium activity for this waste is on the surface, thus
more easily removed than bulk contamination. A study was
made by the JET waste management group and Inutec to deter-
mine the viability of bakeout at 1000 °C under air flow as a tech-
nique for detritiation of carbon fibre composite divertor tiles
(summary available on request) to compliment similar studies
on other fusion related materials [9,12,13–14]. The results of
S54 R. Smith et al. / Journal of Alloys and Compounds 645 (2015) S51–S55
Fig. 3. Schematic representation of one material detritiation process.
these studies indicate that a bakeout under air flow technique is
successful. Other requirements of the MDF are the ability to
track the ILW and LLW waste, sample the waste, break up larger
components and to cycle the process if the LLW limit is not
reached for a given batch.
A commercially available furnace with a 600 L volume that can
reach 1600 °C will be used for the main detritiation stage of the
process. The process is outlined in Fig. 3. After receipt, waste is
sorted, sampled and analysed for tritium content. The waste is then
separated into 40 kg batches placed in barcoded boxes and vented
through the tritium recovery system. The boxes are placed into the
furnace for the main detritiation stage where tritiated water is off-
gassed. A pumping system moves air (to provide the oxygen for
combustion) through the furnace to a bubbler system consisting
of two 200 L water filled drums (where isotopic exchange with
water removes tritium from the air) and out to atmosphere. The
bakeout temperature is chosen such that it is below the melting
point of the material. After bakeout, sampling is done and either
the process is repeated or waste is stored in an LLW rack for dis-
posal. This process can be completed once per day with two fur-
naces used in the facility and operator input only required during
working hours. A feasibility study [15] showed that the process
is achievable, cost effective and safe for waste with the specific
activity in the ranges expected.
7. Challenges facing fusion development with relation to metal
hydrides
Many of the challenges in dealing with metal hydrides are sim-
ilar for dealing with hydrogen species in the fusion community.
The additional hazard experienced in fusion relates to the radiolog-
ical effects of tritium, which readily exchanges with hydrogen in
compounds such as water, so is easily assimilated into living
organisms. This results in a need to carefully monitor for and con-
tain tritium to ensure environmental releases are minimised.
Challenges include:
Methods of storing hydrogen isotopes in a safe and concen-
trated manner: The reaction between hydrogen and uranium in
UBeds is strongly exothermic, which may result in safety limita-
tions if large quantities of gas need to be absorbed in an emer-
gency. The release rate of hydrogen from uranium upon heating
is also limited. The use of uranium, which is a controlled nuclear
material, results in regulatory constraints. Alternative materials
with improved performance are of interest as storage media.
The long term effects of hydrogen diffusion into structural
materials are of interest, including embrittlement and effects on
other material parameters. Relevant materials include stainless
steels, Eurofer, oxide dispersion steels, tungsten and beryllium.
Tritium containment is very important. Hydrogen easily dif-
fuses through many materials, so techniques to form permeation
barriers are required to improve safety, reduce complexity and
increase the lifetime of a plant.
Separation of hydrogen and hydrogen compounds from other
gasses is a key requirement both for low concentrations (such as
diffusion into coolants) and high concentrations (in the tritium fuel
cycle). Novel separation techniques could contribute to fusion
development.
It is more difficult to pump lighter gases such as hydrogen com-
pared to gases such as nitrogen. Experiences with hydrogen pump-
ing technologies are of interest.
Finally, it is worth mentioning that the availability, supply, stor-
age and gas handling of deuterium will be important, and any
issues relating to this specific isotope of hydrogen would be of
interest.
8. Conclusions
The TESG is currently engaged in a series of upgrades in prepa-
ration for DTE2 in 2017 as well as the construction of new facilities
to process waste generated. The experience gained both in this
preparation for, in the execution of DTE2 (Deuterium Tritium
Experiment 2) will be an invaluable resource to the work to come
at ITER and DEMO.
As fusion engineering matures, links with academia outside of
the usual fusion field will become increasingly important. Applying
the knowledge and lessons learnt from relevant fields will benefit
the practical application of fusion technology.
References
[1] R. Lässer, A.C. Bell, N. Bainbridge, P.D. Brennan, B. Grieveson, J.L. Hemmerich, G.
Jones, D. Kennedy, S. Knipe, J. Lupo, J. Mart, A. Perevezentsev, N. Skinner, R.
Stagg, J. Yorkshades, G.V. Atkins, L. Dörr, N. Green, M. Stead, K. Wilson,
Overview of the performance of the jet active gas handling system during and
after DTE1, Fusion Eng. Des. 47 (1999) 173–203.
 R. Smith et al. / Journal of Alloys and Compounds 645 (2015) S51–S55
[2] R. Stagg, J. Hemmerich, R. Lasser, M. Laveyry, J. Lupo, P. Milverton, N. Skinner,
A. Perevezentsev, The intermediate and product storage system for the JET
active gas handling system – inactive commissioning, Fusion Tech. 28 (1995)
1425–1430.
[3] R. Bowman Jr., A. Attalla, NMR studies of the helium distribution in uranium
tritide, Phys. Rev. B 16 (5) (1977) 1828–1843.
[4] Gregory Choppin, Jan-Olov Lilijenzin, Jan Rydberg, Radiochemistry and Nuclear
Chemistry, third ed., Butterworth-Heinemann, 2002.
[5] R. Lasser, A.C. Bell, B. Grieveson, J.L. Hemmerich, R. Stagg, G.V. Atkins, The
analytical gas chromatographic system of the JET active gas handling system –
tritium commissioning and use during DTE1, Fusion Eng. Des. 47 (1999) 173–
203.
[6] D.K. Veirs, G.M. Rosenblatt, Raman line position in molecular hydrogen: H2,
HD, HT, D2, DT, and T2, J. Mol. Spectrosc. 121 (1987) 401–419.
[7] X. Lefebvre, A. Hollingsworth, A. Parracho, P. Dalgliesh, B. Butler, R. Smith, JET
EFDA Contributors, Conceptual design and optimisation for JET water
detritiation system cryo-distillation facility, in: Proceedings of the 24th IAEA
Fusion Energy Conference, San Diego USA, 2012.
[8] N. Bekris, C. Caldwell-Nichols, L. Doerr, M. Glugla, R.-D. Penzhorn, H. Ziegler,
Possible techniques for the detritiation of first wall materials from fusion
machines, J. Nucl. Mater. 307–311 (Part 2) (2002) 1649–1654.
S55
[9] N. Bekris, J.P. Coad, C.H. Skinner, C.A. Gentile, E. Damm, W. Nagele, Thermal
release rate of tritium trapped in bulk and plasma exposed surfaces of carbon
specimens obtained from JET divertor, J. Nucl. Mater. 367–370 (Part B) (2007)
1254–1259.
[10] N. Bekris, J.P. Coad, K. Sugiyama, C. Caldwell-Nichols, T. Tanabe, B. Kloppe, R.
Rolli, JET-EFDA Contributors, Ex-situ tritium removal from JET tiles using RF
inductive heating, Fusion Eng. Des. 83 (7–9) (2008) 1137–1141.
[11] A.N. Perevezentsev, A.C. Bell, L.A. Rivkis, V.M. Filin, V.V. Gushin, M.I. Belyakov,
V.I. Bulkin, I.M. Kravchenko, I.A. Ionessian, Y. Torikai, M. Matsuyama, K.
Wantabe, A.I. Markin, Comparative study of the tritium distribution in metals,
J. Nucl. Mater. 372 (2–3) (2008) 263.
[12] Y. Torikai, D. Murata, R.-D. Penzhorn, K. Akaishi, K. Wantabe, M. Matsuyama,
Migration and release behavior of tritium in SS316 at ambient temperature, J.
Nucl. Mater. 363–365 (2007) 462.
[13] K. Dylst, L. Ortiz, J. Braet, Study on existing detritiation processes and
characterisation of tritiated materials, for application to ITER and DEMO
Final Report, 2007.
[14] S. Rosanvallon, G. Marbach, A.M. Brass, J. Chene, J.P. Daclin, Steel detritiation,
optimization of a process, Fusion Eng. Des. 51–52 (2000) 605–609.
[15] R. Smith, D. Coombs, CCFE Preliminary Proposal for On Site Thermal Treatment
of ILW, UKAEA Internal Document, Sept 2014.