Article pubs.acs.org/ac
Using Thermal Evolution Profiles to Infer Tritium Speciation in Nuclear Site Metals: An Aid to Decommissioning Ian W. Croudace,† Phil E. Warwick,† and Daeji Kim*,†,‡ †
GAU-Radioanalytical, University of Southampton, National Oceanography Centre, Southampton SO14 3ZH, United Kingdom Korea Institute of Nuclear Safety, 19 Guseong, Yuseong, Daejeon, 303-338, Republic of Korea
‡
S Supporting Information *
ABSTRACT: Understanding the association and retention of tritium in metals has significance in nuclear decommissioning programs and can lead to cost benefits through waste reduction and recycling of materials. To develop insights, a range of metals from two nuclear sites and one nonnuclear site were investigated which had different exposure histories. Tritium speciation in metals was inferred through incremental heating experiments over the range of 20−900 °C using a Raddec Pyrolyser instrument. Systematic differences in thermal desorption profiles were found for nonirradiated and irradiated metals. In nonirradiated metals (e.g., stainless steel and copper), it was found that significant tritium had become incorporated following prolonged exposure to tritiated water vapor (HTO) or tritium/hydrogen gas (HT) in nuclear facilities. This externally derived tritium enters metals by diffusion with a rate controlled by the metal composition and whether the surface of the metal had been sealed or coated prior to exposure. The tritium is normally trapped in hydrated oxides lying along grain boundaries. In irradiated metals, an additional type of tritium can form internally through neutron capture reactions. The amount formed depends on the concentration and distribution of trace lithium and boron in the metal as well as the integrated neutron flux. Liberating this kind of tritium typically requires temperatures above 800 °C. The pattern of tritium evolution derived from simple thermal desorption experiments allows reliable inferences to be drawn on the likely origin, location, and phases that trap tritium. Any weakly bound tritium liberated at temperatures of ∼100 °C is indicative of mostly HTO interactions in the metal. Any strongly bound tritium liberated over the range of 600−900 °C is indicative of neutrogenic tritium formed via neutron capture by trace Li and B. Neutron capture by lithium is likely to be more significant than for boron based on lithium’s higher trace element abundance and neutron cross section. The time required for efficient thermal desorption of tritium ultimately depends on the metal composition, its tritium exposure history, integrated neutron flux, sample size, sample geometry, heating rate, and final desorption temperature. etals are important construction materials in fission reactor facilities and account for 22% of the total weight of low-level waste (LLW) and 35% of intermediate level waste (ILW) in the UK.1 Tritium is ubiquitous in nuclear facility construction materials, arising either from neutron capture by trace lithium or boron (Li or B) impurities or through exposure to tritiated water or tritium gas. Any tritium derived from neutron capture by lithium is likely to be more significant based on its trace element abundance and neutron cross section. Deuteriumderived 3H is generally only significant as an external contaminator of metals and is most significant in nuclear power stations where deuterated water is or has been used as a moderator (e.g., CANDU reactors and the UKs prototype SGHWR reactor). Tritiated metal disposal will also be an issue for deuterium−tritium burning fusion reactors (e.g., JET and ITER) as their long-term operation will involve the use of large activities of tritium potentially resulting in large volumes of contaminated metals such as Inconel, stainless steel, copper, beryllium, and aluminum alloys.2−5 The tritium terminology in this paper uses HT (gaseous mixtures of hydrogen and tritium) and HTO (tritiated water or water vapor) acronyms. Any other terms simply repeat specific terms used in cited references.
M
© 2014 American Chemical Society
Determination of tritium requires specialized methods of extraction and measurement because of the low-energy pure β emission of the radionuclide (Emax = 18.6 keV). Also due to its high mobility, 3H cannot be estimated relative to a more readily measurable radionuclide using a “fingerprint” approach. A number of limited nondestructive (in situ) methods have been used for quantifying 3H in metals, including surface activity monitoring6,7 and spectroscopy based on Bremsstrahlung and Xrays induced by interaction of the 3H β-particle with the substrate (BIXS method).8 However, these methods are relatively insensitive and give no insight into the bulk activity. Destructive methods are widely used and more readily permit quantitative determination of 3H. The two main methods involve acid dissolution9−13 and thermal desorption/oxidation.14−18 Whichever extraction method is used for metals, it is critical that the 3H is quantitatively liberated regardless of its speciation or association with the metal. Received: June 5, 2014 Accepted: August 26, 2014 Published: August 26, 2014 9177
dx.doi.org/10.1021/ac502244a | Anal. Chem. 2014, 86, 9177−9185
Analytical Chemistry
Article
Figure 1. Schematic diagram (a−d) showing incorporation, formation, and desorption/decomposition mechanisms in metals for tritium gas, tritiated water, and tritium produced through neutron capture by Li and/or B. The generalized profile for HT−metal interactions (a) is based on work by Calder. (Adapted from ref 21. Copyright (1973), with permission from Elsevier.) k1, surface interaction between water molecule and surface oxyhydroxide layer via hydrogen bonding; k2, tritium exchange between water and surface hydroxide/water; k3, diffusion of surface tritium into bulk via grain boundaries; k4, penetration of HTO into the bulk metal via defects; k5, in situ production of tritium via neutron capture reactions on trace Li or B; k6, tritium gas diffusion into bulk metal along defects/grain boundaries; k7, H/T exchange between tritium gas and adsorbed water/surface hydroxide; k1′, decomposition of surface hydroxide and liberation of tritium; k2′, diffusion of tritium produced in situ from impurity mineral lattice ; k3′, diffusion of bulk tritium to surface via grain boundaries/defects; k4′, decomposition of metal tritide.
Chemical Properties of Metal Surfaces. The surfaces of metals typically (Figure 1) comprise a bulk oxide layer with a terminal layer of oxide or oxyhydroxide which grow to a limited thickness of ca. 2 nm.19 For Fe, Co, Ni, and Cu, the bulk oxide layer comprises low oxidation state oxides with a terminal layer comprising high oxidation state oxides, oxyhydroxide and hydroxides. For metals like Al, Zr, Sn, Mo, W, and Rh, the bulk oxide layer comprises higher oxidation state oxides with a terminal layer of similar high oxidation state oxides, oxyhydroxides and hydroxides.20 The oxide/oxyhydroxide layer is overlain by a chemisorbed water layer. Overlying this is a layer of physisorbed water. The exact composition of the surface layer will be dependent on the metal composition, surface crystal plane orientation, and exposure conditions. Interaction of Tritium with Metals. The mechanism of 3H interaction with metals will be dependent on the origin of the 3H (whether present as HT or HTO or produced in situ via neutron capture), the metal composition, and surface properties. The key interaction pathways are summarized in schematic form in Figure 1a−c. Interaction with HT. The interaction of 3H gas with metals has been reviewed22 and investigated as part of fusion and nuclear weapons’ programs. A typical 3H distribution profile was reported by Calder et al.21 for 304 grade stainless steel (Figure 1a) where three key zones of tritium concentration are marked. Grade 304 is the most commercially important austenitic grade stainless steel closely followed by grade 316 that contains an addition of molybdenum to improve corrosion resistance. The highest 3H activities were associated with the metal surface (zone I21). Exposure of metal surfaces to tritium gas results in H/T exchange with H in surface adsorbed water.
Tritium can pass from this surface into the metal via bulk diffusion and grain boundary diffusion across the surface oxide layer. Activities initially decline with depth before increasing again to a subsurface maximum followed by a progressive decline consistent with classical diffusion (zone II21). At greater depth, a steady 3H concentration was observed associated with grain boundary diffusion (zone III21). Measured tritium permeation rates around 27 °C were considerably higher than expected from extrapolation of high temperature data and was significant even for metals such as Cu and Al which exhibit low H permeability.3 Torikai et al.13 noted that tritium in 316 grade stainless steel was concentrated in the surface (0.05 μm) and then declined sharply to a depth of 7 μm before steadily increasing again until reaching a steady state concentration in the region of 70−80 μm. Removal of the surface 3H layer resulted in its reformation over time through resupply of 3H from the bulk material. Perevezentsev et al.3 also reported a surface maximum in 3H concentrations in stainless steel and Inconel exposed to HT gas. However, for Cu, Al, bronze, and Be, no surface maximum was detected and 3H concentrations decreased continuously with depth. Interaction with HTO. The few experimental studies of HTO−metal interactions have been performed over relatively short exposure periods with the quantity of HTO adsorbed depending on exposure conditions and metal substrate.23,24 The results show the tritium to be initially present as free HTO or HTO weakly adsorbed to the underlying metal surface layer. Tritiated water may bond to metal surfaces exclusively via the oxygen lone pair of electrons where alternate water molecules are bonded to the surface with intervening molecules bound to the surface attached water molecules via H-bonding. In this case, the hydrogen of the intervening molecules points away from the 9178
dx.doi.org/10.1021/ac502244a | Anal. Chem. 2014, 86, 9177−9185
Analytical Chemistry
Article
diffusion coefficients ranged from ∼10−6 cm2/s at 60 °C up to 5 × 10−4 cm2/s at 727 °C (1000 K).27 For Cu and Ni, the diffusion coefficients ranged from 5 × 10−7 at 227 °C to 10−4 cm2/s at 727 °C (1000 K).28 For Al, diffusion coefficients were lower (5 × 10−6 cm2/s at 393 °C) although, at room temperature, the diffusion coefficient (2 × 10−7 cm2/s) was 4 orders of magnitude higher than expected from extrapolation of high temperature diffusion measurements.29 Even for the lowest diffusion coefficient, the mean diffusion path length after 1 h at 327 °C (600 K) was found to be 650 μm suggesting that bulk hydrogen isotopes should be recoverable by thermal desorption. Thermally desorbed tritium predominantly exists as HTO irrespective of the original form of 3H that interacted with or formed in the metal. Metal surfaces have been shown to catalyze both the oxidation of 3H species to HTO in air and any hydrogen isotope exchange between surface-bound and gas phase H isotopic species.3 The current study is important because there have been few previous empirical investigations dealing with the relative evolution of 3H from operationally exposed metals. From a nuclear waste characterization perspective, the thermal desorption characteristics of 3H were expected to provide information on its various interactions and associations. Understanding tritium evolution profiles can effectively guide radioanalytical protocols to ensure accurate waste characterization prior to waste sentencing.
surface (H-up). Alternatively, all water molecules can be bonded to the metal surface with alternating molecules bonding via M− O or M−HO bonds (H-down). In addition, water can dissociate on the surface to form an OH layer. The exact mechanism of adsorption depends on the metal composition and its wettability and is determined by the electronic structure of the metal and crystal plane orientation at the surface.23 In all cases, the bond energies for adsorbed HTO are 0.1−0.4 eV indicating that liberation would occur at low temperatures. Nishikawa et al.24 differentiated between physically adsorbed, chemically adsorbed, and structural forms of surface water on the metal surface, based on the desorption characteristics. Physically adsorbed water readily exchanged with a dry carrier gas; chemically adsorbed water was liberated at temperatures of ca. 100 °C whereas structural waters remained after this procedure. The lowest 3H capacities were observed for Cu whereas Al and 304 stainless steel showed comparable, higher capacities, approximately 5 times greater than for Cu. At low partial pressures ( 95% recovery of 3 H). However, for Al, structural water was stable at 1 × 10−4 mol/ m2 up to 100 °C and then dropped to 6 × 10−5 mol/m2 corresponding to a phase change from β-Al2O3·3H2O to αAl 2 O 3 ·H 2 O. Hammond et al. 26 showed that Cd(OH) 2 dehydrates to CdO in one step at 350 °C. Tritium release rates from bulk metals depend on its location, migration path length, diffusivity, and temperature. Other factors such as the thickness of the surface oxyhydroxide layer, lattice imperfections, dislocations, and grain boundary effects can also impact the observed rate of diffusion. For steels, reported
■
METHODOLOGY A range of irradiated and nonirradiated metals (Table 1) from two nuclear reactors undergoing decommissioning (Material Test Reactor, MTR, Dounreay, UK, and the Steam Generating Table 1. Total 3H Activities Measured Using the Standard Thermal Ramping Protocols for the Metals Investigated fraction
total 3H (Bq/g)b
exposure to HTO
top surface side surface bulk top surface side surface bulk bulk
777 ± 24 1030 350 109 ± 5 44 1 13.0 ± 1
exposure to HTO
bulk
4.9 ± 0.3
exposure to neutrons and HTO exposure to neutrons and HTO
bulk
8230 ± 209
bulk upper Al layer middle Al/ boron carbide layer lower Al layer bulk
4280 ± 116 1080 6540
112 313 ± 15
bulk
9.4 ± 0.8
bulk
4.9 ± 0.3
locationa
source of 3H
stainless steel
nuclear site (Winfrith)
exposure to HTO
copper
nuclear site (Winfrith)
exposure to HTO
aluminum
aluminum
non-nuclear site nuclear site (Winfrith) MTR
boral
MTR
steel
MTR
lead
MTR
cadmium
MTR
material
steel
exposure to neutrons and HTO exposure to neutrons and HTO exposure to neutrons and HTO
a
MTR Materials Test Reactor (Dounreay, UK); Winfrith is the site of the SGHWR (steam generating heavy water reactor). bSample heated from 50 to 900 °C (600 °C for nonirradiated steel) with the ramping cycles as shown in Figure 2. 9179
dx.doi.org/10.1021/ac502244a | Anal. Chem. 2014, 86, 9177−9185
Analytical Chemistry
Article
Figure 2. 3H desorption profiles for different metals using standard thermal ramp rates.
prepared by carefully filing or drilling the bulk sample to avoid significant heat generation that could potentially have resulted in the loss of any weakly bound tritium. Gamma-emitting radionuclides were determined to verify whether metals had been irradiated using a Canberra 40% well-type HPGe γspectrometer. Tritium thermal desorption profiles were determined on all bulk metal samples by collecting the desorbed tritium into
Heavy Water Reactor, SGHWR, Winfrith) and from a nonnuclear site that handled tritium were investigated. Operationally exposed metals were used throughout the study to ensure that the test materials were representative of the materials typically encountered during decommissioning. The irradiated metals originated from the MTR had been exposed to a significant but unspecified neutron flux during the reactor’s operating history. Homogenous, finely divided samples suitable for analysis were 9180
dx.doi.org/10.1021/ac502244a | Anal. Chem. 2014, 86, 9177−9185
Analytical Chemistry
Article
Figure 3. 3H desorption from irradiated and nonirradiated metals using very slow thermal ramping over 24 h for temperatures from 120−300 °C.
HTO recovery and detection limits are 95% and 0.01 Bq/g, respectively.34 Tritium Liberation from Metals Using Total Acid Dissolution. The efficiency of 3H thermal desorption from metals was evaluated by comparing 3H measurements following acid dissolution of a selection of metal samples with those obtained using thermal desorption. Dissolution and desorption tests were conducted at the same time to minimize discrepancies arising from 3H loss during storage (see Table 1, Supporting Information).
bubblers that were changed according to time (not temperature). This sampling approach was followed, in preference to collection according to temperature, to suppress kinetic effects. Temperature was also recorded whenever a bubbler change occurred. Additionally, for nonirradiated stainless steel and copper, desorption profiles were also determined for subsamples taken from the surface and across the cut edge of the metal coupon. Tritium extraction was achieved using a multisample, flowthrough, oxidative heating/combustion system previously described by Warwick et al.34 (Pyrolyser furnace system supplied by Raddec Ltd., Southampton UK). For metal samples that melted within the experimental temperature range (e.g., Al, Pb, and Cd), the sample was loaded onto silica sand in a silica boat. The liberated decomposition products were passed through an intermediate zone and then into a zone containing 10 g of platinised alumina catalyst heated to 800 °C designed to oxidize any combustion products to CO2 and H2O. Any water vapor, including HTO, was trapped in a bubbler containing 20 mL of 0.1 M HNO3. For the determination of tritium desorption profiles, the bubblers were changed every 30 min with subsamples of the bubbler solution being taken for 3H counting. All tritium measurements were performed using Wallac 1220 Quantulus ultra low-level background liquid scintillation counters. The
■
RESULTS Tritium Concentration. Tritium activity concentrations are found to vary significantly between the various metal samples investigated (Table 1). The highest activity concentrations were observed for irradiated aluminum, but high activity concentrations were also observed, unsurprisingly, for the middle B4C (boron carbide) layer of Boral.30−33 For nonirradiated metals, the highest activity concentration was associated with stainless steel and the lowest with Cu. The relative activity concentrations for different metal samples depended on the specific metal, its ambient 3H concentration, the neutron flux (for irradiated metals), and the duration of exposure. Although the specifics of these conditions were largely unknown, it is noteworthy that the 9181
dx.doi.org/10.1021/ac502244a | Anal. Chem. 2014, 86, 9177−9185
Analytical Chemistry
Article
3
higher than that obtained during the standard heating profile. This may indicate that although a significant proportion of 3H in the Boral was present in a weakly bound form the rate of release of this 3H was slow. Tritium Distribution with Depth. Tritium loss according to its inferred location was assessed for Cu and stainless steel (Figure 4). For Cu, the majority of the 3H was associated with the
H activity concentration in the irradiated steel of 313 Bq/g is significantly greater than the predicted value of 12 Bq/g reported by Westall et al.25 Comparison of 3H activity data arising from thermal desorption and acid dissolution confirmed that quantitative recovery of 3H was achieved using thermal desorption (Table S2, Supporting Information). For nonirradiated and irradiated steels, there is good agreement between 3H activity concentrations determined by the two approaches. For Cu, the 3H activity concentrations determined by thermal desorption were higher than for acid dissolution although the reason for this discrepancy is not clear. For irradiated Al sampled from the upper layer of Boral, thermal desorption again gave significantly higher 3H activities which probably indicate minor contamination from the underlying boron carbide layer that did not dissolve in the acid medium. Desorption Profiles. Tritium thermal desorption profiles tended to show two distinct features during slow heating (Figure 2). For most of the metals tested, 3H desorption began between 200 and 300 °C and peaked quite rapidly. For irradiated Al, irradiated Pb and nonirradiated stainless steel, however, 3H desorption occurred over a wider temperature range. For Boral, two distinct stages of 3H desorption were observed consisting of a relatively low temperature release followed by a more prolonged higher temperature release. Phase changes at discrete temperatures and the rate of diffusion of liberated 3H would both impact the observed desorption profiles. To investigate the effect of diffusion kinetics, selected metal samples were heated to a fixed temperature for 24 h and the proportion of 3H liberated was determined (Figure 3 and Table S3, Supporting Information). The observed 3H desorption profiles for different metals can be divided into three empirical groups. Group A (Irradiated Steel and Nonirradiated Copper). Significant 3H desorption occurred at low temperatures (50%) was liberated at 200−300 °C and was probably associated with structural waters. Similar behavior is expected to be observed for nonirradiated steel and Al based on their standard thermal desorption profiles. Nonirradiated Stainless Steel. This exhibited a more prolonged 3H release during the standard thermal desorption test. Heating of the material for prolonged periods at set temperatures confirmed that only 18% of the 3H inventory was liberated at temperatures below 300 °C with 300 °C.
