1. Materials
Ag-ETS-10 (granular, binderless, 16x30 mesh, packing density 950 kg/m³) was prepared and provided by Extraordinary Adsorbents Inc. of Edmonton, Alberta, Canada. Ag-13X (spherical, + 20 mesh, packing density of 970 kg/m3) was obtained from Sigma-Aldrich, Belgium. Nuclearcarb 207C (granular, 6x12 mesh, packing density of 450 kg/m³) was obtained from Chemviron Carbon, United Kingdom, and CarboAct High Purity Carbon (fragmented, 0.01–0.4 cm, packing density of 250 kg/m3) was obtained from Carbo-act International, Netherlands. Ag-ZSM-5 (spherical, particle size 0.2–0.5 mm, packing density of 520 kg/m³) was kindly provided by CEA, France, and was synthetized by Ag exchange of Na-PZ-2/25 obtained from Zeolyst, USA, according to the description given in the thesis of Deliere18. The moisture adsorbent 13X-APG was obtained from UOP CH Sarl, Switzerland, as spherical 8x12 mesh beads.
222Rn was generated and supplied by degassing a PYLON RN-1025 flow-through 3.77 MBq 226Ra source (reference date: 25/04/2014) supplied by PYLON Electronics Inc., Canada. Nitrogen was supplied in N5.0 purity from an in-house LN2 supply tank with a nominal pressure of 8 bar and a dew point of 203 K. Compressed air was supplied from an in-house air compressing station at 6 bar with a dew point of 243 K. Technical grade N4.8 argon was supplied in bottles by Air Products NV, Belgium.
2. Methods
2.1 Synthesis of Ag-ETS-10
ETS-10 was synthesized by the suspension of particulates of anatase (TiO2) in a sodium silicate solution (~ 30% SiO2, 8.5% Na2O). After the addition of ETS-10 seeds and a fluoride mineralizer, the mixture was crystallized for 72 h at 473 K in a 3.8 L Parr stainless steel autoclave at autogenous pressure. After filtering, washing and drying, the crystalline product underwent granulation and was sieved into 16x30 mesh aggregates. The high purity, as synthesized, ETS-10 was loaded into a column and strip exchanged with an excess of silver nitrate in solution at 363 K for 36 h. The silver-exchanged ETS-10 (now Ag-ETS-10) was thoroughly washed with deionized water and dried at 473 K in a forced air oven.
2.2 Experimental Rn adsorption setup
Rn adsorption experiments were performed in a laboratory scale setup schematically depicted in Fig. 5. The trapping device consisted of two redundant moisture traps and two Rn adsorption columns, hygrometers for indication of the moisture content in the system, flow controllers (MFC and VA-flowmeter), a vacuum pump and a radon monitor. The system was built out of stainless steel fittings, valves and tubing (1/4” OD) to assure leak tightness for minimal moisture ingress. A part of the system contained flexible perfluoroalkoxy alkane (PFA) tubing to ease operation.
Supplied N2, air and Ar was reduced in pressure and fed into the experimental system as carrier gas. The flow of the carrier gas in the setup was controlled by a thermal mass flow controller (Red-y smart, Vögtlin Instruments GmbH, Switzerland). The setup was equipped with two redundant moisture traps (MT-1 and MT-2) filled with molecular sieve 13X-APG to reduce the dew point of the incoming gas below 200 K. Redundancy was provided for continuous operation in case of column saturation. Both columns have a volume of approx. 385 cm³ (Øin = 4.05 cm, L = 30 cm) and are made of stainless steel (SS 316). The radon adsorption was performed with one of the adsorption columns RT-1 and RT-2, both having a volume of 17 cm³ (Øin = 1.2 cm, L = 15 cm) and made of stainless steel (SS 316). All columns are designed and manufactured in-house. For all columns, a CF flange with a copper seal on top of the column is installed to ensure leak tightness. The seal is regularly exchanged and a leak test is performed after each column opening via pressurization. Two dew point transmitters (Easidew online, Michell Instruments, United Kingdom) are installed in the setup to measure the dew point of the carrier gas at the exhaust of MT-1/2 and RT-1/2, respectively.
The setup enables regeneration of different sorbents in situ, i.e. while keeping the sorbent in the column. The heating of the stainless steel column was performed via an electrical heating wire connected to a temperature controller that regulates the temperature to a predetermined set-point via a thermocouple. The thermocouple was positioned on the outer wall of the column. The columns and heating wire are thermally insulated with glass fiber tape. The setup was equipped with a vacuum pressure transmitter (pressure range − 1 to 2.5 barg) for pressure monitoring and leak tightness validation the of the setup and the columns in particular after opening and closing.
A (removable) adsorbent column (120 cm³) can be inserted into the experimental loop via quick-connects to enable the degassing of the 222Rn source. Once removed, the loop was closed via stainless steel tubing. The outgoing radon concentration was monitored by sampling of a part of the exhaust stream with a AlphaGUARD Professional Radon Monitor DF2000, Bertin GmbH, Germany. It enables a continuous determination of the volumetric radon concentration with a measuring range from 2 Bq/m³ to 2 MBq/m³ with a data sampling frequency of 1 min. The monitor was equipped with a flow-regulated pump which can be adjusted from 0.05 L/min to 2 L/min, allowing sampling of the exhausted gas stream.
Experiments for environmental radon adsorption were performed in a non-radioactive laboratory at room temperature by using a KNF Laboport N86 vacuum pump and a mass flow controller (Red-y smart, Vögtlin Instruments GmbH, Switzerland) in suction mode through one MT400-4 and one MT120-4 13X-4A moisture traps (both obtained from Agilent, USA) and a glass column (Øin = 1.2 cm, L = 15 cm manufactured at Glasatelier Saillard, Belgium) containing the tested adsorbent. All components were connected by flexible PFA tubing. The exhausted air stream was sampled using the AlphaGUARD DF2000 radon monitor.
2.3 Breakthrough experiments
Breakthrough experiments were performed at laboratory room temperature, which was 293 ± 2 K. A stepwise procedure was applied. At first, the Pylon RN-1025 was degassed to remove accumulated 222Rn in equilibrium with 3.77 MBq 226Ra. In order to control the total 222Rn activity injected in each performed experiment, the degassing was performed with dried laboratory air for 1 hour at a flow rate of 0.5 NL/min, where the flushed 222Rn was adsorbed on a 50 g NuclearCarb 207C column that was subsequently removed. After degassing, radon started accumulating again in the RN-1025 source at a rate of 473 Bq/min. In a next step, the accumulated radon was injected into the selected adsorbent column by flushing the radon source with the desired carrier gas (N2, Ar or air) and flow rate (0.5-5 NL/min) for 5 min. In order to reach the desired activity for injection, the time between the end of the degassing and the end of the injection was adjusted, being 25–35 min for experiments with the activated carbons and Ag-13X and 18–40 h for Ag-ZSM-5 and Ag-ETS-10, respectively. After injection, the radon source was closed and N2, Ar or air was fed to the column until all injected radon has passed through the column (complete breakthrough) and the experiment terminated.
2.4 Material regeneration
Each tested adsorbent underwent a regeneration cycle via heat treatment under inert gas conditions prior to each experiment. In each case, N5.0 nitrogen was used as regeneration gas. The regeneration was performed for at least 15 h at 503 K for Ag-ETS-10, 523 K for Ag-13X and Ag-ZSM-5, while for both activated carbons 423 K was chosen. The moisture adsorbent 13X-APG was regenerated at 553 K. The regeneration of 13X-4A and Ag-ETS-10 for environmental 222Rn adsorption was performed in a muffle furnace (type AAF 11/7, Carbolite, UK) at 503 K overnight and subsequently filled manually into the respective adsorption columns.
2.5 Data treatment
The adsorption coefficient K (m³/kg) for each tested adsorbent were calculated using the formula given as
where t is the mean retention time in [h], m the mass of adsorbent in [kg] and F the volumetric flow rate in [m³/h]. The retention time t is the time of the eluted decay-corrected radon activity to be 50% of the total injected radon activity. For each experimental data set of N measurement points, the elution curves are smoothed via the adjacent averaging method over \(\sqrt{N}\) points. Fitting of the breakthrough curves and the subsequent estimation of t was performed using software data analysis applying the chromatographic plate model function6. The overall uncertainty associated to the measured adsorption coefficient was estimated to be in the order of 10% at a confidence level of k = 1.