A new passive sampler for collecting atmospheric tritiated water vapor

A new passive sampler for collecting atmospheric tritiated water vapor

Atmospheric Environment 154 (2017) 308e317 Contents lists available at ScienceDirect Atmospheric Environment journal homepage: www.elsevier.com/loca...

1MB Sizes 5 Downloads 246 Views

Atmospheric Environment 154 (2017) 308e317

Contents lists available at ScienceDirect

Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv

A new passive sampler for collecting atmospheric tritiated water vapor Bin Feng, Bo Chen*, Weihai Zhuo, Weiyuan Zhang Institute of Radiation Medicine, Fudan University, Shanghai, 200032, China

h i g h l i g h t s  A new passive sampler for collecting environmental HTO was developed.  CFD simulations were carried out to improve the construction of the sampler.  The impact of meteorological factors on a passive HTO sampler was investigated.  The sampler is reliable and applicable for HTO collection in the field.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 4 October 2016 Received in revised form 15 January 2017 Accepted 18 January 2017 Available online 19 January 2017

A new passive sampler was developed for collecting environmental tritiated water vapor. The construction of the sampler was improved according to computational fluid dynamics (CFD) simulations in which the influence on vapor collection by the turbulence inside the sampler was considered. Through changes in temperature from 5  C to 35  C and relative humidity from 45% to 90%, the new sampler revealed stable performance of the sampling rate. Compared with the previous samplers, the new sampler significantly lowered the effect of wind speed. Using the adsorption kinetic curve of the sampler provided in the co-comparison experiments, the quantitative relationship between the mass of adsorbed water and the cumulative absolute humidity exposure was established. Field applications in the vicinity of a nuclear power plant show that the data obtained by the new samplers is consistent with the active measurement. The sampler was preliminarily proven to be reliable and flexible for field investigation of HTO in the atmosphere. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Passive sampler HTO CFD Sampling rate Meteorological factors

1. Introduction Tritium (3H or T), is the only radioisotope of hydrogen (1H) that is measured in programs of atmospheric surveillance due to its potential health hazard (Little and Lambert, 2008; Marang et al., 2011; UNSCEAR, 2012). The atmospheric tritium is mainly generated by natural production, atmospheric weapon tests, and routine or accidental releases from the nuclear industry (Belot, 1986). With the development of fusion technologies, the release of tritium is likely to increase in the near future (Rosanvallon et al., 2016). Atmospheric tritium mainly appears in three forms: tritiated water vapor (HTO), tritiated molecular hydrogen (HT) and tritiated hydrocarbons (CH3T). HTO is a prevalent form of tritium in the atmosphere and has a higher radiological impact on humans than

* Corresponding author. E-mail address: [email protected] (B. Chen). http://dx.doi.org/10.1016/j.atmosenv.2017.01.035 1352-2310/© 2017 Elsevier Ltd. All rights reserved.

HT and CH3T. Due to its volatility and chemical stability, atmospheric HTO is capable of wide dispersion and long-distance transport (Boyer et al., 2009). Hence, for the purpose of assessing the potential influence of tritium on the environment, or for investigating the behavior of HTO in the global hydrologic cycle, it is necessary to conduct a large scale and long term survey on atmospheric HTO concentrations. Among the different approaches used to collect atmospheric HTO, active sampling devices such as the cold trap (Bekris et al., 2001) and the drum bubbler (Rule et al., 1998) were widely applied for the collection of atmospheric water vapor due to their good performance (i.e., high sampling rate, robustness) (Herranz et al., 2011). However, the considerable expense of purchasing, operating, and training technical staff is a disadvantage for these devices, especially in the application of long term and large scale surveys. The dependence on a continuous electricity supply also limits the setting of surveillance points in some countryside regions

B. Feng et al. / Atmospheric Environment 154 (2017) 308e317

(May et al., 2011). In response to the limitations of the active sampling device, a method for passive sampling of atmospheric HTO was developed in the last decades. Because of their low cost, convenience in operation, and electricity independence, the passive samplers have great flexibility in field applications (Seethapathy et al., 2008). These passive devices are capable of covering a larger spatial area with the deployment of large quantities across a broad geographic region simultaneously, which would benefit studies on atmospheric HTO monitoring and health risk evaluation and atmospheric HTO transportation. (Ba and Xu, 2010). According to previous studies, two types of passive sampler were developed for atmospheric HTO sampling. One was first designed by Stephenson for Canadian nuclear facilities in 1984 (Stephenson, 1984). The sampler, using a small collecting chamber and tritium-free H2O as absorbent, was preferable to high-level HTO field monitoring due to its low sampling rate (approximately 5 L air per day). Some researchers used arrays of these samplers at the same sites to collect enough water vapor to reach the requirements for the lower detection limits of environmental monitoring (Stephenson, 1990; Wood et al., 1993; Wood, 1996). However, that led to inconvenient operation and some negligible interaction between different samplers (Yang et al., 2001). The other type of passive sampler had a relatively large collecting chamber containing molecular sieve or silica gel as the adsorbent. One of these devices was the sampler developed by T. Iida et al. that was widely used for nuclear facilities in Japan (Iida et al., 1995). Akata et al. improved the construction and provided another cylindrical sampler for HTO monitoring (Akata et al., 2011). With the great improvement in sampling rate of these samplers, one single sampler of this type could collect enough water vapor for laboratory analysis in most cases. Thus, the interaction effect between samplers can be totally avoided. Some researchers also noted that the sampling rate of a passive sampler would be affected by meteorological factors (i.e., temperature, humidity, and wind speed) (Plaisance et al., 2004; Tuduri et al., 2006; Marang et al., 2011; Guo et al., 2014). However, the meteorological dependence in sampling rate for these samplers, which may ultimately introduce a serious systematic bias, had not been fully discussed in previous studies. Furthermore, due to the limitation of computing capability in 1980s and 1990s, the design of these samplers was mainly based on simplified theoretical calculations or experience (Harper and Purnell, 1987) rather than a fluid dynamics simulation. The diffusion path lengths for these samplers had not been optimized for the meteorological influence affecting the flow field in the collecting chamber. For samplers with the inlet on the top, the situation was even worse because the diffusion length would be gradually shortened with the expansion of the adsorbent. In this study, to reduce the influence of meteorological factors, a new tubular passive sampler was developed for environmental HTO investigation. This sampler retained the advantages of previous passive HTO samplers, such as low cost and convenience in operation. The structural modification and the improvement of the diffusion length were based on computational fluid dynamics (CFD) simulations. This new sampler was compared with previous samplers under different environmental conditions through experiments in the constant temperature and humidity chamber as well as in an indoor wind pipe. To evaluate the performance and flexibility of this new sampler, the co-comparison experiments on the campus of Fudan University and the in-site experiments in the vicinities of a nuclear power plant (NPP) were conducted, respectively. The goal of this study was to provide a new passive sampler which would benefit atmospheric HTO monitoring.

309

2. Materials and methods 2.1. Design of the sampler To overcome the problem of the expansion of adsorption material and to reduce the influence of aerosol particles deposition on the inlet, a suspension type of cylindrical collecting chamber with a sampling hole on the bottom was employed in the new sampler. As a reference for designing this new sampler, some passive samplers with similar structures were considered. These samplers have been used in a large number of environmental monitoring collections for different atmospheric pollutants (NO2, VOC, SVOC, GEM, etc.) and proven to be suitable for long term sampling in the field (Plaisance et al., 2004; Zhang et al., 2012; Guo et al., 2014; Uzmez et al., 2015). The construction of the passive sampler and its component parts are illustrated in Fig. 1. A 4A molecular sieve (MS-4A) in the amount of 300 g was employed in this sampler as an adsorbent for its adsorption capacity with high selectivity and stability (Toci et al., 1995; Kim et al., 2007; Iwai and Yamanishi, 2010). The saturated adsorption capacity was reported to be approximately 20% of its weight for molecular sieve (Iida et al., 1995). A stainless steel cylinder with a 200-mesh screen as the bottom was placed at the upper part of the sampler as the container for the molecular sieve. It is designed to prevent the 1He3H isotopic exchange reaction between adsorbed HTO and the plastic shell of the samplers (Rosson et al., 2000; Iwai and Yamanishi, 2010). The screen is meant to provide a stable contact surface for the adsorbent with the air. A circular opening covered by a membrane filter was used as the sampling inlet in the center of the bottom lid. To meet the demands of different sampling rates, two different sampling inlets with diameters of 43 mm and 13 mm, respectively, were provided. The filter could not only effectively prevent the deposition of aerosol particles on the inner surface or the adsorbent, but could also reduce the influence of wind to the flow field inside the sampler (Sekine et al., 2008). Minimizing the effects of wind turbulence was one of the main aims in designing this sampler. It was reported that suppressing the turbulence near the surface of the adsorbent by creating a relatively stagnant region between the external atmosphere and the adsorbent could significantly reduce the influence of wind on the sampling rate for a tubular passive sampler (Fan et al., 2006). Several experiments have demonstrated that a diffusion length to diameter ratio (L/d) of greater than 2.5 is sufficient to overcome the effects of wind turbulence (Harper and Purnell, 1987; Gair and Penkett, 1995). In this sampler, the diffusion path length was set to 20 cm. To evaluate the performance of this new sampler, two traditional types of samplers were also involved in this study. One is an improved version of Stephenson's sampler which was widely used in China (Yang et al., 2002), called a Type A sampler hereinafter. The other is a sampler used in our previous studies which has similar construction to the one developed by Iida et al. (1995), called a Type B sampler hereafter. The new sampler was named SCPS (Suspended Cylindrical Passive Sampler). SCPS-13 and SCPS-43 were used to represent the new sampler with different sampling inlets. The geometric parameters of all these 3 types of samplers are listed in Table 1. 2.2. CFD simulations To understand the flow features inside the passive sampler, a series of two-dimensional (2D) stimulations were conducted using CFD software Fluent 14.5 (ANSYS Inc. Pennsylvania, USA). To reveal the different flow characteristics inside the SCPS sampler, the diffusion lengths of SCPSs were set to be 8.5 cm, 20 cm,

310

B. Feng et al. / Atmospheric Environment 154 (2017) 308e317

Fig. 1. The construction of the sampler.

 . C2 ¼ 3:5  ð1  εÞ dp  ε3

Table 1 Construction of the three types of passive samplers used in this manuscript. TYPE

Geometric construction

Opening

Adsorbent

A

Cylinder F2.7 cm  6.1 cm Box 24.8 cm  18 cm  9.3 cm Cylinder F8 cm  33 cm

F1.0 cm

MS-4A (5 g) MS-4A (300 g) MS-4A (300 g)

B SCPS

F4.3 cm F4.3 cm (SCPS-43) F1.3 cm (SCPS-13)

where a is the permeability (m2), C2 represents the inertial resistance factor (m1), and dp and ε denote the equivalent diameter of 4A molecular sieve pellets (m) (i.e., the diameter of the sphere for spherical pellets) and the void fraction (%), respectively.

and 25 cm in simulations, respectively. The Type A and Type B samplers were also stimulated using the geometries listed in Table 1. The computer-generated models of all passive samplers (an external fluid domain connected to a passive sampler) were created by AutoCAD®2014, and their geometries were meshed with unstructured quad cells using Gambit 2.4.6 (ANSYS Inc. Pennsylvania, USA). The base grid resolution was selected to be approximately 1/ 100 of the diameter of the passive sampler after several checks for grid-independence results, resulting in approximately 290,000 to 360,000 total number of grid cells for different samplers. Considering the relatively high Reynolds numbers for wind pipes and the important near-wall effect in this case, the standard k-ε turbulent model was used in the CFD simulation. A species transport model with no chemical reaction was utilized to investigate the concentration profile in the samplers in which two components, water vapor at a specified mass fraction and air, were considered (May et al., 2011). In the numerical model, the computational domain was divided into three fluid zones: a bulk zone, a membrane zone and an absorption zone (see Fig. 2). The bulk zone included both an external tube and an inner diffusion tube. The boundary conditions of the external tube were set by a velocity inlet on left side and an outflow on the right side. All two-dimensional simulations were conducted for a range of freestream velocities at the inlet: 0.1, 2, 4, and 6 m$s1, respectively. In the simulations, the absorption zone was considered as an ideal water absorbent with infinite absorbing capability and also a porous medium. To obtain reasonable flow resistance characters, the parameters of the porous medium were calculated by the extended Darcy's law and the Ergun equation (Yang et al., 2013):

a ¼ dp 2  ε 3 and

.  150  ð1  εÞ2

(2)

(1)

2.3. Laboratory experiments To investigate the impacts of meteorological factors on the sampling rate of passive samplers, two groups of laboratory experiments were designed. One is the exposure chamber experiment in a constant temperature and humidity chamber. The level of the meteorological parameters can be precisely controlled: temperature (0  C 100  C, 0.5  C), relative humidity (35%e98%, 2%), and wind speed with a constant parameter of 0.5 m$s1. In the chamber experiment, environments with four different temperatures ranging from 5  C to 35  C were simulated, whereas the impact of humidity was studied at four levels from 45%RH to 90%RH for each temperature. In each batch of the experiments, three replicates of each passive sampler (SCPS-13, SCPS-43, Type A and Type B) were placed in the chamber with the same meteorological parameters for 72 h. The distance between samplers in the chamber was approximately 10 cm. The mass of adsorption material (4A molecular sieve) in each sampler was 300.01 ± 0.02 g (Mean ± SD) for the SCPS and the Type B sampler, and 5.01 ± 0.03 g (Mean ± SD) for the Type A sampler, respectively. The molecular sieve was pretreated for 6 h at 300  C and 99.9 kPa as initial water desorption. The other group of experiments was carried out in a test room for investigating the impacts of wind speed on the sampling rate of passive samplers. During the experiments, the temperature of the room was in the range of 23.6  Ce25.4  C, while the relative humidity fluctuated between 63.2% and 67.4%. Seven simple wind pipes (i.e. series variable axial fans inside 500 mm diameter vent tubes) were placed parallel in the room. The wind speeds for these pipes were set to be 0.1, 0.6, 1.2, 1.8, 2.5, 3.5, 4.5 m$s1, respectively. All types of passive samplers (SCPS-13, SCPS-43, Type A and Type B) were tested in the pipes, respectively. In each experiment, seven passive samplers of the same type were placed in different pipes simultaneously for 7 days. Another sampler with the inlet sealed was also deployed in the room as the blank sample. The experiments were repeated three times for each type of passive sampler. During the laboratory experiments, the temperature and relative humidity were monitored by a ZOGLAB MINI detector

B. Feng et al. / Atmospheric Environment 154 (2017) 308e317

311

Fig. 2. Geometry of CFD simulation.

(ZOGLAB®, Hangzhou, China). The wind speed was recorded using a Hot-Film-Anemometer AR866 (SMARTSENSOR®, Provo, USA). 2.4. Co-comparison experiment To estimate the effective sampling rate and the adsorption kinetic curve of the passive sampler, co-comparison experiments were carried out using the SCPS samplers in parallel with a set of homemade active HTO sampling devices. The active sampler included an air pump, water vapor traps, and a sampling control system. The sampling rate of the active sampler was 1.0 L$min1 calibrated by a soap film flowmeter (Gilian Gilibrator2, Sensidyne, St. Petersburg, USA). The adsorbent (1200 g of 4A molecular sieve) inside the active device was replaced every 48 h working hours. The SCPS-13 and SCPS-43 samplers were deployed in three different sites on the campus of Fudan University from May to July 2016, and placed 1 m above the ground surface. Six replicate samplers were deployed in parallel with the active sampler at each site. The sampling periods were 10, 15, 20 and 25 days, respectively. The effective sampling rate of the passive samplers was estimated by comparison with the results of the active sampler. All the meteorological parameters, including temperature, relative humidity, wind speed, and wind direction, were recorded using a JY-QX mini-type automatic weather station (Unicontrol Ltd, Beijing, China) during the experiment. By recording the uptake weight of the SCPS sampler and its corresponding accumulative absolute humidity, the adsorption kinetic curve of the sampler was obtained. 2.5. Field application To investigate the flexibility of the SCPS sampler, a preliminary in-field experiment was carried out at the vicinity of a nuclear power plant (NPP) in China from June 2016 to September 2016. In the study region the tritium sources of air emissions are mainly released from two 700 MW CANDU reactors. Sampling site 1 is a village located 1 km from the NPP in the downwind region. Site 2, which is 10 km from the NPP, is expected to have relatively low HTO in the atmosphere. The control site named site 3 is on the campus of FUDAN University, 120 km away from the NPP.

Six SCPS-13 samplers were deployed at two monitoring sites near the NPP, as well as at the control site. After a one-month sampling period, all of the samplers were routinely replaced and sent back to the laboratory for further analysis. At each site the metrological data including wind speed, wind direction, temperature, and relative humidity were recorded. 2.6. Laboratory analysis All the samplers were sealed and transported to the laboratory immediately after deployment. The mass of each sampler was weighted by an electronic balance before and after the exposure period. The molecular sieves taken from the samplers were desorbed for 4 h at high temperature and low pressure (300  C, 99.9 kPa) , Czech Republic) using the DJ-500 desorption device (VF, Svitavska (Malara et al., 1999; Gabrus et al., 2015). The desorbed water was distilled using rotary evaporators (YARONG Inc. Shanghai, China). After distillation, an 8 ml water sample was taken in a polythene vial and mixed with 12 ml of liquid scintillation cocktail (UltimaGold™ uLLT, PerkinElmer, Waltham, USA) (Pujol and SanchezCabeza, 1999). The tritium concentration in the 20 ml polythene vial was measured with a liquid scintillation counter LB7 (ALOKA, Tokyo, Japan) for 1000 min. The counting efficiency for tritium was determined by the external standard channel ratio (ESCR), which was calibrated by a set of quenched standard tritium sources referenced to NIST SRM 4947C. The background count rate and counting efficiencies were 1.62 cpm and 29.5%, respectively. The detection limit of the liquid scintillation counter was estimated to be 0.19 cpm. 2.7. Data analysis The data were processed using Microsoft Excel 2013, and the statistical analysis was performed using SPSS 20.0. The significance level for all reported statistics was p ¼ 0.05 by default. Where applicable, values are reported as the mean ± standard deviation (SD) except as noted. In the co-comparison experiments, the collection efficiency of

312

B. Feng et al. / Atmospheric Environment 154 (2017) 308e317

the passive sampler was expressed by the daily increased weight of collected vapor per unit mass of the adsorbent, which was in units of mg$g1$d1, while that of the active sampler was described by the increased weight per sampling volume in units of g$m3. In this study, the effective sampling rate (SR) was used to represent an equivalent volume of air collected by the passive samplers, which was the key factor in evaluating the stability of the samplers. The sampling rate in units of m3$d1, could be calculated as:

SR ¼ Ms =ðCa  TÞ

(3)

where Ms represents the amount of water vapor collected by the sampler during the sampling period (g); Ca and T denote the concentration of water vapor in the equivalent volume (g$m3) and the time of deployment (d), respectively. 3. Results and discussion 3.1. Results of CFD simulations To evaluate the stability of the samplers that differed in diffusion lengths, the profiles of the turbulence intensity inside the samplers are illustrated in Fig. 3. The results indicate that there was no significant turbulence in the collection chamber when the surrounding wind speed outside the samplers was relatively low. However, when the wind speed increased to 2 m$s1, high turbulence intensity was observed in the region near the sampler inlets. For the Type B sampler as well as the SCPS sampler with the 8.5 cm diffusion length, the surface of the adsorbent was too close to the turbulent region. Although the steady-state CFD simulation could not directly estimate the influence of the turbulence on the stability of sampling performance, it is reasonable to consider the shorter diffusion length as one of the negative effects on robustness and flexibility which should be avoided. For the SCPS sampler with the diffusion length greater than 20 cm, there was a buffer region between the adsorbent and the turbulent region resulting in relatively stable adsorption for the water vapor. Balancing the influence of

turbulence and the adsorption efficiency, the diffusion length of the SCPS sampler was designed to be 20 cm. Fig. 4 shows the concentration profiles of water vapor inside different samplers when the surrounding wind speed ranged from 0.1 to 6 m$s1. The results reveal that the vapor concentration occurred as gradient distribution vertically and uniform distribution horizontally near the surface of the adsorbent inside the SCPS and the Type A samplers. However, the concentration distribution inside the Type B sampler was neither horizontally uniform nor vertically a gradient and was directly influenced by the surrounding wind speed. The results indicate that the Type B sampler performance would have a much more serious dependence on the surrounding wind field than the SCPS and Type A samplers. Under the same meteorological conditions (25  C, 70%RH, wind speed ¼ 2 m$s1), the profile lines of the vapor concentration at the center axis of different samplers are plotted in Fig. 5. It is obvious that the slope of the concentration curve for the Type B sampler was much steeper than the other samplers. Based on Fick's first law (Gorecki and Namiesnik, 2002), the adsorption process could be described using the following formula:

dm=dT ¼ dC=dd  D  S

(4)

where C is the vapor concentration in the atmosphere (g$m3), d is the diffusion length (m), D is the diffusion coefficient (m2$s1), m is the mass of the adsorbed water, T is the sampling period (s). The curve slope in Fig. 5 represents the gradient of the vapor concentration, dC/dd. According to Eq. (4), the steeper concentration slope leads to a higher requirement for capacity in the adsorption rate, dm/dT. Thus, the simulation results revealed that the Type B sampler might place a higher demand on the adsorption capacity of adsorption materials. 3.2. Effect of temperature and humidity Several studies have shown that temperature and relative humidity could have potential impacts on the performance of passive samplers during pollutant sampling periods (Uzmez et al., 2015).

Fig. 3. Turbulent intensity contours for different types of passive sampler.

B. Feng et al. / Atmospheric Environment 154 (2017) 308e317

313

Fig. 4. Relative humidity contours for different types of passive samplers.

Absolute humidity (g·m-3)

20

SCPS-13

SCPS-43

Type A

Type B

16

Diffusion length of SCPS

12

Diffusion length of Type B

8

Diffusion length of Type A

4

0 0

0.04

0.08

0.12

0.16

0.2

Distance above the surface of absorption material (m) Fig. 5. The profile lines of the vapor concentration at the center axis of different samplers.

For the HTO passive sampler the meteorological dependence would be worse since the collection of water vapor would be directly influenced by the absolute humidity in the atmosphere. Additionally, the activity of the adsorbent and the diffusion coefficient would increase with the temperature. The results of the sampling rate obtained under controlled conditions in a chamber are summarized in Table 2. Due to the limited sampling time (72 h) and the relatively low sampling efficiency of the Type A sampler, the uptake weights of adsorbent in Type A samplers were less than the lower detection limit, thus the related results were not listed in Table 2. One-way ANOVA analysis was taken for significance test and the result showed that the difference in the sampling rate at different meteorological parameters

was significant (p < 0.001, n ¼ 48) for each type of the passive samplers. Therefore, the effect of temperature and relative humidity on the sampling rate cannot be ignored. To evaluate the flexibility of different samplers in the range of normal meteorological conditions, the coefficient of variation (CV) values were also calculated from Table 2 data as the factor for the stability of performance. The CV value of the sampling rate for the SCPS sampler (SCPS-13, 12.97% and SCPS-43, 10.89%) was much lower than that of the other samplers (18.88%) that revealed relatively stable adsorption and strong robustness. Furthermore, the sampling rates of SCPS and Type B samplers for different absolute humidity are plotted in Fig. 6. The results show that the SCPS sampler had a very stable sampling rate. However, the collection efficiency of the Type

314

B. Feng et al. / Atmospheric Environment 154 (2017) 308e317

Table 2 Sampling rate of different types of passive samplers for chamber exposure experiments. Temperature ( C)

Relative Humidity (%RH)

Absolute Humidity (g$m3)

Sampling rate of passive sampler (m3$d1)

5 5 5 5 15 15 15 15 25 25 25 25 35 35 35 35

45 60 75 90 45 60 75 90 45 60 75 90 45 60 75 90

3.059 4.078 5.098 6.118 5.772 7.696 9.620 11.544 10.362 13.816 17.271 20.725 17.796 23.727 29.659 35.591

0.034 0.044 0.040 0.038 0.037 0.037 0.040 0.040 0.027 0.030 0.034 0.037 0.036 0.031 0.035 0.037

SCPS-13 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

SCPS-43 0.004 0.002 0.002 0.002 0.001 0.004 0.002 0.002 0.002 0.003 0.003 0.003 0.004 0.003 0.002 0.002

0.073 0.086 0.091 0.093 0.070 0.075 0.091 0.089 0.069 0.071 0.085 0.088 0.077 0.074 0.077 0.080

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

Type B 0.004 0.003 0.005 0.001 0.001 0.002 0.006 0.004 0.004 0.003 0.003 0.001 0.001 0.003 0.009 0.003

0.257 0.266 0.369 0.355 0.243 0.244 0.262 0.269 0.200 0.211 0.228 0.247 0.213 0.217 0.233 0.229

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.011 0.008 0.009 0.011 0.002 0.006 0.019 0.026 0.015 0.010 0.021 0.012 0.013 0.011 0.009 0.016

Fig. 6. Sampling rate at different meteorological conditions (The “C, A, and:” represent the mean value, the top and bottom line represent the maximum and minimum value, respectively).

B sampler fluctuated remarkably and had a declining tendency with increasing absolute humidity. The data from the experiments in the chamber confirmed the conclusions of the CFD simulations. The turbulence inside the Type B sampler led to instability of adsorption. The horizontally nonuniform adsorption as well as the great gradient of the vapor concentration vertically in the sampler caused a remarkable dependence of adsorption rate on the partial surface of the adsorbent, which would exceed the capability of the molecular sieve and finally lead to a declining sampling rate for the Type B sampler in high humidity environments. The SCPS sampler effectively solved this problem with the construction improvements and acceptable reduction of efficiency. 3.3. Effect of wind speed To determine the relationship between the surrounding wind speed and the sampling rate, the results of the experiments conducted in the wind pipes are illustrated in Fig. 7. An obvious increase in the sampling rate is observed with increasing wind speed

for the Type B samplers. For other samplers, correlation analysis also showed a significant positive relationship between the sampling rate and the wind speed (p < 0.001 for Type A and SCPS). The results show that the sampling rate increases 19.88% for the Type A sampler and 32.99% for the Type B sampler when the wind speed increased from 0.1 m$s1 to 4.5 m$s1. The rates of increase for SCPS-43 and SCPS-13 samplers are 8.54% and 15.97%, respectively. These results indicate that the influence of surrounding wind speed on the sampling performance could not be ignored for these samplers. Although this problem had not been totally solved, the effect of the surrounding wind on the sampling efficiency could be significantly lowered using these new SCPS samplers. The improvement on the performance stability could be attributed to the structural modifications of these new samplers. Both our CFD simulation and some other reported results (Plaisance et al., 2004; Guo et al., 2014) revealed that the high surrounding wind speed may introduce a turbulence eddy inside the sampler from the inlet. The experimental data confirmed that the appropriate length to diameter ratio for the diffusion chamber of new samplers must be sufficient to overcome the effects of wind

B. Feng et al. / Atmospheric Environment 154 (2017) 308e317

315

Fig. 7. Relationship between the sampling rate and wind speed (The “C, :, -, and A” represent the mean value, the top and bottom line represent the maximum and minimum value, respectively).

turbulence. This result is consistent with previous estimates obtained from the studies of tubular passive samplers (Harper and Purnell, 1987; Gair and Penkett, 1995).

adsorption material; and b is also associated with the factor of interfacial transfer and the mass transfer coefficient in the sampler. For the SCPS-13 and SCPS-43 samplers, the adsorption kinetic curves were expressed as Eq. (6) and Eq. (7), respectively.

3.4. Results of the co-comparison experiments

   Ms ¼ 115:6  1  exp  1:24  105  Evapor

(6)

   Ms ¼ 115:8  1  exp  2:75  105  Evapor

(7)

The adsorption kinetic curves for two SCPS samplers are illustrated in Fig. 9 based on the results of the co-comparison experiments carried out at the campus of Fudan University. The curves were fitted by the equation which is as follows:

   Ms ¼ a  1  exp b  Evapor

(5)

Equations with similar forms are widely used in the adsorption field (Zhang et al., 2012), in which ms (g) is the amount of water vapor accumulated in the adsorption material; Evapor (g$m3$h) represents the gross vapor exposure which is expressed as the cumulative absolute humidity around the sampler; a and b are fitting parameters related to sorption coefficient and volume of

The results in Fig. 8 also indicate that the SCPS-13 and SCPS-43 samplers could provide flexibility on selecting the sampling period. To collect 15 g water (for the 20 ml liquid scintillation vial) and 45 g water (for the 100 ml liquid scintillation vial), the estimated sampling periods for SCPS-13 samplers were 30 d and 90 d, respectively under typical environmental conditions with the absolute humidity of 20 g$m3, whereas the SCPS-43 sampler only needs 15 d and 45 d. For a dry season, it is suggested to use SCPS-43 sampler and

Fig. 8. Adsorption kinetic curves for SCPS samplers.

316

B. Feng et al. / Atmospheric Environment 154 (2017) 308e317

Fig. 9. Comparison of sampling efficiency between passive and active collection.

expand the sampling period according to the historical meteorological data. Assuming the average absolute humidity to be 10 g$m3, the sampling period should be doubled. The MDA (minimum detectable activity) of the liquid scintillation counter LB7 was 0.011 Bq for 1000 min measurement using 20ml polythene vial. Based on the equivalent sampling volume of air for 30 day sampling, the corresponding detection limit for HTO-inair concentrations were estimated to be 0.010 Bq$m3 for SCPS-13 sampler, as well as 0.005 Bq$m3 for SCPS-43 sampler. The relationship between the uptake rates of passive samplers with the average vapor concentration detected by the active device is illustrated in Fig. 9. It shows significant linear correlations for both SCPS-13 and SCPS-43 samplers (p < 0.001). 3.5. Results of the field experiment The results of the field experiment are summarized in Table 3, including the cumulative absolute humidity estimated by the uptake weight of the passive sampler compared with data directly

from meteorological monitoring, and the HTO concentrations acquired by active and passive sampling. The results show that the cumulative absolute humidity calculated according to the uptake weight of the passive sampler and Eq. (6), corresponds well with the one acquired by direct meteorological monitoring. The deviations are within ±6%, which indicates that it is reliable to evaluate the gross vapor exposure without any external meteorological information using the adsorption kinetic curves of the new samplers. The results prove that the new passive samplers can feasibly be applied in field applications independently. The three-month preliminary monitoring of the HTO concentration at three different sites revealed that the results of the passive measurement agreed with the data by active sampling within 16.0%. The HTO concentration observed at Site 1 which is 1 km from the NPP reached the highest value of 0.76 Bq$m3 by passive sampling. The HTO concentrations at Site 2 were in the range of 0.09e0.13 Bq$m3 from June to September 2016, which is slightly higher than the background level measured at the control site

Table 3 Passively and actively measured HTO concentrations for field applications. Sampling Period

01 Junee03 July 2016

Sites

1 2 3

03 Julye03 Aug., 2016

1 2 3

03 Aug.e13 Sep., 2016

1 2 3

a

Relative Present Difference, %.

Uptake Weight (g)

21.65 22.02 18.66 18.81 24.02 23.75 18.01 17.41 17.15 16.51 20.43 19.88 21.63 21.18 23.11 23.57 27.66 26.59

Average Absolute Humidity (g$m3)

HTO Concentration (Bq$m3) a

By passive sampling

By meteorological monitoring

RPD

21.69 22.10 18.41 18.58 24.36 24.06 18.29 17.63 17.34 16.64 21.00 20.38 16.91 16.52 18.21 18.62 22.33 21.34

20.92

3.69 5.66 2.70 1.85 4.19 2.88 5.54 1.72 0.99 3.09 1.26 4.19 1.79 4.05 1.74 4.02 3.29 1.28

18.93 23.38 17.33 17.17 21.27 17.22 17.90 21.62

Passive Data

Active Data

RPDa

0.63 0.65 0.12 0.13 0.11 0.10 0.76 0.72 0.10 0.09 0.07 0.08 0.48 0.45 0.10 0.10 0.06 0.07

0.69

8.71 5.66 5.26 13.04 15.35 3.70 7.72 2.30 5.70 16.01 10.06 5.87 8.59 2.72 4.86 10.89 12.75 4.75

0.12 0.09 0.70 0.11 0.08 0.44 0.09 0.07

B. Feng et al. / Atmospheric Environment 154 (2017) 308e317

(0.07e0.11 Bq$m3). The HTO concentrations fall with distance from the NPP and have remarkable fluctuation in different months, especially for the site near the NPP. To accurately evaluate the environmental effect of atmospheric HTO, it is necessary to carry out a large scale and long term survey of its dynamic distribution. The new sampler developed in this study would be a suitable tool for that investigation. 4. Conclusions A new passive sampler for collecting HTO vapor was developed in this study. With the structural modification based on the CFD simulation, the variation of the sampling rate was controlled within 13% during changes of temperature from 5  C to 35  C and relative humidity from 45% to 90%. Compared with the previous samplers, the new sampler significantly lowered the effect of wind speed, which raised the sampling rates only 8.54% and 15.97% for SCPS-43 and SCPS-13 samplers, respectively, when the wind speed increased from 0.1 m$s1 to 4.5 m$s1. The adsorption kinetic curve obtained in the co-comparison experiments, revealed the quantitative relationship between the mass of adsorbed water and the cumulative absolute humidity exposure. The samplers were successfully used in a preliminary measurement to determinate the atmospheric HTO concentrations at the vicinity of a NPP. The results indicate the new sampler is reliable for HTO collection in the field. The results of this study provide a low-cost method to establish a network for atmospheric HTO investigation and would consequently help to improve our understanding of the global HTO dynamic distribution. Acknowledgments This work was supported by the National Natural Science Foundation of China with the Grant No. 11375048. References Akata, N., Kakiuchi, H., Kanno, K., Shima, N., Hisamatsu, S., 2011. Determination of the atmospheric HTO concentration around the nuclear fuel reprocessing plant in Rokkasho by using a passive type sampler. Fusion Sci. Technol. 60, 1292e1295. Ba, Q., Xu, Y.F., 2010. Input function and simulated distributions of tritium in the North Pacific. Sci. China Earth Sci. 53 (3), 441e453. Bekris, N., Hutter, E., Albrecht, H., Penzhorn, R.D., Murdoch, D., 2001. Cold trapping of traces of tritiated water from the helium loops of a fusion breeder blanket. Fusion Eng. Des. 58e59, 423e428. Belot, Y., 1986. Tritium in plants - a review. Radiat. Prot. Dosim. 16 (1e2), 101e105. tat, P., Badot, P.M., Boyer, C., Vichot, L., Fromm, M., Losset, Y., Tatin-Froux, F., Gue 2009. Tritium in plants: a review of current knowledge. Environ. Exp. Bot. 67 (1), 34e51. Fan, Z.H., Jung, K.H., Lioy, P.J., 2006. Development of a passive sampler to measure personal exposure to gaseous PAHs in community settings. Environ. Sci. Technol. 40 (19), 6051e6057. Gair, A.J., Penkett, S.A., 1995. The Effects of wind-speed and turbulence on the performance of diffusion tube samplers. Atmos. Environ. 29 (18), 2529e2533. Gabrus, E., Nastaj, J., Tabero, P., Aleksandrzak, T., 2015. Experimental studies on 3A and 4A zeolite molecular sieves regeneration in TSA process: aliphatic alcohols dewateringewater desorption. Chem. Eng. J. 259, 232e242. Gorecki, T., Namiesnik, J., 2002. Passive sampling. S0165e9936(02)00407-74 TracTrends Anal. Chem. 21, 276e291. Guo, H.L., Lin, H.M., Zhang, W., Deng, C.Y., Wang, H.H., Zhang, Q.G., Shen, Y.T., Wang, X.J., 2014. Influence of meteorological factors on the atmospheric mercury measurement by a novel passive sampler. Atmos. Environ. 97, 310e315.

317

Harper, M., Purnell, C.J., 1987. Diffusion sampling-a review. Am. Industrial Hyg. Assoc. J. 48 (3), 214e218. Herranz, M., Alegria, N., Idoeta, R., Legarda, F., 2011. Sampling tritiated water vapor from the atmosphere by an active system using silica gel. Radiat. Phys. Chem. 80 (11), 1172e1177. Iida, T., Yokoyama, S., Fukuda, H., Ikebe, Y., 1995. A simple passive method of collecting water-vapor for environmental tritium monitoring. Radiat. Prot. Dosim. 58 (1), 23e27. Iwai, Y., Yamanishi, T., 2010. Development of tritiated vapor absorbent applicable to the atmospheric detritiation system in a nuclear facility. Appl. Radiat. Isotopes 68 (9), 1642e1649. Kim, K.R., Lee, M.S., Paek, S., Yim, S.P., Ahn, D.H., Chung, H., 2007. Adsorption tests of water vapor on synthetic zeolites for an atmospheric detritiation dryer. Radiat. Phys. Chem. 76 (8e9), 1493e1496. Little, M.P., Lambert, B.E., 2008. Systematic review of experimental studies on the relative biological effectiveness of tritium. Radiat. Environ. Biophysics 47 (1), 71e93. Malara, C., Ricapito, I., Edwards, R.A.H., Toci, F., 1999. Evaluation and mitigation of tritium memory in detritiation dryers. J. Nucl. Mater. 273, 203e212. , E., Marang, L., Siclet, F., Luck, M., Maro, D., Tenailleau, L., Jean-Baptiste, P., Fourre Fontugne, M., 2011. Modelling tritium flux from water to atmosphere: application to the Loire River. J. Environ. Radioact. 102 (3), 244e251. May, A.A., Ashman, P., Huang, J.Y., Dhaniyala, S., Holsen, T.M., 2011. Evaluation of the polyurethane foam (PUF) disk passive air sampler: computational modeling and experimental measurements. Atmos. Environ. 45 (26), 4354e4359. Plaisance, H., Plechocki-Minguy, A., Garcia-Fouque, S., Galloo, J.C., 2004. Influence of meteorological factors on the NO2 measurements by passive diffusion tube. Atmos. Environ. 38 (4), 573e580. Pujol, L., Sanchez-Cabeza, J.A., 1999. Optimisation of liquid scintillation counting conditions for rapid tritium determination in aqueous samples. J. Radioanalytical Nucl. Chem. 242 (2), 391e398. Rosson, R., Jakiel, R., Klima, S., Kahn, B., Fledderman, P., 2000. Correcting tritium concentrations in water vapor monitored with silica gel. Health Phys. 78 (1), 68e73. Rosanvallon, S., Torcy, D., Chon, J.K., Dammann, A., 2016. Waste management plans for ITER. Fusion Eng. Des. 109 (B), 1442e1446. Rule, K.R., Larson, S., Kivler, P., Scott, J., 1998. Portable tritium processing using a drum bubbler. Plasma Devices Operations 6 (1e3), 203e210. recki, T., Li, X.J., 2008. Passive sampling in environmental analSeethapathy, S., Go ysis. J. Chromatogr. A 1184 (1e2), 234e253. Sekine, Y., Watts, S.F., Rendell, A., Butsugan, M., 2008. Development of highly Sensitive passive sampler for nitrogen dioxide using porous polyethylene membrane filter as turbulence limiting diffuser. Atmos. Environ. 42 (18), 4079e4088. Stephenson, J., 1984. A diffusion sampler for tritiated-water vapor. Health Phys. 46, 718, 718. Stephenson, J., 1990. Re-evaluation of the Diffusion Sampler for Tritiated Water Vapour. Ontario Hydro Report HSD-SD-90e20. Toci, F., Viola, A., Edwards, R.A.H., Mencrelli, T., Forcina, V., 1995. Sorbent materials for fusion-reactor tritium processing. Fusion Eng. Des. 28, 373e377. Tuduri, L., Harner, T., Hung, H., 2006. Polyurethane foam (PUF) disks passive air samplers: wind effect on sampling rates. Environ. Pollut. 144 (2), 377e383. UNSCEAR, 2012. Source, Effects and Risks of Ionizing Radiation. UNSCEAR. Uzmez, O.O., Gaga, E.O., Dogeroglu, T., 2015. Development and field validation of a new diffusive sampler for determination of atmospheric volatile organic compounds. Atmos. Environ. 107, 174e186. Wood, M.J., Mcelroy, R.G.C., Surette, R.A., Brown, R.M., 1993. Tritium sampling and measurement. Health Phys. 65 (6), 610e627. Wood, M.J., 1996. Outdoor field evaluation of passive tritiated water vapor samplers at Canadian power reactor sites. Health Phys. 70 (2), 258e267. Yang, H.L., Wen, X.L., Wu, B., 2001. Experimental study of performance of passive HTO sampler. Radiat. Prot. Bull. 21 (2), 18e23 (in Chinese, with English abstract). Yang, H.L., Wen, X.L., Wu, B., Yang, H.Y., 2002. Experimental study of performance of passive HTO-in-air sampler. Nucl. Electron. Detect. Technol. 06, 545e548 (in Chinese, with English abstract). Yang, X.F., Zheng, Z.Q.C., Winecki, S., Eckels, S., 2013. Model simulation and experiments of flow and mass transport through a nano-material gas filter. Appl. Math. Model. 37 (20e21), 9052e9062. Zhang, W., Tong, Y.D., Hu, D., Ou, L.B., Wang, X.J., 2012. Characterization of atmospheric mercury concentrations along an urban-rural gradient using a newly developed passive sampler. Atmos. Environ. 47, 26e32.