- Email: [email protected]
Low pressure fusion exhaust gases separation
Fusion Engineering and Design xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Fusion Engineering and Design journal homepage: www.elsevier.com/locate/fusengdes
Low pressure fusion exhaust gases separation ⁎
Domenico De Meisa, , Maria Richettab, Emanuele Serrac, Eric Louradourd a
ENEA, Fusion and Nuclear Safety Department, C. R. Frascati, Via E. Fermi 45, 00044, Frascati (Roma), Italy University of Rome Tor Vergata, Department of Industrial Engineering, via del Politecnico 1, 00133, Roma, Italy ENEA, Department for Sustainability, C.R. Casaccia, Via Anguillarese 301, 00123, Santa Maria di Galeria (Roma), Italy d CTI-Céramiques Techniques Industrielles, 382 Avenue du Moulinas, F-30340, Salindres, France b c
A R T I C LE I N FO
A B S T R A C T
Keywords: Gas selectivity Inorganic porous membranes Ceramic membranes Tokamak demonstration reactor (DEMO) Plasma enhancement gases (PEG) Permeance
Plasma enhancement gases (PEGs) (such as: nitrogen, neon, argon and other inert gases) are injected into the plasma of several tokamaks in order to reduce the power load over the plasma facing components. The exhaust gas in the tokamak demonstration reactor (DEMO) consists of more than 90% of unburned fuel gas (D and T) and the remaining part will be He, PEGs and impurities. In DEMO reactor it is foreseen to recover the fuel gas and PEGs. The research focuses to remove the He from fuel gas and PEGs from fusion reactor. For this purpose, six commercial ceramic membranes have been tested at a permeation apparatus built at ENEA Casaccia laboratories. Single gas permeances for H2, He, Ar and air were measured with a pressure drop across the membranes between 10 Pa up to 100 kPa at room temperature. The selectivities found are low except for the 0.5 nm pore size membranes. The membranes have been supplied by Ceramiques Tecniques et Industrielles (CTI SA, France).
1. Introduction Plasma Enhancement Gases (PEGs) are injected into the tokamak to convert the plasma thermal energy to ultraviolet and soft X-ray radiation. Possible PEGs gases are: nitrogen, neon, argon, xenon and other inert gases. DEMO’s plasma exhaust will contain the impurities coming from the interactions of the plasma with the surfaces, PEGs gases, not burned deuterium and tritium and alfa particles [1–3]. For the DEMO exhaust processing system, the use of inorganic membranes has been recently taken into consideration because of their reduced energy consumption, continuos operation, temperature and radiation stability, modularity and scale-up [4–6]. In the present study six commercial ceramic membranes have been characterized in permeation tests in order to measure their permeances at a feed pressure between 10 Pa to 100 kPa at room temperature. The membranes tested have been given by CTI SA France. The tests have been performed at Dr. Emanuele Serra permeation laboratory in ENEA Casaccia. 2. Experimental set-up The method chosen for the determination of the gas permeance through the membranes is a gas phase technique. The permeation apparatus is constructed using standard stainless steel Ultra High Vacuum (UHV) components. It is described in the references [6–8]. The module
⁎
is made of stainless steel. The two ends of the module are removable to allow for membrane replacement. Gas tightness between the module and the membrane at each end is maintained by 2 fluoroelastomer orings (Viton®). The measurements were conducted using a gas pressure from 10 Pa to 100 kPa. After evacuating the apparatus to an ultra high vacuum so that both sides of the membrane are initially in contact with vacuum, the feed gas side is exposed to a gas at a known fixed pressure while the permeate gas side of the membrane is maintained at very low pressure via a vacuum pump. For this reason at any time, in the permeate gas side, the pressure is always much smaller than that in the feed gas side (usually more than four order of magnitude). Gas permeates through the membrane and is released into the permeate gas side yielding a pressure decrease in the feed gas side. The pressure decrease at the feed side can be measured using three absolute pressure gauges with full-scale readings of 2.2 105 Pa (Pfeiffer APR 262), 103 Pa (Pfeiffer CMR 363) and 105 Pa (Pfeiffer CMR 361). The pressure variation at permeate side can be measured by an absolute pressure gauge with full-scale readings of 105 Pa (Pfeiffer CMR 361) and a 105 Pa-10−7 Pa full range gauge (Pfeiffer PKR 261). Since the volume is calibrated, either the pressure decrease could be converted into an amount of gas in moles permeating through unit area of the membrane (Q(t)) or the rate of pressure decrease could be converted into an amount of gas in moles permeating through unit area of the membrane per second (J(t)). Gas leakage thought the experimental
Corresponding author. E-mail address: [email protected] (D. De Meis).
https://doi.org/10.1016/j.fusengdes.2019.03.012 Received 4 October 2018; Received in revised form 13 December 2018; Accepted 4 March 2019 0920-3796/ © 2019 Elsevier B.V. All rights reserved.
Please cite this article as: Domenico De Meis, et al., Fusion Engineering and Design, https://doi.org/10.1016/j.fusengdes.2019.03.012
Fusion Engineering and Design xxx (xxxx) xxx–xxx
D. De Meis, et al.
apparatus is determined using He gas and a quadrupole mass spectrometer. During the present measurements no leakage from the external air was detected. The tests have been conducted at room temperature. The permeance is then calculated with the following formula:
Pe =
F AΔP
(1) −2
-1
-1
Where, Pe is the Permeance (mol m s Pa ), F is the permeation molar flow (mol s-1), A is the surface area of the membrane (m2), ΔP (=p1-p2) is the pressure difference across the membrane (Pa). The gases studied are H2, He, Ar and air. 3. Membranes tested The six membranes tested have been provided by Ceramiques Tecniques et Industrielles (CTI SA, France). The membranes differ from the top layer pore size and materials: 0.5 nm (hybridsilica), 5 nm (gamma alumina) and 50 nm (zirconia). For each pore size we have a single channel and a multichannel membrane. The ceramic membranes are all 400 mm long. The multi channel (7 channels with 6 mm of the diameter of the channel) has an external diamter of 25 mm. The surface area of the multichannel membranes is: 314 × 6 x 400 × 7 = 005 m2. The single channel has an internal diamter of 6 mm and an external diamter of 10 mm. The surface area of the single channel membrane is: 3.14 × 6 x 400 = 0.0075 m2. We have built the module for the multichannel and a cylindrical metal adapter for the single channel membrane as to use the same module.
Fig. 2. 0.5 nm multichannel CTI SA membrane best fit linearization.
modeled using a parallel Knudsen plus viscous flow model [9]. Determining the best fit line of the experimental values in the plot permeance against (p1+p2) we can find the Knudsen (plus molecular sieving) and viscous contribution (that are respectively the intercept and the slope of the straight line equation) (see Fig. 2 and Annex A). For example, the Knudsen plus molecular sieving permeance for H2 is 1.3 × 10−7 (mol m-2 s-1 Pa-1) (the intercept) while the viscous permeances for H2, expressed in (mol m-2 s-1 Pa-1), is (p1+p2)x4.82 × 10-12 with p1 and p2 in Pa (the slope). Fig. 3 reports the measured gas permeances at room temperature as a function of p1-p2 for the 0.5 nm single channel membrane. Since the selectivity for He and H2 over Ar and air exceeds the Knudsen selectivity, this indicates that molecular sieving is playing an important role in the gas permeances at this pore size. We omit to show the best fit linearization plot for the 0.5 nm single channel membrane and also for the other membranes.
4. Gas permeation tests Permeation tests of the six membranes have been carried out with selected gases at room temperature. We have estimated that the permeances values have maximum 2% of uncertainties. The membranes with pore size of 0.5 nm and 5 nm must be dried under vacuum, or by gas sweeping at high temperature (60–80 °C) during several hours before single gas permeation at room temperature. This was done to eliminate water adsorbed in the membrane pores. 4.1. CTI SA 0.5 nm membranes
4.2. CTI SA 5 nm membranes Fig. 1 reports the measured gas permeances at room temperature as a function of p1-p2 (where p1 is the feed pressure and p2 is the permeate pressure) for the 0.5 nm multichannel membrane. The flux and the permeance through each membrane can be
Figs. 4 and 5 report the measured gas permeances at room temperature as a function of p1-p2 for the 5 nm multichannel and single channel membranes.
Fig. 1. Permeance of 0.5 nm multichannel CTI SA membrane.
Fig. 3. Permeance of 0.5 nm single channel CTI SA membrane. 2
Fusion Engineering and Design xxx (xxxx) xxx–xxx
D. De Meis, et al.
Fig. 4. Permeance of 5 nm CTI multichannel CTI membrane.
Fig. 6. Permeance of 50 nm CTI SA multichannel membrane.
Fig. 5. Permeance of 5 nm single channel CTI SA membrane.
Fig. 7. Permeance of 50 nm single channel CTI SA membrane.
In the multichannel membrane the gas permeance order of He and H2 goes as the inverse molecular weight, consistent with Knudsen diffusion control. In the single channel membrane for all gases the gas permeance order goes as the inverse molecular weight, consistent with Knudsen diffusion control. At higher pressures significant permeance increases are noted for all gases suggesting a viscous flow contribution at this pore size.
Table 1 H2 to gas selectivities for the six membranes (the permeances are obtained from zero pressure intercept); Knudsen theoretical selectivity. Experimental Knudsen selectivity
4.3. CTI SA 50 nm membranes Figs. 6 and 7 report the measured gas permeances at room temperature as a function of p1-p2 for the 50 nm multichannel and single channel membranes. In the 2 membranes (i) both Knudsen and viscous flow contribute to the overall membrane permeance (ii) for all gases the gas permeance order goes as the inverse molecular weight, consistent with Knudsen diffusion control.
Membrane
αH 2He
αH 2Ar
αH 2air
0.5 nm Multi Channel 0.5 nm Single Channel 5 nm Multi Channel 5 nm Single Channel 50 nm Multi Channel 50 nm Single Channel Knudsen theoretical selectivity
1.66 1.70 3.4 1.09 2.59 1.54 1.41
5.78 6.26 6.62 4.29 4.48 3.11 4.47
9.56 13.88 8.2 3.06 4.11 3.42 3.8
that of the multichannel. In fact in the case of the multichannel geometry, the inner channels do not contribute efficiently to the overall gas transport [10].The single channel is expected to perform better than each individual channel in a multichannel membrane. Table 1 depicts H2 to gas selectivities (or separation factors) derived from the following formula:
5. Knudsen separation factors The permeances of the single channel membranes are bigger than 3
Fusion Engineering and Design xxx (xxxx) xxx–xxx
D. De Meis, et al.
αAB =
PeA PeB
1 the permeances were ruled by Knudsen, molecular sieve and viscous diffusion mechanisms; 2 the Knudsen permeances for the membranes are on the order 10−6 10-8 (mol m-2 s-1 Pa-1); 3 molecular sieving is playing an important role in the gas permeances for the 0.5 nm membranes since the selectivity for He and H2 over Ar and air exceeds the Knudsen selectivity; 4 the permeances of the single channel are bigger than that of the multichannel; 5 the selectivities found are low. To overcome this inconvenience various surface membrane modification techniques can be employed to promote other modes of gas transport.
(2)
The Knudsen permeances are obtained from zero pressure intercept. It also reports Knudsen theoretical selectivity derived from the following formula:
αAB = (
MB 1/2 ) MA
(3)
The Knudsen permeation values and the low selectivities found are in line with the literature results for similar membranes [11]. 6. Conclusions Six commercial ceramic membranes made available by Ceramiques Tecniques et Industrielles (CTI SA) France have been characterized in permeation tests in order to measure their permeances at a feed pressure between 10 Pa to 100 kPa at room temperature. We have shown graphically the Knudsen and viscous contribution. We have discovered that:
Acknowledgements The authors thank all colleagues for their help and support. In particular they express their sincere gratitude to Dr. Rich Ciora for his continuous support in this field.
Annex A Let’s consider a porous substrate. Recalling formula (1), the permeance in a porous substrate is:
Pe sub =
Fsub A(p1 − p2 )
In the Knudsen and viscous parallel model [9]:
Pe sub = Pe sub kn + Pe subvisc and Fsub = A Pesub (p1-p2)
Fsub = A(Pe sub kn + Pe subvisc)(p1 − p2) We know that the viscous permeance can be written as:
Pe subvisc =
εηr 2 p1 + p2 8μRTL 2
where Ɛ is the porosity, μ the viscosity (Pa s), η the shape factor, r the pore radius (m), L is the thickness of the membrane (m), R is the gas constant and T is the temperature (K). Let’s pose:
P¯e subvisc =
εηr 2 16μRTL
We have:
Pe subvisc = P¯e subvisc (p1 + p2 ) And finally:
Fsub = P¯e subvisc (p12− p22 ) + Pe subknu (p1 − p2 ) A If we plot the Permeance
(p12−
p22 )
(p1 − p2 )
Fsub A (p1 − p2 )
versus
= (p + p2 ) 1
the slope is P¯e subvisc and the intercept is Pe subknu .
[5] D. De Meis, et al., Microporous inorganic membranes for gas separation and purification, J. Interceram. Int. Ceram. Rev. 67 (4) (2019) 16–21, https://doi.org/10. 1007/s42411-018-0023-2. [6] D. De Meis, et al., Ceramic membranes for the separation of plasma enhancement gases, J. Interceram. Int. Ceram. Rev 67 (6) (2019) 8–13, https://doi.org/10.1007/ s42411-018-0050-z. [7] E. Serra, et al., Hydogen permeation measurements on alumina, J. Am. Ceram. Soc. 88 (January1) (2005) 15–18. [8] E. Serra, et al., Oxygen- and hydrogen-permeation measurements on-mixed conducting SrFeCo0.5Oy ceramic membrane material, Renew. Energy 33 (2008) 241–247 Presented at EMRs 2006.
References [1] T. Nakano, et al., Contribution of Ne ions to radiation enhancement in JT-60U divertor plasmas, J. Nucl. Mater. 438 (2013) S291–S296. [2] M.L. Reinke, et al., Effect of N2, Ne and Ar seeding on Alcator C-Mod H-mode confinement, J. Nucl. Mater. 415 (2011) S340–S344. [3] Y. Igitkhanov, et al., Operational Margins and Impact of Particle Exhaust in DEMO, 4th IAEA DEMO Programme Workshop, KIT, Karlsruhe, 2016, pp. 15–18 Nov. [4] S. Tosti et al., Ceramic membranes for processing plasma enhancement gases, https://doi.org/10.1016/j.fusengdes.2017.01.010.
4
Fusion Engineering and Design xxx (xxxx) xxx–xxx
D. De Meis, et al.
different geometries, J. Membr. Sci. 236 (2004) 101–108. [11] S. Ted Oyama, et al., Review on Mechanisms of gas permeation through inorganic membranes, J. Jpn. Pet. Inst. 54 (5) (2011) 298–309.
[9] A. Caravella, et al., Coupled influence of non ideal diffusion and multilayer asymmetric porous supports on Sieverts law pressure exponent for hydrogen permeation in composite PD-based membranes, Int. J. Hydrogen Energy 39 (2014) 2201–2214. [10] T. Zivkovic, et al., Gas transport efficiency of ceramic membranes:comparison of
5
Contents lists available at ScienceDirect
Fusion Engineering and Design journal homepage: www.elsevier.com/locate/fusengdes
Low pressure fusion exhaust gases separation ⁎
Domenico De Meisa, , Maria Richettab, Emanuele Serrac, Eric Louradourd a
ENEA, Fusion and Nuclear Safety Department, C. R. Frascati, Via E. Fermi 45, 00044, Frascati (Roma), Italy University of Rome Tor Vergata, Department of Industrial Engineering, via del Politecnico 1, 00133, Roma, Italy ENEA, Department for Sustainability, C.R. Casaccia, Via Anguillarese 301, 00123, Santa Maria di Galeria (Roma), Italy d CTI-Céramiques Techniques Industrielles, 382 Avenue du Moulinas, F-30340, Salindres, France b c
A R T I C LE I N FO
A B S T R A C T
Keywords: Gas selectivity Inorganic porous membranes Ceramic membranes Tokamak demonstration reactor (DEMO) Plasma enhancement gases (PEG) Permeance
Plasma enhancement gases (PEGs) (such as: nitrogen, neon, argon and other inert gases) are injected into the plasma of several tokamaks in order to reduce the power load over the plasma facing components. The exhaust gas in the tokamak demonstration reactor (DEMO) consists of more than 90% of unburned fuel gas (D and T) and the remaining part will be He, PEGs and impurities. In DEMO reactor it is foreseen to recover the fuel gas and PEGs. The research focuses to remove the He from fuel gas and PEGs from fusion reactor. For this purpose, six commercial ceramic membranes have been tested at a permeation apparatus built at ENEA Casaccia laboratories. Single gas permeances for H2, He, Ar and air were measured with a pressure drop across the membranes between 10 Pa up to 100 kPa at room temperature. The selectivities found are low except for the 0.5 nm pore size membranes. The membranes have been supplied by Ceramiques Tecniques et Industrielles (CTI SA, France).
1. Introduction Plasma Enhancement Gases (PEGs) are injected into the tokamak to convert the plasma thermal energy to ultraviolet and soft X-ray radiation. Possible PEGs gases are: nitrogen, neon, argon, xenon and other inert gases. DEMO’s plasma exhaust will contain the impurities coming from the interactions of the plasma with the surfaces, PEGs gases, not burned deuterium and tritium and alfa particles [1–3]. For the DEMO exhaust processing system, the use of inorganic membranes has been recently taken into consideration because of their reduced energy consumption, continuos operation, temperature and radiation stability, modularity and scale-up [4–6]. In the present study six commercial ceramic membranes have been characterized in permeation tests in order to measure their permeances at a feed pressure between 10 Pa to 100 kPa at room temperature. The membranes tested have been given by CTI SA France. The tests have been performed at Dr. Emanuele Serra permeation laboratory in ENEA Casaccia. 2. Experimental set-up The method chosen for the determination of the gas permeance through the membranes is a gas phase technique. The permeation apparatus is constructed using standard stainless steel Ultra High Vacuum (UHV) components. It is described in the references [6–8]. The module
⁎
is made of stainless steel. The two ends of the module are removable to allow for membrane replacement. Gas tightness between the module and the membrane at each end is maintained by 2 fluoroelastomer orings (Viton®). The measurements were conducted using a gas pressure from 10 Pa to 100 kPa. After evacuating the apparatus to an ultra high vacuum so that both sides of the membrane are initially in contact with vacuum, the feed gas side is exposed to a gas at a known fixed pressure while the permeate gas side of the membrane is maintained at very low pressure via a vacuum pump. For this reason at any time, in the permeate gas side, the pressure is always much smaller than that in the feed gas side (usually more than four order of magnitude). Gas permeates through the membrane and is released into the permeate gas side yielding a pressure decrease in the feed gas side. The pressure decrease at the feed side can be measured using three absolute pressure gauges with full-scale readings of 2.2 105 Pa (Pfeiffer APR 262), 103 Pa (Pfeiffer CMR 363) and 105 Pa (Pfeiffer CMR 361). The pressure variation at permeate side can be measured by an absolute pressure gauge with full-scale readings of 105 Pa (Pfeiffer CMR 361) and a 105 Pa-10−7 Pa full range gauge (Pfeiffer PKR 261). Since the volume is calibrated, either the pressure decrease could be converted into an amount of gas in moles permeating through unit area of the membrane (Q(t)) or the rate of pressure decrease could be converted into an amount of gas in moles permeating through unit area of the membrane per second (J(t)). Gas leakage thought the experimental
Corresponding author. E-mail address: [email protected] (D. De Meis).
https://doi.org/10.1016/j.fusengdes.2019.03.012 Received 4 October 2018; Received in revised form 13 December 2018; Accepted 4 March 2019 0920-3796/ © 2019 Elsevier B.V. All rights reserved.
Please cite this article as: Domenico De Meis, et al., Fusion Engineering and Design, https://doi.org/10.1016/j.fusengdes.2019.03.012
Fusion Engineering and Design xxx (xxxx) xxx–xxx
D. De Meis, et al.
apparatus is determined using He gas and a quadrupole mass spectrometer. During the present measurements no leakage from the external air was detected. The tests have been conducted at room temperature. The permeance is then calculated with the following formula:
Pe =
F AΔP
(1) −2
-1
-1
Where, Pe is the Permeance (mol m s Pa ), F is the permeation molar flow (mol s-1), A is the surface area of the membrane (m2), ΔP (=p1-p2) is the pressure difference across the membrane (Pa). The gases studied are H2, He, Ar and air. 3. Membranes tested The six membranes tested have been provided by Ceramiques Tecniques et Industrielles (CTI SA, France). The membranes differ from the top layer pore size and materials: 0.5 nm (hybridsilica), 5 nm (gamma alumina) and 50 nm (zirconia). For each pore size we have a single channel and a multichannel membrane. The ceramic membranes are all 400 mm long. The multi channel (7 channels with 6 mm of the diameter of the channel) has an external diamter of 25 mm. The surface area of the multichannel membranes is: 314 × 6 x 400 × 7 = 005 m2. The single channel has an internal diamter of 6 mm and an external diamter of 10 mm. The surface area of the single channel membrane is: 3.14 × 6 x 400 = 0.0075 m2. We have built the module for the multichannel and a cylindrical metal adapter for the single channel membrane as to use the same module.
Fig. 2. 0.5 nm multichannel CTI SA membrane best fit linearization.
modeled using a parallel Knudsen plus viscous flow model [9]. Determining the best fit line of the experimental values in the plot permeance against (p1+p2) we can find the Knudsen (plus molecular sieving) and viscous contribution (that are respectively the intercept and the slope of the straight line equation) (see Fig. 2 and Annex A). For example, the Knudsen plus molecular sieving permeance for H2 is 1.3 × 10−7 (mol m-2 s-1 Pa-1) (the intercept) while the viscous permeances for H2, expressed in (mol m-2 s-1 Pa-1), is (p1+p2)x4.82 × 10-12 with p1 and p2 in Pa (the slope). Fig. 3 reports the measured gas permeances at room temperature as a function of p1-p2 for the 0.5 nm single channel membrane. Since the selectivity for He and H2 over Ar and air exceeds the Knudsen selectivity, this indicates that molecular sieving is playing an important role in the gas permeances at this pore size. We omit to show the best fit linearization plot for the 0.5 nm single channel membrane and also for the other membranes.
4. Gas permeation tests Permeation tests of the six membranes have been carried out with selected gases at room temperature. We have estimated that the permeances values have maximum 2% of uncertainties. The membranes with pore size of 0.5 nm and 5 nm must be dried under vacuum, or by gas sweeping at high temperature (60–80 °C) during several hours before single gas permeation at room temperature. This was done to eliminate water adsorbed in the membrane pores. 4.1. CTI SA 0.5 nm membranes
4.2. CTI SA 5 nm membranes Fig. 1 reports the measured gas permeances at room temperature as a function of p1-p2 (where p1 is the feed pressure and p2 is the permeate pressure) for the 0.5 nm multichannel membrane. The flux and the permeance through each membrane can be
Figs. 4 and 5 report the measured gas permeances at room temperature as a function of p1-p2 for the 5 nm multichannel and single channel membranes.
Fig. 1. Permeance of 0.5 nm multichannel CTI SA membrane.
Fig. 3. Permeance of 0.5 nm single channel CTI SA membrane. 2
Fusion Engineering and Design xxx (xxxx) xxx–xxx
D. De Meis, et al.
Fig. 4. Permeance of 5 nm CTI multichannel CTI membrane.
Fig. 6. Permeance of 50 nm CTI SA multichannel membrane.
Fig. 5. Permeance of 5 nm single channel CTI SA membrane.
Fig. 7. Permeance of 50 nm single channel CTI SA membrane.
In the multichannel membrane the gas permeance order of He and H2 goes as the inverse molecular weight, consistent with Knudsen diffusion control. In the single channel membrane for all gases the gas permeance order goes as the inverse molecular weight, consistent with Knudsen diffusion control. At higher pressures significant permeance increases are noted for all gases suggesting a viscous flow contribution at this pore size.
Table 1 H2 to gas selectivities for the six membranes (the permeances are obtained from zero pressure intercept); Knudsen theoretical selectivity. Experimental Knudsen selectivity
4.3. CTI SA 50 nm membranes Figs. 6 and 7 report the measured gas permeances at room temperature as a function of p1-p2 for the 50 nm multichannel and single channel membranes. In the 2 membranes (i) both Knudsen and viscous flow contribute to the overall membrane permeance (ii) for all gases the gas permeance order goes as the inverse molecular weight, consistent with Knudsen diffusion control.
Membrane
αH 2He
αH 2Ar
αH 2air
0.5 nm Multi Channel 0.5 nm Single Channel 5 nm Multi Channel 5 nm Single Channel 50 nm Multi Channel 50 nm Single Channel Knudsen theoretical selectivity
1.66 1.70 3.4 1.09 2.59 1.54 1.41
5.78 6.26 6.62 4.29 4.48 3.11 4.47
9.56 13.88 8.2 3.06 4.11 3.42 3.8
that of the multichannel. In fact in the case of the multichannel geometry, the inner channels do not contribute efficiently to the overall gas transport [10].The single channel is expected to perform better than each individual channel in a multichannel membrane. Table 1 depicts H2 to gas selectivities (or separation factors) derived from the following formula:
5. Knudsen separation factors The permeances of the single channel membranes are bigger than 3
Fusion Engineering and Design xxx (xxxx) xxx–xxx
D. De Meis, et al.
αAB =
PeA PeB
1 the permeances were ruled by Knudsen, molecular sieve and viscous diffusion mechanisms; 2 the Knudsen permeances for the membranes are on the order 10−6 10-8 (mol m-2 s-1 Pa-1); 3 molecular sieving is playing an important role in the gas permeances for the 0.5 nm membranes since the selectivity for He and H2 over Ar and air exceeds the Knudsen selectivity; 4 the permeances of the single channel are bigger than that of the multichannel; 5 the selectivities found are low. To overcome this inconvenience various surface membrane modification techniques can be employed to promote other modes of gas transport.
(2)
The Knudsen permeances are obtained from zero pressure intercept. It also reports Knudsen theoretical selectivity derived from the following formula:
αAB = (
MB 1/2 ) MA
(3)
The Knudsen permeation values and the low selectivities found are in line with the literature results for similar membranes [11]. 6. Conclusions Six commercial ceramic membranes made available by Ceramiques Tecniques et Industrielles (CTI SA) France have been characterized in permeation tests in order to measure their permeances at a feed pressure between 10 Pa to 100 kPa at room temperature. We have shown graphically the Knudsen and viscous contribution. We have discovered that:
Acknowledgements The authors thank all colleagues for their help and support. In particular they express their sincere gratitude to Dr. Rich Ciora for his continuous support in this field.
Annex A Let’s consider a porous substrate. Recalling formula (1), the permeance in a porous substrate is:
Pe sub =
Fsub A(p1 − p2 )
In the Knudsen and viscous parallel model [9]:
Pe sub = Pe sub kn + Pe subvisc and Fsub = A Pesub (p1-p2)
Fsub = A(Pe sub kn + Pe subvisc)(p1 − p2) We know that the viscous permeance can be written as:
Pe subvisc =
εηr 2 p1 + p2 8μRTL 2
where Ɛ is the porosity, μ the viscosity (Pa s), η the shape factor, r the pore radius (m), L is the thickness of the membrane (m), R is the gas constant and T is the temperature (K). Let’s pose:
P¯e subvisc =
εηr 2 16μRTL
We have:
Pe subvisc = P¯e subvisc (p1 + p2 ) And finally:
Fsub = P¯e subvisc (p12− p22 ) + Pe subknu (p1 − p2 ) A If we plot the Permeance
(p12−
p22 )
(p1 − p2 )
Fsub A (p1 − p2 )
versus
= (p + p2 ) 1
the slope is P¯e subvisc and the intercept is Pe subknu .
[5] D. De Meis, et al., Microporous inorganic membranes for gas separation and purification, J. Interceram. Int. Ceram. Rev. 67 (4) (2019) 16–21, https://doi.org/10. 1007/s42411-018-0023-2. [6] D. De Meis, et al., Ceramic membranes for the separation of plasma enhancement gases, J. Interceram. Int. Ceram. Rev 67 (6) (2019) 8–13, https://doi.org/10.1007/ s42411-018-0050-z. [7] E. Serra, et al., Hydogen permeation measurements on alumina, J. Am. Ceram. Soc. 88 (January1) (2005) 15–18. [8] E. Serra, et al., Oxygen- and hydrogen-permeation measurements on-mixed conducting SrFeCo0.5Oy ceramic membrane material, Renew. Energy 33 (2008) 241–247 Presented at EMRs 2006.
References [1] T. Nakano, et al., Contribution of Ne ions to radiation enhancement in JT-60U divertor plasmas, J. Nucl. Mater. 438 (2013) S291–S296. [2] M.L. Reinke, et al., Effect of N2, Ne and Ar seeding on Alcator C-Mod H-mode confinement, J. Nucl. Mater. 415 (2011) S340–S344. [3] Y. Igitkhanov, et al., Operational Margins and Impact of Particle Exhaust in DEMO, 4th IAEA DEMO Programme Workshop, KIT, Karlsruhe, 2016, pp. 15–18 Nov. [4] S. Tosti et al., Ceramic membranes for processing plasma enhancement gases, https://doi.org/10.1016/j.fusengdes.2017.01.010.
4
Fusion Engineering and Design xxx (xxxx) xxx–xxx
D. De Meis, et al.
different geometries, J. Membr. Sci. 236 (2004) 101–108. [11] S. Ted Oyama, et al., Review on Mechanisms of gas permeation through inorganic membranes, J. Jpn. Pet. Inst. 54 (5) (2011) 298–309.
[9] A. Caravella, et al., Coupled influence of non ideal diffusion and multilayer asymmetric porous supports on Sieverts law pressure exponent for hydrogen permeation in composite PD-based membranes, Int. J. Hydrogen Energy 39 (2014) 2201–2214. [10] T. Zivkovic, et al., Gas transport efficiency of ceramic membranes:comparison of
5