- Email: [email protected]
New experimental setup for continuous mass flux measurement in pervaporation
Accepted Manuscript Title: New experimental setup for continuous mass flux measurement in pervaporation Authors: S.A. Toudji, J.-P. Bonnet, J.-L. Gardarein, E. Carretier PII: DOI: Reference:
S0263-8762(17)30358-1 http://dx.doi.org/doi:10.1016/j.cherd.2017.06.029 CHERD 2732
To appear in: Received date: Revised date: Accepted date:
19-7-2016 19-6-2017 21-6-2017
Please cite this article as: Toudji, S.A., Bonnet, J.-P., Gardarein, J.-L., Carretier, E., New experimental setup for continuous mass flux measurement in pervaporation.Chemical Engineering Research and Design http://dx.doi.org/10.1016/j.cherd.2017.06.029 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
New experimental setup for continuous mass flux measurement in pervaporation S. A. TOUDJI a, J-P. BONNET a,*, J-L GARDAREIN b, E. CARRETIER a a b
Aix Marseille Université, CNRS, Centrale Marseille, M2P2 UMR 7340, 13451, Marseille, France Aix Marseille Université, CNRS, IUSTI UMR 7343, 13453, Marseille, France
* Corresponding author. E-mail address: [email protected]
Highlights: New experimental setup was designed for continuous permeate mass flux measurement in pervaporation Mass flux measured with pressure sensor is validated by comparing to results obtained from conventional gravimetric method Relative high frequency of mass flux measurement allows to consider coupling with heat flux measurement in unsteady state
Abstract Pervaporation is a separation process of liquid mixtures through a thin non-porous membrane. In vacuum pervaporation, the global mass flux is classically estimated by weighing the mass of permeate collected in cold traps. In this work, we propose a new experimental setup that allows a continuous measurement of the mass flux. The new mass flux method measurement was validated for single component permeation (ethanol and water) by comparing mass of permeate collected in cold traps with the level decrease of feed liquid measured with a pressure sensor. This new setup can be useful for laboratory studies dealing with the evolution of mass flux according to different parameters of the process as, for example, the permeate side pressure level or temperature of liquid feed. Keywords: Pervaporation, Unsteady state measurement, Mass flux
1
Introduction Pervaporation is used for the separation of liquid mixtures through a non-porous and
selective membrane. In vacuum pervaporation or sweeping gas pervaporation, the driving force for mass transport is the difference of chemical potential between the liquid feed and the permeate [1]. In the first case the permeate side is kept at reduced pressure using a vacuum 1
pump while in the second case an inert gas sweeps the permeate side [2-4]. Pervaporation was initially developed for the dehydration of alcohol and especially for azeotropic systems [5]. Due to the potential of pervaporation as an energy saving and green process, it is combined with other processes as distillation forming hybrid-processes which are currently used in various applications [6] like: -
Dehydration of solvents [7].
-
Recovery of solvents from aqueous solutions [8-10] or from fermentation broth [11-12] mainly ABE fermentation [13].
-
Separation of organic mixtures [14-15].
Pervaporation differs from other membrane processes because it includes a phase change as it goes from a liquid phase in the feed side into a vapor phase in the permeate side [1]. Generally, in vacuum pervaporation, the permeate is condensed in cold trap and the mass flux J [kg.h-1.m-2] is calculated as the mass of permeate weighed in the cold trap divided by the membrane surface and the time spent to collect this amount of permeate [7]. This method is efficient to have an estimation of the global mass flux but it is unable to provide the mass flux versus the operating time with a small time span. Moreover, the trap change action can be considered as a perturbation for the process. In this work, an original experimental setup has been designed to measure continuously the mass flux and without permeate pressure perturbation. A differential pressure sensor placed in the feed tank provides the hydrostatic pressure with a 10 Hz frequency. The hydrostatic pressure data allow the evolution of the permeate mass flux to be presented at a relatively high frequency. The setup can be used in the classical types of laboratories pervaporation plants. As the amount of permeate is measured thanks to the variation of hydrostatic pressure in the liquid feed side, the permeate side conditions, vacuum or sweeping, should not affect the measurement.
2
2
Experimental setup 6.5
12
TC
25 C Legend: Polyurethane insulation PVC Liquid feed Thermostated enclosure Permeate side (reduced pressure) Membrane + spacer + support Distance and dimensions are expressed in mm
2 b h
11
p = ρ.g.h
150
TT3
0-10 V 0-20 mA 0-10 V
TT2
PT1
0-100 C 0-3 kPa 0-10 kPa
140
8
9
30
2a 150
10
4.a 6 PT2 5 75
42
TT1
4.b 1 3
7
1. Membrane + spacer + support ; 2. Liquid feed tank ; 3. Permeate side ; 4.a,b-5. Cold trap ; 6-8. Pressure sensor ; 7. Vacuum pump ; 9. Data acquisition system ; 10. Computer ; TT1, TT2, TT3. Temperature transmitters
Figure 1. Schematic view of the pervaporation setup
Pervaporation experiments were performed with the setup shown schematically in Figure 1. A commercial PDMS flat sheet membrane from Pervatech is used (1). The total membrane area is 14 cm². To ensure the contact surface between feed liquid and membrane is the same as the contact surface between permeate and membrane, a spacer is placed between the membrane and the stainless steel support. The membrane is supported by the stainless steel support (1 mm thickness) perforated with 25 holes of 6 mm diameter well spatially distributed on the surface (Figure 2).
3
Figure 2. Schematic view of the spacer used between the membrane and the support
During the experiments, the permeate side (3) is continuously maintained at reduced pressure (around 300 Pa) with a vacuum pump (7) ILMVAC-MP 601 Tp. The permeate pressure is continuously monitored with a pressure sensor (6) BROOKS-CMCA that have a vacuum measurement range from 5 to 104 Pa with an accuracy of 0.5 % of full scale (50 Pa). The permeate is condensed in glass cold trap (4.a – 4.b & 5) cooled with liquid nitrogen. It is collected periodically. At the end of the experiment, the three traps (4.a – 4.b & 5) are disconnected from the experimental device in order to be weighed individually. The mass of the empty traps being perfectly known, the permeate mass collected in each trap (∆MB 4a, ∆MB 4b) is calculated by subtracting the mass of the trap after and before permeation. The traps are weighed using a microbalance METTLER TOLEDO-XS1003S with 10 mg precision. Mass collected in security trap (5) is always nil. Pervaporation cell is made of PVC matter with a low thermal conductivity of about 0.16 W.m1
.K-1 to limit heat exchanges with the external environment and with the liquid. The liquid is
filled in both cylindrical tanks (2a) and (2b) above the membrane surface. The tank (2b) is designed with an inner diameter around 6.44 mm to have 3 Pa of differential pressure for 10 mg of permeated liquid (Eq. 1). The total initial volume is 160 mL. A differential pressure sensor (8) Keller PR41X-31 with a measurement range from 0 to 3000 Pa measures the hydrostatic liquid pressure. Using the hydrostatic relation (Eq. 1), the permeate amount could be calculated assuming that evaporation at the top edge of the feed tank is negligible and so the level drop is almost involved by the only permeation phenomenon.
P12
4g M12 D(22 b )
4
Eq. 1
A calibration curve (Figure 3) has been done to limit the error due to the geometric irregularities of the cylindrical tank 2b. Indeed, part 2b was machined in the laboratory with basic machining equipment and D2b, the inner diameter, is not perfectly constant all along the cylinder length. The hydrostatic pressure of liquid has been measured for different volumes added with a micropipettor at 25 °C (corresponding masses are calculated at the temperature considered). A quadratic model has been used to fit the calibration curves. 24 23
Calibration mass, M [g]
22 21 20 19 18 17 16
Experimental for ethanol Quadratic fit for ethanol
15
Experimental for water Quadratic fit for water
14 200
400
600
800
1000
1200
1400
1600
1800
Liquid hydrostatic pressure, P [Pa]
Figure 3. Hydrostatic pressure calibration for pure ethanol and pure water
The global precision of the pressure sensor (8) including the electric noise is 0.5% of the full scale (±15 Pa) that corresponds to ±0.10 g. The mass differences ∆Mp between an initial and final time (ti, tf) can be compared with ∆MB (Eq. 2). The pressure sensor precision (±0.10 g) is sufficient for the quantities of permeate collected in the cold trap varying between 1.5 – 4 g.
M B M B 4a M B 4b
Eq. 2
The entire cell device is insulated using polyurethane foam. All the measurements are simultaneously monitored (10 Hz) using NIcDAQ 9174 central data acquisition (9) connected to a computer (10). The experimental setup is placed inside a thermostated enclosure to control the environment temperature. NAHITA 639 series thermostated enclosure was used with 250 L capacity and 5 °C to 60 °C temperature range with temperature resolution of 0.1 °C. The temperature inside the thermostated enclosure during experiments was maintained at 25 °C.
5
Three experiments were conducted for both pure ethanol and pure water. The experiments last around 1 hour long for ethanol and between 4 – 6 hours for water, because the mass flux of pure water with the organophilic membrane used is very low. For ethanol experiments, the composition of permeate is deduced from density measurements. The density is measured using Anton Paar DMA 5000 M density meter with an accuracy of 5.10-6 g.cm-³ for a temperature accuracy of 0.01 °C. The weight fraction of water in permeate is estimated from a calibration curve relating the weight fraction of ethanol (Wethanol) to the density. The calibration curve has been validated by a chart from Perry's Chemical Engineers Handbook [16].
3
Results and discussion
3.1 Ethanol Results 1300
1800 1300
1200
900 800 700 600 500 400 300 200
1600
1500
0
500
1225
1200
1175
1150
0
50
100
1400
150
200
250
300
350
Time, t [s]
Vacuum pump « on »
1300
1200
Vacuum 1100pump "off" Vacuum pump "on"
100
1250
t4.a → t4.b
Liquid hydrostatic pressure, P [Pa]
Liquid hydrostatic pressure, P [Pa]
1000
0
Liquid hydrostatic pressure, P [Pa]
1275
1700
1100
1000 0 1000
Vacuum pump "off" Vacuum pump "on"
2000 1500
20004000 2500 60003000 8000 3500
Time, t [s]
10000 4000
12000
Time, t [s]
Figure 4. Liquid hydrostatic pressure during pervaporation of pure ethanol (experiment 3)
The feed hydrostatic pressure during the pervaporation of ethanol experiments has been measured and the results are shown on Figure 4. The first 300 s of acquisition shows that the pressure signal is almost constant before the vacuum pump is switched on. It indicates that there is no significant ethanol leak and no significant pervaporation phenomenon. At t=300 s, 6
14000
the vacuum pump is switched on and the pervaporation phenomenon starts. As a result, the liquid level decreases due to its permeation through the membrane and so the hydrostatic pressure decreases. The first shift in pressure (from 1250 to 1150 Pa) is due to the mechanical deformation of the metallic support and so to the deformation of the membrane due to vacuum generated by the pump. We can visually detect that there is a shift in the slopes dP/dt at the time t4.a4.b = 900 s that corresponds to the moment when the valves are switched to collect the permeate from trap 4.a to trap 4.b. Only the data ranging from 300 s to 3600 s (red data in Figure 4) are used. The calibration curve (Figure 3) is used to convert the pressure data into liquid mass data that is contained in the feed tank above the pressure sensor (Figure 5). 18
Mi 17.5
17
Mass, M [g]
16.5
16
15.5
15
Mf
14.5
14
0
500
ti
1000
1500
2000
Time, t [s]
2500
3000
3500
4000
tf
Figure 5. Mass of liquid above the pressure sensor during pervaporation of pure ethanol (experiment 3)
The mass differences ∆MP between the initial and final time (ti, tf) are reported in Table 2. Density measurements show differences between liquid feed and liquid permeate. This can be attributed to water presence in permeate due to humid air leaks that enters in the permeate side system which is at reduced pressure. That phenomenon is frequently encountered in pervaporation [17-18]. Density and composition of permeate are presented in Table 2.
The water mass flow rate qw, defined as Mw/∆t represents the water supplied by humid air leak per time unit, it is estimated around 3.10-5 g.s-1. Hence the mass of permeate weighed in 7
the cold trap is corrected by taking into account the water to estimate the mass of pure ethanol ∆MB
ethanol
to be compared with ∆MP. ∆MB
ethanol
is systematically smaller than the mass
measured by the pressure sensor ∆MP. The deviation D (g), defined as ∆MB ethanol - ∆MP, is a negative value and varies between 0.07 – 0.15 g which correspond to 2 – 5 % of ∆MP. Although this deviation is very low, it can be explained by the evaporation of ethanol on the feed side, near the edge ethanol / air in tank (2b). To estimate the ethanol evaporation rate, several data acquisitions during five hours have been performed with an impervious film in the membrane place. Although dependent on the position of ethanol / air interface in the tank 2b (Figure 1), the average rate of evaporation is about 3.10-5 g.s-1.
3.2 Water results 1800 1775
Liquid hydrostatic pressure, P [Pa]
Liquid hydrostatic pressure, P [Pa]
1700
1600
1500
1750
1725
1700
1675
1650
1400
0
50
100
150
200
250
300
350
Time, t [s]
Vacuum pump « on »
1300
1200
1100
1000
Vacuum pump "off" Vacuum pump "on" 0
2000
4000
6000
8000
10000
12000
14000
Time, t [s]
Figure 6. Liquid hydrostatic pressure during pervaporation of pure water (experiment 3)
The feed hydrostatic pressure during the pervaporation of water experiments has been measured and the results are shown on Figure 6. The first 300 s of acquisition show that the pressure signal is constant before the vacuum pump is switched on. It validates that the water level is constant and so that there is no water leak, no pervaporation and no significant evaporation. At t=300 s, the vacuum pump is switched on and the pervaporation phenomenon 8
starts. As in ethanol case, the first shift in pressure (from 1740 to 1680 Pa) is due to the mechanical deformation of the membrane. The calibration curve (Figure 3) is used to convert the hydrostatic pressure into the mass of liquid contained in the feed tank above the pressure sensor (Figure 7). 23.5
23
Mi
Mass, M [g]
22.5
22
21.5
Mf
21
20.5
0
2000
ti
4000
6000
8000
Time, t [s]
10000
12000
14000
tf
Figure 7. Mass of liquid above the pressure sensor during pervaporation of pure water (experiment 3)
The mass of permeate weighed in the cold trap is systematically greater than the mass measured by the pressure sensor (Table 3). The deviation D (g) is a positive value and increases with the experiment duration. The deviation D (g) divided by the experiments duration gives a water flow rate qw in the same order of magnitude for the three experiments, around 1.10-5 g.s-1, as in the ethanol case (Table 2). So, the water excess in the trap could be attributed to humid air leaks on the permeate side. In this specific case of low permeate fluxes on long duration, the use of a pressure sensor is a better way than the classical gravimetric method because it permit to eliminate the water due to humid air leak.
9
500
3.5
450
3
400
2.5
350
2
300
1.5
250
1
200
p
4
Permeate pressure, P [Pa]
Mass flux, J [kg.h-1.m-2]
3.3 Mass flux estimation
0.5
Mass flux
t4.a → t4.b 0
0
300
600
900
150
Permeate pressure
100 1200 1500 1800 2100 2400 2700 3000 3300 3600
Time, t [s]
Figure 8. Mass flux during pervaporation of pure ethanol (experiment 3)
The very good agreement between mass of permeate measured with the pressure sensor and mass weighed in the cold trap allows to estimate the permeate mass flux using the pressure sensor (Figure 8). In this purpose, a sliding window is used to measure the slope from the mass versus time curve (Figure 5). The window size is 150 s and it scans data from 300 s to 900 s (trap changed), then from 900 s to 3600 s (vacuum pump switched off). As expected, the mass flux increases during first moments of permeation (300 s – 420 s) and reaches a maximum of 3.84 kg.h-1.m-2 when the permeate pressure stabilizes around 190 Pa. For time ranging from 420 s to 900 s, although the permeate pressure is constant, the mass flux decreases due to the liquid temperature drop (close to membrane) [19-20]. Indeed, this temperature reduction implies a reduction of energy available in the liquid leading to a decrease in the mass transport. At t=900 s, the trap changing involves an increase in permeate pressure from 190 Pa to 270 Pa, caused probably by more important air leak on the device with the second trap. The permeate mass flux decreases when the permeate pressure increases. From 1500 s to 3600 s (the end of experiment) the mass flux stabilizes around 2.4 kg.h-1.m-2 as a pseudosteady state seems to be reached. This value is in accordance with permeate mass fluxes reported in literature for pervaporation process [21-25].
10
4
Conclusion: An original and innovative experimental setup has been developed to measure
continuously the mass flux in pervaporation laboratory experiments. The conventional gravimetric method used to measure mass flux in pervaporation serves as reference in this work. This technique needs relatively long period of experiment to have a sufficient mass of permeate to be weighed. Moreover, this technique provides only a global estimation of the mass flux. The conventional value of collected permeate in the trap is compared with the mass of liquid, measured with a pressure sensor, that has left the feed tank. These two masses are supposed to be the same assuming leaks and mass accumulation term in the membrane close to zero. We compare experimentally these two quantities, they are almost the same for all tested conditions (ethanol and water). These results valid the new measurement setup to estimate the permeate mass flux. Moreover, since the pressure sensor is also a transmitter, this setup allows to accede at continuous measurement of the permeate mass flux with relatively high frequency in comparison with the conventional method. This approach opens new perspectives to investigate, with a non-intrusive measurement, the non-stationary regime in the pervaporation process. Works are in progress for: (1) measurement of the temperature field upstream the membrane in order to correlate it with the mass flux; (2) the use of this new technique in liquid binary systems; the experimental setup need to be modified by adding an on-line analyzer (spectrometer, refractometer, …) to allow measurement for mixtures with the same precision as with pure components.
11
References: [1] J.G. Crespo, C. Brazinha, 1 - Fundamentals of pervaporation, in: Pervaporation, Vapour Permeation and Membrane Distillation, Woodhead Publishing, Oxford, 2015, pp. 3-17. [2] C. Brazinha, V.D. Alves, R.M.C. Viegas, J.G. Crespo, Aroma recovery by integration of sweeping gas pervaporation and liquid absorption in membrane contactors, Separation and Purification Technology, 70 (2009) 103-111. [3] C. Vallieres, E. Favre, Vacuum versus sweeping gas operation for binary mixtures separation by dense membrane processes, Journal of Membrane Science, 244 (2004) 17-23. [4] F. Lipnizki, R.W. Field, Integration of vacuum and sweep gas pervaporation to recover organic compounds from wastewater, Separation and Purification Technology, 22–23 (2001) 347-360. [5] A. Jonquières, R. Clément, P. Lochon, J. Néel, M. Dresch, B. Chrétien, Industrial state-of-the-art of pervaporation and vapour permeation in the western countries, Journal of Membrane Science, 206 (2002) 87-117. [6] S.C. George, S. Thomas, Transport phenomena through polymeric systems, Progress in Polymer Science, 26 (2001) 985-1017. [7] P.D. Chapman, T. Oliveira, A.G. Livingston, K. Li, Membranes for the dehydration of solvents by pervaporation, Journal of Membrane Science, 318 (2008) 5-37. [8] T. Uragami, Y. Matsuoka, T. Miyata, Permeation and separation characteristics in removal of dilute volatile organic compounds from aqueous solutions through copolymer membranes consisted of poly(styrene) and poly(dimethylsiloxane) containing a hydrophobic ionic liquid by pervaporation, Journal of Membrane Science, 506 (2016) 109-118. [9] S.B. Kuila, S.K. Ray, Separation of isopropyl alcohol–water mixtures by pervaporation using copolymer membrane: Analysis of sorption and permeation, Chemical Engineering Research and Design, 91 (2013) 377-388. [10] A.J. Toth, P. Mizsey, Methanol removal from aqueous mixture with organophilic pervaporation: Experiments and modelling, Chemical Engineering Research and Design, 98 (2015) 123-135. [11] M. Moussa, V. Athès, Y. Imbert, I. Souchon, O. Vitrac, M.L. Lameloise, Pervaporative Dehydration of Bioethanol using Silica and PVA Membranes: Analysis of Permeation Performances and Effect of Volatile Organic Impurities, Procedia Engineering, 44 (2012) 1173-1176. [12] E. Nagy, P. Mizsey, J. Hancsók, S. Boldyryev, P. Varbanov, Analysis of energy saving by combination of distillation and pervaporation for biofuel production, Chemical Engineering and Processing: Process Intensification, 98 (2015) 86-94. [13] M.F.S. Dubreuil, P. Vandezande, W.H.S. Van Hecke, W.J. Porto-Carrero, C.T.E. Dotremont, Study on ageing/fouling phenomena and the effect of upstream nanofiltration on in-situ product recovery of n-butanol through poly[1-(trimethylsilyl)-1-propyne] pervaporation membranes, Journal of Membrane Science, 447 (2013) 134-143. [14] P. Luis, J. Degrève, B.V. der Bruggen, Separation of methanol–n-butyl acetate mixtures by pervaporation: Potential of 10 commercial membranes, Journal of Membrane Science, 429 (2013) 112. [15] B. Smitha, D. Suhanya, S. Sridhar, M. Ramakrishna, Separation of organic–organic mixtures by pervaporation—a review, Journal of Membrane Science, 241 (2004) 1-21. [16] D.G. Robert H. Perry, Perry's Chemical Engineers' Handbook, Seventh Edition ed. [17] P.T. Sumesh, P.K. Bhattacharya, Analysis of phase change during pervaporation with single component permeation, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 290 (2006) 263-272. [18] C. Vallieres, E. Favre, D. Roizard, J. Bindelle, D. Sacco, New Insights into Pervaporation Mass Transport under Increasing Downstream Pressure Conditions: Critical Role of Inert Gas Entrance, Industrial & Engineering Chemistry Research, 40 (2001) 1559-1565. [19] A. Ito, Y. Feng, H. Sasaki, Temperature drop of feed liquid during pervaporation, Journal of Membrane Science, 133 (1997) 95-102. [20] E. Favre, Temperature polarization in pervaporation, Desalination, 154 (2003) 129-138. 12
[21] I. Blume, J.G. Wijmans, R.W. Baker, The separation of dissolved organics from water by pervaporation, Journal of Membrane Science, 49 (1990) 253-286. [22] X. Lin, H. Kita, K.-i. Okamoto, Silicalite Membrane Preparation, Characterization, and Separation Performance, Industrial & Engineering Chemistry Research, 40 (2001) 9. [23] Y.K. Ong, G.M. Shi, N.L. Le, Y.P. Tang, J. Zuo, S.P. Nunes, T.-S. Chung, Recent membrane development for pervaporation processes, Progress in Polymer Science, 57 (2016) 1-31. [24] X. Zhan, J.-d. Li, J.-q. Huang, C.-x. Chen, PERVAPORATION PROPERTIES OF PDMS MEMBRANES CURED WITH DIFFERENT CROSS-LINKING REAGENTS FOR ETHANOL CONCENTRATION FROM AQUEOUS SOLUTIONS, Chinese Journal of Polymer Science, 27 (2009) 533-542. [25] S. Schnabel, P. Moulin, Q.T. Nguyen, D. Roizard, P. Aptel, Removal of volatile organic components (VOCs) from water by pervaporation: separation improvement by Dean vortices, Journal of Membrane Science, 142 (1998) 129-141.
13
Table 1. Balance precision versus differential pressure sensor Precision M (g) ∆P (Pa) Balance ∆Mb Raw pressure signal Pressure sensor ∆Mp
± 0.01 ± 15 ± 0.10
14
Table 2. Mass of permeate weighed in the cold traps (4.a + 4.b) versus mass of permeate measured by pressure sensor (PT1, pressure sensor (8) in Figure 1) for pure ethanol (raw data) Experiment
1
2
3
3300
3300
3300
3.23
3.36
3.34
3.28
3.33
3.34
0.798834
0.798323
0.799383
Weight fraction in the permeate (-) : Wethanol
0.969
0.971
0.967
Mass of water in the permeate (g) : Mw
0.10
0.10
0.11
Pure ethanol in traps (g) : ∆MB ethanol
3.13
3.26
3.23
D (g) : (∆MB ethanol-∆MP)
-0.15
-0.07
-0.11
D (%)
4.8
2.1
3.4
Water mass flow rate due to humid air leak (10-5 g.s-1) : qw = (Mw/∆t)
3.03
3.03
3.34
Operating time (s) : ∆t = (tf-ti) Permeate mass, 4.a + 4.b – weighed (g) : ∆MB (±0.01 g) Ethanol mass – pressure sensor PT1 (g) : ∆MP = (Mi-Mf) (±0.10 g) Permeate density (g.cm-3)
15
Table 3. Mass of permeate weighed in the cold traps (4.a + 4.b) versus mass of permeate measured by pressure sensor PT1 (pressure sensor (8) in Figure 1) for pure water Experiment
1
2
3
22800
16200
13200
Permeate mass 4.a + 4.b – weighed (g) : ∆MB (±0.01g)
3.68
2.16
1.64
Water mass – pressure sensor PT1 (g) : ∆MP = (Mi-Mf) (±0.10g)
3.38
1.93
1.54
D (g) : (∆MB-∆MP)
0.3
0.23
0.1
D (%)
8.2
10.6
6.1
Water mass flow rate due to humid air leak (10-5 g.s-1) : qw = (D/∆t)
1.32
1.42
0.75
Operating time (s) : ∆t = (tf-ti)
16
S0263-8762(17)30358-1 http://dx.doi.org/doi:10.1016/j.cherd.2017.06.029 CHERD 2732
To appear in: Received date: Revised date: Accepted date:
19-7-2016 19-6-2017 21-6-2017
Please cite this article as: Toudji, S.A., Bonnet, J.-P., Gardarein, J.-L., Carretier, E., New experimental setup for continuous mass flux measurement in pervaporation.Chemical Engineering Research and Design http://dx.doi.org/10.1016/j.cherd.2017.06.029 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
New experimental setup for continuous mass flux measurement in pervaporation S. A. TOUDJI a, J-P. BONNET a,*, J-L GARDAREIN b, E. CARRETIER a a b
Aix Marseille Université, CNRS, Centrale Marseille, M2P2 UMR 7340, 13451, Marseille, France Aix Marseille Université, CNRS, IUSTI UMR 7343, 13453, Marseille, France
* Corresponding author. E-mail address: [email protected]
Highlights: New experimental setup was designed for continuous permeate mass flux measurement in pervaporation Mass flux measured with pressure sensor is validated by comparing to results obtained from conventional gravimetric method Relative high frequency of mass flux measurement allows to consider coupling with heat flux measurement in unsteady state
Abstract Pervaporation is a separation process of liquid mixtures through a thin non-porous membrane. In vacuum pervaporation, the global mass flux is classically estimated by weighing the mass of permeate collected in cold traps. In this work, we propose a new experimental setup that allows a continuous measurement of the mass flux. The new mass flux method measurement was validated for single component permeation (ethanol and water) by comparing mass of permeate collected in cold traps with the level decrease of feed liquid measured with a pressure sensor. This new setup can be useful for laboratory studies dealing with the evolution of mass flux according to different parameters of the process as, for example, the permeate side pressure level or temperature of liquid feed. Keywords: Pervaporation, Unsteady state measurement, Mass flux
1
Introduction Pervaporation is used for the separation of liquid mixtures through a non-porous and
selective membrane. In vacuum pervaporation or sweeping gas pervaporation, the driving force for mass transport is the difference of chemical potential between the liquid feed and the permeate [1]. In the first case the permeate side is kept at reduced pressure using a vacuum 1
pump while in the second case an inert gas sweeps the permeate side [2-4]. Pervaporation was initially developed for the dehydration of alcohol and especially for azeotropic systems [5]. Due to the potential of pervaporation as an energy saving and green process, it is combined with other processes as distillation forming hybrid-processes which are currently used in various applications [6] like: -
Dehydration of solvents [7].
-
Recovery of solvents from aqueous solutions [8-10] or from fermentation broth [11-12] mainly ABE fermentation [13].
-
Separation of organic mixtures [14-15].
Pervaporation differs from other membrane processes because it includes a phase change as it goes from a liquid phase in the feed side into a vapor phase in the permeate side [1]. Generally, in vacuum pervaporation, the permeate is condensed in cold trap and the mass flux J [kg.h-1.m-2] is calculated as the mass of permeate weighed in the cold trap divided by the membrane surface and the time spent to collect this amount of permeate [7]. This method is efficient to have an estimation of the global mass flux but it is unable to provide the mass flux versus the operating time with a small time span. Moreover, the trap change action can be considered as a perturbation for the process. In this work, an original experimental setup has been designed to measure continuously the mass flux and without permeate pressure perturbation. A differential pressure sensor placed in the feed tank provides the hydrostatic pressure with a 10 Hz frequency. The hydrostatic pressure data allow the evolution of the permeate mass flux to be presented at a relatively high frequency. The setup can be used in the classical types of laboratories pervaporation plants. As the amount of permeate is measured thanks to the variation of hydrostatic pressure in the liquid feed side, the permeate side conditions, vacuum or sweeping, should not affect the measurement.
2
2
Experimental setup 6.5
12
TC
25 C Legend: Polyurethane insulation PVC Liquid feed Thermostated enclosure Permeate side (reduced pressure) Membrane + spacer + support Distance and dimensions are expressed in mm
2 b h
11
p = ρ.g.h
150
TT3
0-10 V 0-20 mA 0-10 V
TT2
PT1
0-100 C 0-3 kPa 0-10 kPa
140
8
9
30
2a 150
10
4.a 6 PT2 5 75
42
TT1
4.b 1 3
7
1. Membrane + spacer + support ; 2. Liquid feed tank ; 3. Permeate side ; 4.a,b-5. Cold trap ; 6-8. Pressure sensor ; 7. Vacuum pump ; 9. Data acquisition system ; 10. Computer ; TT1, TT2, TT3. Temperature transmitters
Figure 1. Schematic view of the pervaporation setup
Pervaporation experiments were performed with the setup shown schematically in Figure 1. A commercial PDMS flat sheet membrane from Pervatech is used (1). The total membrane area is 14 cm². To ensure the contact surface between feed liquid and membrane is the same as the contact surface between permeate and membrane, a spacer is placed between the membrane and the stainless steel support. The membrane is supported by the stainless steel support (1 mm thickness) perforated with 25 holes of 6 mm diameter well spatially distributed on the surface (Figure 2).
3
Figure 2. Schematic view of the spacer used between the membrane and the support
During the experiments, the permeate side (3) is continuously maintained at reduced pressure (around 300 Pa) with a vacuum pump (7) ILMVAC-MP 601 Tp. The permeate pressure is continuously monitored with a pressure sensor (6) BROOKS-CMCA that have a vacuum measurement range from 5 to 104 Pa with an accuracy of 0.5 % of full scale (50 Pa). The permeate is condensed in glass cold trap (4.a – 4.b & 5) cooled with liquid nitrogen. It is collected periodically. At the end of the experiment, the three traps (4.a – 4.b & 5) are disconnected from the experimental device in order to be weighed individually. The mass of the empty traps being perfectly known, the permeate mass collected in each trap (∆MB 4a, ∆MB 4b) is calculated by subtracting the mass of the trap after and before permeation. The traps are weighed using a microbalance METTLER TOLEDO-XS1003S with 10 mg precision. Mass collected in security trap (5) is always nil. Pervaporation cell is made of PVC matter with a low thermal conductivity of about 0.16 W.m1
.K-1 to limit heat exchanges with the external environment and with the liquid. The liquid is
filled in both cylindrical tanks (2a) and (2b) above the membrane surface. The tank (2b) is designed with an inner diameter around 6.44 mm to have 3 Pa of differential pressure for 10 mg of permeated liquid (Eq. 1). The total initial volume is 160 mL. A differential pressure sensor (8) Keller PR41X-31 with a measurement range from 0 to 3000 Pa measures the hydrostatic liquid pressure. Using the hydrostatic relation (Eq. 1), the permeate amount could be calculated assuming that evaporation at the top edge of the feed tank is negligible and so the level drop is almost involved by the only permeation phenomenon.
P12
4g M12 D(22 b )
4
Eq. 1
A calibration curve (Figure 3) has been done to limit the error due to the geometric irregularities of the cylindrical tank 2b. Indeed, part 2b was machined in the laboratory with basic machining equipment and D2b, the inner diameter, is not perfectly constant all along the cylinder length. The hydrostatic pressure of liquid has been measured for different volumes added with a micropipettor at 25 °C (corresponding masses are calculated at the temperature considered). A quadratic model has been used to fit the calibration curves. 24 23
Calibration mass, M [g]
22 21 20 19 18 17 16
Experimental for ethanol Quadratic fit for ethanol
15
Experimental for water Quadratic fit for water
14 200
400
600
800
1000
1200
1400
1600
1800
Liquid hydrostatic pressure, P [Pa]
Figure 3. Hydrostatic pressure calibration for pure ethanol and pure water
The global precision of the pressure sensor (8) including the electric noise is 0.5% of the full scale (±15 Pa) that corresponds to ±0.10 g. The mass differences ∆Mp between an initial and final time (ti, tf) can be compared with ∆MB (Eq. 2). The pressure sensor precision (±0.10 g) is sufficient for the quantities of permeate collected in the cold trap varying between 1.5 – 4 g.
M B M B 4a M B 4b
Eq. 2
The entire cell device is insulated using polyurethane foam. All the measurements are simultaneously monitored (10 Hz) using NIcDAQ 9174 central data acquisition (9) connected to a computer (10). The experimental setup is placed inside a thermostated enclosure to control the environment temperature. NAHITA 639 series thermostated enclosure was used with 250 L capacity and 5 °C to 60 °C temperature range with temperature resolution of 0.1 °C. The temperature inside the thermostated enclosure during experiments was maintained at 25 °C.
5
Three experiments were conducted for both pure ethanol and pure water. The experiments last around 1 hour long for ethanol and between 4 – 6 hours for water, because the mass flux of pure water with the organophilic membrane used is very low. For ethanol experiments, the composition of permeate is deduced from density measurements. The density is measured using Anton Paar DMA 5000 M density meter with an accuracy of 5.10-6 g.cm-³ for a temperature accuracy of 0.01 °C. The weight fraction of water in permeate is estimated from a calibration curve relating the weight fraction of ethanol (Wethanol) to the density. The calibration curve has been validated by a chart from Perry's Chemical Engineers Handbook [16].
3
Results and discussion
3.1 Ethanol Results 1300
1800 1300
1200
900 800 700 600 500 400 300 200
1600
1500
0
500
1225
1200
1175
1150
0
50
100
1400
150
200
250
300
350
Time, t [s]
Vacuum pump « on »
1300
1200
Vacuum 1100pump "off" Vacuum pump "on"
100
1250
t4.a → t4.b
Liquid hydrostatic pressure, P [Pa]
Liquid hydrostatic pressure, P [Pa]
1000
0
Liquid hydrostatic pressure, P [Pa]
1275
1700
1100
1000 0 1000
Vacuum pump "off" Vacuum pump "on"
2000 1500
20004000 2500 60003000 8000 3500
Time, t [s]
10000 4000
12000
Time, t [s]
Figure 4. Liquid hydrostatic pressure during pervaporation of pure ethanol (experiment 3)
The feed hydrostatic pressure during the pervaporation of ethanol experiments has been measured and the results are shown on Figure 4. The first 300 s of acquisition shows that the pressure signal is almost constant before the vacuum pump is switched on. It indicates that there is no significant ethanol leak and no significant pervaporation phenomenon. At t=300 s, 6
14000
the vacuum pump is switched on and the pervaporation phenomenon starts. As a result, the liquid level decreases due to its permeation through the membrane and so the hydrostatic pressure decreases. The first shift in pressure (from 1250 to 1150 Pa) is due to the mechanical deformation of the metallic support and so to the deformation of the membrane due to vacuum generated by the pump. We can visually detect that there is a shift in the slopes dP/dt at the time t4.a4.b = 900 s that corresponds to the moment when the valves are switched to collect the permeate from trap 4.a to trap 4.b. Only the data ranging from 300 s to 3600 s (red data in Figure 4) are used. The calibration curve (Figure 3) is used to convert the pressure data into liquid mass data that is contained in the feed tank above the pressure sensor (Figure 5). 18
Mi 17.5
17
Mass, M [g]
16.5
16
15.5
15
Mf
14.5
14
0
500
ti
1000
1500
2000
Time, t [s]
2500
3000
3500
4000
tf
Figure 5. Mass of liquid above the pressure sensor during pervaporation of pure ethanol (experiment 3)
The mass differences ∆MP between the initial and final time (ti, tf) are reported in Table 2. Density measurements show differences between liquid feed and liquid permeate. This can be attributed to water presence in permeate due to humid air leaks that enters in the permeate side system which is at reduced pressure. That phenomenon is frequently encountered in pervaporation [17-18]. Density and composition of permeate are presented in Table 2.
The water mass flow rate qw, defined as Mw/∆t represents the water supplied by humid air leak per time unit, it is estimated around 3.10-5 g.s-1. Hence the mass of permeate weighed in 7
the cold trap is corrected by taking into account the water to estimate the mass of pure ethanol ∆MB
ethanol
to be compared with ∆MP. ∆MB
ethanol
is systematically smaller than the mass
measured by the pressure sensor ∆MP. The deviation D (g), defined as ∆MB ethanol - ∆MP, is a negative value and varies between 0.07 – 0.15 g which correspond to 2 – 5 % of ∆MP. Although this deviation is very low, it can be explained by the evaporation of ethanol on the feed side, near the edge ethanol / air in tank (2b). To estimate the ethanol evaporation rate, several data acquisitions during five hours have been performed with an impervious film in the membrane place. Although dependent on the position of ethanol / air interface in the tank 2b (Figure 1), the average rate of evaporation is about 3.10-5 g.s-1.
3.2 Water results 1800 1775
Liquid hydrostatic pressure, P [Pa]
Liquid hydrostatic pressure, P [Pa]
1700
1600
1500
1750
1725
1700
1675
1650
1400
0
50
100
150
200
250
300
350
Time, t [s]
Vacuum pump « on »
1300
1200
1100
1000
Vacuum pump "off" Vacuum pump "on" 0
2000
4000
6000
8000
10000
12000
14000
Time, t [s]
Figure 6. Liquid hydrostatic pressure during pervaporation of pure water (experiment 3)
The feed hydrostatic pressure during the pervaporation of water experiments has been measured and the results are shown on Figure 6. The first 300 s of acquisition show that the pressure signal is constant before the vacuum pump is switched on. It validates that the water level is constant and so that there is no water leak, no pervaporation and no significant evaporation. At t=300 s, the vacuum pump is switched on and the pervaporation phenomenon 8
starts. As in ethanol case, the first shift in pressure (from 1740 to 1680 Pa) is due to the mechanical deformation of the membrane. The calibration curve (Figure 3) is used to convert the hydrostatic pressure into the mass of liquid contained in the feed tank above the pressure sensor (Figure 7). 23.5
23
Mi
Mass, M [g]
22.5
22
21.5
Mf
21
20.5
0
2000
ti
4000
6000
8000
Time, t [s]
10000
12000
14000
tf
Figure 7. Mass of liquid above the pressure sensor during pervaporation of pure water (experiment 3)
The mass of permeate weighed in the cold trap is systematically greater than the mass measured by the pressure sensor (Table 3). The deviation D (g) is a positive value and increases with the experiment duration. The deviation D (g) divided by the experiments duration gives a water flow rate qw in the same order of magnitude for the three experiments, around 1.10-5 g.s-1, as in the ethanol case (Table 2). So, the water excess in the trap could be attributed to humid air leaks on the permeate side. In this specific case of low permeate fluxes on long duration, the use of a pressure sensor is a better way than the classical gravimetric method because it permit to eliminate the water due to humid air leak.
9
500
3.5
450
3
400
2.5
350
2
300
1.5
250
1
200
p
4
Permeate pressure, P [Pa]
Mass flux, J [kg.h-1.m-2]
3.3 Mass flux estimation
0.5
Mass flux
t4.a → t4.b 0
0
300
600
900
150
Permeate pressure
100 1200 1500 1800 2100 2400 2700 3000 3300 3600
Time, t [s]
Figure 8. Mass flux during pervaporation of pure ethanol (experiment 3)
The very good agreement between mass of permeate measured with the pressure sensor and mass weighed in the cold trap allows to estimate the permeate mass flux using the pressure sensor (Figure 8). In this purpose, a sliding window is used to measure the slope from the mass versus time curve (Figure 5). The window size is 150 s and it scans data from 300 s to 900 s (trap changed), then from 900 s to 3600 s (vacuum pump switched off). As expected, the mass flux increases during first moments of permeation (300 s – 420 s) and reaches a maximum of 3.84 kg.h-1.m-2 when the permeate pressure stabilizes around 190 Pa. For time ranging from 420 s to 900 s, although the permeate pressure is constant, the mass flux decreases due to the liquid temperature drop (close to membrane) [19-20]. Indeed, this temperature reduction implies a reduction of energy available in the liquid leading to a decrease in the mass transport. At t=900 s, the trap changing involves an increase in permeate pressure from 190 Pa to 270 Pa, caused probably by more important air leak on the device with the second trap. The permeate mass flux decreases when the permeate pressure increases. From 1500 s to 3600 s (the end of experiment) the mass flux stabilizes around 2.4 kg.h-1.m-2 as a pseudosteady state seems to be reached. This value is in accordance with permeate mass fluxes reported in literature for pervaporation process [21-25].
10
4
Conclusion: An original and innovative experimental setup has been developed to measure
continuously the mass flux in pervaporation laboratory experiments. The conventional gravimetric method used to measure mass flux in pervaporation serves as reference in this work. This technique needs relatively long period of experiment to have a sufficient mass of permeate to be weighed. Moreover, this technique provides only a global estimation of the mass flux. The conventional value of collected permeate in the trap is compared with the mass of liquid, measured with a pressure sensor, that has left the feed tank. These two masses are supposed to be the same assuming leaks and mass accumulation term in the membrane close to zero. We compare experimentally these two quantities, they are almost the same for all tested conditions (ethanol and water). These results valid the new measurement setup to estimate the permeate mass flux. Moreover, since the pressure sensor is also a transmitter, this setup allows to accede at continuous measurement of the permeate mass flux with relatively high frequency in comparison with the conventional method. This approach opens new perspectives to investigate, with a non-intrusive measurement, the non-stationary regime in the pervaporation process. Works are in progress for: (1) measurement of the temperature field upstream the membrane in order to correlate it with the mass flux; (2) the use of this new technique in liquid binary systems; the experimental setup need to be modified by adding an on-line analyzer (spectrometer, refractometer, …) to allow measurement for mixtures with the same precision as with pure components.
11
References: [1] J.G. Crespo, C. Brazinha, 1 - Fundamentals of pervaporation, in: Pervaporation, Vapour Permeation and Membrane Distillation, Woodhead Publishing, Oxford, 2015, pp. 3-17. [2] C. Brazinha, V.D. Alves, R.M.C. Viegas, J.G. Crespo, Aroma recovery by integration of sweeping gas pervaporation and liquid absorption in membrane contactors, Separation and Purification Technology, 70 (2009) 103-111. [3] C. Vallieres, E. Favre, Vacuum versus sweeping gas operation for binary mixtures separation by dense membrane processes, Journal of Membrane Science, 244 (2004) 17-23. [4] F. Lipnizki, R.W. Field, Integration of vacuum and sweep gas pervaporation to recover organic compounds from wastewater, Separation and Purification Technology, 22–23 (2001) 347-360. [5] A. Jonquières, R. Clément, P. Lochon, J. Néel, M. Dresch, B. Chrétien, Industrial state-of-the-art of pervaporation and vapour permeation in the western countries, Journal of Membrane Science, 206 (2002) 87-117. [6] S.C. George, S. Thomas, Transport phenomena through polymeric systems, Progress in Polymer Science, 26 (2001) 985-1017. [7] P.D. Chapman, T. Oliveira, A.G. Livingston, K. Li, Membranes for the dehydration of solvents by pervaporation, Journal of Membrane Science, 318 (2008) 5-37. [8] T. Uragami, Y. Matsuoka, T. Miyata, Permeation and separation characteristics in removal of dilute volatile organic compounds from aqueous solutions through copolymer membranes consisted of poly(styrene) and poly(dimethylsiloxane) containing a hydrophobic ionic liquid by pervaporation, Journal of Membrane Science, 506 (2016) 109-118. [9] S.B. Kuila, S.K. Ray, Separation of isopropyl alcohol–water mixtures by pervaporation using copolymer membrane: Analysis of sorption and permeation, Chemical Engineering Research and Design, 91 (2013) 377-388. [10] A.J. Toth, P. Mizsey, Methanol removal from aqueous mixture with organophilic pervaporation: Experiments and modelling, Chemical Engineering Research and Design, 98 (2015) 123-135. [11] M. Moussa, V. Athès, Y. Imbert, I. Souchon, O. Vitrac, M.L. Lameloise, Pervaporative Dehydration of Bioethanol using Silica and PVA Membranes: Analysis of Permeation Performances and Effect of Volatile Organic Impurities, Procedia Engineering, 44 (2012) 1173-1176. [12] E. Nagy, P. Mizsey, J. Hancsók, S. Boldyryev, P. Varbanov, Analysis of energy saving by combination of distillation and pervaporation for biofuel production, Chemical Engineering and Processing: Process Intensification, 98 (2015) 86-94. [13] M.F.S. Dubreuil, P. Vandezande, W.H.S. Van Hecke, W.J. Porto-Carrero, C.T.E. Dotremont, Study on ageing/fouling phenomena and the effect of upstream nanofiltration on in-situ product recovery of n-butanol through poly[1-(trimethylsilyl)-1-propyne] pervaporation membranes, Journal of Membrane Science, 447 (2013) 134-143. [14] P. Luis, J. Degrève, B.V. der Bruggen, Separation of methanol–n-butyl acetate mixtures by pervaporation: Potential of 10 commercial membranes, Journal of Membrane Science, 429 (2013) 112. [15] B. Smitha, D. Suhanya, S. Sridhar, M. Ramakrishna, Separation of organic–organic mixtures by pervaporation—a review, Journal of Membrane Science, 241 (2004) 1-21. [16] D.G. Robert H. Perry, Perry's Chemical Engineers' Handbook, Seventh Edition ed. [17] P.T. Sumesh, P.K. Bhattacharya, Analysis of phase change during pervaporation with single component permeation, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 290 (2006) 263-272. [18] C. Vallieres, E. Favre, D. Roizard, J. Bindelle, D. Sacco, New Insights into Pervaporation Mass Transport under Increasing Downstream Pressure Conditions: Critical Role of Inert Gas Entrance, Industrial & Engineering Chemistry Research, 40 (2001) 1559-1565. [19] A. Ito, Y. Feng, H. Sasaki, Temperature drop of feed liquid during pervaporation, Journal of Membrane Science, 133 (1997) 95-102. [20] E. Favre, Temperature polarization in pervaporation, Desalination, 154 (2003) 129-138. 12
[21] I. Blume, J.G. Wijmans, R.W. Baker, The separation of dissolved organics from water by pervaporation, Journal of Membrane Science, 49 (1990) 253-286. [22] X. Lin, H. Kita, K.-i. Okamoto, Silicalite Membrane Preparation, Characterization, and Separation Performance, Industrial & Engineering Chemistry Research, 40 (2001) 9. [23] Y.K. Ong, G.M. Shi, N.L. Le, Y.P. Tang, J. Zuo, S.P. Nunes, T.-S. Chung, Recent membrane development for pervaporation processes, Progress in Polymer Science, 57 (2016) 1-31. [24] X. Zhan, J.-d. Li, J.-q. Huang, C.-x. Chen, PERVAPORATION PROPERTIES OF PDMS MEMBRANES CURED WITH DIFFERENT CROSS-LINKING REAGENTS FOR ETHANOL CONCENTRATION FROM AQUEOUS SOLUTIONS, Chinese Journal of Polymer Science, 27 (2009) 533-542. [25] S. Schnabel, P. Moulin, Q.T. Nguyen, D. Roizard, P. Aptel, Removal of volatile organic components (VOCs) from water by pervaporation: separation improvement by Dean vortices, Journal of Membrane Science, 142 (1998) 129-141.
13
Table 1. Balance precision versus differential pressure sensor Precision M (g) ∆P (Pa) Balance ∆Mb Raw pressure signal Pressure sensor ∆Mp
± 0.01 ± 15 ± 0.10
14
Table 2. Mass of permeate weighed in the cold traps (4.a + 4.b) versus mass of permeate measured by pressure sensor (PT1, pressure sensor (8) in Figure 1) for pure ethanol (raw data) Experiment
1
2
3
3300
3300
3300
3.23
3.36
3.34
3.28
3.33
3.34
0.798834
0.798323
0.799383
Weight fraction in the permeate (-) : Wethanol
0.969
0.971
0.967
Mass of water in the permeate (g) : Mw
0.10
0.10
0.11
Pure ethanol in traps (g) : ∆MB ethanol
3.13
3.26
3.23
D (g) : (∆MB ethanol-∆MP)
-0.15
-0.07
-0.11
D (%)
4.8
2.1
3.4
Water mass flow rate due to humid air leak (10-5 g.s-1) : qw = (Mw/∆t)
3.03
3.03
3.34
Operating time (s) : ∆t = (tf-ti) Permeate mass, 4.a + 4.b – weighed (g) : ∆MB (±0.01 g) Ethanol mass – pressure sensor PT1 (g) : ∆MP = (Mi-Mf) (±0.10 g) Permeate density (g.cm-3)
15
Table 3. Mass of permeate weighed in the cold traps (4.a + 4.b) versus mass of permeate measured by pressure sensor PT1 (pressure sensor (8) in Figure 1) for pure water Experiment
1
2
3
22800
16200
13200
Permeate mass 4.a + 4.b – weighed (g) : ∆MB (±0.01g)
3.68
2.16
1.64
Water mass – pressure sensor PT1 (g) : ∆MP = (Mi-Mf) (±0.10g)
3.38
1.93
1.54
D (g) : (∆MB-∆MP)
0.3
0.23
0.1
D (%)
8.2
10.6
6.1
Water mass flow rate due to humid air leak (10-5 g.s-1) : qw = (D/∆t)
1.32
1.42
0.75
Operating time (s) : ∆t = (tf-ti)
16