Adsorption of Polar and Nonpolar Vapors on Selected Adsorbents: Breakthrough Curves and their Simulation Dr. Robert Eschrich Quantachrome GmbH & Co. KG 2018-04-17 Leipziger Symposium on Dynamic Sorption 2018 sorption heat and energy storage 1 www.dynamicsorption.com
1. Breakthrough Curve Introduction 2. The mixsorb L with Vapor Option 3. Breakthrough Curves of Vapors I. Water Breakthrough Curves on Zeolite, Silica Gel Regenerability and Heat Output II. Toluene Breakthrough Curve on Activates Carbon III. Ethanol/water Mixture Adsorption on Activates Carbon 4. Modelling I. Simulating Water Breakthrough Curves II. Simulating Ethanol Breakthrough Curves 2 www.dynamicsorption.com
Static Volumetric Measurements Sorption takes place in enclosed chamber Pressure is recorded over time Pure Gases only 4 Breakthrough Experiment Sorption takes place in open system Pressure is constant Gas Mixtures only Outlet composition is recorded over time 100 n ads,i = n dosed,i n free,i n dosed,i = p Dose,i V Dose RT n free,i = p Cell,i V Dose + V Cell RT loading mmol g -1 3 2 1 0 0 0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0 p/p 0 n ads = i 0 n ads,i rel. breakthrough / % 80 60 40 20 time-on-stream / s n adsorbed = n in (t)dt n out (t)dt n adsorbed = V in (t) y in(t) V m dt V out (t) y out(t) V m dt 3 www.dynamicsorption.com
Breakthrough Curves Not all Gas Flow Experiments are Breakthrough Experiments! Requirement: Fixed Adsorber Bed gas must not pass the sample without interaction! gas outlet T T detection adsorber c out u out adsorbent heat of adsorption adsorptive T T What is the result of a breakthrough experiment? gas inlet axial dispersion c in u in transport into particle adsorption Time until 5 %, 50 %, of breakthrough is the cycle or production time Integration of the full curve gives saturation capacity of a gas on the adsorbent (equilibrium) Integration until cycle time gives technically usable sorption capacity Shape of the curve contains information about kinetics/mass transfer 4 www.dynamicsorption.com
Different Scales gas outlet detection c out u out adsorbent heat of adsorption T T adsorber adsorptive T T gas inlet axial dispersion c in u in transport into particle adsorption Macroscopic Mesoscopic Microscopic Size of Adsorber Shape of Adsorber Nature of the Fixed Bed Bed Porosity Shape of Particles Textural Properties Surface Characteristics Accessibility 5 www.dynamicsorption.com
Resulting Curves 40 C, 2 L min -1 5 bar (pressurization with N 2 ) Inlet compositions: 5 % CO 2 in N 2 Temperature Maxima Decrease in Flow Direction Increasing Dispersion Area under Temperature Curves increases in Flow Direction Transfer of heat through gas flow 6 www.dynamicsorption.com
Calculating Loadings n adsorbed = n in (t)dt n out (t)dt n adsorbed = V in (t) y in(t) V m dt V out (t) y out(t) V m dt Integrating over the full Curve Integrating over the Curve to e.g. 1 % Breakthrough Saturation Capacity dq = 0.611 mmol g -1 Technically Usable Sorption Capacity dq = 0.445 mmol g -1 7 www.dynamicsorption.com
mixsorb L Fully automated Breakthrough Analyzer Integrated Gas Mixing Including Vapors Up to 40 L/min Gas Flow, up to 10 bar Up to 4 mass flow controllers (MFCs) Up to 2 Evaporators, each capable to supply vapor mixtures Monitoring of gas composition by TCD at the Outlet or Bypass You can attach any additional Analytical Device (e.g. Mass Spec) at the sample port 3P-SIM Simulation Software 8 www.dynamicsorption.com
I. Breakthrough Curves of H 2 O Drying of process gases Important to remove Water and CO2 in Air before Cryogenic Air Separation Water would plug the piping by freezing. Air separation with Pressure Swing Adsorption (PSA) on Zeolites Water has strong affinity to surface Cooling/Condensation Energy Storage Adsorption of water vapor releasing heat of adsorption. Control heat output with dosing of water Regeneration with e.g. thermal solar energy Can be used for cooling as well % RH % RH Adsorber Δp Condensate Heating/Evaporation 9 www.dynamicsorption.com
Breakthrough Curve of H 2 O / N 2 on Zeolite 13X volume fraction y(h 2 O) / % 3.0 2.5 2.0 y(h2o) T1 T2 T3 T4 1.5 1.0 0.5 0 50 100 150 200 250 300 350 400 time-on-stream t / min 80 70 60 50 40 30 20 bed temperature T / C Experimental conditions of a simple breakthrough experiment after Activation at 400 C for 4 h 25 C, Flow rate 4 L min -1 Pressure: 1 bar Standard Adsorber with 80 g of sample Inlet composition: 5 g h -1 H 2 O in N 2 (volume fraction y(h 2 O) = 2.59 %, Relative humidity approx. 80 % @ 25 C) High temperatures during adsorption Loading: 18.9 mmol g -1 Regeneration at 130 C for 3.5 h 10 www.dynamicsorption.com
Breakthrough Curve of H 2 O / N 2 on Zeolite 13X volume fraction y(h 2 O) / % 3.0 2.5 2.0 1.5 1.0 0.5 T1 y(h2o) T2 T3 T4 80 70 60 50 40 30 bed temperature T / C Sample regenerated at 130 C for 3.5 h Same experimental conditions Inlet composition: 5 g h -1 H 2 O in N 2 (volume fraction y(h 2 O) = 2.59 %, Relative humidity approx. 80 % @ 25 C) Loading: 15.4 mmol g -1 lower Unsymmetrical temperature profiles Residual loading before experiment not equally distributed 0 50 100 150 200 250 300 350 400 time-on-stream t / min 20 Overlay 11 www.dynamicsorption.com
Breakthrough Curve of H 2 O / N 2 on Zeolite 13X volume fraction y(h 2 O) / % 3.0 2.5 2.0 1.5 1.0 0.5 regenerated 130 C regenerated 400 C 0 50 100 150 200 250 300 350 400 time-on-stream t / min Loadings: 18.9 mmol g -1 (activated at 400 C) vs. 15.4 mmol g -1 (regenerated at 130 C) Breakthrough Curve shifted to the left Breakthrough curves still have similar shapes Zeolite requires harsh regeneration High temperatures Steep Breakthrough Curves indicate steep isotherms High affinity to water 12 www.dynamicsorption.com
Breakthrough Curve of H 2 O / N 2 on Silica Gel volume fraction y(h 2 O) / % 3.0 2.5 2.0 1.5 1.0 0.5 y(h2o) T1 T2 T3 T4 80 70 60 50 40 30 bed temperature T / C Experimental conditions of a simple breakthrough experiment after Activation at 350 C for 4 h 25 C, Flow rate 4 L min -1 Pressure: 1 bar Standard Adsorber with 80 g of sample Inlet composition: 5 g h -1 H 2 O in N 2 (volume fraction y(h 2 O) = 2.59 %, Relative humidity approx. 80 % @ 25 C) Smaller temperatures peaks Much longer measurement 20 0 100 200 300 400 500 600 700 800 900 1000 1100 time-on-stream t / min Loading: 25.9 mmol g -1 Regeneration at 130 C for 3.5 h 13 www.dynamicsorption.com
Breakthrough Curve of H 2 O / N 2 on Silica Gel volume fraction y(h 2 O) / % 3.0 2.5 2.0 1.5 1.0 0.5 T1 T2 T3 T4 y(h2o) 80 70 60 50 40 30 bed temperature T / C Sample regenerated at 130 C for 3.5 h Same experimental conditions Inlet composition: 5 g h -1 H 2 O in N 2 (volume fraction y(h 2 O) = 2.59 %, Relative humidity approx. 80 % @ 25 C) Loading: 25.1 mmol g -1 almost identical Overlay 20 0 100 200 300 400 500 600 700 800 900 1000 1100 time-on-stream t / min 14 www.dynamicsorption.com
Breakthrough Curve of H 2 O / N 2 on Silica Gel volume fraction y(h 2 O) / % 3.0 2.5 2.0 1.5 1.0 0.5 regenerated 130 C regenerated 400 C 0 200 400 600 800 1000 time-on-stream t / min Loadings: 25.9 mmol g -1 (activated at 350 C) vs. 25.1 mmol g -1 (regenerated at 130 C) Breakthrough Curve changed slope Changing surface chemistry until stable in cycles Regeneration much easier, efficient No high temperature required Regeneration possible by Pressure Reduction But: Breakthrough occurs earlier! 15 www.dynamicsorption.com
Breakthrough Curve of H 2 O / N 2 on Zeolite 13X and Silica Gel 3.0 30 2.5 25 Silica volume fraction y(h 2 O) / % 2.0 1.5 1.0 0.5 Zeolite activated Zeolite regenerated Silica activated Silica regenerated loading q(h 2 O) / mmol g -1 20 15 10 5 0 Zeolite 13X 0 1 2 3 4 5 6 7 8 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 time per volume t V / min cm -3 Materials behave differently in Adsorption/Desorption Cycles Good Agreement with Isotherm data (right hand side) Can we use these curves to get information about stored energy and heating power? 16 www.dynamicsorption.com p / p 0
Immersion Calorimetry Determining the Heat of Adsorption in Liquid Phase 40 35 30 Zeolite 13X»Enthalpy of Adsorption = Wetting + Condensation«h A = h W + h C Enthalpy of wetting Zeolite: 550 J g -1 (g of Adsorbent) Silica Gel: 140 J g -1 (g of Adsorbent) heat flow / mw 25 20 15 10 5 Silica 0 0 5 10 15 20 time t / min Zeolite: 1600 J g -1 (g of water) Silica Gel: 300 J g -1 (g of water) (re-calculated according to water isotherms) Zeolite 13X Silica Gel h W / J g -1 (H 2 O) 1600 300 h C / J g -1 (H 2 O) 2500 2500 h A / J g -1 (H 2 O) 4100 2800 17 www.dynamicsorption.com
Heat Power Comparison Comparing the Heating Power during Adsorption 8.0 7.0 P = (1 rel. Breakthrough) 5 g h 3600 s h h A Heat Power P / Watt 6.0 5.0 4.0 3.0 2.0 Zeolite Silica Gel Zeolite: More Heating Power, but over short duration abrupt drop Silica Gel: Less Heating Power, continuously decreasing longer duration 1.0 0 200 400 600 800 1000 1200 time-on-stream t / min 18 www.dynamicsorption.com
Breakthrough Curve of H 2 O / N 2 on Activated Carbon volume fraction y(h 2 O) / % 3.0 2.5 2.0 1.5 1.0 y(h2o) T1 T2 T3 T4 0.5 0 50 100 150 200 250 300 350 400 450 time-on-stream t / min 44 42 40 38 36 34 32 30 28 26 24 bed temperature T / K 25 C, 4 L min -1 1 bar Inlet composition: 5 g h -1 H 2 O in N 2 (volume fraction y(h 2 O) = 2.59 %, RH approx. 80%) Loading: 1 mmol g -1 Shape of curves can be explained by adsorption and condensation in the pores. Fast breakthrough due to hydrophobic surface Similar to Silica Gel, but Condensation is more pronounced 19 www.dynamicsorption.com
II. Breakthrough Curve of Toluene/N 2 volume fraction y(toluene) / % 2.4 2.1 1.8 1.5 1.2 0.9 0.6 0.3 T1 T2 y(toluene) T3 T4 20 0 10 20 30 40 50 60 70 80 time-on-stream t / min 60 55 50 45 40 35 30 25 bed temperature T / C Activated Carbon D55/1.5 25 C, 4 L min -1 Inlet composition: 20 g h -1 Toluene in N 2 (volume fraction y(toluene) = 2.0 %, p/p 0 = 0.53 (@ 25 C) Large temperature peaks Steep Breakthrough Curve Loading: 2.1 mmol g -1 More similar to H 2 O/Zeolite than H 2 O/Activated Carbon 20 www.dynamicsorption.com
III. Breakthrough Curves of a EtOH/H 2 O Mixture in N 2 on Activated Carbon 3.2 2.8 36 34 Temperature T4 (250 mm) volume fraction / % 2.4 2.0 1.6 1.2 0.8 H 2 O EtOH temperature / C 32 30 28 26 24 0.4 22 0 50 100 150 200 250 time-on-stream / min 20 0 50 100 150 200 250 time-on-stream / min complex breakthrough due to Type-V/Type-I isotherms roll-up effect for H 2 O due to replacement by EtOH 21 www.dynamicsorption.com 2.5 % H 2 O, 1 % EtOH in N 2, Total Flow: 4 Nl min -1 T = 25 C
Procedure 1. Measuring Breakthrough Curve 2. Measuring Pure Component Isotherms: 1 of each adsorptive for isothermal simulations 3 of each adsorptive for non-isothermal simulations 3. Predict Mixture Adsorption with IAST or Multi-Component Models 4. Predict Breakthrough Curves based on Mass- and Energy Balances 5. Evaluate mass transfer by varying k LDF to fit the predicted to the experimental Breakthrough Curves 22 www.dynamicsorption.com
concentration c loading q Mass Transfer coefficient k LDF Convection, Diffusion Adsorptive Film Diffusion Pore Diffusion Adsorbate Simplification Convection, Diffusion Film Diffusion Pore Free Diffusion KNUDSEN Diffusion Surface Diffusion Convection, Diffusion Adsorbent Effective Inner Mass Transfer, Linear Driving Force (LDF) Adsorption Adsorption LDF distance r 23 W. Kast, Adsorption aus der Gasphase: Ingenieurwissenschaftliche Grundlagen und technische Verfahren, 1.Aufl., VCH Wiley Verlag, Weinheim, 1988. www.dynamicsorption.com
Fitting of Breakthrough Curves + Temperatures 8 Experiment 50 8 Simulation 50 volume fraction y(co 2 ) / % 7 6 5 4 3 2 1 T1 T2 T3 T4 y(co 2 ) 48 46 44 42 40 38 36 Temperature / C volume fraction y(co 2 ) / % 7 6 5 4 3 2 1 T1 T2 T3 T4 y(co 2 ) 48 46 44 42 40 38 36 Temperature / C 0 34 0 5 10 15 20 25 time-on-stream t / min 0 34 0 5 10 15 20 25 time-on-stream t / min After fitting the k LDF Mass Transfer Coefficient Good Agreement of Experiment and Simulation in a Standard Breakthrough Experiment (5% CO 2 in N 2 ) Course of Volume Fraction and Temperatures is depicted correctly 24 www.dynamicsorption.com
Fitting of Breakthrough Curves + Temperatures 8 50 volume fraction y(co 2 ) / % 7 6 5 4 3 2 1 T4 y(co 2 ) 48 46 44 42 40 38 36 Temperature / C Overlay of Experiment (Points) and Simulation (Line) 0 34 0 5 10 15 20 25 time-on-stream t / min After fitting the k LDF Mass Transfer Coefficient Good Agreement of Experiment and Simulation in a Standard Breakthrough Experiment (5% CO 2 in N 2 ) Course of Volume Fraction and Temperatures is depicted correctly 25 www.dynamicsorption.com
H 2 O on 13XBFK Modeling with Dual-site Langmuir Sips isotherm model (DSLAI-SIPS) volume fraction H 2 O / % 3.2 2.8 2.4 2.0 1.6 1.2 0.8 0.4 Experiment non-isothermal SIM temperature / C 65 60 55 50 45 40 35 30 T (150 mm) non-isothermal SIM 0 100 200 300 400 500 600 time-on-stream / min 25 0 100 200 300 400 500 600 time-on-stream / min Good correlation between experiment and simulation for Type I isotherms Heat effects can be simulated by using non-isothermal models Future Improvements: Using 2D, instead of 1D models (radial discretization) q eq = q max K 1 p 1 + K 1 p + K 2 p t 1 + K 2 p t 26 Sample: 13XBFK, conditions: 2.5 % H 2 O in N 2, 4 NL min -1, 25 C (80 % RH) www.dynamicsorption.com
H 2 O on Silica Gel Modeling with DSLAI-SIPS Using heats of adsorption Q 1, Q 2, determined by fitting 3 isotherms Using heat of adsorption Q 1, determined by fitting 3 isotherms And Q 2 = Q Evap 3.2 40 3.2 40 volume fraction H 2 O / % 2.8 2.4 2.0 1.6 1.2 0.8 0.4 Exp. y and T SIM y and T 38 36 34 32 30 28 26 temperature / C volume fraction H 2 O / % 2.8 2.4 2.0 1.6 1.2 0.8 0.4 Exp. y and T SIM y and T 38 36 34 32 30 28 26 temperature / C 24 0 200 400 600 800 1000 1200 time-on-stream / min 24 0 200 400 600 800 1000 1200 time-on-stream / min Good correlation between experiment and simulation Better prediction of breakthrough curve with Q 2 = Q evap Better description of temperature curves with Q 1, Q 2 = heats of adsorption (1=LANGMUIR, 2=SIPS) 27 www.dynamicsorption.com
H 2 O on Activated Carbon (Type V Isotherm) Modeling with DSLAI-SIPS volume fraction H 2 O / % 3.2 2.8 2.4 2.0 1.6 1.2 0.8 0.4 Experiment isothermal SIM non-isothermal SIM 0 100 200 300 400 500 600 time-on-stream / min temperature / C 32 31 30 29 28 27 26 25 T (150 mm) non-isothermal SIM 24 0 100 200 300 400 500 600 time-on-stream / min Good correlation between experiment and Simulation for with non-isothermal model Poor Correlation for isothermal models Improvement of Breakthrough fit by using loading-dependent k LDF adsorption + pore filling by condensation But: prediction of temperatures gets worse Sample: Activated Carbon, conditions: 2.5 % H 2 O in N 2, 4 NL min -1, 25 C (80 % RH) Crank J.; The Mathematics of Diffusion, Oxford University Press, 1975. 28 www.dynamicsorption.com
EtOH on Activated Carbon Modeling with DSLAI-SIPS volume fraction EtOH / % 3.2 2.8 2.4 2.0 1.6 1.2 0.8 0.4 Experiment non-isothermal SIM temperature / C 50 45 40 35 30 25 20 15 T non-isothermal SIM 0 10 20 30 40 50 60 70 80 90 100 time-on-stream / min 10 0 10 20 30 40 50 60 70 80 90 100 time-on-stream / min Good correlation between experiment and simulation Future Improvements: Using 2D, instead of 1D models (radial discretization) Using Isosteric heat instead of heat of adsorption 29 www.dynamicsorption.com
Characterization under application-related conditions! mixsorb L is very versatile instrument for application-related studies Vapor Sorption, determine Isotherms, Mixture Isotherms Breakthrough curves of other Adsorptives in the Presence of Water Adsorption Studies of Organic Vapors: VOC adsorption 3P-SIM is a powerful simulation tool for Breakthrough Prediction Future steps: Simulation of vapor mixture adsorption Implementing 2D, instead of 1D models (radial discretization) Implementing isosteric heat into non-isothermal models, where possible 30 www.dynamicsorption.com
Thank you for your attention! 31 Dr. Robert Eschrich, Leipziger Symposium on Dynamic Sorption, 2018-04-17 www.dynamicsorption.com