Adsorption of Water and Ethanol in MFI-Type Zeolites - Langmuir

Subscriber access provided by DUKE UNIV

Article

Adsorption of Water and Ethanol in MFI-type Zeolites Ke Zhang, Ryan P Lively, James D. Noel, Michelle E Dose, Benjamin A McCool, Ronald R Chance, and William J. Koros Langmuir, Just Accepted Manuscript • Publication Date (Web): 08 May 2012 Downloaded from http://pubs.acs.org on May 9, 2012

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Adsorption of Water and Ethanol in MFI-type Zeolites Ke Zhanga, Ryan P. Livelyb, James D. Noela, Michelle E. Doseb, Benjamin A. McCoolb, Ronald R. Chancea,b*, and William J. Korosa a

School of Chemical and Biomolecular Engineering, Georgia Institute of Technology, Atlanta, GA 30332 b

Algenol Biofuels Inc. 28100 Bonita Grande Drive, Bonita Springs, FL 34135

Abstract Water and ethanol vapor adsorption phenomena are investigated systematically on a series of MFI-type zeolites: silicalite-1 samples synthesized via both alkaline (OH-) and fluoride (F-) routes, and ZSM-5 samples with different Si/Al ratios as well as different charge-balancing cations. Full isotherms (0.05-0.95 activity) over the range 25-55°C are presented and the lowest total water uptake ever reported in the literature is shown for silicalite-1 made via a fluoridemediated route wherein internal silanol defects are significantly reduced. At a water activity level of 0.95 (35°C), the total water uptake by silicalite-1 (F-) was found to be 0.263 mmol/g, which was only 12.6%, 9.8% and 3.3% of the capacity for silicalite-1 (OH-), H-ZSM-5 (Si/Al:140) and H-ZSM-5 (Si/Al:15) respectively under the same conditions. While water adsorption shows distinct isotherms for different MFI-type zeolites due to the difference in the concentration, distribution and types of the hydrophilic sites, the ethanol adsorption isotherms present relatively comparable results because of the overall organophilic nature of the zeolite framework. Due to the dramatic differences in the sorption behavior with the different sorbate-sorbent pairs, different models are applied to correlate and analyze the sorption isotherms. An adsorption potential theory was used to fit the water adsorption isotherms on all MFI-type zeolite adsorbents studied. The Langmuir model and Sircar’s model are applied to describe ethanol adsorption on silicalite-1 and ZSM-5 samples, respectively. An ideal ethanol/water adsorption selectivity (α) was estimated for the fluoride-mediated silicalite-1. At 35 oC, α was estimated to be 36 for a 5 mol% ethanol solution in water increasing to 53 at an ethanol concentration of 1 mol%. The adsorption data demonstrate that silicalite-1 made via the fluoride-mediated route is a promising candidate for ethanol extraction from dilute ethanol-water solutions.

1 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 28

1. Introduction Zeolites are widely used as shape selective catalysts and as separation materials. All of these applications depend on the size and chemical nature of the porous network of the zeolite. The application of zeolitic materials for ethanol/water separation has made great progress during the past two decades.1, 2 LTA-type zeolites are ideally suited for the separation of water from organic solutions because they are hydrophilic and their aperture size (0.4 nm) is smaller than almost all organic molecules but larger than water. Zeolite A membranes have already been successfully applied for industrial-scale dehydration with low water content ethanol feeds to produce fuel-grade ethanol in both vapor permeation3 and pervaporation mode.4 However, in the case of an algae-based biofuel processes2,4 where ethanol is the minority component, it is more practical to apply hydrophobic zeolites for the removal of ethanol from water. Among the highly hydrophobic zeolite materials, the MFI-type structure, which includes Al-free silicalite-1 and Alcontaining ZSM-5, are the most widely investigated due to their appropriate pore size, excellent thermal and structural stability, and ease of synthesis. Silicalite-1 is expected to possess an ideally hydrophobic nature due to the elimination of acidic Al sites. Both silicalite-1 and ZSM-5 membranes have demonstrated the capability to ethanol removal from water, as reviewed by Bowen et al.5 The most promising ethanol/water separation performance reported in the literature for MFI-type hydrophobic membranes is for a silicalite-1 pervaporation membrane with a separation factor of 106 at 60 oC for a 5 wt % ethanol-95 wt % water feed.6 In general, the ethanol/water separation factors of high-quality silicalite-1 and highly siliceous ZSM-5 pervaporation membranes are 50 to 100, which are about two orders of magnitude lower than the water/ethanol separation factors of hydrophilic LTA zeolite membranes that are usually several thousands and sometimes over 10,000.5 The difference in separation performance mainly comes from the unfavorable water uptake in the zeolite framework (for the case of ethanol removal from water) and the larger pore size of MFI-type zeolites. For ZSM-5 samples, the strongly hydrophilic Al sites are the source of water attraction, while the existence of structural silanol defects

7

for the theoretically “ideal hydrophobic”

silicalite-1 also allows for some amount of water adsorption. The structural silanol defects (SiOH) consist of terminating silanol groups of the Si-O-Si network wherein oxygen atoms no longer bond to silicon atoms and also from the internal silanol groups where the Si-O-Si bonds 2 ACS Paragon Plus Environment

Page 3 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

are broken. In the latter case, it is also possible that silicon atoms detach from the zeolite framework during synthesis and a maximum four silanol groups can be correspondingly formed per each missing silicon atom and this structure represents a so called “silanol nest”. 8 The nearperfect hydrophobic silicalite-1 crystals can be obtained with an extremely low density of internal defects by using fluorine ions as the mineralizing agent at near neutral conditions, instead of the traditional OH- mineralizing agent at alkaline conditions. 9-11 The overall separation performance of MFI-type zeolites depends on the adsorption and diffusion of ethanol and water molecules into the zeolite framework. As such, it is clear that water and ethanol sorption properties play an important role in determining the separation characteristics of MFI-type zeolite materials. Although these adsorption phenomena are of great interest, we know of no systematic experimental studies and comparisons of water and ethanol adsorption on MFI-type zeolites with different hydrophobicity that includes silicalite-1 derived from a fluoride mediated route. In this work, the vapor adsorption of pure water and ethanol are tested on a series of MFI-type zeolites including silicalite-1 obtained from both the traditional alkaline route and the fluoride route, and ZSM-5 with different Si/Al ratios and different chargebalancing cations. Full isotherms in the temperature range of 25-55°C and activity range of 0.050.95 (where activity is defined as P/P0) are presented here and different models are applied to correlate the data for both water and ethanol adsorption. These adsorption isotherms can be a favorable complement to the body of literature on ethanol/water separations and serve as a reference for simulation studies.12-15 These adsorption isotherms can also provide guidance towards sorbent selection for ethanol-water separation systems such as temperature swing adsorption (TSA) based hollow fiber sorbent technology16, wherein sorbent particles are imbedded in a porous polymer matrix designed to facilitate the adsorption of the target species and exclude the competitive uptake of other species.

2. Experimental and Materials Six different MFI-type molecular sieves were used in the adsorption measurements. ZSM-5 zeolites (CBV 28014, Si/Al~140 and CBV 3024e, Si/Al~15, Zeolyst International) were received in their NH4+ form and are referred as NH4-ZSM-5(140) and NH4-ZSM-5(15). H-ZSM3 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

5(140) and H-ZSM-5(15) were produced by calcination of their NH4 counterparts in air at 823 K for 8h.17 The silicalite-1 samples were prepared in an alkaline medium; these are referred to as silicalite-1(OH-) or in a fluoride medium, which are referred to as silicalite-1(F-). Silicalite1(OH-) was acquired from Sigma-Aldrich and silicalite-1(F-) was prepared in our laboratory. The average particle size is 3-5 µm for silicalite-1(OH-) and 2-3 µm for ZSM-5. Silicalite-1(F-) was prepared in a fluoride-mediated route by hydrothermal synthesis adapted from literature procedures.9, 18, 19 The (OH-) and the (F-) naming system is based on reference to the mineralizing ion in the synthesis procedure. Tetrapropylammonium bromide (TPABr, 1.62 g, 99% Sigma Aldrich) and ammonium fluouride (NH4F, 0.116 g, >99.99% Sigma Aldrich, stored in a desiccator) were first dissolved in 26.88 g of DI water. The TPABr-NH4FH2O mixture was then transferred to a 40mL Teflon sleeve and 4.48 g of Cab-O-Sil M-5 (Cabot Corporation) was slowly added to the mixture, stirring manually for 10 minutes to form a homogenous gel. The Teflon sleeve was then placed in a tightly sealed stainless steel reactor and heated at 180°C for 14 days. The resulting solids were then vacuum filtered and washed with 200mL of water. To remove any un-reacted silica, 30mL of DI water was added to the solids and sonicated for 90 seconds. The slurry was then centrifuged and the water was decanted off. This sonication-centrifugation cycle was repeated a minimum of two more times. The crystals were then dried and calcined using the following profile: ramp 5°C/min, 120°C for 2h, ramp 5°C/min, 550°C for 12h. The F- ion is not incorporated into the silicate-1 framework. The reaction yielded 3.4 g, a 76% yield (based on silica) achieved with an average crystal size of 70 µm x 30 µm x 15 µm. (See Fig. S2 in the Supporting Information.) The pore structure of MFI zeolites was characterized using nitrogen as a probe gas at 77K via a Micrometrics ASAP 2020 adsorption porosimeter equipped with micropore analysis functions. Table 1 summarizes the pore structure information including BET surface areas and pore volumes, estimated by the Horvath-Kawazoe method. As expected, the pore volumes for the protonated ZSM-5 and silicalite samples are the same within experimental error. The NH4-ZSM5 samples show lower pore volume values due to the size and concentration of the NH4+ cations. The vapor adsorption experiments were performed on a VTI-SA vapor sorption analyzer from TA Instruments (New Castle, DE, United States) at temperatures of 25-55oC (water) or 2545oC (ethanol). The vapor activity was controlled automatically by mixing the wet vapor feed 4 ACS Paragon Plus Environment

Page 4 of 28

Page 5 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

with a dry N2 line. As such, N2 serves as a carrier gas for the vapors. The samples “dry mass” was measured under N2 and were at equilibrium with N2 before introduction of the vapors to the sample chamber. The isotherms are obtained for water and ethanol adsorption within 0.05-0.95 activity. Table 1 Pore structure information for MFI-type zeolites used in this study H-ZSM-5(15)

NH4-ZSM-5(15)

H-ZSM-5(140)

NH4-ZSM-5(140)

silicalite-1(OH-)

silicalite-1(F-)

379

361

378

370

379

381

0.195

0.157

0.192

0.174

0.193

0.196

2

BET (m /g) 3

Pore Volume*(cm /g)

* Horvath-Kawazoe method

3. Models and theory Different models are applied to correlate and analyze the sorption isotherms due to the dramatic differences in the sorption behavior with the different sorbate-sorbent systems. The Polanyi potential theory was used to fit the water adsorption isotherms on all MFI-type zeolite adsorbents studied, since it has been shown to be especially useful for adsorption on microporous materials, such as activated carbon with highly nonpolar and hydrophobic adsorption sites20, 21. The adsorption of organic species on organophilic adsorbents is usually of Langmuir type5, 22, so the Langmuir model is used for fitting the ethanol vapor adsorption isotherms on silicalite-1 samples in this study. Sircar’s model23, 24 is used to fit ethanol adsorption isotherms for the more hydrophilic ZSM-5 samples wherein multilayer adsorption is possible at higher activities. 3.1 The Polanyi potential theory In this theory, the adsorption potential ε is defined as the work done by the adsorption forces in delivering the molecules from the gas phase to the sorbed phase on the adsorbent surface.20 For one mole of ideal gas, the adsorption potential ε can be estimated as:

(1)

where P0 is the saturated vapor pressure. Since the adsorption potential is primarily related to temperature-independent forces, ε is largely a function of the volume of the adsorbed phase or a function of the mass amount of the adsorbed phase with the assumption that the density of the 5 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 28

adsorption phase is equal to the liquid density.25 Therefore, the adsorption potential can be rewritten:

(2)

where q is the adsorption amount (mmol/g sorbent or sorbent wt%). Since the adsorption

potential involves the effect of temperature, a plot of q vs. gives a unique temperature

independent relation for a fixed sorbate-sorbent system, which is also known as the “characteristic curve”. The potential theory allows for the prediction of an equilibrium adsorption isotherm at different temperatures once the characteristic curve is obtained. 3.2 The Langmuir model The Langmuir model is expressed as26

(3)

The saturation limit qs and the Henry’s constant K (=bqs) can be obtained by plotting 1/q vs 1/p in the linear form of Langmuir model:

(4)

Therefore, the saturation limit qs is the reciprocal of the intercept and the Henry’s constant is the reciprocal of the slope of equation (4). The parameter b is related to the heat of adsorption (∆Hads) and reflects the degree of adsorption strength on the sorbent surface. The adsorption heat can be calculated from:

∆"#$ %&

(5)

and is equivalent to the isosteric heat of adsorption as long as qs is independent of temperature. 3.3 The Sircar’s model The three-parameter Sircar’s model23, 24 assumes that a multilayer is formed at all activities and the surface area of each layer increases as the activity (and sorbent loading) increases. The model can be expressed as: 6 ACS Paragon Plus Environment

Page 7 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

' ⁄ *' + ⁄ , ( + ⁄ ,

(

∞ + ⁄ *'

(6)

where P/P0 is the activity; - is a temperature-dependent constant describing the sorbate-sorbent

interaction for the first layer and -∞ is the temperature-dependent constant describing the

sorbate-sorbent interaction for the subsequent layers. Sircar suggested that the parameters - and qs in the model can be obtained using low activity adsorption data by setting -∞ to zero so that

only a monolayer is assumed to be formed. Then equation (6) is expressed as a Langmuir-type equation as:

' ⁄

*'( + ⁄ (

(7)

,

Equation (7) can also be rearranged in its linear form:

'

(

∙

'( +

(8)

'(

Therefore, the parameters C1 and qs can be calculated from the slope and intercept values by plotting (1/q) vs (Po/P). The third parameter -∞ can be estimated by the best fit of all the data. In

Sircar’s model, the temperature function forms of - and -∞ are not defined. Therefore, they

have to be calculated separately for different temperatures. 3.4 The isosteric heat of adsorption The isosteric heat of adsorption is derived from the Clausius-Clapeyron equation as:27 01234 5

678 6&

(9)

9

where 01234 is the isosteric heat of adsorption and 9 indicates constant equilibrium adsorption

quantities. The isosteric heat of adsorption 01234 can be calculated by measuring adsorption isotherms at different temperatures and employing the thermodynamic relationship of equation (9). A plot of lnP

against (1/T) at constant adsorption uptakes yields a straight line with a slope equal to (ΔH/R).


Langmuir

4. Results and discussion 4.1 Water adsorption Water vapor adsorption isotherms at 35 oC for the MFI-types zeolites are illustrated in Figure 1. The tests at other temperatures in this study exhibit similar behaviors (adsorption isotherms at other temperatures can be found in the Supporting Information, Fig. S3). Silicalite-1 (OH-) and highly siliceous ZSM-5 (140) samples exhibit Type III isotherms according to BDDT classifications, whereas silicalite-1 (F-) exhibits a Type V isotherm due to its continued water uptake above saturation pressure.13, 15 These isotherms are formed due to stronger sorbate-sorbate interactions, i.e. water-water in this case, than sorbate-sorbent interactions where water molecules contact the hydrophobic zeolite framework. Therefore, the absolute value of adsorption enthalpy for the first layer is likely smaller than the liquefaction enthalpy of water, which explains the weak adsorption observed in the initial stage of adsorption at the low activity regions. As the adsorption proceeds, the adsorbed water molecules likely act as seeds for the formation of water clusters, which results in a self-enhanced adsorption phenomenon.

10

adsorption amounts (mmol/g adsorbent)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 28

H-ZSM5 (15) NH4-ZSM5 (15) 8

H-ZSM5 (140) NH4-ZSM5 (140) -

silicalite (OH )

6

-

silicalite (F ) 4

2

0 0.0

0.2

0.4

0.6

0.8

1.0

P/Po

Fig.1 Water adsorption isotherms for MFI-type molecular sieves at 35 oC Except for the silicalite-1 (F-), the data points are the average of two measurements, the error bars indicating the variation. The silicalite (F-) data points are the average of five measurements; the error bars indicate standard deviations which are smaller than the symbols.


Page 9 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

The increase in framework Al content enhances the zeolite’s hydrophicility and thus significantly promotes water adsorption as shown for ZSM-5 (15) in Fig.1. This has the additional effect of transforming the isotherm shape towards a Type-II isotherm with a steeper initial slope than a Type III or V isotherm. The dependence of water adsorption on framework aluminum content for zeolites has been well established and the general characteristic is that more water can be sorbed for zeolites with higher Al content due to the partially ionic and hydrophilic sites generated by the tetrahedrally coordinated trivalent Al.28-31 This expected adsorption behavior is observed in this work with the total water uptake being significantly higher for ZSM-5 (15) than for ZSM-5 (140) by roughly the factor of 10 expected from the difference in the concentration of acid sites. Furthermore, it was also found that the cations in the framework also affect the total water loading. For both ZSM-5 samples, the adsorbed water amounts were found to always be higher in their H-forms than their counterpart NH4-form at identical experimental conditions and this difference becomes larger when more acidic sites are present in the framework. The ammonium ion occupies more of the free space within the zeolite framework, as is already clear from Table 1. In fact, the sorption ratios for the ZSM-5(15) sample (about 1.2) and the ZSM-5(140) sample (about 1.1) are almost exactly the same for water and nitrogen (Table 1). The correlation of zeolite pore structure with this ion-inhibition effect is discussed later in the ethanol adsorption section, since similar sorption results also occur with ethanol adsorption in ZSM-5 (15) with different cations. The initial interaction stoichiometry between water and the hydrophilic Al sites can be estimated from the adsorption uptakes by extrapolating the first plateau region in the isotherm to infinitely small vapor pressure.32 The results are listed in Table 2 in terms of “H2O/Al” calculated by the ratio of the number of sorbed water molecules to the number of Al sites in one MFI unit cell. According to the definition of unit cell (UC) for MFI-type structures ([Mn+]x/nAlxSi96-xO192), Al/UC is an intrinsic property for ZSM-5 (15) and ZSM-5 (140), the values being 6 and 0.68 respectively. The H2O/UC ratio is obtained by the intercept of the line in the first plateau region at low activities. As shown in Table 2, the H2O/Al ratios for H-form ZSM-5 samples lie between 1.8 and 2, while the two NH4-form ZSM-5 ones have lower H2O/Al ratios of 1 to 1.3. These results indicate that there is likely a steric effect for the ammonium ions in ZSM-5 that hinders the sorbate-sorbent interaction at the onset of adsorption. Monte Carlo simulations reveal that the Al sites in Na-ZSM-5 zeolites coordinate two water molecules per site 9 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 28

at low coverage, which correspondingly promotes water clustering in the vicinity of these sites.7 This is consistent with the results for H-ZSM-5 in this study. It is also substantiated by temperature-programmed-desorption of water from an H-ZSM-5 (13) sample33 that the dimeric water cluster is formed at lower loadings while the tetrameric cluster tends to be developed at maximum uptake. Results for the two NH4 materials suggest closer to one water molecule per Al site. Table 2 Initial water cluster size for different MFI-type molecular sieves at different temperatures. Al/UC is the number of aluminum sites per unit cell. Also shown is the number of water molecules per unit cell and per aluminum site in the low pressure limit. Adsorbents

T/oC

Al/UC

H2O/UC

H2O/Al

H-ZSM5 (140)

25

0.68

1.2

1.8

35

0.68

1.2

1.8

45

0.68

1.2

1.8

25

0.68

0.89

1.3

35

0.68

0.84

1.2

45

0.68

0.80

1.2

25

6

12.0

2

35

6

11.0

1.8

45

6

10.7

1.8

25

6

6.4

1.1

35

6

6.4

1.1

45

6

6.2

1.0

25

N/A

0.12

N/A

35

N/A

0.11

N/A

45

N/A

0.11

N/A

25

N/A

0.57

N/A

35

N/A

0.56

N/A

45

N/A

0.51

N/A

NH4-ZSM5 (140)

H-ZSM5 (15)

NH4-ZSM5 (15)

-

Silicalite (F )

-

Silicalite (OH )

Both silicalite-1 samples exhibit lower water uptakes than ZSM-5. The most hydrophobic sorption response observed occurs in the fluoride-mediated silicalite-1 sample. For example, at the highest activity studied in this work (0.95) at 35oC, the adsorption uptake in silicalite-1 (F-) is 0.263 mmol/g, accounting for only 12.6%, 9.8% and 3.3% of the capacity for silicalite-1 (OH-), H-ZSM-5 (140) and H-ZSM-5 (15) at the same condition, which corresponds to 2.08, 2.68 and 8.04 mmol/g, respectively. Note that the highly siliceous ZSM-5 (140) is already fairly 10 ACS Paragon Plus Environment

Page 11 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

hydrophobic as evidenced by the relatively small water adsorption amounts. The hydrophobic nature of pure-silica silicalite-1 (F-) becomes clearer when comparing the two silicalite-1 samples. At an activity of 0.05 and 35°C, the water uptake is only 0.024 mmol/g for silicalite-1 (F-) whereas the water uptake is 0.138 mmol/g for silicalite-1 (OH-). At more saturated conditions (0.90 activity, 35°C), silicalite-1 (F-) adsorbs 0.176 mmol water per gram, which is approximately one tenth of the measured capacity of silicalite-1 (OH-), 1.75 mmol/g. To the best of our knowledge, the silicalite-1 (F-) in this study demonstrates the lowest water uptake ever experimentally reported for adsorption in MFI-type zeolites. The hydrophobicity of silicalite-1 (F-) can also be demonstrated by the initial water cluster size listed in Table 2. The H2O/UC values for silicalite-1 samples are obtained by extrapolating the first plateau region in the isotherms to zero pressure, which is the same method as discussed previously for the H2O/Al estimation. The initial H2O/UC for silicalite-1 (F-) is about 0.12 at 25 oC. The distinct water adsorption properties for the two silicalite-1 zeolites can be attributed to the difference in their internal structural silanol defects. Though both would have external terminating silanol defects, their impact for the relatively large crystals in this study is expected to be small.38 The ideal hydrophobic silicalite-1 (F-) crystals can be obtained with fewer internal defects by using fluoride as the mineralizing agent at near neutral conditions, while the silicalite1 prepared by the traditional routes at alkaline conditions with OH- as the mineralizing agent usually have considerable internal silanol defects created on removal of the charge balancing centers for the template tetrapropylammonium cations (i.e., nominally 4 per UC).34 In the fluoride route, the F- can offset the template ions in a way that generates few internal defects.34 The structural silanol defects can be identified by the nuclear magnetic resonance (NMR) technique. By comparing 29Si cross-polarization magic-angle-spinning (CP-MAS) NMR spectra with its corresponding 29Si MAS NMR spectra, the existence of small amounts of defects (nondetectable in 29Si MAS NMR) can be revealed by an enhanced peak in 29Si CP-MAS NMR that results from the cross polarization of Si-OH with 1H. The fluoride-derived silicalite-1 shows no additional peak in its

29

Si CP-MAS NMR spectra which implies the presence of negligible

amounts of silanol defects.10, 35 On the contrary, the silicalite-1 synthesized by the traditional alkaline method usually involves the appearance of extra peaks that are ascribed to the ≡Si-O-framework defects. Moreover, for the commercial silicalite-1 (OH-) sample from Aldrich, it has been previously found that more hydrophilic defects (≡Si-O--Na+-(H2O)n groups) are also 11 ACS Paragon Plus Environment

Langmuir

present.12 While these NMR experiments are not conducted in this study, similar trends are expected to hold for our materials. The difference in the silanol defect densities and/or the defect types accounts for the different water adsorption behaviors for silicalite-1 (F-) and silicalite-1 (OH-) shown in Fig.1, especially at activities higher than 0.6. Figure 2 summarizes the water vapor adsorption isotherms for silicalite-1 molecular sieves in the literature32, 36, 37 and the ones obtained in this study. It is evident that there is a discrepancy in the water adsorption amounts for different silicalite-1 samples, which could be due to the difference in the density, type and distribution of structural defects. Cheng et al.38 functionalized the internal surface of silicalite-1(OH-) with aliphatic alcohols and substantiated a decreased water uptake on the 1-butanol-functionalized materials due to the esterification reaction that converted the hydrophilic defect sites into hydrophobic ones. Their work was consistent with the expected 4 silanol nests/UC.

silicalite (F-) this work

2.5

adsorption amounts (mmol/g)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 28

silicalite (OH-) this work Thompson et al.[36] Olson et al. [32] Oumi et al. [37]

2.0

1.5

1.0

0.5

0.0 0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

P/Po

Fig.2 Water adsorption isotherms for silicalite-1 molecular sieves. The three literature data sets are all nominally silicalite-1 (OH-) materials.

The existence of these silanol defects also makes it difficult to simulate the water vapor adsorption behavior at the molecular level. In a grand canonical Monte Carlo (GCMC) simulation with a defect-free silicalite-1 model, the water vapor uptake can be either significantly overestimated or underestimated compared to experimental results.14 For a more accurate 12 ACS Paragon Plus Environment

Page 13 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

simulation, the internal silanol defects in silicalite-1 framework are taken into account in a series of recent publications. Trzpit et al.12 created “strong” and “weak” silanol nests by modifying the partial charges placed on Si, O and H atoms for water adsorption. Ahunbay7 and Yazaydin et al.8 both reported the critical silanol defect concentration of two silanol nests per UC above which the water adsorption increased significantly compared to the defect-free silicalite-1 model. However, their simulated water adsorption experiments at high vapor pressures exhibit a difference of one order of magnitude for identical defect concentrations of four silanol nests/UC. The different potential models used in the simulation may result in this inconsistency as suggested by the authors. Nevertheless, another factor is the silanol defect distribution throughout the zeolite framework. The silanol defect is thought to originate from the template with a theoretical value of four TPA cations per UC in the as-made MFI structures.29 However, it is found that silicalite usually contains a variable amount of internal defects, whose concentration depends on the concentration of Na and Al impurities acting as mineralizing agents.39 Bordiga et al.39 identified different types of silanol defects in silicalites by a detailed infrared spectroscopy study, which involves isolated silanol, free silanol in cluster and silanols interacted with hydrogen bonds. Moreover, all these defect species can be internal silanols with a distribution. Therefore, it is reasonable to conclude that the defect distribution also plays an important role in determining the water adsorption. A significant increase in water uptake is likely to occur when the silanol defects are close enough that the initially adsorbed water molecules around silanol defects can act as seeds to attract more water molecules. The increase of defect quantity alone may not always lead to enhanced water adsorption when the silanol defects are not adjacent to each other. This subtlety can also be reflected by the experimental data in Fig.1 where the water uptakes for silicalite-1 (OH-) are systematically lower than those for ZSM-5 (140). Note that the Al concentration of ZSM-5 (140) is 0.68 Al/UC and the Al sites are more hydrophilic than the silanol defects.7 Therefore, the silanol defect concentration in silicalite-1 (OH-) should theoretically be smaller than 0.68 per UC. However, at this small defect concentration, water condensation occurred within silicalite-1 (OH-) at high activities in this study. But this has not been revealed by those simulation studies in which a significant increase in water adsorption only happens to silicalite-1 with defect concentrations higher than two silanol nests per UC (four silanol groups per silanol nest). This may be due to the presence of ≡Si-O--Na+-(H2O)n groups in silicalite-1 (OH-), which are believed to be more hydrophilic than silanol defects. Therefore, it is 13 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 28

still a challenging task to identify an ideal simulation model for the water/silicalite-1 system due to the intricacy of defect density, defect type and distribution of the experimentally obtained silicalite-1 zeolite. The isosteric heats of adsorption for the six MFI-type zeolites are calculated and summarized in Fig.3. Our results indicate that for silicalite-1 samples, the heats of adsorption increase with increasing adsorption amounts, which reflects their hydrophobic nature. The low heats of sorption imply that low-activity water physisorption within defect-free silicalite-1 (F-) is only marginally thermodynamically spontaneous. For the defect-free silicalite-1 (F-), the initial adsorbed water at low loadings is attributed mainly to physisorption of the water molecules within the hydrophobic framework. The physisorbed water molecules then act as seeds for the subsequent water clustering leading to stronger water-water interactions that result in a significant increase in the heat of adsorption with increasing water loading until the zeolite framework is homogeneously covered by water molecules. As shown in Fig.3, the heat of adsorption of silicalite-1 (F-) increases from 13.9 kJ/mol at the water loading of 0.025 mmol/g, which is in the range of physisorption values, to the final stable value of about 44 kJ/mol. The variation of adsorption heats for silicalite-1 (OH-) has a similar trend with that of silicalite-1 (F-) but exhibiting an obviously higher heat of adsorption initially and a less steep slope, likely due to the existence of silanol defects which act as weak hydrophilic sites. Contrary to the trend for silicalite-1, the isosteric heats of adsorption decrease with increasing water loading when the strongly hydrophilic Al sites are introduced into the zeolite framework. At low activities, strongly exothermic adsorption preferentially occurs at the acidic alumina sites that have much higher sorbate-sorbent interaction energy than the sorbate-sorbate interaction energies. As the water activity increases, the water molecules interact either with previously adsorbed molecules around Al sites or with the less-energetic framework similar as the physisorption period for pure-silica silicalite-1, which in both cases lead to a gradual decrease in the adsorption heats until the zeolite structure is homogeneously covered. This same trend is clearly shown in Fig.3 for ZSM-5 (140) in both its NH4 and H form. However, the heat of adsorption at low activities cannot be calculated due to the testing condition limitations of our sorption apparatus. The lower limit of controllable activity for the vapor adsorption equipment is 0.02, an activity at which significant amounts of water have already been adsorbed for the ZSM14 ACS Paragon Plus Environment

Page 15 of 28

5 samples used in this work, especially for the one with a higher density of Al sites, i.e. the ZSM-5 (15). For this sorbent, the initial water uptake at an activity of 0.02 is over 1.1 mmol/g, which makes the calculation of its adsorption heats at lower loadings unfeasible. A previous study on the interaction of water with H-ZSM-5 showed that the heat of adsorption of a H-ZSM5 (37.5) sorbent can be even higher than 100 kJ/mol initially,32 which reflects the highly exothermic nature of the interaction of water with unhydrated ions.

50

Isosteric heat of adsorption (kJ/mol)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

40

30

H-ZSM5 (15) NH4-ZSM5 (15) H-ZSM5 (140) NH4-ZSM5 (140)

20

-

Silicalite (OH ) -

Silicalite (F ) 10 0

1

2

3

4

5

Q (mmol/g)

Fig.3 Isosteric heats of adsorption for water adsorption in different MFI-type molecular sieves. Each data point represents the average of at least two determinations. For clarity, error bars reflecting the minor variability of the duplicates have been omitted.

We find that the heats of adsorption for all MFI-type zeolites converge as the water loadings increase to stable values of 43-46 kJ/mol, which lies well within the range of the enthalpy of water condensation (44.0 kJ/mol at 25oC; 42.7 kJ/mol at 55oC). The same convergence occurs for both silicalite-1 forms and ZSM-5 forms. This is generally consistent with Monte Carlo simulation results of water adsorption in hydrophobic MFI zeolites; the adsorption heats of defect-free silicalite-1 and Al-containing Na-ZSM-5 exhibit an opposite variation trend but finally converge with the increasing water loadings.7 Finally, the water vapor uptake for the MFI-type molecular sieves at all the temperatures and activities tested in this work are plotted as a function of the adsorption potential (See Fig. S1, 15 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 28

supporting information). This demonstrates that the adsorption data for each sorbent conform to one respective characteristic curve, which indicates that the potential theory model can reasonably describe the water-MFI sorption system in this study. As discussed earlier, the primary advantage of the potential theory is that the characteristic curve is temperature independent and one can easily predict the adsorption uptake at any activity and temperature by simply calculating the corresponding adsorption potential.

4.2 Ethanol adsorption Figure 4 shows ethanol vapor adsorption isotherms at 35 oC for the MFI-type molecular sieves. (Adsorption isotherms at other temperatures can be found in the Supporting Information, Fig. S4) The MFI-type zeolites studied here exhibit an overall organophilic nature where the near-saturation adsorption coverage is typically achieved at pressures far below the saturation vapor pressure. On the contrary, water uptake into these zeolites tends to increase with pressure up to the saturation pressure (Fig.1). These experiments also reveal that the difference in the absolute ethanol uptake for these zeolites is not nearly as significant as for the case where water is the sorbate. The similarity between ethanol and water adsorption responses is that both ethanol and water molecules tend to be preferentially adsorbed on the hydrophilic sites in the zeolite structure. Therefore, for pure vapor adsorption tests, the MFI-type molecular sieve with higher Al contents should in theory adsorb more ethanol than the one with lower Al sites due to the small induced dipole ethanol possesses. In this way, the expected sequence of ethanol uptake at the same activity and temperature should be ZSM-5 (15) > ZSM-5 (140) > silicalite-1. The isotherms in Fig.5 indeed demonstrate this relationship between adsorption amounts and Al contents as H-ZSM-5 (15) > H-ZSM-5 (140) > silicalite-1. The only exception is ZSM-5 (15) in its ammonium form, which possesses an obviously lower ethanol adsorption capacity, similar to its water adsorption response (Fig.1). The ammonium ions likely restrain ethanol adsorption and this effect becomes more obvious at low Si/Al ratios, as in the case of the ZSM-5 (15). As shown in Table 1, the existence of ammonium ions reduces the pore volume of NH4-ZSM-5 (15) by approximately 20% compared to H-ZSM-5 (15). This steric inhibition on adsorption is quantitatively related to the pore volume. Ethanol uptake in H-ZSM-5 (15) was found to be about 20% higher than NH4-ZSM-5 (15). The water adsorption response on H-ZSM-5 (15) and NH416 ACS Paragon Plus Environment

Page 17 of 28

ZSM-5 (15) exhibits similar behavior and the percentage differences in adsorption amounts becomes stable around 20% when the activity is higher than 0.4.

3.5

adsorption amounts (mmol/g adsorbent)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

3.0 2.5 2.0 1.5 1.0

H-ZSM5 (15) NH4-ZSM5 (15)

0.5

H-ZSM5 (140) NH4-ZSM5 (140) -

silicalite (OH ) 0.0

-

silicalite (F ) 0.0

0.2

0.4

0.6

0.8

1.0

P/Po

Fig. 4 Ethanol adsorption isotherms for MFI-type molecular sieves at 35 oC.The data points are the average of two measurements, the error bar indicating the variation.

The silicalite-1 (F-) and silicalite-1 (OH-) samples display practically identical ethanol adsorption responses due to the overall organophilic nature of the pure-silica MFI-type framework. The weakly hydrophilic structural defects in silicalite-1 (OH-) apparently have no discernable influence on ethanol loadings even though their presence affects the water adsorption significantly. The isotherms for both silicalite-1 (F-) and silicalite-1 (OH-) are of the Type-I form which can be described by the ideal Langmuir model (Fig. 5). As shown in Table S1 (supporting information) lists the Langmuir model parameters for ethanol adsorption in silicalite-1 (F-) and silicalite-1 (OH-). The saturation limit (qs) values lie in 11.3-11.7% (sorbent wt%) and slightly decrease with increasing temperature. For non-porous and macroporous adsorbents, the saturation limit is equivalent to the monolayer coverage on the surface of adsorbents, provided capillary condensation does not occur in the case of macroporous adsorbents. When used to describe microporous materials, the Type-I isotherm reflects micropore filling within the


Langmuir

microporous adsorbents such as MFI-type molecular sieves, where the saturation limit corresponds to the total volume of micropore filling. Therefore, the saturation limit-which is ideally the maximum amount of molecules adsorbed into the pore volume of the sorbent-is expected to be temperature-independent as long as the zeolite structure and the total pore volume do not change with temperature. The saturation limit is only expected to be dependent on the sorbate molecular size.5 The slight decrease observed in the saturation limit with increasing temperature could likely be attributed to the much greater coefficient of expansion of the sorbate (ethanol) than that of the sorbent (zeolite framework).22 The adsorption heats of ethanol on silicalite-1 (F-) and silicalite-1 (OH-) are 35.8 kJ/mol and 36.5 kJ/mol respectively, as calculated by equation (5).

12

10

8

q (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 28

6

4

2

-

0

silicalite (F ) Experiment 0.0

0.2

0.4

0.6

0.8

1.0

P/Po

Fig. 5 The experimental data and Langmuir fit for ethanol adsorption in silicalite-1(F-) at 35 oC. The experimental points are the average of two determinations.


Page 19 of 28

0.096

-1

0.092

1/q (wt%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

0.088

0.084

0.080

25C 30C 35C 40C

0.076

0.072 0.0

0.5

1.0 3

1.5 -1

2.0

2.5

(1/P) × 10 (Pa )

Fig. 6 The (1/q) vs (1/p) plots for the Langmuir fit of H-ZSM-5(15) adsorption responses (average of two determinations)

For the ZSM-5 molecular sieves, the ethanol adsorption isotherms do not follow the Langmuir model at activities higher than 0.5; rather, there is a concave-up adsorption response at higher activities (Fig. 4). The relationship of 1/q versus 1/p for ethanol adsorption into H-ZSM-5 (15) is shown in Fig. 6 and is also representative for all the other ZSM-5 samples tested in this study. It is evident that the Langmuir model does not provide a satisfactory description of the ethanol adsorption isotherms due to the obvious curvature at higher vapor pressures. As discussed above, the Type-I isotherm generally reflects the micropore filling phenomenon on micropore sorbents with the saturation limit being the micropore volume. Multilayer adsorption is possible when the activity is further increased after saturation limit is achieved, most likely due to a densification of the sorbate phase towards a more liquid-like density. To mitigate the overprediction of the BET model at higher activities, Sircar23, 24 developed a three parameter isotherm equation with the assumption that a multilayer is formed at any activity and the surface area of each layer increases as the vapor activity increases, as shown in equation (6). In Fig. 7, the ethanol adsorption isotherms and modeling are illustrated for H-ZSM-5. Sircar’s model provides a fairly accurate description of the isotherms with only a slight overprediction at high ethanol activities for H-ZSM-5 (15). 19 ACS Paragon Plus Environment

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Ethanol adsorption amounts (mmol/g)

Langmuir

Page 20 of 28

3.0

2.5

2.0

1.5

1.0

0.5

H-ZSM-5 (15) experiment H-ZSM-5 (140) experiment

0.0 0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

P/Po

Fig. 7 Sircar model fit for ethanol adsorption in H-ZSM-5 molecular sieves at 35 oC. The experimental points are the average of two determinations.

Table S2 (supporting information) lists the Sircar Model parameters for ethanol adsorption on H-ZSM-5 molecular sieves obtained as described earlier in reference to Equations (6) and (8). It should be noted that these parameters work very well for reproducing the experimentally obtained ethanol adsorption isotherms, as shown in Fig. 7. However, direct quantitative comparison is problematic, possibly due to the difficulty in determining the three parameters precisely. In this model, the parameters - and -∞ indicate the strength of the adsorption and higher values reflects stronger bonding. In this way, these two parameters are expected to decrease with increasing temperature. In Table S2, it seems that only - for H-ZSM-

5 (15) follows this temperature dependence, while - values for H-ZSM-5 (140) are more or less

the same despite any changes in temperature. Moreover, there is no evident temperature dependence of -∞ for both H-ZSM-5 samples with relative larger standard deviations compared

to those for - . Sircar suggested that the parameters in the model can be estimated using adsorption data at low activities by setting -∞ to be zero where only a monolayer is assumed to

form and simplify the model to its Langmuir type, as in equation (8). Therefore, the parameters - and qs in the model can then be obtained and the third parameter -∞ can be determined by the

best fit of the adsorption data using the - and qs values. The parameters in Table S2 are calculated by this method and the calculation starts with the low activity data (0.05-0.3). 20 ACS Paragon Plus Environment

Page 21 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

However, our results indicate that a slight change in qs could significantly alter the - values without affecting the overall regression of the data. For example, if the calculation starts with a narrower activity window of 0.05-0.25 or 0.05-0.2, the - values can change significantly but still provide accurate isotherm reproduction with a relatively invariable value for qs. The unexpected temperature dependence of -∞ might be due to the model assumption that the sorbate-sorbent interaction energy is constant after the first layer and the available adsorption sites are fixed. For the H-ZSM-5 molecular sieves with Al sites, it is plausible that the adsorption energies are different for different layers because of the heterogeneity of the framework incorporated with acidic H (Al) sites. In this way, it is suggested that the model can still have a reasonable prediction of low activity vapor sorption before the saturation limit (monolayer) is reached, but may show deviations at high activities since -∞ is not really constant.24 The estimation of - also has an influence on the determination of -∞ . Despite the challenges in accurately determining the model parameters, Sircar’s model is still advantageous because of its ability to accurately predict ethanol/ZSM-5 adsorption isotherms in a concise and physically meaningful way.

4.3 Ideal Ethanol/water sorption selectivity The water and ethanol vapor adsorption isotherms reported here indicate that hydrophobic MFI-type zeolites like silicalite-1 (F-), silicalite-1 (OH-) and ZSM-5 (140) have the potential to be applied for ethanol removal from water processes due to their unfavorable water adsorption response. The ideal vapor phase ethanol/water sorption selectivity (α) at 35 oC is estimated using the pure-vapor adsorption isotherms within the ethanol concentration range (mol %) of 1-5% in water. The ethanol/water sorption selectivity is defined as: 40 ?

'@ABC,EFGH$ ⁄'I#AHF,EFGH$

(10)

'@ABC,J#KEF ⁄'I#AHF,J#KEF

where C is the concentration of ethanol or water in the adsorbed phase or in the vapor.


Langmuir

100

Ideal ethanol/water sorption selectivity

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 28

10

-

silicalite-1(F ) -

silicalite-1(OH ) H-ZSM5(140)

1 1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

Ethanol concentration (mol%)

Fig. 8 Ideal ethanol/water sorption selectivity at 35 oC for different hydrophobic MFI-type zeolites

Raoult's law and Henry’s law are applied to estimate the vapor pressures of water and ethanol, respectively. The Henry’s law constant of ethanol solubility in water is derived as: 41 LC & C 5MN.P

L

Q ∙ & R 5MN.P

(11)

where Tc is the temperature dependence constant (K) and Tc is 6600 K, KH(298.15) is the Henry's law constant for ethanol in water at 298.15 K (mol/kg.bar) and KH(298.15) equals to 190 mol/kg.bar. As shown in Fig. 8, silicalite-1 (F-) demonstrates excellent ethanol removal capability and its ideal ethanol/water sorption selectivity is an order of magnitude higher than those for silicalite-1 (OH-) and H-ZSM-5 (140). Furthermore, the sorption selectivity increases from 36 to 53 when the ethanol concentration decreases from 5 to 1 mol% (balance water), which is advantageous for the ethanol removal in algae-based biofuel technologies since the raw ethanol product is usually quite dilute.42 Silicalite-1 (OH-) and H-ZSM-5 (140) zeolites show mild ideal sorption selectivities of 4.5-8.3 and 3.8-7.0 in the same liquid concentration range (1-5 mol%) of ethanol in water. It is worth mentioning the difference between ethanol/water separation factors and sorption selectivities discussed above. For a pervaporation membrane process, an ethanol/water separation factor can be defined as43: 22 ACS Paragon Plus Environment

Page 23 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

ST⁄U

VH ⁄VI WH ⁄WI

∗

Y

?T⁄U ∙ H∗ ∙ H I

(12)

YI

where ye is the mole fraction of ethanol vapor in a permeate stream, yw is the mole fraction of water vapor in a permeate stream, xe is the mole fraction of ethanol in a feed liquid, xw is the mole fraction of water in a feed liquid, ?T⁄U is the permeability selectivity (that can be obtained ∗

Y

by sorption selectivity times diffusion selectivity), H∗ H is the volatility factor, P* is the Y I

I

saturation pressure, and γ is the activity coefficient. Therefore, included within the separation factor are the permselectivity and relative volatilities of the two components. Since ethanol has larger saturation pressure and larger activity coefficients (for dilute ethanol solutions) than water, the ethanol/water separation factor is usually one order of magnitude larger than permeability selectivity. For example, we report here an ethanol/water vapor sorption selectivity of 36 in silicalite-1(F-) for 95 mol%-5 mol% water-ethanol mixture at 35oC. Assuming an ethanol/water diffusion selectivity of 1.0 (we actually find 2.0 from the time responses) and recognizing that the volatility factor is 11.6 at this temperature, we find an ethanol/water separation factor of 418. While binary vapor sorption is beyond the scope of this investigation, for the nearly defect-free silicalite-1 (F-), we expect the competitive ethanol/water sorption selectivity to be lower than the pure-vapor sorption selectivity with the hypothesis that the sorbed ethanol molecules will act as water sorption site “seeds” due to the adsorbed ethanol molecule’s exposed -OH tail. However, it is still evident that silicalite-1(F-) is quite promising for ethanol removal from water and has potential application in temperature swing adsorption (TSA) hollow fiber sorbent technology.16, 44, 45

5. Conclusions Water and ethanol vapor adsorption phenomena are investigated systematically on six different MFI-type zeolites including silicalite-1 (F-), silicalite (OH-), H-ZSM-5 (140), NH4ZSM-5 (140), H-ZSM-5 (15) and NH4-ZSM-5 (15) at temperatures of 25-55 oC. Increasing framework Al content enhances the zeolite hydrophicility and the corresponding water uptake. The purely-siliceous silicalite-1 molecular sieves have lower water uptakes than ZSM-5 ones. The silicalite-1 (F-) exhibits ideal hydrophobic nature with the lowest observed water uptakes


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 28

due to the reduction of internal defects by using F- as the mineralizing agent at near neutral synthesis conditions, while the silicalite-1 prepared by the traditional route using alkaline conditions with OH- as the mineralizing agent usually have considerable internal silanol defects. At 0.9 water activity at 35°C, silicalite-1 (F-) adsorbs 0.176 mmol water per gram, approximately 10% and 3% of the total capacity of the silicalite-1 (OH-) and ZSM-5 (15), respectively. While the water adsorption response shows distinct vapor uptakes for different MFI-type zeolites due to the difference in the nature of hydrophilic sites, the ethanol adsorption presents relatively comparable results due to the overall organophilic nature of the zeolite framework. Thus the silicalite-1 (F-) is very promising for dilute ethanol extraction from water because due to extremely unfavorable water adsorption but comparable ethanol uptakes compared to other MFItype hydrophobic molecular sieves. An excellent ideal vapor phase ethanol/water sorption selectivity of 36 is shown for silicalite-1 (F-) at 35 oC for a 5 mol% ethanol solution in water and the ideal selectivity increases to 53 when the ethanol concentration decreases to 1 mol%. In the same liquid concentration range (1-5 mol%) of ethanol solution in water, the ideal sorption selectivity lies in 4.5-8.3 and 3.8-7.0 for the silicalite-1 (OH-) and H-ZSM-5 (140) zeolites.

Acknowledgements This material is based upon work supported by the Department of Energy under Award Number DE-FOA-0000096 and also supported by Algenol Biofuels. The authors thank Joshua Thompson and Megan Lydon for useful discussions on the synthesis of silicalite-1 using the fluoride mediated route. W.J. Koros thanks Award no. KUS-I1-011-21 made by King Abdullah University of Science and Technology (KAUST) for financial support. Disclaimer: This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, 24 ACS Paragon Plus Environment

Page 25 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. Supporting Information The supporting information contains the characteristic curves for water uptake into the MFI-type samples, SEM images of the silicalite-1(F-), water isotherms on MFI-type zeolites at 25-55 oC, ethanol adsorption isotherms on MFI-type zeolites at 25-45 oC, Langmuir parameters for ethanol sorption in silicalite-1, and Sircar’s model parameters for ethanol in ZSM-5. This information is available free of charge via the Internet at http://pubs.acs.org/. Corresponding author *Tel.: +1 404 385 1931. Fax: +1 404 385 2683. E-mail: [email protected]

References 1. Caro, J.; Noack, M.; Kolsch, P.; Schafer, R., Zeolite membranes - state of their development and perspective. Microporous Mesoporous Mater. 2000, 38, (1), 3-24. 2. Caro, J.; Noack, M., Zeolite membranes - Recent developments and progress. Microporous Mesoporous Mater. 2008, 115, (3), 215-233. 3. Masanobu Aizawa; Suguru Fujita; Yoshinobu Takaki; Kazuhiro Yano; Shimizu, T.; Toshihiro Asari; Jun Yano; Takaharu Yagi; Sawamura, K.; Shinoya, K. In Development of Hitz zeolite membrane dehydration system for bio-Ethanol production, AIChE annual meeting, Salt Lake City, UT, 2010; Salt Lake City, UT, 2010. 4. Morigami, Y.; Kondo, M.; Abe, J.; Kita, H.; Okamoto, K., The first large-scale pervaporation plant using tubular-type module with zeolite NaA membrane. Sep. Purif. Technol. 2001, 25, (1-3), 251260. 5. Bowen, T. C.; Noble, R. D.; Falconer, J. L., Fundamentals and applications of pervaporation through zeolite membranes. J. Membr. Sci. 2004, 245, (1-2), 1-33. 6. Lin, X.; Chen, X. S.; Kita, H.; Okamoto, K., Synthesis of silicalite tubular membranes by in situ crystallization. AIChE Journal 2003, 49, (1), 237-247. 7. Ahunbay, M. G., Monte Carlo Simulation of Water Adsorption in Hydrophobic MFI Zeolites with Hydrophilic Sites. Langmuir 2011, 27, (8), 4986-4993. 8. Yazaydin, A. O.; Thompson, R. W., Molecular simulation of water adsorption in silicalite: Effect of silanol groups and different cations. Microporous Mesoporous Mater. 2009, 123, (1-3), 169-176.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

9. Chezeau, J. M.; Delmotte, L.; Guth, J. L.; Gabelica, Z., Influence of synthesis conditions and postsynthesis treatments on the nature and quantity of structural defects in highly siliceous MFI zeolitesA high-resolution solid-state Si-29 NMR-study. Zeolites 1991, 11, (6), 598-606. 10. Axon, S. A.; Klinowski, J., Nuclear-magnetic-resonance studies of the synthesis of zeolite ZSM-5 by the fluoride method. Appl. Catal. A-Gen. 1994, 111, (1), 29-39. 11. Gualtieri, M. L.; Gualtieri, A. F.; Prudenziati, M., Seeded growth of TPA-MFI films using the fluoride route. Microporous Mesoporous Mater. 2008, 111, (1-3), 604-611. 12. Trzpit, M.; Soulard, M.; Patarin, J.; Desbiens, N.; Cailliez, F.; Boutin, A.; Demachy, I.; Fuchs, A. H., The effect of local defects on water adsorption in silicalite-1 zeolite: A joint experimental and molecular simulation study. Langmuir 2007, 23, (20), 10131-10139. 13. Eroshenko, V.; Regis, R. C.; Soulard, M.; Patarin, J., Energetics: A new field of applications for hydrophobic zeolites. J. Am. Chem. Soc. 2001, 123, (33), 8129-8130. 14. Desbiens, N.; Boutin, A.; Demachy, I., Water condensation in hydrophobic silicalite-1 zeolite: A molecular simulation study. J. Phys. Chem. B 2005, 109, (50), 24071-24076. 15. Desbiens, N.; Demachy, I.; Fuchs, A. H.; Kirsch-Rodeschini, H.; Soulard, M.; Patarin, J., Water condensation in hydrophobic nanopores. Angew. Chem.-Int. Edit. 2005, 44, (33), 5310-5313. 16. Lively, R. P.; Chance, R. R.; Kelley, B. T.; Deckman, H. W.; Drese, J. H.; Jones, C. W.; Koros, W. J., Hollow Fiber Adsorbents for CO2 Removal from Flue Gas. Ind. Eng. Chem. Res. 2009, 48, (15), 7314-7324. 17. Bowen, T. C.; Vane, L. M., Ethanol, acetic acid, and water adsorption from binary and ternary liquid mixtures on high-silica zeolites. Langmuir 2006, 22, (8), 3721-3727. 18. Chezeau, J. M.; Delmotte, L.; Guth, J. L.; Soulard, M., High-resolution solid-state Si-29 and C-13 NMR on highly siliceous MFI-type zeolites synthesized in nonalkaline fluoride medium. Zeolites 1989, 9, (1), 78-80. 19. Fyfe, C. A.; Brouwer, D. H.; Lewis, A. R.; Chezeau, J. M., Location of the fluoride ion in tetrapropylammonium fluoride silicalite-1 determined by H-1/F-19/Si-29 triple resonance CP, REDOR, and TEDOR NMR experiments. J. Am. Chem. Soc. 2001, 123, (28), 6882-6891. 20. Yang, R. T., Gas separation by adsorption processes. Imperial College Press: 1987; Vol. 1. 21. Hashi, M.; Tezel, F. H.; Thibault, J., Ethanol Recovery from Fermentation Broth via Carbon Dioxide Stripping and Adsorption. Energy Fuels 2010, 24, 4628-4637. 22. Cavalcante, C. L.; Ruthven, D. M., Adsorption of branched and cyclic paraffins in silicalite. 1. Equilibrium. Ind. Eng. Chem. Res. 1995, 34, (1), 177-184. 23. Sircar, S., New isotherm for multilayer adsorption of vapors on nonporous adsorbents. Adsorpt. Sci. Technol. 1985, 2, 23. 24. Lee, J. Y.; Westgate, P. J.; Ladisch, M. R., Water and ethanol sorption phenomena on starch. AIChE Journal 1991, 37, (8), 1187-1195. 25. Chang, H.; Yuan, X. G.; Tian, H.; Zeng, A. W., Experimental investigation and modeling of adsorption of water and ethanol on cornmeal in an ethanol-water binary vapor system. Chem. Eng. Technol. 2006, 29, (4), 454-461. 26. Ruthven, D. M.; Kaul, B. K., Adsorption of aromatic hydrocarbons in NaX zeolite. 1. Equilibrium. Ind. Eng. Chem. Res. 1993, 32, (9), 2047-2052. 27. Sircar, S., Recent developments in macroscopic measurement of multicomponent gas adsorption equilibria, kinetics, and heats. Ind. Eng. Chem. Res. 2007, 46, (10), 2917-2927. 28. Chen, N. Y., Hydrophobic properties of zeolites. J. Phys. Chem. 1976, 80, (1), 60-64. 29. Flanigen, E. M.; Bennett, J. M.; Grose, R. W.; Cohen, J. P.; Patton, R. L.; Kirchner, R. M.; Smith, J. V., Silicalite, a new hydrophobic crystalline silica molecular-sieve. Nature 1978, 271, (5645), 512-516. 30. Olson, D. H.; Haag, W. O.; Lago, R. M., Chemical and physical properties of the ZSM-5 substitutional series. J. Catal. 1980, 61, (2), 390-396.


Page 26 of 28

Page 27 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

31. Nakamoto, H.; Takahashi, H., Hydrophobic natures of zeolite ZSM-5. Zeolites 1982, 2, (2), 6768. 32. Olson, D. H.; Haag, W. O.; Borghard, W. S., Use of water as a probe of zeolitic properties: interaction of water with HZSM-5. Microporous Mesoporous Mater. 2000, 35-6, 435-446. 33. Olson, D. H.; Zygmunt, S. A.; Erhardt, M. K.; Curtiss, L. A.; Iton, L. E., Evidence for dimeric and tetrameric water clusters in HZSM-5. Zeolites 1997, 18, (5-6), 347-349. 34. Camblor, M. A.; Villaescusa, L. A.; Diaz-Cabanas, M. J., Synthesis of all-silica and high-silica molecular sieves in fluoride media. Top. Catal. 1999, 9, (1-2), 59-76. 35. Axon, S. A.; Klinowski, J., Synthesis and characterization of defect-free crystals of MFI-type zeolites. Appl. Catal. A-Gen. 1992, 81, (1), 27-34. 36. Giaya, A.; Thompson, R. W., Single-component gas phase adsorption and desorption studies using a tapered element oscillating microbalance. Microporous Mesoporous Mater. 2002, 55, (3), 265274. 37. Oumi, Y.; Miyajima, A.; Miyamoto, J.; Sano, T., Binary mixture adsorption of water and ethanol on silicalite. In Impact of Zeolites and Other Porous Materials on the New Technologies at the Beginning of the New Millennium, Aiello, R.; Giordano, G.; Testa, F., Eds. Elsevier Science Bv: Amsterdam, 2002; Vol. 142, pp 1595-1602. 38. Cheng, C.-H.; Bae, T.-H.; McCool, B. A.; Chance, R. R.; Nair, S.; Jones, C. W., Functionalization of the internal surface of pure-silica MFI zeolite with aliphatic alcohols. J. Phys. Chem. C 2008, 112, (10), 3543-3551. 39. Bordiga, S.; Ugliengo, P.; Damin, A.; Lamberti, C.; Spoto, G.; Zecchina, A.; Spano, G.; Buzzoni, R.; Dalloro, L.; Rivetti, F., Hydroxyls nests in defective silicalites and strained structures derived upon dehydroxylation: vibrational properties and theoretical modelling. Top. Catal. 2001, 15, (1), 43-52. 40. Lively, R. P.; Dose, M. E.; Thompson, J. A.; McCool, B. A.; Chance, R. R.; Koros, W. J., Ethanol and water adsorption in methanol-derived ZIF-71. Chem. Commun. 2011, 47, (30), 8667-8669. 41. Snider, J. R.; Dawson, G. A., Tropospheric light alcohols, carbonyls, and acetonitrile: concentrations in southwestern United-States and Henry law data. J. Geophys. Res.-Atmos. 1985, 90, (ND2), 3797-3805. 42. Luo, D.; Hu, Z.; Choi, D. G.; Thomas, V. M.; Realff, M. J.; Chance, R. R., Life Cycle Energy and Greenhouse Gas Emissions for an Ethanol Production Process Based on Blue-Green Algae. Environ. Sci. Technol. 2010, 44, (22), 8670-8677. 43. Qiu, W. L.; Kosuri, M.; Zhou, F. B.; Koros, W. J., Dehydration of ethanol-water mixtures using asymmetric hollow fiber membranes from commercial polyimides. J. Membr. Sci. 2009, 327, (1-2), 96103. 44. Lively, R. P.; Leta, D. P.; DeRites, B. A.; Chance, R. R.; Koros, W. J., Hollow fiber adsorbents for CO2 capture: Kinetic sorption performance. Chem. Eng. J. 2011, 171, (3), 801-810. 45. Lively, R. P.; Mysona, J. A.; Chance, R. R.; Koros, W. J., Formation of Defect-Free Latex Films on Porous Fiber Supports. ACS Appl. Mater. Interfaces 2011, 3, (9), 3568-3582.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Ethanol and Water sorption in defect-free MFI 338x270mm (96 x 96 DPI)

ACS Paragon Plus Environment

Page 28 of 28

Adsorption of Water and Ethanol in MFI-Type Zeolites - Langmuir

Recommend Documents