Insights into Supramolecular Sites Responsible for ... - ACS Publications

Mar 15, 2017 - compounds with greatest affinity not easiliy removed with water and ethanol ..... these cavities are the host sites for guest adsorptio...
0 downloads 0 Views 968KB Size
Subscriber access provided by University of Newcastle, Australia

Article

Insights into Supramolecular Sites Responsible for Complete Separation of Biomass-Derived Phenolics and Glucose in Metal-Organic Framework NU-1000 Mizuho Yabushita, Peng Li, Kathleen A Durkin, Hirokazu Kobayashi, Atsushi Fukuoka, Omar K. Farha, and Alexander Katz Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.7b00045 • Publication Date (Web): 15 Mar 2017 Downloaded from http://pubs.acs.org on March 18, 2017

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 33

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

Insights into Supramolecular Sites Responsible for Complete Separation of Biomass-Derived Phenolics and Glucose in Metal-Organic Framework NU-1000 Mizuho Yabushita,†,‡ Peng Li,§ Kathleen A. Durkin,┴ Hirokazu Kobayashi,‡ Atsushi Fukuoka,*,‡ Omar K. Farha,*,§,ǁ and Alexander Katz*,† †

Department of Chemical and Biomolecular Engineering, University of California, Berkeley,

California 94720, United States ‡

Institute for Catalysis, Hokkaido University, Sapporo, Hokkaido 001-0021, Japan

§

Department of Chemistry, Northwestern University, Evanston, Illinois 60208, United States



College of Chemistry, University of California, Berkeley, California 94720, United States

ǁ

Department of Chemistry, Faculty of Science, King Abdulaziz University, Jeddah 22254, Saudi

Arabia

ACS Paragon Plus Environment

1

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 33

ABSTRACT

The molecular origins of adsorption of lignin-derived phenolics to metal-organic framework NU1000 are investigated from aqueous solution, as well as in competitive mode with glucose present in the same aqueous mixture. A comparison of adsorption equilibrium constants (Kads) for phenolics functionalized with either carboxylic acid or aldehyde substituents demonstrated only a slight increase (less than a factor of 6) for the former according to both experiments and calculations. This small difference in Kads between aldehyde and carboxylic-acid substituted adsorbates is consistent with the pyrene unit of NU-1000 as the adsorption site, rather than the zirconia nodes, while at saturation coverage, the adsorption capacity suggests multiple guests per pyrene. Experimental standard free energies of adsorption directly correlated with the molecular size, and electronic structure calculations confirmed this direct relationship. The underlying origins of this relationship allude to noncovalent π–π interactions as being responsible for adsorption, the same interactions present in the condensed phase of the phenolics, which to a large extent govern their heat of vaporization. Thus, NU-1000 acts as a preformed aromatic cavity for driving aromatic guest adsorption from aqueous solution, and does so specifically, without causing detectable glucose adsorption from aqueous mixture – thereby achieving complete

glucose-phenolics

separations.

The

reusability

of

NU-1000

during

an

adsorption/desorption cycle was good, but some of the phenolic compounds with greatest affinity could not be easiliy removed with water and ethanol washes at room temperature. A competitive adsorption experiment gave an upper bound for Kads for glucose of at most 0.18 M-1, which can be compared with Kads for the phenolics investigated here, which fell in the range of 443 M-1 to 42,639 M-1. The actual value of Kads for glucose may be much closer to zero given the fact that we have failed to observe any glucose uptake with NU-1000 as adsorbent.

ACS Paragon Plus Environment

2

Page 3 of 33

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

INTRODUCTION Bioethanol as a gasoline additive offers a means to close the carbon cycle created by burning fuel, since plant biomass synthesis uses atmospheric CO2 as a carbon source via photosynthesis.1 Emerging second-generation bioethanol is derived from lignocellulosic biomass – the most abundant form of biomass on earth – and in contrast to first-generation bioethanol, which is derived from starch, instead uses that portion of the plant based on cellulose and hemicellulose fractions (that together account for typically 80% of the plant biomass), which cannot compete for food usage.2,3 In 2016, the U.S. production of second-generation bioethanol totaled 374 million liters, a tiny fraction of the currently more than 50 billion liters of corn ethanol produced nationally.4 Thus, one can foresee room for strong growth of cellulosic bioethanol, as it partially replaces its predecessor, which federal legislation (e.g., the Renewable Fuel Standard) facilitates.5 Cellulosic bioethanol production requires physicochemical pretreatment in order to render polysaccharides locked in biomass feedstock to be more accessible to enzymes and microorganisms for further depolymerization.6,7 However, an undesired yet unavoidable consequence of such pretreatment is the release of aromatic compounds into aqueous solutions, along with sugars. Such compounds consist of sugar-derived 5-hydroxymethylfurfural (HMF) and furfural, as well as several lignin-derived aromatics, which are released when ether bonds in lignin are broken. These aromatic compounds inhibit enzymes that are used as catalysts for biomass depolymerization into fermentable sugars such as glucose and xylose, as well as

ACS Paragon Plus Environment

3

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 4 of 33

microorganisms during subsequent downstream processing (e.g., ethanol fermentation), although there has been much recent progress in creating more tolerant microorganisms.8,9 Prior studies of aromatic-molecule inhibition of biocatalysts have concluded that increasing degrees of inhibitor hydrophobicity tend to increase the potency of poisoning the active site, a result that is intuitively satisfying given the fact that most if not all active sites within the enzyme interior are hydrophobic.10 This trend explains why hydrophobic and large aromatic compounds are potent poisons, such as gallic acid and tannic acid.11,12 The latter essentially completely inhibit all activity of cellobiohydrolase 1 and β-glucosidase 1 at a concentration as low as 0.25 mM in aqueous solution.13 In general, there are a multitude of lignin-derived aromatics that serve as enzyme inhibitors, and many of these have been demonstrated as such (e.g., 4-hydroxybenzaldehyde, 4-hydroxybenzoic acid, vanillin, vanillic acid, syringaldehyde, and syringic acid).8,13–15 For example, only 3 mM of vanillin has been reported to halve the fermentation efficiency of Saccharomyces cerevisiae.14 Because smaller enzyme loadings become deactivated to a greater extent given a certain amount of lignocellulosic feedstock after physicochemical pretreatment, with deactivation often being observed to be greater than 50% relative to uninhibited values, such poisoning ultimately requires more costly enzymes to achieve the same degree of biomass depolymerization. Recently, we demonstrated metal-organic framework (MOF) NU-100016 as a unique material that completely separated sugar-derived aromatics such as HMF and furfural from an aqueous mixture that also contained monomeric sugars, in a single equilibrium stage of contact.17 While no sugar (i.e., glucose, xylose, and fructose) adsorption was observed with NU-1000, we demonstrated that this could not be achieved with conventional amorphous carbon adsorbents, which at concentrations low enough to avoid saturation of sites by the aromatic adsorbates,

ACS Paragon Plus Environment

4

Page 5 of 33

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

exhibited a significant amount of sugar adsorption as well. That is to say, for the conventional amorphous carbon adsorbents, the selectivity of competitive adsorption favoring aromatic adsorbates was severely decreased – to the point of favoring sugars at low concentrations of the aromatic adsorbates, where the ratio of sugar to aromatic adsorbate concentrations became large (e.g., 300). This type of lack of perfect selectivity between sugars and aromatics is also demonstrated with hydrophobic zeolite adsorbents, which exhibit uptake of sugar in singlecomponent isotherm data.18 In contrast, NU-1000 was shown to be selective even at these low aromatic concentrations – not causing any measurable sugar adsorption even with a ratio of 300 between sugar and aromatic adsorbate.17 Here, in this manuscript, we wish to further understand the importance of molecular structure on the selectivity of sugar-aromatic separations by NU-1000 and generalize our results to also include more potent lignin-derived aromatic molecules as competitive adsorbates. We thus investigate the adsorption of a comparative set of six aromatic adsorbates (i.e., 4hydroxybenzaldehyde, 4-hydroxybenzoic acid, vanillin, vanillic acid, syringaldehyde, and syringic acid) on NU-1000, both as pure components as well as competitively, in an aqueous mixture with glucose—a fermentable sugar. Our results demonstrate the role of NU-1000 as a material that provides sacrificial hydrophobic pockets, which, with enzyme-like selectivity, can be used to remove aromatic toxins from an aqueous mixture without loss of precious glucose from the same solution. To investigate this separation further, we designed and implemented an experiment consisting of competitive adsorption of furfural and glucose, to estimate a possible upper bound for the relative affinity (i.e., adsorption equilibrium constant) of glucose to NU1000, while realizing that this number may be nearly zero. We propose this experiment as being generally useful for benchmarking adsorbents for their ability to perform complete sugar-

ACS Paragon Plus Environment

5

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 33

aromatics separations from aqueous solution. We also demonstrate the first kinetic separation involving NU-1000 in a small-column format and perform electronic structure calculations to understand trends in adsorption affinity of each of these compounds, since previously, these calculations were able to elucidate the weak interaction between adsorption site in NU-1000 and sugars.19 The results of these experiments and calculations allow us to understand those molecular features of the adsorbate that contribute to bonding with NU-1000, in a way that enables prediction of bonding of new and uninvestigated aromatic compounds to NU-1000.

EXPERIMENTAL SECTION Adsorption of Phenolics on NU-1000. NU-1000 was initially prepared by following the reported large-scale synthetic procedure.20 In this adsorption study, six phenolic compounds were tested as adsorbates: 4-hydroxybenzaldehyde; 4-hydroxybenzoic acid; vanillin; vanillic acid; syringaldehyde; and syringic acid, all of which were supplied from Sigma-Aldrich. An aqueous stock solution containing each phenolic compound shown above was prepared in a volumetric flask, and the stock solution was used as-is or after dilution in water. For competitive adsorption experiments, six phenolic compounds and glucose (Sigma-Aldrich) were dissolved in the same aqueous solution, where the initial concentrations of each phenolic compound and glucose were 3 mM and 500 mM, respectively. Then, 5 mg of NU-1000 or activated carbon material MSC-30 (Kansai Coke & Chemicals) was dispersed in 1.5 mL of the aqueous solution. The suspension was ultrasonicated for 1 min to disperse the adsorbent powder well, vortexed for at least 30 min at 297 K, and then filtered through a polytetrafluoroethylene (PTFE) membrane (0.2 µm mesh) in a syringeless filter device Mini-UniPrep (Whatman). The amount of residual

ACS Paragon Plus Environment

6

Page 7 of 33

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

compound in the liquid filtrate was quantified by high-performance liquid chromatography (HPLC, Shimadzu, Prominence HPLC System, refractive index detector) equipped with an Aminex HPX-87H column (Bio-Rad, ø7.8 × 300 mm, mobile phase 5 mM H2SO4 0.8 mL min-1, column temperature 328 K), with an absolute calibration method. The subtraction of mass of phenolic or sugar compound quantified by HPLC from that of compound charged offered an adsorption uptake. For the reuse experiment involving NU-1000, the filtered solid recoverd from the first adsorption experiment was washed with 1 mL of water 10 times followed by 1 mL of ethanol 20 times on a solid-phase extraction column (Thermo Scientific, HyperSep Silica, silica bed 200 mg), to extract adsorbed species, followed by drying under vacuum in a desiccator overnight. Then, the dried powder was used for a second run of adsorption, in the same usual manner described above. All isotherms of phenolics adsorption on NU-1000 observed in this study exhibit Type I behavior (i.e., Langmuirian, vide infra). The Langmuir equation (eq 1) gives the adsorption equilibrium constant (Kads) and adsorption capacity (Qmax) for each phenolic compound.

Q=

CKadsQmax 1 + CK ads

(1)

where C is the equilibrium concentration and Q is the adsorption uptake. This equation can be transformed into eq 2, which corresponds to the Langmuir plot (see Figure S1 in the Supporting Information). C C 1 = + Q Qmax K adsQmax

(2)

ACS Paragon Plus Environment

7

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 8 of 33

Separation of Phenolics and Glucose by NU-1000 Column. To prepare a column consisting of NU-1000, 5 mg of the MOF was dispersed in 0.5 mL of water, and the suspension was ultrasonicated for 1 min to disperse the solid particles well. The suspension was charged in a bottom part of Mini-UniPrep device, and then an upper part of this device equipped with a PTFE membrane (0.2 µm mesh) was inserted into the bottom part, to prepare a wet thin-column consisting of NU-1000 on the membrane. After removal of liquid filtrate with a pipette and subsequent removal of the upper part containing the NU-1000 column, 0.2 mL of aqueous mixture containing six phenolic compounds described above (3 mM for each phenolic compound) and glucose (500 mM) was charged into the bottom part, followed by inserting the upper part again, in order to expose the NU-1000 column to the aqueous mixture. The liquid fraction was collected by a pipette. Afterward, the column was washed with 0.1 mL of water up to five times by the same procedure. The collected liquid fraction was analyzed by HPLC (vide supra), to determine the amount of compounds captured and removed from the aqueous solution by the NU-1000 column. Electronic Structure Calculations. Geometry optimizations were performed with Gaussian 09 revision D0121 using the M06-2X functional and the cc-pVDZ basis set. The pyrene moiety, which was previously used to investigate the similar adsorption systems,19 was truncated from NU-1000 as obtained from the Cambridge Structural Database.16 We considered a range of conformers for each species with respect to orientation of functional groups in both the adsorbed and isolated states. The results reported are from the lowest energy conformer of these, confirmed as minima by frequency calculations. Results are calculated with a Polarizable Continuum Model (PCM) using the integral equation formalism variant. Maestro22 was used to calculate Connolly surfaces where a spherical probe with a radius of 1.4 Å (mimicking water), is

ACS Paragon Plus Environment

8

Page 9 of 33

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

rolled over the molecule, giving the solvent accessible area with respect to the van der Waals radii of the atoms.

RESULTS AND DISCUSSION Figure 1 shows adsorption isotherms of six phenolic compounds on NU-1000 at 297 K, consisting of 4-hydroxybenzaldehyde, 4-hydroxybenzoic acid, vanillin, vanillic acid, syringaldehyde, and syringic acid, which are depicted in Figure 1D and were chosen as adsorbates because of their relevance to lignocellulose-derived hydrolysates.8,14,15 The observed Type I (i.e., Langmuirian) isotherms in Figure 1 are consistent with previously observed adsorption isotherms of furanics (i.e., HMF, and furfural) with NU-1000,17 and also similar to the steep adsorption isotherms observed for aromatics (i.e., vanillin, HMF, and furfural) within hydrophobic pockets of zeolites.18 When compared with the previously observed Type II adsorption isotherms for sugar dimers such as cellobiose and lactose on NU-1000,19 the Langmuirian trend observed here for phenolics adsorption indicates higher relative affinity to NU-1000.23,24 The Langmuir parameters—adsorption equilibrium constant (Kads) and adsorption capacity (Qmax)—as determined from the transformed isotherm data (see Figure S1 in the Supporting Information) are summarized in Table 1. These data for the three aldehydes (i.e., 4hydroxybenzaldehyde, vanillin, and syringaldehyde) and three carboxylic acids (i.e., 4hydroxybenzoic acid, vanillic acid, and syringic acid) demonstrate the gradual increase of Kads

ACS Paragon Plus Environment

9

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 33

values from 443 ± 79 M-1 to 17947 ± 3974 M-1 for aldehydes (2431 ± 653 M-1 to 42639 ± 6229 M-1 for carboxylic acids), upon increasing the number of methoxy substituents in the series. A comparison of the Kads for the aldehyde and carboxylic-acid variant of each adsorbate (i.e., 4hydroxybenzaldehyde versus 4-hydroxybenzoic acid; vanillin versus vanillic acid; and syringaldehyde versus syringic acid) demonstrates a relative increase favoring the carboxylic acid, in a relative ratio of 2.4 – 5.5.

Figure 1. Adsorption isotherms of phenolics on NU-1000 at 297 K: (A) 4-hydroxybenzaldehyde and 4-hydroxybenzoic acid; (B) vanillin and vanillic acid; (C) syringaldehyde and syringic acid; and (D) structure of phenolics used in this study. The Langmuir plots are shown in Figure S1 in the Supporting Information. The dashed lines exhibit the isotherms replicated by the Langmuir parameters in Table 1.

ACS Paragon Plus Environment

10

Page 11 of 33

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Table 1. Langmuir parameters for aromatics adsorption on NU-1000 at 297 K. 23 SAa Kadsc Distanceb Qmaxd Qmax,mole 24 Qmax,mol/pyrenef 25 Adsorbate 2 -1 -1 -1 /Å /Å /M /mg gNU-1000 /mmol g NU-1000 26 27 28 4-Hydroxybenzaldehyde 133.6 3.32 ± 0.08 443 ± 79 354 ± 26 2.90 ± 0.21 4.34 ± 0.31 29 30 4-Hydroxybenzoic acid 139.9 3.25 ± 0.03 2431 ± 653 325 ± 15 2.35 ± 0.11 3.52 ± 0.16 31 32 Vanillin 159.2 3.27 ± 0.03 8017 ± 1680 371 ± 27 2.44 ± 0.18 3.65 ± 0.27 33 34 Vanillic acid 165.5 3.29 ± 0.05 20407 ± 5251 391 ± 20 2.33 ± 0.12 3.49 ± 0.18 35 36 Syringaldehyde 185.3 3.26 ± 0.08 17947 ± 3974 340 ± 21 1.86 ± 0.11 2.78 ± 0.16 37 38 Syringic acid 191.6 3.24 ± 0.04 42639 ± 6229 373 ± 16 1.88 ± 0.08 2.81 ± 0.12 39 40 5-Hydroxymethylfurfuralg 136.0 3.26 ± 0.04 120 ± 16 240 ± 3 1.90 ± 0.02 2.84 ± 0.03 41 42 Furfuralg 109.1 3.24 ± 0.06 28 ± 6 467 ± 28 4.86 ± 0.29 7.28 ± 0.43 43 44 aSurface area of aromatic compound, estimated from the optimized geometry by calculation (see Experimental). 45 b Distance between aromatic adsorbate and pyrene unit, determined from the optimized geometry (see Figure 2). 46 c Adsorption equilibrium constant. dAdsorption capacity on a mass basis. eAdsorption capacity on a mole basis. fMolar 47 -1 16 g 48 ratio of adsorption capacity to pyrene unit in NU-1000 (i.e., 0.668 mmol gNU-1000 ) . Langmuir parameters were 17 49 previously reported. 50 51 52 53 54 55 In light of the adsorption data above, a relevant question that we wished to address is 56 57 the location of the adsorption site in NU-1000, since this site could in principle involve either the 58 59 60 ACS Paragon Plus Environment

11

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 12 of 33

hydrophobic pyrene linker or the hydrophilic zirconia node, which is decorated with Zr–OH (zirconol) functionality, as native oxide surfaces, or, possibly, a combination of both. We previously compared the adsorption of hydrophobic molecules such as toluene and slightly less hydrophobic phenol from aqueous solution, on the surface of native (silanol-rich) silica as well as hydrophobically modified silica, which was functionalized with site-isolated grafted calix[4]arene supramolecular host cavities in the cone conformation.25 The latter were shown to have equilibrium constants for toluene adsorption from aqueous solution that were greater by a factor of at least 100 fold (10 fold for phenol) over the native silica surface. These data as well as the observed invariability of the nitrobenzene adsorption isotherm from aqueous solution (when normalized on a fraction of calixarene-sites-occupied basis, for adsorbents spanning a large variation in calixarene-site density)26 led us to conclude that the adsorption of hydrophobic aromatic molecules in the aromatic cavity of the grafted calix[4]arene site was much more favorable than on the surface of silica.

This prior insight is relevant to the question of adsorption site location in NU-1000, pertaining to the two sites mentioned above. For NU-1000, previously, the measured Kads for HMF of 120 M-1 was only 6-fold lower than Kads on a high surface-area carbon material consisting of MSC-30,17 and 14-fold higher than on a hyper-crosslinked polymer consisting of biphenyl repeat units.27 In addition, based on data above, the measured Kads for vanillin from aqueous solution is 1.4-fold higher than for a non-ionic polymeric resin Sephabeads SP206,28 and, when compared with amorphous carbon, Kads for vanillic acid adsorption from aqueous solution is only 4.6-fold lower for NU-1000.29 Altogether, within the context of the expected trends alluded to above favoring small aromatic molecule adsorption on hydrophobic rather than native-oxide surfaces, the data above are consistent with the pyrene repeat units rather than the

ACS Paragon Plus Environment

12

Page 13 of 33

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

zirconia nodes being the adsorption sites. Our logic to support this statement is made by demonstrating that were it not the case, it would lead to a contradiction: the observed Kads in that case would be expected to deviate by more than an order of magnitude from that on amorphous carbon and crosslinked organic polymer resins, especially for the more hydrophobic adsorbates such as vanillin, which is not observed. Thus, we posit that the relevant adsorption site to consider is the pyrene repeat unit, rather than the zirconia node.

Yet in view of several examples involving carboxylate coordination to the zirconia nodes of NU-1000,30,31 we were compelled to look into the possibility of interaction with the zirconia nodes further, for the adsorbates in Figure 1. Our hypothesis was that if similar coordination of carboxylate-containing adsorbates to the zirconia nodes of NU-1000 was a significant driving force for adsorption, then we should observe a large differentiation in the Kads values when comparing the aldehyde and carboxylate functionalization of the adsorbates in Figure 1. This was based on the hypothesis that the aldehyde-functionalized adsorbates would interact with the nodes much weaker than the carboxylate-functionalized ones, because of a lack of negative charge (which could be used to electrostatically interact with bound water on the zirconia node, or, alternatively, with a Lewis-acid center on the node32). This hypothesis was borne out when comparing the adsorption of vanillin and vanillic acid on a zirconia reference material (see Figure S2 in the Supporting Information for data and a description of this material). On this material, the adsorption of vanillic acid (carboxylate-functionalized adsorbate) was represented by a Kads that was 63-fold higher compared with vanillin (aldehyde-functionalized adsorbate). However, in the case of NU-1000, data in Table 1 only demonstrate a 2.5-fold increase in Kads favoring vanillic acid over vanillin. Similar differences are observed across the line with 4-hydroxybenzaldehyde and 4-hydroxybenzoic acid (5.5-fold difference in Kads

ACS Paragon Plus Environment

13

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 33

favoring acid) as well as syringaldehyde and syringic acid (2.5-fold difference in Kads favoring acid) in Table 1. These small differences in Kads for carboxylic-acid and aldehyde functionalized variants of the adsorbate cannot be rationalized on the basis of carboxylate-substituent coordination to the zirconia nodes, which, if present, would necessitate much larger differences in Kads. Thus, this comparison further supports the lack of zirconia-node involvement as an adsorption site, in the adsorption experiments of Figure 1. We explain the difference between our outcome and prior experiments where carboxylate coordination to the zirconia nodes of NU1000 was observed, by the fact that we are conducting adsorption experiments in water whereas all other examples of carboxylate coordination have involved aprotic organic solvents.30,31 According to this explanation, water acts as a competitive adsorbate, which decreases the strength of carboxylate coordination to the zirconia nodes of NU-1000. Having said that, we cannot completely rule out the possibility of minor interactions with the zirconia nodes, in addition to the pyrene repeat units – but we can state with certainty that the zirconia nodes in and of themselves are not driving aromatic molecule adsorption, based on the observed relative affinities from aqueous solution described above.

To further elucidate the differences in Kads observed between carboxylic-acid and aldehyde functionalized adsorbates, we performed molecular modeling of the interactions between each adsorbate and a substituted pyrene linker of NU-1000, which was previously used to represent the minimally complex adsorption site of NU-1000.19 Electronic structure calculations using this site previously successfully predicted a lack of glucose adsorption while demonstrating sufficient interaction with the β-linked dimer of glucose, cellobiose, to support observed adsorption of the latter.19 Figure 2 shows geometry-optimized minimum energy configurations of interactions between the various aromatic guests investigated here, and the

ACS Paragon Plus Environment

14

Page 15 of 33

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

minimally complex unit previously used to model adsorption to NU-1000 described above. A nearly cofacial interaction (dihedral angles of 171° – 179° were measured) between the aromatic guests and the pyrene units are observed, which is similar to those observed in other systems involving π–π interactions.33,34 Regardless of benzoic acid domains present in the pyrene linker, which may cause steric hindrance against an adsorbate molecule, the distances between the pyrene unit and adsorbate are 3.24 ± 0.04 Å – 3.32 ± 0.08 Å (Table 1), shorter than the expected van der Waals distance of 3.40 Å (i.e., double of the radius of a carbon atom). These shorter distances therefore are consistent with an interaction between the aromatic adsorbate and pyrene unit via π–π interaction. According to previous reports investigating π–π aromatic interactions via electronic structure calculations,33,35,36 substituents do not impact the strength of the interaction by altering electron density of the aromatic rings involved, but instead they act independently and interact directly with the pyrene unit.

We investigated the scaling of the interactions described above as a function of the adsorbate size, because we previously observed a direct scaling between the types of van der Waals interactions described above (i.e., π–π interactions) and size in related adsorption of aromatic guests from aqueous solution in preorganized π-rich cavities of grafted calixarene adsorbents.26 In the latter system, this led us to propose that the energetics of the interactions between adsorbent and adsorbate (i.e., host and guest, respectively) were the same as those that govern interactions between adsorbate molecules in the condensed phase and that are represented by the heat of adsorbate vaporization.26 Based on the calculated surface areas for each adsorbate, data in Figure 3A show a direct correlation between adsorption Gibbs free energy (∆G°ads) as calculated from data in Table 1, and the molecular surface area of each compound. A similar trend is observed in Figure 3B based on electronic structure calculations; in fact, the slopes in

ACS Paragon Plus Environment

15

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 33

Figures 3A and 3B are nearly identical, which is suprising given anticipated favorable entropy of adsorption due to both host and guest aqueous desolvation,37,38 which is not accounted for in the calculations. Data in Figure 3 show a lack of different scaling (i.e. a lack of vertically offset lines for blue aldehydes versus red carboxylic acids data points in Figure 3), which could otherwise be indicative of carboxylate coordination to the zirconia nodes – thereby reaffirming conclusions above regarding the lack of such coordination. However, crucially, in both cases, the similar and direct scaling of ∆G°ads with size of adsorbate (i.e., adsorption of larger aromatic compounds occurs with greater affinity on NU-1000) means that adsorption is driven by the cavities (pore space) that is adjacent to each pyrene unit of NU-1000, by interactions that are similar in type and energy as observed in the condensed phase and are largely represented by the heat of vaporization for each adsorbate. In our systems, these cavities are the host sites for guest adsorption, which become preorganized and self-assembled during the NU-1000 synthesis.

ACS Paragon Plus Environment

16

Page 17 of 33

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

Figure 2. Optimized geometry of aromatics-pyrene supersystems. Legends: grey = carbon atoms of aromatic compound; black = carbon atoms of pyrene unit; red = oxygen atoms; and white = hydrogen atoms.

Figure 3. Parametric plot of ∆G°ads as a function of surface area of phenolic compound, based on (A) experimental data at 297 K and (B) computational data. Legend: blue = aldehydes and red = carboxylic acids.

ACS Paragon Plus Environment

17

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 18 of 33

The preceding analysis dealt with interactions responsible for adsorption of aromatic guests to host sites on NU-1000, consisting of pyrene connecting units, and these ultimately control the value of Kads and thereby partially control the slope of the isotherm in the low concentration regime. In the following section, we analyze molecular structural factors that control the saturation capacity of each guest to NU-1000. Starting with the smallest and unsubstituted molecules in our set shown in Table 1, the adsorption capacities of 4hydroxybenzaldehyde and 4-hydroxybenzoic acid based on mole were measured to be 2.90 ± 0.21 mmol gNU-1000-1 and 2.35 ± 0.11 mmol gNU-1000-1, respectively. These values decreased upon further substitution with a methoxy group (1.86 ± 0.11 mmol gNU-1000-1 for syringaldehyde and 1.88 ± 0.08 mmol gNU-1000-1 for syringic acid). Such an adsorption capacity decrease can be rationalized by the increase of adsorption cross-sectional area, similar to reasoning being previously invoked in other adsorption systems.39 Now generalizing to all adsorbates in our set, the molar ratios of adsorbate at saturation coverage (that is, adsorption capacity on a mole basis, Qmax,mol) normalized to the number of pyrene units is calculated to be 2.78 – 4.34, based on data in Table 1. This requires at least 1.4 molecules simultaneously adsorbed on a single pyrene face at saturation coverage, with both pyrene faces accessible for adsorption. Based on this number being greater than unity as well as the lack of available space in Figure 2 to easily accommodate a second molecule in a monolayer configuration, we surmise that either the adsorbates need to pucker further or, alternatively, they may stack in multilayers, in which a second or higher layer does not interact directly with the pyrene unit but rather with adsorbates beneath. The latter alternative is deemed unlikely based on the Type I shape of the isotherms in Figure 1.23,24 There is no evidence of a condensation or pore filling phenomenon, which would be indicative of

ACS Paragon Plus Environment

18

Page 19 of 33

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

multilayers such as a Type II or other shape. This suggests that multiple aromatic adsorbates locate on the pyrene unit in a T-shaped edge-to-face configuration,33–35,40 where the edge of the aromatic adsorbates (i.e., substituent or C–H dipole) vertically interacts with the pyrene face. Such a configuration minimizes the adsorption cross-sectional area and enables several molecules to simultaneously interact with a single pyrene unit face.

While the calculations shown in Figure 2 represent a more dilute surface coverage than that corresponding to Qmax, for reasons described above, we reaffirm that the nature and energetics of the noncovalent interactions between guest and pyrene host site is very similar to that between guests themselves. This is represented by the trend in Figure 3, which shows a similar scaling based on size with the standard free energy of adsorption, as would be expected based on the heat of vaporization for the same molecules. Therefore, even if modifying the computational molecular model to account for a Qmax adsorption regime, where there are multiple aromatic guests per pyrene host site, we do not expect a large change in the direct nature of the trends exhibited in Figure 3. This is because any new interactions brought about by the additional guest molecules between themselves and/or the pyrene unit would also scale proportional to the guest size.

The favorable interaction between NU-1000 and phenolic compounds is in stark contrast to sugar monomers that previously showed no affinity with the MOF.17,19 Such a significant difference in affinity offers us an opportunity to investigate competitive adsorption, where both phenolic compounds and glucose are present in the same aqueous mixture. If NU1000 adsorbs only phenolic compounds without causing any measurable glucose uptake, this selective molecular recognition opens the door to applications in fermentation pretreatment for

ACS Paragon Plus Environment

19

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 20 of 33

second-generation bioethanol production, to remove phenolics, which poison microorganisms and enzymes that catalyze ethanol fermentation from a sugar feedstock.8,14,15

Figure 4 represents the uptakes of glucose (initial concentration of 500 mM) and six phenolics (initial concentration of 3 mM for each phenolic compound) from an aqueous mixture, when using one of two adsorbents, consisting of either NU-1000 or amorphous carbon MSC-30, under competitive adsorption conditions at 297 K. This aqueous mixture is meant to simulate the small amount of phenolics in the presence of a large excess of glucose, as would be present in an actual sugar feedstock after biomass pretreatment, during second-generation ethanol production.14,15 Even in the presence of a large excess of glucose, NU-1000 selectively adsorbs 99 ± 0.3% of the phenolics originally present in the aqueous solution (corresponding to 171 ± 0.3 mg gNU-1000-1 of total phenolics uptake), while causing no observable glucose uptake. Upon recovering the NU-1000 material after adsorption, the material was washed with water (1 mL ten times) followed by ethanol (1 mL twenty times) on a filter to regenerate adsorption sites, and it was subsequently reused in a second adsorption test. This regenerated NU-1000 preserves the integrity of the MOF and exhibits the same high degree of selectivity in the reuse test by removing >99% of the phenolic compounds originally present in the aqueous solution, without adsorbing any glucose (see the data of second cycle in Figure 4). In comparison, although MSC30 also removes almost all phenolic compounds, 226 ± 4 mg gMSC-30-1 of glucose is adsorbed at the same time, which represents 1.3-fold higher glucose uptake relative to total phenolic uptake by mass. Upon analyzing the filtrate of the washes of the NU-1000 after the first adsorption experiment (see Figure S3 in the Supporting Information), while the weaker adsorbing 4hydroxybenzaldehyde was completely removed, other more strongly interacting guests (i.e., 4hydroxybenzoic acid, vanillin, and vanillic acid) were partially removed, with almost no

ACS Paragon Plus Environment

20

Page 21 of 33

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

extraction of very strongly interacting guests such as syringaldehyde and syringic acid. Future work will focus on the development of selective adsorption systems with full regeneration of sites after each adsorption/desorption cycle.

Figure 4. Competitive adsorption of glucose and six phenolic compounds on 25 mg of adsorbents at 297 K. Initial concentrations of glucose and each phenolic compound are 500 mM and 3 mM, respectively. Prior to the second cycle, NU-1000 was regenerated via washing with water and ethanol.

We further conducted the separation of glucose and six phenolic compounds at room temperature by a short column consisting of 5 mg of NU-1000. Figure 5 shows the adsorbed fraction of glucose and phenolics, when washing the column with 0.1 mL of water up to five times. Glucose is recovered in nearly quantitative yield of 98 ± 2% during this separation experiment, whereas 88 ± 3% of phenolics are captured and removed from aqueous mixture by the column. These captured phenolics can be gradually eluted when washing the NU-1000

ACS Paragon Plus Environment

21

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 22 of 33

column with water (0.1 mL of water added five times to the column), and specifically, weaklybonded phenolics (i.e., 4-hydroxybenzaldehyde and 4-hydroxybenzoic acid) are easily washed out, relative to strongly-bonded compounds, which is consistent with the Kads trend in Table 1. After 0.5 mL of total water wash in 0.1 mL increments, 16 ± 4% of phenolics are recovered. As already described above, the phenolic compounds remaining on NU-1000 can be partially extracted with ethanol, and the regenerated NU-1000 is capable of being reused. This data demonstrates the potential application of NU-1000 for the removal of phenolic inhibitors from fermentation broths in a column format.

Figure 5. Separation of glucose and six phenolic compounds by NU-1000 column. Figure 5A shows the adsorbed fractions of glucose and total phenolics, and Figure 5B represents those of

ACS Paragon Plus Environment

22

Page 23 of 33

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

each phenolic compound. The column was washed with 0.1 mL of water up to five times, and the abscissa of both figures represents total volume of washing water.

Finally, having observed the complete separation of glucose and aromatic compounds above, we wished to determine a maximum possible upper bound for the glucose equilibrium constant for adsorption to NU-1000. Given just how weakly glucose adsorbs to NU-1000, we are cognizant of the fact that the actual value of the glucose equilibrium constant for adsorption to NU-1000 may be nearly zero. Yet we are compelled to provide some guidance as to what its maximum possible value is, and do so here with an approach that relies on a competitive adsorption experiment to obtain an upper bound for Kads of glucose. During competitive adsorption, the ratio of Kads values for two adsorbates in the dilute coverage regime (assuming same Qmax,mol for both adsorbates – see Supporting Information for details) is given by:

K ads, Glc K ads, Arm

f Glc 1 − f Glc = f Arm 1 − f Arm

(3)

where Kads,Glc and Kads,Arm are the adsorption equilibrium constants of glucose and aromatics, respectively, and fGlc and fArm are the fractions of glucose and aromatics, respectively, which have been removed via adsorption from the aqueous mixture. We chose furfural as the aromatic compound that shows lowest affinity to NU-1000 according to data in Table 1, which results in greater sensitivity for determining a small value of Kads,Glc in eq 3, based on an experimentally accessible fGlc. Thus, we conducted the competitive adsorption of glucose and furfural on NU-

ACS Paragon Plus Environment

23

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 33

1000 at 297 K (initial concentration of both glucose and furfural was 50 µM). We observed 89% of furfural initially present to be adsorbed, and no measurable decrease in the glucose solution concentration when comparing before and after NU-1000 treatment of the aqueous mixture. Based on this data and eq 3, we estimate the Kads,Glc value to be smaller than 0.18 ± 0.02 M-1 while noting that its actual value may indeed be much smaller than this and nearly zero.

CONCLUSIONS

Using a combination of experiments and electronic structure calculations, we investigate the noncovalent interactions that are responsible for aromatic guest adsorption to pyrene host sites of NU-1000, using a comparison involving six phenolic compounds. While NU-1000 exhibits high affinity for the adsorption of these phenolics from aqueous solution, it also demonstrates complete separation of the phenolics from an aqueous mixture consisting of a large excess of glucose, in both a batch and column format with good reusability. Molecular modeling calculations show the importance of π–π interactions between the guest and pyrene host site of NU-1000, by virtue of direct interaction between aromatic substituents of the guest and πelectrons and/or C–H dipole at a nearby vertex of the pyrene site. The direct scaling of the experimental standard Gibbs free energy of adsorption with respect to the guest molecule size is consistent with results from electronic structure calculations, and points to the pyrene units rather than the zirconia nodes as the supramolecular host site for adsorption. The similar nature of interactions responsible for adsorption relative to those interactions involved in the heat of vaporization for these aromatic guests suggests the same direct scaling, even when modifying the

ACS Paragon Plus Environment

24

Page 25 of 33

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

molecular modeling calculations to accommodate several guests per pyrene host site, as observed experimentally at saturation coverage, where at least 1.4 guest molecules per pyrene face were measured at saturation coverage.

ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.XXXXXXX. Langmuir plots for phenolics adsorption on NU-1000, adsorption isotherms on a zirconia reference material, desorbed fraction of each phenolic compound adsorbed on NU-1000 via washing with water and ethanol, and derivation method of eq 3 from competitive adsorption data (PDF)

AUTHOR INFORMATION Corresponding Authors *E-mail: [email protected] *E-mail: [email protected] *E-mail: [email protected] Notes

ACS Paragon Plus Environment

25

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 26 of 33

The authors declare no competing financial interest.

ACKNOWLEDGMENTS This research was supported by the funding from the Office of Basic Energy Sciences of the Department of Energy (DE-FG02-05ER15696). O.K.F. gratefully acknowledges support from the Inorganometallic Catalyst Design Center, an Energy Frontier Research Center, funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, under Award DESC0012702.

REFERENCES (1)

Wyman, C. Handbook on Bioethanol: Production and Utilization; Taylor & Francis:

Washington, DC, 1996. (2)

Youngs, H.; Somerville, C. Best practices for biofuels. Science 2014, 344, 1095–1096.

(3)

Service, R. F. Cellulosic ethanol at last? Science 2014, 345, 1111.

(4)

Ethanol Producer Magazine – The Latest News and Data About Ethanol Production.

http://www.ethanolproducer.com/plants/listplants/US/Existing/All (accessed January 4, 2017). (5)

Ziolkowska, J. R. Prospective technologies, feedstocks and market innovations for ethanol

and biodiesel production in the US. Biotechnol. Rep. 2014, 4, 94–98.

ACS Paragon Plus Environment

26

Page 27 of 33

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)

Mosier, N.; Wyman, C.; Dale, B.; Elander, R.; Lee, Y. Y.; Holtzapple, M.; Ladisch, M.

Features of promising technologies for pretreatment of lignocellulosic biomass. Bioresour. Technol. 2005, 96, 673–686. (7)

Yang, B.; Wyman, C. E. Pretreatment: the key to unlocking low-cost cellulosic ethanol.

Biofuels, Bioprod. Biorefin. 2008, 2, 26–40. (8)

Almeida, J. R. M.; Modig, T.; Petersson, A.; Hähn-Hägerdal, B.; Lidén, G.; Gorwa-

Grauslund, M. F. Increased tolerance and conversion of inhibitors in lignocellulosic hydrolysates by Saccharomyces cerevisiae. J. Chem. Technol. Biotechnol. 2007, 82, 340–349. (9)

Parawira, W.; Tekere, M. Biotechnological strategies to overcome inhibitors in

lignocellulose hydrolysates for ethanol production: review. Crit. Rev. Biotechnol. 2011, 31, 20– 31. (10) Yang, Q.; Pan, X. Correlation between lignin physicochemical properties and inhibition to enzymatic hydrolysis of cellulose. Biotechnol. Bioeng. 2016, 113, 1213–1224. (11) Ximenes, E.; Kim, Y.; Mosier, N.; Dien, B.; Ladisch, M. Inhibition of cellulases by phenols. Enzyme Microb. Technol. 2010, 46, 170–176. (12) Ximenes, E.; Kim, Y.; Mosier, N.; Dien, B.; Ladisch, M. Deactivation of cellulases by phenols. Enzyme Microb. Technol. 2011, 48, 54–60. (13) Mhlongo, S. I.; den Haan, R.; Viljoen-Bloom, M.; van Zyl, W. H. Lignocellulosic hydrolysate inhibitors selectively inhibit/deactivate cellulase performance. Enzyme Microb. Technol. 2015, 81, 16–22.

ACS Paragon Plus Environment

27

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 28 of 33

(14) Delgenes, J. P.; Moletta, R.; Navarro, J. M. Effects of lignocellulose degradation products on ethanol fermentations of glucose and xylose by Saccharomyces cerevisiae, Zymomonas mobilis, Pichia stipitis, and Candida shehatae. Enzyme Microb. Technol. 1996, 19, 220–225. (15) Klinke, H. B.; Thomsen, A. B.; Ahring, B. K. Inhibition of ethanol-producing yeast and bacteria by degradation products produced during pre-treatment of biomass. Appl. Microbiol. Biotechnol. 2004, 66, 10–26. (16) Mondloch, J. E.; Bury, W.; Fairen-Jimenez, D.; Kwon, S.; DeMarco, E. J.; Weston, M. H.; Sarjeant, A. A.; Nguyen, S. T.; Stair, P. C.; Snurr, R. Q.; Farha, O. K.; Hupp, J. T. Vapor-Phase Metalation by Atomic Layer Deposition in a Metal–Organic Framework. J. Am. Chem. Soc. 2013, 135, 10294–10297. (17) Yabushita, M.; Li, P.; Kobayashi, H.; Fukuoka, A.; Farha, O. K.; Katz, A. Complete furanics–sugar separations with metal–organic framework NU-1000. Chem. Commun. 2016, 52, 11791–11794. (18) Ranjan, R.; Thust, S.; Gounaris, C. E.; Woo, M.; Floudas, C. A.; von Keitz, M.; Valentas, K. J.; Wei, J.; Tsapatsis, M. Adsorption of fermentation inhibitors from lignocellulosic biomass hydrolyzates for improved ethanol yield and value-added product recovery. Micropor. Mesopor. Mater. 2009, 122, 143–148. (19) Yabushita, M.; Li, P.; Bernales, V.; Kobayashi, H.; Fukuoka, A.; Gagliardi, L.; Farha, O. K.; Katz, A. Unprecedented selectivity in molecular recognition of carbohydrates by a metal– organic framework. Chem. Commun. 2016, 52, 7094–7097.

ACS Paragon Plus Environment

28

Page 29 of 33

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

(20) Wang, T. C.; Vermeulen, N. A.; Kim, I. S.; Martinson, A. B. F.; Stoddart, J. F.; Hupp, J. T.; Farha, O. K. Scalable synthesis and post-modification of a mesoporous metal-organic framework called NU-1000. Nat. Protocols 2016, 11 (1), 149–162. (21) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M. J.; Heyd, J.; Brothers, E. N.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A. P.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09; Gaussian, Inc.: Wallingford, 2009. (22) Schrödinger Release 2016-2: Maestro; Schrödinger, LLC: New York, 2016. (23) Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.; Pierotti, R. A.; Rouquérol, J.; Siemieniewska, T. Reporting physisorption data for gas/solid systems with special reference to the determination of surface area and porosity (Recommendations 1984). Pure Appl. Chem. 1985, 57, 603–619. (24) Thommes, M.; Kaneko, K.; Neimark, A. V.; Olivier, J. P.; Rodriguez-Reinoso, F.; Rouquerol, J.; Sing, K. S. W. Physisorption of gases, with special reference to the evaluation of

ACS Paragon Plus Environment

29

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 30 of 33

surface area and pore size distribution (IUPAC Technical Report). Pure Appl. Chem. 2015, 87, 1051–1069. (25) Katz, A.; Da Costa, P.; Lam, A. C. P.; Notestein, J. M. The First Single-Step Immobilization of a Calix-[4]-arene onto the Surface of Silica. Chem. Mater. 2002, 14, 3364– 3368. (26) Notestein, J. M.; Katz, A.; Iglesia, E. Energetics of Small Molecule and Water Complexation in Hydrophobic Calixarene Cavities. Langmuir 2006, 22, 4004–4014. (27) Detoni, C.; Gierlich, C. H.; Rose, M.; Palkovits, R. Selective Liquid Phase Adsorption of 5-Hydroxymethylfurfural on Nanoporous Hyper-Cross-Linked Polymers. ACS Sustainable Chem. Eng. 2014, 2, 2407–2415. (28) Zabkova, M.; Otero, M.; Minceva, M.; Zabka, M.; Rodrigues, A. E. Separation of synthetic vanillin at different pH onto polymeric adsorbent Sephabeads SP206. Chem. Eng. Process. 2006, 45, 598–607. (29) Richard, D.; Núñez, M. L. D.; Schweich, D. Adsorption of complex phenolic compounds on active charcoal: Adsorption capacity and isotherms. Chem. Eng. J. 2009, 148, 1–7. (30) Deria, P.; Mondloch, J. E.; Tylianakis, E.; Ghosh, P.; Bury, W.; Snurr, R. Q.; Hupp, J. T.; Farha, O. K. Perfluoroalkane Functionalization of NU-1000 via Solvent-Assisted Ligand Incorporation: Synthesis and CO2 Adsorption Studies. J. Am. Chem. Soc. 2013, 135, 16801– 16804.

ACS Paragon Plus Environment

30

Page 31 of 33

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) Rimoldi, M.; Nakamura, A.; Vermeulen, N. A.; Henkelis, J. J.; Blackburn, A. K.; Hupp, J. T.; Stoddart, J. F.; Farha, O. K. A metal–organic framework immobilised iridium pincer complex. Chem. Sci. 2016, 7, 4980–4984. (32) Boisvert, J.-P.; Persello, J.; Foissy, A.; Castaing, J.-C.; Cabane, B. Effect of surface charge on the adsorption mode of sodium poly(acrylate) on alumina-coated TiO2 used as coating pigment. Colloids Surf. A: Physicochem. Eng. Asp. 2000, 168, 287–296. (33) Sinnokrot, M. O.; Sherrill, C. D. High-Accuracy Quantum Mechanical Studies of π−π Interactions in Benzene Dimers. J. Phys. Chem. A 2006, 110, 10656–10668. (34) Hunter, C. A.; Lawson, K. R.; Perkins, J.; Urch, C. J. Aromatic interactions. J. Chem. Soc. Perkin Trans. 2 2001, 651–669. (35) Ringer, A. L.; Sinnokrot, M. O.; Lively, R. P.; Sherrill, C. D. The Effect of Multiple Substituents on Sandwich and T-Shaped π–π Interactions. Chem. Eur. J. 2006, 12, 3821–3828. (36) Wheeler, S. E. Understanding Substituent Effects in Noncovalent Interactions Involving Aromatic Rings. Acc. Chem. Res. 2013, 46, 1029–1038. (37) Southall, N. T.; Dill, K. A.; Haymet, A. D. J. A View of the Hydrophobic Effect. J. Phys. Chem. B 2002, 106, 521–533. (38) Yabushita, M.; Kobayashi, H.; Hasegawa, J.; Hara, K.; Fukuoka, A. Entropically Favored Adsorption of Cellulosic Molecules onto Carbon Materials through Hydrophobic Functionalities. ChemSusChem 2014, 7, 1443–1450.

ACS Paragon Plus Environment

31

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 32 of 33

(39) Darmograi, G.; Prelot, B.; Layrac, G.; Tichit, D.; Martin-Gassin, G.; Salles, F.; Zajac, J. Study of Adsorption and Intercalation of Orange-Type Dyes into Mg–Al Layered Double Hydroxide. J. Phys. Chem. C 2015, 119, 23388–23397. (40) Adams, H.; Carver, F. J.; Hunter, C. A.; Morales, J. C.; Seward, E. M. Chemical DoubleMutant Cycles for the Measurement of Weak Intermolecular Interactions: Edge-to-Face Aromatic Interactions. Angew. Chem., Int. Ed. 1996, 35, 1542–1544.

ACS Paragon Plus Environment

32

Page 33 of 33

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

Table of Contents

ACS Paragon Plus Environment

33