High-Throughput Continuous Flow Synthesis of Nickel Nanoparticles

Nov 7, 2016 - Frederick G. Baddour , Vanessa A. Witte , Connor P. Nash , Michael B. Griffin , Daniel A. Ruddy , and Joshua A. Schaidle. ACS Sustainabl...
1 downloads 8 Views 1MB Size
Subscriber access provided by UNIV OF REGINA

Article

High-Throughput Continuous Flow Synthesis of Nickel Nanoparticles for the Catalytic Hydrodeoxygenation of Guaiacol Emily J. Roberts, Susan E. Habas, Lu Wang, Daniel A. Ruddy, Erick A. White, Frederick G. Baddour, Michael B. Griffin, Joshua A. Schaidle, Noah Malmstadt, and Richard L. Brutchey ACS Sustainable Chem. Eng., Just Accepted Manuscript • DOI: 10.1021/ acssuschemeng.6b02009 • Publication Date (Web): 07 Nov 2016 Downloaded from http://pubs.acs.org on November 17, 2016

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.

ACS Sustainable Chemistry & Engineering 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 16

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

ACS Sustainable Chemistry & Engineering

High-Throughput Continuous Flow Synthesis of Nickel Nanoparticles for the Catalytic Hydrodeoxygenation of Guaiacol Emily J. Roberts,1 Susan E. Habas,2 Lu Wang,3 Daniel A. Ruddy,2 Erick A. White,2 Frederick G. Baddour,2 Michael B. Griffin,2 Joshua A. Schaidle,*,2 Noah Malmstadt,*1,3 Richard L. Brutchey*,1 1

Department of Chemistry, 840 Downey Way, University of Southern California, Los Angeles, CA 90089-0744, USA 2

National Bioenergy Center, National Renewable Energy Laboratory, 15013 Denver West Parkway, Golden, CO 80401-3305, USA 3

Mork Family Department of Chemical Engineering and Materials Science, University of Southern California, 925 Bloom Walk, Los Angeles, CA 90089-1211, USA Email: [email protected]

Keywords: nickel nanoparticles, microreactor, millifluidics, ex-situ fast pyrolysis, hydrodeoxygenation, lignin Abstract The translation of batch chemistries to high-throughput flow methods addresses scaling concerns associated with the implementation of colloidal nanoparticle (NP) catalysts for industrial processes. A literature procedure for the synthesis of Ni-NPs was adapted to a continuous millifluidic (mF) flow method, achieving yields >60%. Conversely, NPs prepared in a batch (B) reaction under conditions analogous to the continuous flow conditions gave only a 45% yield. Both mF- and B-Ni-NP catalysts were supported on SiO2 and compared to a Ni/SiO2 catalyst prepared by traditional incipient wetness (IW) impregnation for the hydrodeoxygenation (HDO) of guaiacol under ex situ catalytic fast pyrolysis conditions (350 °C, 0.5 MPa). Compared to the IW method, both colloidal NPs displayed increased morphological control and narrowed size distributions, and the NPs prepared by both methods showed similar size, shape, and crystallinity. The Ni-NP catalyst synthesized by the continuous flow method exhibited similar H-adsorption site densities, site-time yields, and selectivities towards deoxygenated products compared to the analogous batch-prepared catalyst, and outperformed the IW catalyst with respect to higher selectivity to lower oxygen content products and a 31-fold decrease in deactivation rate. These results demonstrate the utility of synthesizing colloidal Ni-NP catalysts using flow methods that can produce >27 g/day of Ni NPs (equivalent to >0.5 kg of 5 wt% Ni/SiO2), while maintaining the catalytic properties displayed by the batch equivalent.

1. Introduction

1

ACS Paragon Plus Environment

ACS Sustainable Chemistry & Engineering

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

The realization of sustainable routes to liquid fuels from biomass relies on the effective transformation of highly oxygenated biomass into hydrocarbon products with properties suitable for blending into existing fuels, as well as the development of low-cost, earth-abundant catalysts that can promote these transformations over extended catalyst lifetimes.1,2 Addressing these challenges will require advanced catalysts with controlled active site(s) that promote desired transformations (e.g., deoxygenation, hydrogenation and C–C coupling) while resisting deactivation, and that can be produced at relevant scales with high throughput. Over the past two decades, nanoparticle (NP) syntheses have been developed to address these catalytic challenges through controlled size and morphology, which has resulted in promising catalytic performance and a greater understanding of fundamental structure-function relationships that drive the development of next-generation catalysts.3,4,5 Opportunities to cost-effectively and sustainably produce biomass-derived liquid fuels lie at the convergence of: process design and catalytic reaction engineering for biomass utilization,2 rational design of novel catalysts (e.g., highly defined NPs), development of robust and versatile synthetic routes,6 and the high-throughput production of NPs to be used at scale.7,8,9 Ex situ catalytic fast pyrolysis (CFP) of lignocellulosic biomass (e.g., woody and herbaceous feedstocks) is a promising approach to the sustainable production of blendable liquid fuels.1,2 In the ex situ CFP process, pyrolysis vapors are catalytically upgraded prior to condensation to provide a bio-oil with improved quality and stability that can undergo downstream hydroprocessing. The key objectives of upgrading are deoxygenation, minimization of carbon losses to coke and non-condensable gases, and hydrogenation, along with C–C coupling reactions, to shift the product slate towards the distillaterange. This process requires higher temperatures (350-500 ˚C), lower pressures (near atmospheric), and lower H2 concentrations (stoichiometric) than typical hydrotreating reactions. 10 The most common heterogeneous upgrading catalyst, transition metal-promoted MoS2,11,12 exhibits high rates of deactivation in the presence of water and requires a continuous feed of sulfide reagents to maintain catalyst stability leading to sulfur-containing species in the fuel that must be removed.10,13,14 Noble metal catalysts (e.g., Pt and Pd) are robust and do not require supplemental co-feeds to maintain activity; however, they require costly, less abundant elements.15,16,17 Consequently, we have identified a need to investigate cost-effective catalysts that promote hydrogenation, deoxygenation, and C–C coupling under ex situ CFP conditions while maintaining stability.18 Targeted catalyst design requires the correlation of catalyst features with reactivity. Solutionphase routes to colloidal NPs enable the preparation of particles with well-defined features (size, shape, and composition), and the ability to synthetically tailor these attributes. The homogeneous features inherent to solution-synthesized colloidal NPs can be controlled and related back to catalytic performance to elucidate structure-function relationships. In contrast, catalysts prepared by traditional impregnation methods can be composed of a wide array of particle sizes, shapes, and compositions on a support material.19 Common large-scale synthetic routes, such as incipient wetness (IW) impregnation followed by calcination and/or reduction, generally require high temperatures and are support-dependent processes. Recently, Griffin et al. reported that earth-abundant colloidal Ni NP catalysts exhibit high conversion and selectivity towards fuel molecules with low oxygen content relative to IW analogues.20 Such results, in addition to advances in the understanding of morphologydependent catalytic properties, underscore the need to further develop well-defined, colloidal NP catalysts, which ultimately must be prepared at scale.21,22 Currently, colloidal NP catalysts are synthesized using batch reactors, typically on a small scale (10-100 mL). In order to produce a significant quantity of colloidal NP catalyst to be viable for industrial biomass pyrolysis vapor upgrading, these synthetic systems need to be scaled. Scaling volume or reagent concentrations of such batch reactors can lead to inhomogenities in mass and thermal transport, which leads to significant variability between batches as a result of changes to the nucleation and growth rates.23 Moreover, inconsistencies in reaction parameters can have strong

2

ACS Paragon Plus Environment

Page 2 of 16

Page 3 of 16

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

ACS Sustainable Chemistry & Engineering

effects on the resulting morphologies that govern catalytic properties. Alternatively, continuous flow microfluidic (µF) and millifluidic (mF) methods have been identified as a solution to the aforementioned issues associated with large-scale production. Monodisperse particles with increased yields can be obtained as a result of the superior heat and mass transport properties provided by the high surface-area-to-volume ratios inherent to channels with micron-scale diameters.19 The reduced reaction volume promotes fast heating and cooling, which minimizes energy inputs and reaction times, leading to more sustainable processes. Unlike batch reactions, flow methods are amenable to automation, affording reproducibility with high throughput and high fidelity.7,24,25,26 Additionally, operating in small volumes reduces environmental and safety risks associated with typical batch processes in terms of exposure to chemicals and potential spills.27 Flow processes are amenable to sustainable scale-up, process intensification, and solvent/reagent recycling processes.28,29 Therefore, the green chemistry aspects of the research reported here are multifold: in addition to facilitating the catalytic production of biofuels, catalyst preparation is inherently more sustainable. While continuous flow methods have been employed in a variety of colloidal NP syntheses,30,31,32 there have been few reports on the utilization of the resulting NPs for catalysis, particularly for earth abundant, low-cost materials like Ni.33,34 Herein, we report the high-throughput, mF synthesis of well-defined, colloidal Ni-NPs for the hydrodeoxygenation (HDO) of guaiacol (2-methoxyphenol), a model compound representative of the major lignin-derived products found in the fast pyrolysis vapor from lignocellulosic biomass. The performance of the mF-synthesized Ni-NPs towards HDO under ex situ CFP conditions is compared to Ni-NPs synthesized by a batch reaction, and a Ni catalyst prepared by IW impregnation. Characterization of both the unsupported colloidal NPs and SiO2-supported catalysts was conducted in order to elucidate the effect that the synthetic method has on catalyst properties. Compared to the IW-synthesized Ni catalyst, the colloidal NPs displayed increased catalytic stability and higher selectivity towards deoxygenated products. Overall, the mF synthesis of colloidal Ni-NPs produced gram quantities of catalyst with comparable activity to the batch-synthesized Ni-NPs.

2. Results and Discussion 2.1 mF synthesis of Ni-NPs The high-throughput, flow synthesis of Ni-NPs was performed in a mF reactor (Scheme 1) based on a literature synthesis by Carenco et al.35 A constant precursor flow rate is achieved through a feedback loop between an analytical balance and pressurized gas. This balance monitors the flux of the precursor solution (Ni(acac)2, oleylamine, octadecene, and trioctylphosphine) in real time, and the computer-controlled system adjusts the pressure to maintain a constant flow rate. Before introduction of the reactant stream into the convection oven, the precursor solution passes through an integrated one-way valve designed to prevent backflow caused by evolution of volatile products downstream. Upon entering the convection oven pre-heated to 220 ˚C, rapid NP nucleation occurs as a result of the fast temperature ramp rate yielding a dark-colored suspension.36 At the reaction temperature, gas is evolved from the system at a rate of 317 mL h–1, causing the stream to separate into discrete plugs (Supporting Information). The plugs are isolated from one another, preventing axial dispersion, while internal circulation within each plug in the coiled tubing facilitates thorough mixing.37,38 Upon exiting the reactor, Ni-NP growth is halted by rapid heat dissipation from the reaction stream. The flow rate was maintained at 133 mL h–1 of solution (measured residence time of 16 min) to yield Ni-NPs denoted as mF-Ni-NPs. The yield of the mF-Ni-NPs was estimated to be 62% relative to the Ni(acac)2 precursor by gravimetric analysis. Based on the flow rate and yield, the

3

ACS Paragon Plus Environment

ACS Sustainable Chemistry & Engineering

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

throughput of a single channel mF device is >27 g of Ni-NPs per day, which would equate to >0.5 kg of 5 wt% SiO2 supported Ni-NP catalyst. As a result of the superior heat and mass transport properties inherent to a mF system, the residence time needed to generate similarly sized NPs is significantly shorter than the 2 h reaction time reported for the small-batch synthesis, as reported by Carenco et al.35 The reduced reaction volume also promotes fast heating and cooling, which minimizes energy inputs and reaction times. In an effort to synthesize Ni-NPs in a batch reaction under analogous conditions to the mF system with respect to total reaction volume, time, and temperature (albeit with a slower heat ramp rate), a batch reaction was performed at 220 ˚C and thermally quenched after 16 min, to yield NPs denoted as B-Ni-NPs in a lower yield of 45%. We attribute the higher yield of the continuous flow synthesis to result from the superior heat transfer and mixing conditions compared to batch. Scheme 1. mF reactor system for the continuous flow production of Ni NPs.

2.2 Characterization of colloidal Ni-NPs Powder X-ray diffraction (XRD) patterns of the resulting mF- and B-Ni-NPs exhibited the diffraction peaks of face-centered cubic structure Ni (Figure 1a,b). Three peaks at 44.6˚, 51.9˚, and 76.4˚ 2θ can be indexed to the (111), (200), and (220) reflections, respectively. A lattice parameter of a = 3.52 Å was calculated for both the mF-Ni and B-Ni-NPs, which is in agreement with bulk Ni (PDF# 99000-2639). Because Ni is more susceptible to oxidation than noble metals, it is important to note that no crystalline impurity phases of cubic NiO, hexagonal Ni(OH)2, or the β- or γ-phase of NiOOH were observed. Transmission electron microscopy (TEM) revealed a spherical morphology for both systems (Figure 1c,d). The mF-Ni-NPs possessed slightly larger sizes (11.1 ± 3.1 nm) than the analogous batch prepared samples (8.8 ± 2.4 nm). Both methods gave a polydispersity with standard deviations about the mean diameter (σ/d) of 27%. The average crystallite size determined by Scherrer analysis was 6.4 nm and 5.5 nm for the mF-Ni and B-Ni NPs, respectively. This difference in particle and crystallite sizes has been previously reported for colloidal Ni-NPs, and results from their polycrystallinity.35 Highresolution TEM micrographs displaying lattice fringes are provided in Figure 1g,h and further corroborate the NP crystallinity. A d-spacing of 0.20 nm was calculated from the observed lattice fringes, corresponding to the (111) set of lattice planes.

4

ACS Paragon Plus Environment

Page 4 of 16

Page 5 of 16

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

ACS Sustainable Chemistry & Engineering

Figure 1. (a,b) XRD patterns with reference peaks for Ni (PDF# 99-000-2639), (c,d) TEM micrographs, and (e,f) corresponding histograms of colloidal mF-Ni (11.1 ± 3.1 nm; σ/d = 27%) and B-Ni-NPs (8.8 ± 2.4 nm; σ/d = 27%), denoted by green and purple colors, respectively. (g,h) High-resolution TEM micrographs displaying lattice spacing of the (111) planes. 2.3 Preparation and characterization of SiO2-supported Ni-NPs The resulting colloidal Ni-NPs synthesized with the mF-device and under analogous batch conditions were subsequently supported on amorphous SiO2 targeting a metal loading of 5 wt%. This was achieved by adding a Ni-NP in CHCl3 dispersion dropwise to a stirring suspension of SiO2 in CHCl3.

5

ACS Paragon Plus Environment

ACS Sustainable Chemistry & Engineering

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

After 12 h of stirring, the resulting catalyst was recovered by centrifugation and dried overnight in vacuo. The final metal loading for each catalyst was determined by ICP-OES and is given in the Supporting Information, Table S1. The application of colloidal NPs in industrial catalysts is extremely rare, compared to NP catalysts prepared by precipitation or impregnation methods. Therefore, we also employed a conventional bulk IW impregnation technique to prepare a control catalyst for comparison, denoted as IW-Ni. TEM micrographs confirmed that all supported Ni NP catalysts were well dispersed on the SiO2 support (Figure 2a,b). Similar average diameters and polydispersities were observed for the mF-Ni NPs (11.8 ± 3.0 nm; σ/d = 25%) and B-Ni-NPs (9.2 ± 2.2 nm; σ/d = 25%) with respect to their presupported colloidal suspensions, demonstrating that, as expected, the supporting procedure did not affect the NP size or size distribution. The resulting IW-Ni particles exhibited an irregular spherical morphology with an average diameter of 15.4 ± 7.4 nm (Figure 2c). In addition, particles with diameters >20 nm were observed throughout the support and there was a two-fold increase in polydispersity (σ/d = 48%) relative to the colloidal Ni-NPs. One of the reasons that solution-phase syntheses exhibit increased morphological control is the incorporation of stabilizing agents. Changes in the oleylamine and trioctylphosphine concentrations have previously been identified to have large influences on the resulting Ni-NP morphologies; the presence of primary amines can increase the reduction rate, while long-chain phosphines control particle size through coordination to the Ni0 surface.35 Accordingly, the solution-synthesized materials displayed a more regular size and morphology as compared to the IW-Ni particles. Bulk and surface characterization techniques were employed to characterize the supported Ni-NPs. XPS spectra of the Ni2p region of the supported Ni-NPs is provided in Figure 2g and the full survey scans are provided in the Supporting Information (Figure S1). Assignment of this region is often complicated by surface oxidation of Ni into higher-valent NiO, NiOOH, and Ni(OH)2.39 The Ni2p3/2 and 2p1/2 doublets, and corresponding satellite peaks of these oxidized species, have binding energies that are close to that of Ni0 metal, causing overlap of the corresponding multiplets and shake-up structures. The B-Ni-NPs exhibited a Ni2p3/2 peak binding energy of 852.9 eV with a spin-orbit splitting of 17.3 eV, matching closely with literature values for Ni0.39,40,41 The peaks at higher binding energies (~862 eV) are the respective satellite peaks for Ni2p3/2. In addition to the primary 2p3/2 and 2p1/2 Ni0 peaks, the mF-Ni-NPs displayed a peak at 855.9 eV, which can be attributed to Ni(OH)2.42 The IW-Ni NPs Ni2p spectrum was indistinct, which may result from reduced surface area-to-volume ratio due to the larger particle size. Qualitatively, the broad peak that spans the Ni0, NiO, and Ni(OH)2 regions suggests that there is a combination of Ni0 and oxidized Ni species on the surface of the IW-Ni catalyst. Additionally, XPS spectra of the colloidal Ni-NPs displayed peaks in the N1s region that are expected for the presence of OAm; the I-Ni NPs did not display this N1s peak as a result of the lack of OAm in the catalyst preparation (Supporting Information, Figure S2). XRD analysis of the supported catalysts suggests that the oxidized Ni species observed via XPS are limited to an amorphous or thin surface layer, as all samples only exhibited characteristic diffraction peaks corresponding to the fcc structure expected for Ni0 (Figure 2h). The diffuse amorphous background observed at lower 2θ is attributed to the amorphous SiO2 support. Consistent lattice parameters along with the absence of maxima corresponding to oxidized Ni species, further confirm that no change occurred during supporting of the colloidal Ni-NPs. It is also important to note that the presence of oxidized Ni surface species are likely to be completely reduced to Ni0 prior to catalytic testing as part of catalyst activation process (Supporting Information, Figure S3).

6

ACS Paragon Plus Environment

Page 6 of 16

Page 7 of 16

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

ACS Sustainable Chemistry & Engineering

Figure 2. TEM micrographs and their corresponding size histograms of silica supported (a,d) mF-Ni-NPs (11.8 ± 3.0 nm), (b,e) B-Ni-NPs (9.2 ± 2.2 nm), and (c,f) IW-Ni particles (15.4 ± 7.4 nm). (g) Highresolution XPS spectra of the Ni2p region and (h) XRD of mF-Ni-NPs (green), B-Ni-NPs (purple), and IWNi particles (pink). The solid and dashed lines in the XPS spectra refer to the Ni0 and Ni(OH)2 main peak binding energies, respectively. The H-adsorption site density of the supported Ni catalysts was measured using H2 chemisorption after in-situ reduction at 350 ˚C (based on catalytic reaction temperature) for the NiNPs and 450 ˚C (based on temperature programmed reduction) for IW-Ni. Comparable H-adsorption site densities of 36 µmolH* gcat–1 for mF-Ni-NPs and 32 µmolH* gcat–1 for B-Ni-NPs were determined. In contrast, the IW-Ni catalyst had a low H-adsorption site density of 5.4 µmolH* gcat–1 that is attributed to the larger particle size. The similar H-adsorption site densities for the mF-Ni-NPs and B-Ni-NPs agree with the structural characterization data (Table S1) and further highlight the efficacy of the mF approach to prepare NPs that are quantitatively similar to those prepared in batch.

7

ACS Paragon Plus Environment

ACS Sustainable Chemistry & Engineering

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

2.4 Catalytic performance in the deoxygenation of guaiacol It was recently demonstrated that a batch-prepared Ni-NP catalyst outperformed an analogous IW-Ni catalyst in terms of activity and selectivity when comparing values at 8 h time on stream in the deoxygenation of guaiacol under ex situ CFP conditions.20 A conceptual process design with technoeconomic analysis of the ex situ CFP route at commercial scale has identified that catalyst lifetime is likely more important than initial activity.18 Considering this, the three Ni/SiO2 catalysts prepared here were evaluated in the deoxygenation of guaiacol under ex situ CFP conditions (350 °C, 0.5 MPa) to compare activity, selectivity, and deactivation. No effort was made to remove the stabilizing ligands on the surface of the NP-catalysts prior to reaction. Catalyst activity was first assessed by comparing the conversion of guaiacol to liquid products (Figure 3a). The IW-Ni catalyst deactivated rapidly in the initial 250 min, decreasing from 52 to 8.2% conversion. In contrast, the B-NiNP catalyst displayed a minimal decrease in conversion from 59 to 52% over 420 min, and the mF-NiNP catalyst decreased from 72 to 34%. The conversion data at time on stream past 200 min support the stark difference in activity between the IW-Ni catalyst and colloidal Ni-NP catalysts reported previously.20 Further insight into catalyst activity was obtained by normalizing the product conversion rate to H-adsorption site density to give site-time yield (STY) values (Figure 3b). The STY values were determined at conversion levels between 5.4 and 72%, and thus, the results do not necessarily reflect the intrinsic rates of each material. Instead, they allow for a comparative analysis of the catalysts at a single reaction condition relevant to ex situ CFP. The IW-Ni catalyst exhibits the highest initial STY (1.9 s–1), but rapidly deactivates to 0.20 s–1. The B-Ni-NP catalyst demonstrated a lower initial STY of 0.31 s– 1 , but exhibited minimal deactivation. The mF-Ni-NP catalyst demonstrated an initially comparable STY value (0.33 s–1) compared to B-Ni-NP, but the moderate deactivation resulted in a lower STY (0.15 s–1) compared to B-Ni-NP (0.27 s–1) at 420 min. The relative deactivation rates were quantified using linear fits to STY versus time, employing the full datasets for both NP catalysts and the initial 200 min data for the rapidly deactivating IW-Ni catalyst. Compared to the minimal deactivation exhibited by the BNi-NP catalyst, a 110-fold faster deactivation rate was observed for the IW-Ni catalyst, and a 3.6-fold faster deactivation rate was observed for the mF-Ni-NP catalyst. Comparing the IW-Ni and mF-Ni catalysts, the two catalysts having synthetic methods that are more amenable to scale-up, these data highlight a 31-fold decrease in deactivation rate exhibited by the mF-Ni-NP catalyst relative to IW-Ni. Considering the significant difference in STY deactivation rates, we propose two hypotheses: (1) There may be two different active site types on these materials, where IW-Ni has both high-activity sites that deactivate quickly and lower-activity sites that are more stable. The NP catalysts do not possess the high-activity sites, possibly due to ligand blocking, but consist of the lower-activity sites that resist deactivation. (2) The active sites are the same on the IW and NP catalysts, but those on the IW catalyst deactivate rapidly while those on the NP catalysts are more resistant to deactivation.

8

ACS Paragon Plus Environment

Page 8 of 16

Page 9 of 16

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

ACS Sustainable Chemistry & Engineering

Figure 3. (a) Guaiacol conversion as a function of time on stream and (b) product STY as a function of time on stream for mF-Ni-NPs (green), B-Ni-NPs (purple), and IW-Ni particles (pink) under ex situ CFP conditions (350 ˚C, 0.5 MPa).

9

ACS Paragon Plus Environment

ACS Sustainable Chemistry & Engineering

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

Guaiacol is an interesting model compound due to its multiple functional groups that enable investigation into the relative selectivity for a variety of reactions (e.g., demethylation, demethoxylation, direct deoxygenation, hydrogenation and hydrodeoxygenation). Thus, the reaction network associated with the deoxygenation of guaiacol can be complex.16,20,43,44 Common initial reaction steps have been proposed where either demethylation (DME), demethoxylation (DMO), or direct deoxygenation (DDO) yield catechol, phenol, or anisole, respectively (Supporting Information, Scheme S1). These products can react further to yield the fully deoxygenated aromatic product benzene and ring-hydrogenated cyclohexanone, cyclohexanol, and cyclohexane. The molar selectivity of the organic phase products (i.e., with carbon numbers ≥5) observed from the Ni catalysts explored here is presented in Figure 4. The products are grouped as containing no oxygen (e.g., cyclohexane, benzene), one-oxygen (e.g., phenol, anisole, cyclohexanol, cyclohexanone), and two-oxygen atoms (e.g., catechol, 2-methoxycyclohexanol, 1,2-dimethoxybenzene). The individual product selectivity data are presented in the Supporting Information, Table S2. Low selectivity (