Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Article
Pd supported on Carbon Nitride Boosts the Direct Hydrogen Peroxide Synthesis. Rosa Arrigo, Manfred E. Schuster, Salvatore Abate, Gianfranco Giorgianni, Gabriele Centi, Siglinda Perathoner, Sabine Wrabetz, Verena Pfeifer, Markus Antonietti, and Robert Schloegl ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.6b01889 • Publication Date (Web): 01 Sep 2016 Downloaded from http://pubs.acs.org on September 6, 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 Catalysis 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 18
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 Catalysis
Pd Supported on Carbon Nitride Boosts the Direct Hydrogen Peroxide Synthesis Rosa Arrigo,* †‡ Manfred E. Schuster,‡ Salvatore Abate,§ Gianfranco Giorgianni,§ Gabriele Centi,§ Siglinda Perathoner,§ Sabine Wrabetz, ‡ Verena Pfeifer, ‡ Markus Antonietti, δ Robert Schlögl, ‡†. †
Max-Planck-Institut für Chemische Energiekonversion, Stiftstrasse 34-36, 45470 Mülheim an der Ruhr, Germany
‡
Fritz-Haber Institut der Max-Planck Gesellschaft, Faradayweg 4-6, 14195 Berlin, Germany
§
Università degli Studi di Messina, V.le F. Stagno D'Alcontres 31, 98166 Messina, Italy
δ
Max-Planck-Institut für Kolloid und Grenzflächenforschung, Am Mühlenberg 1 OT Golm, 14476, Potsdam, Germany
*Correspondence to:
[email protected] ABSTRACT: Herein, the development of an improved Pd on carbon nitride catalyst for the direct H2O2 synthesis from the elements is reported. Microcalorimetric CO chemisorption is used for characterizing the chemical speciation of the Pd selective and unselective sites. Selectivity trends among the samples suggest that a bare metal surface with a differential heat of CO chemisorption ranging between 140 and 120 kJ*mol-1 is responsible for the total O2 hydrogenation, while a maximum threshold value of differential heat of CO chemisorption of approximately 70 kJ*mol-1 is necessary for the partial hydrogenation of O2 to H2O2. Such low differential heat of CO chemisorption indicates a low exposed metallic Pd surface subjected to electron-withdrawing from the surrounding ligands, i.e. the N functional group on the carbon support. With respect to Ncontaining carbon nanotubes, carbon nitrides provide: higher concentration of N sites; a flexible network of π-conjugated polymeric subunits with sp3 linking subunits; a flakes-like morphology with high exposure of reactive C edge terminations. This results in a more effective kinetic stabilization of the electronically modified Pd species.
KEYWORDS: Heterogeneous Catalysis · Palladium · Carbon nitrides · NK edge NEXAFS · Hydrogen Peroxide INTRODUCTION The simplification of a chemical process into a direct synthesis is one of the objectives of the sustainable chemistry, which in the case of the synthesis of hydrogen peroxide (H2O2), faces two major challenges: it requires the reactor to operate at any time out of the H2/O2 explosion limits; it requires the development of stable catalysts, which are able to achieve high productivities of H2O2 while avoiding unwanted, facile water formation (Scheme S1). As shown by the large variety of patents [1] (Table S1) and scientific publications (Table 1) [1-10] this synthesis requires action at different levels: from the optimization of the catalyst formulation to the experimental reaction conditions, i.e. T, gas composition, use of stabilizing additive as well as the fluid-dynamic condition realized in the reactor. Despite its fast deactivation, Pd on carbon catalysts provide very high activity and selectivity. [2] Tuning the surface chemistry of the carbon support has proved to be an effective strategy to further improve the catalytic performance and lifetime. [2,3,5-6] By tracing the pronounced change of the catalyst structure accompanying the dramatic change of the catalytic performance, we previously established the active and selective form of Pd that is realized by site-blocking of electronically modified Pd0 nanoparticles (NPs) on N-containing carbon nanotubes (NCNTs). [6] In this study, we attempt to develop a synthetic strategy to maximize the relative abundance of the selective sites. The approaches considered are: the minimization of the Pd loading in a synthetic procedure that
favours Pd/N interactions; the use of carbon support with high concentration of N species, thermally stable at temperature around 773K. We used carbon nanotubes (CNTs) treated in NH3 at 873 K (NCNT873) and 473 K (NCNT473) [11] and, CxN synthesized according to [12] and thermally annealed at 773 K. We report a more efficient and long-lasting kinetic stabilization of the selective Pd species achieved by using CxN support. By means of high-resolution synchrotron-based X-ray absorption and photoelectron spectroscopy (XPS), transmission electron microscopy (TEM), electron energy loss spectroscopy (EELS) and microcalorimetric CO chemisorption, we attribute the special performance of Pd on CxN to the higher N abundance as well as the lower graphitic structural order of the CxN, if compared to NCNTs. The dynamic interactions between the Pd NPS and the CxN support induced by the reactive environment guarantee the electronic modification of the metal surface exposed to O2 and H2, being this a necessary condition for selective hydrogenation.
EXPERIMENTS Preparation of the support: The functionalization of the CNT (Pyrograf. Inc) was carried out as reported in [11]. Accordingly, 10 g of CNTs (grade PR24PS) were suspended in 500 ml HNO3 conc. (70% Sigma-Aldrich), heated to 373 K and kept at this temperature for 2 h followed by filtering, rinsing and drying overnight at 393 K. Afterwards 10 g of the oxidized CNT were treated in
1 ACS Paragon Plus Environment
ACS Catalysis
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
NH3 (200 ml/min) for 4 h to obtain NCNT. Two different NCNT batches, namely NCNT473 and NCNT873, were obtained by NH3 treatment at 473 K and 873 K, respectively. These two batches of NCNTs were used to immobilize the Pd nanoparticles. The carbon nitride sample was prepared following the procedure in [12]. Accordingly, molten cyanamide (1 g, 24 mmol; Sigma-Aldrich) was heated and stirred at 343 K, and a 1 g of a 40 wt % dispersion of SiO2 particles (Ludox HS40, Aldrich) in water was added dropwise to establish a 40 vol% porosity. The resulting transparent mixture was then heated at a rate of 4.5 K min−1 over 2 h to reach a temperature of 848 K and then kept at this temperature for another 4 h. The resulting yellow powder was treated with 25 ml of 4 M NH4HF2 for two days at room temperature to remove the silica template in a closed polypropylene bottle flushed with nitrogen. The powder was then centrifuged and washed three times with distilled water and twice with ethanol. Finally the powder was dried at 343 K under vacuum for several hours. We refer to this support as fresh CxN. The fresh CxN support was subsequently thermally annealed in helium at 773 K for 2 h. In this manuscript we report the structural characterization of the annealed support. The thermally annealed CxN was used to immobilize Pd nanoparticles via sol immobilization according to the procedure below and the catalytic data are herein reported for this catalyst. The fresh support (before the thermal annealing) was also used to immobilize Pd nanoparticles according to the procedure described below, however the catalytic data are not shown because the catalyst was inactive.
Preparation of 5 g of Pd monometallic samples (theoretical 2 wt%) via sol immobilization (SI): Na2PdCl4 from Aldrich (99.99% purity), NaBH4 (>96% purity) from Fluka and polyvinyl alcohol (PVA) (Mw = 13,000–23,000, 87–89% hydrolyzed) from Aldrich were used. First, aqueous solutions of PVA (1 wt%) and NaBH4 (0.1 M) were prepared. 0.94 mmol solid Na2PdCl4.2H2O and 10 ml of PVA solution were added to 250 ml of H2O and mechanically stirred with a stirring rate of 1000 rpm. After 3 min of stirring, 40 ml of a freshly prepared 0.1 M solution of NaBH4 was added. The colloidal solution was then acidified with 2 ml of H2SO4 1 M and the functionalized NCNT (4.9 g) were quickly added to the solution for the immobilization. The so-obtained suspension was kept under vigorous mechanical stirring for 1 h. The samples were then filtered, washed and dried overnight at 353 K. Samples notations are as following: Pd1.8%873 and Pd1.7%473. A similar procedure was used to synthetize Pd catalysts 0.3 wt % on NCNT873 and CxN freshly prepared and annealed at 770K. PVA (0.01 g) was added to 100 ml of distilled water and kept under stirring for 5 minutes. Afterwards, 0.14 mmol of Na2PdCl4 were added. Following, 0.014 g of solid NaBH4 were added to the solution under stirring at 1250 rpm for 30 minutes. Finally 2 ml of H2SO4 1 M were added. The sol was then added to each support (NCNT, 4.98 g) under stirring for 30 minutes. The catalyst was then filtered, rinsed with distilled water and dried overnight at 353 K. The samples are denoted as following: Pd0.3%NCNTSI and Pd0.3%CxN.
Page 2 of 18
Preparation of Pd monometallic on NCNT873 samples (theoretical 0.3 wt %) via incipient wetness impregnation (IWI) method: The nanotubes were impregnated with an aqueous solution of Pd(NO3)2 (10ml/gcat). For this purpose the required volume of a commercial solution of Pd(NO3)2 (sigma Aldrich 8.5% Pd wt/wt) was added to a 0.01 M HNO3 solution and dropwise added to the NCNT. The slurry was then sonicated in an ultrasonic bath for 10 sec and afterwards the solvent was evaporated slowly at room temperature in air for 48h. Thereafter, the samples containing the metal precursor were oxidized in 5% O2 at 473 K for 2h and reduced in 5% H2 at 473 k for 2h. The sample is denoted as Pd0.3%NCNTIWI.
Catalytic test: All catalysts were tested at room temperature in the catalytic hydrogenation of O2 to H2O2 in a stirred stainless steel reactor coated with Teflon (capacity 300 ml) containing 35 mg or 70 mg of catalyst in a fine powder form and 125 ml of anhydrous CH3OH as a reaction medium. 125 µL of H2SO4 were added for H2O2 stabilization. The autoclave operates in semi-batch conditions, e.g. continuous feed of the gas to the well mixed autoclave. The concentration of H2 was 9 vol%, and the O2/H2 ratio was 7. The ballast gas was CO2. The feed is continuously bubbled through the reaction medium at room temperature and at a total pressure of 3 MPa. Stirring (1300 rpm) was started after reaching the pressure of 3 MPa, and experiments were carried out for 180 min, monitoring both H2O2 and H2O concentration in the liquid phase, and H2 and O2 concentrations in the gas phase. Gas analysis was performed using a gas chromatograph unit (Agilent 3000A equipped with a Molsieve 5 Ǻ column using Argon as carrier gas). The reaction products were analyzed by potentiometric titrations of H2O2 (Metrohm, 794 Basic Trino) and H2O (Metrohm, 831 KF Coulometer), respectively. The supports were also tested but have not shown activity. The productivity reported in Figure 3 is calculated as ([H2O2]t+1 -[H2O2]t) / Pd loading (mg) · t (h). The cumulative productivity reported in Figure S4 is calculated as ([H2O2] / Pd loading (mg). The selectivity is calculated as [H2O2] / [H2O2] · [H2O2]
Catalyst Characterization: High-resolution transmission electron microscopy (HRTEM) and high angle annular dark field scanning transmission electron microscopy (HAADF-STEM) have been performed with a TitanCs300 microscope from FEI equipped with Cs-corrector from CEOS. The samples were dry-deposited on a holey carbon film supported on a Cu grid. CO-chemisorption was carried out with a SETARAM MS70 Calvet calorimeter combined with a custom-designed high vacuum and gas dosing apparatus. An absolute pressure transducer (MKS Baratron type 121) measures pressure variations of 0.003 mbar in the dosing volume of 139 ml (provided the box temperature does not vary by more than ±1.5 K) and allows (as a conservative estimate) dosing as low as 0.02 µmol into the sample cell. All samples were first pressed under low pressures (125 MPa) and cut into small
2 ACS Paragon Plus Environment
Page 3 of 18
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 Catalysis
pellets, which were sieved to a diameter of 0.4-0.6 mm and then placed into the volumetric cell. After evacuation at 313 K for 17 h to a pressure ≤10−8 mbar, the cell was closed and the CO pulses were dosed stepwise at 313 K. Pressure (mbar), adsorption temperature (°C) and the heat signal (V) were recorded. We have adopted the calorimetric sign criterion as positive energetic quantity for an exothermic process. The pressure-controlled dosing systems with calibrated volume allows for the detection of the amount of adsorbed molecules (adsorption isotherm) as well as differential heat of adsorption (integral heat/amount of adsorbed molecules) and gives the possibility to elucidate the distribution of the adsorption sites along the range of adsorption heats. XPS and NEXAFS measurements were carried out at the ISISS end station and beamline at BESSY II. The freshly prepared powders or used samples collected from the reaction vessel and exposed to atmospheric environment were pre-formed in pellet and directly exposed to vacuum (10-7 mbar) in the XPS chamber without any pretreatment. The samples proved to be stable as time related dynamics of the XPS peak were not observed upon exposure to vacuum and the X-ray beam. XPS measurements were performed applying a suitable excitation energy corresponding to a kinetic energy (KE) of the photo-emitted electrons of 600 eV for the core levels Pd3d, C1s, O1s and N1s. The energy pass Ep was normally set to 20 eV. The Pd3d envelopes were fitted using Casa XPS software after subtraction of a Shirley background. The fitting of the spectra was done constraining the peak position by ±0.05eV. The fitting of the Pd3d consists of four components Pd0 and Pd1 with Doniach-Sunjic (DS) line shape and Pd2 and Pd3 with Gaussian-Lorentzian (GL) line-shape. The area ratios between the Pd3d5/2 and Pd3d3/2 spin orbit split transitions was constrained to the theoretical value of 3:2 and the distance between the two spin orbit split transition was 5.3 eV. Binding energies (BEs) were referenced to the Fermi edge recorded after each core level measurement. Quantification of the elemental composition was carried out according to homogeneous model distribution. For quantification the spectra have been normalized to the impinging photon flux. Auger Electron Yield (AEY) NEXAFS spectra were recorded with an analyzer setting of 353 eV electron kinetic energy (KE) and 100 eV pass energy (Ep), while the beamline setting was exit slit (ES) 111µm and fix focus constant (cff) 1.4 (cff=cosα/cosβ). The kinetic energy window was chosen such to avoid photoelectrons moving through the NEXAFS spectrum while sweeping the excitation energy, while broad Ep was necessary to obtain reasonable intensity. The exit slit value chosen enables an optimal compromise between high photon intensity and good spectral resolution. The higher order suppression operation mode of the monochromator was applied (fix focus constant cff=1.4) to avoid contributions to the background in NEXAFS spectra that might complicate intensity normalization of the spectra on impinging photon flux.
RESULTS AND DISCUSSION The surface chemistry of the carbon supports was characterized by XPS and NEXAFS. A broad distribution of N spe-
cies is introduced on the surface of CNTs by the NH3 treatment (Figure S1a): at 473 K, sp2/sp3 linear or cyclic aliphatic CNH moieties are prevalent (N2 and N3 components in the XP spectra, respectively), while at 873 K the abundance of substitutional N is enhanced, i.e. pyridine-like N (N1) and graphitic N (N4). [11, 13] The assignment of resonances in the N K edge NEXAFS spectra (Figure 1) according to ref. 13 is as follows: three resonances R1, R2, R3 between 399 eV and 402 eV, are caused by 1s→π* electronic transition for N atoms in N-C bonds. The resonance due to 1s→ σ* transition is found at around 408 eV (R6). The R1, R2 and R3 resonances occur at around 398.6 eV, 399.4 eV and around 401~ 401.4 eV, respectively. The R1 resonance has been assigned to nitrogen atoms with two carbon neighbors in pyridinelike configuration; the R2 resonance is attributed to sp2/sp3 C=N-H bond configuration and the resonance R3 to the socalled „graphitic“ three-fold nitrogen atom. The R5 resonance at 403.8 eV is present on the CNT oxidized and can be attributed to N-O bonds resonances. Similarly, N in CxN is present as aromatic N species (R1 and R3 in Figure 1 and N1and N4 in Figure S1b, respectively) and mainly in a chemical bond configuration which resembles the sp3/sp2 -C=N-H↔=C-N-H enamine/imine tautomer moieties observed for NCNT473 (R2 in Figure 1 and N2/N3 in Figure S1b). On the CxN, the resonance R4 at ~ 402.3 eV is particularly intense. This resonance is indeed typical for carbon nitrides and was assigned to the sp3, three-fold N atoms bridging the basic structural units. [14] The structural assessment of the resonances in the NK edge spectrum is depicted in Scheme 1 for the most commonly proposed structure of CxN. [14] XPS elemental analysis indicates that the CxN phase presents inhomogeneous N abundance (ranging from 8 at% to 16 at% in Figure S1) and a poorly graphitic C-enriched surface (Figure S2). The TEM image in Figure 2 shows flakeslike CxN particles in which polycrystalline graphitic domains exist (see selected area electron diffractogram in the inset). The more bulk sensitive C1s core excitation EELS spectrum of CxN is compared in Figure 2b to highly oriented pyrolitic graphite (HOPG), which is considered a reference for an sp2 hybridized carbon structure without inplane defects (black line) and to NCNTs synthesized via catalytic chemical vapor deposition as in ref. 13 (green line). The spectra have been normalized to the σ* resonance for comparison after power-law background subtraction and plural scattering removal. In HOPG and NCNT_CCVD, the π* resonance at around 285.6 eV and the σ* resonance at around 292 eV, originates from 1s→ π* and 1s→ σ* electron excitations of the C=C bond (in the ring), respectively. The spectral features between 286 and 290 eV are assigned to the Rydberg transitions, which are C1s-π* transition for C-R bond where R is a N, H or O atom characteristic for functionalized aromatic groups. [15] Deviations from the in-plane structural order of HOPG produces a decrease in the π*/σ* signal ratio and broadening of the π*resonance, while the fine structure above 290 eV becomes unresolved. Therefore, the CxN phase possesses a sp2-bonding configuration but without long range order. Moreover, the Rydberg transition region is more intense in comparison with HOPG and NCNTs indicating a higher amount of functional groups in sp3 configuration. The quantification of the N and C abundance from the EELS spectra was 38 at% and 62 at%,
3 ACS Paragon Plus Environment
ACS Catalysis
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
respectively, while O was not detected in the bulk of the sample (in contrast to the surface as shown in the surface sensitive XP spectra in Figure S3).
Page 4 of 18
a)
b)
Figure 1: Auger electron yield NK edge NEXAFS spectra for pristine CxN (black line), NCNT873 (magenta line), NCNT473 (blue line), NCNT synthetized via catalytic chemical vapor deposition as in ref. 13 (green line), CNT oxidized with HNO3 (orange line). The intensity of the CxN spectrum is divided by 5 for a better comparison with the other spectra. Figure 2: TEM image of the pristine CxN and corresponding selected area electron diffractogram (SAED) in the inset (a); energy electron loss spectra (b).
Scheme 1: Simplified CxN structural unit reporting all the N species as observed in the NEXAFS spectrum: R1 is pyridine N 2 3 (in red); R2 is sp /sp C=N-H in terminal amine moieties (in blue); R3 is attributed to graphitic-type N (in green); R4 is at3 tributed to sp N atoms bridging the basic structural units (in orange).
The sequence of the formation of the various N species through the condensation of reagent molecules is exemplified in scheme S2 of the supporting information. Thus, we propose a bulk structure of this CxN that is close to the structure proposed for g-C3N4 and composed of a network of sp3 N atoms linking π-conjugated polymeric subunits. The subunits contain smaller amount of N atoms than in the proposed tri-s-triazine ring subunits of g-C3N4, [14, 16] and are here formed of condensed rings, with limited aromatic π-delocalization. N is mainly present as sp2/sp3 C=N-H↔=C-N-H enamine/imine N functionalities rather than the pyridine N. The above described CxN phase is covered by a C enriched over-layer with predominant sp3 character.
Colloidal Pd metal NPs stabilized by polyvinyl alcohol (PVA) were immobilized on the NCNTs as well as on the CxN. Our initial attempt [6] was devoted to achieving high Pd loading with the aim of exploiting all the N functionalities on the carbon surface, which serve as anchoring sites for the NPs. This yielded 1.8 wt% Pd on NCNT873 (denoted as Pd1.8%873) and 1.7 wt% on NCNT473 (denoted as Pd1.7%473). Those samples showed impressive H2O2 productivity after a short induction period (Figure 3a). In particular, Pd1.8%873 shows an outstanding productivity and selectivity after 20 minutes of reaction. The H2O2 productivity of Pd1.8%873 increased further to a maximum of 400 µmol*mgPd-1*h-1 accompanied by a decrease of selectivity due to water formation (Figure 3b, Table 1, Figure S4a). Pd1.7%473 shows a similar "bell-shape" activity profile, but inferior catalytic performance characterized by a lower H2O2 productivity and a rapid increase in water formation (Figure S4b, Table 1). The different catalytic behavior of the two catalysts clearly suggests that only some of the N species are beneficial and, they may be responsible for the formation of small “raft-like” Pd NPs (i.e. see Figure S5), which are more abundant on NCNT873. [5] To reach for higher selectivity, we have synthesized two additional catalysts in which the Pd loading was kept very low, aiming for selective immobilization of the Pd NPs on the strongly interacting N sites of NCNT873. The Pd was loaded by sol-immobilization (Pd0.3%873SI) as well as by incipient wetness impregnation (IWI) (Pd0.3%873IWI).
4 ACS Paragon Plus Environment
Page 5 of 18
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 Catalysis
Table 1. Catalytic performance of Pd catalysts. a)
n
Catalyst
This work[a] Pd1.8%873 Pd1.7%473 Pd0.3%873SI Pd0.3%873IWI Pd0.3%CxN b)
r
S(%)
reaction time (min)
O2:H2 ratio
Mod e
248 239 57 10 140 82
93 39 48 38 68 58
20 20 20 20 20 90
7 7 7 7 7
SB SB SB SB SB
1384 143 175 [f] 0.3 0.6[f] 61
NA NA >98 >70 70 96
0.5 30 30 240 60 30
2
B
2 15 12 5
B B SB B
Product (µmol H2O2 -1 -1 mgCat h )
[b]
Literature [7] Pd1%{Au}/C Pd2.5%Au/C [2][c] [d][8] Pd0.6%/XCPd5%/C[e][9] [10] Pd1%Sn/TiO2
-1
Figure 3: H2O2 productivity normalised for the Pd loading (a) and selectivity (b) for: Pd1.8%873 (black square), Pd1.7%473 (red circle) and Pd0.3%CxN (blue triangle).
Scanning electron microscopy (SEM) and energy dispersive X-ray (EDX) analysis confirmed the immobilization of the Pd NPs (Figure S6-S7) with the expected metal loading (Figure S8). The reduction of the Pd loading did not lead to high activity and selectivity (Figure S4c-d). For such a low loading, the sol immobilization method requires an optimization of synthesis conditions to improve the NPs dispersion (not shown), which was not further investigated here. The reason for the poor activity of the impregnated sample will be clarified later on. The outstanding effect of N in the carbon matrix is corroborated by the catalytic performance of the Pd0.3%CxN also reported in Figure 3 and in Figure S4e. Despite the very low loading, this catalyst enabled the highest productivity of 460 µmol*mgPd-1*h-1 after 20 minutes of reaction. In terms of selectivity, at long reaction time this catalyst outperforms any other catalysts we have investigated, demonstrating an incomparable stability under the harsh reaction conditions. We further verified that the CxN alone is inactive in this reaction as well as the immobilized Pd NPs on the freshly prepared CxN (i.e. without annealing at 770K). Understanding the nanostructure dynamics behind the time evolution of the catalytic performance of Pd1.8%873, Pd1.7%473 and especially the relatively high selectivity at long reaction time of Pd0.3%CxN catalysts is crucial in order to guide the optimization of catalyst synthesis. The particle size distribution by High Angle Annular Dark Field Scanning Transmission Electron Microscopy (HAADF-STEM) of the fresh Pd/NCNTs ranges from subnanometer clusters to ~3nm NPs (Figure S9-10), where Pd1.8%873 exhibits a smaller average particle size (1.7±0.5 nm) than Pd1.7%473 (2.3±0.5 nm).
Reaction conditions: [a] catalyst amount 0.28 g ml for samples with 1.7%-1.8% Pd loading and 0.7 g ml-1 for 0.3% Pd loading; H2SO4: CH3OH = 0.001 (vol); O2:H2:CO2 = 63:9:28 vol%; total pressure 3 MPa; RT; [b] If not differently indicated, the values of productivity from the selected publications were con-1 -1 verted in µmol H2O2 mgCat h for a convenient comparison of the results in this work with literature data; [c] 10 mg catalyst, feed gas O2:H2=2, 4 MPa, 2 °C, 0.5 hours, CH3OH/H2O as solvent, pretreated by 2% CH3COOH for 30 min; [d] 50 mg 0.6 %wt Pd, atmospheric pressure, T=10 °C, Ratio O2/H2=15 and N2 as balance, C2H6O/H2SO4 as solvent;[e] O2/H2=12, CH3OH/NaBr, RT, 0.9 MPa, N2 as balance; [f] productivity (P) expressed in wt%. SB=semibatch: B=batch.
Figure 4: HAADF-STEM image of fresh Pd0.3%CxN (a) and used after 180 minutes of reaction (d); secondary electrons (SE) image of fresh (b) and used after 180 min of reaction (e) and overlapped HAADF-STEM/SE image for fresh (c) and used (f) samples. Green circle in (d) indicates very small nanoparticles; particles highlighted in red are examples of nanoparticles inside the pores of the CxN.
Already after 12 minutes of reaction, the particle size distribution underwent a significant change (Figure S9 and Table 2) with an increase of NPs size due to sintering. The comparison of the HAADF-STEM images for the fresh Pd0.3%CxN and the used one after 180 minutes of reaction
5 ACS Paragon Plus Environment
ACS Catalysis
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
(Figure 4d-f and Figure S11) also show a considerable sintering of the NPs to form large agglomerates (Figure S11), although small nanoparticles are still present (as for example those inside the green circle in Figure 4d). Note that this state corresponds to a comparatively high selectivity. Interestingly, in the used sample, some NPs are contained inside the rough, porous-like surface of the CxN (particles highlighted in red in Figure 4d), whilst on the fresh sample, the NPs are evenly covering the surface (Figure 4 a-c). The efficient embedding of the NPs at the CxN surface and the partial encapsulation by C over-layer, is clearly visible in the HRTEM of the used sample (Figure 5) and confirmed by the angular tilting experiment performed for a NP in the used sample (Figure S12).
Page 6 of 18
kJ/mol. This value is close to the value reported [18] for Pd (111) single crystal surface. CO chemisorption after 12 minutes of reaction for the
Pd1.8%873 and Pd1.7%473 shows the exposure of Pd surface (Table 2), which is, however, lower than that estimated by the TEM data (Table 2), indicating the partial removal of PVA by reaction with the solvent.[5]
Figure 5: HRTEM of Pd0.3%CxN after 180 minutes of reaction. Note the partial encapsulation of the NPs in the CxN matrix.
CK edge EELS in Figure 2 shows that the structure of the CxN was affected by the harsh environment of the direct H2O2 synthesis: the increase of the π* resonance intensity indicates an increased contribution from the π-conjugated domains (removal of the amorphous C-enriched domain). Moreover, we found that the surface N abundance increases in the CxN from the initial 16 at% on the fresh support to 50 at% on the used Pd catalyst (Figure S3) indicating the loss of the surface C enriched phase. We assume that this is a consequence of the reaction of the support with the reactive environment (H2O2, H2, O2) and the nanoparticles itself, which facilitate the mobility of the nanoparticles until they reach a more stable location. CO chemisorption was used to probe the exposed metal surface. The differential heats of adsorption as a function of the CO coverage are reported in Figure 6 for all the samples investigated, while the corresponding adsorption isotherms are reported in Figure S13. The exposed surface area for each sample extracted from the adsorption isotherm in Figure S13 applying the Langmuir model are summarized in table 2. Due to the initial coverage of the Pd NPs by PVA, CO (likewise O2 during the reaction) cannot access the Pd surface for chemisorption,[5] which explains the induction time in the H2O2 productivity observed for this type of catalysts (Figure 3). For example, the differential heat measured after the exposure of the fresh Pd0.3%CxN sample to CO is about 20 kJ/mol, typical for physisorption process. [17] In contrast, the unprotected metal surface of the fresh Pd0.3%873IWI gives rise to a much higher value of differential heat of CO adsorption with an initial value of 140 kJ/mol rapidly evolving to a plateau-like profile with a value of 120
Figure 6: Differential heat of chemisorption as function of CO uptake at 313 K for: Pd1.8%873 used after 12 minutes of reaction (black square); Pd1.7%473 used after 12 minutes of reaction (red circle); Pd0.3%873IWI fresh (magenta rhombus); Pd0.3%CxN fresh (pale blue triangle); 1st cycle of CO adsorption on Pd0.3%CxN nd used after 180 minutes of reaction (green rhombus) and 2 cycle of CO adsorption after desorption step (blue triangle).
Table 2. Samples characterization data. Catalyst
a
b
f
Pd1.8%873SI f Pd1.7%473SI g Pd0.3%CxNSI g Pd0.3%873IWI h Pd0.3%CxNSI
dPd [nm] 1.9±1.5 2.9±0.9 3.4±1.0 nd 5.0±2.0
c
Sth. 2 [m /g] 1.65 0.75 0.15
d
SCO 2 [m /g] 0.2 0.34 0.03 0.05 0.16
∆Hads._CO [kJ/mol] 67 96 20 140 i 185 h 80
e
[a] 12 min of reaction, [b] average NPs size determined by STEM, [c] Pd surface area estimated by the STEM data assuming semispherical exposure of NPs, [d] monolayer capacity of CO adsorption at 313 K estimated by the Langmuir model. [e] Initial differential heat of CO adsorption at 313 K, [f] used sample after 12 minutes of reaction, [g] fresh sample, [h] sample after 180 minutes of reaction, [i] first adsorption cycle, [h] second adsorption cycle. (nd) non determined.
The differential heat of chemisorption is also much lower than the 100 kJ/mol reported for unprotected 2 nm Pd NPs. [17] [18] This can be interpreted in terms of an electron withdrawing effect exerted by the species coordinated at the surface atoms of the Pd NPs. These species can be both the functional groups of the support surface as well as the residual PVA overlayer. Specifically, CO chemisorbs on Pd with a σ-donation from the nonbonding sp-hybridized electron pair on carbon to the empty d-orbitals of the metal and, a π-back donation from the filled d-orbitals of the metal to the π∗-antibondig orbital of the CO molecule. The more pronounced the π-backdonation, the stronger the PdCO bond. Thus a depletion of the Pd d-states electrons
6 ACS Paragon Plus Environment
Page 7 of 18
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 Catalysis
caused by the surrounding chemical environments manifests itself as weakening of the CO binding energy. In the case of the used Pd0.3%CxN after 180 minutes of reaction, the exposed surface substantially increases to 0.16 m2/g after 180 minutes of reaction. It is worth noticing that the exposed surface area determined by CO chemisorption is very similar to the average estimated semispherical exposure by TEM (Table 2) for the used sample after 180 minutes of reaction. Moreover, we observe a more pronounced heterogeneity in the energetic distribution of chemisorption sites. The initial differential heat of chemisorption is very high in the first CO pulse (185 kJ/mol) and reaches a value of about 70 kJ/mol in the second pulses, followed by a rapid decay (Figure 6). Pd0.3%CxN after 180 minutes of reaction is also characterized by a quite broad distribution of particle sizes (Table 2). We assume that NPs with large exposed Pd domain account for the differential heat of 185 kJ/mol. If compared to the value reported for the bare Pd surface, such a high value could be due to the additional contribution from the heat of reactions, such as CO reaction with Pd subsurface impurities accumulated during the H2O2 synthesis (i. e. O, H) or CO disproportionation. The differential heat of 70 kJ/mol is due to chemisorption on electronically modified Pd domains by the surrounding environment. A second adsorption cycle after desorption of the CO from the first cycle proves the irreversible transformation of the high energy chemisorbing sites (185 kJ/mol). Particularly, in the second cycle, we observe an initial differential heat of chemisorption of 80 kJ/mol. We consider that the energetic distribution of adsorption sites in the second cycle is more representative of CO chemisorption on the Pd sites of this catalyst. Thus, the correlation of the heat with the dimension of the Pd domains and the extent of the electron withdrawing effect exerted by the surrounding environment is more appropriate. In analogy to the very poorly performing Pd0.3%873IWI, intuitively a lower value of differential heat corresponds to a state of high selectivity (see Pd1.8%873 vs Pd1.7%473 after 12 minutes of reaction for instance). A similar correlation is seen for the used Pd0.3%CxN, whose lower energy chemisorbing sites than the bare metal surface account for the relatively higher selectivity observed also after 180 minutes of reaction. A simple mathematical analysis of the selectivity data with respect to the population of CO chemisorbing sites (equation specified in supporting information) indicates an approximate threshold value of CO chemisorption of 70 kJ/mol: below this value, a selective path to H2O2 occurs; above this value, H2O formation is favored. This data analysis also suggests that CO chemisorption below 40 kJ/mol is irrelevant for the H2 and O2 activation. Those values could be used to model selective Pd sites and predict better catalysts for the H2O2 direct synthesis. The evolution of the electronic structure of the Pd NPs upon reaction was studied previously by photoemission spectroscopy [6] for the two higher loaded catalysts Pd1.8%873 and Pd1.7%473 and here reported in supporting information for convenience (Figure S14). Table S2 summarizes the elemental composition of the two catalysts Pd1.8%873 and Pd1.7%473 upon reaction: while the N species are stable in the reactive environment (the N content remains un-
changed), the Pd content is reduced especially after 180 min of reaction. The electronic state of the Pd species is reflected in the chemical shifts observed in the Pd 3d XP spectra. The reactivity profiles of the two samples correlates with the life time of the minor component at 336.9 eV (Figure S14, Table S2), a Pdn+ species [6], which is an interfacial species located at the surface of electronically modified Pd0 NPs (335.5 eV). It is worthy noticing that the similar Pdn+/Pd0NP ratio for the Pd1.8%873 fresh and upon 12 minutes of reaction indicates a similar surface/volume ratio and thus a high stability of the small metal NPs for this sample. However, the formation of large Pd0 NPs by sintering of the initially present small Pd0 NP (335.5 eV) upon reaction is also manifested by the occurrence of bulk-like Pd0 at 335 eV. The instability of other Pd0 NPs through dissolution in liquid phase is also observed as an enhancement of the Pd2+ species (337.9 eV) after 12 minutes of reaction (Figure S14, Table S2), which can redeposit on the carbon surface if surface ligands are available (in Pd1.7%473 more than in Pd1.8%873). The significant change of the catalytic performance of these catalysts synthesized from preformed NPs are most likely triggered by the removal of PVA that leads to the exposure of Pd sites,[5] which is followed by the observed sintering, partial dissolution and deposition of Pd species. The beneficial effect of the protective capping layer is clearly demonstrated by the poor catalytic performance of the Pd0.3%873IWI (Figure S4d). In a previous study we found that the Pdn+ species at the interface are assured by the presence of C and O surface and subsurface impurities. [6] Those surface/subsurface impurities of the Pdn+/Pd0 NP modulate the adsorption of the reactants and the electron transfer from the metal to the adsorbate. Those results are also in agreement with recent mechanistic insights into the direct H2O2 synthesis, suggesting that protection of undercoordinated sites increases the selectivity. [19-20] A plausible mechanism which was described in the literature [13] is that the C and O impurities in the NPs subsurface limit the H uptake to the external catalyst surface and, by preventing the H diffusion into the subsurface the partial hydrogenation of the O=O bond H2O2 is realized. We further consider that the surface species (e.g. PVA, CO2, SO4 2-) might act as ligands for Pd, modifying the electronic structure and/or partially poisoning adsorption sites similarly to the effect of Sn on Pd supported on oxides. [10]
However, the protection against excessive metal exposure and the electronic modification exerted by the C overlayer residual from the PVA is short-lived, its detachment from the Pd surface results in a change of the selectivity for most of the NPs from partial to total O2 hydrogenation. The N species on the carbon support are more effective and can induce the required electronic modification and stabilize the small NPs against the thermodynamically favored Pd agglomeration for longer reaction time. However the chemical nature of the N species is pivotal for exerting this function. The Figure S15 shows the N1s spectra for the Pd1.7%473 and Pd1.8%873 fresh and used. While there is not substantial qualitative and quantitative difference in the N species upon reactive H2O2 atmosphere for both samples, the N1s peak-line shape upon metal immobilization is sig-
7 ACS Paragon Plus Environment
ACS Catalysis
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
nificantly modified for the sample Pd1.8%873 in contrast to the Pd1.7%473 (Figure S1a vs Figure S15). These heterocyclic substitutional N species on the NCNT873 are involved in the chemical bonding with the Pd [13] (the relative abundance of N1 decreases upon immobilization while the N2 increases in Figure S15) and they are particularly beneficial for the stabilization of the small Pd0 NPs (< 2nm) enabling a persistent electron withdrawing effect on the Pd as expressed by the invariant Pdn+/Pd0NP ratio. Acidic and basic O and N moieties such as amide (N2) and lactam (N3) are abundant on Pd1.7%473 were Pd apparently establish a chemical bond through the O atom (as suggested by the similarity of the N1s line shape before (Figure S1a) and after Pd immobilization (Figure S15)). The Pd-O bond is probably not stable against hydrogenolysis caused by the H spill-over from the Pd NPs during the H2O2 synthesis. Thus, on the one hand, there are more NPs sintered and, on the other hand, there are more Pd2+ species likely caused by chelating with the surface functional groups. On CxN, heterocyclic substitutional N species are highly abundant (Figure S1) and can establish a strong chemical bond with a larger number of Pd NPs, which resists the harsh environmental conditions of the H2O2 direct synthesis. Interestingly, we also found that the surface N abundance increases in the CxN indicating the loss of the surface C enriched phase upon reaction, thus providing additional anchoring sites for mobile NPs on the one hand and, partially encapsulating the nanoparticles on the other one. We postulate that the Pd surface/subsurface impurities needed for partial hydrogenation are probably generated/regenerated in situ by reaction with the support. In summary, we have shown that for the direct synthesis of H2O2 from H2 and O2, structurally optimized CxN represents an ideal support for Pd, enabling lowering of its loading and increasing the catalyst lifetime. Strong chemical bonding of Pd with N species and surface protection by carbonaceous impurities are the factors explaining the stable performance of CxN and open up a completely new strategy of catalyst design. To achieve high activity and selectivity for long reaction time, not only the active and selective state of the catalyst needs to be identified but also the dynamic transformation of the catalyst under reaction conditions. Accordingly, catalyst formulations with “sacrificial supports” that under reaction enable the dynamic regeneration of the active and selective state have high potential for enhancing the catalysts lifetime and reduce the metal loading.
ASSOCIATED CONTENT Supporting Information Additional characterization results are summarized in the Supporting Information (PDF). The Supporting Information is available free of charge on the ACS Publications website.
AUTHOR INFORMATION Corresponding Author *
[email protected].
Present Addresses
Page 8 of 18
† Diamond Light Source Ltd., Diamond House, Harwell Science & Innovation, Campus, Didcot, Oxfordshire OX11 0DE. All authors have given approval to the final version of the manuscript.
ACKNOWLEDGMENT The authors thank Gisela Weinberg for Scanning Electron Microscopy investigations. HZB is acknowledged for allocation of synchrotron radiation beam time. This work was carried out in the frame of MAXNET Energy.
REFERENCES (1) Centi, G.; Perathoner, S.; Abate, S.; Direct synthesis of Hydrogen Peroxide: Recent Advances, in Modern heterogeneous oxidation catalysis: Design, reaction and characterization, Edited by N. Mizuno, WILHEY VCH 2009, Verlag Weinheim, Ch. 8, p. 253. (2) Edwards, J. K., Solsona, B.; Ntainjua, E. N.; Carley, A. F.; Herzing, A. A.; Kiely, C. J.; Hutchings, G. J., Science 2009, 323, 1037–1041. (3) Biasi, P.; Canu, P.; Menegazzo, F.; Pinna, F.; Salmi, T. O., Ind. Eng. Chem. Res., 2012, 51, 8883–8890. (4) Pashkova, A.; Dittmeyer, R.; Kaltenborn, N.; Richter, H., Chem. Eng. J. 2010, 165, 924-933. (5) Arrigo, R.; Wrabetz, S.; Schuster, M. E.; Wang, D.; Villa, A.; Rosenthal, D; Girgsdies, F.; Weinberg, G.; Prati, L.; Schlögl, R., Phys. Chem. Chem. Phys. 2012, 14, 10523-32. (6) Arrigo, R.; Schuster, M. E.; Abate, S.; Wrabetz, S.; Amakawa, K.; Teschner, D.; Freni, M.; Centi, G.; Perathoner, S.; Hävecker, M.; Schlögl, R., ChemSusChem 2014, 7, 179–194. (7) Pritchard, J.; Piccinini, M.; Tiruvalam, R.; He, Q.; Dimitratos, N.; Lopez-Sanchez, J. A.; Morgan, D. J.; Carley, A. F.; Edwards, J. K.; Kiely, C. J.; Hutchings, G. J.; Catal. Sci. Technol. 2013, 3, 308–317. (8) Liu, Q. ; Bauer, J. C. ; Schaak, R. E.; Lunsford, J. H., Angew. Chem. Int. Ed. 2008, 47, 6221–6224. (9) Moreno, T.; Serna, J. G.; Plucinski, P.; Sanchez-Montero, M. J. ; Cocero, M. J. , Appl. Catal. A: General 2010, 386, 2833. (10) Freakley, S. J; He, Q.; Harrhy, J. H; Lu, L.; Crole, D. A.; Morgan, D. J.; Ntainjua, E. N.; Edwards, J. K.; Carley, A. F.; Borisevich, A. Y.; Kiely, C. J.; Hutchings, G. J., Science 2016, 351, 6276965-8. (11) Arrigo, R.; Hävecker, M.; Wrabetz, S.; Blume, R.; Lerch, M. ; McGregor, J.; Parrott, E. P. J.; Zeitler, J. A.; Gladden, L. F. ; Knop-Gericke, A.; Schloegl, R.; Su, D. S., J. Am. Chem. Soc. 2010, 132, 9616–9630. (12) Groenevolt, M.; Antonietti, M.; Adv. Mater. 2005, 17, 1789–1792. (13) Arrigo, R.; Schuster, M. E.; Xie, Z.; Yi, Y.; Wowsnick, G.; Sun, L. L.; Hermann, K. E.; Friedrich, M.; Kast, P.; Hävecker, M.; Knop-Gericke, A.; Schlögl, R., ACS Catal 2015, 5 (5), 2740-2753. (14) Zheng, Y.; Jiao, Y. ; Zhu, Y. ; Li, L. H.; Han, Y. ; Chen, Y. ; Du, A. ; Jaroniec, M.; Qiao, S. Z., Nature Communications 5, 3783 doi:10.1038/ncomms4783 (2014). (15) Schuster, M. E.; Haevecker, M.; Arrigo, R.; Blume, R.; Knauer, M.; Ivleva, N. P.; Su, D. S.; Niessner, R.; Schlogl, R., J. Phys. Chem. A 2011, 115, 2568-2580. (16) Zhang, Y.; Pan, Q.; Chai, G.; Liang, M.; Dong, G.; Zhang, Q.; Qiu, J.; Scientific Reports, 2013, 3, DOI: 10.1038/srep01943. (17) Guerrero-Ruiz, A., Topic in Catal. 19 (202) 303. (18) Flores-Camacho, J. M.; Fischer-Wolfarth, J.-H.; Peter, M.; Campbell, C. T.; Schauermann, S.; Freund, H.-J.; Phys. Chem. Chem. Phys. 2011, 13, 16800–16810.
8 ACS Paragon Plus Environment
Page 9 of 18
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 Catalysis
(19) Plauck, A.; Stangland, E. E.; Dumesic, J. A.; Mavrikakisa, M., PNAS 2016, 113, E1973. (20) Wilson, N. M.; Flaherty, D. W., JACS 2016, 138, 574– 586.
9 ACS Paragon Plus Environment
ACS Catalysis
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 18
Table of Contents artwork
10 ACS Paragon Plus Environment
Page 11 of 18
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 Catalysis
TEM image of the pristine CxN and corresponding selected area electron diffractogram (SAED) in the inset (a); 180x180mm (144 x 144 DPI)
ACS Paragon Plus Environment
ACS Catalysis
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
energy electron loss spectra (b). 171x119mm (96 x 96 DPI)
ACS Paragon Plus Environment
Page 12 of 18
Page 13 of 18
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 Catalysis
H2O2 productivity (a) and selectivity (b) for: Pd1.8%873 (black square), Pd1.7%473 (red circle) and Pd0.3%CxN (blue triangle). 56x39mm (300 x 300 DPI)
ACS Paragon Plus Environment
ACS Catalysis
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
H2O2 productivity (a) and selectivity (b) for: Pd1.8%873 (black square), Pd1.7%473 (red circle) and Pd0.3%CxN (blue triangle). 56x39mm (300 x 300 DPI)
ACS Paragon Plus Environment
Page 14 of 18
Page 15 of 18
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 Catalysis
HAADF-STEM image of fresh Pd0.3%CxN (a) and used after 180 minutes of reaction (d); secondary electrons (SE) image of fresh (b) and used after 180 min of reaction (e) and overlapped HAADF-STEM/SE image for fresh (c) and used (f) samples. Green circle in (d) indicates very small nanoparticles; particles highlighted in red are examples of nanoparticles inside the pores of the CxN. 254x168mm (150 x 150 DPI)
ACS Paragon Plus Environment
ACS Catalysis
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
HRTEM of Pd0.3%CxN after 180 minutes of reaction. Note the partial encapsulation of the NPs in the CxN matrix. 722x722mm (72 x 72 DPI)
ACS Paragon Plus Environment
Page 16 of 18
Page 17 of 18
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 Catalysis
Differential heat of chemisorption as function of CO uptake at 313 K for: Pd1.8%873 used after 12 minutes of reaction (black square); Pd1.7%473 used after 12 minutes of reaction (red circle); Pd0.3%873IWI fresh (magenta rhombus); Pd0.3%CxN fresh (pale blue triangle); 1st cycle of CO adsorption on Pd0.3%CxN used after 180 minutes of reaction (green rhombus) and 2nd cycle of CO adsorption after desorption step (blue triangle). 170x120mm (96 x 96 DPI)
ACS Paragon Plus Environment