Edge selective growth of MCp2 (M=Fe,Co,Ni) precursors on Pt

Subscriber access provided by YORK UNIV

Article

Edge selective growth of MCp2 (M=Fe,Co,Ni) precursors on Pt nanoparticles in atomic layer deposition: a combined theoretical and experimental study Yanwei Wen, Jiaming Cai, Jie Zhang, Jiaqiang Yang, Lu Shi, Kun Cao, Rong Chen, and Bin Shan Chem. Mater., Just Accepted Manuscript • DOI: 10.1021/acs.chemmater.8b03168 • Publication Date (Web): 07 Dec 2018 Downloaded from http://pubs.acs.org on December 8, 2018

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 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 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.

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 36 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

Chemistry of Materials

Edge selective growth of MCp2 (M=Fe,Co,Ni) precursors on Pt nanoparticles in atomic layer deposition: a combined theoretical and experimental study Yanwei Wen1, #, Jiaming Cai2,#, Jie Zhang1, Jiaqiang Yang1, Lu Shi1, Kun Cao2, Rong Chen2,*, and Bin Shan1,* 1State

Key Laboratory of Materials Processing and Die and Mould Technology and

School of Materials Science and Engineering, Huazhong University of Science and Technology, Wuhan 430074, China 2State

Key Laboratory of Digital Manufacturing Equipment and Technology and

School of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wuhan 430074, China # Y.W. Wen, J.M. Cai contribute equally to this work.

Abstract: Recent experiments about the selective coating of transition metal oxide on Pt nanoparticles have aroused great interests in molecular catalysis for the promotion of both activity and stability. In this work, first-principles calculations combined with micro-kinetics methods are employed to shed light on the edge selective growth mechanism of 3d-transition metal oxide on Pt nanoparticles in atomic layer deposition (ALD) from the metal cyclopentadienyl precursors (MCp2, M=Fe,Co,Ni). The MCp2 decomposition on the surface of Pt nanoparticles exhibits robust preferential growing following the order of edge > (100) > (111), which indicates that edges are naturally selected to be covered and the (111) facets could survive towards the MCp2 precursors. The preferred deposition on edge site is attributed to a more favorable splitting path for the precursors.

On the other hand, competing reactions make the overall reaction rates

of MCp2 precursors on edge sites follow the order of NiCp2 > FeCp2> CoCp2. Moreover, the reaction rate analysis indicates the edge selectivity of MCp2 on Pt nanoparticles is temperature dependent and a high temperature will suppress the selectivity between different sites. FeCp2 could maintain high selectivity in a wide temperature range among the three precursors. The theoretical predictions about the

*

Author to whom correspondence should be addressed. E-mail: [email protected], [email protected]. 1

ACS Paragon Plus Environment

Chemistry of Materials 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

edge selective growth of MCp2 are confirmed by the Fourier transform infrared (FTIR) measurements of CO signals on successive ALD coated Pt nanoparticles. The combination of theoretical and experimental study demonstrates the robust edgeselective growth of MCp2 on Pt nanoparticles, which may open up new avenue for the design of metal-oxide composite catalyst with specific site passivation.

1. Introduction Due to the excellent catalytic performance in extensive reactions, Pt-based nanoparticle with well-dispersed and uniform size has been widely used in fields such as combustion, photocatalysis, fuel cells and exhaust purification [1, 2, 3, 4]. As known, the activity of Pt nanoparticle is closely related to its size. When operated under harsh conditions (high temperature or high pressure), Pt nanoparticles experience severe degradation of the activity due to sintering into larger ones. An effective approach to improve the stability of Pt nanoparticle catalyst is by encapsulating it with metal oxides. Among the techniques of conformal coating, atomic layer deposition (ALD) shows remarkable advantages for its precise control of thickness on the atomic scale. For example, elaborately protective layers of Al2O3 [5], ZnO [6], FeO [7], TiO2 [8], SnO2 [9] have been successfully deposited on Pt nanoparticles by ALD. The modified structures are demonstrated to show enhanced sintering resistance compared with the non-coated counterparts upon thermal aging. However, complete encapsulation may cover the active sites of the surface, which degrades the catalytic performance of Pt nanoparticles. Many works have been done to achieve a porous protection layer to avoid the loss of active sites [5, 7, 10, 11]. In general, special attention should be paid to the growth conditions or post-processing to expose those active sites and the complicated treatment poses great challenges in its practical application. It’s worth noting that recent experiments demonstrated that some metal oxides can be deposited on the desired sites of seed materials by selective ALD. For example, Lu et al. found that precursor trimethyl aluminum (TMA) preferred to grow on the step sites of Pd nanoparticles, while such preferential growth was absent on Pt nanoparticles 2


Page 2 of 36

Page 3 of 36 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


[12, 13, 14]. Cao et al. utilized the difference of the binding strength of Ce(thd)4 on Pt (100) and (111), and achieved the sole deposition of CeOx on Pt nanoparticles (111) facets, naturally leaving the active (100) facet available for CO oxidation [15]. As an active protective layer, transition metal oxides have attracted intensive attentions for its synergistic effect in catalysis. Cyclopentadienyl metal (MCp2) is a typical kind of precursor to grow transition metal oxides by ALD. Singh et al. investigated the growth of nickelocene (NiCp2) and ferrocene (FeCp2) on different substrates and found they preferred to be deposited on oxygen-dissociated active noble metals (Pt, Ir) rather than inert SiO2 and Al2O3. [16] Qin et. al. showed that the low coordinated Pt sites could be blocked selectively by FeOx coating and the interaction between FeOx and Pt nanoparticles largely improved the selective hydrogenation of cinnamaldehyde.[17] In our recent work, only a few ALD cycles of NiOx on Pt nanoparticles led to dramatically improved stability without the degradation of CO oxidation activity. [18] It was proposed that the small amount of NiOx preferentially covered the edge sites of Pt nanoparticles, which hindered the Ostwald ripening of the nanoparticles[19]. However, the selective principle of MCp2 (M=Fe,Co,Ni) precursors on Pt nanoparticle is still open and the selective growth mechanism has not been explored up to now. Clarifying these questions is of great significance to construct diverse interface nanostructures of metaloxide catalyst with both excellent activity and stability. In this work, typical surface structures of Pt nanoparticle are considered to interact with three types of MCp2 precursors. Different dissociation paths of the MCp2 precursors are examined and it is found that the dissociation energy barriers on Pt facets follow a robust order of edge < (100) < (111), suggesting that MCp2 precursors are preferably deposited on edge sites and the (111) facet is the most inert. By comparing the dissociation processes, it is found that the edge opens a favorable splitting path for the precursors, which lowers down the energy barrier.

Moreover, it is found that the

selectivity of the preferential growth is temperature dependent. It is predicted that the selectivity of NiCp2 is sensitive to temperature and the edge selective growth is greatly suppressed at a higher temperature, which is confirmed by the experiment. FeCp2 could keep significant difference of the deposition rate among these three typical sites in a 3



Page 4 of 36

wide temperature range. Our ALD experiments of MOx coated Pt nanoparticles and CO Fourier transform infrared (FTIR) test demonstrate the CO adsorption peak of low coordinated sites decrease firstly, followed by (100) and (111). These results corroborate the preferential growth order predicted by the theoretical study.

Our work

sets up a road map for designing the selective growth of metal oxide on metal facets, which may find numerous applications in the field of catalysis.

2. Method 2.1 Computational model and method The pristine surface of nanoparticles consists of different facets, where Pt atoms in these facets and edges are in different chemical surroundings with respect to coordinated numbers. Fig. 1(a) shows a typical configuration of the Pt nanoparticle, where most of the surface consists of low-indices facets such as (100) and (111). It is notice that puckered atoms chains are formed at the joint of two facets, which is named edge sites as marked yellow. Here we use low-indices facets (100) and (111) to imitate the regular surface of Pt nanoparticles. While the joint atomic chains of (100) and (111) facets are constructed to mimic the edge sites of nanoparticles (see Fig. 1(a)). 4×4 supercells with five atomic layers are used to simulate the MCp2 growth on the Pt (111) and (100) slabs. While a 13.75 Å × 14.04 Å rectangular unit cell with eight atomic layers is used to model the growth of MCp2 on the edge slab (Fig. 1(b)). The top three layers Pt atoms are allowed to relax during calculation.

To eliminate the interaction

between the Pt slab and its periodic images, we create a vacuum distance larger than 17 Å for the supercell geometry. Our calculations are performed by using the first-principles plane-wave pseudopotential formulation [20, 21, 22] as implemented in the Vienna ab-initio Simulation Package (VASP). The exchange-correlation functional is in the form of Perdew-Burke-Ernzerhof (PBE) [23] with the generalized gradient approximation (GGA). The cutoff energy of 400 eV for the plane-wave basis, and k-mesh of 3×3×1 are applied for the three Pt slabs ((111), (100), and edge) to ensure the energy convergence to 1 meV and the residual force acting on each atom is less than 0.05 eV/Å. 4


Page 5 of 36 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 adsorption energy (E_a) of a precursor on the Pt surface is defined by the formula: E_a= E_*ads – E_* - E_MCp2, where E_*ads is total energy of Pt slab with precursor adsorbed, E_* and E_MCp2 are the energies of the Pt slab and precursor molecule, respectively. The decomposition energy (E_d) is calculated by E_d= E_*dec – E_* - E_MCp2, where E_*dec is total energy of the surface with precursor decomposed. Thus E_a represents the binding strength between the precursor and the Pt surface, while E_d estimates the reaction heat of the MCp2 and Pt surface. The climbing image nudged elastic band (CI-NEB) [24] is used to search the minimum energy path (MEP) of the precursor dissociation on the Pt surface and locate the transition state between two local minima. Six intermediate images are interpolated between the adsorbed state (AS) and decomposed state (DS). To evaluate the reaction activity of the precursor on the Pt surface, we propose a microscopic model about deposition process of the precursor, which includes the adsorption, dissociation of the precursors, and the further oxidation of the precursor intermediates as follows: k1   MCp2* MCp2  *  

(1)

( MCp ~ Cp )* MCp2* 

(2)

( MCp ~ Cp )*  O3  MOx*  CO2  H 2O

(3)

k1

k2

k2

Since the oxidation of the ligand -Cp is fast (Equ. (3)) with O3 pulse, the model can be simplified into a model involving the adsorption (Equ. (1)) and dissociation routes (Equ. (2)). It is noticed that the dissociation step (Equ. (2)) is dramatically exothermic, thus the reaction rate of reverse process (k-2) is small and can be safely neglected in our study. The rate expressions of adsorption and dissociation processes are shown:

r1  k1PMCp2*  k 1 MCp * 2

r2  k2 MCp *  k 2 ( MCp ~Cp )* =k2 MCp * 2

2

Where θi implies the coverage of corresponding intermediate. ki and k-i correspond to the rate constants of the forward and reverse reaction in the elementary step i. 5



Page 6 of 36

P(MCp2) denotes the partial pressure of the precursors, which is set as the 0.001 atm according to the experimental reference. Collision theory is employed to describe the forward rate of molecular adsorption by: S

k1 

2 mA k BT

where σ is the surface area of the adsorption site, mA is the mass of the adsorbent, and S is the dimensionless sticking coefficient. The backward rate of molecular desorption is written as: E_a

k-1 =

k BT k B T . e h

While the forward rate of MCp2 dissociation in equation (2) can be expressed as : E

b k T k 2 = B e k BT h

Where Eb is the reaction energy barrier of the corresponding reaction. A steady state approximation about MCp2* is introduced in the micro-kinetics analysis:  MCp * 2

t

=r1  r2 =0  r1  r2 

*

 MCp

* 2



k2  k 1 k1PMCp2

There are three types of species occupying the Pt surface sites, adsorbed MCp2*, free site * and decomposed MCp*. Under a certain ALD reaction condition, the first half reaction of ALD leads to a steady coverage of the product MCp* on the surface, which can be described by a parameter  ( 0    1 ). Thus the sum of the coverage of adsorbed MCp2* and free site *can be written as:

*   MCp  1    a ( 0  a  1 ) * 2

(4)

Based on the discussion above, the overall reaction rate of MCp2 precursors on Pt substrate can be calculated by the following equation:

6


Page 7 of 36 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


r =a

k2 k1 PMCp2 k1 PMCp2 +k1 +k2

(5)

2.2 Experimental sections To verify the theoretical predictions, Pt nanoparticles are firstly deposited on the Al2O3 substrate by ALD. The Al2O3 sphere supports (NanoDur, 99.5%) with the average size of 40 nm and specific surface area of 32~40 m2/g were purchased from Alfa Aesar, which were usually used as the supports for the design and preparation of model catalysts. The Pt ALD were performed in a home-made ALD reactor for surface coating of nanoparticles based on the principle of fluidized bed. The base pressure of ALD reactor was about 1.2 torr. About 200 mg nanoparticles were loaded in the designed powder holder with the top and bottom being sealed by stainless steel screen. The nanoparticles were firstly fluidized by 500 mL/min of 11 vol. % O3 (generated by O3 generator using ultrahigh purity O2 (99.999%)) at the reaction temperature before ALD reaction. The Pt ALD were performed at 200 º C with the Pt precursor of trimethyl(methylcyclopentadienyl)platinum (MeCpPtMe3, 98%, Sigma-Aldrich) and oxygen source of 11% vol. O3 balanced by O2. The Pt precursor was held in a stainless steel bottle, which was kept at 65 º C during Pt ALD. The pulse time and purge time for MeCpPtMe3 were 200 s and 200 s. The pulse pressure of MeCpPtMe3 was about 0.1 torr. O3 was introduced to the reactor with the pulse time and purge time of 200 s and 200 s. The pulse pressure of O3 was about 4 torr. Note that, all Pt/Al2O3 in this paper are prepared with two Pt ALD cycles. Thereafter, the reactor temperature was set to 150 º C for NiOx ALD with the precursors of bis(cyclopentadienyl)nickel (NiCp2, 98%, Strem Chemicals) and 11 vol. % O3 balanced by O2. The pulse time and purge time for NiCp2 and O3 were identical 60 s and 120 s. For FeOx ALD, The pulse time and purge time of FeCp2 and O3 were set as 200s and 300s at 180 º C. The FeCp2 source bottle was heated to 100 º C. CoOx ALD was carried out in a commercial (SUNALE™ R200, Picosun) reactor at 150 º C. CoCp2 was held in a stainless steel bubbler maintained at 100 º C during deposition. The chamber background pressure was controlled at 500 Pa. The entire ALD sequence consisted 1.6 s pulse of CoCp2 and 2 s 7



pulse of O3 with 8 s purge of N2 introduced in between. The crystal structures of samples were characterized by Field Emission Transmission electron microscopy (FTEM, Tecnai G2 30). The in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) spectra of CO adsorption on MOx coated Pt nanoparticles were collected by a wide-band mercury-cadmium-telluride (MCT-A) detector, which was implemented on a Nicolet iS50 Fourier transform infrared (FTIR) spectrometer from Thermo Fisher Scientific. Some details for the sample preparation can be found in our previous work [18]. The samples were transferred to a DRIFTS cell that provides the means for heating and gas exposures. They were first pretreated at 300 °C (from room temperature, 10 °C/min) by 50 sccm of O2 for 30 minutes to remove possible organic species in the ALD process. Then the temperature was cooled down to 30 °C. The background spectra were collected in a flow of 50 sccm of N2 at 30 °C. CO adsorption spectra were collected after treatment by 50 sccm of 1% CO. After the spectra are stable, a flow of 50 sccm of N2 was used to remove physically adsorbed CO and the CO chemical adsorption spectra could be obtained. 3. Result and discussion 3.1 Growth model of MCp2 on Pt nanoparticles The coordination numbers of Pt atom on pristine edge, (100) and (111) are 7, 8 and 9 respectively. The vacancy formation energies of Pt atom on edge, (100) and (111) of the nanoparticles are calculated to be 0.36, 0.59 and 1.20 eV based on our firstprinciples method, which agree with the order of the surface energy [25] and can serve as a good indicator of the instability of Pt atoms. The depositions of MOx (M=Fe,Co,Ni) on these three facets are considered. The initial few ALD cycles of metal oxides deserve special attention because it involves the nucleation of the metal oxide, which plays an important role on the selective growth on Pt nanoparticles. Once the Pt nanoparticle is covered by a complete layer of metal oxide, the following self-limiting reaction would take place on the metal oxide instead of on the Pt facets and selective growth diminishes on the MOx substrate.

As a result, we focus on the initial ALD cycle of MOx growth.

One complete ALD cycle of MOx growth consists of two self-limiting reactions, which are the adsorption/decomposition of MCp2 and oxidation of remaining precursor. The 8


Page 8 of 36

Page 9 of 36 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


schematic diagram of the ALD process is plotted in Fig. 2(a) and the process can be expressed by the equations listed below: MCp2 + Pt →MCp ― Pt + Cp MCp ― Pt + O3→MO ― Pt + CO + H2O First self-limiting reaction is more complex since it involves the adsorption and decomposition of MCp2 with a variety of ligands and intermediate products. The second one refers to an oxidation of the remaining -Cp ligand and such reaction is generally fast and irreversible under experimental conditions, which is not the rate determining step in the whole cycle. Thus the first self-limiting reaction involved the adsorption and decomposition of the precursor is investigated in detail in following sections. Similar approximations have also been applied to study the ALD process of Pt and HfO2 on graphene [26, 27, 28], Ru on ruthenium surfaces [29], metal oxides on oxidized and hydrogen-terminated Si(001) surface [30] and so on [31, 32]. As the MCp2 precursor has two -Cp ligands, it is intuitive that single or double Cp groups can be dissociated during the decomposition process. According to the adsorption orientation (Fig. 2(b)), MCp2 precursor can lie down or stand on the Pt surface, which leads to the subsequent slipping or splitting dissociation pathways in Fig. 2(b). Moreover, the dehydrogenation of Cp [33] is also considered as supplement to the slipping pathway since predehydrogenation is another possible reaction route before the dissociation of MCp- (see Fig. S5 in SI).

3.2 Structural and energetic properties of MCp2 on Pt facets Since the MCp2s (M=Fe, Co, Ni) exhibit similar configurations for the adsorption and decomposition structures on each facet, we take NiCp2 as a representative example for discussion. Fig. 3(a) and 3(b) show the adsorption & decomposition configurations of the precursor NiCp2 on Pt (111) and edge along the slipping path. Those configurations on (100) is not shown for brevity as the configurations are similar. For the adsorption state, it is found NiCp2 lies down on the surface with -Cp ring facing the Pt plane and two ortho-carbon atoms are bonded to two neighboring Pt atoms. The distance between the Pt and C atoms are calculated to be 2.13 Å, 2.15 Å and 2.14 Å on 9



the Pt (100), (111) and edge slabs, respectively. Such short distances imply chemical bonds are formed between NiCp2 and Pt surface. The corresponding E_a of NiCp2 on these facets are −2.42, −2.11 and −2.37 eV, respectively. The charge density difference contour of NiCp2 adsorbed on Pt (100) is showed in Fig. S1. It’s clear that chemical bonds are formed between C and Pt atoms, which contributes to the strong adsorption energies. Thereafter, the top ligand -NiCp is expected to slip to a neighboring 3-fold site of Pt (111) to get a stable decomposition state. Along the splitting path as shown in Fig. 3(c) and (d), the NiCp2 precursor is first adsorbed to stand on the Pt (111) and edge with two -Cp ligands perpendicular to the substrate. The adsorption energies are −1.05, −0.97 and −0.98 eV for the (111), (100) and edge slabs, respectively. We notice E_a of precursors standing on these facets are much weaker than those lying on them. It is reasonable since the inert H atoms of -Cp ligand cannot form chemical bonds with the Pt slabs in the latter case. The nearest H-Pt distance is about 2.86 Å. Together with the charge transfer contour (Fig. S1), it is concluded that the weak van der Waals interactions are responsible for the lower E_a in the standing mode. The following transition state involves a detachment of the two -Cp groups simultaneously (Fig. S2). Two Cp ligands are dissociated until they cover the Pt slab and the residual transition metal atom bonds to the Pt atom at the eventually decomposed state. Table I summarizes the E_a and E_d of MCp2 on the three facets for CoCp2 and FeCp2. It is found that the adsorption energies of CoCp2 and FeCp2 with lying down configurations are close to those standing ones on each facet. Moreover, the binding strength of precursor on different facets exhibits slight difference for CoCp2 or FeCp2. By comparing the decomposition energies of MCp2, it can be seen that the slipping mode is much more stable than the splitting one on each facet. This is reasonable because strong M-Pt chemical bonds are involved in the decomposed state and slipping mode provides a full contact of M and the Pt surface, which gains a larger reaction heat than that of splitting mode. If we compare the E_d of different MCp2 precursors in one column, it is found the reaction heat follows an order of NiCp2 > CoCp2 > FeCp2, which indicates the intrinsic activity of the precursor when interacting with the Pt slabs. If we focus on the E_d on different facets by rows, it is found the (100) slab exhibits similar 10


Page 10 of 36

Page 11 of 36 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


reaction heat with the edge sites. The (111) slab shows much lower reaction heat, which implies its reduced reactivity towards the MCp2 precursor.

3.3 Dissociation energy paths of MCp2 on Pt facets To evaluate the possibility of the dissociation process, the MEP path and energy barriers of MCp2 dissociation on the three Pt facets are calculated by CI-NEB method. Fig. 4(a) gives the free energy diagrams of the initial state (IS), adsorption state (AS), transition state (TS) as well as the decomposition state (DS) for NiCp2 along slipping (right) and splitting paths (left), where the detailed configurations can be found in Fig. S2. Along the slipping path, the energy barriers of NiCp2 dissociation on edge, (100) and (111) are calculated to be 0.49 eV, 0.66 eV and 1.76 eV respectively, while those barriers for splitting mode are 0.36 eV, 1.51 eV and 1.41 eV. It’s clear that the decomposition of NiCp2 on the edge slab in splitting mode (0.36 eV) is the most favorable among the three facets, which indicates the NiCp2 show preferential edge selective growth on Pt nanoparticles. Combined with the two paths, it suggests a preferential dissociation order of the NiCp2 precursor on these facets: edge > (100) > (111). Moreover, the NiCp2 dissociations on the (100) prefers the slipping path according to the energy barriers, while those barriers along these two paths are comparative on (111). The adsorption energies and energy barriers along the DH path are shown in Table S1. The most favorable path would not change and the following calculated reaction rate remains the order of edge > 100 >111. Fig. 4(b) and 4(c) show the free energy diagrams of the IS, AS, TS and DS for CoCp2 and FeCp2 along the two paths. Common features can be found in Fig. 4(a), 4(b) and 4(c). Firstly, we notice the dissociations of CoCp2 and FeCp2 exhibit similar edge selective growth along each path (blue curves) with that of NiCp2. Along the slipping path, the reaction barriers on (100) are comparable to those on edge. It is found the decomposition energies of MCp2 on edge and (100) facets are close to each other as mentioned above, which are more exothermic than those of (111).

While for splitting

mode, the open routes of MCp2 on (100) and (111) plane are similar and the energy barriers are quite high. The transition state involves the dissociation of Cp-M-Cp and 11



Page 12 of 36

the edge site provides a favorable splitting mode without too much repulsion between the precursor and Pt slabs. As a result, the energy barriers are drastically lowered compared with those plane structures (100) and (111). Table I also summarizes the dissociated energy barriers of the precursor on these facets. Firstly, these minimum energy barriers on different facets shows that MCp2 precursors exhibit a robust preferential dissociation order of edge > (100) > (111). To find the origin of the edge selectivity, the optimal dissociation modes are carefully examined on each facet. It is found the precursors prefer to be dissociated along slipping path with much lower energy barriers on (100). While on (111), the slipping modes exhibit comparatively high energy barriers with those splitting ones. Thus, we can conclude that the slipping mode is generally the preferred dissociation route on these planar slabs (100) and (111). On edge slab, the raised Pt atomic chain on edge offers a sharp blade to split the precursor and thus it opens a convenient splitting way for the dissociation of MCp2, which is unavailable on the (100) and (111).

3.4 The reaction rate diagram within independent model The reaction rate is calculated to quantitatively evaluate the activity of MCp2 deposited on these three facets along each path. Besides the energy barrier E_b, the adsorption energy E_a contributes to the final reaction rate by changing the branching ratios of the forward reaction vs the backward desorption.

A large adsorption energy

would suppress the reversed reaction rate (k-1), leading to an altered reaction rate. According to equation (5), it is straight forward to calculate reaction rates along all paths on different facets by taking them independently in the model.

It is noticed that

the reaction rate along splitting path of NiCp2 (4.6*105) is only slightly larger than that along slipping path (4.4*105). To get a clear picture of the contribution of E_a and E_b to the reaction rates along each path, Fig. 5 show the calculated reaction rate diagram by including the energy barriers and adsorption energies. The temperature is set to 150 ℃, which corresponds the ALD growth temperature of the MCp2 in experiments. The warm color indicates the region with high reaction rate, while the cold color represents the region with low reaction rate. The bar gives the order of the reaction rate. The lower 12


Page 13 of 36 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


left corner with red denotes that lower E_a and E_b lead to the higher activity, while weak binding and high energy barrier will yield low activity. A total of nine MCp2 dissociation routes on Pt facets are marked in this diagram. It’s clear that these points are divided into three groups: (111) resides green and blue (upper left region), (100) takes yellow and brown (the lower middle region), while edge sites locates the red (lower left region). The color difference indicates the obvious preferential growth on the edge for the three MCp2 precursors. By choosing the more favorable path, it is found the reaction rates of the precursors on each facet exhibit the order NiCp2 > CoCp2 > FeCp2 within the independent model. In real experiments, the two paths may compete with each other and change the reaction rates on each facet, which would be discussed in the next section. It is noticed the reaction rate roughly correlates with the energy barrier E_b as the contours color grade is distributed along the y-axis. However, the reaction rate of FeCp2 on edge is lower than those of CoCp2 and NiCp2 even if it has a lower energy barrier. This is attributed to its weak adsorption energy on edge site, which illustrate the role of E_a in the reaction rate.

3.5 Micro-kinetics analysis with competing effect Slipping and splitting pathways may compete with each other on the same facet in ALD process. To better describe the competing reaction along two pathways, the equation of the coverage of the available reaction sites (Equ. (4)) should be revised as：

*   A1 + A 2  a *

(

*

* is the coverage of free site,  A1 is the coverage of MCp * adsorbed with 2 *

slipping model and

 A2

*

is the coverage of MCp2* with splitting model). Similarly, we

can derive a revised version of reaction rate of slipping path (r2) and splitting path (r4), with competitive adsorption appropriately considered:

r2  k2 A1* 

ak2 k1 PA (k3 +k4 ) (k1 +k2 )(k3 +k4 )  (k3 +k4 )k1 PA  k3 PA (k1 +k2 ) 13



r4  k4 A 2* 

Page 14 of 36

ak4 k3 PA (k1 +k2 ) (k1 +k2 )(k3 +k4 )  (k3 +k4 )k1 PA  k3 PA (k1 +k2 )

k1, k-1 and k2 are the reaction rate constants of the adsorption, desorption and dissociation process for slipping path, while k3, k-3 and k4 are those constants for splitting path. It is assumed that adsorption of the precursor is random for the two paths, thus we have

k1  k3 

S 2 mA k BT

under a specific condition. Moreover, the

configurations of the two adsorbed species A1* and A2* are considerably different with large adsorption energies and we assume the inter-transformations between the two species are negligible in the reaction network. A competing factor r2/r4 can be defined to indicate which pathway is more favorable,

r2 ak2 k1 PA (k3 +k4 ) k2 (k3 +k4 )  = r4 ak4 k3 PA (k1 +k2 ) k4 (k1 +k2 )

(6)

A large/small value of competing factor suggests the slipping/splitting model is more favorable. The reaction rates of NiCp2 on the three facets are listed in Table 2. The coverage θ*, θA1* and θA2* of NiCp2 are also shown to illustrate the competing adsorption of each species. It can be seen that the value of r2/r4 on edge equals to 1 and those for (100) and (111) are about 105. Those values indicate contribution from the slipping path dominates on (100) and (111) facet, while splitting path and slipping path contributes equally on the edge. The conclusion is in qualitative agreement with the prediction from the independent model. Due to the favorable adsorption energy and low energy barriers on edge, both slipping and splitting paths yield fast transition to the product state and show high reaction rates with order of ~105. Analytically, low energy barriers (0.36 eV and 0.49 eV) yield k2 and k4 >> k-1 and k-3, making Equ. (6) asymptotically approach 1. Along (111) and (100) facets, the overall picture is different. As indicated in Table 2, the high energy barrier along splitting path on (100) hinders the transition of adsorbed species to the product and the residual A2* covers all the reaction sites, which poisons of the surface (free site θ* ~1.7*10-6). Thus the reaction rate of the other path (slipping) is significantly depressed from ~104 (independent case) to 1.6 due to the loss of 14


Page 15 of 36 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


available reaction sites when competing reaction is considered. The (111) facet follows a similar decrease of the reaction rate due to the poisoning of adsorbed species in slipping mode. The case for CoCp2 is similar with NiCp2 and the values of r2/r4 are 1, 1011 and 104 on edge, (100) and (111) facets, respectively. This indicates slipping modes are also dominant on (100) and (111) for CoCp2. The poisoning of adsorbents on each facet is more severe than that of NiCp2, and the total reaction rate on edge, (100) and (111) are 101, 10-8, and 10-14, respectively. FeCp2 is the only case among three precursors investigated that has a r2/r4 value of 10-3 on edge, which means the splitting mode is dominant. The more favorable reaction paths on (100) and (111) are again the slipping modes. The total reaction rates on edge, (100) and (111) are on the order of 104, 10-1, and 10-22, respectively. It is worth noting that the low adsorption energies of FeCp2 imply weak bindings of reactants and a low coverage is kept on the facet. In this case, poisoning will not happen and the competing effect is negligible.

3.6 Selectivity of MCp2 on Pt facets within competing micro-kinetics model With the consideration of competing reactions along the two paths, the overall reaction rates of the MCp2 precursors on the three facets are calculated at 425K in Fig. 6 and the lowest reaction rate of FeCp2 on (111) is on the order of 10-21. To evaluate the selectivity of the preferential growth on each facet, we choose the order difference of the reaction rate as an indicator. The orders of the reaction rates (log(r)) for NiCp2 on edge, (100) and (111) are 5.9, 0.21 and -8.2, respectively. The order difference between edge and (100) is ~106, which indicates that the preferential growth of edge site comparing to (100) facet is prominent. However, the (100) facet still exhibits a relatively high reaction rate, which implies that the deposition on (100) may occur along with the coating of edge sites. While the value on (111) is about 106 orders lower than (100), indicating a slow deposition rate on (111) facet.

The rate orders of FeCp2 are

104 and 10-1 for edge and (100), which is a little smaller than those of NiCp2. The order difference is close to that of NiCp2, indicating a similar selectivity with NiCp2. For CoCp2, the reaction rates on edge and (100) are 101 and 10-8, much lower than those of 15



FeCp2 and NiCp2. Thus the reaction rates of the three precursors on edge and (100) follow the sequence of NiCp2 > FeCp2 > CoCp2. It is noticed the predicted order by competing model is different from that of independent one with the FeCp2 and CoCp2 reversed, which suggests that competing reaction significantly affects the real reaction rates. Each precursor exhibits a rigorous preferential growth order with edge > (100) > (111) and the large order differences imply distinct selectivity in the three facets for the MCp2 precursors. The temperature effect on the selectivity of MCp2 precursors is further investigated. Fig. 7(a), (b) and (c) show the log(r) of NiCp2, CoCp2 and FeCp2 on the different facets at 325K, 425K and 525K, respectively. To make the order difference more clear between edge and (100), we take reckon that the rate order less than 10-10 is inert and set it to zero in Fig. 7. We find the selectivity is sensitive to the increase of temperature for NiCp2. The order difference of the reaction rate for NiCp2 decreases from 8.9 to 5.7 and 3.2 as the temperature increases from 325K to 425K and 525K, which implies the selectivity decreases about three orders per 100K. The CoCp2 exhibits the order difference between the two facets decrease from 9.6 to 7.8 as the temperature increases from 425K to 525K. The order differences between edge and (100) for FeCp2 are 8.0, 5.8 and 4.7 at the three temperatures. These values reveal that the selectivity degrades as the increase of temperature. Especially, the selectivity of NiCp2 is the most sensitive to the change of temperature and the selectivity may vanishes at high temperature. In contrast, the FeCp2 shows a strong temperature resistance on the selectivity since only one order changes when temperature increase to 425K to 525K. The three large values of the order differences at these temperatures also indicate the selectivity of FeCp2 can be well kept in a wide temperature range.

3.7 Experimental validation of the selective growth To confirm the selective growth of MCp2 on different facets, we use the precursors MCp2 and O3 to deposit different cycles of MOx on Pt nanoparticles and measure the CO adsorption on each sample. For NiOx and FeOx ALD, the number of cycles are 0, 1, 2, 4 and 6(FeOx)/8(NiOx), respectively. While the ALD cycles for CoOx are 0, 20, 16


Page 16 of 36

Page 17 of 36 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


50, 75 and 100, respectively.

Ellipsometry is used to measure the thickness and

growth rate of MOx on a planar Pt film on silicon wafer. Then the same ALD process is used on Pt nanoparticles and the thickness is estimated from the growth rate on planar samples. Fig. 8 shows the FTIR signals of CO chemisorption on different ALD cycles of NiOx coated Pt nanoparticles. The three fitting peaks of linear adsorption curves are assigned into CO adsorbed on the

(111) (green), (100) (pink) and low coordinated

sites (blue) according to the previous works.[7, 34] Here we suppose the low coordinated sites to be the edges since the vertex site only take a small piece of the surface area of the nanoparticle. We notice the peak area indicates the remained Pt sites available for CO. It is found that the FTIR signal of edge sites gradually decrease as the ALD cycles of NiOx increase. While the peak area ratio of (111): (100) exhibits a reverse order when further ALD NiOx are applied. The FTIR signals of CoOx and FeOx coated Pt nanoparticles follow a similar trend to that of NiOx (See Fig. S3 and S4). Fig. 9 shows the calculated ratios of exposed Pt sites among the three facets upon different ALD cycles for Ni, Co and Fe oxides. The cycle numbers are transferred to the rough thickness according to the ellipsometry measurement.

As shown in Fig. 9(a), 0.06 nm

thick NiOx on Pt nanoparticles leads to 25% loss of edge sites for CO adsorption. While the intensity of Pt (100) and (111) adsorption does not change much, which implies that a preferential growth of NiOx on edge sites. As the thickness increases to 0.12 nm, the intensity of edge sites almost diminishes due to the full cover of NiOx. Moreover, the ratio of (111):(100) increases to some degree, which indicates that some Pt (100) sites are terminated by NiOx. Further deposition of NiOx cause a successive decrease of the Pt (100) intensity. The cases for CoOx and FeOx on Pt nanoparticles exhibit similar trends as indicated in Fig. 9(b) and 9(c). These experimental results demonstrate a robust selective deposition order of MCp2 on Pt nanoparticles with edge > (100) > (111), which is in good agreement with our theoretical predictions. As (111) facet is much more inert than the edge and (100), it is assumed that the growth of precursor on (111) is negligible at the first few cycles and thus the FTIR peak area on (111) is set as a reference. Fig. 10 gives the peak area ratios of edge:(111) and (100):(111) as the increase of cycle numbers. First of all, the slopes of edge:(111) and 17



(100):(111) roughly indicate the relatively growth rates of MOx on edge and (100) per cycle and it is clearly seen that the rates show the same order of NiCp2 > FeCp2 > CoCp2, which is in good agreement with the calculated reaction rates on both the edge and (100) facets according to the competing reactions model. High temperature ALD growth of NiCp2 is employed to investigate the temperature effect on the selectivity between these facets. Fig. 11 shows the FTIR peaks of pristine Pt nanoparticles, 1 cycle NiOx coated samples at 425K and 475K, respectively. Different from Fig. 11(b) where the signal of (100) is well kept, it is found in Fig.11(c) that the signal of (100) exhibits a significant decline along with the decreases of the peak of edge after higher temperature ALD growth. This implies the selectivity of NiCp2 is obviously depressed at high temperature, which agrees well with our theoretical prediction.

4. Conclusion The adsorption and dissociation of transition metal precursors MCp2 (M=Ni, Co, Fe) on Pt nanoparticles are investigated based on first-principles calculations. The reaction energy barriers show that ALD growth of MCp2 preferentially blocks the edge site rather than (100) and (111) facets at first. Combined with micro-kinetics analysis, the selective growth order of MOx on Pt facets is revealed with edge > (100) > (111). The calculated reaction rates within competing paths imply the activity of the precursors on edge sites follows an order of NiCp2 > FeCp2> CoCp2. Moreover, it is found the selectivity of precursors on different facet is temperature dependent. It is predicted that the selectivity of NiCp2 between edge and (100) would be suppressed at high temperature, while FeCp2 could maintain high selectivity in a wide temperature range. ALD experiments and FTIR measurement confirm the selective growth order of MCp2 on these facets. Our work sheds light on the selective growth of MCp2 on Pt nanoparticles theoretically and provides a fine-controlled, preferable protection on Pt nano-catalysts, with utilization of its maximum active sites at the same time.

18


Page 18 of 36

Page 19 of 36 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


ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website at XXX. Configurations and charge density differences of NiCp2 on Pt (100), configurations of MCp2 adsorption state, transition state and decomposition state on Pt facet along two paths, FTIR of linear CO adsorption spectra on CoOx and FeOx coated Pt nanoparticles, dehydrogenation paths of NiCp2 on Pt surface(PDF).

ACKNOWLEDGEMENTS This work is supported by the National Natural Science Foundation of China (Grants 51871103, 51702106, 51572097, 51575217 and 51835005), the China Postdoctoral Science Foundation (Grant 2017M622433), Fundamental Research Funds for the Central Universities, HUST (2018KFYYXJJ031).

R. Chen acknowledges the

Thousand Young Talents Plan and the Recruitment Program of Global Experts. Calculations were done at the Texas Advanced Computing Center (TACC) at The University of Texas at Austin (http://www.tacc.utexas.edu).

19



Page 20 of 36

Tables Table 1. The adsorption, decomposition energies and energy barriers (eV) of MCp2 (M= Ni, Co, Fe) on Pt (100), (111) and edge slabs along slipping and splitting paths Facet MCp2

Ni

Co

Fe

100

111 split

slip

edge

mode

slip

split

slip

split

E_a

-2.42

-1.05

-2.11

-0.97

-2.37

-0.98

E_d

-5.08

-2.73

-3.45

-1.76

-4.82

-3.54

E_b

0.66

1.51

1.76

1.41

0.49

0.36

E_a

-1.83

-1.78

-1.65

-1.70

-1.72

-1.67

E_d

-4.63

-1.71

-2.71

-0.71

-4.18

-2.83

E_b

0.82

2.71

2.25

2.66

0.60

0.99

E_a

-0.45

-0.27

-0.22

-0.26

-0.30

-0.16

E_d

-3.29

-0.31

-1.26

0.67

-2.80

-1.46

E_b

1.05

2.12

2.60

2.51

0.65

0.27

Table 2 Competition of adsorption and reaction of NiCp2 precursors along slipping and splitting paths (a=0.5) Facet

θ

r

edge

111

100

θA1*

0.031

0.5

1.2*10-5

θA2*

0.001

2.2*10-10

0.5

θ*

0.47

6.7*10-15

1.7*10-6

r2

4.35*105

6.2*10-9

1.6

r4

4.35*105

3.8*10-14

5.7*10-6

1

1.6*105

2.8*105

r2/r4

20


Page 21 of 36 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 Captions

Figure 1. (a) the surface structure of Pt nanoparticle. Low-indices facets (100) and (111) are imitated the regular surface, while joint atomic chains of (100) and (111) mimics the edge sites of nanoparticles, (b) side-views from b-axis of the (100), (111) and edge slabs.

Figure 2. (a) Schematic illustration of initial reaction process of MOx ALD. (b) The possible decomposition pathways of NiCp2 precursor.

Figure 3. Configurations of the MCp2 adsorption state (AS) to decomposition state (DS) on (a) (111), (b) edge along slipping path and on (c) (111), (d) edge along splitting path.

Figure 4. Calculated energy diagrams for precursor (a) NiCp2, (b) CoCp2 and (c) CoCp2 reacted on the Pt(111), (100), and edge along slipping and splitting paths.

Figure 5. Reaction rate contours of MCp2 on Pt(111), Pt(100), and edge sites with respect to the adsorption energies (E_a) and energy barriers (E_b) of the precursors on different facet.

Figure 6. The order of reaction rate of MCp2 (M=Fe, Co, Ni) on (100), (111) and edge sites.

Figure 7. The temperature effect on the selectivity of MCp2 precursors (a)Ni, (b) Co (c) Fe, on (100), (111) and edge sites.

Figure 8. The fitting curves of linear CO adsorption spectra on NiOx coated Pt nanoparticles with different ALD cycles (a) 1, (b) 2, (c) 4, and (d) 8.

Figure 9. The remained ratio of the edge, (100) and (111) sites upon ALD coating 21



thickness of (a) NiOx, (b) CoOx, (c) FeOx on the Pt nanoparticles from the FTIR measurement.

Figure 10. Peak area ratios of edge:(111) and (100):(111) for NiOx, CoOx and FeOx on Pt nanoparticles upon ALD cycles.

Figure 11. FTIR peaks of (a) pristine, (b) 1cycle NiOx at 425K and (c) 1cyle NiOx at 475K coated Pt nanoparticles.

22


Page 22 of 36

Page 23 of 36 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


Figures

Figure 1

23



Figure 2

24


Page 24 of 36

Page 25 of 36 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 3

25


Chemistry of Materials 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 26


Page 26 of 36

Page 27 of 36 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 4

27



Figure 5

28


Page 28 of 36

Page 29 of 36 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 6

Figure 7

29




Page 30 of 36

Page 31 of 36 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 8

Figure 9

31




Page 32 of 36

Page 33 of 36 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 10

Figure 11

References 1 King, J. S.; Wittstock, A.; Biener, J.; Kucheyev, S. O.; Wang, Y. M.; Baumann, T. F.; Giri, S. K.; Hamza, A. V.; Baeumer, M.; Bent, S. F. Ultralow Loading Pt Nanocatalysts Prepared by Atomic 33



Layer Deposition on Carbon Aerogels. Nano Lett. 2008, 8, 2405–2409. 2 Nie, Y.; Li, L.; Wei, Z. Recent advancements in Pt and Pt-free catalysts for oxygen reduction reaction. Chem. Soc. Rev. 2015, 44, 2168–2201. 3 Li, X.; Bi, W.; Zhang, L.; Tao, S.; Chu, W.; Zhang, Q.; Luo, Y.; Wu, C.; Xie, Y. Single-Atom Pt as Co-Catalyst for Enhanced Photocatalytic H2 Evolution. Adv. Mater. 2016, 28, 2427–2431. 4 Antolini, E. Structural Parameters of Supported Fuel Cell Catalysts: the Effect of Particle Size, Inter-particle Distance and Metal Loading on Catalytic Activity and Fuel Cell Performance. Appl. Catal. B: Environ. 2016, 181, 298–313. 5 Liang, X.; Li, J.; Yu, M.; McMurray, C. N.; Falconer, J. L.; Weimer, A. W. Stabilization of Supported Metal Nanoparticles Using an Ultrathin Porous Shell. ACS Catal. 2011, 1, 1162–1165. 6 Lei, Y.; Lee, S.; Low, K.-B.; Marshall, C. L.; Elam, J. W. Combining Electronic and Geometric Effects of ZnO-Promoted Pt Nanocatalysts for Aqueous Phase Reforming of 1-Propanol. ACS Catal. 2016, 6, 3457–3460. 7 Hu, Q.; Wang, S.; Gao, Z.; Li, Y.; Zhang, Q.; Xiang, Q.; Qin, Y. The Precise Decoration of Pt Nanoparticles with Fe Oxide by Atomic Layer Deposition for the Selective Hydrogenation of Cinnamaldehyde. Appl. Catal. B: Environ. 2017, 218, 591–599. 8 Celebioglu, A.; Ranjith, K. S.; Eren, H.; Biyikli, N.; Uyar, T. Surface Decoration of Pt Nanoparticles via ALD with TiO2 Protective Layer on Polymeric Nanofibers as Flexible and Reusable Heterogeneous Nanocatalysts. Sci. Rep. 2017, 7, 13401. 9 Marichy, C.; Ercolano, G.; Caputo, G.; Willinger, M. G.; Jones, D.; Rozière, J.; Pinna, N.; Cavaliere, S. ALD SnO2 Protective Decoration Enhances the Durability of a Pt Based Electrocatalyst. J. Mater. Chem. A 2016, 4, 969-975. 10 Yi, H.; Du, H.; Hu, Y.; Yan, H.; Jiang, H.-L.; Lu, J. Precisely Controlled Porous Alumina Overcoating on Pd Catalyst by Atomic Layer Deposition: Enhanced Selectivity and Durability in Hydrogenation of 1,3-Butadiene. ACS Catal. 2015, 5, 2735−2739. 11 Lu, J.; Liu, B.; Greeley, J. P.; Feng, Z.; Libera, J. A.; Lei, Y.; Bedzyk, M. J.; Stair, P. C.; Elam, J. W. Porous Alumina Protective Coatings on Palladium Nanoparticles by Self-Poisoned Atomic Layer Deposition. Chem. Mater. 2012, 24, 2047−2055. 12 Lu, J.; Fu, B.; Kung, M. C.; Xiao, G.; Elam, J. W.; Kung, H. H.; Stair, P. C. Coking- and Sintering-Resistant Palladium Catalysts Achieved Through Atomic Layer Deposition. Science 2012, 335, 1205−1208. 13 Feng, H.; Lu, J.; Stair, P. C.; Elam, J. W. Alumina Over-coating on Pd Nanoparticle Catalysts by Atomic Layer Deposition: Enhanced Stability and Reactivity. Catal. Lett. 2011, 141, 512–517. 14 Lu, J.; Liu, B.; Guisinger, N. P.; Stair, P. C.; Greeley, J. P.; Elam, J. W. First-Principles Predictions and in Situ Experimental Validation of Alumina Atomic Layer Deposition on Metal Surfaces. Chem. Mater. 2014, 26, 6752–6761. 15 Cao, K.; Shi, L.; Gong, M.; Cai, J.; Liu, X.; Chu, S.; Lang, Y.; Shan, B.; Chen, R. Nanofence Stabilized Platinum Nanoparticles Catalyst via Facet-Selective Atomic Layer Deposition. Small 34


Page 34 of 36

Page 35 of 36 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


2017, 13, 1700648. 16 Singh, J. A.; Thissen, N. F.W.; Kim, W.-H.; Johnson, H.; Kessels, W. M.M.; Bol, A. A.; Bent, S. F.; Mackus, A. J.M. Area-Selective Atomic Layer Deposition of Metal Oxides on Noble Metals through Catalytic Oxygen Activation. Chem. Mater. 2018, 30, 663−670. 17 Hua, Q.M.; Wang, S.; Gao, Z.; Li, Y.Q.; Zhang, Q.; Xiang, Q.; Qin, Y. The precise decoration of Pt nanoparticles with Fe oxide by atomic layer deposition for the selective hydrogenation of cinnamaldehyde. Appl. Catal. B: Environ. 2017, 218, 591–599. 18 Cai, J.; Zhang, J.; Cao, K.; Gong, M.; Lang, Y.; Liu, X.; Chu, S.; Shan, B.; Chen, R. Selective Passivation of Pt Nanoparticles with Enhanced Sintering Resistance and Activity toward CO Oxidation via Atomic Layer Deposition. ACS Appl. Nano Mater. 2018, 1, 522−530. 19 Hansen, T. W.; DeLaRiva, A. T.; Challa, S. R.; Datye, A. K. Sintering of Catalytic Nanoparticles: Particle Migration or Ostwald Ripening? Acc. Chem. Res. 2013, 46, 1720–1730. 20 Kresse, G.; Hafner, J. Ab initio Molecular Dynamics for Liquid Metals. Phys. Rev. B 1993, 47, 558. 21 Kresse, G.; Hafner, J. Ab initio Molecular-Dynamics Simulation of the Liquid-Metal– Amorphous-Semiconductor Transition in Germanium. Phys. Rev. B 1994, 49, 14251. 22 Kresse, G.; Furthmüller, J. Efficiency of ab-initio Total Energy Calculations for Metals and Semiconductors Using a Plane-Wave Basis Set. Comput. Mater. Sci. 1996, 6, 15–50. 23 Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865. 24 Henkelman, G.; Uberuaga, B. P.; Jónsson, H. A Climbing Image Nudged Elastic Band Method for Finding Saddle Points and Minimum Energy Paths. J. Chem. Phys. 2000, 113, 9901–9904. 25 Vitos, L.; Ruban, A.V.; Skriver, H.L.; Kollár. J. The surface energy of metals. Surf. Sci. 1998, 411,186-202. 26 Karasulu, B.; Vervuurt, R. H. J.; Kessels, W. M. M.; Bol, A. A. Continuous and Ultrathin Platinum Films on Graphene Using Atomic Layer Deposition: a Combined Computational and Experimental Study. Nanoscale 2016, 8, 19829-19845. 27 Kim, K.; Lee, H.-B.-R.; Johnson, R. W.; Tanskanen, J. T.; Liu, N.; Kim, M.-G.; Pang, C.; Ahn, C.; Bent, S. F.; Bao, Z. Selective Metal Deposition at Graphene Line Defects by Atomic Layer Deposition. Nat. Commun. 2014, 5, 4781. 28 Oh, Il-K.; Tanskanen, J.; Jung, H.; Kim, K., Lee, M. J.; Lee, Z.; Lee, S.-K.; Ahn, J.-H.; Lee, C. W.; Kim, K.; Kim, H.; Lee, H.-B.-R. Nucleation and Growth of the HfO2 Dielectric Layer for Graphene-Based Devices. Chem. Mater. 2015, 27, 5868−5877. 29 Phung, Q. M.; Pourtois, G.; Swerts, J.; Pierloot, K.; Delabie, A. Atomic Layer Deposition of Ruthenium on Ruthenium Surfaces: A Theoretical Study. J. Phys. Chem. C 2015, 119, 6592−6603. 30 Longo, R. C.; McDonnell, S.; Dick, D.; Wallace, R. M.; Chabal, Y. J.; Owen, J. H. G.; Ballard, J. B.; Randall, J. N.; Cho, K. Selectivity of Metal Oxide Atomic Layer Deposition on Hydrogen Terminated and Oxidized Si(001)-(2×1) Surface. J. Vac. Sci. Technol. B 2014, 32, 03D112. 35



31 Elliott, S. D. Atomic-scale Simulation of ALD Chemistry. Semicond. Sci. Technol. 2012, 27, 074008. 32 Elliott, S. D.; Dey, G.; Maimaiti, Y.; Ablat, H.; Filatova, E. A.; Fomengia, G. N. Modeling Mechanism and Growth Reactions for New Nanofabrication Processes by Atomic Layer Deposition. Adv. Mater. 2016, 28, 5367–5380. 33 Aaltonen, T.; Ritala, M.; Sajavaara, T.; Keinonen, J.; Leskelä, M. Atomic Layer Deposition of Platinum Thin Films. Chem. Mater. 2012, 24, 1752-1761 34 Hoxha, F.; Schimmoeller, B.; Cakl, Z.; Urakawa, A.; Mallat, T.; Pratsinis, S. E.; Baiker, A. Influence of Support Acid–base Properties on the Platinum-Catalyzed Enantioselective Hydrogenation of Activated Ketones. J. Catal. 2010, 271, 115–124. TOC

36


Page 36 of 36

Edge selective growth of MCp2 (M=Fe,Co,Ni) precursors on Pt

Recommend Documents