Article pubs.acs.org/Langmuir
Gaining Control over Radiolytic Synthesis of Uniform Sub-3nanometer Palladium Nanoparticles: Use of Aromatic Liquids in the Electron Microscope Patricia Abellan,*,†,‡ Lucas R. Parent,‡,§ Naila Al Hasan,∥ Chiwoo Park,⊥ Ilke Arslan,‡ Ayman M. Karim,# James E. Evans,▽ and Nigel D. Browning‡ †
SuperSTEM Laboratory, SciTech Daresbury Campus, Keckwick Lane, Daresbury WA4 4AD, United Kingdom Fundamental and Computational Sciences Directorate, ∥Institute for Integrated Catalysis, and ▽Environmental Molecular Science LaboratoryPacific Northwest National Laboratory, Post Office Box 999, Richland, Washington 99352, United States § Department of Chemistry & Biochemistry, University of California, San Diego, La Jolla, California 92093, United States ⊥ Department of Industrial and Manufacturing Engineering, Florida State University, Tallahassee, Florida 32306, United States # Department of Chemical Engineering, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061, United States ‡
S Supporting Information *
ABSTRACT: Synthesizing nanomaterials of uniform shape and size is of critical importance to access and manipulate the novel structure−property relationships arising at the nanoscale, such as catalytic activity. In this work, we synthesize Pd nanoparticles with well-controlled size in the sub-3 nm range using scanning transmission electron microscopy (STEM) in combination with an in situ liquid stage. We use an aromatic hydrocarbon (toluene) as a solvent that is very resistant to high-energy electron irradiation, which creates a net reducing environment without the need for additives to scavenge oxidizing radicals. The primary reducing species is molecular hydrogen, which is a widely used reductant in the synthesis of supported metal catalysts. We propose a mechanism of particle formation based on the effect of tri-n-octylphosphine (TOP) on size stabilization, relatively low production of radicals, and autocatalytic reduction of Pd(II) compounds. We combine in situ STEM results with insights from in situ small-angle X-ray scattering (SAXS) from alcohol-based synthesis, having similar reduction potential, in a customized microfluidic device as well as ex situ bulk experiments. This has allowed us to develop a fundamental growth model for the synthesis of size-stabilized Pd nanoparticles and demonstrate the utility of correlating different in situ and ex situ characterization techniques to understand, and ultimately control, metal nanostructure synthesis.
■
and sensing.13,14 Typically, the redox synthesis of particles with well-controlled size involves the introduction of a protective stabilizer during particle formation: surfactants, polymers, dendrimers, and ligands are common additives used.3,15,16 Additionally, factors such as nature and concentration of reducing agent and precursors, nucleation sites, and processing temperature also determine the nucleation and growth kinetics of metallic nanocrystals of uniform size and shape.15,17−19 Methods that exploit radiation chemistry have been previously applied to the synthesis of metallic nanomaterials,6,20−24 including Pd nanostructures.25 Radiolytic synthesis has also been successful in preparing metal-supported clusters for catalytic applications.26−28 These synthesis routes use the
INTRODUCTION The chemical and physical properties of metallic nanomaterials can differ significantly from those of the bulk. These differences arise from the relatively large number of surface atoms per particle volume and from quantum confinement effects on the nanoscale,1−3 and they can be tuned during synthesis.4 For example, the catalytic efficiency of metallic nanoparticles and clusters is known to be strongly particle-size-dependent.3,5,6 True control of particle size during synthesis is, therefore, crucial to achieving optimized performance and to precisely investigate structure−property relationships. For the specific case of Pd, nanoparticles are widely used as catalysts and electrocatalysts,7,8 where high efficiency in various fuel cell types has been demonstrated. Pd is an especially effective catalyst, for instance, of ethanol oxidation in alkaline media9,10 since it is not easily poisoned by carbon monoxide and, hence, is a key material for applications such as hydrogen storage11,12 © 2016 American Chemical Society
Received: November 15, 2015 Revised: January 7, 2016 Published: January 7, 2016 1468
DOI: 10.1021/acs.langmuir.5b04200 Langmuir 2016, 32, 1468−1477
Article
Langmuir
relatively constant. Platinum spacers of 70 nm thickness were built on one of the chips by using focused ion beam (FIB) lithography in a dual-beam FIB/SEM (FEI Quanta). Prior to sample loading, the silicon nitride chips were submitted to plasma cleaning for 1 min in argon atmosphere at 150 mTorr pressure. Loading of the sample in the fluid stage was carried out inside an MBraun glovebox filled with purified argon where the solution was allowed to flow into the cell at 5 μL/min for 15 min. Prior to solution flow, argon was allowed to flow directly from the environment for 10 min to ensure no residual oxygen was present in the lines. The electron beam current values reported have been measured by use of the screen dose meter of the microscope, previously calibrated to give the value at the sample plane by use of the calibration curve shown in Figure S1. Calibration of the screen dose meter current reading of the microscope was done by use of an analytical holder equipped with a Faraday cup near the specimen location at the tip of the holder (Gatan, Inc.).47 An electron current of 5.7 pA and pixel dwell time of 3 μs were used for all STEM data sets, giving rise to values of the electron dose rate of 31 and 116 e−/Å2 per frame for 450K× and 910K× magnification, respectively. Formation kinetics were not accurately detected for magnifications below 450K×. A direct conversion of electron dose rates into radiation chemistry units yields 3.1 × 107 Gy/s (450K×) and 11.5 × 107 Gy/s (910K×), where the stopping power for 300 kV electrons in toluene is 2.352 MeV·cm2/g51 and total frame time was 3.78 s. In situ data sets 1−4 are included as Movies S1, S2. S3, and S4, respectively. The liquid thickness was not measured in situ. Thin liquid layer thicknesses of 100−150 nm were considered for our volume calculation in Figure 4 (see Supporting Information for further details and the effect of thickness uncertainty in the calculations). Precursor Solution Preparation and Nanoparticle Chemical Synthesis. Toluene (99.8%) and recrystallized palladium(II) acetate (99.98%) were purchased from Sigma−Aldrich (St. Louis, MO). Palladium acetate was dissolved in toluene in septum-sealed scintillation vials to maintain an airtight environment and degassed with ultra-high-purity helium (99.999% purity from Oxarc) prior to addition of the capping agent, TOP. The concentration of Pd in toluene was 5 mM for in situ STEM experiments as well as ex situ bulk synthesis and 50 mM for in situ small-angle X-ray scattering (SAXS) experiments. The precursor to capping agent ratio was 1:1.5 Pd/TOP. Storage and addition of TOP (Aldrich, 90%) was carried out in an MBraun glovebox supplied with nitrogen. Hexanol (anhydrous, ≥99%) was also purchased from Sigma−Aldrich (St. Louis, MO) and sparged with ultra-high-purity helium prior to solution preparation. Small-Angle X-ray Scattering. SAXS studies of TOP-capped Pd nanoparticle bulk synthesis were carried out at the Argonne National Laboratory (ANL) Advanced Photon Source (APS) with the pinhole setup at beamlines 12-ID-C and 12-ID-B with incident X-ray energies of 18 and 12 keV, respectively. In situ SAXS data was also collected in a microfluidic reactor with similar results. The reacted solution from the microfluidic reactor (at 50 mM) was diluted with toluene to 5 mM and drop-cast on a lacey carbon film supported on a copper transmission electron microscopy (TEM) grid (Ted Pella, Inc.) to investigate the reaction products by STEM. From the time evolution of particles from SAXS, a linear fit was used to obtain the nucleation rate (arbitrary units/s) as the slope of the number of particles versus time at the early reaction times. The slope was scaled by a factor estimated by taking into account the initial concentration of solution, final average nanoparticle size, and final number of nanoparticles (where full conversion of the precursor after an extended time period is assumed). A step-by-step estimation of the initial nucleation rate from SAXS for similar reactions and details on microreactor fabrication and operation, SAXS data acquisition and analysis, and materials preparation can be found elsewhere.38 Multitarget Particle Tracking Analysis. The multitarget particle tracking approach has been previously applied to analysis of video frames of silver nanoparticles grown via in situ liquid STEM44,52 and has provided insights into the attack mechanisms that occur during degradation of electrolyte solutions in Li ion batteries.53 First, an image segmentation algorithm was applied to successive images in
chemical effects of the absorption of high-energy radiation, typically 60Co γ-rays or 2−20 MeV electrons,29,30 by precursorcontaining liquids to form nanostructures by reproducing a selective reducing/oxidizing environment. Conditions to generate free radicals of strong reducing potential can be found and exploited for different applications.31,32 Some strengths of radiation chemical synthesis are that the nuclei formed are homogeneously distributed in the whole volume, growth rate can be easily controlled, and synthesis can be performed at room temperature.6,33 New characterization methods have been developed to investigate colloidal synthesis and gather real-time information at the solid−liquid interface.34−38 Among them, in situ liquid cells used in combination with (scanning) transmission electron microscopy [(S)TEM] uniquely allow for direct imaging of the formation of nanoparticles with millisecond temporal resolution and subnanometer spatial resolution.39−44 Control of the experimental parameters and reproducibility of the results requires an understanding of the effects of high-energy electrons on the solution.45−49 Thus far, most experiments in the fluid stage involve the use of water as a solvent. In water, the radiation chemical yields are relatively large where highly reactive oxidizing and reducing radicals and species are created in about equal amounts.50 Thus, reproducing net reducing conditions for particle growth involves the addition of substances that convert primary radicals into free reducing radicals (use of OH• scavengers, for instance).20,50 Since the electron beam in the STEM is used as both irradiation source and imaging probe to resolve dynamic processes, larger incident electron doses may be required to achieve high-magnification imaging in the case of small (
