Fabrication of Glass-Insulated Ultramicrometer to Submicrometer

Subscriber access provided by Kaohsiung Medical University

Article

Fabrication of Glass Insulated Ultra- to Sub-Micron Carbon Fiber Electrodes to Support a Single Nanoparticle and Nanoparticle Ensembles in Electrocatalytic Investigations Cesar A Ortiz-Ledón, and Cynthia G. Zoski Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b02785 • Publication Date (Web): 09 Oct 2018 Downloaded from http://pubs.acs.org on October 11, 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 35 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

Analytical Chemistry

Fabrication of Glass Insulated Ultra- to Sub-Micron Carbon Fiber Electrodes to Support a Single Nanoparticle and Nanoparticle Ensembles in Electrocatalytic Investigations César A. Ortiz-Ledón and Cynthia G. Zoski*

Center for Electrochemistry, Department of Chemistry, The University of Texas at Austin, Texas, 78712, United States

Corresponding Author * C.G. Zoski; Telephone: (512) 475-9698; Email: [email protected]

1 ACS Paragon Plus Environment

Analytical Chemistry 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

ABSTRACT We report a procedure to fabricate glass-insulated carbon fiber ultramicroelectrodes (CF UMEs) with sizes of 10 µm and 1 µm in diameter. CF UMEs of theses sizes are commonly insulated with polymer or epoxy resins, which typically have pinholes and are less stable and difficult to polish. Through judicious choice of polishing materials, the fabrication procedure reported here leads to CF UMEs with a surface geometry that, within experimental error, is close to that of an inlaid disk. We demonstrate how the hardness factor of commonly-used polishing materials may alter the geometry of CF UMEs, in which carbon has a larger hardness factor compared to other metals (i.e., Pt or Au). Atomic force microscopy (AFM) and steady state voltammetry (SSV) were used to characterize the CF UME surface and record the electrochemical response, respectively. These results were compared to theoretical values for an inlaid disk UME. Pt nanoparticle (NP) collision experiments using electrocatalytic amplification were used to deposit an exact number of Pt NPs on a CF UME surface. Surface roughness was found to significantly decrease the sticking of Pt NPs on the CF surface, compared to the theoretical collision frequency. However, decreasing surface roughness through judicious polishing led to good agreement between experimental and theoretical collision frequencies of NPs on a CF UME surface, leading to the fabrication of NP ensembles of UME dimensions (UME-NPE) and a single nanoparticle UME (SNP-UME). These electrodes were used to record and analyze SSVs for the hydrogen evolution reaction (HER) in acidic media.


Page 2 of 35

Page 3 of 35 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


INTRODUCTION Carbon materials have historically been used in electrochemical studies due to their low cost and wide potential window resulting from electrochemical inertness to inner sphere reactions.1,2 Commonly-used carbon materials in electrochemistry include glassy carbon, highly oriented pyrolytic graphite (HOPG), graphene, carbon nanotubes, electron-beam-deposited carbon (eC) and boron doped diamond (BDD).3-7 These carbon materials have become routine in experimental setups where macroelectrodes with dimensions on the order of millimeters to centimeters are used. In experiments requiring electrodes of micron and sub-micron dimensions, carbon fiber (CF) is the carbon-based material of choice due to the availability of diameters of 10-40 µm.8 When sealed in glass, CFs of these sizes are generally referred to as ultramicroelectrodes (UMEs) due to possible diameters < 25μm.8 For example, Wightman fabricated cylindrical CF UMEs by aspirating a fiber (6-10 μm diameter) into a glass capillary and pulling the assembly on a vertical pipette puller to generate a CF wire protruding from the pulled capillary which was sealed at the CF-glass junction and used for detection of neurotransmitters.9,10 Wightman also fabricated inlaid disk CF UMEs by aspirating a CF (10 μm diameter) into a glass capillary, sealing with epoxy resin, and carefully cutting the protruding CF.11,12 More recently, Mauzeroll fabricated CF UMEs for scanning electrochemical microscopy (SECM) by pre-pulling a glass capillary using a horizontal pipette puller, inserting a 7 μm diameter CF into the pre-pulled capillary, pulling the CF and glass assembly together, and sealing with a heating coil. CF UMEs fabricated by this method have the advantage of a thin-glass-insulated CF (i.e., RG  3, where RG refers to insulator/conductor ratio) without the use of epoxy.13 Another CF UME fabrication method involves heat sealing a CF under vacuum in a thick glass capillary, leading to a more robust CF UME of  8 μm diameter that can be



polished with alumina.14 Typically steady state voltammetry (SSV) using an outer sphere redox mediator (e.g., ferrocene methanol) is performed in characterizing the diameter of the CF.8 CF UMEs of nanometer dimensions can also be fabricated, for example by flame15 or electrochemical etching16 of the CF followed by insulating with electrophoretic paint or a polymer film. Due to the fragility of the CF in these nm-sized CF UMEs, mechanical polishing of the surface becomes challenging and often impossible. Flame-etched CFs have been secured in quartz nanopipettes, but the fragile etched portion of the exposed CF of tip diameter  50-200 nm and shaft length from  500-2000 nm was insulated with a polymer coating instead of sealing in glass.17 More recent methods of fabricating nanometer carbon electrodes include pyrolysis of butane inside pulled quartz nanopipettes.18,19 Carbon disk nanoelectrodes (NEs) have also been fabricated by chemical vapor deposition (CVD) of carbon from methane inside a pulled nanopipette and using a focused ion beam (FIB) to mill the exposed surface followed by characterization using SECM.20 In addition to the great effort necessary in fabricating carbon NEs, extreme care must be taken during experiments to avoid serious damage due to electrostatic discharge.20,21 The use of carbon electrodes in supporting metal NPs to quantitatively measure electrocatalytic activity, either of a single NP or an ensemble of NPs, has increased in the past decade. For example, Kucernak electrodeposited single Pt NPs on polymer coated electrochemically-etched carbon NEs with an effective electrode radius ≤ 10 nm and studied inner sphere reactions such as the oxygen reduction reaction (ORR) and the hydrogen oxidation reaction (HOR) where Pt NP sizes ranged from 55 to 3600 nm and 36 to 12500 nm respectively.22,23 Bard’s group used flame etched conical carbon electrodes insulated with electrophoretic paint to detect and study single metal NPs.24 Recently, Mirkin used CVD to


Page 4 of 35

Page 5 of 35 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


fabricate a carbon NE with an effective radius of 3 nm (i.e., determined from the limiting current of a steady-state voltammogram) where he observed hydrogen adsorption/desorption features on Au clusters in 0.1 M HClO4 not observed at larger electrodes.25 Unwin’s group deposited Au NPs on a TEM grid using scanning electrochemical cell microscopy (SECCM) and NP collisions detected through hydrazine oxidation and recorded steady-state voltammograms on isolated Au NPs on the grid which were correlated with TEM characterization of the individual NPs.26 Carbon surfaces are easily prone to organic contamination which can impact electrocatalytic measurements. Amemiya’s group extensively studied the origin and consequences of organic contamination using SECM on HOPG and other graphitic surfaces such as graphene and ultraflat electron beam deposited carbon.27-29 Their results demonstrate that organic contamination decreases electrochemical activity and that ultrahigh pure solutions are needed when using HOPG or any other carbon surface. Amemiya reported organic contamination effects in terms of total organic carbon (i.e., TOC ≤ 4 ppb).30 Discrepancies between theoretical and experimental collision frequencies have been reported for NP collisions at UME surfaces. For example, differences reported by Bard for Pt NP collisions at a 8 µm diameter CF UME sealed in glass compared to that treated with piranha solution were attributed to the state of the CF UME surface.31 Identical Pt NP collision experiments performed on boron doped diamond (BDD) UMEs resulted in blip responses compared to the staircase responses observed by Bard.32 Blip responses may occur when Pt NPs collide on a surface but do not permanently stick, or due to deactivation of adsorbed Pt NP leading to a rapid decrease in the current.33 Compton reported the influence of the surface of their cylindrical CF UME on the poor electrical connection observed with a Pd NP.34 Compton also reported blip responses for Au NP in comparison to staircase responses observed on BDD-



UMEs.35 Other groups using 8 μm diameter CF UME sealed in epoxy reported Pd NP blip responses rather than staircase responses.36 We describe here the fabrication of inlaid disk CF UMEs of sizes 10 μm and 1μm in diameter sealed in borosilicate glass. Typically, the glass/CF seal is not strong and has to be reinforced with epoxy resin.37,38 This weak seal arises from differences in the thermal expansion coefficient between glass and CF.39 The fabrication procedure reported here involves sealing CF in borosilicate glass followed by dedicated polishing materials and procedures, and surface analysis using atomic force microscopy (AFM). The AFM analysis gives information on electrode roughness, which was found to be  9 nm for 10 and 1 µm diameter sizes. The use of these CF UMEs in electrocatalysis investigations was explored first through Pt NP collisions to compare experimental and theoretical frequencies of single events. The resulting Pt UME-NPEs were then used to investigate their use in recording and analyzing SSVs for the hydrogen evolution reaction (HER). We also performed collision experiments to observe a single Pt NP on a CF UME of 1 μm diameter and performed HER on the adsorbed single NP. EXPERIMENTAL SECTION Materials. All chemicals were of analytical grade and used as received. Potassium chloride (KCl), Sulfuric Acid High Purity (H2SO4), Sodium Phosphate Dibasic (Na2HPO4) and Sodium Phosphate Monobasic (NaH2PO4) and Sodium borohydride (NaBH4) were obtained from Fisher Scientific. Ferrocenemethanol 97%, Hydrazine 35% (N2H4), Sodium Perchlorate (NaClO4), Perchloric Acid High purity (HClO4), Hexachloroplatinic acid (H2PtCl6), Citric acid (C6H8O7), Sodium citrate (NaC6H7O7), Ascorbic acid (C6H8O6), and Fumed Silica were obtained from Sigma-Aldrich. Sodium Sulfate (Na2SO4) was purchased from Alfa Aesar. Water (18.2 MΩ cm) from an Integral3 Milli-Q system (Millipore Co., Bedford, MA) with TOC values between 2-3 6 ACS Paragon Plus Environment

Page 6 of 35

Page 7 of 35 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


ppb was used. Diamond paste of grades 1 µm, 0.25 µm, and 0.1 µm were obtained from Ted Pella Inc. Alumina powder of different grades 1 µm, 0.3 µm, and 0.05 µm were obtained from Buehler, LTD. Argon gas was of ultrahigh purity grade. Instrumentation. Electrochemical experiments were performed on a CHI 660B electrochemical workstation and in a CHI 220B Faraday cage. AFM images were recorded with a Bruker Dimension Fastscan microscope using a Peak Force tapping in air mode coupled with Scan Asyst and an Asylum Research MFP-3D AFM. The CF UME was held in a brass homemade AFM holder with a Teflon screw to tighten the UME to avoid any loose movement. The holder-CF UME assembly was screwed onto the AFM stage. DLS measurements were performed using a Malvern Zetasizer Nano S (Malvern Instruments, Malvern UK) as described previously.40 Details and results of these DLS measurements are shown in the Supporting Information. UME Fabrication. Electrodes were prepared individually by heat-sealing a 10 µm diameter CF (AMOCO PERFORMANCE PRODUCTS THORNEL P55), or Au and Pt 10 µm diameter wires (Goodfellow Inc. USA) in a borosilicate glass capillary (OD=2.0 mm; ID=1.16 mm, Sutter Instrument) under vacuum. The capillary was filled with silver epoxy to make connection to a copper wire. The silver epoxy was left for 24 h at 120 ºC to cure. A series of grit paper (400, 600, 800 and 1200) were used to expose the CF. Different procedures for polishing were followed to observe the different effects on the CF surface. In the first polishing procedure, a clean microcloth with one of the γ (gamma) alumina suspensions starting from the coarsest to the finest particle size (i.e. 1 µm > 0.3 µm > 0.05 µm) was used to polish the CF surface after it was exposed using grit paper. Each alumina suspension was prepared by adding alumina powder to water until a white milky color was obtained. In the second procedure, the



CF surface was polished with diamond paste (1 µm, 0.25 µm and 0.1 µm) with each grade applied on a clean microcloth, after exposing the CF surface with the various grit papers. A saturated suspension of 0.007 µm fumed silica (Sigma-Aldrich) was made by adding 5 g of silica to 400 mL of ultrapure water (TOC ≤ 2 ppb). The polishing treatment with silica suspension was applied for 40 seconds. While flushing with ultrapure water on a new microcloth, it was observed that extensive polishing longer than 40 s removed some of the insulator (borosilicate glass) to produce a protruded CF surface. A flame etch procedure was followed in fabricating submicron CF electrodes where a CF protruding from a borosilicate glass capillary was quickly etched by passing through a gas/oxygen flame.15 The etched tip was then checked under an optical microscope to ensure a sharp tip with a diameter on the order of 1μm. The open end of the glass capillary was sealed to avoid any damage to the etched tip, which was pulled deep inside, approximately 1 cm away from the end of the capillary. A heating coil and vacuum were used to seal the fiber into the glass capillary. Long exposure times on the heating coil produces bubbles that can damage the CF on both CF UME sizes used in this work. To avoid this, one needs to continuously check the sealing process and remove the tip as soon as it is sealed. Submicron CFs were exposed following reported methodology. 41,42 In order to facilitate AFM imaging on submicron CFs, the insulating glass was sharpened from approximately 2 mm to 0.2-0.3 mm. Although this sharpening was not necessary for electrochemical studies, the reduced glass insulation made it easier to find the surface when positioning the AFM tip on submicron CF UMEs. For AFM analyses, CF UME were fabricated using the same procedure, but the length was adjusted to 3.2 cm for ease of positioning on the AFM stage.


Page 8 of 35

Page 9 of 35 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 success rate for CF UME fabrication is ≥ 90%. However, the presence of bubbles on the carbon-glass seal will cause the structural integrity to be compromised resulting in high capacitance, resistance behavior, and leakage (Figure S1). Therefore, we strongly recommend discarding any CF UME with bubbles generated during fabrication. The success rate for submicron sized CF UME was ≥ 70 %. The crucial steps for this fabrication procedure are the CF flame-etching and exposure once sealed. For example, one should be able to observe a very sharp flame-etched CF under the microscope to verify a tip size of ~1 µm. Since polishing impacts the final electrode size, exposing the surface must be done carefully. A high impedance meter coupled with careful polishing facilitates fabrication of a CF UME in the nanometer scale range, as demonstrated in White’s group with Pt glass sealed wires.41 AFM Characterization. CF UMEs were characterized using AFM, where height profiles and 3D images were analyzed to observe the effect of the different polishing treatments. Each CF UME was analyzed after each of the γ-alumina, diamond paste, and fumed silica treatments. The AFM tapping mode coupled with Scan Asyst was used. This mode combines the benefits of tapping and contact modes through intermittent contact with the sample. Another important feature of this mode is that it automatically sets the critical parameters for achieving a good image of an inhomogeneous sample. AFM images were acquired using a homemade AFM holder that was screwed into the stage as shown in Figure S2. Two CF UMEs were used and three repetitions of each polishing procedure for each of the CF UMEs were performed. NP Synthesis. Citrate capped NPs were synthesized following a seed-mediated growth procedure for NPs of (32±3) nm diameter.43 Seed particles were synthesized by adding 7 mL of 3.8 mM H2PtCl6 to 90 mL of boiling ultrapure water with stirring. After 1 min, 2.2 mL of a



solution containing 1% sodium citrate and 0.05% citric acid was added followed by a quick addition of 1.1 mL of freshly prepared 0.08% NaBH4 while the solution was stirring. The reaction was stirred and heated for 10 min and then cooled to room temperature. To prepare larger NPs, 1 mL of the previous seed solution was added to 29 mL of ultrapure water followed by addition of 0.090 mL of 0.2 M H2PtCl6 and 0.5 mL of a solution containing 1.25% ascorbic acid and 1% sodium citrate. The temperature of the solution was increased slowly (~ 5 0C/min) to the boiling point, and then refluxed for 1 hr. The NPs were washed by centrifuging at 14,000 rpm using a Sorval microcentrifuge (Thermo Electronic Corp.) for 10 min; the precipitate was re-suspended in ultrapure water to make a stock NP suspension of calculated 7.9 nM. The concentration of this stock NP suspension was determined to be 7.2  0.2 nM using DLS. An aliquot of this suspension (34 µL) was injected into the electrochemical cell containing 25 mL of solution in order to obtain calculated NP concentrations in the 10 pM range. This calculated NP concentration is affected experimentally by NP aggregation, as we showed previously.40 NPs were synthesized to generate a new NP stock solution to perform each new series of experiments. Newly synthesized NPs and generated NP stock solutions were never more than 5 days in age. TEM NP Characterization. NPs were also characterized by Transmission Electron Microscopy (TEM). A 3 µL aliquot of a Pt NP stock solution was drop cast onto a 200 mesh carbon film coated TEM grid (Electron Microscopy Sciences). Electrochemical Experiments. CF UMEs were evaluated using the reversible ferrocenemethanol oxidation mediator in a three-electrode setup with a Pt wire as the counter electrode and Ag/AgCl (KCl sat’d) as the reference electrode. Steady-state voltammograms (SSVs) were recorded in solutions of 1 mM ferrocenemethanol in 0.1 M KCl on CF UMEs polished using the different treatments.


Page 10 of 35

Page 11 of 35 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


Nanoparticle Collision Experiments. NP collision experiments were performed using hydrazine oxidation in a solution containing 15 mM N2H4 + 5 mM PBS (pH 7.00) and setting the CF UME potential at -0.1 V vs. Ag/AgCl for a total time of 100 seconds after NP injection into the electrochemical cell. We showed previously that this solution composition resulted in good agreement between the theoretical and experimental collision frequency of individual nanoparticles undergoing collisions and demonstrated that Pt NP aggregation in this solution had the least impact on NP diffusion coefficient and concentration values which are directly proportional to collision frequency.40 At a potential of -0.1 V, the hydrazine oxidation current on CF is  0 A in comparison to the limiting oxidation current reached on each Pt NP that collides on the CF surface. This behavior reflects the sluggish kinetics on CF in comparison to reversible kinetics for hydrazine oxidation on Pt. An aliquot of Pt NP stock solution (i.e., to give a nominal 10 pM NP concentration in the cell) was injected after a stable hydrazine oxidation background current was achieved. Ar was bubbled into the solution for de-aeration as well as to distribute the injected NP throughout the solution. After 10 s, the Ar bubbler was removed from the solution and positioned above the solution to maintain an Ar blanket. Each NP collision experiment was repeated six times with three different CF UME to give a total of three sets of six independent experiments. A freshly made NP stock solution was used for each set of six independent measurements on a specific UME. Each CF UME was polished prior to a collision experiment. Every experiment was performed by repeating the three polishing procedures (i.e., polishing with γ-alumina suspensions, polishing with diamond paste and ultimately polishing with diamond paste followed by the silica suspension). A NP collision experiment was performed on a Au UME to compare the frequency results with the CF UME. We previously showed good agreement between experimental and theoretical collision frequency on 10 µm



diameter Au UMEs in 15 mM N2H4 + 5 mM PBS (pH 7.00).40 Thus, NP collisions based on hydrazine oxidation on a Au UME serves as a benchmark against which NP collisions on a CF UME can be compared. For single NP experiments, a CF UME of 500 nm radius was used, and the collision experiment was stopped as soon as one NP collision was observed. A data acquisition time of 50 ms was used to detect single NP events in all NP collision experiments. Electrocatalysis at Small NP Ensembles and at a Single NP. The electrocatalytic activity of Pt NPs was evaluated using the HER in acidic solution. In this experiment, NP collisions were observed for 200 seconds after injection on a CF UME polished with diamond paste and fumed silica in a solution with 15 mM N2H4 + 5 mM PBS. The resulting Pt NP modified CF UME was imaged by AFM. SSVs were recorded in 2 mM H2SO4 + 0.1 M Na2SO4 under Ar at 0.1 Vs-1 from 0.0 V to -1.2 V vs. Ag/AgCl. Background subtraction was performed by recording a SSV on a bare CF UME. In recording SSVs at single Pt NPs, solutions of 20 mM HClO4 + 0.2 M NaClO4 were used. This concentration is far below the limit for hydrogen bubble formation as reported in previous studies.44 CF UMEs of a ≈ 500 nm were used for these experiments to decrease the CF background current. RESULTS AND DISCUSSION Surface Characterization of Bare CF UMEs. CF UMEs were characterized using SSV with FcMeOH as the redox mediator and AFM to detect changes in the SSV limiting currents and surface images respectively due to polishing materials consisting of γ-alumina suspensions, diamond pastes, or diamond pastes followed by a silica NP (7 nm diameter) suspension. Figure 1a and 1b show AFM images of the CF UME after polishing with a diamond paste/silica NP suspension and with γ-alumina suspensions respectively. Figure 1c shows the height profiles


Page 12 of 35

Page 13 of 35 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


from the cross sections indicated by the red lines of the AFM images of Figure 1a and 1b and Figure S3a. These height profiles show that the CF of the CF UME polished with γ-alumina suspensions protrudes ~ 210 nm above the surface of the glass insulator, ~ 50 nm for the CF UME polished with diamond paste, and ~ 9 nm for the CF UME polished with diamond paste followed by the silica NP suspension. These results are consistent with hardness factors of 7 Mohs for fumed silica, 8 Mohs for γ-alumina powder, 10 Mohs for diamond paste, 2.5 Mohs for Au, 3.5 Mohs for Pt, (i.e., as reported in a table from Ted Pella, Inc.), and 6.5-7.5 Mohs for carbon fiber.45 Borosilicate glass has a hardness factor between 5-6 Mohs,46 which indicates that it is harder than Au or Pt, but softer than CF, and softer than any of the polishing materials. In terms of electrode hardness, CF > Pt > Au, and in terms of polishing material hardness, diamond paste > γ-alumina > fumed silica. Thus γ-alumina is hard enough to remove the insulating borosilicate glass but not hard enough to remove the CF with the same efficiency, leading to the rounded 210 nm protrusion above the glass surface. In contrast, the hardness factor of diamond paste is large enough to remove the insulating borosilicate glass and the CF efficiently, leading to the sharper 50 nm protrusion above the glass surface. The addition of a final polishing step with fumed silica, slowly removes more of the CF and borosilicate glass to smooth the surface and reduce the CF protrusion to the observed 9 nm above the glass surface. Table S1 compares values for the height of the CF above the insulating glass of CF UME for each of the polishing materials used. In agreement with the AFM results show in Figure 1a-c and Figure S3a, CF UME polished with alumina have a CF height of (200±10) nm above the insulating glass surface, based on six independent experiments using AFM. In contrast, CF UME polished with diamond paste resulted in a smaller CF height of (50±10) nm above the glass insulation, with the smallest



Page 14 of 35

CF height of (9±1) nm occurring after polishing with diamond paste followed by a silica suspension. The effect of the polishing procedure on surface roughness of the CF UMEs was also evaluated from recorded AFM images using AFM Nanoscope Analysis, Bruker Inc. software by selecting the total area of the CF. The software measures the height of the different oscillations on the surface and performs a statistical analysis to generate an average surface roughness over the entire surface. The results are reported in Table S2. On CF UMEs polished with alumina suspensions, surface peaks and valleys oscillated between (32±2) nm, between (12±1) nm for polishing with diamond pastes, and between (5.4 ±0.5) nm for polishing with diamond pastes followed by a silica suspension. These analyses indicate that polishing CF UMEs with diamond pastes followed by a silica suspension can lead to a surface on which NPs greater than 5 nm in diameter can be observed. Figure 1d shows steady state voltammograms recorded on a CF UME polished with alumina suspensions only, with diamond pastes, and diamond pastes/silica suspension, recorded after AFM imaging of the surface. The limiting current increased from 1.37 nA for a CF UME polished with diamond pastes/silica to 1.46 nA when the same CF UME was polished with alumina suspensions, compared to a value of 1.38 nA calculated from the limiting current for an inlaid disk UME47 (1)

𝑖𝑙𝑖𝑚 = ―4𝑛F𝐷𝐶𝑎 where n is the number of electrons, F is Faraday’s constant, D is the diffusion coefficient of

FcMeOH (7.0  10-6 cm2/s), C is the concentration and a is the radius. For truncated geometries, such as the one observed when polishing with alumina suspensions, the equation derived in Oldham’s work for a finite hemicylinder (i.e., to within 10%) may be more appropriate in describing the limiting current. This equation is given by Eq. 248 14 ACS Paragon Plus Environment

Page 15 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(2)

𝑖𝑙𝑖𝑚 = ― (1 ± 0.1)𝑛𝐹𝐷𝐶 2𝜋𝐴 where the area of a hemicylinder is given by

(3)

𝐴 = 𝜋𝑎(2ℎ + 𝑎)

and h represents the height of the hemicylinder. For CF UME polished with alumina, a limiting current of (1.6 ± 0.1) nA for FcMeOH oxidation based on 6 repetitions was observed experimentally, in good agreement with a theoretical current Eq. 2 for a hemicylinder of (1.59 ± 0.08) nA with a = (5.13 ± 0.06) μm and h= (200 ±10) nm, but outside the range of experimental error for a limiting current of (1.38 ± 0.02) nA calculated for an inlaid disk (i.e., Eq. 1) of the same radius. With diamond paste polishing, the experimental limiting current was measured as (1.40 ± 0.02) nA with a = (5.13 ± 0.06) μm, in good agreement with a theoretical limiting current Eq. 1 for an inlaid disk of (1.38 ± 0.02) nA, but outside the range of experimental error for a limiting current of (1.5 ± 0.1) nA calculated for a hemicylinder of a = (5.13 ± 0.06) μm and h = (50 ± 10) nm. The experimental limiting current of a CF UME polished with diamond pastes and silica was (1.39 ± 0.02) nA, in good agreement with a theoretical limiting current Eq. 1 of (1.38 ± 0.02) nA for an inlaid disk of radius a = (5.13 ± 0.06) μm, but again outside the range of experimental error for a limiting current of (1.5 ± 0.1) nA, calculated using Eq. 2 for a hemicylinder where a = (5.13 ± 0.06) μm and h = (9 ± 1) nm. These comparisons indicate that the geometry for this CF UME polished with diamond pastes and silica resembles an inlaid disk and not a hemicylinder. Data is summarized in Table S3. The overlapping of the normalized steady-state voltammograms shown in Figure S3b is indicative of reversible kinetics. This is further confirmed by fitting the SSVs with a theoretical reversible SSV generated from Eq. S1. Details are discussed in the Supporting Information.



The CF UMEs described here are insulated in glass rather than by a surrounding polymer coating or electrophoretic paint as reported for other CF UMEs.11,12,15,16 Moreover, the polishing procedure reported here leads to CF UMEs which have a smooth surface, as determined by AFM, in contrast to those insulated with electrophoretic paint or polymer coatings where surface renewal is achieved by continuously cutting the end of the fiber, often resulting in a rough surface unless a focused ion beam (FIB) is used.21,49 Flame-etched CF UMEs with diameters on the order of 1 μm were also fabricated and analyzed using SSV and AFM, as shown in Figure S4. Figure S4c shows that the CF surface is within  5 nm of the glass insulation, with a roughness of 2.0 ± 0.7 nm when polished with diamond paste/silica. A diameter of 1.3 μm was found from the limiting current of the SSV (Figure S4b), in close agreement with 1.4 μm obtained from the AFM height profile (Figure S4c). Table S4 shows AFM and electrochemical characterization parameters for fabricated submicron CF UMEs and Figures S4-S8 show the data for experiments and electrodes reported in Table S4. In the following sections we demonstrate the use of these glass insulated CF UMEs in Pt NP collision experiments through electrocatalytic amplification and illustrate the use of the resulting Pt NP modified CF UMEs in electrocatalysis. NP Collisions on CF UMEs: Effect of Polishing Materials. The effect of CF UME polishing materials on Pt NP stochastic collisions through electrocatalytic amplification was evaluated using hydrazine oxidation. Hydrazine oxidation is an inner sphere reaction which is favorable on Pt surfaces but not on carbon surfaces where large overpotentials are observed, as shown in Figure S9.31,32 Based on this substantial difference in electrocatalytic behavior, CF surfaces can be used as a conductive substrate for Pt NP collisions. In these experiments, the CF


Page 16 of 35

Page 17 of 35 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


UME is held at a potential where hydrazine oxidation is kinetically slow and the current is negligible, compared to the more facile kinetics on Pt where the current has reached a diffusion limiting value. When a Pt NP collides on the CF surface and sticks, a current step from the baseline value to that corresponding to the limiting current of the Pt NP that sticks on the surface occurs. As Pt NPs continue to collide on the surface and stick, a staircase current response occurs which increases negatively in the anodic direction. Figure 2 shows NP collision experiments where a CF UME was held at -0.1 V vs. Ag/AgCl and a collision time of 100 s after NP injection was used to analyze the arrival of Pt NPs on CF UMEs polished with different materials. From TEM we observed a size distribution of Pt NPs of (34±5) nm diameter as shown in Figure S10. The expected NP collision frequency in our experiments for these Pt NPs was calculated from the flux equation for molecules diffusing towards an inlaid disk UME31 fdif = 4DNPCNPaNA

(4)

where D is the diffusion coefficient of Pt NPs in hydrazine + PBS solution (i.e., measured as (6.74 ± 0.09 10-8 cm2/s using DLS, Table S5), C is the concentration of NP (i.e., measured as 5.1 ± 0.3 pM using DLS, Table S5), a = (5.13 ± 0.06) μm is the radius of the UME as determined in the previous section, and NA is Avogadro’s number. Using these values, a frequency of (0.43 ± 0.03) Hz was calculated, in agreement with our previous report where a frequency of (0.36 ± 0.05) Hz was found.40 Experimental frequencies were calculated from: 𝑓𝑒𝑥𝑝 =

# 𝑜𝑓 𝑐𝑜𝑙𝑙𝑖𝑠𝑖𝑜𝑛 𝑒𝑣𝑒𝑛𝑡𝑠

(5)

𝑡𝑖𝑚𝑒

Figure 2 shows that the current due to NP collisions as a function of time for the CF UME polished with diamond paste/silica shows a step-like behavior which increases negatively as each NP sticks on collision, corresponding to a collision frequency of (0.39±0.03) Hz in good 17 ACS Paragon Plus Environment


agreement with the theoretical frequency (Eq. 4). We attribute this behavior to a CF UME that has a geometry more closely resembling an inlaid disk UME with a comparatively smoother surface on which NPs can collide and stick. This behavior contrasts with the NP collisions observed on CF UME polished using only alumina or diamond paste where significantly fewer collisions are seen ((0.04  0.02) Hz and (0.05  0.02) Hz, respectively) and the CF UME current returns to the background level after about 75 s. For alumina polishing, the CF UME geometry is closer to a hemicylinder with a surface roughness that exceeds the size of the colliding NPs and prevents NPs that do collide from sticking. For diamond paste polishing, the resulting geometry behaves like a disk UME, but has a roughness factor on the order of about half the size of colliding NPs which appears to prevent colliding NPs from sticking. The data is summarized in Table S6; NP collisions on two additional CF UME are shown in Figure S11 in the Supporting Information. Similar behavior based on surface considerations was reported by Bard on 8 μm diameter CF UMEs who also found that the collision frequency increased when treated with piranha solution and attributed this behavior to an increase in surface defects leading to enhanced adsorption of NP, but still well below the predicted collision frequency.31 MacPherson observed blip responses from Pt NP collisions on boron doped diamond compared to the staircase response shown in this work.32 We also compared NP collision behavior on a CF UME to that on a Au UME of comparable size where both UMEs were polished using diamond paste/silica (Figure S12). The Au UME also exhibited staircase current behavior, indicating that the NPs stick on collision with a collision frequency of (0.41±0.02) Hz, in close agreement with that found for the CF UME and the predicted value from Eq. 4.


Page 18 of 35

Page 19 of 35 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


In the following sections, we use AFM imaging to show the NPs adsorbed on CF UME surfaces after NP collision experiments. We also use the resulting CF UME-NPEs in recording SSVs in acid solution for the HER. Finally, we demonstrate the capture of a single NP on a CF UME of nanometer radius, its evaluation through AFM, and the electrocatalytic activity of the single adsorbed NP for the HER. Electrocatalytic Activity of NP Ensembles on CF UMEs. An important aspect of this work is the use of smooth CF surfaces to deposit a known number of Pt NPs through NP collision experiments on which electrocatalytic properties of inner sphere reactions can be investigated. Experiments including AFM surface characterization and HER electrocatalysis in acid media were performed to demonstrate that the fabricated CF UME-NPEs can be used in evaluating electrocatalytic activity. In an electrochemical cell containing 15 mM N2H4 + 5 mM PB (pH 7.00), a small aliquot of a freshly prepared Pt NP stock solution was injected and Pt NP collisions were monitored for 200 s after injection of the NP solution. A SSV was recorded in a different electrochemical cell of comparable size containing 2 mM H2SO4 + 0.1 M Na2SO4 under Ar. Figure 3 shows the sequence of the individual experiments. NP collisions (Figure 3a) were recorded on a freshly polished CF UME surface. A total of six independent experiments made with the same freshly prepared Pt NP stock solution led to an experimental NP collision frequency of 0.37 ± 0.02 Hz. This frequency corresponds to an average of 74 ± 4 NPs that deposit on a 10 µm diameter CF UME and is in reasonable agreement with the predicted frequency (0.43± 0.03 Hz) and experimental frequency (0.39 ± 0.03) Hz in the previous section. These results demonstrate the importance of a smooth CF surface for NP collision experiments. Figure 3b shows a SSV for the HER in acidic media, indicating that Pt NPs adsorbed on the surface are electrochemically active.



Page 20 of 35

Experimental limiting currents were analyzed using an equation developed for randomly spaced NPs on a UME surface50

𝑖L,UME ― NPE =

4𝑛𝐹𝐷𝐶𝑎 1+

(6)

(1 ― 𝑓)𝑅𝑁𝑃 𝜋𝑙𝑛2𝑓𝑎

where n, F, D, C and a were defined earlier, RNP is the radius of the spherical NP, and f is the fraction of active area = NRNP2/a2. The proton diffusion coefficient was measured from the limiting current of a SSV recorded on a 10 μm diameter Pt UME in 2 mM H2SO4 + 0.1 M Na2SO4 and found to be (8.0 ± 0.2  10 -5) cm2 s-1. For f

Fabrication of Glass-Insulated Ultramicrometer to Submicrometer

Recommend Documents