Surface Fouling of Ultrananocrystalline Diamond Microelectrodes

Subscriber access provided by Kaohsiung Medical University

Article

Surface Fouling of Ultrananocrystalline Diamond Microelectrodes During Dopamine Detection: Improving Lifetime Via Electrochemical Cycling An-Yi Chang, Gaurab Dutta, Shabnam Siddiqui, and Prabhu Arumugam ACS Chem. Neurosci., Just Accepted Manuscript • DOI: 10.1021/acschemneuro.8b00257 • Publication Date (Web): 04 Oct 2018 Downloaded from http://pubs.acs.org on October 14, 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 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Chemical Neuroscience

Surface Fouling of Ultrananocrystalline Diamond Microelectrodes During Dopamine Detection: Improving Lifetime Via Electrochemical Cycling An-Yi Chang, Gaurab Dutta, Shabnam Siddiqui and Prabhu U. Arumugam*

Institute for Micromanufacturing, 911 Hergot Ave, Louisiana Tech University, Ruston, Louisiana 71272, USA *Corresponding author. Tel: 318 257 5122. E-mail: [email protected]

ACS Paragon Plus Environment

1

ACS Chemical Neuroscience 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 In this work, we report the electrochemical response of a boron-doped ultrananocrystalline diamond (BDUNCD) microelectrode during long-term dopamine (DA) detection. Specifically, changes to its electrochemical activity and electroactive area due to DA-by products and surface oxidation are studied via scanning electron microscopy, energy dispersive spectroscopy, electrochemical impedance spectroscopy and silver deposition imaging (SDI). The fouling studies with amperometry (AM) and fast scan cyclic voltammetry (FSCV) methods suggest that the microelectrodes are heavily fouled due to poor DA-dopamine-o-quinone cyclization rates followed by a combination of polymer formation and major changes in their surface chemistry. SDI data confirms the presence of the insulating polymer with sparsely-distributed tiny electroactive regions. This resulted in severely distorted DA signals and a 90% loss in signal starting as early as 3 hrs for AM and a 56% loss at 6.5 hrs for FSCV. This underscores the need for cleaning of the fouled microelectrodes if they have to be used long-term. Out of the three in vivo suitable electrochemical cycling cleaning waveforms investigated, the standard waveform (−0.4 V to +1.0 V) provides the best cleaned surface with a fully retained voltammogram shape, no hysteresis, no DA signal loss (a 90±0.72 nA increase) and the smallest charge transfer resistance value of 0.4±0.02 MΩ even after 6.5 hrs of monitoring. Most importantly, this is the same waveform that is widely used for in vivo detection with carbon fiber microelectrodes. Future work to test these microelectrodes for more than 24 hrs of DA detection is anticipated.

Keywords: Ultrananocrystalline diamond, fouling, electrochemical, cleaning, dopamine, sensitivity


Page 2 of 38

Page 3 of 38 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


1.

INTRODUCTION

Chronic monitoring of extracellular neurochemicals is critical to the understanding of several brain disorders.1,2 Studies have shown that abnormal levels of neurochemicals (e.g. dopamine, DA, serotonin, glutamate, adenosine, GABA) are linked to Parkinson’s3,4, epilepsy5,6, Alzheimer’s7-9 and many other brain disorders. Recently, treatment methods such as Deep Brain Stimulation (DBS) have emerged as a successful alternative10,11 to medications and standard therapeutic interventions.10-14 Clinical DBS treatment is an iterative process in which stimulation parameters such as stimulus frequency, amplitude, and pulse duration are controlled in an openloop configuration to regulate neurochemical release. But to maximize benefits and reduce side effects a closed-loop approach that applies neurochemical feedback to guide stimulation parameters is preferred.15 Electrochemical (EC) techniques such as amperometry (AM) and fast scan cyclic voltammetry (FSCV) are routinely used to directly measure changes in neurochemical levels.16-18 Such neurochemical measurements are usually performed rapidly with good sensitivity and selectivity with carbon microelectrodes.19, 20 The carbon fiber microelectrode (CFM) with its small size (~510 µm diameter) is the current gold standard electrode for neurochemical sensing.21,22 When combined with extended-scan FSCV (>+1.2 V), detection limits in the nanomolar range are obtained for several important neurochemicals.23-25 Among the electroactive neurochemicals, DA plays a critical role in the central nervous, system. The seminal work of Ralph Adams demonstrated DA detection in vitro and in vivo by electrochemical methods.26 For example, FSCV was successfully used to detect DA in the brain of an anesthetized rat selectively27. Unfortunately, the DA oxidation on carbon electrodes fouls its surface, which results in a



significant reduction of its oxidation current (i.e. detection signal). For example, a 50% fouling was observed at the CFM within 2 hrs of AM28 or FSCV detection.29 A 35% reduction in sensitivity was observed at hydrogenated conical-tip carbon electrodes.30 Another disadvantage of CFM is its relatively small “faradaic electrochemical potential window” that easily electrolyzes water, and thus prevents the detection of chemicals that exhibit relatively higher oxidation potentials (e.g. adenosine, +1.4 V).31 The fouling (rate) of the electrodes is generally dependent on the chemical concentration, the applied voltage or potential at the working electrode, and the time duration of the applied voltage, which is the detection or monitoring time of the chemical. The CFMs can be renewed by extended-scan FSCV treatments with increased sensitivity32,33 but was accompanied by the corresponding drawback of a loss of electrode surface due to chemical etching.34 Some of the other approaches that have been utilised to overcome fouling are the application of novel waveforms, pretreatments (laser, flame, electrochemical) and/or tailoring of surface functional groups.35-38 Fouling can severely diminish the electrode’s response by passivating the electroactive sites.39 This limits its useful lifetime to only few hours in many cases.28-30 Advanced carbon materials28,37,40-44 and their composites are utilized to mitigate some of the drawbacks of CFMs. It is well-known in the literature that on carbon electrodes, the initial stages of DA fouling involve 2-electron quasireversability oxidation of DA into dopamine-o-quinone (DOQ). A higher pH allows a faster DOQ desorption rate producing ample amounts of nonprotonated DOQ and thereby facilitating DA/DOQ cyclization at the electrode interface.28 A favorable cyclization with a faster DOQ desorption rate should yield a higher DA sensitivity through an increased availability of DA adsorption sites resulting in an enhanced DA adsorption.45 In contrast, in less favorable environments, the cyclization reaction proceeds to a


Page 4 of 38

Page 5 of 38 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


surface-adherent passivation film such as melanin28, 46-48 that is widely responsible for electrode fouling.28 Among the advanced carbon electrodes, boron-doped diamond (BDD) thin films that are chemically stable with a wide faradaic electrochemical potential window, low background currents, excellent dimensional stability and low concentrations of polar surface functional groups with less surface passivation and conductivity loss are demonstrating great promise for neurochemical

monitoring.49-54

Boron-doped

ultrananocrystalline

diamond

(BDUNCD)

possesses a higher resistance towards fouling than other forms of carbon electrodes.55 Recent work in our group and others showed the basic BDUNCD microelectrode characteristics such as their unmatched dimensional stability after several million FSCV cycles and their in vivo capability.34, 56 To date, these electrodes have been used for neurochemical recordings in vitro and in vivo.56 However, there are very few electrode studies that were targeted towards chronic (>4 hrs) detection that is challenging because current electrodes are highly susceptible to surface fouling. Physical treatments such as mechanical polishing, vacuum heat treatment, laser activation, flame etching have shown increased analyte adsorption and electron-transfer kinetics at carbon electrodes but the main disadvantage is that they cannot be used in

in vivo

experiments.33 Another treatment that has proven to be useful is electrochemical cycling (ECC) or etching that exposes the electrode surface to anodic (up to +1.4 V) and cathodic (up to −0.6 V) potentials with different voltage or potential waveforms and/or scan rates.57-59 The main advantages of ECC are its ability to etch the carbon surface oxidatively and to continuously regenerate a fresh surface and prevent electrode fouling, which made the CFMs remain active for long time periods.33 This makes ECC a very useful cleaning tool for the detection of



catecholamines since the state of the surface can be restored by a simple electrochemical procedure, a key requirement for chronic in vivo monitoring. CFMs are routinely chemically etched by cycling between −0.5 V and +1.3 V (this range is called: “extended waveforms”) in order to etch and alter the outermost surface and then expose new electroactivate sites and to retain their chemical sensitivity. However, similar studies to understand the nature of BDD electrode fouling or the effect of ECC on cleaning and maintaining sensitivity have not been performed thus far. Thus, the motivation and the goal of the current research is to develop 250-µm diameter BDUNCD microelectrodes (Figure 1A,B) that can detect neurochemicals chronically (>4 hrs) with minimal loss of electrode sensitivity (i.e. an ability to generate large distinct redox current peaks for a given neurochemical) and electron-transfer kinetics (i.e. an ability to generate the redox current peaks at lower applied voltages) using ECC cleaning method60 with waveforms56 that are in vivo suitable. 2.

RESULTS AND DISCUSSION

Surface Fouling Studies of BDUNCD Microelectrodes using AM Method. BDUNCD surface consists of a mixture of sp3 diamond grains and unstructured sp3 and sp2 amorphous carbon grain boundaries with rich surface functional groups61,62 that are known to greatly affect the rate of electron transfer processes.63 BDUNCD films are generally grown in a mixture of methane and hydrogen with a CH4/H2 ratio of ~5%50 providing enhanced adsorption sites through π−π interactions that promote adsorption of DA and their oxidation products which enhances the rate of electrode surface fouling.50 AM that has been applied successfully to study the nature of electrode fouling28 is employed here for the first time to develop a basic understanding of the BDUNCD surface fouling. During this study an electrode potential of +0.3


Page 6 of 38

Page 7 of 38 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


V, which is 50 mV higher than the DA oxidation potential, was applied for 4 hrs in a 100 µM DA solution prepared in 1X PBS buffer solution in “ no flow” condition using a static cell. Intermediate FSCV data (−0.4 V to +1.0 V, 400 V s−1, 10 Hz) was measured at 0th, 1st, 3rd and 4th hr over the 4 hrs period of intentional fouling of the microelectrode (Figures 2A-E). Table S1 (refer the supporting information) shows the FSCV parameters and values recorded during the AM study. The −0.4 V to +1.0 V “standard” waveform was chosen because the previous studies suggest that a −0.4 V cathodic/resting potential between scans is optimal in terms of attaining maximum DA sensitivity.19,64 A +1.0 V anodic limit provides stability in DA signal gain and faradic response together with an improved temporal resolution and S/N ratio. Also, the standard waveform is widely used to detect DA with CFMs.19 After 1st hr of AM (Figure 2C), the subsequent FSCV forward “anodic” DA signal is 1.2±0.12 µA (brown curve), which indicates a favorable DA/DOQ cyclization. An anodic peak shift of greater than 200 mV from 628±12.56 mV at 0th hr to 832±16.64 mV at 1st hr is in agreement with the presence of DA/DOQ couple on the electrode surface,45 which resulted in a higher DA sensitivity indicating desirable DOQ desorption rates as observed in similar studies.45,65 At the 3rd and 4th hrs of AM (Figures 2D,E), the FSCV signal is severely distorted with large hysteresis and no clear anodic peaks. Overall, a 90% decrease in the DA oxidation signal accompanied by a reduction in cathodic signal was observed after 4 hrs. This is due to the formation of passivating polymer films. Thus, with time, there is a less favorable environment for the cyclization reaction that led to the formation of the polymer such as melanin.46-48 One other possible cause for the reduction in anodic signal and the associated hysteresis could be due to unfavourable surface chemistry, which is sparsely studied in the BDUNCD literature.66 We also studied the fouling rate in AM method under flow conditions using microfluidics. For this, the microelectrode was first



integrated into a custom microfluidic cell (Figure 4A, details in the supporting information), where the DA solution was allowed to flow over the microelectrode surface. The microelectrode was intentionally fouled by flowing a 100 µM DA solution at 0.2 mL/min in the microfluidic channel for 30 min (henceforth referred to as one AM DA fouling/monitoring cycle). Overall, a >95% reduction in the DA signal after 4 hrs of fouling was observed (Figure S1, see the supporting information), which suggests that the rate is severe and is similar to “no flow” conditions. The energy dispersive x-ray spectroscopy (EDS) studies confirmed the presence of polymers such as polydopamine (PDA) and melanin on the 4 hr fouled BDUNCD surface. Compared to an as-deposited or as-microfabricated “clean” BDUNCD surface (control, Figure 3A), the fouled surface showed a 4× increase in nitrogen (N) and oxygen (O) atomic weight percentages, i.e. from 1.22 ±0.11 to 4.39 ±0.2 and 0.22 ±0.04 to 4.45 ±0.07, respectively (n = 5) (Figure 3B). This increase in N and O content is expected on the severely fouled surface because of the DA by-products.67 A similar increase in sodium (Na) and chlorine (Cl) atomic weight percentages arising from the 1X PBS buffer was also observed. An increase in N/C ratio (at. wt %) from ~ 0.015 to 0.074 and O/C ratio (at. wt%) from ~0.017 to 0.123 is due to an accumulation of N and O surface functionalities, which are known to be abundant in DA oxidative by-products. SDI was employed to investigate the changes in BDUNCD’s electrochemical activity during fouling. Compton and colleagues have utilized the SDI technique to study the electrochemical deposition of silver particles on BDD electrodes.68 Recently, our group chose SDI to investigate the electrical heterogeneity of carbon nanotube (CNT) modified BDUNCD microelectrodes.61 FESEM images were obtained at the 0th, 1st, 3rd and 4th h (Figures 2B-E, inset). As-deposited BDUNCD electrodes are inherently heterogeneous in terms of their electrical conductivity and


Page 8 of 38

Page 9 of 38 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


electrochemical activity due to non-uniform distribution of boron dopants in grains and grain boundaries. A highly uniform distribution of Ag particles across the as-deposited BDUNCD microelectrode surface (2.04±0.06 particles/µm2) reveals similar electroactivity across the electrode surface (Figure 2B, inset). The nearly spherical Ag particles (∼500±55 nm) confirms a diffusion-controlled particle growth mechanism, which is expected on hydrophobic surfaces such as an as-deposited hydrogen-terminated BDUNCD electrode (Figure S2 A1-A3). After 1 hr of fouling, the electrode surface tends to became hydrophilic due to oxygenation69,70 (confirmed from EDS studies), which resulted in Ag dendritic clusters (3-4.5±0.45 µm) (Figure 2C, inset). In addition, smaller Ag particles were randomly distributed based on the local electrical conductivity values. For example, more conductive regions or facets have a distribution of 1.37±0.16 particles/µm2 mostly in the form of agglomerated large-sized particles (500±70 nm). While, less conductive facets have a distribution of 1.7±0.39 particles/µm2 with much smallersized particles (250±40 nm), which could be a sign of surface oxidation (Figure S2 B1-B3). After 3 hrs of fouling, little agglomeration of Ag particles on the electrode surface suggests the presence of an insulating film intermingled with small very sparsely-distributed electroactive regions resulting in smaller Ag particle deposition (111±8.1 nm) with a particle density of 2.04±0.2 particles/µm2 (Figure S2 C1-C3). After 4 hrs of fouling, more Ag particles grew with a particle density of 3.02±1.03 particles/µm2 and with considerably smaller-sized particles (46±8.74 nm) (Figure S2 D1-D3). Thus, the fouling at the BDUNCD microelectrode is similar to a CFM in terms of passivation film formation and loss of electrode area. However, in addition, the BDUNCD microelectrode undergoes surface oxidation with undesirable surface chemistry, which tends to reduce the electroactivity of the electrode surface resulting in lower sensitivity for DA detection.



Surface Fouling Studies of BDUNCD Microelectrodes using FSCV Method Applying a similar methodology as the previous section, the baseline fouling characteristics were investigated during FSCV cycling19,33,56,64 without using any cleaning or reactivation methods. The microelectrode was intentionally fouled by flowing a 100 µM DA solution at 0.2 mL/min in the microfluidic channel and FSCV cycled (−0.4 V to +1.0 V, 400 V s-1, 60 Hz) for 30 min (henceforth referred to as one FSCV DA monitoring

cycle). A total of 13 such cycles

(equivalent to 1.4 million FSCV cycles) were performed with intermittent FSCV measurements at 10 Hz after each cycle (Figure 4B, C). At 0th hr i.e. on as-deposited control surface, the DA anodic peak current (Ipa) recorded was 1.00±0.06 µA with a Ipa/Ipc ratio of ~1 (Figure 4B, green curve). The anodic and cathodic peak potential of ~0.6±0.01 V and ∼0.25±0.01 V, respectively are in good agreement with the literature.33,56,64 Until the 4th cycle (equivalent to 400,000 FSCV cycles), both anodic and cathodic currents increased sharply (brown curve). The current increases could be due to the mild oxidation of the surface that introduced favorable functional groups such as weakly bonded carbonyl (C=O) or carboxyl (−COOH) groups that are known to enhance DA adsorption and oxidation currents.55 The resting potential at −0.4 V during successive scans also exerts a positive effect on the current compared to the −0.5 V and −0.6 V (Figure S3, see the supporting information). The currents immediately reduced after 7 cycles and a 40% decrease was observed at cathodic potentials after 13 cycles when compared to −0.4 V data. After the 4th cycle, the anodic current gradually decreased by 56% at the end of the 13th cycle (black curve), which could be due to both polymer-induced fouling and surface oxidationinduced fouling. Sabine and colleagues reported that such extended anodic and cathodic potentials alters the BDD surface chemistry and negatively affects the signals with large hysteresis and lower currents.71 The shift in the anodic peak potential from +0.6 V to +0.8 V


Page 10 of 38

Page 11 of 38 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


signifies the presence of DA/DOQ redox centers and hydroquinone moieties65. Further shifts in the anodic peak potential towards more positive values until the 8th cycle (equivalent to ~860,000 FSCV cycles) and the gradual decrease in anodic current (Figure 4B, red curve) suggests that there is an increased formation of the polymer film. The SEM images of the SDI studies (Figures 4D,E) showing nearly spherically shaped Ag particles at the 0th cycle and the sparse agglomeration of Ag particles with randomly distributed smaller particles at 13th cycle (Figures 4F,G) suggests a fouling mechanism that is very similar to the AM method described in the previous section. The FSCV fouling is less severe than that of the AM method. In FSCV method, the microelectrode was able to detect DA for up to 6.5 hr with only a 56% signal loss,. However, in AM method, there was almost a total signal loss in 4 hrs. The reduced fouling rates in FSCV method could be due to formation of a non-continuous thin DA-based polymer film with electrochemically active indole species. Effect of ECC Cleaning on BDUNCD Surface Fouling It is evident that BDUNCD microelectrodes also requires frequent cleaning (similar to CFMs) for long-term (>4 hrs) DA measurements. Therefore, to improve the useful lifetime of BDUNCD microelectrodes, an investigation of the effect of ECC on reducing fouling was conducted. Three FSCV waveforms, −0.4 V to +1.0 V (standard), −0.5 V to +0.8 V (shortened) and −0.5 V to +1.2 V (extended) that are in vivo suitable were studied. Additionally, these waveforms facilitate an understanding of the effect of anodic (+0.8 V vs. +1 V vs. +1.2 V) and cathodic/resting (−0.4 V vs. −0.5 V vs. −0.6 V) potentials on extending the electrode lifetime. After each FSCV DA monitoring or fouling cycle, the microelectrode was cleaned by flowing 1X PBS buffer solution at 0.16 mL/min in the microfluidic channel and FSCV cycling (−0.4 V to +1.0 V, 400 V s-1, 60 Hz) for 5 min (henceforth called “one FSCV cleaning cycle”). Again, a total of 13 fouling-



cleaning cycles were performed with intermittent FSCV measurements (0.4 V to +1.0 V, 400 V s-1, 10 Hz) after each fouling-cleaning cycle (Figures 5A-E). Based on the changes in the anodic peak current and peak potential values and the quality of the FSCV signal in terms of hysteresis and clear current peaks during the 13 cycles, the standard waveform ECC cleaning (Figure 5A) was determined to be the best method of cleaning fouled BDUNCD microelectrodes. The trend in the fouling characteristics (Figure 5D, green dots) is similar to the microelectrode that had no cleaning (Figure 5D, red dots). However, the initial increase in the anodic peak current was extended up to 8 cycles instead of 4 cycles and the current at the end of 13th cycle is still 90±0.72 nA higher than that of the 0th cycle (Table S2). Interestingly, the shift of initial anodic peak potential (Figure 5E, green dots) for standard ECC cleaning over 13 cycles is ~9% as compared to ~24% for that of the microelectrode that had no ECC cleaning (Figure 5E, red dots). This reflects retention of favorable surface chemistry that promotes high signal current values (Table S2). This could be due to either limited oxidation of less-stable non-diamond impurites at the grain boundary19 or an efficient removal of DA by-products, which needs experimental verification. The FSCV signal quality is retained even after 13th cycles with no hysteresis and clear anodic and cathodic peaks. In the case of shortened waveform ECC cleaning (Figure 5B), the low anodic potential of +0.8 V retained more DOQ on the electrode surface and also in creating a greater initial increase in the anodic currents due to a greater percentage of DOQ reduction to DA. The mild anodic potential of +0.8 V decreased the surface oxidation and altered the surface functional groups that are favorable in enhancing the DA adsorption rate. However, after the 3rd cycle, the anodic current decreased sharply (Figure 5D, purple dots) as compared to the standard waveform ECC cleaning. This could be due to the difficulty in removing the oxidative by-products at mild anodic potentials and the formation of −OH terminated regions at


Page 12 of 38

Page 13 of 38 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


high cathodic potentials that act as high-affinity catechol anchor groups. This resulted in unfavourable cyclization and contributed towards further passivation of the surface.69,70 The anodic current decreased by ~30% at the end of the 13th cycle (Table S2) and the FSCV signal is distorted with large hysteresis and a poor anodic peak. In the case of extended waveform ECC cleaning (Figure 5C), the high anodic potential of +1.2 V resulted in an enhanced DA adsorption rate due to surface becoming more hydrophilic. Interestingly, the high anodic potential of +1.2 V provided the largest initial increase in the anodic current (4.0±0.2 µA, Figure 5D, brown dots ) (Table S2), which could be due to a combination of electrochemically active surface functional groups aiding DA adsorption and an efficient removal of the DA by-products. After the 4th cycle, the anodic current decreased sharply (Figure 5D, brown dots) similar to that of shortened waveform ECC cleaning. This could be due to excessive surface oxidation at high anodic potentials and the formation of −OH terminated regions at high cathodic potentials that are again unfavourable to cyclization, enhancing the possibility of further oxidation and polymerization to melanin (confirmed by a higher charge transfer resistance value as discussed in the next section). The anodic current decreased by 38% (Table S2) at the end of the 13th cycle and the FSCV signal is distorted with large hysteresis and a poor anodic peak. The trend of anodic peak potential change during the different cycles of the shortened and extended waveforms (Figure 5E, purple/brown dots) is similar to that of the microelectrode that had no cleaning (Figure 5E, red dots) but with a slightly larger increase in the peak potential values (Table S2). The statistical analysis of the data was performed using two-way ANOVA for all four conditions (three ECC cleaning waveforms vs. no cleaning) (n=3, P 4 hrs) DA monitoring are reported. SEM, EDS, SDI and EIS data confirm that the BDUNCD surface is fouled due to fouling that arises from DA by-products such as melanin (similar to fouling at CFM) and fouling that arises from surface oxidation and formation of unfavourable surface functional groups that has high affinity towards DA derivatives resulting in poor reversibility for DA/DOQ cyclization. The AM fouling studies suggest that the BDUNCD microelectrodes are


Page 16 of 38

Page 17 of 38 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


heavily fouled due to a combination of polymer formation and surface oxidation, which results in severely distorted FSCV signals and a >90% loss in signal and which unfortunately starts as early as 3 hrs into a period of continuous DA monitoring in “no flow” condition. The FSCV fouling studies suggest that the FSCV method can be employed to detect DA for up to 6.5 hrs with only a 56% signal loss under flow conditions. On the contrary, the electrode fouled almost completely in AM method under similar flow conditions. This clearly underscores the need for cleaning the fouled BDUNCD microelectrodes irrespective of the detection method if they are to be used for long-term measurements. ECC cleaning methods were investigated by applying three waveforms (standard, shortened, extended) that are all in vivo suitable. Among them, the standard waveform (−0.4 V to +1.0 V) provides the best cleaned surface even after 6.5 hrs monitoring with a fully retained voltammogram shape, no hysteresis, no DA signal loss (a 90±0.72 nA increase) and the smallest Rct value of 0.4±0.016 MΩ. Most importantly, this is the same waveform that is widely used to detect DA using a CFM. Future work for testing the microelectrodes for more than 24 hrs DA monitoring is anticipated. METHODS Commercial grade 4-inch silicon wafer was used for growth of BDUNCD films covered by a 1 µm thick silicon dioxide (SiO2) layer deposited by HFCVD (Hot filament chemical vapor deposition).44,76 The BDUNCD growth process and related microfabrication for BDUNCD with nine individually addressable planar disk microelectrodes of 250 μm diameter in a 3×3 microelectrode array (MEA) format is discussed in detail by Siddiqui et al.44 The electrical conductivity of the film was < 0.1 Ω cm based on 4-probe measurements (Pro4, Lucas Labs, Gilroy, CA) with ∼3000 parts per million (0.3%) boron-to-carbon mole ratio as confirmed by the service provided (Advanced Diamond Technologies, Inc., IL) who performed the BDUNCD



deposition. Electrical isolation of connecting pads of these MEAs were confirmed with a twopoint probe multimeter to ensure proper SiO2 passivation for further characterization.34 All the MEAs were configured as working electrodes in the electrochemical analysis and were precleaned in ethanol for 30 s and dried in nitrogen before use. The MEAs were also electrochemically cleaned in 0.05 M sulfuric acid by cycling them from −0.2 V to +0.8 V at a scan rate of 0.1 Vs-1 to ensure reproducible MEA surfaces free from any surface impurities and with a stabilized background (charging) current. The details of the PDMS microfluidic system bonding to the BDUNCD MEA chip is shown in the supporting information (Figure S4). All electrochemical experiments on the microfluidic channel were performed in two electrode configuration with Pt wire (1.5 mm dia) as a reference/counter electrode (Alfa Aesar) using an Autolab potentiostat (PGSTAT 302N, Metrohm USA). The potentiostat is equipped with a Frequency response analyzer 2, ECD and multiplexing modules along with NOVA 1.10.3 software for data acquisition and analysis. A 100 μM DA solution prepared in 1X PBS buffer was used. For EIS studies, the Nyquist plots were obtained in a custom-made static Teflon cell where a 4 mm diameter O-ring area was exposed to access the BDUNCD MEAs. A 5mM K4Fe(CN)6/5mM K3Fe(CN)6 solution prepared in 1M KCl was used here in a three electrode configuration. And the data was recorded between 100 kHz and 100 mHz with a 10 mV ac signal amplitude and +0.23 V (Open Circuit Potential, OCP) with a Platinum coil (Alfa Aesar) as a counter electrode and a saturated calomel electrode (SCE, Accumet, New Hampshire, USA) as the reference electrode.75 Deionized (DI) water was prepared through a three-filter purification system from Continental Water Systems (Modulab DI recirculator, service deionization polisher). Fouling of MEAs were carried out in the static cell (instead of the microfluidic cell) using a two electrode AM method. On the other hand, MEAs integrated in microfluidic flow cell


Page 18 of 38

Page 19 of 38 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


(MFC) were fouled through controlled FSCV cycling. During MFC integrated MEAs FSCV cycling, the resting potential was −0.4 V with continuous exposure to DA. SDI imaging was realized by Ag particle electrodeposition on the BDUNCD surface by biasing at −0.5 V in a Teflon cell with an O-ring area applying the two electrode method where a platinum coil served as the counter/reference electrode with a deposition time of 200 s for each 250 µm BDUNCD microelectrode in 3 mM of silver nitrate (AgNO3) solution prepared in 0.01 M perchloric acid77. Samples after Ag deposition were briefly dried in nitrogen gas. Perchloric acid was used to improve the stability of AgNO3 and to effect a uniform deposition of Ag particles.78 The surface morphology of as-deposited BDUNCD and Ag deposited on different ECC cleaned and fouled BDUNCD surfaces was imaged and analyzed by FESEM (Hitachi S-4800) along with EDS (integrated in FESEM) for measurement of active and passivated surfaces. All chemicals were reagent grade and purchased from Sigma Aldrich Chemical Co. The chemicals were employed as received unless otherwise specified. All solutions were freshly prepared on the day of the experiment.

ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website at DOI: XYZ. AUTHOR INFORMATION Corresponding Author Email: [email protected]



ORCID Prabhu U. Arumugam: 0000-0001-9942-7036 Author Contributions P.U.A. conceived of the work, all of the authors collectively designed experiments, prepared the manuscript, and collected and analyzed data. Funding This work was supported by the Louisiana Board of Regents-RCS support fund (LEQSF 201417)-RD-A-07) to A.C, G.D and P.U.A) and National Science Foundation OIA/EPSCoR grant (1632891 to S.S). Notes The authors declare no competing financial interest. ABBREVIATIONS BDD, Boron doped diamond BDUNCD; boron doped ultrananocrystalline diamond, CV, cyclic voltammetry; FSCV, fast-scan cyclic voltammetry; CFM, carbon-fiber microelectrode; EIS, electrochemical impedance spectroscopy; PBS, phosphate-buffered saline; DA, dopamine; FIGURE CAPTIONS Figure 1. SEM images showing a 3×3 BDUNCD MEA with nine individually addressable 250µm diameter microelectrodes. Scale bars for (A, B): 1 mm and 100 µm, respectively. Figure 2. (A) BDUNCD microelectrode fouling studies using AM method. A +0.3 V was applied for 4 h in a 100 µM DA solution prepared in 1X PBS buffer (blue curve). (B-E) FSCV


Page 20 of 38

Page 21 of 38 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


th

st

rd

th

measurements taken at 0 , 1 , 3 and 4 h (blue, brown, green, red curves). The scan rate is 400 V/s. The standard waveform (−0.4 V to +1.0 V) is applied. The voltammograms is shown after the background subtraction. The arrows show the direction of the FSCV scan. (Inset) SEM images of Ag particle deposited electrochemically on the BDUNCD microelectrode surface. The scale bar is 500 nm. Figure 3 (A) EDS spectra of as-deposited BDUNCD surface tracking presence of elements like Boron (B), Carbon (C), Nitrogen (N), Oxygen (O), Sodium (Na), Silicon (Si) and Chlorine (Cl) on the surface. (B) shows the average atomic weight percentages of elements B, C, N, O, Na, Si and Cl that were collected from five different spots within the same microelectrode surface. (C) EDS spectra of BDUNCD surface that was taken after applying a +0.3 V for 4 h amperometry in deoxygenated DA solution. Inset is the chemical structure of melanin film formed after DA polymerization. Inset molecular structure of dopamine derived melanin. (D) shows the average atomic weight percentages and its change with respect to as deposited surface for elements like B, C, N, O, Na, Si and Cl which were collected from five different spots within the same microelectrodes (n=3) surface after exposure to DA solution for 4 h during amperometry. Figure 4. Effect of BDUNCD surface fouling measured by FSCV DA signals. (A) An optical image of the experimental setup showing the BDUCND chip integrated with the custom microfluidics, micropumps and the electrochemical cell used to house the microelectrode setup. The details of the microfluidic cell is shown in the supporting information. (B) FSCV voltammograms taken after different fouling cycles. Legends: as-deposited (green curve), after th

th

th

4 cycle (brown), 8 cycle (red) and 13 cycle (black). The arrows show the direction of the FSCV scan. The voltammogram is shown after the background subtraction in 1X PBS buffer solution. (C) The effect of surface fouling on the DA anodic peak current. The FSCV



Page 22 of 38

voltammograms are recorded using a 250 µm BDUNCD microelectrode in 100 µM DA solution prepared in 1X PBS buffer solution. The scan rate is 400 V/s, the waveform is −0.4 V to +1.0 V th

at 10 Hz (n = 3). Inset: SEM images of Ag particles deposited on to the as-deposited, 0 cycle th

BDUNCD surface (D, E - green box) and a fouled, 13 cycle BDUNCD surface (F, G - black box). The scale bars are 100 µm and 5 µm, respectively. Figure 5. Effect of ECC cleaning waveforms on fouled BDUNCD surfaces. FSCV voltammograms measured after applying continuous FSCV waveform (−0.4 V to +1.0V, 400 V/s) at 60 Hz was applied for 30 min in 100 µM DA with 0.2 ml/min flow rate for the intentional fouling. And then followed by ECC cleaning cycle with three waveforms (A-C) standard, shortened and extended waveforms, respectively for 5 min in 1X PBS. All voltammograms are th

th

shown after the background subtraction in 1X PBS. Legends: 0 cycle (green), 4 cycle (brown), th

th

8 cycle (red) and 13 cycle (black). The arrows show the direction of the FSCV scan. (D) The changes in the anodic peak currents and (E) peak potentials. Legends: no ECC cleaning (red), standard waveform (green), shortened (purple) and extended (brown). Two-way ANOVA was performed (n=3, P

Surface Fouling of Ultrananocrystalline Diamond Microelectrodes

Recommend Documents