Article pubs.acs.org/est
Toward Tailored Functional Design of Multi-Walled Carbon Nanotubes (MWNTs): Electrochemical and Antimicrobial Activity Enhancement via Oxidation and Selective Reduction Leanne M. Gilbertson,† David G. Goodwin, Jr.,§ André D. Taylor,† Lisa Pfefferle,† and Julie B. Zimmerman*,†,‡ †
Department of Chemical and Environmental Engineering, Yale University, New Haven, Connecticut 06520-8286, United States School of Forestry and Environmental Studies, Yale University, New Haven, Connecticut 06520, United States § Department of Chemistry, Johns Hopkins University, Baltimore, Maryland 21218, United States ‡
ABSTRACT: Multiwalled carbon nanotubes (MWNTs) are utilized in a number of sectors as a result of their favorable electronic properties. In addition, MWNT antimicrobial properties can be exploited or considered a potential liability depending on their intended application and handling. The ability to tailor electrochemical and antimicrobial properties using economical and conventional treatment processes introduces the potential to significantly enhance product performance. Oxygen functional groups are known to influence several MWNT properties, including redox activity. Here, MWNTs were functionalized with oxygen groups using standard acid treatments followed by selective reduction via annealing. Chemical derivatization coupled to X-ray photoelectron spectroscopy was utilized to quantify specific surface oxygen group concentration after variable treatment conditions, which were then correlated to observed trends in electrochemical and antimicrobial activities. These activities were evaluated as the potential for MWNTs to participate in the oxygen reduction reaction and to have the ability to promote the oxidation of glutathione. The compiled results strongly suggest that the reduction of surface carboxyl groups and the redox activity of carbonyl groups promote enhanced MWNT reactivity and elucidate the opportunity to design functional MWNTs for enhanced performance in their intended electrochemical or antimicrobial application.
■
INTRODUCTION
In order to evaluate the influence of surface modifications on MWNT properties and behavior, direct correlations must be established between specific functional groups and identified trends in physicochemical properties. Quantifying precise amounts and types of functional groups on carbon nanomaterials poses a challenge. Currently available methods include thermogravimetric analysis (TGA), temperature-programmed desorption (TPD), Boehm titration, Raman spectroscopy, infrared spectroscopy (IR), nuclear magnetic resonance spectroscopy (NMR), electron energy-loss spectroscopy (EELS), near edge X-ray absorption fine structure (NEXAFS), energy dispersive X-ray analysis (EDAX), and X-ray photoelectron spectroscopy (XPS). 11−16 Each technique has limitations and are oftentimes used in combination to provide semiquantitiative data on the relative amount of functionalization.17 In order to overcome these limitations, a procedure involving chemical derivatization (CD) coupled to XPS (CDXPS) was established. CD-XPS is utilized to selectively tag oxygen functional groups and determine their relative
The unique properties of pristine and functionalized multiwalled carbon nanotubes (MWNTs) are studied extensively for both their theoretical merit and to guide the development of novel products and product improvements across many sectors, including energy, electronics, and healthcare.1−9 MWNT reactivity, in particular the ability to facilitate electron transfer, is of particular interest to many of these applications, from either an electrochemical or antimicrobial activity perspective. The ability to tailor these activities using economical and conventional treatment processes introduces the potential to significantly enhance product performance. While a previous comprehensive characterization of oxygen functionalized MWNTs (O-MWNTs) annealed from 400−900 °C identified a significant trend in antimicrobial and electrochemical activity that correlated with the maximum annealing temperature, the report did not establish correlation with the composition of surface oxygen functional groups.10 In that study, a stepwise increase in O-MWNT activity was observed up to 800 °C followed by a significant decrease upon annealing at 900 °C. This data provided an interesting trend but did not clearly elucidate the mechanism that controls the changes in chemical activity upon annealing. © 2014 American Chemical Society
Received: Revised: Accepted: Published: 5938
January 27, 2014 April 20, 2014 April 22, 2014 April 22, 2014 dx.doi.org/10.1021/es500468y | Environ. Sci. Technol. 2014, 48, 5938−5945
Environmental Science & Technology
Article
Table 1. Compiled Treatment Conditions and Characterization Data Including Relative Electrochemical Activity (Half-Wave Potential (E1/2) and Electron Transfer Efficiency (ETE)) and Total % Surface Oxygena MWNT sample
manufacturer
acid treatment (AT) conditions
AT1 AT1−600 AT1−900 AT2 AT2−400 AT2−600 AT2−900 AT2−1400 NL-AR NL-400 NL-600 NL-900 NL-18% NL-25% NL-70%
CheapTubes CheapTubes CheapTubes CheapTubes CheapTubes CheapTubes CheapTubes CheapTubes NanoLabs NanoLabs NanoLabs NanoLabs NanoLabs NanoLabs NanoLabs
1 h 70% HNO3 1 h 70% HNO3 1 h 70% HNO3 2 h 70% HNO3 2 h 70% HNO3 2 h 70% HNO3 2 h 70% HNO3 2 h 70% HNO3 H2SO4/HNO3 By H2SO4/HNO3 By H2SO4/HNO3 By H2SO4/HNO3 By 18% HNO3 25% HNO3 70% HNO3
annealing temp (He) 600 °C 900 °C 400 °C 600 °C 900 °C 1400 °C
manufacturer manufacturer manufacturer manufacturer
400 °C 600 °C 900 °C
E1/2 (V)b
ETE
Tptal %O
−0.215 −0.200 −0.232 −0.219c −0.221c −0.127c −0.244c −0.244 −0.235 −0.215 −0.214 −0.238 −0.213 −0.217 −0.213
2.92 2.20 2.56 2.65c 2.28c 2.47c 2.12c 2.38 2.62 2.64 2.63 2.45 1.87 2.44 2.65
7.6 4.6 3.5 1.3 0.8 8.5 4.4 3.0 1.2 4.2 5.4 7.9
a
Abbreviations used to identify each sample: AT1 (acid treated for 1 h); AT2 (acid treated for 2 h); NL (NanoLabs); AR (as-received); X00 indicates the maximum annealing temperature maintained for 1 h; X% indicates the strength (% w/w) of acid used during acid treatment. bChanges in E1/2 values >0.01 are italicized indicating significant right shifts in the respective polarization curve (1600 rpm). cElectrochemical activity results previously reported.10
carbon active sites are restored. The liberated ROS rapidly oxidizes additional GSH to GSSG.36 In another study, systematic changes in surface oxygen content was shown to significantly influence the redox properties of different OMWNTs.10 Yet, there remains a lack of empirical evidence that identifies the specific functional group associated with the observed trend in reactivity. In this study, CD-XPS is employed as the primary characterization technique to identify COOH, C−OH, and CO groups on the surface of MWNTs treated under various conditions to systematically alter the oxygen functional group population. The identification of these groups seeks to resolve the type of O-group(s) associated with the significant increase in electrochemical and antimicrobial activity measured by ORR and GSH oxidation experiment, respectively. The results suggest that as the total amount of surface oxygen decreases with increased annealing temperature, the relative amount of each individual group also changes. Furthermore, the specific distribution of the oxygen functional group population correlates with changes in electrochemical and antimicrobial activities. The potential to tailor these properties using economically feasible conventional treatment processes is promising for the advancement of numerous CNT applications and for reducing unintended toxicological consequences.
concentration more accurately than can be done through peak fitting of the C(1s) spectrum of a carbon nanotube sample. Thus, far, methods have been developed to detect carboxylic acid (COOH), hydroxyl (C−OH), and carbonyl (CO) functional groups and have been used to determine how different oxidative treatments impact the relative amount of these three oxygen groups on CNT surfaces.14,15,18,19 The ability of CNTs to donate or withdraw electrons between the carbon surface and participating species, which can be influenced by the presence of surface oxygen functional groups, enables them to catalyze reactions.20 The electrochemical oxygen reduction reaction (ORR) is used as an indicator of relative CNT reactivity.21−26 In ORR, reduction of molecular oxygen occurs at the surface of CNTs following a two electron pathway.27−29 It is hypothesized that the redox reaction occurs predominantly on the edge planes and at defect sites in the basal plane of carbonaceous species, and that oxygen functional groups influence this reaction.7,30,31 Several reports claim that certain types of oxygen functional groups, those containing the carbonyl (CO) moiety, are responsible for enhanced catalytic and electrochemical activity, which may be attributed to their basicity.16,20,27,31−33 Yet, identifying specific surface functionality on CNTs is challenging and further resolution of these oxygen groups is necessary before establishing correlations or causation. CNTs are also known to participate in the production of reactive oxygen species (ROS), the importance of which has been explored for its pertinence to biological activity. Commonly, ROS concentration within a system is quantified directly using indicator molecules.34,35 An alternate method involves measuring the potential for CNTs to promote the oxidation of glutathione (GSH) to its disulfide form (GSSG), which has been used as an indicator of relative CNT antimicrobial activity.10,36,37 While the precise details surrounding the CNT-GSH interactions are not fully resolved, Liu et al. have proposed a mechanism.36 First, they suggest the necessity of a surface oxygen group to facilitate the GSH oxidation reaction. The mechanism involves direct reduction of CNT surface oxides by GSH followed by liberation of ROS as the
■
EXPERIMENTAL SECTION
Materials. Glutathione (GSH, 98%, ACROS Organics), 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB, ≥ 98%, SigmaAldrich), TRIS (ultrapure, American Bioanalytical), sodium bicarbonate (Fisher BioReagents), potassium hydroxide (pellets, MACRON Fine Chemical), dimethyl sulfoxide (DMSO, Sigma-Aldrich), Nafion (DuPont DE520) were all used as received. Derivatizing reagents 2,2,2-trifluoroacetic anhydride (TFAA), 2,2,2-trifluoroethanol (TFE), di-tert-butylcarbodiimide (DTBC), and 2,2,2-trifluoroethylhydrazine (TFH) were purchased from Sigma−Aldrich and used as received. All reagents were 99% pure, with the exception of TFH, which was obtained as a 70% solution (by volume) in water. 5939
dx.doi.org/10.1021/es500468y | Environ. Sci. Technol. 2014, 48, 5938−5945
Environmental Science & Technology
Article
elsewhere.10,26 Briefly, a stable suspension of MWNTs was prepared, 10 μL were deposited onto a glassy carbon electrode (7.5 μg cm−2 loading), and the samples were then allowed to dry. Cyclic voltammetry was performed in nitrogen saturated 1 M KOH followed by ORR measured in oxygen saturated 1 M KOH at increased rotating speeds (400, 625, 900, 1600, and 2500 rpm). Experiments are run in at least triplicate to obtain a reliable sample average.
MWNT Preparation. Pristine MWNTs were obtained from CheapTubes (Burlington, VT, CCVD, >95 wt %, 10−20 nm diameter, 10−20 μm length) and were treated in-house via reflux in nitric acid (HNO3, 15.7 M) for either 1 or 2 h, which served as the starting material for the AT1 and AT2 annealed sample sets. Both pristine and oxygenated MWNTs were obtained from NanoLab Inc. (Waltham, MA 15 ± 5 nm diameter, 1−5 μm length, >95% purity).38,39 The pristine MWNTs served as the starting material for the NL-18−70% sample set. Treatment conditions are described elsewhere.19 The oxygenated MWNTs from NanoLab were treated by the manufacture via reflux in concentrated sulfuric−nitric acid mix and served as the starting material for the NL annealed sample set.39 All annealed samples were heated under helium (He) at the indicated maximum temperatures for 1 h (10° per min heating rate).10 Compiled treatment conditions for all MWNT samples are outlined in Table 1. MWNT Characterization. Sample Purity. The potential for metal catalyst impurities to influence CNT toxicity has been shown.40,41 Yet, the presence of metal catalysts alone is not sufficient to induce a toxic effect, rather the metal particles must be bioavailable.40 Thermogravimetric analysis for the in-house acid treated samples reveal percent mass residual less than 2% while the reported purity of the NanoLabs purchased OMWNTs is >95%. These samples serve as the starting material within each sample set and minimal reactivity is universally observed. X-ray Photoelectron Spectroscopy (XPS). XPS was utilized to determine the total surface oxygen content of each sample and fluorine content upon reaction of targeted oxygen functional groups with chemical derivatization reagents. Approximately 3 mg of MWNTs were dusted onto copper tape (0.5 × 0.5 cm2) until the tape was completely covered. Samples were loaded into a PHI 5600 XPS system (Pbase < 5 × 10−9 Torr) and analyzed using Mg Kα X-rays (1253.6 eV) and a high energy electron energy analyzer, operating at a constant pass-energy of 58.7 eV with a scan rate of 0.125 eV/step (50 ms/step). At each point, the spot size was controlled with a slot aperture of 800 × 2000 μm2 to maximize the photoelectron count. XP spectra were processed using commercially available software (CasaXPS) and integration of the relevant peak areas was used for quantification. Further method details are outlined elsewhere.15,18,19 Chemical Derivatization XPS Analysis (CD-XPS). In order to evaluate the surface oxygen functional group concentrations, samples were derivatized with 2,2,2-trifluoroethanol (TFE) with di-tert-butylcarbodiimide (DTBC) and pyridine, trifluoroacetic anhydride (TFAA), and 2,2,2-trifluoroethylhydrazine (TFH) to quantify carboxylic acid (COOH), hydroxyl (OH) and carbonyl (CO) groups, respectively. Method details are outlined elsewhere.15,19 Acellular Assay. MWNT-Mediated Glutathione (GSH) Oxidation. O-MWNTs were evaluated for their potential to mediate the oxidation of GSH as previously described.10 Briefly, the concentration of GSH was monitored over a 6 h period of time spectrophotometrically using Ellman’s reagent. The percent loss of GSH was calculated compared to the control, which contained no MWNTs. Average and standard deviations are determined from replicated triplicate experiments. Electrochemical Measurements by the Oxygen Reduction Reaction (ORR). MWNT electrochemical activity was determined using a rotating disc electrode (RDE) technique in 1 M KOH. Method details are described
■
RESULTS AND DISCUSSION In this study, the physicochemical properties of MWNTs were systematically altered using oxidative acid treatment followed by selective reduction via annealing (heat treatment under inert conditions). Strong oxidizing agents, such as nitric and sulfuric acid (HNO3 and H2SO4), are commonly used to minimize residual metal oxide catalysts and add oxygen functional groups to the surface of CNTs. Here, multiple treatment conditions were employed to introduce varying initial concentrations of total surface oxygen and oxygen group type. Annealing is utilized to reduce surface oxygen functional groups and to heal tube defects upon reaching high enough temperatures.42 Annealing temperatures used in this study are based upon previously identified significant changes in electrochemical and antimicrobial activity (400, 600, and 900 °C).10 An additional temperature of 1400 °C was evaluated to ensure the continued downward trend in reactivity at temperatures greater than 900 °C. MWNTs were obtained from multiple vendors and batches to demonstrate that the observed trends are conserved and to ensure the robust application of our findings. The treatment conditions as well as results from ORR (electrochemical activity) and XPS (total surface oxygen) are summarized in Table 1. Oxygen Functional Group Identification via CD-XPS. Several studies have documented the thermal stability of oxygen functional groups on the surface of carbonaceous materials and provide empirical evidence of the groups present after annealing at various maximum temperatures.13,17,33,43 For a compiled reference table of decomposition temperatures, refer to Szymanski et al. and Figueiredo et al.33,43 In general, the more labile groups, such as carboxyl groups, decompose at lower temperatures (∼400 °C), followed by hydroxyl (∼700 °C) and carbonyl groups (∼900 °C).33,43 In addition to decomposition of surface oxygen groups with increased maximum annealing temperature, it has been reported that partial decomposition of carboxylic acid and hydroxyl groups results in their thermal conversion to cyclic lactones, lactols and anhydrides. 17,33 Relevant oxygen functional groups are identified in Figure 1. Chemical derivatization coupled to XPS (CD-XPS) has been successfully employed to identify the relative presence of carboxyl, hydroxyl, and carbonyl oxygen groups present on the surface of carbonaceous materials.15,18,19 XPS as well as CDXPS were utilized in this study to quantify the total amount of surface oxygen and determine changes in oxygen group population under the different treatment conditions, respectively. The % [O]COOH, % [O]OH, and %[O]CO were calculated using an established method.15,18,19 These values are compiled in a stacked column plot (Figure 2) to enable visualization of the relative amounts of each group under the different treatment conditions. The % [O]Other is composed of groups such as ethers, esters, lactones, lactols and anhydrides. Current CD-XPS methods are unable to tag these groups specifically. 5940
dx.doi.org/10.1021/es500468y | Environ. Sci. Technol. 2014, 48, 5938−5945
Environmental Science & Technology
Article
A similar trend is observed for the NanoLabs (NL) annealed sample set, including a decrease in total oxygen content with increased annealing temperature, a significant decrease in COOH groups at 400 °C, and further COOH reduction at 600 °C at which point the carbonyl group is the most abundant type of oxygen functional group. Compared to the CheapTubes annealed sample set, the decrease in COOH groups upon annealing at 400 °C is greater for NanoLabs O-MWNTs. In addition, the manufacturer acid treatment, involving a mix of HNO3:H2SO4, resulted in a greater initial amount of CO surface groups than the in-house HNO3 treatments for CheapTubes MWNTs and %[O]CO remains constant throughout the annealing treatment. Pristine MWNTs were treated with HNO3 of increasing strength (18−70% w/w) under otherwise identical conditions. The strength of HNO3 used in the oxidative treatment process is known to influence the surface oxygen content, particularly the total amount of oxygen introduced to the surface.19 CDXPS results confirm that the total amount of oxygen increased with increasing % w/w acid treatment. Furthermore, this increase in total oxygen is primarily as COOH and “other” oxygen-containing groups, consistent with results in the literature.16,19 The combined CD-XPS results of the annealed and acid treated MWNT sample sets reveal opposing trends in total oxygen and oxygen functional group populations under increased treatment conditions. This provides an important opportunity to establish the relationships between oxygen functional group type, total surface oxygen, and the observed trends in electrochemical and antimicrobial activities. O-MWNT Electrochemical Activity. Reduction of oxygen on carbon surfaces has been studied extensively.28,29 In alkaline conditions, like those utilized in this study (1 M KOH), oxygen reduction by carbon surfaces proceeds primarily via the two electron pathways (O2 + H2O + 2 e− → HO2− + OH−).28 This redox reaction is proposed to occur predominantly at the edge planes and at defect sites in the basal plane according to the following reaction mechanism:28,29
Figure 1. Schematic representation of pertinent oxygen functional groups that can be present on the surface of O-MWNTs as a result of acid and annealing treatment.
For the CheapTubes (AT) annealed sample set, the total percent oxygen decreases with increased annealing temperature and the relative amount of each oxygen group also changes. In particular, the proportion of carboxyl groups decreased significantly upon annealing at 400 °C and further decreased at 600 °C. The decomposition of carboxyl groups at ∼400 °C is in agreement with literature values.13,17,33,43 Upon annealing at 900 °C and beyond, the majority of oxygen has been removed from the surface.33,43 As such, the total remaining percent oxygen is too low and the inherent experimental error too high to resolve the relative concentration of the three functional groups by CD-XPS at these higher temperatures with good accuracy. The primary source of this systematic error is associated with the nature of the fluorine reagents used in this method. Fluorine can be slightly overcounted due to its polarity, enabling it to stick to a small extent to the XPS chamber. This can contribute to slightly higher fluorine counts on samples (generally
