Quantitative Detection of Single Walled Carbon Nanotube in Water

Dec 10, 2012 - Ψ Department of Chemical Engineering, Auburn University, Auburn, Alabama 36849, United States. # Civil and Environmental Engineering, ...
0 downloads 9 Views 4MB Size
Article pubs.acs.org/est

Quantitative Detection of Single Walled Carbon Nanotube in Water Using DNA and Magnetic Fluorescent Spheres Linda C. Mota,⊥ Esteban E. Ureña-Benavides,Ψ Yeomin Yoon,# and Ahjeong Son⊥,* ⊥

Department of Civil Engineering, Auburn University, Auburn, Alabama 36849, United States Department of Chemical Engineering, Auburn University, Auburn, Alabama 36849, United States # Civil and Environmental Engineering, University of South Carolina, Columbia, South Carolina 29208, United States Ψ

S Supporting Information *

ABSTRACT: Carbon nanotubes (CNTs) possess unique properties that have led to an increase in their research and usage for a wide variety of fields. This growing demand of CNTs poses a major public health risk given its unregulated release into the environment. Unfortunately there is a significant information gap on the actual quantity of CNTs in the environment due to limitation of existing detection methods. This is mainly owing to the ubiquitous carbon chemistry of CNT. In response we developed a method (CNTcapture method) that is able to structurally differentiate CNT from other interference carbon materials in an aqueous medium. The affinity between single walled nanotubes (SWNTs) and specific single stranded DNA (ssDNA) was employed to capture SWNTs in water. SWNT-specific separation was obtained via magnetic separation. Dual fluorescent labels attached to sandwich ssDNA probes were used for quantification. The specific affinity between DNA and SWNTs was verified and no significant side-interactions were observed. With optimized incubation duration (30 min) and buffer composition (10−7 % sodium dodecyl sulfate and pH 7.9), a calibration curve of quantification (R2 = 0.90) was obtained with a range of SWNT concentration (0.05−10 μg/mL) against graphene as a planar analog. Comparison to other spectroscopy based methods was carried out to highlight the specificity and sensitivity of the presented method for CNT detection in aquatic sample.



INTRODUCTION Carbon nanotubes (CNTs) have demonstrated extraordinary physical and chemical properties since their discovery two decades ago and became candidates for numerous applications such as nanocomposites, energy storage, microelectronic, and medical devices.1−4 Engineered nanomaterials, including CNTs, are beginning to find their way into several commercially available and commonly discarded products; including plastics, papers, textiles, cosmetics, sunscreens, and sporting goods.1−5 A previous study projects the worldwide market for products containing nanomaterials such as CNTs as $1 trillion by 2015.6 The alarming and escalating scale at which engineered nanomaterials are being produced and used in our daily lives parallels that of asbestos more than a century ago. As history has a tendency to repeat itself, it is not surprising that health implications of CNTs have begun to surface, especially since there is an apparent lack of control and regulation in its uses. CNTs can be present as single-walled carbon nanotubes (SWNTs) or multiwalled carbon nanotubes (MWNTs). SWNTs are a one atomic layer thick hollow cylinder of carbon with diameters in the order of a nanometer and lengths ranging from hundreds of nanometers to micrometers.7 Unfortunately, the same attributes of SWNTs that enabled the manufacturing of high performance products, also make them difficult to be removed from the organism when ingested or inhaled. © 2012 American Chemical Society

The fate of CNTs in aquatic environments has invoked significant concern over its safety and environmental implications. These concerns pertain to its toxicity associated with cardiopulmonary diseases,8,9 biopersistence,10 and pathogenicity.10 Potential health risks of CNTs have been demonstrated in mice,11 pigs,12 and human skin.13 The toxicity resembles asbestos 14 and possibly results in lung disease, like mesothelioma (lung cancer), which is caused via inducing oxidative stress.12 There are many pathways where nonsoluble contaminants in aquatic environment can find their way into the atmosphere. CNTs can become airborne contaminants when CNT laden water is allowed to dry on surfaces exposed to winds, such as seasonal creeks, riverbeds and farmlands. It is probable that contaminants such as the CNTs can be carried off together with fine dust by winds after they have been dried. Since it is more likely for CNTs to be first released into an aquatic environment before making its way to the atmosphere, it is therefore critical to detect and quantify the degree of CNTs present in the aquatic environment before it escalates into an environmental and public health catastrophe. Received: Revised: Accepted: Published: 493

September 10, 2012 December 4, 2012 December 7, 2012 December 10, 2012 dx.doi.org/10.1021/es303671u | Environ. Sci. Technol. 2013, 47, 493−501

Environmental Science & Technology

Article

during sonication. All dilutions for the different samples and experiments were prepared from these stock solutions. Field Emission Scanning Electron Microscopy (FESEM). An FE-SEM (JSM-7000F, JEOL, Tokyo, Japan) was used to observe SWNTs and graphene nanoparticles. Samples were prepared for SWNTs and graphene separately at 10 μg/ mL in water. The tubes were sonicated as previously discussed and dispensed on an aluminum pan and allowed to air-dry for two hours. SWNTs and graphene were observed in the FESEM at an accelerating voltage of 20 kV. Gel-Electrophoresis. To assess possible interaction (e.g., dimerization or polymerization) of probe and signaling DNA during incubation, gel-electrophoresis was performed to measure the size of post incubation DNA. This possible interaction was hypothesized as a nondominant affinity binding in the CNT-capture method and it was confirmed by gelelectrophoresis. The length of each probe and signaling DNA is 40 bp, thus, any dimer (80 bp) or polymer (120 bp or more) would be observed in the gel electrophoresis if such dimerization or polymerization occurred during incubation. Samples tested include: 50 μM probe DNA, 50 μM signaling DNA, a mixture of 50 μM of both DNA, 5 μg/mL SWNT only (control), a mixture of 50 μM of both DNA and 5 μg/mL SWNT, and sodium dodecyl sulfate (SDS) containing phosphate buffer (control). Samples were incubated overnight in a mix-plate (MixMate, Eppendorf, Hamburg, Germany) at 1500 rpm. Gel-electrophoresis was performed in 2% agarose gel to separate DNA fragments based on DNA size (bp) at 100 V for 90 min in the gel-electrophoresis system (Biorad, Hercules, CA). After ethidium bromide staining the gel image was acquired via Kodak Gel Logic 100 gel imaging system with a Fisher Carestream Health molecular imaging system. Zeta Potential. The dispersion quality of SWNTs was evaluated by zeta potential measurements as an indirect assessment of SWNT−SWNT binding affinity. Along with DNA dimerization, the SWNT aggregation is also considered as one of possible side-interactions in the CNT-capture method, which should not be dominant compared to SWNT-DNA affinity. However, the aggregation effects, common in CNTs present in water (or aquatic environments), are attributed to the total van der Waals interaction between the nanotubes.19 Thus the investigation of aggregation effect during incubation is critical to the success of SWNT quantification. SWNTs were incubated with DNA and compared to the SWNT only sample by measuring the zeta potential of the samples using a Zetasizer Nanoseries (Nano-ZS, Malvern Instruments, Malvern, UK). Two sets of SWNT samples in duplicate (0.1, 1, 10, and 100 μg/mL) were prepared in 1.5 mL of 10−7 % SDS containing deionized water (pH 5.6) by sonication (8 W, 8 min),20 and serial dilution. For one of the sets, the samples were incubated in the mix-plate for 30 min with 20 μL of 10 μM probe DNA for each sample. After the incubation, 750 μL of each sample was transferred to a clear zeta cell (DTS 1060C, Malvern Instruments, Malvern, UK) for zeta potential measurement. Data was analyzed using DTS (Nano) Software v5.10 (Malvern Instruments). The measured electrophoretic mobility (μ) is converted to zeta potential using the Smoluchowski approximation.21 100 runs were performed per reading after 2 min of equilibration for each sample, one reading per duplicate sample. Fourier Transform Infrared Spectroscopy. FTIR was employed to identify a possible bonding or interaction between SWNTs and DNA as the main affinity binding of SWNT detection. FTIR was also used to assess a possible dimerization

Despite the likelihood of CNT contamination in aquatic systems, current detection technologies are inadequate for the verification of CNT contamination. Existing carbon chemistry based tests are not able to differentiate CNTs from other carbon based compounds that are in abundance in the environment. Differentiation via ultracentrifugation,15 by virtue of the CNTs’ inertia mass, is currently used in laboratories to separate CNTs from the solution. However, this method is not specific to CNTs. Any non-CNT suspended particles that have greater inertia mass than the CNTs will also be separated together with the CNTs via centrifugation. Optical methods such as near-infrared- 16 or UV-vis 17 spectroscopy may not yield a specific response from CNTs as its specific wavelength can overlap with that of other organic compounds or nanomaterials. Since CNTs are not soluble in water, they behave as a suspension and the samples are very heterogeneous, containing nanotubes from various diameters, as well as lengths. This also makes CNT detection using common analytical equipment (e.g., chromatography) nearly impossible. Another significant limitation for CNT quantification is the lack of a detection limit or when provided, the range of concentrations is not environmentally relevant.18 In order to overcome the limitation of existing methods and to address the information gap related to the quantity and proliferation of CNTs in the environment, the objective of this study was to develop a detection method for CNTs (i.e., SWNTs in this study) using DNA and magnetic fluorescence spheres (CNT-capture method). Briefly, SWNTs are specifically captured with G/T alternating single stranded DNA (ssDNA) and the complex is magnetically separated. Quantification is achieved via normalization of fluorescence measurements from ssDNA and magnetic fluorescent spheres. In order for the presented CNT-capture method to function as a SWNT specific detection method, the bonding affinity between ssDNA and SWNT must be dominant, which was verified by Fourier transform infrared spectroscopy (FTIR). In addition, side-interactions between other components (i.e, SWNT−SWNT or DNA−DNA) must be negligible and it was verified via measurement of zeta potential, FTIR, and gel electrophoresis. Method parameters such as incubation time and buffer composition were also investigated and used for the SWNT quantification in the proposed method. The specificity and sensitivity of the CNT-capture method for SWNT detection was further highlighted by running it alongside other potential methods of Raman, near-infrared and ultraviolet (UV)-vis spectroscopy.



EXPERIMENTAL SECTION Preparation of Carbon Nanomaterials. SWNTs and graphene were purchased from Cheap Tubes (Brattleboro, VT). SWNTs were reported by the manufacturer to be 90% pure, 500−2000 nm in length, an outer diameter of 1−2 nm. Graphene nanoplatelets were reported for research grade, less than three layers, 99% purity, and less than 3 nm of thickness with a greater than 750 m2/g in surface area. Graphene was used as a planar analog for the proposed CNT-capture method to show the SWNT specificity of the method. A stock solution of 1000 μg/mL was prepared for each, SWNTs and graphene, by adding 0.015 g of each nanoparticle into 15 mL of deionized water in glass vials. Each solution was sonicated at 10 W for 15 min using a misonix microprobe sonicator XL-2000 series (Qsonica, Newtown, CT). All sonication was performed while placing the tube in an iced water bath to cool down the samples 494

dx.doi.org/10.1021/es303671u | Environ. Sci. Technol. 2013, 47, 493−501

Environmental Science & Technology

Article

Figure 1. The schematic diagram of the proposed carbon nanotube detection using magnetic fluorescent sphere and sandwich ssDNA wrapping technology. CNT and MS depict carbon nanotube and magnetic fluorescent spheres, respectively.

(Sigma-Aldrich). A detailed chemistry for covalent bonding between magnetic sphere and probe DNA is illustrated in Figure 1. The samples were placed in a mix-plate at 1500 rpm for two hours at ambient temperature. Afterward, the samples were washed twice with phosphate buffer (0.1 M, pH 7.4) using DynaMag - 2 magnet (Invitrogen, Carlsbad, CA). As a result, the complex of magnetic sphere and probe DNA were resuspended in 250 μL of 10−7 % SDS containing phosphate buffer. All the incubation and washing steps were performed with an aluminum foil cover to avoid light exposure. SWNT Quantification by the Magnetic Sphere and DNAs. Varying amounts (0, 0.025, 0.05, 0.1, 0.25, 0.5, 1, 2.5, 5, and 10 μg/mL) of SWNTs and graphene (control) were prepared in SDS-phosphate solution for a final volume of 750 μL. SDS composition in buffer (10−1 to 10−8 %) and pH (5.9− 7.9) of phosphate buffer were varied to obtain an optimum buffer solution. Prior to the serial dilution for lower concentration, the suspension was sonicated for 8 min at 8 W to avoid aggregation of SWNTs. Twenty μL of 10 μM signaling DNA (40 bp G/T alternating ssDNA) labeled with Cy5 was added to each tube containing SWNTs. Subsequently, the magnetic sphere-probe DNA complex was added to a range of SWNT solutions and signaling DNA to allow for capturing of SWNTs by both probe and signaling DNA as illustrated in Figure 1. Sample nanotubes were placed in the mix-plate at 1500 rpm and incubated for 5 min −8 h, covered with the aluminum foil. After incubation, the nanotubes were placed in the magnet and washed as previously described. The contents were resuspended in 200 μL of phosphate buffer and transferred for fluorescence measurement using a MDS Spectramax M2 microplate reader. All fluorescence readings (λem = 675 nm, λex = 642 nm for Cy5; λem = 478 nm, λex = 260 nm for yellow fluorescence of magnetic fluorescent spheres) were taken five times for each sample. In order to normalize the different numbers of particle-SWNT complex among reactions, the Cy5 signal from signaling DNA was divided by the yellow

between DNA as side interaction during SWNT detection as described earlier. Attenuated total reflectance (ATR) at 45 degree incidence angle was performed using a Nicolet FT-IR spectrometer (iS10, Thermo Fisher, Waltham, MA). A germanium single bounce crystal plate was used to measure the samples (iTR/iD5, Thermo Fisher). The data was analyzed using OMNIC 8.1.210, samples were run at 64 scans and the resolution set to 4 cm−1. The spectra were corrected for baseline and ATR correction. In order to prepare the samples for ATR analysis the DNA was precipitated. In a 1.5 mL centrifuge tube containing 150 μL of 100 μM DNA or a mixture of 75 μL of each 100 μM ssDNA, 750 μL ethanol, 100 μL 3 M sodium acetate (Sigma−Aldrich, St. Louis, MO) was added, vortexed and placed inside a freezer at −20 °C for one week. Afterward the solution was centrifuged at 14 000 rpm for 10 min by Eppendorf centrifuge 5418. The supernatant was carefully discarded; 0.5 mL of ethanol was added for washing purposes and centrifuged at 14 000 rpm for 10 min. Supernatant was carefully removed, and the precipitate was allowed to air-dry. The precipitate for each sample was placed carefully in the germanium plate for ATR analysis. SWNT powder was placed in the plate until the germanium was covered. Conjugation of Magnetic Fluorescence Spheres with Probe DNA. Magnetic fluorescence spheres (1% w/v, 1−1.4 μm) with both carboxylic functional groups and encoded yellow fluorescence were purchased from Spherotech (Lake Forest, IL). The probe DNA were 5′ aminated ssDNA with G/ T alternating sequences (total 40 bp) and commercially synthesized (IDT, Coralville, IA). This specific sequence and length were chosen based on previous literature for the best dispersion of SWNTs.20 Similarly to previous studies,22,23 a covalent bond was made between the carboxylic magnetic spheres (at a final concentration of 0.005%) and the aminated probe DNA (500 μL of 0.4 μM probe DNA) with an aid of ethylcarbodiimide hydrochloride and N-hydroxy-succinimide 495

dx.doi.org/10.1021/es303671u | Environ. Sci. Technol. 2013, 47, 493−501

Environmental Science & Technology

Article

fluorescence signal from the magnetic spheres. Using the fluorescence of magnetic spheres for normalization is the unique feature of the developed method as discussed elsewhere.22,23 SWNT Quantification by Other Potential Methods. To show the potential CNT detection capability of existing methods and to compare it to the CNT-capture method, three conventional methods were selected. Raman spectroscopy was chosen because it has been commonly used for CNT characterization.20 Previous studies indicated that unique peaks of individually dispersed CNTs were observed at wavelengths of around 1200−1300 nm (visible-near-infrared range) via near-infrared spectroscopy20,21,24 and around 567 nm via UVvis spectroscopy.25 SWNT samples of various concentrations (0, 10, 50, 100, 250, 500 μg/mL) were prepared and sonicated in a total volume of 1.5 mL with 10−7 % SDS solution. Prepared samples were used for Raman, near-infrared, and UV-vis spectroscopy to obtain a calibration curve for SWNT detection as a comparison to the CNT-capture method. A detailed description for three methods used is described in the Supporting Information. The range and equation of quantification, correlation coefficient (R2), and specificity to SWNT detection were also determined for each method tested.

the sample that contained SWNTs and ssDNA (Figure 2b), a subtraction of SWNT spectrum from the spectrum of the mixture was performed in order to assess and compare the peaks of the ssDNA. Figure 2b shows the two main areas for DNA analysis, the PO2− backbone peaks, as well as the bases. The peaks in the PO2− backbone were slightly shifted from 1061 to 1057 cm−1 for the symmetric stretch and from 1226 to 1206 cm−1 for the asymmetric stretch, as compared to ssDNA. This shift in the peak’s position can be attributed to a possible change of conformation of the ssDNA in order to interact with the SWNTs. Previous studies have demonstrated a shift in PO2− peaks that was attributed to a possible change of conformation in the DNA backbone to adjust for interaction.28 In the spectrum of Figure 2b, the peak corresponding to the CO (carbonyl) functional groups in the aromatic bases can be seen. Since the CO is an integral part of the aromatic rings, a shift in the peak from 1688 to 1592 cm−1 is evidence of π−π interaction of the aromatic bases and the nanotubes’ surface. π-stacking can result from interactions between the delocalized electrons in the p-orbitals of the carbon ring on the nanotubes surface and the DNA. The presence of strongly polarizing heteroatoms has major influence in electrostatic interactions. In the case of DNA, due to the electronegativity of functional groups located in the ring of the bases, the ring becomes slightly more positive, enabling electrostatic interaction with the electrons in the p-orbitals on the CNT surface.29 Studies have been reported where a shift in this carbonyl peak can be observed and is attributed to an interaction of the gold surface with the DNA.30 In a study where SWNTs were dispersed using ssDNA through sonication, the findings suggest a helical interaction between SWNTs and ssDNA evidenced by a higher intensity of peaks in the areas of 1234 cm−1 for PO2− and from 1720 to 1640 cm−1 for the bases.31 The results described here support a preferential interaction between SWNTs and ssDNA. Negligible Dimerization between Probe DNA and Signaling DNA. As part of the development of the presented CNT-capture method, it is important to verify that the side interactions (e.g., DNA−DNA, SWNT−SWNT) between the components are negligible. To evaluate if dimerization (or polymerization) occurred between DNAs, the samples were incubated and the size of the DNA was assessed and compared to the individual ssDNA. As shown in Figure 3a, there was no significant shift in the observed bands between the mixture of ssDNA and individual ssDNA (40 bp). Since the sequences of the ssDNA are not complementary to each other, it was expected that no dimerization would be observed. ATR-FTIR was used to further assess possible bond formation or functional group interaction. It has been shown that the FTIR spectrum of DNA can be divided in two main areas: the PO2− backbone, that can be found from 1250 to 950 cm−1; and the bases of the DNA, where the purine and pyrimidine ring can be observed around 1600 to 1500 cm−1; and the CO, C−N, CC inside the aromatic ring can be seen from 1750 to 1600 cm−1.28,32 As demonstrated in Figure 3b, we can observe the previously mentioned peaks of the PO2− symmetric stretch at 1061 cm−1 and the asymmetric PO2− stretch at 1226 cm−1. The pyrimidine and purine ring was found at 1481 cm−1 and the CO of the aromatic ring was observed at 1688 cm−1. In the mixed ssDNA there was no significant shift of peaks as exhibited in Figure 3c, suggesting that there was no side interaction between the two ssDNA. From the data it can be



RESULTS AND DISCUSSION Dominant Affinity between ssDNA and SWNTs. The affinity binding interaction between SWNTs and ssDNA was assessed to verify that it is the main binding mechanism of the CNT-capture method. SWNTs and ssDNA were mixed and ATR-FTIR was performed to evaluate interactions between each other. As observed in Figure 2a, SWNT only sample exhibited a large hump over most of the spectrum attributed to a high light absorbance and lack of functional groups.26,27 For

Figure 2. Affinity investigation between SWNTs and DNA bonding via FTIR spectra: (a) SWNT only (control) and (b) the mixture of SWNTs and probe DNA. As depicted by arrows, the peaks at 1057, 1206, and 1592 cm−1 indicate the potential bonding between SWNTs and DNA during incubation for CNT detection. 496

dx.doi.org/10.1021/es303671u | Environ. Sci. Technol. 2013, 47, 493−501

Environmental Science & Technology

Article

Figure 3. Investigation of DNA−DNA interaction (potential dimerization as a side reaction) via gel electrophoresis and FTIR analysis. In the gel picture of (a) lane 1:20 bp incremented DNA ladder; lane 2: 50 μM probe DNA (40 bp); lane 3: 50 μM signaling DNA (40 bp); lane 4: the mixture of probe and signaling DNA; lane 5: 10 ppm SWNT only (control); lane 6: the mixture of probe DNA and SWNT; lane 7: SDS (negative control). The arrow in lane 4 indicates no dimerization occurred when two DNAs were mixed. The identical FTIR spectra of (b) the probe DNA only and (c) the mixture of probe and signaling DNA also indicated no side interaction between DNAs.

μg/mL of SWNTs (data not shown). This similar trend has already been demonstrated, where the lower concentrations of different nanoparticles (e.g., Ludox silica, MWNTs, gold) exhibited a significant decrease of negative charge. This decrease of negative charge was due to the significant dilution of sample and the zeta potential signal was attributed to noise in the sample.33 The samples with concentrations less than 1 μg/mL are thus considered to be below the detection limit of the instrument. It is important to note that preliminary samples (data not shown) were made where phosphate buffer (0.1 M, pH 7.4) was used as the dispersant, but the zeta potential was unable to be measured potentially because of the significant differences in electrical conductivity in the phosphate buffer at high phosphate concentration. Quantification of SWNTs by the CNT-capture method. Prior to SWNT quantification demonstration by the CNTcapture method, optimum incubation duration and buffer compositions (pH and SDS) were determined and presented in Supporting Information Figure S1 (Detailed results are described in the SI). Overall the incubation time and buffer compositions determined for the CNT-capture method were to be at 30 min incubation time, 10−7 % SDS in the phosphate buffer with pH 7.9, respectively. The determined parameters were applied for the proposed CNT-capture method, where ssDNA trapped the SWNTs in an aqueous solution. Graphene was used as a planar analog to SWNTs to assess the specificity of the detection method for SWNTs since graphene exhibits a planar structure in contrast to the cylindrical structure of SWNTs. As a result of quantification, the R2 over the concentration range of 0.05−10 μg/mL was calculated to be 0.90 for SWNTs as observed in Figure 5a. In contrast, graphene as a planar analog to SWNTs showed a very low R2 value, suggesting that the presented CNT-capture method is very specific to cylindrical structures such as SWNTs, and not for planar structures. As previously mentioned, this observation can be attributed to a stronger interaction between ssDNA and SWNTs; even though graphene has been reported to bind to ssDNA by π stacking interactions.34,35 We hypothesize that the

concluded that there was no dimerization which occurred between the ssDNA during incubation. Differential Dispersion of SWNTs in the Presence or Absence of ssDNA. Another critical aspect that can hinder SWNT-DNA binding is the interaction between SWNTs (i.e., aggregation). The zeta potential was measured to evaluate the effect of ssDNA in the dispersion of SWNTs in our system. The findings showed that in the presence of ssDNA, the zeta potential becomes more negative (i.e., stable) in a range of SWNTs (0.1−100 μg/mL), resulting in a higher dispersion of the SWNTs (Figure 4). The result indicated that there was no significant aggregation among SWNTs during incubation for CNT-capture method. Interestingly, the results showed a concentration dependent effect on the zeta potential for the lower concentrations of SWNTs. There was a significant decrease of negative charge in both samples observed in the less than 1 μg/mL concentration range, even more evident in the concentrations of 0.001−0.01

Figure 4. Investigation of SWNT−SWNT interaction (potential aggregation) by zeta potential measurements at a range of concentrations of SWNT with and without DNA. Error bars indicate standard error of the mean (SEM) of duplicate samples. 497

dx.doi.org/10.1021/es303671u | Environ. Sci. Technol. 2013, 47, 493−501

Environmental Science & Technology

Article

0.91 (Figure 6d). However the quantification range of SWNTs at near-infrared spectroscopy was about 50−250 μg/mL, which are higher than the predicted occurrence (ranging from 0.05 to 0.15 μg/mL) of CNTs reported in the sludge of wastewater treatment plants.39 This indicates that near-infrared spectroscopy is not feasible to measure the potential environmental concentration of CNTs due to the high detection limit. Besides the high detection limit, the major limitation of spectroscopy methods is the inability to distinguish between different sample components.18 In other words the methods lack specificity. The third conventional technique (UV-vis spectroscopy) for SWNT detection tested specificity using a mixed sample with SWNTs and graphene. The spectrum of the mixture of SWNT and graphene samples was assessed and it was found that the peak absorbance for both types of carbon nanoparticles was 284 nm (see the arrows in Figure 6e). It was interesting to note that the spectrum of graphene showed significantly lower absorbance values in contrast to SWNTs of the same concentration which can be attributed to the difference of morphology between a cylinder and a planar structure. For the calibration curve of SWNT only samples the calculated R2 was 0.98, which was a good fit in the range of 0.1−100 μg/mL SWNTs (Figure 6f). However, in the sample where SWNTs and graphene were mixed as a 1:1 ratio at the same concentration, a lower absorbance was observed compared to the SWNT only data (Figure 6f). It is important to note that both samples contain the same amount of SWNTs; however, the absorbance of the mixed sample is lower, presumably due to the presence of an interference material (i.e., graphene), which may hinder light absorption by the SWNTs. Overall, based on the effect of graphene on absorption by the SWNTs, it appears that UV-vis spectroscopy does not have the ability to distinguish SWNTs in presence of different carbon nanomaterials. The analytical results of each method for SWNT quantification are shown in Table 1. It can be observed in Table 1 the comparison of the different methods was assessed based on: range of quantification, equation of quantification, selectivity, and R2 for each sample. As compared to three potential CNT detection methods employed above, the proposed CNT-capture method showed the highest sensitivity (limit of quantification = 0.05 μg/mL SWNTs) as well as high selectivity with respect to graphene. The CNT-capture method also has the potential to quantify at higher SWNT concentrations than the concentration ranges used in the presented results. However, the possible limiting factor for higher SWNT concentration is the amount of ssDNA immobilized to magnetic spheres. In other words, the range of quantification for SWNT can be increased by further optimization. In summary, the proposed CNT-capture method using magnetic spheres and DNA demonstrated the best specificity and appropriate sensitivity to quantify SWNTs in aqueous solutions. This study is a first step toward developing an adequate CNT detection method. Actual water samples would likely involve aggregated CNT bundle, therefore further experimentation is necessary to demonstrate the ability to analyze these samples. We anticipate this development would contribute significantly to the detection and quantification of low concentrations of CNTs in aquatic environments. This technique could potentially play a major role in the pursuit to study the impact and implications of nanomaterials in the environment. This methodology can become available for field

Figure 5. (a) SWNT quantification by the proposed CNT-capture method. Five fluorescence readings were taken for each sample. Error bars indicate SEM of five fluorescence readings. Graphene was used as a planar analog to SWNTs to assess the specificity of the CNT-capture method. FE-SEM images indicate structural difference of target nanoparticles: (b) cylindrical structure of SWNT and (c) the stacked sheets of graphene.

cylindrical structure of SWNTs allow for a larger interaction area and consequently stronger binding. The difference of the molecular structure between SWNTs and graphene was shown in FE-SEM images in Figure 5b and c. Figure 5b shows the cylindrical structure of SWNTs and Figure 5c showed the stacked sheets of graphene. Comparison to Other Potential Methods for SWNTs. Several methods that are theoretically feasible for quantification of CNTs were evaluated: Raman spectroscopy, near-infrared spectroscopy, and UV-vis spectroscopy. Raman spectroscopy was used to detect CNTs and the corresponding Raman peaks are seen in Figure 6a. The G-band around 1600 cm−1 (i.e., the representative of sp2 hybridization of SWNTs, indicated by the arrow in Figure 6a) was selected to make a calibration curve of a range of SWNT concentration as shown in Figure 6b. The calculated R2 (0.18) of Raman spectroscopy was very low, suggesting that this method is inadequate to quantify SWNTs. In other words, the Raman was not able to perform a quantitative detection of SWNTs in a range of concentrations. Similarly, Raman spectroscopy has been used to detect the presence of CNTs in cells,36 daphnids,37 and amphibian’s intestines;38 however, they are not quantitative analyses of the CNTs present in the samples. Near infrared spectroscopy was used to evaluate the detection capabilities for SWNTs. The spectrum of SWNTs in the near-infrared range demonstrated several absorbance peaks (Figure 6c). Each peak at various wavelengths was evaluated by making a calibration curve of a range of SWNT concentrations, where the R2 was calculated (data not shown). After the evaluation of the different peaks and R2, it was found that the wavelength that exhibited the highest R2 was at 1297 nm (depicted by the arrow in Figure 6c). The calibration curve at 1297 nm wavelength gave an R2 = 498

dx.doi.org/10.1021/es303671u | Environ. Sci. Technol. 2013, 47, 493−501

Environmental Science & Technology

Article

Figure 6. CNT quantification by potential other methods: (a), (b) Raman spectroscopy; (c), (d) Near infrared spectrometry; (e), (f) UV-vis spectrometry. Samples were made in duplicate for each method. *Graphene + SWNT comprised of SWNT and graphene at the same concentration indicated in x-axis.

studies of CNT contamination, and may also be adopted by

Table 1. Comparison of Detection Methods for CNT method Raman spectroscopy NIR spectroscopy UV-vis spectroscopy CNT-capture methoda a

range of quantification, CNT, x (μg/mL) 10−250 50−250 0.1−100 0.05−10

various agencies to facilitate policy making and product equation of quantification y (intensity) = 37 x + 11002 y (ODb) = 0.024 x − 0.600 y (OD) = 0.016 x + 0.090 y (fluorescence) = 0.161 x + 0.984

R2

CNT selectivity

regulation, as well as, disposal of nanomaterials including

0.18

yes

CNTs. Elucidation of potential interferences by nonspecific

0.91

no

0.98

no

0.90

yes

binders and environmental factors should be considered in detail to further enhance the capabilities of the proposed method.



ASSOCIATED CONTENT

S Supporting Information *

b

The method was developed in this study. OD stands for optical density of absorbance.

This information is available free of charge via the Internet at http://pubs.acs.org/. 499

dx.doi.org/10.1021/es303671u | Environ. Sci. Technol. 2013, 47, 493−501

Environmental Science & Technology



Article

(16) Cherukuri, P.; Bachilo, S. M.; Litovsky, S. H.; Weisman, R. B. Near-Infrared fluorescence microscopy of single-walled carbon nanotubes in phagocytic cells. J. Am. Chem. Soc. 2004, 126, 15638−15639. (17) Li, Z. F.; Luo, G. H.; Zhou, W. P.; Wei, F.; Xiang, R.; Liu, Y. P. The quantitative characterization of the concentration and diseprsion of multi-walled carbon nanotubes in suspension by spectrophotometry. Nanotechnology 2006, 17, 3692−3698. (18) Petersen, E. J.; Zhang, L.; Mattison, N. T.; O’Carroll, D. M.; Whelton, A. J.; Uddin, N.; Nguyen, T.; Huang, Q.; Henry, T. B.; Holbrook, R. D.; Chen, K. L. Potential release pathways, environmental fate, and ecological risks of carbon nanotubes. Environ. Sci. Technol. 2011, 45, 9837−9856. (19) Girifalco, L. A.; Hodak, M.; Lee, R. S. Carbon nanotubes, buckyballs, ropes, and a universal graphitic potential. Phys. Rev. B 2000, 62, 13104−13110. (20) Zheng, M.; Jagota, A.; Strano, M. S.; Santos, A. P.; Barone, P.; Chou, S. G.; Diner, B. A.; Dresselhaus, M. S.; McLean, R. S.; Onoa, G. B.; Samsonidze, G. G.; Semke, E. D.; Usrey, M.; Walls, D. J. Structurebased carbon nanotube sorting by sequence-dependent DNA assembly. Science 2003, 302, 1545−1548. (21) White, B.; Banerjee, S.; O’Brien, S.; Turro, N. J.; Herman, I. P. Zeta-potential measurements of surfactant-wrapped individual singlewalled carbon nanotubes. J. Phys. Chem. C 2007, 111, 13684−13690. (22) Kim, G. Y.; Son, A. Development and characterization of a magnetic bead-quantum dot nanoparticles based assay capable of Escherichia coli O157:H7 quantification. Anal. Chim. Acta 2010, 677, 90−96. (23) Kim, G. Y.; Wang, X.; Ahn, H.; Son, A. Gene quantification by the NanoGene assay is resistant to inhibition by humic acids. Environ. Sci. Technol. 2011, 45, 8873−8880. (24) Zheng, M.; Jagota, A.; Semke, E. D.; Diner, B. A.; McLean, R. S.; Lustig, S. R.; Richardson, R. E.; Tassi, N. G. DNA-assisted dispersion and separation of carbon nanotubes. Nat. Mater. 2003, 2, 338−342. (25) Tan, Y.; Resasco, D. E. Dispersion of single-walled carbon nanotubes of narrow diameter distribution. J. Phys. Chem. B 2005, 109, 14454−14460. (26) Nakamura, T.; T. Ohana, T.; Ishihara, M.; Hasegawa, M.; Koga, Y. Chemical modification of single-walled carbon nanotubes with sulfur-containing functionalities. Diam. Relat. Mater. 2006, 16, 1091− 1094. (27) Martínez-Rubí, Y.; Guan, J.; Lin, S.; Scriver, C.; Sturgeon, R. E.; Simard, B. Rapid and controllable covalent functionalization of singlewalled carbon nanotubes at room temperature. Chem. Commun. 2007, 5146−5148. (28) Le-Tien, C.; Lafortune, R.; Shareck, F.; Lacroix, M. DNA analysis of a radiotolerant bacterium Pantoea agglomerans by FT-IR spectroscopy. Talanta 2007, 71, 1969−1975. (29) Yarotski, D. A.; Kilina, S. V.; Talin, A. A.; Tretiak, S.; Prezhdo, O. V.; Balatsky, A. V.; Taylor, A. J. Scanning tunneling microscopy of DNA-wrapped carbon nanotubes. Nano Lett. 2009, 9, 12−17. (30) Petrovykh, D. Y.; Kimura-Suda, H.; Whitman, L. J.; Tarlov, M. J. Quantitative analysis and characterization of DNA immobilized on gold. J. Am. Chem. Soc. 2003, 125, 5219−5226. (31) Hu, C.; Zhang, Y.; Bao, G.; Zhang, Y.; Liu, M.; Wang, Z. L. DNA functionalized single-walled carbon nanotubes for electrochemical detection. J. Phys. Chem. B. 2005, 109, 20072−20076. (32) Jovanović, S. P.; Marković, Z. M.; Kleut, D. N.; Romcević, N. Z.; Trajković, V. S.; Dramićanin, M. D.; Todorović Marković, B. M. A novel method for the functionalization of gamma-irradiated single wall carbon nanotubes with DNA. Nanotechnology 2009, 20, 445602. (33) Tantra, R.; Schulze, P.; Quincey, P. Effect of nanoparticle concentration on zeta-potential measurement results and reproducibility. Particuology 2010, 8, 279−285. (34) He, S.; Song, B.; Li, D.; Zhu, C.; Qi, W.; Wen, Y.; Wang, L.; Song, S.; Fang, H.; Fan, C. A graphene nanoprobe for rapid, sensitive, and multicolor fluorescent DNA analysis. Adv. Funct. Mater. 2010, 20 (3), 453−459.

AUTHOR INFORMATION

Corresponding Author

*Phone. +1 (334) 844-6260; fax. +1 (334) 844-6290; e-mail. [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS



REFERENCES

We appreciate receiving the permission for use of the ATRFTIR (Dr. Virginia Davis), and zeta-sizer (Dr. Yucheng Feng and Dr. Dongye Zhao) at Auburn University. This work was supported by National Science Foundation (CAREER award #1054768).

(1) Baughman, R. H.; Zakhidov, A. A.; de Heer, W. A. Carbon nanotubes-the route toward applications. Science 2002, 297, 787−792. (2) Harutyunyan, A. R.; Pradhan, B. K.; Surnanasekera, G. U.; Korobko, E. Y.; Kuznetsov, A. A. Carbon nanotubes for medical applications. Eur. Cells Mater. 2002, 3, 84−87. (3) O’Connell, M. J., Carbon Nanotubes: Properties and Applications; CRC Press: Boca Raton, FL, 2006. (4) Valcarcel, M.; Simonet, B. M.; Cardenas, S.; Suarez, B. Present and future applications of carbon nanotubes to analytical science. Anal. Bioanal. Chem. 2005, 382, 1783−1790. (5) Ball, P. Roll up for the revolution. Nature 2001, 414, 142−144. (6) Roco, M. C. Environmentally responsible development of nanotechnology. Environ. Sci. Technol. 2005, 39, 106A−112A. (7) Iijima, S.; Ichihashi, T. Single-shell carbon nanotubes of 1-nm diameter. Nature 1993, 363, 603−605. (8) Lam, C.; James, J. T.; McCluskey, R.; Arepalli, S.; Hunter, R. L. A review of carbon nanotube toxicity and assessment of potential occupational and environmental health risks. Crit. Rev.Toxicol. 2006, 36, 189−217. (9) Pope, C. A. I.; Burnett, R. T.; Thurston, G. D.; Thun, M. J.; Calle, E. E.; Krewski, D.; Godleski, J. J. Cardiovascular mortality and longterm exposure to particulate air pollution: Epidemiological evidence of general pathophysiological pathways of disease. Circulation 2004, 109, 71−77. (10) Pacurari, M.; Yin, X. J.; Zhao, J.; Ding, M.; Leonard, S. S.; Schwegler-Berry, D.; Ducatman, B. S.; Sbarra, D.; Hoover, M. D.; Castranova, V.; Vallyathan, V. Raw sigle-wall carbon nanotubes induce oxidative stress and activate MAPKs, AP-1, NF-kB, and Akt in normal and malignant human mesothelial cells. Environ. Health Perspect. 2008, 116, 1211−1217. (11) Lam, C. W.; James, J. T.; McCluskey, R.; Hunter, R. L. Pulmonary toxicity of single-wall carbon nanotubes in mice 7 and 90 days after intratracheal instillation. Toxicol. Sci. 2004, 77, 126−134. (12) Huczko, A.; Lange, H.; Całko, E.; Grubek-Jaworska, H.; Droszcz, P. Physiological test of carbon nanotubes: Are they asbestoslike? Fullerene Sci. Technol. 2001, 9, 251−254. (13) Shvedova, A. A.; Castranova, V.; Kisin, E. R.; Schwegler-Berry, D.; Murray, A. R.; Gandelsman, V. Z.; Maynard, A.; P, B. Exposure to carbon nanotube material: Assessment of nanotube cytotoxicity using human keratinocyte cells. J. Toxicol. Environ. Health A 2003, 66, 1909− 1926. (14) Poland, C. A.; Duffin, R.; Kinloch, I.; Maynard, A.; Wallace, W. A. H.; Seaton, A.; Stone, V.; Brown, S.; MacNee, W.; Donaldson, K. Carbon nanotubes introduced into the abdominal cavity of mice show asbestos-like pathogenicity in a pilot study. Nat. Nanotechnol. 2008, 3, 423−428. (15) O’Connell, M. J.; Bachilo, S. M.; Huffman, C. B.; Moore, V. C.; Strano, M. S.; Haroz, E. H.; Rialon, K. L.; Boul, P. J.; Noon, W. H.; Kittrell, C.; Ma, J.; Hauge, R. H.; Weisman, R. B.; Smalley, R. E. Band gap fluorescence from individual single-walled carbon nanotubes. Science 2002, 297, 593−596. 500

dx.doi.org/10.1021/es303671u | Environ. Sci. Technol. 2013, 47, 493−501

Environmental Science & Technology

Article

(35) Husale, B. S.; Sahoo, S.; Radenovic, A.; Traversi, F.; Annibale, P.; Kis, A. ssDNA binding reveals the atomic structure of graphene. Langmuir 2010, 26 (23), 18078−18082. (36) Romero, G.; Rojas, E.; Estrela-Lopis, I.; Donath, E.; Moya, S. E. Spontaneous confocal Raman microscopy–a tool to study the uptake of nanoparticles and carbon nanotubes into cells. Nanoscale Res. Lett. 2011, 6, 429. (37) Roberts, A. P.; Mount, A. S.; Seda, B.; Souther, J.; Qiao, R.; Lin, S.; Ke, P. C.; Rao, A. M.; Klaine, S. J. In vivo biomodification of lipidcoated carbon nanotubes by Daphnia magna. Environ. Sci. Technol. 2007, 41, 3025−3029. (38) Mouchet, F.; Landois, P.; Sarremejean, E.; Bernard, G.; Puech, P.; Pinelli, E.; Flahaut, E.; Gauthier, L. Characterisation and in vivo ecotoxicity evaluation of double-wall carbon nanotubes in larvae of the amphibian Xenopus laevis. Aquat. Toxicol. 2008, 87, 127−137. (39) Gottschalk, F.; Sonderer, T.; Scholz, R. W.; Nowack, B. Modeled environmental concentrations of engineered nanomaterials (TiO(2), ZnO, Ag, CNT, Fullerenes) for different regions. Environ. Sci. Technol. 2009, 43, 9216−9222.

501

dx.doi.org/10.1021/es303671u | Environ. Sci. Technol. 2013, 47, 493−501