Rigid versus Flexible Ligands on Carbon Nanotubes for the

However, an excessive level of this trace element in the body may result in ..... the combined spectroscopic and electrical behavior of bare SWNTs (Fi...
1 downloads 0 Views 460KB Size
Download PDF
Article pubs.acs.org/Macromolecules

Rigid versus Flexible Ligands on Carbon Nanotubes for the Enhanced Sensitivity of Cobalt Ions Pingping Gou,†,‡ Nadine D. Kraut,† Ian Matthew Feigel,† and Alexander Star*,†,‡ †

Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260, United States National Energy Technology Laboratory, U.S. Department of Energy, Pittsburgh, Pennsylvania 15236, United States

‡

S Supporting Information *

ABSTRACT: Carbon nanotubes have shown great promise in the fabrication of ultracompact and highly sensitive chemical and biological sensors. Additional chemical functionalization schemes can controllably improve selectivity of the carbon nanotube-based sensors; however, the exact transduction mechanism is still under debate. In this article we detail the synthesis and selective response of single-walled carbon nanotubes (SWNTs) functionalized with polyazomethine (PAM) polymer toward the application of a specific trace metal ion detector. The response of the polymer system was compared to shape persistent macrocycle (MAC) comprised of identical ion coordination ligands. While ion detection with rigid MAC/SWNT chemiresistor was comparable to bare SWNT, flexible PAM offers significant SWNT signal amplification, allowing for picomolar detection of Co2+ ions with both selectivity and a fast response. We hypothesized that rearrangement of the flexible PAM on the SWNT network is a sensing mechanism which allows for ultrasensitive detection of metal ions. The electron transfer and polymer rearrangement on the SWNT were studied by a combination of optical spectroscopy and electrical measurementsultimately allowing for a better understanding of fundamental mechanisms that prompt device response.

■

INTRODUCTION

Metal ions are potential toxins in the environment as well as in the body. Even metals essential for human health can lead to adverse effects if accumulation becomes too high. For example, cobalt plays an essential role in the metabolism of ions and formation of hemoglobin; also, it is a key constituent of vitamin B12. The daily requirement for cobalt has been estimated to be up to 40 μg. However, an excessive level of this trace element in the body may result in deafness, blindness, and hypothyroidism to name a few.14 CNT-based sensors have been successful at detecting many metal ions, including Zn2+, Pb2+, Hg2+, Cu2+, Ni2+, As3+, and Cd2+ in nanomolar to micromolar concentrations.18−22 This work presents an ultrasensitive CNT-based electrical conductivity device for the detection of Co2+. Selective electrochemical sensing has been achieved through the functionalization of the CNT surface with enzymes and other molecules that have specific interactions with the intended target.1−4 Noncovalent functionalization is ideal since this method does not disrupt the conjugation of the graphitic structure that makes up a carbon nanotube and gives the material its electronic properties.23−32 A commonly used technique for this noncovalent attachment is to take advantage of van der Waals forces, specifically π−π stacking. This strategy has been used for the attachment of cyclodextrin moieties,31 nanoparticles,26 polymer,27 lysozyme,25 and glycolipids23 to

Many useful and sensitive techniques currently exist for chemical analysis and recognition, though the push to create even smaller and more sensitive systems is always upon us. Advances in the field of nanotechnology have helped to make the possibility of device miniaturization within reach. Carbon nanotubes, CNTs, have shown great potential for chemical sensing on a miniaturized scale.1−4 Current research efforts largely focus on real-time, in-situ monitoring of environmental and biological analytes of interest.5−9 For example, detection of the persistent organic pollutant 2,4,5-trichlorobiphenyl (TCB) by a selective interaction between TCB and cyclodextrin decorated carbon nanotubes has been recently demonstrated for use in real-time environmental monitoring.10 The ability to monitor metabolites and other analytes within the body and brain is being explored through the investigation of implantable biosensors.11 An implanted nanodevice would be able to continuously monitor chemical levels within the body without the need for the extraction and lengthy preparation typically used with a bodily sample.11−13 Such a device would allow the early detection of disease through precursor sensing or could also monitor other implanted foreign objects such as a prosthetic or an aesthetic enhancement that could leak a potential toxin into the body over time.14 Also, implantable devices could quickly measure bioaccumulation chemicals in fish and wildlife without having to kill and process the animals of interest.15−17 © 2013 American Chemical Society

Received: January 17, 2013 Revised: January 22, 2013 Published: February 5, 2013 1376

dx.doi.org/10.1021/ma400113m | Macromolecules 2013, 46, 1376−1383

Macromolecules

Article

CNTs. In this paper we use the π−π stacking interactions to attach two different forms of Co2+ chelating molecular ligands to single-walled carbon nanotubes (SWNT). A rigid, shape-persistent macrocycle, salophen macrocycle (MAC), has been previously reported for its coordination with metal ions.33,34 In this work, we designed and synthesized a linear version of MAC, termed polyazomethine (PAM), which offers similar metal chelating properties while having significant morphology differences, specifically it is open-chain and flexible. The π-conjugated backbone of these molecules make them ideal ligand systems for studying with CNT devices as they offer a great advantage in noncovalent functionalization.23−32 The flexible nature of the PAM ligand in comparison to the rigid MAC ligand offers new insight into the electronic sensing properties of functionalized CNTs. Previous research on CNTbased sensors has primarily focused on the presence or absence of the target analyte. The research presented here brings to light a difference in sensing capabilities due to additional conformational changes that occur on the CNT side walls. Most recently, Choi et al. have electronically measured the conformational changes of a single lysozyme tethered to a CNT due to its interaction with peptidoglycan.25 While the CNT device could monitor the lysozyme conformational change, this change occurs at a distance from the CNT side wall to allow room for interaction with the peptidoglycan. We show how rearrangement of noncovalently attached ligands, due to metal ion complexation, leads to different CNT sensor signals. The PAM ligands must undergo conformational change to bind a single Co2+ ion and in turn elicit an optoelectronic response. The metal ion binding event occurs with the rigid MAC ligand, but since this ligand does not require cooperation with another ligand, there is little to no rearrangement upon Co2+ chelation. Results show that the flexibility of PAM contributes to a more sensitive response to Co2+ over a larger dynamic range in comparison to MAC.

(3) following the reported procedure.33 Polyazomethine (PAM) polymer, an open-chain compound analogous to MAC, was synthesized from the same dialdehyde 1 but using 1,4-dioctyloxy-2,5-phenylenediamine (2) as the diamine. PAM polymer, a repeating chain version of MAC, was obtained as a brown-red powder that was miscible in many organic solvents such as THF, CHCl3, and CH2Cl2. Its good solubility can be partially attributed to the presence of octyloxy groups incorporated in the polymer backbone. The chemical structure of PAM was confirmed by 1H NMR spectroscopy (Supporting Information, Figure S1), FTIR spectroscopy (Figure S2), and elemental analysis. Furthermore, UV−vis absorption spectroscopy was utilized to examine electron transitions in PAM both in solution and as a solid film (Figure S3). The HOMO− LUMO difference of PAM can be estimated to be ca. 2.1 eV, determined from the intercept of the normalized absorption and emission spectra (Figure S4). Metal Ion Coordination. In order to elucidate the mechanism of metal ion coordination with MAC and PAM molecules, optical titrations were performed in order to determine the number of metal ions that complex with each ligand. It has been previously demonstrated that the salophen macrocycle can form a stable complex with three transition metal ions, and its triangular conformation remains largely unchanged after the complexation.34 On the other hand, PAM has a more flexible backbone which can undergo conformational changes and can chelate metal ions using different chains (macromolecules). Such interchain coordination of metal ions should result in a cross-linked polymer network, which in turn should result in a change in optoelectronic properties of the polymer because of improved orbital overlap of π-conjugated chains.35,36 The assumed complexation modes were investigated by UV−vis spectroscopy and photoluminescence (PL) emission spectroscopy (Figure 1). Addition of Co2+ ions to a PAM solution in THF (10−4 M) lead to an ∼140 nm red-shift of the polymer main peak, indicating that the resulting polymer backbone has a greater π-electron delocalization and a reduced band gap (Figure 1A). Assembly of π-conjugated PAM polymer chains through the metal coordination is assumed to be responsible for the observed change in the optoelectronic properties. The inset of Figure 1A displays the nonlinear dependence of the absorption intensity on the concentration of Co2+ at λmax (631 nm) with saturation occurring at a metal ion to PAM ratio of 0.3. These results suggest the formation of the previously discussed cross-linked polymer network upon metal complexation. Emission spectroscopy was also utilized to study metal complexation, specifically fluorescence quenching was observed with the addition of Co2+ ions upon PL titration (Figure 1B). The quenching may arise from energy transfer in interstrand interactions. The quenching is saturated with a metal ion to polymer unit ratio around 0.3, which is consistent with UV−vis titration. In contrast to PAM, UV−vis spectra recorded during metal complexation of MAC (Figure 1C) shows only a modest red-shift (∼30 nm) upon complexation. Additionally, complexation stoichiometry of three cobalt ions with one MAC molecule is in good agreement with the literature results.33,34 The relatively smaller red-shift can be attributed to the rigid structure largely unaffected by complexation compared to the large conformational change of PAM polymer. The observed change in UV−vis absorption spectra of PAM upon complexation with transition metals can be applied toward the design of fluorescent and colorimetric sensors.

■

RESULTS AND DISCUSSION Synthesis and Characterization. Salophen macrocycle (MAC) was prepared (Scheme 1) by polycondensation of 3,6diformylcatechol (1) and 1,2-dioctyloxy-4,5-phenylenediamine

Scheme 1. Synthesis of Polyazomethine (PAM) Polymer and Salophen Macrocycle (MAC)

1377

dx.doi.org/10.1021/ma400113m | Macromolecules 2013, 46, 1376−1383

Macromolecules

Article

Figure 1. (A) UV−vis titration of PAM polymer with CoCl2. The inset depicts the intensity of PAM absorption at 630 nm versus the relative concentration of Co2+ ions and reveals that ca. 0.33 Co2+ ions complex with each polymer repeat unit. (B) Photoluminescence (PL) emission spectra (at 500 nm excitation) upon titration of PAM with CoCl2. The inset depicts the dependence of the PAM emission at 621 nm on the relative Co2+ concentration. (C) UV−vis titration of MAC with CoCl2. Inset: the titration curve is consistent with the presented structure of the complex with 3 equiv of Co2+ ions per macrocycle. (D) UV−vis−NIR absorption of PAM film (red), after treatment with CoCl2 (10−3 M in MeCN) (blue) and then after dipping into MeCN:ethylenediamine (9:1) solution (purple) to remove Co2+ ions. Inset: optical image of a PAM film with three drops of Co2+ ions along its diagonal distinctly showing the red to blue color change. (E) Representation of MAC (left) and PAM (right) upon complexation with Co2+ ions.

While Co2+ solution has a weak blue color hardly detected by a naked eye, when added to the deep red PAM solution, the entire mixture becomes a very intense blue color (Figure 2). To study this effect spectroscopically, a polymer film was prepared by spin-coating a solution of PAM in CHCl3 onto a quartz plate. Drop-casting of 10 μL of 10−3 M Co2+ solution in MeCN onto the polymer film led to 100 nm red-shift in UV−vis absorption of the film and an obvious color change from red to blue (Figure 1D). The color and maximum wavelength both recover upon treatment with MeCN/ethylenediamine (9:1). Ethylenediamine strongly chelates Co2+ ions breaking the crosslinked network. The decrease in absorbance of the film after treatment arises from some film dissolution in MeCN. The inset in Figure 1D is a photograph of the PAM film after addition of Co2+ ions. The proposed structures of the ligands complexed with Co2+ ions are given in Figure 1E. The three binding sites of MAC are open, and complexation does not affect the structure or orientation of the ligand; however, in

Figure 2. Photograph of vials with PAM in THF (orange color), after addition of Co2+ (blue color), PAM/SWNT in THF (red-brown color), bare SWNT in THF (SWNT precipitation is visible after 24 h), and PAM/SWNT with Co2+ (PAM/SWNT precipitate immediately after addition of Co2+).

order for PAM to fully chelate the metal ions, numerous chains are necessary. 1378

dx.doi.org/10.1021/ma400113m | Macromolecules 2013, 46, 1376−1383

Macromolecules

Article

Figure 3. (A) Schematic drawing of a carbon nanotube/polymer device used in this study. Single-walled carbon nanotubes (SWNT) and a noncovalently functionalized ligand deposited on a quartz plate forming a transparent yet conducting thin film. (B) Photograph of a quartz plate with two Al foil electrodes connected to the film using silver paint. Magnification of the PAM/SWNT via scanning electron microscopy (SEM) reveals the morphology of the nanotube/polymer film. (C) UV−vis−NIR absorption spectra of SWNT (blue), PAM (red), and PAM/SWNT (green). Inset: density of states (DOS) diagram showing the electron transitions in SWNTs. (D) Optical response of PAM/SWNT to Co2+ ions. Specifically, the S11 SWNT band increases upon addition of 10−9 M Co2+ ions. (E) Optical response of MAC/SWNT to 10−5 M Co2+ ions where the change in the S11 transition band is significantly less. (F) Electrical conductance (G) of PAM/SWNT film in air and in varying concentrations (M) of Co2+ ions. (G) Real-time conductance measurement of MAC/SWNT to varying concentrations of Co2+ ions. (H) Relative response plot comparing Co2+ ion sensitivity of PAM/SWNT to MAC/SWNT. PAM/SWNT shows the greatest response over the largest dynamic range, and additionally, the response scales with the logarithm of the concentration of ions.

In a control experiment, a thin film of PAM was exposed to acidic (HCl) and basic (NH3) vapors (Figure S5). It can be observed that the protonation and deprotonation of the PAM thin film upon treatment with HCl and NH3 resulted in a reversible color change with a significantly smaller shift (∼20 nm). Noncovalent Interactions between Ligands and SWNTs. An insight into the mechanism of interaction of PAM and SWNTs is related to the stabilization of SWNTs in organic solvent. Sonication of SWNTs in a THF solution containing PAM polymer resulted in a stable SWNT suspension which exhibited no precipitation over several days, while SWNT suspensions in THF alone tends to aggregate and precipitate after 24 h (Figure 2). The stability of the suspension is directly related to interactions between PAM polymer and

SWNTs. The hydrophobic aromatic backbone of PAM polymer forms π−π stacking interactions with SWNT sidewalls, while the octyloxy groups provide solubility for the PAM/SWNT composite in organic solvent. Similar polymeric structures such as poly(m-phenylenevinylene), PmPV, have previously been used to suspend SWNTs in organic solvents.29,30 Comparatively, sonication of SWNTs in THF solution containing MAC resulted in a suspension with a brown-red color, but SWNTs aggregated in 24 h (not shown). This difference in SWNT dispersibility suggests that MAC has weaker interactions with SWNTs than PAM polymer. Addition of Co2+ ions into the PAM/SWNT composite suspension resulted in a sudden precipitation of all SWNTs from the solution, presumably due to cross-linking PAM polymer through Co2+ ions (Figure 2). 1379

dx.doi.org/10.1021/ma400113m | Macromolecules 2013, 46, 1376−1383

Macromolecules

Article

Figure 4. (A) Real-time conductance measurements of bare SWNT for varying concentration of Co2+ ions. (B) UV−vis−NIR absorption spectra of bare SWNT (red) and after addition of 10−3 M Co2+ ions (blue).

polymer onto the SWNT film, the S11 band has a reduced intensity and a red-shift of about 20 nm which can be also attributed to π−π stacking interactions between carbon nanotube networks and polymer backbone. The S22 band was not shifted but did experience a reduction in intensity. The M11 band was masked by the PAM absorption at UV−vis region, so any changes were not evident. The observed reduction in both S11 and S22 suggests that the polymer coating changes the dielectric environment of the SWNTs. Since PAM itself is not a conducting polymer due to its nonplanar polymer backbone, PAM-modified SWNT forms a dielectric layer. Specifically, charge carriers do not flow through the polymer layer, but transfer between polymer layer and SWNT networks upon molecular interactions. Another important observation is that the conductance of the SWNT film is suppressed greatly upon the polymer modification from 628 ± 2 μS for bare SWNT films to 341 ± 2 μS for PAM/SWNT complex (not shown). This observation is consistent with the reduction in both S11 and S22, indicating a dielectric layer formed by PAM attachment. UV−vis−NIR absorption spectra of SWNT films before and after addition of MAC ligand are presented in Figure S7. Co2+ Detection with SWNT Devices. The thin film devices were tested for the detection of metal ions. Ligandmodified SWNT thin films were brought into contact with Co2+ ions, and UV−vis−NIR absorption and electrical conductance data were recorded (Figure 3). Co2+ ions were deposited onto the chip from MeCN and allowed to dry in ambient conditions. The UV−vis−NIR spectra of the ligand/ SWNT thin film before and after addition of Co2+ ions are shown in Figure 3D,E. The complexation of Co2+ ion with the PAM/SWNT thin film results in a small red-shift for PAM signal and increase of the absorption intensity of the nanotube S11 transition (Figure 3D). However, as the concentrations of Co2+ are much lower than those used for solution titration, the color change of the PAM films was difficult to discern by the naked eye. These observations suggest that studying the optical properties alone does not provide an effective means for detecting the metal ions at very low concentrations. In addition, the changes for the S22 transition were miniscule. For the rigid analogue, MAC-modified SWNTs, very little changes in the spectrum were observed upon addition of Co2+ ions, besides a miniscule increase in absorption intensity of the S11 transition upon Co2+ addition (Figure 3E). Taking advantage of the electrical configuration of the detector device, real-time conductance measurements were taken for addition of cobalt ions at varying concentrations (Figure 3F−H). The conductance of PAM/SWNT decreased upon addition of Co2+ ions with concentrations ranging from picomolar to millimolar range (Figure 3F). The response in

Additionally, the color of the polymer solution changed from red to blue, similarly to previously discussed results. The presence of the ligands on the SWNTs was confirmed and characterized by atomic force microscopy (AFM) (Figure S6). The SWNTs showed a marked increase in diameter from ∼1 nm (bare SNWTs) to ten or several tens of nanometers after PAM modification. As a comparison, AFM imaging revealed that the average diameters for MAC adsorbed on SWNT sidewalls ranged approximately from ∼5 to 10 nm. Thus, AFM studies confirmed that the metal complex ligands, PAM and MAC, both associate strongly with SWNTs through noncovalent interactions. Furthermore, thin films of SWNT deposited onto quartz substrates were studied both before and after deposition of PAM polymer. A network of SWNT was deposited onto a quartz plate to produce a transparent yet conductive thin film, to which the ligand was subsequently added (Figure 3A). The SWNT thin films were fabricated as described previously in the literature.37,38 A solution of SWNTs in DMF was spray-cast onto quartz plates via a spray gun. Then electrodes were formed by applying Al metal tape on the edge of the quartz plate and connecting them to the SWNT network with silver paint (Figure 3B). Subsequently, either PAM or MAC was drop-cast from a solution of THF onto the SWNT network. The SWNT solution must be applied to the quartz surface independently of either ligand to ensure the percolation threshold is reached creating SWNT networks for optimal device conductance.39−42 Fabrication of SWNT network device on a transparent surface allowed for combined spectroscopy and electrical measurements of the device.37,38 Scanning electron microscopy (SEM) images of the SWNT thin film after PAM were drop-cast onto SWNT networks are shown in Figure 3B. The absorption spectrum for pristine SWNT clearly depicts the first two semiconducting and metallic electronic transitions, referred to as S11, S22, and M11, respectively. The inset in Figure 3C schematically illustrates these transitions in a density of states diagram. The origin and significance of SWNT transition bands have been heavily studied and reported in detail.43−45 All UV−vis−NIR spectra are normalized at 1400 nm, where no SWNT absorption bands appear. There is one major absorption band observed for PAM polymer with a maximum wavelength value around 535 nm; the SWNT film does not show any prominent peaks in this region (Figure 3C). When PAM polymer was deposited on top of the SWNT film, the major absorption peak was broadened and redshifted by 35 nm (λmax value of 570 nm), presumably as a result of noncovalent π−π interactions between the SWNTs and the π-conjugated PAM backbone. Moreover, the first two semiconducting transitions bands of SWNTs were affected when complexed with PAM polymer. Upon addition of PAM 1380

dx.doi.org/10.1021/ma400113m | Macromolecules 2013, 46, 1376−1383

Macromolecules

Article

SWNT conductance (ΔG/G 0 ) as a function of Co 2+ concentration ([Co2+]) (Figure 3H) scales with the logarithm (base 10) of the concentration of Co2+ ions. It should be recalled here that color changes of PAM film upon Co2+ addition were only visible at relatively high concentrations (10−4 M); thus, they were not observed in this experiment. These results suggest that the PAM/SWNT thin film electrical device shows great potential for future development of SWNTbased metal ion sensor devices. The MAC/SWNT complex conductance measurements revealed that the film conductance decreased with addition of cobalt ions (Figure 3G); however, the change in conductance was significantly less than the PAMbased device, and the response was only linear over the range of 10−6−10−3 M. The conductance of bare SWNTs when exposed to Co 2+ ions increased in real time with increasing concentrations of Co2+ ions (Figure 4). This response is opposite to those observed on PAM/SWNT and MAC/SWNT devices, indicating that chelation of Co2+ with the PAM or MAC ligands is responsible for the decrease in the conductance of the film. Plotting the relative response as a function of the concentration of Co2+ ions of PAM/SWNT and MAC/SWNT together reveals that that sensitivity (magnitude of slope of the curve) is much greater for the PAM/SWNT complex. The conductance change of the PAM/SWNT device is linear with respect to the log [Co2+] over all the ranges tested, while the change of MAC/SWNT is 10 times diminished and has a much narrower dynamic range (10−6−10−3 M). The calculated limits of detection for the PAM/SWNT and MAC/SWNT devices are 10−20 and 10−7 M, respectively. One possible explanation is that the PAM polymer has a large conformational change upon metal ion complexation. The conformational change of the polymer disturbs the carbon nanotube networks through π−π stacking interactions, thus amplifying the electrical signal change. In contrast, the conformation of macrocycle molecule is little affected by metal ion complexation, thus smaller influence on SWNT networks giving rise to its inherently lower sensitivity. PAM complexation has been also studied with Cu2+ ions. Although PAM demonstrates a large absorbance red-shift (∼160 nm) with Co2+ because of the polymer conformational change (vide supra), addition of Cu2+ ions leads presumably to a smaller conformational change with an absorbance red-shift of only ∼40 nm (Figure S8). Therefore, the optical/electrical device was tested for varying concentration of Cu2+ ions in order to further understand the transduction mechanism of the PAM/SWNT device. UV−vis−NIR absorption spectroscopy reveals a decrease in the absorption intensity of the S11 transition band upon addition of 10−3 M Cu2+ ions (Figure S9). At the same time, the film conductivity increased upon increasing concentrations of Cu2+ ions from the picomolar to millimolar range. Presuming that Cu2+ works as an electron acceptor, its addition onto the PAM/SWNT film induces electron withdrawal from SWNT valence band, thus reducing the S11 transition band. Meanwhile, the increased amount of hole carriers results in the enhancement of the film conductance with increasing concentration of copper ions. However, the conformational change of PAM upon coordination of Cu2+ is much smaller than coordination with Co2+, which elucidates the origin of the smaller response, supporting the reasoning behind PAM/SWNT thin film device selectivity to cobalt ions. Proposed Transduction Mechanism for Co2+ Detection. Insight into the transduction mechanism for cobalt ion

detection can be obtained by examining the combined spectroscopic and electrical behavior of bare SWNTs (Figure 4). When control studies with bare SWNT were performed, the response was observed to be reversed compared to PAM functionalized SWNTs. Although the PAM/SWNT conductivity response decreases with increasing Co2+ concentration, the bare SWNT conductivity increases (Figure 4A). Recently, Dionisio et al. demonstrated a similar response reversal due the cavitand receptor functionalization on SWNTs in comparison to bare SWNTs.46 The increased conductance on a bare p-type SWNT upon increasing concentrations of cobalt ions is expected for positively charged ions adsorbed on p-type SWNTs because they increase the concentration of p-type carriers (hole carriers); at the same time, electron density in the SWNT is reduced and thus decreases the intensity of the S11 transition band (Figure 4B). On the contrary, quite the opposite effect was observed for the ligand/SWNT complexes. Specifically, the exposure to Co2+ ions enhanced the S11 transitions and decreased the overall conductance of the device. The increase in absorption intensity of the S11 transition band is indicative of electron donation into the top of SWNT valence band. This would also decrease the number of hole carriers which explains the decrease in the film conductance. Moreover, the conformational changes of PAM polymer upon cobalt coordination deform the SWNT network through π−π stacking interactions and thus amplify conductance signals as a result. This possible explanation is supported by the significant lowering of the detection limit for the PAM based device compared to the rigid MAC. Cyclic voltammetry (CV) experiments were also performed to investigate the electronic mechanism upon Co2+ complexation with the PAM polymer (Figure S10). Upon addition of Co2+ a significant alteration in the CV curves was observed. This change in the electrochemical behavior correlates with a change in the relative donating ability of the polymer complex in comparison to PAM without Co2+. Figure 5 schematically illustrates the difference between the three systems tested for Co2+ ion sensitivity. Co2+ ions attach to

Figure 5. Schematic representation of the proposed mechanism of detection. Upon addition of Co2+ ions to MAC/SWNT the ions fill any available macrocycles, which offers no advantage over bare SWNT with respect to changes in electrical conductance. On the other hand, addition of Co2+ ions to PAM/SWNT will invoke a rearrangement of the polymer to bind with the metal ion resulting in a larger enhancement in signal.

the sidewalls of SWNTs, resulting in changes in electrical and optical properties. Additionally, for MAC/SWNT, Co2+ ions will complex with available macrocycles, resulting in similar changes in the properties of SWNTs as bare SWNTs. However, when PAM is functionalized onto SWNTs, addition of Co2+ ions results in a rearrangement of the polymer on the SWNT 1381

dx.doi.org/10.1021/ma400113m | Macromolecules 2013, 46, 1376−1383

Macromolecules

Article

Synthesis of Polyazomethine (PAM) Polymer. Freshly prepared 1,4-dioctyloxy-2,5-phenylenediamine (109 mg, 0.3 mmol) was added to a solution of 3,6-diformylcatechol (50 mg, 0.30 mmol), LiCl (55 mg, 1.2 mmol), and 4-dimethylaminopyridine (DMAP) (3.7 mg, 0.03 mmol) in a mixture of 3 mL of hexamethylphosphoramide (HMPA) and 3 mL of anhydrous methylpyrrolidone (NMP). The reaction mixture (deep orange red color) was stirred for 48 h under ambient conditions. 23 μL (0.15 mmol) of tert-butylaniline was added to the solution, and the resultant mixture was stirred for an additional 6 h. The resulting precipitate (polymer) was extracted from MeOH (200 mL) and washed thoroughly with MeOH and distilled water. PAM polymer was further purified by washing with MeOH in a Soxhlet extractor for 12 h to afford a brown-orange solid (83 mg, 56%). νCN, 1614 cm−1; νO−C−O 1215 cm−1. 1H NMR (CDCl3) δ: 12.71 (2H, s, OH), 8.63 (2H, s, CHN), 7.48 (4H, d, aromatic CH), 4.03 (4H, m, OCH2), 1.89 (m, CH2), 1.35 (m, CH2), 0.88 (t, 3JHH = 6.3 Hz, CH3). The number-average molecular weight (Mn = 24 600 g/ mol) and polydispersity index (PDI = 1.4) of polymer were determined in THF by gel permeation chromatography (GPC) against polystyrene standards in THF. Anal. Calcd for (C30H42N2O4)n: C, 72.84; H, 8.56; N, 5.66. Found: C, 73.08; H, 8.60; N, 5.71. Device Fabrication. Nanotube devices were fabricated as described previously.37,38 Briefly, SWNTs were suspended in DMF using ultrasonication and then spray-casted onto heated 1 in. × 1 in. fused quartz (SiO2) plates (0.0625 in. thick; Quartz Scientific; reported specific resistance of 10 × 1018 Ω cm−3 at 20 °C) using a commercial air gun (Iwata, Inc.). The plates, which served as the device substrates, were cleaned prior to SWNT deposition with acetone and then subsequently rinsed with water and dried under compressed air. Electrodes were fabricated by placing Al tape and Ag paint (SPI supplies) on top of the SWNT network, leaving a spacing of 1.5 cm to create an unobstructed optical window. One drop of PAM (10 μM) in THF was subsequently deposited onto the SWNT network. For comparison of PAM to its corresponding macrocycle, MAC (1.4 × 10−4 M) was drop-cast onto the network instead of PAM. Metal Ion Interactions. Cobalt chloride (CoCl2) solutions in MeCN were prepared at eight different concentrations in the 10−12− 10−2 M range. A 10 μL drop of the CoCl2 solution was added to the center of the SWNT/PAM film and allowed to evaporate. UV−vis− NIR absorption spectroscopy and electrical conductance measurements were collected for 1000 s. The evaporation of MeCN drop took a few seconds, and the signals were steady within 1000 s.

surface, amplifying changes in electrical and optical properties of SWNTs.

■

CONCLUSION In conclusion, ultrasensitive Co2+ ion detection has been presented based on the noncovalently functionalized PAM/ SWNT optoelectronic device. Detection of Co2+ ions in the picomolar range has been demonstrated with a fast response time and high selectivity. Both spectroscopic and electrical signals agree, and their comparison gives insight into the mechanism of detection. While a visible color change is observed upon Co2+ binding to the devices, this optical detection method is not sufficient at very low analyte concentrations. The low detection limit of this device shows promise for its use as an implantable Co2+ sensor. Specifically, this device could be used to monitor arthroprosthetic cobaltism resulting from prosthetic hip implants that may cause toxic effects due to leeching Co2+ over time.14 In comparing the flexible ligand PAM to the rigid ligand MAC, we learn that the rearrangement of the PAM on the surface of the SWNT plays an important role in the sensitivity of the Co2+ detector. The rigid ligand MAC did not show a response significantly different from the bare SWNT upon metal ion binding. These results support the conclusion that Co2+ interaction alone is not the whole story when considering the electrochemical response in this SWNT device.

■

EXPERIMENTAL SECTION

Materials and Instrumentation. SWNTs (prepared by the arc discharge technique)47 were purchased from Carbon Solutions, Inc. (P2-SWNT). Chemicals were purchased from Sigma-Aldrich and used as received. Solvents were dried, distilled, and stored under argon. All reactions were carried out in a N2 atmosphere unless otherwise stated. 3,6-Diformylcatechol,48 1,2-dioctyloxy-4,5-phenylenediamine,49 and 1,4-dioctyloxy-2,5-phenylenediamine were synthesized following published procedures (see Supporting Information for details). FT-IR spectra were recorded either as thin films on NaCl plates or as KBr pellets on an Avatar 380 Nicolet FT-IR spectrometer. Proton nuclear magnetic resonance (1H NMR) spectra were recorded on a Bruker DPX 300 NMR spectrometer (300 MHz) at 25 °C. Deuterated solvent was used as the lock, and the residual solvent was used as an internal standard. Molecular weights of polymers were determined by using a Waters gel permeation chromatography system equipped with a Water 510 HPLC pump, phenogel columns, and a Waters 410 differential refractometer at a flow rate of 0.4 mL/min. All molecular weights were measured against polystyrene standards in THF. UV−vis−NIR spectroscopic measurements were recorded on a PerkinElmer Lambda 900 spectrophotometer. Additionally, the SWNT network conductance versus time, G(t), was monitored by applying a bias voltage of 50 mV through the network and measuring the resultant current using a Keithley model 2400 source meter interfaced with LabView software. Synthesis of Salophen Macrocycle (MAC). MAC was synthesized following a published procedure.33 1,2-Dioctyloxy-4,5phenylenediamine (1.9 g, 6.0 mmol) was dissolved in 60 mL of 1:1 degassed CHCl3:MeCN under a N2 environment. 3,6-Diformylcatechol (1.0 g, 6.0 mmol) was added, changing the appearance of the solution from colorless to deep red. After heating at 90 °C while refluxing for 2 h, the solution was cooled down to room temperature, yielding red needles of salophen macrocycle, which were isolated on a funnel, washed with chilled MeCN, and dried under vacuum. Yield: 92 mg, 71%. 1H NMR (CDCl3) δ: 8.56 (6H, s, CHN), 7.01 (6H, s, aromatic CH), 6.78 (6H, d, aromatic CH), 4.05 (12H, t, 3JHH = 5.6 Hz, OCH2), 1.88 (12H, m, CH2), 1.52 (12H, m, CH2), 1.32 (24H, m, CH2), 0.91 (t, 3JHH = 6.0 Hz, CH3).

■

ASSOCIATED CONTENT

* Supporting Information S

Additional details about synthesis of polyazomethine (PAM) polymer and its characterization by NMR, FTIR, UV−vis− NIR, and emission spectra as well as response of the polymer to acidic and basic vapors; additional figures investigate interactions of PAM and salophen macrocycle (MAC) with SWNTs, cyclic voltammetry data, sensor reversibility studies, and control experiments using Cu2+ ions. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Tel (412) 624-6493; Fax (412) 624-4027; e-mail astar@pitt. edu. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was performed in support of ongoing research in sensor systems and diagnostics at the National Technology Laboratory (NETL) under RES Contract DE-FE0004000. 1382

dx.doi.org/10.1021/ma400113m | Macromolecules 2013, 46, 1376−1383

Macromolecules

■

Article

(35) Cerrada, P.; Oriol, L.; Pinol, M.; Serrano, J. L. J. Am. Chem. Soc. 1997, 119, 7581−7582. (36) Yang, C. J.; Jenekhe, S. A. Chem. Mater. 1991, 3, 878−887. (37) Kauffman, D. R.; Shade, C. M.; Uh, H.; Petoud, S.; Star, A. Nat. Chem. 2009, 1, 500−506. (38) Kauffman, D. R.; Star, A. J. Phys. Chem. C 2008, 112, 4430− 4434. (39) Bekyarova, E.; Itkis, M. E.; Cabrera, N.; Zhao, B.; Yu, A.; Gao, J.; Haddon, R. C. J. Am. Chem. Soc. 2005, 127, 5990−5995. (40) Fuhrer, M. S.; Nygard, J.; Shih, L.; Forero, M.; Yoon, Y.-G.; Mazzoni, M. S. C.; Choi, H. J.; Ihm, J.; Louie, S. G.; Zettl, A.; McEuen, P. L. Science 2000, 288, 494−497. (41) Hu, L.; Hecht, D. S.; Grüner, G. Nano Lett. 2004, 4, 2513− 2517. (42) Lee, H. S.; Yun, C. H.; Kim, S. K.; Choi, J. H.; Lee, C. J.; Jin, H.J.; Lee, H.; Park, S. J.; Park, M. Appl. Phys. Lett. 2009, 95, 134104/1− 134104/3. (43) Aihara, J.-i. J. Phys. Chem. A 1999, 103, 7487−7495. (44) Bachilo, S. M.; Strano, M. S.; Kittrell, C.; Hauge, R. H.; Smalley, R. E.; Weisman, R. B. Science 2002, 298, 2361−2366. (45) Kataura, H.; Kumazawa, Y.; Maniwa, Y.; Umezu, I.; Suzuki, S.; Ohtsuka, Y.; Achiba, Y. Synth. Met. 1999, 103, 2555−2558. (46) Dionisio, M.; Schnorr, J. M.; Michaelis, V. K.; Griffin, R. G.; Swager, T. M.; Dalcanale, E. J. Am. Chem. Soc. 2012, 134, 6540−6543. (47) Journet, C.; Maser, W. K.; Bernier, P.; Loiseau, A.; de la Chapelle, M. L.; Lefrant, S.; Deniard, P.; Lee, R.; Fischer, J. E. Nature 1997, 388, 756−758. (48) Vickery, E. H.; Pahler, L. F.; Eisenbraun, E. J. J. Org. Chem. 1979, 44, 4444−4446. (49) Kim, D.-H.; Choi, M. J.; Chang, S.-K. Bull. Korean Chem. Soc. 2000, 21, 145−147.

REFERENCES

(1) Jacobs, C. B.; Peairs, M. J.; Venton, B. J. Anal. Chim. Acta 2010, 662, 105−127. (2) Feigel, I. M.; Vedala, H.; Star, A. J. Mater. Chem. 2011, 21, 8940− 8954. (3) Kauffman, D. R.; Star, A. Angew. Chem., Int. Ed. 2008, 47, 6550− 6570. (4) Allen, B. L.; Kichambare, P. D.; Star, A. Adv. Mater. 2007, 19, 1439−1451. (5) Esser, B.; Schnorr, J. M.; Swager, T. M. Angew. Chem., Int. Ed. 2012, 51, 5507−5507. (6) Xiao, N.; Venton, B. J. Anal. Chem. 2012, 84, 7816−7822. (7) Kim, J. P.; Hong, S.; Sim, S. J. J. Nanosci. Nanotechnol. 2011, 11, 4182−4187. (8) Ding, M.; Tang, Y.; Gou, P.; Reber, M. J.; Star, A. Adv. Mater. 2011, 23, 536−540. (9) Takase, M.; Yoneyama, Y.; Murata, M.; Hibi, K.; Ren, H.; Endo, H. Anal. Bioanal. Chem. 2012, 403, 1187−1190. (10) Kong, L.-T.; Wang, J.; Meng, F.-L.; Chen, X.; Jin, Z.; Li, M.-Q.; Liu, J.-H.; Huang, X.-J. J. Mater. Chem. 2011, 21, 11109−11115. (11) Vaddiraju, S.; Tomazos, I.; Burgess, D. J.; Jain, F. C.; Papadimitrakopoulos, F. Biosens. Bioelectron. 2010, 25, 1553−1565. (12) Barone, P. W.; Baik, S.; Heller, D. A.; Strano, M. S. Nat. Mater. 2005, 4, 86−92. (13) Sung, J.; Barone, P. W.; Kong, H.; Strano, M. S. Biomaterials 2009, 30, 622−31. (14) Tower, S. Alaska Med. 2010, 52, 28−32. (15) Hogenboom, A. C.; van Leerdam, J. A.; de Voogt, P. J. Chromatogr., A 2009, 1216, 510−519. (16) Kataoka, H.; Saito, K. J. Pharm. Biomed. Anal. 2011, 54, 926− 950. (17) Speltini, A.; Sturini, M.; Maraschi, F.; Profumo, A.; Albini, A. TrAC, Trends Anal. Chem. 2011, 30, 1337−1350. (18) Brege, J. J.; Gallaway, C.; Barron, A. R. J. Phys. Chem. C 2007, 111, 17812−17820. (19) Crespo, G. A.; Macho, S.; Bobacka, J.; Rius, F. X. Anal. Chem. 2009, 81, 676−681. (20) Forzani, E. S.; Li, X.; Zhang, P.; Tao, N.; Zhang, R.; Amlani, I.; Tsui, R.; Nagahara, L. A. Small 2006, 2, 1283−1291. (21) Kim, T. H.; Lee, J.; Hong, S. J. Phys. Chem. C 2009, 113, 19393−19396. (22) Wanekaya, A. K. Analyst 2011, 136, 4383−4391. (23) Assali, M.; Leal, M. P.; Fernandez, I.; Baati, R.; Mioskowski, C.; Khiar, N. Soft Matter 2009, 5, 948−950. (24) Balasubramanian, K.; Burghard, M. J. Mater. Chem. 2008, 18, 3071−3083. (25) Choi, Y.; Moody, I. S.; Sims, P. C.; Hunt, S. R.; Corso, B. L.; Perez, I.; Weiss, G. A.; Collins, P. G. Science 2012, 335, 319−324. (26) Hu, L.; Zhao, Y.-L.; Ryu, K.; Zhou, C.; Stoddart, J. F.; Grüner, G. Adv. Mater. 2008, 20, 939−946. (27) Imin, P.; Imit, M.; Adronov, A. Macromolecules 2011, 44, 9138− 9145. (28) Liu, S.; Shen, Q.; Cao, Y.; Gan, L.; Wang, Z.; Steigerwald, M. L.; Guo, X. Coord. Chem. Rev. 2010, 254, 1101−1116. (29) Star, A.; Liu, Y.; Grant, K.; Ridvan, L.; Stoddart, J. F.; Steuerman, D. W.; Diehl, M. R.; Boukai, A.; Heath, J. R. Macromolecules 2003, 36, 553−560. (30) Star, A.; Stoddart, J. F.; Steuerman, D.; Diehl, M.; Boukai, A.; Wong, E. W.; Yang, X.; Chung, S.-W.; Choi, H.; Heath, J. R. Angew. Chem., Int. Ed. 2001, 40, 1721−1725. (31) Zhao, Y.-L.; Hu, L.; Stoddart, J. F.; Grüner, G. Adv. Mater. 2008, 20, 1910−1915. (32) Zhao, Y.-L.; Stoddart, J. F. Acc. Chem. Res. 2009, 42, 1161− 1171. (33) Gallant, A. J.; MacLachlan, M. J. Angew. Chem., Int. Ed. 2003, 42, 5307−5310. (34) Leung, A. C. W.; MacLachlan, M. J. J. Mater. Chem. 2007, 17, 1923−1932. 1383

dx.doi.org/10.1021/ma400113m | Macromolecules 2013, 46, 1376−1383