Structure of Polydopamine: A Never-Ending Story? - ACS Publications

Jul 22, 2013 - Faculty of Chemistry and Chemical Engineering, Babes-Bolyai University Cluj-Napoca, Arany Janos Straße 11, 400028 Cluj-Napoca,. Romani...
4 downloads 5 Views 2MB Size
Download PDF
Article pubs.acs.org/Langmuir

Structure of Polydopamine: A Never-Ending Story? Jürgen Liebscher,*,†,‡ Radosław Mrówczyński,†,‡ Holger A. Scheidt,§ Claudiu Filip,† Niculina D. Hădade,∥ Rodica Turcu,† Attila Bende,† and Sebastian Beck‡ †

National Institute of Research and Development for Isotopic and Molecular Technologies, Donath 65-103, RO-400293 Cluj-Napoca, Romania ‡ Department of Chemistry, Humboldt-Universität zu Berlin, Brook-Taylor-Street 2, 12489 Berlin, Germany § Institut für Medizinische Physik und Biophysik, Universität Leipzig, Härtelstraße 16-18, 04107 Leipzig, Germany ∥ Faculty of Chemistry and Chemical Engineering, Babes-Bolyai University Cluj-Napoca, Arany Janos Straße 11, 400028 Cluj-Napoca, Romania S Supporting Information *

ABSTRACT: Polydopamine (PDA) formed by the oxidation of dopamine is an important polymer, in particular, for coating various surfaces. It is composed of dihydroxyindole, indoledione, and dopamine units, which are assumed to be covalently linked. Although PDA has been applied in a manifold way, its structure is still under discussion. Similarities have been observed in melanins/eumelanins as naturally occurring, deeply colored polymer pigments derived from L-DOPA. Recently, an alternative structure was proposed for PDA wherein dihydroxyindoline, indolinedione, and eventually dopamine units are not covalently linked to each other but are held together by hydrogen bonding between oxygen atoms or π stacking. In this study, we show that this structural proposal is very unlikely to occur taking into account unambiguous results obtained by different analytical methods, among them 13C CPPI MAS NMR (cross-polarization polarization−inversion magic angle spinning NMR), 1H MAS NMR (magic angle spinning NMR), and ES-HRMS (electrospray ionization high-resolution mass spectrometry) for the first time in addition to XPS (X-ray photoelectron spectroscopy) and FTIR spectroscopy. The results give rise to a verified structural assignment of PDA wherein dihydroxyindole and indoledione units with different degrees of (un)saturation are covalently linked by C−C bonds between their benzene rings. Furthermore, proof of open-chain (dopamine) monomer units in PDA is provided. Advanced DFT calculations imply the arrangements of several PDA chains preferably by quinone−hydroquinone-type interactions in a parallel or antiparallel manner. From all of these results, a number of hypotheses published before could be experimentally supported or were found to be contradictory, thus leading to a better understanding of the PDA structure.

■

INTRODUCTION Because the catechol motif occurring in DOPA-containing mussel adhesive proteins was found to be responsible for the enormous power of shells for adhesion to all kinds of surfaces, catechol-containing molecules and polymers were applied to coat various materials such as metals, metal oxides, nonmetal oxides, silica, ceramics, polymers, and nanomaterials.1−3 Among the catechol-containing coatings, polydopamine (PDA) has attracted much attention. It can be easily synthesized by exposing dopamine (3,4-dihydroxyphenylethylamine) to air in basic aqueous buffer solutions.1 PDA has been used to coat metals,1 oxides,1 ceramics,1 polymers,4 carbon nanotubes,5 and recently magnetite nanoparticles.6−8 The coating is very resistant and has gained interest for its promising biomedical applications, such as in drug delivery, biosensing, and interfacing with cells.2 PDA allows easily functionalization through reactions with amino- or mercapto-nucleophiles.1 It is assumed that carbonyl or α,ß-unsaturated carbonyl moieties act as electrophilic sites for these types of nucleophiles. In this way, biotin, fluorescent labels, proteins, enzymes, DNA, and also “clickable” azido groups were covalently linked to polydopamine surfaces.6,9−12 Despite the wide application of PDA, it is © 2013 American Chemical Society

amazing that the structure has not yet been fully explored. Numerous different structures have been proposed in the literature, as exemplified in Figure 1.1,13 These proposals include homopolymers of indoles (1, 2) or polymers built up by different indole moieties in various oxidation states (8). The connection between the monomer units was proposed to occur not only via the benzo moiety but also via the pyrrole fragment (e.g., 3−5). It is assumed that dopamine first forms the indole skeleton by oxidative ring closure before the monomer units are connected by dehydrogenative C−C bond formation.1,13 However, primary C−C bond connections of dopamine were also proposed (6 and 7).14,15 Furthermore, PDA structures were deduced wherein the monomer units consist of two fused indole rings (9). Analogies to PDA structures were noticed with eumelanins (synonymous with melanins),16,17 which are pigments occurring naturally in hair or skin. They result from tyrosine or LDOPA as a precursor. Although it was initially thought that Received: May 30, 2013 Revised: July 4, 2013 Published: July 22, 2013 10539

dx.doi.org/10.1021/la4020288 | Langmuir 2013, 29, 10539−10548

Langmuir

Article

Figure 1. Some structures of PDA 1−10 and eumelanin 11−14 proposed in the literature.

This conclusion was based on the results obtained using several analytical methods, such as 13C NMR, FTIR, UV, and EPR spectroscopy, and powder X-ray diffraction as well as chemical oxidation studies. However, as already surmised in our previous results6 on the structural investigation of PDA/ magnetite nanoparticles, this structural proposal is questionable. Hong et al. proposed that PDA consists of mixtures of covalently bound indole units and physical trimers wherein two dopamine molecules and one dihydroxyindole are held together by π stacking and hydrogen bonding.26 Herein we report investigations of the PDA structure using various analytical methods, which provide new experimental evidence that supports the structural assignment of 15 (Figure 8) of PDA. Our results are in agreement with some structures previously described in the literature but are in disagreement with others, such as that proposed by Dreyer et al.25

melanin structures obtained by the oxidation of L-DOPA, dopamine, or 5,6-dihydroxyindole were the same (RaperMason theory),16 this assumption was disproved later.14 Synthetic model compounds of eumelanins can be obtained by the oxidative coupling of 5,6-dihydroxyindole-2-carboxylic acid.18 PDA analogous structures, such as 11, were proposed for eumelanins.19 In addition, different skeletons involving linkages of the indole units via position 3 (12, also hypothesized for PDA)1 and condensed polyindoles of type 13 formed by an assumed Diels−Alder reaction20 were postulated for eumelanins. On the basis of DFT calculations, cyclic tetramer 14 was also proposed to be a eumelanin protomolecule.21 Furthermore, an indication was found that eumelanins consist of pancakelike piles of lower oligomers (up to eight indole units).22 Beside structures of PDA and eumelanins based on the covalent linkage of monomer units as evidenced by mass spectrometry,23,24 there are also some reports that describe the involvement of H bonding rather than C−C linkages. Thus, very recently, Dreyer et al.25 proposed H bonding between indoline-derived quinoid monomers and indoline catechol-like units building up the PDA, such as structure 10 (Figure 1).

■

EXPERIMENTAL SECTION

Materials. L-Dopamine-hydrochloride, TRIS buffer, and phosphate buffer were purchased from Sigma-Aldrich (Germany). 10540

dx.doi.org/10.1021/la4020288 | Langmuir 2013, 29, 10539−10548

Langmuir

Article

the dielectric constant, ζ is the zeta potential, η is the viscosity, and f(κa) is the Henry function (Smoluchowski approximation f(κa) = 1.5; Hückel approximation f(κa) = 1). Computational Methods. The equilibrium geometry of four-unit indole structures were obtained using DFT calculations with the M062X31 exchange-correlation function and considering the TZVP32 tripleζ basis set with polarization functions implemented in the Gaussian 0933 program package. For the intermolecular interaction energy, calculations with the def2-TZVP34,35 basis set were used. Because of the huge number of atomic components, larger indole unit networks could not be properly described in the framework of the ab initio or the conventional DFT molecular theories. In contrast, the densityfunctional-based tight-binding method combined with the selfconsistent charge technique (SCC-DFTB)36 can be considered to be an adequate solution for treating large biologically interesting or nanoscaled materials with good accuracy, similar to results of highlevel theoretical methods.37−39 Accordingly, the geometry optimization of stacking-chain polymers with two chains of four-unit indole structures was obtained using the SCC-DFTB method implemented in the DFTB+40,41 program, where the empirical correction for dispersion forces was also taken into account. The molecular graphics (figures of the molecular structures) were built using Avogadro42 software.

Polydopamine was obtained according to literature procedures in TRIS buffer or phosphate buffer.1 Dopamine hydrochloride (1g, 5.3 mmol) was dissolved in 500 mL of 10 mM buffer (TRIS at pH 8.5 or phosphate at pH 8.5) and stirred in air for 20 h. The resulting black precipitate was separated by centrifugation at 4000 rpm, washed with water, and centrifuged again. This washing step was repeated three times, and the solid was dried under vacuum at 40 °C overnight. In the case of TRIS buffer, higher yields of solid PDA were obtained (170 mg) than in phosphate buffer (40 mg). Although it is known that the properties of PDA can depend on the preparation conditions,26−28 it turns out that both samples show identical NMR and FTIR spectra and MS data. Spectroscopic Methods. The 1H NMR spectra were obtained on a Bruker Avance III 600 MHz NMR spectrometer (Bruker Biospin, Rheinstetten, Germany). A 4 mm double-resonance MAS probe was used in MAS NMR spectroscopy. To obtain the 1H MAS NMR spectrum, PDA was dissolved in D2O/DMSO-d6 and was used to fill a disposable insert (Bruker Biospin). A MAS frequency of 5000 Hz was applied. The liquid-state 1H NMR spectrum of PDA in TFA-d solution was measured on a 5 mm TXI probe head with a z gradient using a standard watergate W5 sequence for solvent suppression.29 The 90° 1 H pulse was 13.75 μs. All spectra were measured at 303 K. 13 C solid-state (13C ss) NMR spectra were acquired by using two distinct experimental setups. One corresponds to a standard CP-MAS experiment, whereas for the other the so-called cross-polarization polarization-inversion (CPPI) experimental scheme30 was employed. The 13C ss-NMR spectra were recorded at a 125.73 MHz Larmor frequency with a Bruker AVANCE III 500 spectrometer operating at room temperature. Contact pulses of 2 ms and 50 μs were considered for direct CP, whereas the polarization inversion (PI) time τinv was adjusted for the complete depolarization of aromatic CH carbons in rigid organic solids (Results and Discussion). Both spectra were acquired under high-power proton decoupling (100 kHz) by using the two-pulse phase-modulation (TPPM) sequence. To get a reasonable S/N ratio, 60 000 transients were averaged with a recycle delay of 1 s. Calibration was carried out with respect to the CH3 line in TMS (tetramethylsilane), through an indirect procedure that uses the carbonyl 13C line (176.5 ppm) of ß-glycine as an external reference. High-resolution mass spectra (HRMS) were recorded on an LTQ ORBITRAP XL mass spectrometer (ThermoScientific) using positive ion mode electrospray (ES+) ionization. The instrument was externally calibrated according to the manufacturer’s recommendations. The samples were introduced into the spectrometer by direct infusion at a flow rate of 5 μL/min. The conditions used were as follows: spray voltage, 5 kV; sheath and auxiliary gas flow, 35 and 7 arbitrary units, respectively; capillary temperature, 275 °C; capillary voltage, 45 V; and tube lens voltage, +250 V. The number of microscans was set to three in all experiments. Chemical surface analysis of PDA was performed by X-ray photoelectron spectroscopy (XPS). XPS spectra were recorded using a SPECS XPS spectrometer equipped with a dual-anode Al/Mg X-ray source, a PHOIBOS 150 2D CCD hemispherical energy analyzer, and a multichanneltron detector with vacuum maintained at 1 × 10−9 Torr. The AlK Kα X-ray source (1486.6 eV) operated at 200 W was used for XPS investigations. The XPS survey spectra were recorded at 30 eV pass energy with 0.5 eV/step. The high-resolution spectra for individual elements were recorded by accumulating 10−15 scans at a 30 eV pass energy and 0.1 eV/step. The powdered sample was pressed on an indium foil to perform the XPS measurements. Sample surfaces were cleaned prior to analyses by argon ion bombardment (500 V). Data analysis and curve fitting were performed using Casa XPS software with a Gaussian−Lorentzian product function and nonlinear Shirley background subtraction. FTIR spectra were recorded in KBr pellets using a Jasco FTIR 610 spectrophotometer. A Malvern Zetasizer Nano-ZS and secondgeneration PALS (phase analysis light scattering), called M3PALS, were used to measure the zeta potential. The zeta potential was determined indirectly via measurement of the electrophoretic mobility Ue and using the Henry equation: Ue = ((2εζf(κa))/3η), where ε is

■

RESULTS AND DISCUSSION PDA was synthesized by the reported procedures,1 exposing dopamine hydrochloride to air under basic conditions in TRIS or phosphate buffer. FTIR spectroscopic investigations showed no differences in the PDA products obtained by the two methods (Supporting Information). However, it is worth mentioning that the polymerization occurred much slower in phosphate buffer and the yields in terms of black precipitated PDA were much lower. This is attributed to the formation of a considerable portion of soluble products, which were not investigated in the present study but were investigated before by Hong et al.26 Obviously, these soluble materials possess different structures and properties and were shown to function as intermediates in the formation of insoluble PDA pigments under the proper conditions. In this paper, the indole units shown in Figure 2 were taken into consideration as monomer building blocks for PDA. Most

Figure 2. Possible monomer indole and dopamine units M of PDA and monomers MH. 10541

dx.doi.org/10.1021/la4020288 | Langmuir 2013, 29, 10539−10548

Langmuir

Article

of these structures are based on the hitherto accepted assumption that dopamine forms indole-like structures upon oxidation. It has to be mentioned that M2/MH2, M3/MH3, and M4/MH4 as well as M5/MH5, M6/MH6, and M7/MH7 are tautomers. However, it has to be taken into consideration that monomer units M7 and M8 can be found in PDA chains where the amino group is not connected to the aromatic ring to form an indole structure as already proposed in the literature.7,9,14,15,43−46 Oxidative coupling of catechol units or other phenol derivatives connecting two or more benzene rings with each other by the formal loss of two hydrogen atoms is a well-known reaction and thus can also occur in the case of the oxidation of dopamine. In fact, it was previously shown that thermal treatment of PDAcoated surfaces causes the cyclization of such free aminoethyl phenyl groups into indole units.44 NMR investigations under magic angle spinning conditions (MAS), among them 1H NMR (in DMSO-d6/D2O) measurements for the first time (Figure 3), revealed that both 2,3-

Figure 4. 13C ss-NMR spectra of PDA (TRIS sample) recorded at 14 kHz MAS frequency by using the standard CP-MAS sequence at long (2 ms, black line) and short (80 μs, blue line) contact times and CPPI sequences (red line), respectively. The A−H labels of the dominant NMR lines correspond to the assignment given for structure 15 and also in ref 25 (see structure 10 in Figure 1) whereas the arrows denote smaller lines that can most probably be assigned to the CH2 and aromatic CH carbons in M7/M8 and M1/M3 monomer units, respectively.

units M7 and M8. Carbonyl C-atom signals of M1, M2, M4, M6, or M7 were found at about 170 −180 ppm, and phenolic C−OH carbon atom signals (as in M3, M4, M5, M6, or M8) were found at about 145 ppm. Other signals in the 110−130 ppm region can be assigned to benzo C atoms in any of the M substructures. Tautomer M6 can be ruled out because a signal at around 60 ppm, which is typical for the saturated bridgehead CH, is missing. To obtain unambiguous proof of whether protons are attached to the benzene rings of PDA, a more detailed analysis based on cross-polarization (CP) experiments was carried out under MAS at a 14 kHz sample spinning frequency. The 13C ss CP-MAS spectrum, shown in black in Figure 4, resembles fairly well the spectrum reported by Dreyer et al.25 as well as Della Vecchia et al.15 (For 13C CP-MAS spectra of natural melanins, see ref 47.) In particular, all reported major signals can also be identified in our case. However, the larger MAS frequency employed in the present study enabled the observation of other features, which could not be previously noticed because of the superposition between central lines and spinning side bands. Thus, partially overlapping 13C lines were distinguished at 27 and 104 ppm (labeled with arrows in Figure 4). They could be assigned to benzylic CH2 (M7, M8) and to CH in position 3 of the pyrrole ring of indoles (M1, M3), respectively. Because their intensity is low, these structural units can occur only to a lesser extent. Della Vecchia et al. came to the same conclusion as far as noncyclized units were concerned.15 To enable a direct comparison with the structure proposed in Dreyer’s paper, the 13C spectral assignment through the A−G labeling scheme (structure 10 in Figure 1) is also used here (structure 15 in Figure 8). As compared to the previously reported 13C spectrum, some of the signals we obtained (Figure 4) have slightly different chemical shifts: 33 ppm (G), 42 ppm (H), 118/120 ppm (D/C), 130 ppm (E), 144 ppm (F), and 172 ppm (A/B). The observed differences occur, most probably because the CP-MAS spectrum recorded at a 14 kHz MAS frequency is free of overlaps with spinning sidebands.

Figure 3. Solid state 1H MAS NMR spectrum (600.1 MHz) of PDA (TRIS sample) in D2O/DMSO-d6 solution at a MAS frequency of 5000 Hz at 303 K.

dihydroindole and indole moieties exist in PDA as exemplified in structures M1−M6 and open-chain aminoethyl compounds M7 and M8 can also be involved. This assumption is in agreement with relevant peaks in the 1H NMR spectrum (Figure 3) for aliphatic protons (M2, M4−M8) at about 1.4 (CH2−C), 2.9 (CH2−C or CH2−N), and 4 ppm (CH2−N) and aromatic CH signals (M1, M3) at about 7 and 8.1 ppm (indole positions 3 and 2, respectively). The aliphatic signals appeared to be slightly shifted as a result of anisotropy effects exerted by the aromatic moieties. It has to be stressed that peaks around 6.6 ppm, typical for monomer 2,3-dihydroindoles MH with H atoms in positions 4 and 7, as would display structure 10 proposed by Dreyer et al.,25 were not observed. These results can be underlined by the liquid-state 1H NMR spectrum of PDA in TFA-d (Figure S17 in SI). Peaks characteristic of 2,3-dihydroindole and indole fragments M were also found in the 13C CP MAS NMR spectrum (Figure 4, black curve). Signals at 30 and 40 ppm may be assigned to the carbon atoms of the partially saturated fivemembered ring in M2, M4, M5, and M6 but also to the carbon atoms belonging to the aminoethyl moieties in noncyclized 10542

dx.doi.org/10.1021/la4020288 | Langmuir 2013, 29, 10539−10548

Langmuir

Article

Figure 5. ES(+)-HRMS spectrum of PDA (0.1 mg/mL) in 97/2/1 methanol/DMSO/TFA. (a) Comparison of the experimental spectrum (top) and simulated isotopic patterns (bottom) for PDA tetramers with different degrees of saturation (4Q+6 to 4Q+18). (b) HR-MS peak of a PDA octamer (8Q+36).

In Figure 4, the conventional 13C CP-MAS spectrum (black) is compared to a short-contact-time CP-MAS spectrum (blue), which is dominated by 13C signals of protonated carbon atoms, and to a CPPI (cross-polarization polarization inversion) spectrum (red) adjusted for the complete cancelation of aromatic CH signals, whereas the dominant resonances are generated by quaternary carbons. From this perspective, the spectra marked in blue and red can be considered, to some extent, to be due to the splitting of the long-contact-time CPMAS spectrum (black in Figure 4) into two distinct contributions of protonated carbon atoms and quaternary carbon atoms, respectively. The short contact time 13C CPMAS spectrum (blue) was recorded using an 80 μs contact time, which represents an average of the maximum 1H → 13C transferred polarization in (quasi-rigid) CH and CH2 moieties. As expected, all of the aliphatic CH2 lines show up in the low ppm region of the spectrum. However, the aromatic region of the short-contact-time spectrum (blue) displays only two resonances at 104 and 118 ppm as compared to the conventional CP MAS spectrum (black). Therefore, to clarify this issue, a complementary analysis by CPPI was mandatory.48 CPPI extends the conventional CP-MAS approach in the sense that the direct 1H → 13C polarization transfer is followed by a reversed 13C → 1H transfer period for which the duration, τinv, can be adjusted in such a way that selected 13C nuclei are completely depolarized and their NMR signals are eliminated. Assuming that the molecules are quite rigidly packed in the solid state, the τinv = τ0j values required for complete depolarization of different Cj carbons can be associated with particular moieties to which they belong. For protonated carbon atoms, the characteristic τ0 values are short (35−60 μs) and depend mostly on the C−H bond lengths and the number of protons. For quaternary carbons, the τ0 values are longer (>150 μs). Here, for each particular carbon site, the exact τ0 value is a measure of the distances from the closest nonbonded protons. Because the polarization inversion period τinv exceeds a characteristic τ0j value, the polarization of the corresponding Cj

carbon will start to build up in the opposite direction, thus reversing the Cj NMR line in the spectrum: its “negative” amplitude will increase with further increasing τinv. To determine appropriate parameters for recording the 13C CPPI spectrum of PDA as shown in Figure 4 (red), several test CPPI experiments on Lisinopril (a relatively rigid phenylcontaining organic solid) were carried out in combination with numerical simulations of polarization inversion curves on representative CHn (n = 2−6) spin systems, as performed by Tripon et al.49 for aliphatic moieties. For this model system, it was established that the CH carbons in the phenyl ring fully depolarize at τ0 values ranging from 48 to 53 μs so that an average value of τinv = 50 μs was chosen to record the CPPI spectrum of PDA, which can also be assumed to be fairly rigid. At this polarization inversion time, the 13C lines at 104 and 118 ppm disappeared (Figure 4, in red), whereas all of the others had reduced intensities, which correlated with partial depolarization or repolarization in the opposite direction of the corresponding 13C nuclear spins. This is an important result indicating that the two carbon lines labeled C/D in the publication of Dreyer et al.25 cannot both be protonated. If this were the case, then the line at 120 ppm would also disappear in our CPPI spectrum. All of these NMR results give evidence of PDA structure 15 (Figure 8). The signals at 118 ppm may be assigned to aromatic tertiary carbon atoms (CH) as found in position F of openchain dopamine units M7 and M8 in cyclotetramers (Figure 6) and in terminal monomer units of oligomer chains (positions T in structure 15, see also the oligomer structures verified by HR MS in Figure 6). Quaternary carbon atoms at 120 ppm can be assigned to benzo positions D at bridging points (positions 4 and 7 for indole units) between monomer units. Dreyer et al.25 concluded from their CP MAS experiment (with additional evolution time) that both signals of C atoms in positions C and D are tertiary. However, the signal reduction that they observed is not proof of the covalent attachment of the protons to the respective C atoms. The phenomenon can be also caused by the proximity of those particular C atoms with protons in the 10543

dx.doi.org/10.1021/la4020288 | Langmuir 2013, 29, 10539−10548

Langmuir

Article

Figure 6. HR-MS peaks for tetramers and an octamer of PDA found by high-resolution ES(+)-MS.

unambiguously proves the existence of open-chain dopamine units, because otherwise it is impossible to host 18 additional H atoms. It is noteworthy that no difference in the ES(+) spectra of PDA samples obtained in TRIS or phosphate buffer was observed (Table S1 and Figures S2−S6 in SI). However, it has to be mentioned that the HRMS signal at m/ z 595.1841 (4 Q + 12H) also matches a cyclotetramer, similar to structure 14 postulated as a protomolecule for natural eumelanin.21 XPS spectra of PDA on polymer supports43,52,53 or on silica24 were previously measured. We obtained XPS spectra from nonsupported PDA (Figure 7 and Figure S16 in SI) giving evidence for structure 15. The O 1s spectrum (Figure S16 in SI) shows two major contributions one from CO (530.4 eV)

vicinity (2 to 3 Å), which is very likely to occur in PDA as a result of the dense packing of different PDA chains. (See also our quantum chemical calculations below.) This conclusion is in accordance with the 13C CPPI spectrum (Figure 4) where nonprotonated C atoms (C, D) are found for the benzo ring at postion 4 or 7 of indole structures that are adjacent to aromatic C−O or CO moieties of the dopamine units. Out of all the nonprotonated carbon atoms, C/D has the strongest coupling with other protons in the vicinity, which is the only explanation for the fact that the intensity of the line at 120 ppm is reduced to a larger extent as compared to the extent of reduction of the E/F (130/144 ppm) and A/B (172 ppm) 13C signals. Apart from this qualitative discussion, a more quantitative interpretation of the relative intensities in terms of local structural features is difficult because of structural heterogeneities both inside PDA chain 15 and at the level of its supramolecular aggregates. MALDI-MS proved to be a useful tool in the analysis of melanin50,51 and PDA-coated glass.23 We used ES (+)-MS to investigate the structure of PDA. Here, interesting peaks could be detected for oligomers with different degrees of saturation (i.e., the involvement of monomer units M1 to M8). Moreover, a particular feature observed for these oligomers was a conserved mass difference of 2. Some examples are shown in Figures 5 and 6 (see also Table S1 and Figures S2−S6 in SI) for PDA tetramers (4Q+6 to 4Q+18) and a PDA octamer (4Q +36). The same pattern of signals was observed for PDA trimers (peaks corresponding to oligomers 3Q to 3Q+16) and pentamers (peaks assigned to oligomer species 5Q+8 to 5Q +22; see SI). Naturally, the hydrogen atoms found at sites of saturation can be distributed in several ways in the oligomer chain, and additional tautomers can also exist. However, the appearance of the peak at m/z 601.2316 (Figure 5a)

Figure 7. High-resolution XPS spectra of N 1s core levels of PDA. 10544

dx.doi.org/10.1021/la4020288 | Langmuir 2013, 29, 10539−10548

Langmuir

Article

Figure 8. General structural proposal 15 for PDA (only two possibilities out of other tautomers are shown) and less likely alternative possibilities of bridging in PDA.

and the other from OH (532.1 eV), whereas another component of very low intensity located at 536.1 eV was ascribed to COOH, stemming from sample contamination by the environment, which is a known phenomenon in XPS. The best fit for the C 1s spectrum (Figure S16 in SI) was obtained with three components corresponding to the carbon atoms: C− C/CHx (284.1 eV), C−N/C−OH (285.2 eV), and CO (287.9 eV). The N 1s spectrum (Figure 7) contains only one contribution found at 399.4 eV, which is typical for amine NH. It is worth mentioning that imine nitrogen atoms as shown in the structure proposed by Shalev23 were not found in the HRXPS spectrum. Thus, structures 1, 8, 14, M4, and MH4 can be ruled out. The XPS spectra allowed us to obtain values of the atomic ratios of N/C = 0.095 and N/O = 0.3, which are smaller than the theoretical values for pure dopamine (N/C = 0.125 and N/O = 0.5). This result indicates the presence of more O and C atoms in the sample than expected, and this is most probably also due to sample contamination. Up to this point it can be stated that PDA exists as oligomer mixture 15 that can occur in different tautomeric forms (e.g., 15′). This is in agreement with previous assumptions that PDA and also naturally occurring eumelanins consist of mixtures, which in the latter case were termed “chemical disorders” by Meredith et al.54 Monomer units containing aminoethyl group M7 or M8 are presumably less abundant in PDA as already indicated in the 13C MAS NMR investigations (Figure 4). This is in agreement with the zeta potential of −44 mV determined for the PDA-TRIS sample (Figure S18 in SI). Obviously, the negative charges found in catechole units (acidic OH groups) of PDA strongly predominate the positive charges of aminoethyl groups, indicating that the latter occur to a lesser extent. In this context, it is worth mentioning that phenolformaldehyde resins exhibited a zeta potential of −38 mV.55

Possible mechanisms of PDA formation were previously proposed. They comprise oxidative oligomerizations of noncyclized dopamine units followed by oxidative ring closure to indole units or a reversed sequence wherein dopamine forms dihydroxyindoles prior to oxidative coupling to PDA.15 These proposals also allow us to explain our structural assessment of 15 for PDA. Possible connections of monomer units by 2,4-, 2,7-, or 2,2bridging as proposed in the literature19,26,47 (Figure 8) are not in contradiction with our results but are less likely to occur if the absence of signals between 6 and 7 ppm in the 1H NMR spectra is considered. Finally, the question arises how the monomer units are arranged and if there is H bonding within the polymer chain. DFT calculations of PDA chains built up of four different Mi monomers imply a linear chain structure where the molecular planes of consecutive monomer units are rotated by about 45− 70°. The four-unit PDA chain built from M3-M2-M3-M2 is presented in Figure S12 in the SI. To obtain information about the self-assembling pattern of PDA, we calculated double polymer chains, with each consisting of eight monomers (SI). The geometry optimization gave rise to four major combinations of two different PDA chains. For simplicity, only a central trimer part is shown in Figure 9 and Figures S14−S15 in SI. The first configuration is characterized by full stacking of the PDA molecular planes (Figure 9), where the monomers are oriented in a parallel manner. In the second case (Figure S13 in SI), the molecular planes also show a stacking configuration, but the PDA monomers are oriented antiparallel with respect to the five-membered rings, and in this way, the OH fragments can easily form intermolecular O−H···O-type interactions. In this configuration, the full stacking overlap cannot be observed anymore, but just a partial overlap of the rings. The third configuration (Figure S14 in SI) is based on the stacking 10545

dx.doi.org/10.1021/la4020288 | Langmuir 2013, 29, 10539−10548

Langmuir

Article

Figure 9. Parallel-stacking association pattern of the PDA chains.

overlap of the five-membered rings of every second monomer unit. This structure is further stabilized by two nonlinear H interactions of the N−H···O type between transversal PDA chains. A T-shaped form is found in the fourth configuration (Figure S15 in SI), where a C−H bond of the first PDA unit is oriented almost perpendicular to the molecular plane of the other PDA chain. Furthermore, different combinations of all four basic orientations are possible. To estimate the occurrence of these four chain configurations we have computed the strength of the intermolecular interaction between 2 × 2 dopamine units for each case. The strongest interaction was found for the parallel stack configuration, ΔE= −19.29 kcal/ mol, followed by the antiparallel configuration, ΔE= −14.75 kcal/mol. For the other two-chain configurations, much lower interaction energy values were obtained: ΔE = −8.85 kcal/mol for the third geometry and ΔE = −4.56 kcal/mol for the Tshaped configuration case. The two most stable arrangements in parallel or antiparallel stacking fashion are in agreement with the fact that PDA can form stable radicals,25 a property that is also found in simple quinhydrone. The supramolecular arrangement of PDA chains might also be important for explaining the black color of the polymer. However, for eumelanin, evidence was found by real-space microscopic analysis that the photophysical properties are derived from the primary chemical structure rather than by supramolecular organization.56 As already shown from quantum chemical calculations (vide supra), a similar situation was found to occur with the well-known quinhydrone where chargetransfer (CT) complexation between reduced and oxidized forms exists. In PDA, multiple CT can occur between adjacent chains where the hydroquinone units of one chain interact with the quinone units of the adjacent chain (Figure 10) possibly assisted by additional H-bonding, finally resulting in a 3D network. In fact, onionlike supramolecular nanoaggregates were proposed for PDA.56 Because different building blocks are randomly distributed in PDA chain 15, there could be regions where CT interaction is not possible because quinone units face quinone units or hydroquinone units face hydroquinone units. Such situations are shown in Figure 10 on the right-hand side. Those parts of the chains probably do not stay too far away from each other in order to maintain the edge-to-face interaction (Figure S15 in SI). However, the system could minimize such less-favorable constellations by switching from one tautomer to another (M2 to M3 and M5 to M7). Eventually misfits, such as those shown on the right side of Figure 10, are alternatively omitted by a kind of template reaction in the oxidation/dehydrogenation of dopamine to PDA leading to preferred CT situations. Branching in PDA

Figure 10. CT-complex formation between PDA chains and minimization of misfits by adopting complementary tautomers. (Top) Optimal tautomer distribution of monomer units. (Bottom) Misfit of the two units on the right-hand side because of an unfavorable constellation. Possible additional H bonding is not shown.

chains or 2,2-, 2,5-, or 2,7-bridging would disturb such arrangements and hence should be rare. As far as the structure proposed here is concerned, one has to be aware that the properties of PDA can depend on the reaction conditions (e.g., the oxidants used15,26−28), and thus our findings genuinely concern solid PDA obtained by air oxidation in TRIS or phosphate buffer.

■

CONCLUSIONS Important new contributions to the hitherto not fully understood structure of PDA, a widely used polymer for coating surfaces, were provided by the application of several analytical methods, among them 13C CPPI MAS NMR spectroscopy, 1H MAS NMR spectroscopy, and high-resolution ES(+)-MS for the first time. We demonstrated that PDA cannot consist of single indoline units just held together by hydrogen bonding as proposed in the literature. Instead, C−C connections between the monomer units exist. The occurrence of PDA oligomers in a different state of (un)saturation was proved by HR-MS. This method also evidenced the existence of dopamine units, wherein aminoethyl chains are found rather than five-membered N-heterocycles in the oligomer PDA chain. On the basis of these facts, a structural model of PDA was developed, consisting of mixtures of different oligomers wherein indole units with different degrees of (un)saturation and open-chain dopamine units occur. Intermolecular interactions of PDA chains were rationalized by DFT calculations, giving preference to CT interactions between o-quinoid and catechol units. On the basis of the profound experimental results, it was possible to give proof of hypothesized structural proposals for PDA found in the literature on one hand and to render other structural models unlikely on the other hand. In this way, important progress in the elucidation and understanding of the structure of PDA was achieved.

■

ASSOCIATED CONTENT

S Supporting Information *

1

H NMR spectrum, elemental analysis, FTIR spectra, HRMS data, DFT calculations of octamers, and zeta potential 10546

dx.doi.org/10.1021/la4020288 | Langmuir 2013, 29, 10539−10548

Langmuir

Article

reaction process on the polydopamine membranes and its corrosion protection properties for 304 stainless steel. J. Mol. Struct. 2010, 982, 152−161. (14) Binns, F.; King, J. A. G.; Mishra, S. N.; Percival, A.; Robson, N. C.; Swan, G. A.; Waggott, A. Studies related to chemistry of melanins 0.13. Studies on structure of dopamine-melanin. J. Chem. Soc. C 1970, 2063−2070. (15) Della Vecchia, N. F.; Avolio, R.; Alfe, M.; Errico, M. E.; Napolitano, A.; d’Ischia, M. Building-block diversity in polydopamine underpins a multifunctional eumelanin-type platform tunable through a quinone control point. Adv. Funct. Mater. 2013, 23, 1331−1340. (16) Ito, S. A chemist’s view of melanogenesis. Pigm. Cell Res. 2003, 16, 230−236. (17) Peter, M. G.; Förster, H. On the structure of eumelanins identification of constitutional patterns by solid-state NMR-spectroscopy. Angew. Chem., Int. Ed. 1989, 28, 741−743. (18) Pezzella, A.; Vogna, D.; Prota, G. Atropoisomeric melanin intermediates by oxidation of the melanogenic precursor 5,6dihydroxyindole-2-carboxylic acid under biomimetic conditions. Tetrahedron 2002, 58, 3681−3687. (19) d’Ischia, M.; Napolitano, A.; Pezzella, A.; Meredith, P.; Sarna, T. Chemical and structural diversity in eumelanins: unexplored biooptoelectronic materials. Angew. Chem., Int. Ed. 2009, 48, 3914−3921. (20) Swift, J. A. Speculations on the molecular structure of eumelamin. Int. J. Cosm. Sci. 2009, 31, 143−150. (21) Kaxiras, E.; Tsolakidis, A.; Zonios, G.; Meng, S. Structural model of eumelanin. Phys. Rev. Lett. 2006, 97, 218102−218106. (22) Zajac, G. W.; Gallas, J. M.; Cheng, J.; Eisner, M.; Moss, S. C.; Alvaradoswaisgood, A. E. The fundamental unit of synthetic melanin a verification by tunneling microscopy of X-ray-scattering results. Biochim. Biophys. Acta, Gen. Subj. 1994, 1199, 271−278. (23) Shalev, T.; Gopin, A.; Bauer, M.; Stark, R. W.; Rahimipour, S. Non-leaching antimicrobial surfaces through polydopamine bioinspired coating of quaternary ammonium salts or an ultrashort antimicrobial lipopeptide. J. Mater. Chem. 2012, 22, 2026−2032. (24) Bernsmann, F.; Ponche, A.; Ringwald, C.; Hemmerle, J.; Raya, J.; Bechinger, B.; Voegel, J. C.; Schaaf, P.; Ball, V. Characterization of dopamine-melanin growth on silicon oxide. J. Phys. Chem. C 2009, 113, 8234−8242. (25) Dreyer, D. R.; Miller, D. J.; Freeman, B. D.; Paul, D. R.; Bielawski, C. W. Elucidating the structure of poly(dopamine). Langmuir 2012, 28, 6428−6435. (26) Hong, S.; Na, Y. S.; Choi, S.; Song, I. T.; Kim, W. Y.; Lee, H. Non-covalent self-assembly and covalent polymerization co-contribute to polydopamine formation. Adv. Funct. Mater. 2012, 22, 4711−4717. (27) Bertazzo, A.; Costa, C. V. L; Allegri, G.; Favretto, D.; Traldi, P. Application of matrix-assisted laser desorption/ionization mass spectrometry to the detection of melanins formed from dopa and dopamine. J. Mass Spectrom. 1999, 922−929. (28) Bernsmann, F.; Ball, V.; Addiego, F.; Ponche, A.; Michel, M.; Gracio, J. J. D.; Toniazzo, V.; Ruch, D. Dopamine-melanin film deposition depends on the used oxidant and buffer solution. Langmuir 2011, 27, 2819−2825. (29) Liu, M. L.; Mao, X. A.; Ye, C. H.; Huang, H.; Nicholson, J. K.; Lindon, J. C. Improved WATERGATE pulse sequences for solvent suppression in NMR spectroscopy. J. Magn. Reson. 1998, 132, 125− 129. (30) Palmas, P.; Tekely, P.; Canet, D. Local-field measurements on powder samples from polarization inversion of the rare-spin magnetization. J. Magn. Reson. Ser. A 1993, 104, 26−36. (31) Zhao, Y.; Truhlar, D. G. The M06 suite of density functionals for main group thermochemistry, thermochemical kinetics, noncovalent interactions, excited states, and transition elements: two new functionals and systematic testing of four M06-class functionals and 12 other functionals. Theor. Chem. Acc. 2008, 120, 215−241. (32) Schafer, A.; Huber, C.; Ahlrichs, R. Fully optimized contracted gaussian-basis sets of triple zeta valence quality for atoms Li to Kr. J. Chem. Phys. 1994, 100, 5829−5835.

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

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Fax: +40 264 420042. Author Contributions

H.A.S., C.F. and N.H. contributed equally to this work. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the Romanian Ministry of Education and Research under research program POS-CCEMETAVASINT, project no. 550/2010. We acknowledge Dr. Ioan Bratu for FTIR (INCDTIM Cluj-Napoca) measurements and the group of Prof. Dr. Ladislau Vecas (Academy of Science, Branch Timisoara) for zeta potential determination.

■

ABBREVIATIONS PDA, polydopamine; CPPI MAS NMR, cross-polarization polarization-inversion magic angle spinning NMR; MAS NMR, magic angle spinning NMR; ES-MS, electrospray ionization mass spectrometry; XPS, X-ray photoelectron spectroscopy

■

REFERENCES

(1) Lee, H.; Dellatore, S. M.; Miller, W. M.; Messersmith, P. B. Mussel-inspired surface chemistry for multifunctional coatings. Science 2007, 318, 426−430. (2) Lynge, M. E.; van der Westen, R.; Postma, A.; Stadler, B. Polydopamine-a nature-inspired polymer coating for biomedical science. Nanoscale 2011, 3, 4916−4928. (3) Ye, Q.; Zhou, F.; Liu, W. M. Bioinspired catecholic chemistry for surface modification. Chem. Soc. Rev. 2011, 40, 4244−4258. (4) Zhu, L. P.; Yu, J. Z.; Xu, Y. Y.; Xi, Z. Y.; Zhu, B. K. Surface modification of PVDF porous membranes via poly(DOPA) coating and heparin immobilization. Colloids Surf., B 2009, 69, 152−155. (5) Fei, B.; Qian, B. T.; Yang, Z. Y.; Wang, R. H.; Liu, W. C.; Mak, C. L.; Xin, J. H. Coating carbon nanotubes by spontaneous oxidative polymerization of dopamine. Carbon 2008, 46, 1795−1797. (6) Mrowczynski, R.; Turcu, R.; Leostean, C.; Scheidt, H. A.; Liebscher, J. New versatile polydopamine coated functionalized magnetic nanoparticles. Mater. Chem. Phys. 2013, 138, 295−302. (7) Zhang, M.; He, X. W.; Chen, L. X.; Zhang, Y. K. Preparation of IDA-Cu functionalized core-satellite Fe3O4/polydopamine/Au magnetic nanocomposites and their application for depletion of abundant protein in bovine blood. J. Mater. Chem. 2010, 20, 10696−10704. (8) Si, J. Y.; Yang, H. Preparation and characterization of biocompatible Fe3O4@polydopamine spheres with core/shell nanostructure. Mater. Chem. Phys. 2011, 128, 519−524. (9) Lee, H.; Rho, J.; Messersmith, P. B. Facile conjugation of biomolecules onto surfaces via mussel adhesive protein inspired coatings. Adv. Mater. 2009, 21, 431−434. (10) Zhou, W. H.; Lu, C. H.; Guo, X. C.; Chen, F. R.; Yang, H. H.; Wang, X. R. Mussel-inspired molecularly imprinted polymer coating superparamagnetic nanoparticles for protein recognition. J. Mater. Chem. 2010, 20, 880−883. (11) Ren, Y. H.; Rivera, J. G.; He, L. H.; Kulkarni, H.; Lee, D. K.; Messersmith, P. B. Facile, high efficiency immobilization of lipase enzyme on magnetic iron oxide nanoparticles via a biomimetic coating. BMC Biotechnol. 2011, 11, 63. (12) Liu, Q. Z.; Wang, X. L.; Yu, B.; Zhou, F.; Xue, Q. J. Self-healing surface hydrophobicity by consecutive release of hydrophobic molecules from mesoporous silica. Langmuir 2012, 28, 5845−5849. (13) Yu, F.; Chen, S. G.; Chen, Y.; Li, H. M.; Yang, L.; Chen, Y. Y.; Yin, Y. S. Experimental and theoretical analysis of polymerization 10547

dx.doi.org/10.1021/la4020288 | Langmuir 2013, 29, 10539−10548

Langmuir

Article

(33) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, J.; et al. Gaussian 09, Revision C.01. Gaussian Inc.: Wallingford, CT, 2009. (34) Weigend, F.; Ahlrichs, R. Balanced basis sets of split valence, triple zeta valence and quadruple zeta valence quality for H to Rn: design and assessment of accuracy. Phys. Chem. Chem. Phys. 2005, 7, 3297−3305. (35) Schuchardt, K. L.; Didier, B. T.; Elsethagen, T.; Sun, L. S.; Gurumoorthi, V.; Chase, J.; Li, J.; Windus, T. L. Basis set exchange: a community database for computational sciences. J. Chem. Inf. Model. 2007, 47, 1045−1052. (36) Elstner, M.; Porezag, D.; Jungnickel, G.; Elsner, J.; Haugk, M.; Frauenheim, T.; Suhai, S.; Seifert, G. Self-consistent-charge densityfunctional tight-binding method for simulations of complex materials properties. Phys. Rev. B 1998, 58, 7260−7268. (37) Elstner, M.; Jalkanen, K. J.; Knapp-Mohammady, M.; Frauenheim, T.; Suhai, S. Energetics and structure of glycine and alanine based model peptides: approximate SCC-DFTB, AM1 and PM3 methods in comparison with DFT, HF and MP2 calculations. Chem. Phys. 2001, 263, 203−219. (38) Elstner, M.; Jalkanen, K. J.; Knapp-Mohammady, M.; Frauenheim, T.; Suhai, S. DFT studies on helix formation in Nacetyl-(L-alanyl)(n)-N’-methylamide for n=1−20. Chem. Phys. 2000, 256, 15−27. (39) Elstner, M.; Hobza, P.; Frauenheim, T.; Suhai, S.; Kaxiras, E. Hydrogen bonding and stacking interactions of nucleic acid base pairs: A density-functional-theory based treatment. J. Chem. Phys. 2001, 114, 5149−5155. (40) DFTB+ home page. http://www.dftb-plus.info. (41) Aradi, B.; Hourahine, B.; Frauenheim, T. DFTB+, a sparse matrix-based implementation of the DFTB method. J. Phys. Chem. A 2007, 111, 5678−5684. (42) Hanwell, M. D.; Curtis, D. E.; Lonie, D. C.; Vandermeersch, T.; Zurek, E.; Hutchison, G. R. Avogadro: an advanced semantic chemical editor, visualization, and analysis platform. J. Cheminform. 2012, 4, 17. (43) Pan, F. S.; Jia, H. P.; Qiao, S. Z.; Jiang, Z. Y.; Wang, J. T.; Wang, B. Y.; Zhong, Y. R. Bioinspired fabrication of high performance composite membranes with ultrathin defect-free skin layer. J. Membr. Sci. 2009, 341, 279−285. (44) Proks, V.; Brus, J.; Pop-Georgievski, O.; Vecernikova, E.; Wisniewski, W.; Kotek, J.; Urbanova, M.; Rypacek, F. Thermalinduced transformation of polydopamine structures: an efficient route for the stabilization of the polydopamine surfaces. Macromol. Chem. Phys. 2013, 214, 499−507. (45) Crescenzi, O.; Kroesche, C.; Hoffbauer, W.; Jansen, M.; Napolitano, A.; Prota, G.; Peter, M. G. Synthesis of dopamines labeled with C-13 in the alpha-side or beta-side chain position and their application to structural studies on melanins by solid-state NMRspectroscopy. Liebigs Ann. Chem. 1994, 563−567. (46) Ju, K. Y.; Lee, Y.; Lee, S.; Park, S. B.; Lee, J. K. Bioinspired polymerization of dopamine to generate melanin-like nanoparticles having an excellent free-radical-scavenging property. Biomacromolecules 2011, 12, 625−632. (47) Adhyaru, B. B.; Akhmedov, N. G.; Katritzky, A. R.; Bowers, C. R. Solid-state cross-polarization magic angle spinning C-13 and N-15 NMR characterization of sepia melanin, sepia melanin free acid and human hair melanin in comparison with several model compounds. Magn. Reson. Chem. 2003, 41, 466−474. (48) Rovnyak, D. Tutorial on analytic theory for cross-polarization in solid state NMR. Concept Magn. Reson. A 2008, 32A, 254−276. (49) Tripon, C.; Aluas, M.; Filip, X.; Filip, C. Polarization transfer from remote protons in C-13 CP/MAS. J. Magn. Reson. 2006, 183, 68−76. (50) Napolitano, A.; Pezzella, A.; Prota, G.; Seraglia, R.; Traldi, P. Structural analysis of synthetic melanins from 5,6-dihydroxyindole by matrix-assisted laser desorption ionization mass spectrometry. Rapid Commun. Mass Spectrom. 1996, 10, 468−472.

(51) Napolitano, A.; Pezzella, A.; Prota, G.; Seraglia, R.; Traldi, P. A reassessment of the structure of 5,6-dihydroxyindole-2-carboxylic acid melanins by matrix-assisted laser desorption ionization mass spectrometry. Rapid Commun. Mass Spectrom. 1996, 10, 204−208. (52) Li, B.; Liu, W. P.; Jiang, Z. Y.; Dong, X.; Wang, B. Y.; Zhong, Y. R. Ultrathin and stable active layer of dense composite membrane enabled by poly(dopamine). Langmuir 2009, 25, 7368−7374. (53) Xi, Z. Y.; Xu, Y. Y.; Zhu, L. P.; Wang, Y.; Zhu, B. K. A facile method of surface modification for hydrophobic polymer membranes based on the adhesive behavior of poly(DOPA) and poly(dopamine). J. Membr. Sci. 2009, 327, 244−253. (54) Meredith, P.; Powell, B. J.; Riesz, J.; Nighswander-Rempel, S. P.; Pederson, M. R.; Moore, E. G. Towards structure-property-function relationships for eumelanin. Soft Matter. 2006, 2, 37−44. (55) Sun, L. M.; Li, M. Y.; Lin, M. Q.; Peng, B.; Guo, J. X. Dispersion properties of a water-soluble phenol-formaldehyde resin. J. Disper. Sci. Technol. 2009, 30, 605−608. (56) Watt, A. A. R.; Bothma, J. P.; Meredith, P. The supramolecular structure of melanin. Soft Matter. 2009, 5, 3754−3760.

■

NOTE ADDED IN PROOF In parallel to this publication a recent review was published summarizing information about the structure of PDA, naturally not including the results presented here: Dreyer, D. R.; Miller, D. J.; Freeman, B. D.; Paul, D. R.; Bielawski, C. W. Perspectives on poly(dopamine). Chem. Sci. 2013, DOI 10.1039/c3sc51501.

10548

dx.doi.org/10.1021/la4020288 | Langmuir 2013, 29, 10539−10548