Fluorescence‐Lifetime Imaging and Super‐Resolution Microscopies Shed Light on the Directed‐ and Self‐Assembly of Functional Porphyrins onto Carbon Nanotubes and Flat Surfaces

Abstract Functional porphyrins have attracted intense attention due to their remarkably high extinction coefficients in the visible region and potential for optical and energy‐related applications. Two new routes to functionalised SWNTs have been established using a bulky ZnII‐porphyrin featuring thiolate groups at the periphery. We probed the optical properties of this zinc(II)‐substituted, bulky aryl porphyrin and those of the corresponding new nano‐composites with single walled carbon nanotube (SWNTs) and coronene, as a model for graphene. We report hereby on: i) the supramolecular interactions between the pristine SWNTs and ZnII‐porphyrin by virtue of π–π stacking, and ii) a novel covalent binding strategy based on the Bingel reaction. The functional porphyrins used acted as dispersing agent for the SWNTs and the resulting nanohybrids showed improved dispersibility in common organic solvents. The synthesized hybrid materials were probed by various characterisation techniques, leading to the prediction that supramolecular polymerisation and host–guest functionalities control the fluorescence emission intensity and fluorescence lifetime properties. For the first time, XPS studies highlighted the differences in covalent versus non‐covalent attachments of functional metalloporphyrins to SWNTs. Gas‐phase DFT calculations indicated that the ZnII‐porphyrin interacts non‐covalently with SWNTs to form a donor–acceptor complex. The covalent attachment of the porphyrin chromophore to the surface of SWNTs affects the absorption and emission properties of the hybrid system to a greater extent than in the case of the supramolecular functionalisation of the SWNTs. This represents a synthetic challenge as well as an opportunity in the design of functional nanohybrids for future sensing and optoelectronic applications.


General details for materials characterization
Transmission electron microscopy (TEM) images were obtained with Gatan Dualvison digital camera on a JEOL 1200EXII transmission electron microscope coupled with Energydispersive X-ray spectroscopy (point resolution, 0.16 nm).
Raman spectroscopy was carried out on a Renishaw inVia Raman spectroscopy. The specimens were either in solid state or dispersed in pure water (MilliQ) or water: ethanol 1:1 mixture. During the measurement, the carbon nanomaterials samples were deposited on an aluminum plate substrate. The input wavelength was set at 514 nm. More than 10 times accumulations were generally applied in Raman spectroscopy measurements, and the beam was focused in at least three different positions across the specimen and these spectra were averaged to obtain batch-representative peaks and most reliable results.
Fluorescence spectroscopy measurement was carried out on a Perkin Elmer Luminescence spectrophotometer LS 55. The concentration of sample applied to fluorescence spectroscopy can be vary between 10 -5 -10 -7 M depend on the strength of the fluorescence. UV-vis spectroscopy was carried out by using a Perkin-Elmer Lambda 35 spectrometer.
The Atomic Force Microscopy measurements were carried out a Digital Instruments Multimode Atomic Force Microscope with IIIa controller. All AFM measurements were obtained under tapping mode and using Silicon Probes (Nascatec GmbH model NST-NCHFR). The AFM samples were deposited onto freshly cleared mica substrate by spin coating (Laurell Technologies WS-400, 3000rpm).
Fourier transform Infrared spectra were obtained by using a Perkin-Elmer 1000 FT-IR spectrometer.
XP spectra of calcined samples were collected using a VG Escalab II spectrometer using aluminum Kα radiation (1486.6 eV) and a hemispherical analyzer for detection of electrons.
The Pass Energy was set at 50 eV for survey scan of the sample, and 20 eV for the more intense scans of specific areas. The resulting spectra were analyzed using CasaXPS peak fitting software, and sample charging corrected setting the C 1s (C-C) signal at 284.5 eV.

Single crystal X-ray diffraction
Free base porphyrin did not show the peak in the 2.5-3.5 min, and presented a main peak at 14.697 min, which means the free base porphyrin was majorly pure. For Zn(II)-porphyrin, similar measured result can be observed: there was a main absorption peak at 14.707 min. Figure S4. HPLC of Zn(II)-poprhyin and free base porphyrin

NMR
Solution multinuclear NMR spectra were recorded on a Bruker Avance 500MHz spectrometer. 1 Figure S11. 1 H DOSY NMR spectrum (500 MHz, 298 K, CDCl3) of free base porphyrin precursor of (1). Diffusion coefficient = 4.57x 10 -10 m 2 s -1 suggestion lack of meaningful aggregation levels in solutions at the ca 5 mM conc. needed to record a spectrum. Figure S12. 1 H DOSY NMR spectrum (500 MHz, 298 K, CDCl3) of (1). Diffusion coefficient = 1.62x 10 -9 m 2 s -1 which seems to indicate a limited ability of this Zn(II) prophyrin to aggregate in solution even at the ca 5 mM conc. needed to record a spectrum. Figure   hydrogen, nitrogen and carbon from the SWNTs). Figure S10(b) shows more detailed surface morphology information that after complexation with compound 1, the relatively clear and smooth SWNTs surface became rugged and there were some aggregations formed on it.

Characterisation of the Covalently linked Zn(II)-porphyrin and SWNTs complex.
Perceptibly, the side-wall of SWNT was significantly 'roughened' by the coverage of another material, which indicated the presence of compound 1 that coated on the surface of SWNT. [1], [2], [3], [4] The length of the SWNTs was not altered and the intact SWNTs structure indicated the controlled time sonication didn't introduce any surface defect. The TEM images can also indicate that the SNWTs were better dispersed in ethanol with a few bundles persist.

(a) (b)
Additional TEM images of non-covalently linked SWNTs@Zn(II)-porphyrin are presented in main text.
Figure S15: (a) Illustration of supramolecular assembly between porphyrins and singlewalled carbon nanotubes in agreement to that emerging from Taku Hasobe's research; [5] (b) Schematic diagram showing the attachment of a non-covalently linked porphyrin 1 within the complex.
The compound 1 may contribute to the formation some complex bundles, as seen by TEM.
Taku Hasobe's research_ENREF_32, [5] suggests that in this, and related, porphyrin@SWNTs systems, the porphyrin molecules can be used as a molecular glue by virtue of their π-π stacking on the surface of SWNTs to combine two functionalised SWNTs together and form a larger self-assembled structure. The work by Taku Hasobe indicated that the attaching angle of porphyrin and SWNTs was 0 degrees. This indicates that the combination of porphyrin and SWNTs occurs via a face-face orientation and the two systems in the donor-acceptor nanocomplex are very closely held together, and may well apply to the systems studied herein too.  together. It can be seen from the SEM images in Figure S17 and S18, which correspond to the non-covalently linked compound 1@SWNTs complex that surface morphology information can be derived, specifically that the coating with compound 1 was uniform and covered the entire aromatic surface of SWNTs. After the complexation, there appeared aggregates onto the SWNTs surface.

Solid state investigation by AFM
Full surface morphology information of this complex nanohybrid material can be achieved by combining TEM, SEM and AFM information.  The AFM spectroscopy measurements of non-covalently linked compound 1@SWNTs were carried out to further study the surface morphology information of the complexes.
For the non-covalently linked compound 1@SWNTs complex, AFM measurement was introduced by dropping compound 1@SWNTs complex dispersion solution onto a HOPG substrate, the mica substrate was also tested in the sample preparation progress with spin coating, but it did not seem to provide meaningful images.
The high qualified complex dispersion was prepared though prolonged sonication and centrifugations. The solvent combination for complex dispersion was 1:1 ethanol and chloroform at ca 1 mg/mL concentration, chosen not only to keep the continuity but also to be easily evaporated after dropping onto the HOPG substrate. From those images shown above, the compound 1 uniformly covers the entire aromatic surface of SWNTs and the surface decorated SWNTs could be relatively better dispersed compared to the raw intact SWNTs, for which it has not been possible to obtain meaningful images. The 3D AFM image indicated the fact that the individual functionalised carbon nanotube has ca. 9 nm height, although taller objects were identified and could not be imaged.
The AFM measurement was carried out via dropping the complex suspension onto a HOPG substrate. Because of the different electron charges and hydrophilic properties, a mica substrate measurement was attempted but failed in AFM sample preparation and characterisation process. Compared with non-covalently linked complex which presented a ca.
9 nm height, the covalently complex presented a slightly higher height. It is reasonable to assume that the height difference between covalently and non-covalently linked complex was due to the different linking space.      The intact SWNTs Raman spectra presents the G band at 1581 cm -1 (primary graphitic mode) and G band at 1289 cm -1 (defect band). Meanwhile, the Raman spectra of precursor emerging after the Bingel reaction containing functionalised SWNTs exhibits a G band at 1585 cm -1 and a D band at 1289cm -1 . It is commonly accepted that the ID/IG band intensity ratio reflects the degree of crystallinity of graphitic carbon structure in carbon nanotubes. It can be seen that the ID/IG band intensity ratio of Bingel reaction functionalised SWNTs (0.16) has been an increase compared to the ID/IG band intensity ratio of the intact free SWNTs

UV-vis titrations and fitting of association binding constant.
In order to confirm that the π-π interactions between porphyrin and carbon single-walled nanotubes and quantify their supramolecular assembling, a host-guest UV-Vis titration [6] was carried out by using coronene, C24H12, coronene, (termed Guest) as model for sp 2 carbon system and Zn-Porphyrin (termed Host). The formation of these HG 1:1 stoichiometry can be described by the following bimolecular equilibrium: The stepwise binding 1:1 constants therefore can be described as:  Table S5   Table S5. Binding constant Ka 1:1 is in M -1 . Standard error of estimated data (SEy) and convariance of fit (Covf) are reported and calculated according reference [6] K1:1 (M -1 ) SEy Covf Zn-Phorphyrin 46200 0.0038 0.0032 The UV-vis titration data indicated enhanced absorptions relative to the peak at 421 nm and 412 nm. A broader peak at 546 nm was also observed in the non-covalently linked complex filtrate, which corresponds to the secondary absorption band of intact free porphyrin. Figure S27 shows the UV-vi absorption titration of coronene and Zn(II)-porphyrin while the insert image is the magnified UV-vis titration spectrum. Plots of the experimental ΔA of Zn(II)-Porphyrin at λabs = 412 nm (blue markers), λabs = 421 nm (red markers) and fitting curves (blue and red solid lines for 1:1 fitted isotherm) are also reported hereby.       (98.7 %) and 2 = 476.1 ps (1.3%) respectively (910 nm), but chi 2 was larger than 1.5. The experiment was therefore carried out at 810 nm 2P excitation instead, which was found to be significantly more suitable for this porphyrin system (vide infra).

STED Super Resolution Microscopy
3D STED Microscopy was carried out on a Leica TCS SP8 3X Gated STED (Leica, Mannheim, Germany). [7] The microscope is equipped with a pulsed supercontinuum white light excitation laser at 80Mhz (NKT, Denmark), and a continuous wavelength STED lasers at 592nm and a pulsed at 775nm. Experiments with Zn(II)Porphyrin (1) were done exciting at 488nm, their emission was depleted at 775nm, and the emission was collected around 580nm employing Leica HyD detectors. For the STED 3D reconstruction, each individual XY plane of the whole structure was scanned using the same conditions as mentioned above. Each individual XY plane image was deconvolved using the Huygens Software package (SVI, Netherlands), the final surface rendered 3D reconstruction was done employing the 3D imaging Leica software, LAX.  Figure S49. Addition of coronene (1M solution in toluene) and washing process of the 1:1 toluene : Chloroform suspentions non-covalent and covalent linked Zn(II)-porphyrin@SWNTs (2) and (6).

The coronene challenge experiment
As discussed in the main text, coronene was introduced as a competitive candidate likely to disrupt the interactions between the SWNTs and Zn(II)-porphyrin systems This experiment sheds light on the nature of Zn-porphyrin-SWNTs and clarify its energy transfer process. Due to the strong delocalized electron density, coronene was expected to display a stronger interaction with porphyrin molecule compared with SWNTs. Coronene was added into the non-covalent and covalent linked Zn(II)-porphyrin@SWNTs suspensions and then the mixtures were washed with toluene several times. The final products were, then,collected by filtration (Scheme 1). The solids were redispersed in a 1:1 ethanol and chloroform. Both the filtrates and solids were kept for further analysis, as discussed below. In particularly, Figure   50a shows the (a) (b) Figure 50. Competition experiment: a) UV-vis spectra of Zn(II)-porphyrin, pristine SWNTs, non-covalent Zn(II)-porphyrin@SWNTs nanohybrids (2) and coronene, covalent Zn(II)-porphyrin@SWNTs nanohybrids (6) and coronene. After competitive exchanges with coronene, all the products were collected by filtration and dispersed in a 1:1 chloroform mixture (1 μM); b) UV-vis spectra of the filtrate of non-covalent Zn(II)-porphyrin@SWNTs (2) and covalent Zn(II)-porphyrin@SWNTs nanohybrids (6) after competitive exchanges with coronene UV-vis spectra of the Zn(II)-porphyrin, non-covalent and covalent linked Zn(II)-porphyrin@SWNTs and SWNTs after treatment with coronene and toluene. The characteristic absorption band at 436 nm of the non-covalent Zn(II)-porphyrin@SWNTs (2) complex disappeared and the final products exhibit a similar UV-vis absorption spectrum as the intact SWNTs suspension. This suggests that after treating the system with coronene and washing with toluene, the Zn(II)-porphyrin was successfully displaced from the SWNTs surface. In contrast, the characteristic absorption profile of the covalent Zn(II)-porphyrin@SWNTs complex (6) can be seen even after coronene treatment. Figure 50b shows the UV-vis measurement of Zn(II)-porphyrin, covalent and non-covalent linked Zn(II)-porphyrin@SWNTs filtrates after treatment with coronene/toluene and as well as free coronene. UV-vis absorption spectrum of the non-covalent linked Zn(II)-porphyrin@SWNTs filtrate shows that the absorption band at 420 nm was enhanced and the absorption band at 450 nm disappears followed by the appearance of broader peak also emerged at 546 nm. The latter corresponds to the secondary absorption band of intact Zn(II)-porphyrin. These results demonstrate that coronene and Zn(II)-porphyrins were linked together to form a new noncovalent complex. Zn(II)-porphyrin can, therefore, be displaced from the SWNTs walls only when coronene is added. Such experiments point out the difference of the chemical nature of the covalent and non-covalent interactions between porphyrin and SWNTs. SWNTs is formed and driven by - stacking interactions. During the assembling process, the Soret band of unreacted Zn(II)-porphyrin (1) disappeared as a result of the formation of a supramolecular non-covalent Zn(II)-porphyrin@SWNTs complex (2). The charge transfer band at 430 nm observed upon UV-Vis titration also corresponds to the maximum UV-Vis absorption of the non-covalently linked complex (2) shown in Figure 11 of the manuscript.

SWNTs-Porphyrin UV-vis titration
The maximum UV-Vis absorption for the complex (6) was observed at 420 nm due to a redshift of the Soret band and no charge transfer bands were observed.