**Carbon Nanomaterials for Therapy, Diagnosis, and Biosensing**

Editors

**Antonino Mazzaglia Anna Piperno**

MDPI • Basel • Beijing • Wuhan • Barcelona • Belgrade • Manchester • Tokyo • Cluj • Tianjin

*Editors* Antonino Mazzaglia University of Messina Italy

Anna Piperno University of Messina Italy

*Editorial Office* MDPI St. Alban-Anlage 66 4052 Basel, Switzerland

This is a reprint of articles from the Special Issue published online in the open access journal *Nanomaterials* (ISSN 2079-4991) (available at: https://www.mdpi.com/journal/nanomaterials/ special issues/carbon nano biosen).

For citation purposes, cite each article independently as indicated on the article page online and as indicated below:

LastName, A.A.; LastName, B.B.; LastName, C.C. Article Title. *Journal Name* **Year**, *Volume Number*, Page Range.

**ISBN 978-3-0365-4511-0 (Hbk) ISBN 978-3-0365-4512-7 (PDF)**

© 2022 by the authors. Articles in this book are Open Access and distributed under the Creative Commons Attribution (CC BY) license, which allows users to download, copy and build upon published articles, as long as the author and publisher are properly credited, which ensures maximum dissemination and a wider impact of our publications.

The book as a whole is distributed by MDPI under the terms and conditions of the Creative Commons license CC BY-NC-ND.

## **Contents**


#### **Annalaura Cordaro, Giulia Neri, Maria Teresa Sciortino, Angela Scala and Anna Piperno** Graphene-Based Strategies in Liquid Biopsy and in Viral Diseases Diagnosis Reprinted from: *Nanomaterials* **2020**, *10*, 1014, doi:10.3390/nano10061014 .............. **139**

### **About the Editors**

#### **Antonino Mazzaglia**

Antonino Mazzaglia is the Research Director at the National Council of Research-Institute for the Nanostructured Materials (CNR-ISMN) at the University of Messina (Italy). He is a member of the Chemical Science PhD Board at the University of Messina. He joined the International, European, and National Advisory Boards on Cyclodextrins and the Italian Society of Photobiology. He was the Chair of the International Summer School on Cyclodextrins (2013) and the Bridge Virtual Meeting (2021) and was the Chair of the XX International Cyclodextrin Symposium (20th ICS). He has been a visiting scientist and professor in Europe (University of Exeter, Kings College London, University of Normandie, etc.) and in various Italian Universities. He has been an invited plenary and key-note speaker at several conferences (ICS, ESP; ICPP, etc.), a Member of the Italian Research Evaluation Board (VQR) for chemistry and an Editorial Board Member of *Nanomaterials* and *IJMS*. He is the author 103 peer-reviewed papers, 7 book chapters, and 1 monographic review, accomplishing an H-index of 30 and 2336 citations (SCOPUS source). His research expertise is focused on bio-soft hybrid materials based on cyclodextrins, porphyrinoids and photosensitisers, polysaccharides, polymers, and carbon nanoplatforms that have stimuli-responsiveness properties for therapy, diagnosis, and biosensing.

#### **Anna Piperno**

Anna Piperno earned a Master's degree in Pharmaceutical Chemistry and Technology in 1994 and a PhD in Pharmaceutical Science in 1999 at the University of Messina (Italy). She obtained a permanent position at Messina University as a researcher in 2001 and is currently a full professor in organic chemistry at the University of Messina. Since 2001, she has carried out teaching activities in organic chemistry and has supervised several students and doctoral fellows. She is a member of the Chemical Science PhD Board at the University of Messina. She began her research activities by studying organic synthesis methodologies and the chemistry of heterocyclic compounds. Specifically, her work was focused on the design and synthesis of new compounds that interfere with viral replication or cell death/proliferation. Successively, her research interests have been extended to the functionalization of carbon nanomaterials and biopolymers for applications in drug delivery, regenerative medicine, and biosensing. She actively collaborates with pharmaceutical industries on projects in the field of drug delivery and liquid biopsy. Currently, she is the PI of the MSCA Doctoral Networks "STRIKE" project and of NATO's "VIPER" project. She is the author of 123 papers that have been published in peer-reviewed journals and book chapters (H index = 36, citations = 3143, Scopus).

### *Editorial* **Carbon Nanomaterials for Therapy, Diagnosis and Biosensing**

**Antonino Mazzaglia 1,\* and Anna Piperno 2,\***


In carbon nanomaterial design, the fine-tuning of their functionalities and physicochemical properties has increased their potential for therapeutic, diagnostic and biosensing applications [1–3]. In this Special Issue, articles or mini reviews on nanoplatforms originating from the synergistic combination of carbon-based nanomaterials (i.e., nanotubes, graphene, graphene oxide, carbon quantum dots, nanodiamond, etc.) and various functional molecules such as drugs, natural compounds, biomolecules, polymers, metal nanoparticles and macrocycles relevant in drug delivery, in multi-targeted therapy, in theragnostics, as scaffolds in tissue engineering and as a sensing material, have been selected for publication.

Trapani et al. investigated the ability of multiwalled carbon nanotubes (MWC-NTs) covalently modified with polyamine chains of various length (ethylenediamine (EDA) and tetraethylenepentamine (EPA)) to induce the J-aggregation of meso-tetrakis (4-sulfonatophenyl)porphyrin (TPPS) in different experimental conditions. The authors reported that, in mild acidic conditions, TPPS porphyrin easily self-assembles into Jaggregates, showing peculiar extinction bands in the visible region (λ ∼= 493 nm) and in the therapeutic window (λ ∼= 710 nm), together with an emission band in the red spectral region. The results of this study describe the experimental conditions in which to obtain stable TPPS J-aggregates in medium mimicking physiological conditions for a stimuliresponsive therapeutic action upon irradiation on their extinction bands and fluorescence probes in cellular environments [4].

In the design of diagnostic nanoplatforms, aiming to improve the surface-enhanced Raman spectroscopy (SERS) effect, Neri et al. proposed a new graphene/gold nanocomposite composed of gold nanoparticles (AuNPs), produced by pulsed laser ablation in liquids (PLAL), and a nitrogen-doped graphene platform (G-NH2) obtained by direct delamination and chemical functionalization of graphite flakes with 4-methyl-2-p-nitrophenyl oxazolone, followed by the reduction of p-nitrophenyl groups. This approach allowed the authors to study SERS properties of graphene loaded with pure AuNPs without the influence of capping agents, surfactants, or salt produced in the chemical reduction of gold ions. The SERS platform was tested for its ability to detect Rhodamine 6G and Dopamine as molecular probes at a concentration around 1 μM. The platform showed good stability and the ability to reproduce Raman signals without degradation although its sensitivity was low. Overall, the feasibility of the proposed method opens up the field to further research on improving the detection limits of molecular probes interacting with loaded AuNPs [5].

Nowadays, new therapeutic approaches using carbon nanomaterials have become very attractive. In this scenario, Pennetta et al. investigated the formation of Doxorubicin (DOX) nano-conveyors as a stacked drug-delivery system for application in cancer treatment. The innovative nanoplatform was obtained by functionalizing single- and multi-walled carbon nanotubes (CNT and MWCNT, respectively) by cycloaddition reaction between carbon nanotubes and a pyrrole-derived compound. Two different adducts between CNT

**Citation:** Mazzaglia, A.; Piperno, A. Carbon Nanomaterials for Therapy, Diagnosis and Biosensing. *Nanomaterials* **2022**, *12*, 1597. https://doi.org/10.3390/ nano12091597

Received: 23 April 2022 Accepted: 27 April 2022 Published: 9 May 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

and pyrrole polypropylene glycol (PPGP) were prepared: the supramolecular adduct (CNT/PPGPs) and the covalent one (CNT/PPGPc). The supramolecular interactions were studied on the basis of molecular dynamics simulations, and by monitoring the emission and the absorption spectra of DOX. The biological studies revealed that two of the synthesized nanoplatforms are effectively able to obtain DOX within A549 and M14 cell lines and to enhance the cell mortality at a much lower effective dose of DOX. This work paves the way for the facile functionalization of carbon nanotubes by exploiting the "pyrrole methodology" for the development of novel technological carbon-based drug-delivery systems [6].

One of the main widespread concerns regarding using carbon-based materials as alternative nanobiomaterials for cancer therapy is their inherent cytotoxicity, which remains debated, with studies demonstrating contradictory results. Garriga et al. investigated the in vitro toxicity of various carbon nanomaterials in human epithelial colorectal adenocarcinoma (Caco-2) cells and human breast adenocarcinoma (MCF-7) cells. Carbon nanohorns (CNH), carbon nanotubes (CNT), carbon nanoplatelets (CNP), graphene oxide (GO), reduced graphene oxide (GO) and nanodiamonds (ND) were systematically evaluated and compared using Pluronic F-127 as a dispersant agent. Carbon nanomaterial exposure affected the cell viability in the following order: CNP < CNH < RGO < CNT < GO < ND, with a pronounced effect on the more rapidly dividing Caco-2 cells. Hydrophobicity and morphological features are the main causes of decreases in cell viability, enhanced levels of ROS (radical oxygen species) and apoptosis/necrosis. In this study, ND showed low toxicity thanks to a lack of ROS levels and was efficient in the loading of hydrophilic drugs, such as DOX, by assembling in the ND surface or within the pores. On the other hand, this study evidenced the low toxicity of CNT and RGO, and the high camptothecin (CPT) loading because of the strong π–π stacking interactions. Altogether, despite the various obstacles that still have to be overcome before considering carbon nanomaterials suitable as drug carriers (i.e., the potential long-term toxicity), this study highlighted a screening and risk-to-benefit assessment and, together with drug-loading efficiency studies, is fundamental to the development of advanced multi-functional carbon nanomaterials for cancer theragnostic applications [7].

Nanodiamonds with detonation origins were investigated by Claveau et al. as delivery systems for anti-cancer therapy in vivo models. The authors studied the ability of cationic hydrogenated detonation nanodiamonds to carry active small interfering RNA (siRNA) in a mice model of Ewing sarcoma, which is bone cancer of young adults due to the *EWS-FLI1* junction oncogene in the majority of patients. Labeled nanodiamonds obtained using radioactive tritium gas instead of hydrogen gas allowed the authors to investigate the trafficking of nanodiamonds throughout mouse organs and their excretion as urine and feces. Moreover, the ability of siRNA to inhibit the expression of the oncogene *EWS-FLI1* in tumor-xenografted mice was demonstrated. Overall, this study represents a substantial step towards the use of ultra-small solid nanoparticles for the delivery of nucleic acid in vivo [8].

In the framework of carbon nanomaterial development for therapeutic purposes, Lee et al. reported a new type of carbon dot (CDOT) nanoparticle as a new antiplatelet agent. The inhibition of platelet activation is considered a potential therapeutic approach for the treatment of arterial thrombotic diseases; therefore, maintaining platelets in their inactive state has gained much attention. CDOT could actively inhibit human platelet activation by suppressing some crucial mechanisms (e.g., PKC activation, and Akt, JNK1/2 and p38 MAPK phosphorylation) with no in vitro cytotoxicity. This in vivo study revealed that the CDOTs had an antithrombotic effect on the ADP-induced pulmonary thromboembolic mice model by reducing mortality and by preserving the normal bleeding tendency in mice. Altogether, these results suggest that a direct application of CDOTs may contribute to the development of new antiplatelet drugs for the treatment of arterial thromboembolic diseases [9].

In the application of graphene nanomaterials for dental regenerative engineering, Di Carlo et al. proposed the covalent functionalization of a graphene oxide (GO)-decorated cortical membrane (Lamina®) in the promotion of the adhesion, growth and osteogenic differentiation of DPSCs (Dental Pulp Stem Cells). The GO-decorated Laminas demonstrated an increase in the roughness of Laminas and a reduction in toxicity and did not affect the membrane integrity of DPSCs. In conclusion, this study showed that the GO covalent functionalization of Laminas was effective; was relatively easy to obtain; and favored both the proliferation rate of DPSCs, probably due to the capacity of GO to adsorb proteins present in the medium, and the deposition of calcium phosphate. Overall, this study is promising because the proposed material holds potential as a useful substrate in facilitating in vivo bone regeneration [10].

In antibacterial applications, Nicosia et al. synthesized novel NanoHybrid Systems based on graphene, polymers and AgNPs (namely, NanoHy-GPS) using an easy microwave irradiation approach free of reductants and surfactants. The fine-tuned hybrid system combines the properties of polymers, graphene and AgNPs as a potential on-demand antimicrobial coating system. Polymers play key roles in ensuring the coating compatibility of the graphene platform, making it adaptable for a specific substrate. The driving force of this strategy is the tuning of the interfacial interactions towards targeted substrates, thus optimizing the homogeneity of the dispersion of the GO derivatives within specific polymer matrices. The formulation of functionalized graphene with AgNPs entrapped in suitable polymers resulted in a doubly beneficial effect: an increase in graphene processability and achievement of graphene-enriched antimicrobial nanobiomaterials. NanoHy-GPS was proposed as a potential alternative to common antibacterial agents, which leak into the environment and/or within organisms' tissues [11].

Finally, Cordaro et al. proposed a review aimed at providing a comprehensive and exhaustive summary of the contributions of graphene-based nanomaterials to liquid biopsy. Liquid biopsy is considered an innovative method that has provided surprising perspectives in the early diagnosis of severe diseases such as cancer, metabolic syndrome, and autoimmune and neurodegenerative disorders and in monitoring their treatment. Although nanotechnology based on graphene has been poorly applied for the rapid diagnosis of viral diseases, the extraordinary features of graphene (i.e., high electronic conductivity, large specific area and surface functionalization) can also be exploited for the diagnosis of emerging viral diseases, such as coronavirus disease 2019 (COVID-19) [12].

The variety of applications covered by the nine articles published in this Special Issue of *Nanomaterials* "Carbon Nanomaterials for Therapy, Diagnosis and Biosensing" is proof of the growing attention on the use of carbon nanomaterials in the biomedical/pharmaceutical field in recent years. We hope that the readers enjoy reading these articles and find them useful for their research and for advancing carbon nanomaterials from the laboratory to clinical nanomedicine. Finally, we acknowledge all of the authors who contributed their work to this Special Issue as well as the editorial board of *Nanomaterials* for all of their support.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### **Novel Nanohybrids Based on Supramolecular Assemblies of Meso-tetrakis-(4-sulfonatophenyl) Porphyrin J-aggregates and Amine-Functionalized Carbon Nanotubes**

#### **Mariachiara Trapani 1, Antonino Mazzaglia 1,\*, Anna Piperno 2,3, Annalaura Cordaro 1,2, Roberto Zagami 1, Maria Angela Castriciano 1,\*, Andrea Romeo 1,2,4 and Luigi Monsù Scolaro 1,2,4**


Received: 13 February 2020; Accepted: 26 March 2020; Published: 2 April 2020

**Abstract:** The ability of multiwalled carbon nanotubes (MWCNTs) covalently functionalized with polyamine chains of different length (ethylenediamine, EDA and tetraethylenepentamine, EPA) to induce the J-aggregation of meso-tetrakis(4-sulfonatophenyl)porphyrin (TPPS) was investigated in different experimental conditions. Under mild acidic conditions, protonated amino groups allow for the assembly by electrostatic interaction with the diacid form of TPPS, leading to hybrid nanomaterials. The presence of only one pendant amino group for a chain in EDA does not lead to any aggregation, whereas EPA (with four amine groups for chain) is effective in inducing J-aggregation using different mixing protocols. These nanohybrids have been characterized through UV/Vis extinction, fluorescence emission, resonance light scattering and circular dichroism spectroscopy. Their morphology and chemical composition have been elucidated through transmission electron microscopy (TEM) and scanning transmission electron microscopy (STEM). TEM and STEM analysis evidence single or bundles of MWCNTs in contact with TPPS J-aggregates nanotubes. The nanohybrids are quite stable for days, even in aqueous solutions mimicking physiological medium (NaCl 0.15 M). This property, together with their peculiar optical features in the therapeutic window of visible spectrum, make them potentially useful for biomedical applications.

**Keywords:** porphyrin; J-aggregates; carbon nanotubes; nanohybrids

#### **1. Introduction**

Carbon nanotubes (CNTs) are intriguing materials with applications ranging from nanotechnology-related devices (i.e., in electronics, energy storage, water treatment, as sensor/biosensor) [1] to drug/probes delivery systems for therapy and diagnosis [2,3].

Functionalized CNTs are widely used to reduce the intrinsic toxicity of "as produced" (pristine) CNTs by increasing the tolerability and the biodegradability in vivo [4,5]. Furthermore, opportunely modified and sized multiwalled CNTs (MWCNTs) are not retained in the organs and can be easily cleared by body excretion [6].

Functional nanomaterials based on CNTs were designed as therapeutic enhancers by combining CNTs with different systems, such as cyclodextrins, biomolecules or porphyrinoids [3]. Recently, some of us reported antiviral- and plasmid/delivery systems [7,8] based on properly functionalized MWCNTs, investigating also their intracellular fate. Similarly to other nanomaterials based on carbon for multi-targeted therapies and imaging [9,10], CNTs functional nanomaterials were endowed with unique properties generated by the synergic actions of components [3].

Non-covalent modification of CNTs with porphyrinoids is a well-investigated strategy to modulate the environment of chromophores, thus improving the light absorption and emission features [11], charge-transport [12] or energy transfer properties [13] in view of bio-labeling and light harvesting applications. In particular, it is well-known as π-stacking of CNTs with hydrosoluble porphyrins provides donor-acceptor complexes with efficient energy transfer [14]. Porphyrin free bases or metallo porphyrins/DNA supramolecular systems undergo strong charge transfer with semiconducting CNTs [15]. Enhanced photoconductivity has been reported for J- and H- porphyrin aggregates (head-to-tail or head-to-head molecules stacking, respectively) obtained in solution by interaction with single-walled CNTs (SWCNTs) [16] or at solid state with double-walled CNTs (DWCNTs) [17] or by decorating MWCNTs film [18]. Moreover, recently it was demonstrated that photoluminescence properties of a hybrid material assembled by formation of J-aggregates of benzo[e]indocarbocyanine (BIC) on SWCNTs can be modulated by selecting cis- or trans- isomer of the dye: the first one quenches the photoluminescence by strong interaction with CNTs, whereas the second one forms free J-aggregates characterized by photoluminescence bands of practical use in biomedical imaging [19]. Indeed, it is well known that J-aggregates feature very narrow red-shifted absorption bands, showing renewed optical, photophysical, and structural properties vs. monomer [20]. In this framework, multifunctional nanotheranostic based on J-type aggregates of cyanine [21,22], bacterio-pheophorbide [23], chlorine [24], and Bodipy [25] were proposed due to their excellent photothermal and/or NIR absorbing features for applications in imaging guided therapy (i.e., photoacustic imaging). However, these J-type aggregates generally need to be entrapped in liposomes or dispersed in surfactants to increase their solubility and bio-availability.

Within the incoming research of composite nanomaterials, our interest has been addressed to J-aggregates of meso-tetrakis-(4-sulfonatophenyl)porphyrin (TPPS) exhibiting peculiar optical features [26–29]. Such features can be fine-tuned depending on the strategy adopted to obtain the structure, i.e., by selecting appropriate polyamines as scaffolds [30–35] or by tailoring nanomaterials (i.e., metal nanoparticles) [36–40] or triggering porphyrin J-aggregation by modulating pH and/or ionic strength [41–44]. TPPS J-aggregates, in line of principle, would not necessitate further manipulation/encapsulation to explicate their properties within cells or tissues. These aggregated species could lead to stimuli-responsive therapeutic action upon irradiation on their extinction bands, fluorescence probing in cellular environments, or refilling of the dye upon eventual J-aggregates disassembly [45,46] in biological sites [47].

Regarding the design of SWCNTs/TPPS nanohybrids, assembly between TPPS with amine-conjugated SWCNTs has been obtained in water, pointing out that the photophysical properties of porphyrin are largely influenced by length of CNTs amine chain [48] whereas TPPS J-aggregates on SWCNTs have been prepared in organic solvents [49].

Herein, we report on the formation of relatively stable TPPS J-aggregates wrapped to covalently amine-modified MWCNTs in aqueous solution. We anticipate that, in the presence of tetraethylenepentamine-functionalized MWCNTs (MWCNT-EPA) in mild acidic conditions, TPPS porphyrin easily self-assembles into J-aggregates exhibiting peculiar extinction bands in the visible region (i.e., -493 nm) and in the therapeutic window (i.e., -710 nm), together with an emission band in the red spectral region for potential phototherapeutic and/or photodiagnostic applications. Conversely, ethylenediamine-modified MWCNTs (MWCNT-EDA) do not induce J-aggregate formation due to EDA structural features vs. EPA ones. The MWCNT-EPA/TPPS J-aggregates are stable in mimicking

physiological medium (NaCl = 0.15 M), thus opening the route for their potential application in dual therapeutic/diagnostic assessment.

#### **2. Materials and Methods**

The 5,10,15,20-tetrakis(4 sulfonatophenyl)porphyrin (TPPS), MWCNTs, tert-butyl 2-aminoethylcarbamate (EDABoc), tetraethylenepentamine (EPA), N-(3-dimethylaminopropyl)-N- -ethylcarbodiimide hydrochloride (EDC), 1-Hydroxybenzotriazole (HOBt), Ninhydrin test kit, other solvents and reagents were purchased from Sigma-Aldrich Chemicals (Milan, Italy). The stock porphyrin aqueous solutions were freshly prepared and their concentrations were determined using the extinction coefficient at the Soret maximum (<sup>ε</sup> = 5.33 <sup>×</sup> 105 M−1cm−<sup>1</sup> at λ = 414 nm). All the reagents were used without further purification and all the solutions were prepared in dust free Milli-Q water (Merck, Darmstadt, Germany).

Carboxylated multiwalled carbon nanotubes (MWCNT-Ox) were prepared by the oxidation of MWCNTs (mean diameter 5–10 nm; average length 10–20 μm) with sulfuric acid/nitric acid (3:1 *v*/*v*, 98% and 69%) according to the protocol previously reported [50]. MWCNT-Ox (100 mg) in 25 mL of dry DMF were sonicated for 30 min, then EDC (112 mg, 0.58 mmol) and HOBt (40 mg, 0.30 mmol), were added and the black suspension was stirred at room temperature for one hour. EDABoc (55 mg, 0.24 mmol) was added and the reaction was stirred for 72 hours. Water/ethanol (1:1 mixture) was added and the crude reaction mixture was filtered under vacuum (Millipore, 0.1 μm), washed with an excess of water/ethanol and finally diethyl ether. The resulting MWCNT-EDABoc were dispersed in dioxane (10 mL) and treated with 5 mL of HCl 4 M at room temperature for 4 h. The mixture was filtered under vacuum (Millipore, 0.1 μm) and the precipitate was treated with 5 mL of triethylamine/water (1:4), thus obtaining MWCNT-EDA. This was washed several times with water/ethanol by successive bath sonication and centrifugation (8000 rpm 10 min) procedures and finally dried at 60 ◦C to give 60 mg of material. The amount of free amine groups on MWCNT-EDA was estimated by Kaiser test (0.22 mmol/g).

MWCNT-EPA were prepared by the coupling of MWCNT-Ox and EPA, in presence of EDC/HOBt, according to the amidation reaction procedure above described. The resulting MWCNT-EPA were washed several times with water/ethanol by successive bath sonication and centrifugation (8000 rpm 10 min) procedures and finally dried at 60 ◦C. Termogravimetric analysis (TGA) data indicated a weight loss of about 4.3% at 500 ◦C, which roughly corresponds to 0.22 mmol/g of EPA. The ninhydrin assay indicated an amount of free amino groups of 0.44 mmol/g.

Primary amine loadings were measured spectroscopically using the colorimetric Kaiser conditions [51,52]. Commercial Kaiser test kit is composed of three solutions as it follows: (a) 0.5 g/mL of phenol in absolute EtOH; (b) 2 mL of potassium cyanide 1 mM (aqueous solution) dissolved in 98 mL of pyridine; (c) 0.05 g/mL of ninhydrin in absolute EtOH. Briefly, 0.5 mg of MWCNT-EDA or MWCNT-EPA were treated in sequence with 75 μL of solution (a), 100 μL of solution (b) and 75 μL of solution (c). The dispersion was sonicated in a water bath and then was heated at 120 ◦C for 5 min, diluted with 4750 μL of absolute EtOH and centrifugated at 14,000 rpm. The absorbance at 570 nm of supernatant was correlated to the amount of free amine groups on MWCNTs surface (NH2 loading (mmol/g)); using the following equation:

$$[frec\,\text{amines}] = ([\text{Abs}] \times \text{dilution} \times 1000)/(\varepsilon \times \text{sample weight} \times \text{optical path}) \tag{1}$$

where dilution was fixed to 5 mL, optical path was 1 cm; sample weight was 0.5 mg; extinction coefficient (ε) was 15,000 M−<sup>1</sup> cm<sup>−</sup>1.

Dispersions of MWCNT-EDA and MWCNT-EPA (0.43 mg/mL) were prepared in 10 mM citrate buffer by bath sonication for 20 min. For the experiments, a volume of 100 μL has been used and diluted to a final concentration of 0.02 mg/mL. The interaction of TPPS with the two batches of MWCNTs has been investigated in citrate buffer solution (10 mM, pH 2.4) following two different mixing order procedures: (i) porphyrin-first protocol (PF) and (ii) porphyrin- last protocol (PL), consisting of the addition of a proper volume of MWCNTs dispersion to a diluted TPPS solution in citrate buffer and of the addition of TPPS from a stock solution to a diluted MWCNTs dispersion, respectively [43,53]. In some experiments MWCNT- EPA in citrate buffer has been previously mixed with NaCl (0.15 M), followed by the addition of TPPS. In all the experiments, the final concentration of TPPS was 5 μM.

UV-Vis spectra have been collected on a diode-array spectrophotometer Agilent model 8452. The circular dichroism (CD) spectra were recorded on a JASCO J-720 spectropolarimeter, equipped with a 450 W xenon lamp. CD spectra were corrected both for the cell and buffer contributions. A Jasco mod. FP-750 spectrofluorometer has been used to record fluorescence emission and Resonance Light Scattering (RLS) spectra. Emission spectra were not corrected for the absorption of the samples and a synchronous scan protocol with a right angle geometry was adopted for collecting RLS spectra [54]. All the aqueous dispersions were analysed by using a 1 cm optical path cuvette.

TGA was performed by using PerkinElmer Instruments Pyris1 TGA at a heating rate of 10 ◦C/min over the range from room temperature (r.t) to 1000 ◦C under N2 atmosphere.

A TEM, JEM2100 LaB, working at 100 kV, and a digital Scanning transmission electron microscopy (STEM) set with BF & DF STEM Detectors plus SE/BSE detector (University of Exeter, UK) were used to investigate morphology of MWCNTs and MWCNTs/TPPS J-aggregates. Samples were prepared by evaporating ten drops of the aqueous dispersions of the investigated system more days after mixing (1–3 days) on 300 mesh holey-carbon coated copper grids.

#### **3. Results and Discussion**

#### *Synthesis and Characterization of Amine Multiwalled Carbon Nanotubes*

Amine multiwalled carbon nanotubes, MWCNT-EDA and MWCNT-EPA, were prepared by coupling of carboxylated MWCNTs (MWCNT-Ox) with tert-butyl 2-aminoethylcarbamate (EDABoc) or tetraethylenepentamine using EDC/HOBt in DMF according to Figure 1A,B, respectively. MWCNT-Ox were prepared by the oxidation (HNO3/H2SO4 1:3, 6 h, 60 ◦C) of commercially available multiwalled carbon nanotubes according to a previously reported procedure [8,50].

**Figure 1.** Schematic representations of MWCNT-EDA (**A**) and MWCNT-EPA (**B**) synthetic procedures.

The degree of functionalization of MWCNTs was investigated by TGA analysis (Figure 2) and the primary amines loadings was determined using the colorimetric Kaiser test [51,52].

**Figure 2.** TGA profiles of MWCNT-Ox (dark line) and MWCNT-EPA (red line) (**A**). TGA profiles of MWCNT-EDABoc (dark line) and MWCNT-EDA (red line) (**B**). In the insets DTG curve of MWCNT-EPA (A) and MWCNT-EDABoc (B). TGA analyses were carried out under N2 atmosphere.

The TGA curve of MWCNT-Ox shows a gradual weight loss of about 15.5% at 500 ◦C (Figure 2A). MWCNT-EPA TGA profile displays two weight loss steps in the range 100–400 ◦C, likely due to decomposition of polyamine alkyl chain (inset of Figure 2A). From TGA data, a weight loss of about 4.3% at 500 ◦C which roughly corresponds to 0.22 mmol/g of EPA (Figure 2A) has been estimated. By the correlation with the absorbance at 570 nm using the ninhydrin assay, an amount of free amino groups of 0.44 mmol/g has been determinated, probably suggesting a role of secondary amine groups in the amidation reactions (Figure 1B).

According to literature data, a higher thermal stability of MWCNTs containing free amine groups has been detected with respect to the MWCNTs sample containing Boc-amine groups (MWCNT-EDA vs. MWCNT-EDABoc, Figure 2B) [55,56]. The significant weight loss of MWCNT-EDABoc in the range 100–300 ◦C (see DTG, inset Figure 2B) can be attributed to the thermal decomposition and rearrangement of the tert-butoxyl groups. On the basis of TGA data, it was not realistic to determine the degree of functionalization in terms of weight loss (Δm ≈ 0.9–1%) [57]. Thus, we have estimated the amount of free amine functional groups by the Kaiser test (0.22 mmol/g).

TEM analyses of functionalized MWCNTs indicated that the chemical functionalization with amine groups preserved the characteristic morphology of multiwalled tubes scaffold. An average external diameter of ~10 nm, corresponding to an average number of 8–10 layers were found. MWCNT-ox appeared strongly aggregates in bundles (Figure 3A), whereas well distinct isolated MWCNTs are observed in TEM image of amine functionalized MWCNTs (MWCNT-EDA, Figure 3B and MWCNT-EPA, Figure 3C). Moreover, all the functionalized MWCNTs (MWCNT-Ox, MWCNT-EPA and MWCNT-EDA) were shortened by oxidation: the length was reduced from the micrometre (pristine MWCNTs, see Supplementary Figure S1) to nanometre scale [8,50].

**Figure 3.** TEM images of MWCNT-Ox (**A**), MWCNT- EDA (**B**) and MWCNT-EPA (**C**).

In order to obtain MWCNT-EDA/TPPS J-aggregates hybrids, the two aforementioned mixing order protocols (both PF and PL) have been used under mild acidic condition. Whatever of the procedure employed, we found that the extinction features of free diacid porphyrin (B band centered at 434 nm and Q bands at 592 and 645 nm) remained unchanged even after 1 day (Figure 4A). No TPPS J aggregates were formed and no interaction between TPPS and MWCNT-EDA was revealed. This behavior could be ascribable to the presence of only one pendant amino group for chain in EDA, and this observation agrees with previous results on the role of the amine chain length in inducing the aggregation of TPPS [53].

**Figure 4.** UV-Vis spectra of MWCNT-EDA (black line) and upon TPPS addition (red line), and 1 day after mixing (blue line) (**A**), and of MWCNT-EPA (black line), upon TPPS addition (red line), and 1 day after mixing (blue line) (**B**). Experimental conditions: [TPPS] = 5 μM; MWCNT-EDA or MWCNT-EPA = 0.02 mg/mL; 10 mM citrate buffer at pH 2.4; PL protocol; T = 298 K.

On the other hand, upon the addition of the porphyrin to the MWCNT-EPA dispersion, the UV-Vis profile shows the spectral signatures both of the diacid form of TPPS (B-band at 434 nm) and of J-aggregates evidenced by their typical extinction band arising at 489 nm. During the time, we observed the decrease of the intensity of the diacid band accompanied by the increase of the intensity of J-band, which furthermore undergoes a bathochromic shift from 489 to 493 nm. After one day, at the end of the aggregation process, UV-Vis spectrum of MWCNT-EPA/TPPS exhibits the B and Q bands ascribable to the residual monomeric diacid form of TPPS and J-aggregates (Figure 4B). It is noteworthy that the formation of J-aggregates does not occur under the same experimental conditions in absence of MWCNT-EPA, but it can be forced by decreasing the pH of the medium [42]. On the bases of the experimental evidences, we suggest that only EPA functionalized MWCNTs are able to trigger the TPPS aggregation process. This could be ascribable to the occurrence of an initial electrostatic interaction among a sufficient number of positively charged protonated amino groups on the CNTs surface and negatively charged sulfonated groups present in the periphery of the dyes [31,58]. Moreover, the observed red-shift of the extinction B-band of J-aggregates in the time could suggest a rearrangement of the aggregates due to an interaction with different amine-modified carbon nanotubes or their location in a different microenvironment with respect to the aqueous solution.

The emission spectrum (λexc 455 nm), at the end of the aggregation process, shows the typical fluorescence emission of the diacid form of TPPS centered at 669 nm. Further, upon excitation on the J-aggregates band (λexc 493 nm), an emission band at 717 nm can be also detectable (Figure 5). The different ratio of the bands intensity at 669 and 717 nm at the two distinctive excitation wavelengths is due to the fluorescence emission generated from the aggregated species. This evidence is confirmed by the related excitation spectra (Figure 5 inset) showing spectral features for both the diacidic TPPS and J-aggregates. However, the monomeric porphyrin is the predominant species in the excitation profile at both emission wavelengths, due to its longer lifetime value with respect to the J-aggregate one [59,60].

**Figure 5.** Fluorescence emission spectra of MWCNT-EPA/TPPS J-aggregates system 1 day after mixing (dashed and solid lines recorded at λexc = 455 and 493 nm, respectively) and, in the inset, the corresponding excitation spectra (dashed and solid lines recorded at λem = 669 and 717 nm, respectively). Experimental conditions: [TPPS] = 5 μM; MWCNT-EPA = 0.02 mg/mL; 10 mM citrate buffer at pH 2.4; PL protocol; T = 298 K.

The RLS spectra recorded soon after mixing shows a very sharp peak in the red region of the extinction band which increases in intensity at the end of the aggregation process. These findings agree with our previous results [42], pointing to the formation of self-assemblies of electronically coupled porphyrins, which cause a large enhancement of the resonant light scattering [54] at the red-edge of the extinction peak (Figure 6).

J-aggregates of TPPS induced by MWCNT-EPA were stable in acidic aqueous dispersion for a day or more after preparation. In difference with our previous findings on polyamine-mediated J-aggregates [30,31,61], the optical profiles are in line with the usual Frenkel exciton theory, rather than in terms of an extended network formed by the J-aggregates and amine modified MWCNTs in which dipole–dipole coupling among single porphyrins takes place.

**Figure 6.** RLS spectra of MWCNT-EPA (black line), after TPPS addition (red line), and 1 day after mixing (blue line). Experimental conditions: [TPPS] = 5 μM; MWCNT-EPA = 0.02 mg/mL; 10 mM citrate buffer at pH 2.4; PL protocol; T = 298 K.

CD spectra were recorded after freshly mixing of the components and at the end of the aggregation process (Figure 7). As expected for achiral MWCNT-EPA, CD spectrum is silent before the addition of the chromophoric species, so confirming the absence of optical activity for amine modified CNTs. On the other hand, when TPPS was added a slight bisegnate positive Cotton effect in the aggregates absorption region has been observed. At the end of the aggregation process, an increase in intensity and a red shift of the CD profile were observed. This behavior, observed by means of spectroscopic and light scattering techniques, is due to the formation of large and rearranged structures as the result of the interactions among porphyrin aggregates and functionalized MWCNTs.

**Figure 7.** CD spectra of MWCNT-EPA (black line), soon after TPPS addition (red line), and 1 day after mixing (blue line). Experimental conditions: [TPPS] = 5 μM; MWCNT-EPA = 0.02 mg/mL; 10 mM citrate buffer at pH 2.4; PL protocol; T = 298 K.

In agreement with the spectroscopic characterization, representative TEM images of MWCNT-EPA/TPPS J-aggregates pointed to the coexistence of both MWCNT-EPA and TPPS J-aggregates in the same area (Figure 8 and Supplementary Figure S2). In particular, TPPS J-aggregates with an average diameter of 45 nm and length of 250–500 nm seem to be wrapped by separate nanotubes or bundles of MWCNT-EPA having an average external diameter of about 15 nm.

**Figure 8.** TEM images of MWCNT-EPA/TPPS J-aggregates at low (**A**) and high resolution (**C**): (**B**,**D**) corresponds to STEM analysis of total (C/N/O/S) and C/S merging of elements distribution respectively taken within the dashed region of the assemblies in (**C**) (for individuals colours of elements refers to (**B**) in Figure S3; see Materials and Methods for preparation conditions).

STEM analysis of MWCNT-EPA/TPPS J-aggregates shows the total elemental distribution pointing out the presence of carbon and oxygen for MWCNT-EPA, and carbon, oxygen, sulfur and nitrogen for TPPS J-aggregates (Supplementary Figure S3). Interestingly, in the marked area (Figure 8C), the total elements merging (Figure 8B) appears to be similar to the carbon/sulfur merging (Figure 8D). These results evidence the co-localization of TPPS J-aggregates and carbon nanotubes in the investigated samples. Since self-organization phenomenon is a hierarchical process, it is well known as the morphology of final aggregates can be controlled by the mixing order protocol [43]. In this framework, we performed experiments by adding amine carbon nanotubes to diacid TPPS (PF protocol). Surprisingly, at the end of the aggregation process, all the spectroscopic (Supplementary Figures S4 and S5) and morphological features (Supplementary Figure S6) show no change with respect to that observed by previous reagent mixing order protocol (PL). Because, in the case of the self-aggregation of neat TPPS in acidic conditions [43], the mixing order protocol is strictly related to the occurrence of porphyrin nucleation phenomena, here, we are prone to think that MWCNT-EPA could act as nucleation centers, inducing dye aggregation independently by the mixing order protocol.

In order to verify the stability of MWCNT-EPA/TPPS J-aggregates in mimicking physiological medium (NaCl 0.9% *w*/*w* - 0.15 M), the system has been prepared by firstly dispersing MWCNT-EPA in NaCl 0.15 M aqueous solution and then adding TPPS. Under these conditions, the spectroscopic evidences of the final system remain almost unchanged (Figure 9) with respect to the unsalted solutions thus confirming the formation of chromophoric assemblies. The premixing of MWCNTs and NaCl, followed by the addition of porphyrin, seems to lead to a larger amount of J-aggregates due to the ionic strength effect [62]. Generally, optical stability for J-aggregates is difficult to achieve. Therefore, the use of surfactans or the entrapment in liposomes of the dye forming J-aggregates were experienced in literature [21]. In our case, MWCNTs induce, whatever the preparation procedure, the formation of stable J-aggregates able to retain their optical properties even after more days (1–3 days). Therefore, no further manipulation to preserve their pristine optical properties was necessary.

**Figure 9.** UV- Vis spectra (**A**) and RLS ( **B**) of MWCNT-EPA (black line), upon TPPS addition (red line) and 4 h after mixing (blue line); (**C**) Fluorescence emission spectra of MWCNT-EPA/TPPS J-aggregates (dashed and solid lines at λexc = 455 and 492 nm, respectively) and (**D**) the corresponding excitation spectrum ( dashed and solid lines at λem = 669 and and 717 nm, respectively). Experimental conditions: [TPPS] = 5 μM; MWCNT-EPA = 0.02 mg/mL; 10 mM citrate buffer at pH 2.4; [NaCl] = 0.15 M, PL protocol; T = 298 K.

Altogether, the combination of drug carrier ability of MWCNTs with the theranostic properties of porphyrins could allow the development of MWCNTs/TPPS J-aggregates nanohybrids for applications in biomedical field. Unlike from the others families of hybrid carbon nanomaterials [9], the potential applications in biological/pharmaceutical field of carbon based nanomaterials- porphyrins appear still scarcely investigated, especially for J-aggregates self-assemblies as well as their intracellular trafficking, therapeutic and imaging properties. In the literature, it was observed that hybrids nanomaterials based on porphyrinoids [24,63] were prepared at pH different from physiological conditions, and then treated with cells. In this context, future work will be devoted to studying the biocompatibility and cellular uptake of our hybrid MWCNTs/J-aggregates supramolecular systems. With these perspectives in mind, this research could lay the groundwork for the incoming biological assessment of MWCNT-EPA/TPPS J-aggregates.

#### **4. Conclusions**

MWCNTs can be easily functionalized by covalently introducing pendant amino-groups on their surface. In this paper we used ethylenediamine (EDA) and tetraethylenepentamine (EPA), which after coupling with carboxylic groups on the exterior walls, led to one and four protonable amino groups for chain, respectively. Under mild acidic conditions, the diacid form of TPPS is able to electrostatically bind to the surface and eventually aggregate. In line with our previous investigations on the ability of polyamines to trigger the aggregation of TPPS, the presence of only one pendant amino group (EDA) is not enough to induce the formation of TPPS J-aggregates, whatever the mixing protocol. On the other hand, when EPA functionality is present, these species are effective to generate stable MWCNT-EPA/TPPS J-aggregates nanohybrids and their general spectroscopic features are rather independent on the mixing protocol. A similar behavior was observed in solutions mimicking physiological medium (NaCl - 0.15 M), whereby stable nanohybrids were also obtained. These systems exhibit remarkable optical features, and in this perspective could be considered for potential applications in phototherapy (by irradiating on extinction bands at 491 nm and/or at 709 nm) and/or bio-imaging (by exploiting the fluorescent emission band at 716 nm). In this respect, this class of amine-modified MWCNTs could be investigated as carriers of J-aggregates in biological environment. All these optical and structural properties make MWCNT-EPA/TPPS J-aggregates appealing for further considerations in theranostic.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2079-4991/10/4/669/s1, Figure S1: TEM image of pristine MWCNTs; Figure S2: TEM images of MWCNT-EPA/TPPS J-aggregates; Figure S3: STEM of MWCNT-EPA/TPPS J-aggregates; Figure S4: UV-Vis spectra of an aqueous solution of TPPS J-aggregates/MWCNT-EPA (PF protocol) and MWCNT-EPA TPPS J-aggregates (PL protocol); Figure S5: CD spectra of an aqueous solution of TPPS J-aggregates/MWCNT-EPA (PF protocol) and MWCNT-EPA/TPPS J-aggregates (PL protocol); Figure S6: TEM images of TPPS J-aggregates/MWCNT-EPA (PF protocol).

**Author Contributions:** Conceptualization, A.M., M.A.C., A.R. and L.M.S.; investigation, M.T., A.M., A.P. and A.C.; data curation, M.T., A.M. and A.P.; writing—original draft preparation, M.T., A.M. and L.M.S.; writing—review and editing, A.M., R.Z., M.A.C. and A.R.; visualization, R.Z., M.A.C. and A.R.; Supervision, L.M.S. All authors discussed the results and commented on the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by CNR (Project ISMN-CNR: Materials and Dispositives for Health and Life Quality) for financial support.

**Acknowledgments:** We are grateful to Yanqui Zhu and Hong Chang (University of Exeter) for kind assistance with TGA, TEM and STEM analyses.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **SERS Sensing Properties of New Graphene**/**Gold Nanocomposite**

#### **Giulia Neri 1, Enza Fazio 2,\*, Placido Giuseppe Mineo 3,4, Angela Scala <sup>1</sup> and Anna Piperno 1,\***


Received: 19 July 2019; Accepted: 28 August 2019; Published: 30 August 2019

**Abstract:** The development of graphene (G) substrates without damage on the sp2 network allows to tune the interactions with plasmonic noble metal surfaces to finally enhance surface enhanced Raman spectroscopy (SERS) effect. Here, we describe a new graphene/gold nanocomposite obtained by loading gold nanoparticles (Au NPs), produced by pulsed laser ablation in liquids (PLAL), on a new nitrogen-doped graphene platform (G-NH2). The graphene platform was synthesized by direct delamination and chemical functionalization of graphite flakes with 4-methyl-2-*p*-nitrophenyl oxazolone, followed by reduction of *p*-nitrophenyl groups. Finally, the G-NH2/Au SERS platform was prepared by using the conventional aerography spraying technique. SERS properties of G-NH2/Au were tested using Rhodamine 6G (Rh6G) and Dopamine (DA) as molecular probes. Raman features of Rh6G and DA are still detectable for concentration values down to 1 <sup>×</sup> <sup>10</sup>−<sup>5</sup> M and <sup>1</sup> <sup>×</sup> <sup>10</sup>−<sup>6</sup> M respectively.

**Keywords:** graphene/gold nanocomposite; SERS; Dopamine; Rhodamine 6G

#### **1. Introduction**

Since their discovery, graphene materials (G), due to their outstanding physicochemical properties [1], have generated huge interest in numerous fields including biomedicine, electronics, sensing, energy, etc. [2–8]. They have been proposed as drug delivery systems for photothermal [9] and photodynamic therapy [10], as scaffold in tissue engineering [11], and as materials for biosensing [12,13]. Recently, G and its functionalized derivatives have been investigated as substrates for SERS (surface enhanced Raman spectroscopy) applications [14,15], a versatile technique that enables the rapid detection of various types of molecules [16,17].

Metal nanoparticles (i.e., Cu, Ag, gold nanoparticles (Au NPs)) are the most extensively studied SERS-active substrates since their collective electronic excitations, namely surface plasmons, are very interesting for a large variety of applications. Localized surface plasmon resonance excitation in Ag and Au NPs produces strong extinction and scattering spectra, resulting in amplification of the electric field (E) near the particle surfaces such that |E| <sup>2</sup> can be 100–10,000 times greater in intensity than the incident field, which acts on a spatial range of 10–50 nm. These effects are mainly influenced by two factors: (i) NPs morphology (in terms of size and shape) and (ii) local dielectric environment [18–20].

Two main mechanisms are involved in Raman signal enhancement: The electromagnetic mechanism (EM), due to the strong amplification of the local EM field [21], and the chemical effect (CM) that involves the creation of new electronic states generated by the interaction between the metal and the molecules adsorbed on it [22]. Such new electronic states allow for resonant Raman scattering processes; the control of the distances among the localized surface plasmons, on a sub-nanometer scale, is a critical parameter to control the inter-particle optical coupling and therefore, the efficiency of SERS response [23].

Recently, engineered G have attracted a huge amount of attention as platforms for biological SERS sensing [24]. G offer a large flat surface to adsorb molecules through π–π interactions determining the manner in which molecules bind with the surface which, in turn, determines the symmetry of the molecules and the effective charge transfer [25]. However, G alone provides a limited enhancement factor [26] while the combination of specifically designed G with metallic NPs is an interesting strategy to obtain new materials with synergistic effect and improved SERS sensing performance. The high compatibility of G with metal noble NPs is mainly due to: (i) Transparency to laser light and localized plasmonic fields; (ii) high thermal conductivity; and (iii) appropriate dielectric strength that confines the plasmonic field [27,28].

Despite the potentiality of these hybrid systems, the critical point of G-based SERS substrates regards the development of chemical strategies avoiding the damage of the G sp<sup>2</sup> network, keeping an electron high mobility and, at the same time, enabling the tuning of the interactions with plasmonic surface to finally enhance SERS effects. Generally, 2D materials were obtained by liquid chemical exfoliation of related 3D stratified bulk materials, processes that required the presence of intercalation agents and ultrasonication treatment. To overcome the long processing times and to guarantee the quality of 2D substrates, the exfoliation methods have been continuously implemented [29,30]. Recently, we have developed a straightforward method for the direct delamination of graphite flakes into functionalized G with preserved sp<sup>2</sup> network [31]. G-MNPO platform (Figure 1) [31], obtained by solvent-free 1,3-dipolar cycloaddition reaction of 4-methyl-2-*p*-nitrophenyl oxazolone with graphite, was selected for the development of a new nitrogen-doped graphene network (G-NH2). The amine groups, obtained by reduction of *p*-nitrophenyl group on the Δ-1-pyrrolidine rings, were envisaged as anchoring sites for Au NPs. Here, we report the synthesis and characterization of graphene/gold nanocomposite (G-NH2/Au) obtained by mixing G-NH2 and Au NPs. Au NPs were produced by pulsed laser ablation in liquids (PLAL) technique that allowed the production of metal NPs in a variety of solvents with tuned size and optical properties [32,33]. No surfactant is needed to stabilize the colloids obtained by PLAL, and the NPs are extremely pure without any post-synthesis treatment [31]. To the best of our knowledge, no data have been reported in the literature about the SERS properties of G/Au platforms, where Au NPs were produced by PLAL technique.

The chemical composition and the morphology of G-NH2 and G-NH2/Au platforms were investigated by micro-Raman and X-ray photoelectron (XPS) spectroscopies, scanning transmission electron microscopy (STEM), and thermogravimetric analysis (TGA).

**Figure 1.** Schematic representation of G-MNPO and G-NH2. Chemical structure of Rhodamine 6G (Rh6G) and Dopamine (DA).

The G-NH2/Au dispersion was transferred onto a glass slide to obtain a uniform nanostructure thick film and its SERS properties were tested using Rhodamine 6G (Rh6G) and Dopamine (DA) as molecular probes (Figure 1).

#### **2. Materials and Methods**

#### *2.1. Materials*

Graphite flakes, Dopamine, Rhodamine 6G, solvents, and other reagents were purchased from Sigma Aldrich, (Milan, Italy); gold target (high purity, 99.99%) was purchased from Mateck srl (Jülich, Germany).

#### *2.2. Synthesis of G-NH2*

G-MNPO was prepared according to the synthetic method already reported [31]. Further, 240 mg of G-MNPO, (0.09 mmol of NO2) were homogenously dispersed in 30 mL H2O by sonication (30 min). NaBH4 (100 mg, 2.63 mmol) was added and the reaction mixture was stirred at 80 ◦C for 12 h. Afterwards, the reaction mixture was cooled to room temperature (r.t.), acidified to pH 3 by addition of a HCl 1M solution, and stirred for 1 h at r.t. G-NH2 was recovered by filtration under vacuum (Millipore 0.1 μm) and it was purified by washing with 1:1 water/ethanol mixture. Finally, the residue was dried at ~60 ◦C to recover 185 mg of G-NH2.

#### *2.3. Synthesis of Au NPs by PLAL*

Au water colloids were prepared according to previously reported procedure [34] using the 532 nm second harmonic emission wavelength of a Nd:YAG laser (Tempest- Laser Point srl, Milan Italy) operating at a repetition rate of 10 Hz (pulse length: 5 ns).

#### *2.4. Synthesis of G-NH2*/*Au*

First, 20 mL of Au NPs were added to a dispersion of G-NH2 (52 mg) in water (2 mL), obtained by sonication for 10 min, and the mixture was ultrasonicated (65% W) for 30 min. The reaction mixture was filtered at reduced pressure (Millipore 0.1 μm), the solid was repeatedly washed with water, and after drying at ~60 ◦C, 44 mg of G-NH2/Au were recovered.

#### *2.5. Preparation of G-NH2*/*Au SERS Platform*

The aqueous dispersion of G-NH2/Au (5 mg/mL) was deposited onto a glass slide using the conventional aerography spraying technique. The aerography spraying system is made up by a high-pressure air brush with interchangeable nozzles of different sizes. During the deposition, the nozzle is continuously moved to ensure a uniform distribution on the substrate. The spraying is carried out in a deposition chamber equipped with a heated substrate holder and an excess vapors removal system to guarantee standard and reproducible conditions. The GNH2/Au SERS platform was tested for Rhodamine 6G (Rh6G) at concentrations of 1 <sup>×</sup> <sup>10</sup><sup>−</sup>3, 2 <sup>×</sup> <sup>10</sup><sup>−</sup>4, 5 <sup>×</sup> <sup>10</sup>−<sup>5</sup> M and for Dopamine (DA) at concentrations of 1 <sup>×</sup> 10−3, 2 <sup>×</sup> 10−4, 5 <sup>×</sup> 10−5, and 5 <sup>×</sup> 10−<sup>6</sup> M. Rh6G and DA solutions were prepared using deionized water. The excitation sources were the 532 nm and 638 nm diode laser lines. The substrates were dipped in these solutions for 30 min and then taken out for free drying, after which the surface enhanced Raman (SERS) spectra were collected.

#### *2.6. Samples Characterization*

Thermal gravimetric analysis (TGA) profiles were acquired Perkin-Elmer Pyris TGA7 in the temperature range of 50–1000 ◦C. G-NH2 or G-NH2/Au (about 5 mg) were placed in a platinum pan and kept at 25 ◦C under a 60 mL min−<sup>1</sup> air flow until balance stabilization (balance sensitivity was 0.01 mg), and subsequently heated with a scan rate of 10 ◦C min−<sup>1</sup> under the same air flux. The calibration of instrument was settled according to previously reported procedure [31].

X-ray photoelectron spectroscopy was used to determine the surface elemental composition of the material and their bonding configurations. The spectra were acquired using a K-Alpha system (Thermo-Scientific, Germany) equipped with a monochromatic Al-Kα source (1486.6 eV), and operating in constant analyzer energy mode (pass energy: 200 eV), according to previously reported protocol [35]. Samples (G-NH2, G-NH2/Au, AuNPs) were deposited on a nickel grid to carry out scanning transmission electron microscopy (STEM) using a ZEISS instrument Merlin-Gemini 2 column (Merlin-Gemini, Germany), operating at primary voltage of 30 kV and at the working distance of 4 mm.

Raman spectra were acquired using the Horiba XploRA spectrometer (HORIBA Instruments, Milan, Italy) coupled with an optical microscope equipped with the 50X and 100X objectives. The excitation wavelengths used were 532 nm and 638 nm coming from solid diode lasers. The integration time was varied from 5 to 120 s, with an accumulation time of 2s, in order to optimize the signal to noise.

UV-vis optical absorption spectra of the Au and G-NH2/Au samples were recorded using quartz cells and a Perkin Elmer (Lambda 750 model) spectrometer (Perkin Elmer, Milan, Italy) working in the 300–900 nm range.

#### **3. Results**

#### *3.1. G-NH2*/*Au SERS Platform*

SERS platform based on graphene/gold nanocomposite (G-NH2/Au) was obtained through a procedure involving: (i) Synthesis of G-NH2 by reduction of G-MNPO; (ii) Preparation of G-NH2/Au and (iii) Deposition of G-NH2/Au onto a glass slide by an aerography spraying probe (Figure 2).

**Figure 2.** Preparation of G-NH2/Au surface enhanced Raman spectroscopy (SERS) platform.

G-MNPO was prepared according to the synthetic method previously reported [31], carrying out the cycloaddition reaction at the molar ratio of 1:7 flake graphite/oxazolone. The experimental conditions for G-MNPO synthesis were optimized to obtain a substrate with a large surface area and a high degree of functionalization (0.037 mmol of NO2/100 mg). G-MNPO was reduced with NaBH4 and converted in the protonated salt by treatment with hydrochloric acid. The cationic centers on G-NH2 surfaces increased the G dispersibility in water and guaranteed a better interaction with Au NPs. G-NH2/Au nanocomposite was obtained by mixing, under ultrasonication treatment, the aqueous dispersion of G-NH2 with the freshly prepared colloidal dispersion of Au NPs [34]. Finally, the aqueous dispersion of G-NH2/Au was deposited onto the glass slide.

#### *3.2. Characterization of Graphene*/*Gold Nanocomposite (G-NH2*/*Au)*

The content of Au NPs on G was estimated by TGA under air atmosphere (Figure 3). TGA profiles of G-NH2 and G-NH2/Au showed a high thermal stability without significant weight loss under 600 ◦C, indicating the absence of labile oxygen-containing functional groups. TGA profile of G-NH2 showed a decomposition between 750 ◦C and 900 ◦C, with a complete decomposition of carbon at temperatures higher than 900 ◦C; whereas the G-NH2/Au profile showed a lower decomposition temperature between 600 ◦C and 800 ◦C and the decomposition of carbon content became remarkable

at 800 ◦C. The lower thermal decomposition of G-NH2/Au compared to G-NH2 could be attributed to the presence of Au NPs that increased the interlayer spacing and porosity of G-NH2/Au. The residual mass of 7.29% indicated the loading of Au NPs on G-NH2/Au nanocomposite.

**Figure 3.** TGA profiles of G-NH2 and G-NH2/Au under air atmosphere.

Detailed information about the functionalities on G surfaces were obtained by XPS analysis. The wide scan spectra of G-NH2 and G-NH2/Au were reported in Figure 4a, with the Au 4f profile in the inset. This profile was characterized by well-separated spin-orbit components (Δ = 3.7 eV) where the Au 4f peak was centered at the binding energy of 84.0 eV, which is characteristic of the metal Au species. The Au, C, O, and N relative atomic percentages are reported in Table 1. The Au weight content percentage calculated by XPS was in good agreement with TGA data (Table 1). The N 1s high-resolution profile of G-NH2 (Figure 4b) showed the presence of two peaks centered at about 400 eV, attributed to N=C and –NH3 <sup>+</sup> species, and at 407 eV due to NO2. The lower contribution of the peak at 407 eV in G-NH2 sample, compared with G-MNPO (20.05% vs. 41.18%, see Figure 4b and Table 1), indicated a good reduction of nitro groups into amino groups. The decrease of the oxygen content after the reduction reaction (19% vs. 7.4%, see Table 1) was connected with the changes observed by N 1s profile.

**Figure 4.** (**a**): XPS wide scan of G-NH2 and G-NH2/Au samples and line shapes of Au 4f (inset). (**b**): N 1s photoelectron deconvoluted line shapes of G-MNPO, G-NH2, and G-NH2/Au.


**Table 1.** Atomic content percentage for G-MNPO, G-NH2, and G-NH2/Au samples as determined by XPS analysis and N 1s percentage determined by deconvolution of XPS N 1s band. Weight content percentage of G-NH2/Au calculated by XPS values (at the bottom).

C 1s profiles of G-MNPO, G-NH2, and G-NH2/Au were deconvolved considering six spectral components: A main contribution at 284.5 eV attributed to C=C/C–C in the aromatic ring, and four other contributions, at higher binding energies, corresponding to carbon atoms bonded to nitrogen (C–N) and oxygen (C–OH, C–O, C=O) centered at 285.2, 286.3, 288.7, and 288.9 eV, respectively. The contribution at about 291.0 eV referred to π–π\* bonds (Figure 5).

**Figure 5.** C 1s photoelectron deconvoluted line shapes: (**a**) G-MNPO, (**b**) G-NH2, (**c**) G-NH2/Au. (**d**) O 1s photoelectron deconvoluted line shapes of G-MNPO (blue), G-NH2 (violet), G-NH2/Au (brown).

Morphological information about the size and distribution of Au NPs within the G layers was obtained by electron microscopy analyses. STEM images (Figure 6) showed homogeneously distributed exfoliated G layers and various dimensional transparent sheets, in several portions of the sample, stacked onto each other, with a thickness of about 2–3 nm. Moreover, Au NPs characterized by an average size of 15 nm were mainly distributed at the edges of the G layers.

**Figure 6.** STEM images. (**A**) Au NPs produced by the green pulsed laser ablation technique in water at the laser fluence F of 1.5 J/cm2 and the irradiation time t of 20 min. The NPs are nearly spherical in shape with a mean diameter of 15 nm; (**B**) G-NH2 network with exfoliated G layers and transparent sheets, stacked onto each other, with a thickness of about 2–3 nm; (**C**,**D**) G-NH2/Au platform, with Au NPs embedded within the overlapped thin layers of graphene.

In order to investigate the SERS enhancement of G-NH2/Au, Raman spectroscopy analysis was performed (Figure 7). The Raman spectrum of G-NH2 showed the G and 2D feature bands at 1580 and 2720 cm−1, respectively (Figure 7). The very weak D-peak was indicative of the high quality of G and the 2D band splitting indicated the presence of a multilayers system. All these Raman contributions were also evident in the G-NH2/Au nanocomposite, however some relevant differences were detected (Figure 7). Firstly, the increase of the intensity of all the peaks was observed, including a D band centered at about 1350 cm−1, as a result of a certain degree of disorder induced by Au NPs insertion within G layers. The strong electric field gradient induced by the metallic NPs determined an overall change of the dipole moment during the vibration, even in the absence of a polarizability change. On the other hand, when G and Au NPs were in close proximity, some Raman forbidden peak appeared, namely the D' and the D+G contributions at about 1616 cm−<sup>1</sup> and 2925 cm<sup>−</sup>1, respectively. These evidences can be determined by: (i) The insertion of Au NPs on G-NH2 platform, mainly at the edges of G layers (functionalized area of G layers) as suggested by computational studies [31] and (ii) reduced size of layers due to the mechanical effect of ultrasonication treatment adopted for the preparation of the nanocomposite. Moreover, the decrease of IG/I2D ratio, from 1.58 to 0.98, pointed out a better exfoliation of G-NH2/Au with respect to G-NH2; the insertion of Au NPs between G-NH2 layers probably promoted their separation. Finally, the shifting of G and 2D bands suggested the anchorage of Au NPs on G surface. Raman signal was collected at several different sample locations to take into account the Au spatial homogeneity distribution within the nanoplatform. No significant

changes were observed from one point to another one, which indicated that Au NPs were almost uniformly distributed within and/or on G layers.

**Figure 7.** Raman spectra of G-MNPO, G-NH2, G-NH2/Au; the deconvolution of the D and G bands of the G-NH2/Au is reported in the inset.

#### *3.3. SERS Properties of G-NH2*/*Au Platform*

In order to test the SERS properties, the platforms (G-NH2 and G-NH2/Au) were immersed for 30 min in Rh6G aqueous solutions at different concentrations (1 <sup>×</sup> <sup>10</sup><sup>−</sup>3, 2 <sup>×</sup> <sup>10</sup><sup>−</sup>4, 5 <sup>×</sup> <sup>10</sup>−<sup>5</sup> M) and then air dried. Raman spectra were acquired using two different excitation diode laser lines (532 nm and 638 nm). UV-vis spectroscopy was exploited to determine the appropriate laser wavelength for resonant excitation of the localized surface plasmon. In fact, SERS is more effective when incident radiation falling on the nanostructured substrate is completely absorbed by metal NPs, so that excitation of the localized surface plasmon can take place. The field enhancement is greatest when the plasmon frequency is in resonance with the incident radiation. In Figure 8b, the optical absorption spectra of the freshly prepared Au NPs and of G-NH2/Au were reported. Au NPs in water showed the characteristic Au surface plasmon resonance (SPR) band at 522 nm, due to the coherent oscillations of surface electrons interacting with an external electromagnetic field; whereas a red-shift (from 522 to 548 nm) and a decrease of the SPR intensity was observed in the G-NH2/Au sample, suggesting an increase of the spatial distance between each Au NPs and the others, due to their dispersion into each G foil and/or within the G layers. Moreover, a charge transfer from Au NPs to G occurred, resulting in a decrease in electron density, which, in turn, contributed to the red-shift and the intensity decrease of the SPR band. Moreover, it is well known that the coating of gold surface with graphene modifies the propagation constant of surface plasmon polariton (SPP), thereby changing the sensitivity to refractive index change and, in turn, the optical response of the entire system [36].

SERS spectra, acquired using the 638 nm laser excitation, showed the well-defined Raman Rh6G peaks at about 615, 777, 1189, 1314, 1366, 1513, and 1651 cm−<sup>1</sup> (Figure 8a). The feature at 615 cm−<sup>1</sup> was assigned to the C–C–C in-plane bending mode, the peak at 777 cm−<sup>1</sup> to the C–H out-of-plane bending mode and the residual peaks to the aromatic stretching vibrations of C atoms. Raman features were clearly observable at 10−<sup>3</sup> M concentration and less evident, but still visible, for lower Rh6G concentration values (down to 1 <sup>×</sup> 10−<sup>5</sup> M). On the other hand, if the Rh6G aqueous solution was deposited onto a G-NH2 bare platform (i.e., without Au NPs), no Raman activity was detected even at a 10−<sup>3</sup> M concentration. The Raman spectrum of G-NH2 was characterized only by broad asymmetric and low-intensity bands, at around 1580 cm−<sup>1</sup> (referred as G band) and near 1330 cm−<sup>1</sup> (referred as D band), typical of carbon-based materials. The Rh6G SERS spectra obtained on the G-NH2/Au platform were very similar to that obtained by using a substrate made from Au nanostructured film (Figure 8b). As a final remark, we observed that, by using a 532 nm laser excitation, no Raman signals could be collected in all the tested conditions. This unusual behavior can be explained taking into account the red-shifted observed SPR optical absorption, which certainly reduced the SERS effect. Summarizing, G layers positively influenced both the EM and the CM coupling, enhancing the SERS process due to the interesting optical properties, nanostructures high surface/volume ratio, and a great affinity between G and Au NPs.

The ability of the G-NH2/Au nanocomposite to detect the biomolecules was tested using DA in a label-free configuration (Figure 9). DA is adsorbed on G surface through π–π stacking interactions [37]. The high surface area of G supporting the DA adsorption and diffusion processes was the primary condition for the efficient sensing of DA by SERS.

**Figure 8.** (**a**) SERS spectra of Rh6G (10<sup>−</sup>5, 10−<sup>4</sup> and 10−<sup>3</sup> M) onto G-NH2/Au platform; G-NH2 platform (10−<sup>4</sup> M, black line) and AuNPs film (10−<sup>4</sup> M Rh6G reported in the inset), by using the 638 nm laser excitation. (**b**) Optical absorption spectra of freshly prepared Au NPs (red line) and G-NH2/Au (blue line, 0.4 mg/mL).

**Figure 9.** SERS spectra of DA tested at different concentrations (10<sup>−</sup>6, 10−5, 10−4, and 10−<sup>3</sup> M) onto the G-NH2/Au platform along with the control test onto the G-NH2 platform (10−<sup>3</sup> M, black line), on the left. The Raman intensity signal trend onto DA concentration is reported on the right. The excitation is the 638 nm laser line.

SERS spectra of DA showed several characteristic peaks centered at about 608, 767, 1349 cm−1. It is worth noting that other weak Raman features in the 1050–1300 cm−<sup>1</sup> and 1500–1800 cm−<sup>1</sup> regions

can be detected despite the remarkable Raman background. The observed peaks were ascribed to the outside surface deformation of breathing, bending, and stretching vibrations of CH ring, bending vibration of NH, and aromatic C=C, respectively [38]. It is plausible that both EM and CM were involved in SERS signals, as already observed in reduced graphene oxide/silver nano-triangle sol substrate [39,40].

Au NPs between G layers behaved as "hot spots", which allowed the detection of DA Raman signals, not observed in the G-NH2 bare platform. Since the distribution of Au NPs within G sheets played a fundamental role in determining SERS response and that it is known from the literature that one of the problems that still remain in question is the reproducibility and the repeatability of the spectra at low concentrations, we acquired SERS spectra in different points of the SERS substrate and at different times. We observed, point to point, very minimal variations in the intensity of some DA characteristic peaks without the degradation of nanocomposites.

#### **4. Discussion**

The results of this work has demonstrated the great potentiality for SERS applications of functionalized G obtained by covalent modification [41]. The cycloaddition protocol (i.e., 1,3-dipolar cycloaddition between mesoionic compounds and graphite) [31] furnished a G network decorated with Δ-1-pyrrolidine rings, mainly in the edge defected sites. This approach incorporated several interesting advantages: (i) Avoided the damage of sp<sup>2</sup> G network; (ii) provided functionalized G with high degree of functionalization (i.e., 4.6%) and the NO2 group on pyrrolidine rings was reduced in good yield (XPS data indicated the reduction of almost half of the nitro groups in NH2 groups); (iii) the amine groups assisted the anchorage of Au NPs produced by PLAL on G surface; (iv) the dispersibility in water of the G-NH2 nanocomposite was enough for its deposition onto glass slide by aerography spraying technique. Au NPs included on G resulted in a hybrid nanocomposite (G-NH2/Au) that combined the stronger plasmonic-based EM of Au with the superior stability, adsorption, and quenching of G. This nanocomposite revealed strong interactions between the two entities. From its spectral features, the origin of these interactions could be attributed primarily to the strong electric field gradient induced by Au NPs that determined an overall change of the dipole moment during the vibration, even in the absence of a polarizability change. The SERS activity of the assembled Au NPs with the graphene platform is justified in terms of "hot spots". Au NPs within the G-NH2 structure are "confined" to a certain region sensitive to the Raman scattering [42,43]; localization of light as surface plasmons in noble metal nanostructures enables their potential role in antennas, single molecule detection, and surface-enhanced Raman. Light localization by graphene structure induces the change of the electron structure of molecules due to their direct interaction with the surface in the first adsorbed layers. However, there is no "chemical enhancement" and the one in SERS is associated with very strong change of the electric field, when one moves away from the surface [44]. In this work, G-NH2/Au nanocomposite was used to identify the dye Rh6G and the neurotransmitter DA. DA is a catecholamine that plays a significant role in the functioning of central nervous, vascular, hormonal systems and its abnormal variation concentration in vivo has been linked to serious neurological diseases. The direct SERS quantification of DA in biological fluids remains a great challenge due to the low concentration (<10−<sup>10</sup> M) and the high complexity of biological matrix. Raman features of Rh6G and DA were still detectable for concentration values down to 1 <sup>×</sup> 10−<sup>5</sup> M and 1 <sup>×</sup> 10−<sup>6</sup> M, respectively, although the sensibility of our system was found lower than the graphene-based SERS substrates reported in the literature for both analytes [24,45]. From our studies, it emerged that an improvement of G-NH2/Au sensibility is imperative before proposing it as substrate for the detection of DA in biological matrix. We hypothesized that the detection limit of G-NH2/Au nanocomposite could be improved by tuning the DA absorption properties on G and by setting the features, size, and shape of plasmonic noble metal NPs.

#### **5. Conclusions**

In summary, we investigated the SERS properties of a new graphene/gold nanocomposite (G-NH2/Au) obtained by combining Au NPs produced by PLAL technique with G covalently functionalized (G-NH2). After the chemical modification of G, the SERS platform was obtained by loading Au NPs on the G-NH2 surface and deposition of G-NH2/Au nanocomposite onto the glass slide by an aerography spraying technique. The chemical composition and the morphology of nanocomposites were investigated by micro-Raman XPS, STEM, and TGA analyses. STEM analyses showed transparent graphene sheets, with various dimensions, stacked onto each other, with a thickness of about 2–3 nm. Au NPs were detected as uniform spherical structures, with an average size of 15 nm, mainly distributed at the edges of the G layers. A good Au NPs loading was estimated by TGA and XPS analysis (i.e., 7.29% and 7.5%, respectively). This strategy allowed us to study SERS properties of G loaded with pure Au NPs without the influence of capping agents, surfactants, or salt produced in the chemical reduction of gold ions. SERS platform was tested to identify the dye Rh6G and the neurotransmitter DA; Raman features of Rh6G and DA are still detectable for concentration values down to 1 <sup>×</sup> <sup>10</sup>−<sup>5</sup> M and 1 <sup>×</sup> <sup>10</sup>−<sup>6</sup> M, respectively.

In conclusion, our platform possessed good stability and capability to reproduce the Raman signals without degradation although with low sensibility. Considering the feasibility of our method, further study will be devoted to improving the DA detection limits to refine the absorption properties of G-NH2 and the plasmonic effect of loaded noble metal NPs.

**Author Contributions:** Conceptualization, A.P. and E.F.; methodology, G.N., E.F., P.G.M., A.S. and A.P.; validation, G.N., E.F., P.G.M., A.S. and A.P.; formal analysis, G.N., E.F., P.G.M., A.S. and A.P.; investigation, G.N., E.F. and P.G.M.; data curation, G.N., E.F., P.G.M., A.S. and A.P.; writing—original draft preparation, A.P. and E.F.; writing—review and editing, G.N., E.F, P.G.M., A.S. and A.P.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
