Mn²⁺-mediated energy transfer process as a versatile origin of photoluminescence in graphene oxide

Abstract

Many factors affect the photoluminescence (PL) of graphene oxide (GO). The influence of Mn²⁺ on the PL in GO was systematically investigated when the universal heteroion was introduced in the common preparation based on the Hummers method, and for the first time we propose that a Mn²⁺ mediated energy transfer process may be a versatile origin in GO.

---

Luminescent graphene oxide (GO) has attracted great interest due to its tunable optical properties,¹ and excellent compatibility in optoelectronics² and biology.³ Different synthetic strategies have a different influence on the photoluminescence (PL) of GO,^4,7 and the preparation of chemically derived GO is facile and endows PL emissions from ultraviolet to near infrared.⁵ The emission of GO sheets or dots emerging in the range from red to NIR⁶ is effective for biological environments due to its weak damage and strong penetration to tissues and organs; moreover, blue emission in GO is also widely reported.⁷

The PL mechanism of GO, either red or blue emission, is not clearly understood yet. Many factors, such as the pH value,⁸ solvent,⁹ and reduction,¹⁰ affect the PL energy or intensity markedly. The unclear structure of GO,¹¹ and the different preparation methods^5b also make the PL mechanism more complicated. There have been several explanations on the PL of GO. Chhowalla proposed that the red emission in GO comes from a disorder-induced localized state related transition¹ and considered the blue emission from electron–hole recombination among sp² clusters enclosed by sp³ defects,^7a whereas Wu et al.^7b attributed the blue emission to intrinsic zigzag edges with a carbene-like triplet ground state. Ajayan reported a mirror-symmetry in the PL and PLE spectra of GO sheets, and proposed that the emission from GO is a quasi-molecule fluorescence.¹² While Chu proposed that chemical reduction plays a critical role in the blue PL of GO.¹⁰ A deep understanding of the PL behaviour of GO is necessary and urgent to obtain good control over the PL of GO and eventually enable to be practically applied in optoelectronic devices.

As mentioned above, the synthetic strategies⁴ and dielectric environment around GO suspensions have great impact on the PL of GO. However, the influence of certain ions introduced in the chemical preparation process, on the PL of GO is rarely studied,¹³ and it is necessary to be investigated in order to elucidate the PL mechanism of GO. Here we carried out a systematic investigation on the PL of GO and explored the influence of Mn²⁺ on the PL behaviour, which widely exists in chemically derived GO. It is interesting to find a strong excitation-independent blue emission and stationary PLE peak around 330 nm and 450 nm in a highly acidic GO solution, which is induced by a Mn²⁺ mediated energy transfer process and identified by the long decay lifetime. To the best of our knowledge, this is the first report on a Mn²⁺ mediated energy transfer process as a versatile origin of PL in GO, and it is believed that this work will provide a new concept on the PL of GO and will be helpful in understanding the PL mechanism of CDGO.

GO is synthesized via a modified Hummers method.¹⁴ In a typical procedure, 5.0 g of graphite powder (325 mesh, 99.8%, Alfa Aesar) was preoxidized in 7.5 mL of 98% H₂SO₄ with 2.5 g K₂S₂O₈ and 2.5 g P₂O₅ over 6 hours at 80 °C. The product was washed by deionized water to make it neutral. The filter cake is dried at 60 °C overnight. 500 mg preoxidized graphite is mixed with an equal mass of NaNO₃ (Sigma-Aldrich) in 20 mL concentrated sulfuric acid in a three-neck flask, which was placed in an ice bath and kept under mild stirring. 3.0 g KMnO₄ (Alfa-Aesar) is added very slowly to keep the temperature below 10 °C. After fully mixing, 20 mL deionized water is added dropwise. The reaction is carried out at 35 °C for 6 hours, followed by which 100 mL deionized water is poured into the flask. The mixture is heated at 95 °C for 15 min, and then 10 mL H₂O₂ (SinoChem) aqueous solution is injected to quench the excess KMnO₄. The brown solution becomes bright yellow, indicating the reduction of Mn⁷⁺ ion. The solution is cooled down to room temperature naturally. The above solution is centrifuged at 8500 rpm for 15 min. The supernatant is collected and referred to as GO-1. The precipitates are washed by deionized water and centrifuged at 8500 rpm for 15 min. After 5 cycles of washing, the final solid precipitates are collected and dried at 60 °C for 2–3 days. The dried products are dispersed in deionized water to form about a 1 mg mL⁻¹ solution, which is referred to as GO-5. Further purification does not influence the PL of GO (see Fig. 4d); thus, GO-5 is chosen as the counterpart of GO-1. The Raman spectra were recorded by an HR800 Jobin Yvon Raman spectrometer with a 514 nm excitation laser under ambient conditions. The PL and PLE spectra are recorded by a Cary-5000 fluorescence spectrometer. The time-resolved PL test is performed on a FLS920 fluorescence spectrometer (Edinburgh Instruments) by means of time-correlated single-photon counting (TCSPC), adopting a pulsed diode laser operating at a photon energy of 3.06 eV (405 nm) as the excitation source. The X-ray photoelectron spectroscopy (XPS) measurements were performed on a Thermo ESCALAB 250 with a monochromatized Al Kα source.

We find that the luminescence behaviour of the GO solution after washing once and five times is obviously different, while their Raman spectra (Fig. 1c) are nearly identical. The typical PL and PLE spectra of GO-1 and GO-5 are shown in Fig. 1a and b, respectively. It can be seen that the PL peak is red-shifted from ∼460 nm in GO-1 to ∼600 nm in GO-5 and the emission intensity is lowered after five times of washing. The PLE spectrum of GO-1 shows only one peak at 330 nm, while in GO-5 it exhibits two peaks around 450 nm and 540 nm. The Raman spectra of GO-1 and GO-5 show the typical D band (∼1350 cm⁻¹) and G band (∼1580 cm⁻¹) of graphene oxide, with a very close D/G ratio of ∼1.08. This indicates that there is no clear difference in the defect concentration between GO-1 and GO-5.

Fig. 1 PLE (red) and PL (black) spectra of (a) GO-1 and (b) GO-5. (c) Raman spectra of GO-1 (black) and GO-5 (red).

We performed excitation-dependent PL measurements on GO-1 and GO-5 (as shown in Fig. 2) to further investigate their difference. We recorded a series of PL spectra on GO-1 with the excitation wavelength ranging from 280 nm to 570 nm (Fig. 2a) and found that like luminescent polyaromatic molecules, the excitation dependent PL is also observed in our luminescent GO solution. The PL emission is continuously red shifted with the increasing excitation wavelength. PLE spectra detected from 420 nm to 610 nm are shown in Fig. 2b. All of the PLE spectra are dominated by a major peak centered around 330 nm, and a tail emerged when measured after 480 nm, whereas no obvious shift of the PLE peak wavelength is observed. PLE intensity reaches the maximum at the measured emission wavelength of 460 nm. One feature of the PLE spectra of GO-1 is the long tail extending toward low energy. In Fig. 2b, a very weak peak around 400 nm is observed in nearly in all of the PLE spectra and a moderately strong peak around 450 nm is observed in the PLE spectra with a measured emission wavelength longer than 480 nm. The weak peak around 400 nm is probably the low-frequency vibrational Raman signal from water.¹⁵ The peak around 450 nm is also not shifted obviously. The relationship between the excitation and emission wavelength is also extracted as shown in Fig. 2c. We note that when the excitation energy is lower than 3.2 eV, a nearly linear dependence of the excitation and emission wavelength is shown. While the excitation energy is higher than 3.2 eV (∼390 nm), the emission peak wavelength almost remains constant, exhibiting an excitation independent emission, which may indicate a different emission mechanism. This will be discussed later.

Fig. 2 (a) Normalized PL spectra of GO-1 at different excitation wavelengths (from 280 nm to 570 nm); (b) PLE spectra of GO-1 (emission wavelength monitored from 420 nm to 610 nm); (c) dependence of emission energy on excitation energy. The data are extracted from (a). Solid red line exhibits a linear decrease of emissive photon energy with excitation photon energy; (d) normalized PL spectra of GO-5 at different excitation wavelengths (from 450 nm to 620 nm); (e) PLE spectra of GO-5 (emission wavelength monitored from 420 nm to 620 nm) The sharp peaks at ultraviolet light range is induced by the double frequency modes of corresponding detected wavelength; (f) dependence of emission energy on excitation energy. The data are extracted from (d). Solid red line exhibits a linear decrease of emissive photon energy with excitation photon energy.

Analogously, the PL and PLE spectra (Fig. 2d–f) of GO-5 are examined. In Fig. 2d, with the excitation wavelength ranging from 450 nm to 620 nm, a typical excitation dependent emission from yellow to red is observed. Here, no excitation independent emission like the blue emission peak at low excitation wavelength in GO-1 is observed. The PLE spectra (Fig. 2e) exhibit rather different fine structures from that of GO-1. Here, a very broad PLE band almost covering the entire visible light range is observed. Two peaks around 450 nm and 550 nm are overlaid on the envelope. The PLE peak around 450 nm, which appeared in GO-1, is also observed here and likewise, and it does not shift when the measured emission wavelength is varied, which indicates that it may have the same origin as GO-1. When the peak around 550 nm is investigated carefully (Fig. 2f), a well-fitted linear relationship of the excitation and emission wavelength is observed.

The effect of pH value on the luminescence of a GO solution is also investigated to further investigate the difference between GO-1 and GO-5. A highly concentrated NaOH aqueous solution is employed to adjust the pH value of GO-1 and GO-5 due to the intrinsic acidity of the GO solution. There are two reasons for the employment of highly concentrated NaOH aqueous solution: (1) the addition of a solid alkaline powder directly would result in a rather high local alkaline concentration and a rapid increase in temperature, which will facilitate the precipitation of colloid GO sheets. (2) If the concentration is low, the concentration effect on the absorbance and photoluminescence intensity is pronounced. We added the highly concentrated NaOH aqueous solvent under stirring to GO to obtain the GO solutions with different pH values and to weaken the temperature and concentration effects. The pH value dependent PL spectra of GO-1 and GO-5 are shown in Fig. 3. We note that when the highly concentrated NaOH aqueous solution is added to GO-1, a large amount of bubbles is produced and some black precipitates emerge, but disappeared immediately, while the precipitates never dissolved when the pH was over 6. However, if some acidic solution, such as HCl or HNO₃, was added, the precipitates become soluble. We speculate that the black precipitate is a kind of metal salt instead of GO due to two aspects: (1) normally carbon-based nanomaterials do not have reversible solubility in acid-alkaline solutions. (2) While the amount of precipitates did not increase with further addition of NaOH solution, the supernatant still showed obvious absorbance bands with GO-characteristics. However, we cannot rule out the probability that the coprecipitation of an insoluble metal salt and GO. XPS is performed to explore the composition of the precipitates. The sample is prepared via heating at 80 °C until no liquid is left. A Pale black powder is obtained ground. From the XPS spectra of this solid sample (Fig. 4), more than 10 at% Mn is detected. To verify that the content of Mn can be tuned by washing, GO-3 and GO-7 (denoted as GO washed 3 and 7 times by DI-water) is also inspected together with GO-1 and GO-5. Mn is detectable in GO-3 (with content of ∼3 at%) while it is nearly undetectable in GO-5 and GO-7, suggesting that Mn can be tuned to some extent by washing. According to the 2p_3/2 peak position¹⁶ (∼642.0 eV) and the oxidation properties of manganese heptoxide and the solubility of Mn ions in acid-alkali solutions,¹⁷ we suggest that MnO₂ is the most likely candidate of the solid precipitation and the soluble counterparts at lower pH values are certain Mn²⁺ salts. The pH-dependent PL spectra of GO-1, as shown in Fig. 3a, are dominated by a blue emission band centered at 460 nm. The emission wavelength is not shifted obviously with the increasing pH value. The PL emission intensity increases, reaches the peak at the pH of 4.06, and then decreases at higher pH values. Especially, when the pH is close to 6, the PL intensity sharply drops and becomes nearly unresolved. The black precipitate, which emerged in GO-1, was not observed among the whole pH range in GO-5, which agreed well with our previous speculation that of relatively high concentration of Mn²⁺ is involved in GO-1, while with five times washing, GO-5 contains barely any Mn²⁺ and is nearly undetectable from XPS. Nevertheless, we can also not directly perorate that Mn²⁺ has no influence on the PL of GO-5 as PL is extremely sensitive to defects/impurities, and even trace amounts can be significant. We find the decreasing trend of the PL intensity with pH elevation fluctuates, which may be attributed it to the double pK_a or dynamic structure in colloidal GO solutions.¹⁸ The PL spectra (Fig. 3b) of GO-5 show a yellow emission band around 570 nm in acidic solution. The PL intensity undergoes a sharp drop when the pH value is over 5. When pH value is over 8, the yellow-orange emission is severely weakened and is unobservable; moreover, a band at ∼460 nm is gradually resolved. This indicates that the PL of GO-5 is actually the sum of the strong yellow-orange band and weak blue band. While for GO-1, it is the reverse, namely, the PL is the sum of the strong blue band and weak yellow-orange band.

Fig. 3 pH-dependent PL spectra of (a) GO-1 and (b) GO-5, respectively.

Fig. 4 XPS spectra of GO washed for different times. (a) C 1s and (b) Mn 2p core level spectra of GO-1. (c) C 1s core level spectrum of GO-5. (d) Full spectra of GO-3, GO-5 and GO-7 (denoted as GO washed 3, 5 and 7 times). Yellow stripe in (d) is the location of the Mn 2p line, indicating that nearly no Mn is detected in GO-5 and GO-7, while it is detectable in GO-3.

Time-resolved PL decay (TRPL, Fig. 5) is carried out to probe the carrier recombination process in order to further explore the underlying difference between GO-1 and GO-5. Fig. 5a shows that the carriers in GO-1 have a lifetime of 2.5 ns and 1.4 ns in GO-5. It should be noted that the PL decay in GO is rather fast and the decay lifetime is of the order of picoseconds;^1,6,10,13 the relatively long lifetime here indicates that some extra charge carrier transport or recombination process is involved. Taking all of the above observed results into comprehensive consideration, we proposed that Mn²⁺ acts as an intermediate state in the carrier decay process, which can well explain the long decay lifetime obtained here. We also monitored the PL decay at different emission wavelengths in GO-1 (Fig. 5b) and GO-5 (Fig. 5c). It was found that a longer emission wavelength exhibits relatively longer carrier lifetime (Fig. 5d). In particular, at 600 nm, the carrier lifetime of GO-1 is almost equal to that of GO-5. This again indicates that it is likely that the emission at 600 nm in GO-1 and GO-5 has an identical origin.

Fig. 5 TRPL of GO-1 and GO-5. (a) Comparison of decay process of GO-1 (red line) and GO-5 (blue line) at 550 nm. Black line is the instrumental response function (IRF) of the TRPL test system; TRPL monitoring at different PL lines of GO-1 (b) and GO-5 (c). IRF curve is identical to (a); (d) dependence of decay lifetime for blue (black dot line) of GO-1 and yellow-to-orange (blue dot line) emission of GO-5.

It is well established that the Mn²⁺ mediated emission in semiconductor nanocrystals¹⁹ or inorganic phosphors²⁰ is often located around 590 nm, which corresponds to a regular d–d transition (commonly called ⁴T₁ to ⁶A₁ transition) of Mn²⁺. However, in our samples, with a high concentration Mn²⁺ as in GO-1, the blue emission is dominant while the yellow-orange emission around 590 nm is not observed. Moreover, the blue emission vanishes and the yellow-orange emission appears when there is barely any Mn²⁺. The transition energy of Mn²⁺ may be very different here in GO compared to those in nanocrystals or phosphors as the position of the emission band of Mn²⁺ is strongly dependent on the field strength around Mn²⁺. Allowing for the excitation-dependent emission and the very short decay lifetime here, we speculate that the blue emission around 460 nm in GO-1 and red emission around 590 nm in GO-5 is not the result of the Mn²⁺ ligand field transition, while a Mn²⁺ mediated energy transfer is probably responsible for the PL emission.

Although the proposed explanation that disorder-induced localized defect states¹ or (un)protonated effects at carbene-like zigzag sites^7b can explain the variable PL between 470 nm and 600 nm with the reduction of GO and the reversibility of blue emission with pH tuning, they are not applicable here and cannot well explain some of the unique PL behaviour observed in this work. Chu's claim¹³ that Mn²⁺ plays an important role in the improved low-energy PL from the Mn²⁺-bonded rGO also cannot explain the excitation-independent blue emission and the stationary 330 nm and 450 nm PLE peaks. By the combination of the Mn²⁺ mediated energy transfer process and the pH-dependent reversible solubility of Mn²⁺-related salts, the scenario observed can be well explained as illustrated in Fig. 6. As proposed in our previous study,²¹ the PL of GO is derived from the recombination of e–h pairs localized in the isolated sp² clusters within the sp³ oxygen-functional matrix and the related defect states, and the different size and distribution of sp² clusters consistent with variable bandgap^7a and defect states, corresponding to the excitation-dependent emission in GO. The stationary PLE peaks are observed around 330 nm and 450 nm, corresponding to the ⁶A₁ to ⁴T₁ and ⁶A₁ to ⁴T₂ transition of Mn²⁺, respectively. When the excitation wavelength (λ_ex) varies from long to short wavelengths, the photons are first absorbed by the defect states and the orange light is emitted, which corresponds to the long wavelength emission in GO-1 and GO-5; thus, the emission lifetime at 600 nm is identical for GO-1 and GO-5. As λ_ex is shorter than 450 nm, in addition to the absorption of GO, the transition from ⁶A₁ to ⁴T₁ in Mn²⁺ initiates and the absorbed energy and is transferred to the defect states in GO. When λ_ex is shorter than 390 nm (∼3.2 eV), the Mn²⁺ mediated state (⁶A₁ to ⁴T₂ transition) is gradually excited dominantly and this energy then is transferred quickly to the sp² clusters in GO and then relax to result in the blue PL emission. This explains the excitation-independent emission in GO-1 when λ_ex is shorter than 390 nm (Fig. 2c). The large energy can hardly transfer to the defect states due to their large energy deviation; thus, the 330 nm PLE peak cannot be observed in GO-5. The mechanism of the Mn²⁺ ion mediated emission can be understood by the classic energy transfer theory²² about donor–acceptor pairs. The extent of energy transfer is determined by the distance between GO and Mn²⁺, and also the extent of spectral overlap. The rate of energy transfer k_T(r) is given by k_T(r) = (R₀/r)⁶/τ_D, where r is the distance between GO and Mn²⁺, and τ_D is the lifetime of GO in the absence of energy transfer. In the case of Mn²⁺ adsorption, we believe the distance is small enough for the energy transfer, although the exact distance in such a system is still unknown. As a corroboration of the proposed Mn²⁺-mediated energy transfer, the lifetimes of the PL mediated by Mn²⁺-related energy transfer process in GO-1 and GO-5 are relatively longer than the typical picoseconds decay in GO. The reversible solubility of Mn²⁺-related salts in acid-alkali conditions clarifies the black precipitates observed in GO-1 when the pH is high. The coprecipitation of Mn²⁺-related salts and GO is also possible as the PL intensity is sharply dropped when the pH is over 6.

Fig. 6 Schematic multi-level diagram explaining the whole PL scenario in GO associated with the Mn²⁺-mediated energy transfer process. The blue and yellow-orange emissions are attributed to sp² clusters and defect/localized states, respectively. The ⁶A₁ to ⁴T₁ and ⁶A₁ to ⁴T₂ transition of Mn²⁺ correspond to the ∼450 nm and ∼330 nm PLE peak, respectively. The energy transfer from Mn²⁺ to the defect states or sp² clusters results in the different PL.

Conclusions

In summary, an excitation-independent blue emission and two stationary PLE peaks around 330 nm and 450 nm are found to be related to the Mn²⁺ mediated energy transfer process, and this Mn²⁺ mediated energy transfer is firmly involved in purified GO, which contains barely any Mn²⁺. TRPL provides credible evidence that Mn²⁺ acts as an intermediate state in the carrier recombination process. Our work indicates for the first time that Mn²⁺ mediated energy transfer is likely a versatile origin of the PL in GO, which may widely exist and play a part in the PL of GO. Though the physics of the Mn²⁺ related state should be further investigated, we believe this work on the influence of the “intrinsic” heteroatom derived readily from the universal chemical preparation on the PL of GO will remind us to care more about the chemical environment around GO when elucidating the PL mechanism, and the results on the Mn²⁺ mediated energy transfer will be helpful to understand the PL mechanism of GO.

Acknowledgements

This work was supported by the NSFC (no. 51372223), the Program for Innovative Research Team in University of Ministry of Education of China (no. IRT13037), the Science and Technology Department of Zhejiang Province (no. 2010R50020), and the APNSF (no.10040606Q11).

Notes and references

1. C. T. Chien, S. S. Li, W. J. Lai, Y. C. Yeh, H. A. Chen, I. S. Chen, L. C. Chen, K. H. Chen, T. Nemoto, S. Isoda, M. Chen, T. Fujita, G. Eda, H. Yamaguchi, M. Chhowalla and C. W. Chen, Angew. Chem., Int. Ed., 2012, 51, 6662.
2. (a) S. Gilje, S. Han, M. Wang, K. L. Wang and R. B. Kaner, Nano Lett., 2007, 7, 3394; (b) P. Kumar, L. S. Panchakarla, S. V. Bhat, U. Maitra, K. S. Subrahmanyam and C. N. R. Rao, Nanotechnology, 2010, 21, 385701; (c) P. Kumar, K. S. Subrahmanyam and C. N. R. Rao, Int. J. Nanosci., 2011, 10, 559; (d) P. Kumar, K. S. Subrahmanyam and C. N. R. Rao, Mater. Express, 2011, 1, 252.
3. J. L. Li, H. C. Bao, X. L. Hou, L. Sun, X. G. Wang and M. Gu, Angew. Chem., Int. Ed., 2012, 51, 1830.
4. (a) P. Kumar, RSC Adv., 2013, 3, 11987; (b) P. Kumar, B. Das, B. Chitara, K. S. Subrahmanyam, K. Gopalakrishnan, S. B. Krupanidhi and C. N. R. Rao, Macromol. Chem. Phys., 2012, 213, 1146.
5. (a) L. Cao, M. J. Meziani, S. Sahu and Y. P. Sun, Acc. Chem. Res., 2013, 46, 171; (b) K. P. Loh, Q. Bao, G. Eda and M. Chhowalla, Nat. Chem., 2010, 2, 1015.
6. (a) X. Sun, Z. Liu, K. Welsher, J. T. Robinson, A. Goodwin, S. Zaric and H. Dai, Nano Res., 2008, 1, 203; (b) T. Gokus, R. R. Nair, A. Bonetti, M. Böhmler, A. Lombardo, K. S. Novoselov, A. K. Geim, A. C. Ferrari and A. Hartschuh, ACS Nano, 2009, 3, 3963.
7. (a) G. Eda, Y. Y. Lin, C. Mattevi, H. Yamaguchi, H. A. Chen, I. S. Chen, C. W. Chen and M. Chhowalla, Adv. Mater., 2010, 22, 505; (b) D. Pan, J. Zhang, Z. Li and M. Wu, Adv. Mater., 2010, 22, 734.
8. J. L. Chen and X. P. Yan, Chem. Commun., 2011, 47, 3135.
9. J. I. Paredes, S. Villar-Rodil, A. MartÍnez-Alonso and J. M. D. TascÓn, Langmuir, 2008, 24, 10560.
10. Z. Gan, S. Xiong, X. Wu, T. Xu, X. Zhu, X. Gan, J. Guo, J. Shen, L. Sun and P. K. Chu, Adv. Mater. Res., 2013, 1, 926.
11. (a) A. Lerf, H. He, M. Forster and J. Klinowski, J. Phys. Chem. B, 1998, 102, 4477; (b) W. Gao, L. B. Alemany, L. Ci and P. M. Ajayan, Nat. Chem., 2009, 1, 403.
12. C. Galande, A. D. Mohite, A. V. Naumov, W. Gao, L. Ci, A. Ajayan, H. Guo, A. Srivastava, R. B. Weisman and P. M. Ajayan, Sci. Rep., 2011, 1, 85.
13. Z. X. Gan, S. J. Xiong, X. L. Wu, C. Y. He, J. C. Shen and P. K. Chu, Nano Lett., 2011, 11, 3951.
14. W. S. Hummers and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339.
15. D. M. Carey and G. M. Korenowski, J. Chem. Phys., 1998, 108, 2669.
16. T. Yanagida, H. Tanaka, T. Kawai, E. Ikenaga, M. Kobata, J. J. Kim and K. Kobayashi, Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 73, 132503.
17. K. R. Koch and P. F. Krause, J. Chem. Educ., 1982, 59, 973.
18. (a) A. M. Dimiev, L. B. Alemany and J. M. Tour, ACS Nano, 2013, 7, 576; (b) B. Konkena and S. Vasudevan, J. Phys. Chem. Lett., 2012, 3, 867; (c) S. Kochmann, T. Hirsch and O. S. Wolfbeis, J. Fluoresc., 2012, 22, 849.
19. (a) D. J. Norris, N. Yao, F. T. Charnock and T. A. Kennedy, Nano Lett., 2001, 1, 3; (b) N. Pradhan and X. Peng, J. Am. Chem. Soc., 2007, 129, 3339.
20. (a) C. J. Duan, A. C. A. Delsing and H. T. Hintzen, Chem. Mater., 2009, 21, 1010; (b) N. Da, M. Peng, S. Krolikowski and L. Wondraczek, Opt. Express, 2010, 18, 2549.
21. H. He, H. Li, T. Zhang, L. Sun and Z. Ye, RSC Adv., 2014, 4, 18141.
22. J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer, 3rd edn, 2006.

---