Nanoscale View Article Online
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
PAPER
View Journal | View Issue
Cite this: Nanoscale, 2014, 6, 3384
Amino-functionalized graphene quantum dots: origin of tunable heterogeneous photoluminescence† G. Sandeep Kumar,‡b Rajarshi Roy,‡a Dipayan Sen,a Uttam Kumar Ghorai,b Ranjit Thapa,c Nilesh Mazumder,a Subhajit Sahab and Kalyan K. Chattopadhyay*ab Graphene quantum dots are known to exhibit tunable photoluminescence (PL) through manipulation of edge functionality under various synthesis conditions. Here, we report observation of excitation dependent anomalous m–n type fingerprint PL transition in synthesized amino functionalized graphene quantum dots (5–7 nm). The effect of band-to-band p*–p and interstate to band n–p induced transitions led to effective multicolor emission under changeable excitation wavelength in the functionalized system. A reasonable assertion that equi-coupling of p*–p and n–p transitions activated the heterogeneous dual mode cyan emission was made upon observation of the PL spectra. Furthermore, investigation of incremented dimensional scaling through facile synthesis of amino functionalized quantum graphene flakes (20–30 nm) revealed it had negligible effect on the modulated PL pattern. Moreover, an effort was made to trace the origin of excitation dependent tunable heterogeneous photoluminescence through the framework of energy band diagram hypothesis and first principles analysis. Ab initio results suggested formation of an interband state as a manifestation of p
Received 9th October 2013 Accepted 2nd December 2013
orbital hybridization between C–N atoms at the edge sites. Therefore comprehensive theoretical and experimental analysis revealed that newly created energy levels can exist as an interband within the
DOI: 10.1039/c3nr05376h
energy gap in functionalized graphene quantum structures yielding excitation dependent tunable PL for
www.rsc.org/nanoscale
optoelectronic applications.
Introduction Graphene and its chemical derivatives have been the center of attention in nanotechnology during the latter half of the last decade exhibiting superlative electronic and mechanical properties.1–3 On the other hand, the development of uorescent graphene oxide based materials has been in demand for usage in diverse opto-electronic applications.4,5 But slowly and steadily coming into the fray PL applications based on chemically synthesized graphene quantum dots (GQDs) have gathered prominence in recent years.6–10 Due to quantization effect, one can expect interesting PL behavior in graphene based quantum dots.11–13 In an expanding technological scenario, several techniques like surface passivation methods and edge termination with functional groups inside these structures are signicantly
a
Thin Films and Nanoscience Laboratory, Dept. of Physics, Jadavpur University, Kolkata-700032, India. E-mail: [email protected]
b
School of Materials Science and Nanotechnology, Jadavpur University, Kolkata700032, India
c
SRM Research Institute, SRM University, Kattankulathur, 603203, Tamil Nadu, India
† Electronic supplementary 10.1039/c3nr05376h
information
(ESI)
available.
‡ These authors contributed equally to the present work.
3384 | Nanoscale, 2014, 6, 3384–3391
See
DOI:
contributing towards broad tunable PL by varying the excitation wavelength only.14–18 Several mechanisms have been proposed to explain this type of PL behavior, such as incomplete solvation of the excited state of the species,19,20 size distribution for nanocrystallites,21 level splitting by doped ions.22 However, the origin of tunable PL in edge functionalized graphene quantum dots is yet to be well understood. As mentioned earlier, previously reported results showed that due to excitation redshi, continual reduction of PL intensity and tunable emission was observed when the amount of functionalization was varied at different temperature conditions in graphene quantum dots.13,16 But in most of these cases, emissions due to band to band transitions were considered, whereas the roles of interstate or defect induced emissions have not been extensively probed. For functionalized systems there will be a presence of certain interbands in the structures depending on the synthesis conditions and the degree of functionalization while their positions in the optical band gap will determine the nature of tunable PL. Considering a generic case, heterogeneous emissions are expected for heterogeneous systems; however previous reports have shown that due to variation in the degree of amino functionalization the direct band gap has been manipulated in functionalized graphene quantum dots. It was inferred that more the degree of amino edge functionality shrinkage of the
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
Paper
band gap took place with a redshi of PL9,16 and in most of the cases heterogeneous emissions were not at all considered. Our current motivation for this work emphasizes probing how the presence of direct and intersystem bands will affect the nature of excitation dependent PL in functionalized graphene based quantum structures. Hence with this view in mind, in this article we report excitation wavelength dependent tunable heterogeneous PL in amino functionalized graphene quantum dots (5–7 nm) and detailed explanation is given in connection with energy band diagram analogy and ab initio approach. To understand whether this PL conversion had any size dependence, we synthesized amino edge terminated quantum graphene akes (QGFs) (20–30 nm). From rst principles calculations, we have investigated the possible origin of newly created energy levels and the governing nature of hybridization as an outcome of amino edge functionalization using a prototype structural model. Furthermore, calculation of PL quantum yield using an analytical approach, energy transfer efficiency and time decay were carried out to ascertain the impact of functionalized quantum dots as diverse carbon based luminescent material.
Experimental Sample preparation For the synthesis of GO, a modied Hummer's method approach was employed from graphite powder. As synthesized, the resultant product was then vacuum dried. The obtained GO with a loading of (1 mg ml1) was further dissolved in DI water and exfoliated for 30 min in a high power sonicator. The resultant solution was processed subsequently to lter out larger bulk particles using a high speed centrifuge for 15 min at 25 000 rpm and the supernatant was collected to obtain a yellow colored stable hydrophilic suspension of exfoliated uniform graphene oxide layers.23,24 For GQDs synthesis, 10 ml of GO solution (1 mg ml1 loading), 10 ml of DI water and 6 ml of ammonia solution were added together. This mixture was stirred for 30 min, following which ultrasonication was carried out in a 250 W high power sonicator for another 30 min. Next, the resultant solution was transferred into a Teon lined autoclave and heated at 150 C for 5 h by hydrothermal treatment. Furthermore, this solution was naturally brought down to room temperature and subsequently the GQDs were collected by ltration through a 0.22 mm Teon membrane. The resultant supernatant was heated at 100 C for another 1 h to vapor out excess ammonia. The nal resultant solution was centrifuged at 25 000 rpm before further characterizations were carried out.15,16 For QGFs synthesis, 10 ml of GO solution, 10 ml of DI water and 12 ml of ammonia solution was added together in a beaker. This mixture was stirred for 30 min, followed by ultrasonication in a 250 W high power sonicator for another 30 min. Next, this resultant solution was heated at 100 C for 4 h in open atmosphere, following which the solution was brought down to room temperature and ltered through a 0.22 mm Teon membrane and the supernatant was collected. The resultant supernatant was heated at 100 C for another 1 h to vapor out excess
This journal is © The Royal Society of Chemistry 2014
Nanoscale
ammonia. The nal resultant solution was centrifuged at 25 000 rpm to precipitate the large akes before further characterizations were carried out. Characterizations For synthesis of the graphene quantum structures (GQDs and QGFs), a facile chemical route was utilized following the standard protocols. Extensive structural and compositional analysis were performed for our samples using high resolution transmission electron microscopy (HRTEM), Fourier transform infrared spectroscopy (FTIR), X-ray photoelectron spectroscopy (XPS) and Raman spectroscopy. Additionally, UV-vis characterizations were conducted to observe the optical absorption edge in the quantum structures. Photoluminescence and time decay analysis were carried out to investigate the excitation wavelength dependent tunable heterogeneous photoluminescence conversion along with computation of time decay proles for corresponding band to band and interstate to band transitions. Ab initio calculations were performed using the CASTEP computational algorithm. For more information on the computational modeling details, ess the ESI.†
Results and discussions TEM and Raman analysis In Fig. 1(a)–(h), HRTEM images for GQDs and QGFs are shown. We observed that well distributed uniform circular GQDs were formed with an average diameter 5–7 nm. For the amino edge terminated quantum graphene akes (QGFs), it was found that anisotropic akes were formed due to open atmosphere synthesis conditions with dimensions ranging in between 20 and 30 nm. From the high-resolution lattice images for both quantum structures (GQDs, QGFs), showed that the hexagonal lattice constant ‘a’ was 0.25 nm.25 Fig. 1(i) demonstrates Raman spectra for GQDs and QGFs respectively. The spectra for the Raman bands (D and G) represent the distinctive feature encountered for sp2 carbon nanostructures. The peak near 1590 cm1 can be assigned to G band is generally attributed to the E2g phonon of C sp2 atoms, while the peak near 1350 cm1 is generally assigned to D band which originates from a breathing k-point phonon with A1g symmetry which is related to local defects and disorders. It is quite evident that the relative D/ G ratio is greater in QGFs than GQDs along with considerable broadening of D band FWHM. This can be due to more structural disorders being present26,27 in QGFs due to open atmosphere synthesis conditions. The relative D/G intensity for GQDs and QGFs were found to be around 0.81 and 0.85 respectively which is quite consistent for these structures. UV-vis and PL data analysis Fig. 2(a) and (b) represent the combined UV-vis absorption spectrum and PL spectra under variable excitation wavelengths for amino-functionalized GQDs and QGFs respectively. From UV-vis spectrum, a broad absorption region was observed for both structures. Moreover, in QGFs we observed a similar sort of PL modulation to that of the functionalized GQDs. The
Nanoscale, 2014, 6, 3384–3391 | 3385
View Article Online
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
Nanoscale
Paper
(a–c) represent HRTEM images for GQDs and (d–f) for QGFs. (b and e) represent high resolution lattice images for GQDs and QGFs. The surface scan histogram of the high resolution lattice images for both quantum structures shown in (c and f) the surface line scans of the structures where the distance in between two consecutive peak to peak points represent the hexagonal lattice constant ‘a’ 0.25 nm. (g and h) Corresponding particle size distribution fractions for GQDs and QGFs (i) Raman spectra for GQDs and QGFs.
Fig. 1
(a and b) UV-vis absorption (black line) and excitation dependent tunable heterogeneous photoluminescence modulation spectra for amino-functionalized GQDs and QGFs (blue, cyan and green line).
Fig. 2
emission peak intensity in the green region was normalized to show how peak intensity in the blue region got suppressed with
3386 | Nanoscale, 2014, 6, 3384–3391
redshi of excitation wavelength for both structures. Nonetheless, a similar trend in PL emission pattern and UV absorption in both structures can be attributed to amino functionalization with negligible dependence on the size of the quantum structures. We believe that due to the presence of amino edge functional groups and C]C sp2 carbon domains different electronic transition states will appear due to bonding and antibonding of molecular orbitals. We attribute this tunable PL due to zigzag effect (band to band p*–p transition 429 nm) and interstate to band transitions (n–p transitions 497 nm) resulting from amino edge functionalization. Primarily p*–p transitions dominated at lower wavelength regimes. As excitation wavelength incremented to higher regimes, n–p transitions dominated while p*–p transitions got suppressed. However, equi-coupling of p*–p and n–p induced transitions resulted in dual mode cyan emission at excitation 365 nm. Although excitation dependent tunable PL is not uncommon, instead in
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
Paper
this scenario a broad range tunable PL could be observed without modifying the structural attributes through variation of pH or the solvent medium. Non-radiative oxygenated moieties were replaced by amino functional groups which are acting as new radiative recombination centers and enabling tunable PL response.28,29 Hence, effective emission from blue to cyan to green under variable excitations was observed. The excitation wavelength dependent heterogeneous PL modulation resulting from amino edge functionalization is in accordance with our proposed energy band diagram model (Fig. 3). Further investigation of the consistence of emission wavelength for both emissions (blue and green) with increasing excitation for the GQDs is shown in Fig. 4(a). It was found that the position of the emission peak wavelength remained constant as the excitation varied from 310 to 450 nm for both blue and green regions. From the derivative plot of intensity with respective emission peaks versus excitation wavelength as shown in Fig. 4(b), it could be suggested that continual suppression of blue emission with increment of excitation occurred while the peak corresponding to green emission remained unaltered throughout the excitation range. The intersection point of the two Y-axis intensities corresponding to two distinct emissive regions demarcates the origin of cyan emission further corroborating the obtained photoluminescence spectra for GQDs. Fig. 4(c) shows a ratio of the normalized intensities of peak 1 (blue) to peak 2 (green) and peak 2 to peak 1 with respect to excitation varying between 310 and 450 nm.
First principles calculations For verication of the proposed energy band model, we systematically carried out rst principles calculations and observed the creation of a new energy level due to amino group substitution at the edges of the graphene quantum structure (GQS) within the theoretical framework to validate our hypothesis. All technical computational details relating to the analysis using rst principles calculations have been provided in the characterization section of the ESI page.† To ensure a step by step identication technique towards formation of the interband, primarily three different structures were considered in our theoretical calculations as shown in Fig. 5 to model our analysis. For realistic modeling of our structures, a
Fig. 3 Schematic representation of the proposed energy band diagram model to explain p*–p and n–p transitions.
This journal is © The Royal Society of Chemistry 2014
Nanoscale
typical DV 5 8 5 lattice defect was introduced in the middle of the GQS basal plane. The combined total density of states (TDOS) plot for all structures is shown in Fig. 5(d). The dashed vertical line near zero energy designates the reference position for the Fermi level (Ef). All calculated results using rst principles calculations were subjected to a scissor correction factor of 2.5 and found to be in good agreement with the experimental results. From the TDOS results, it was clear that two separate states are created within the HOMO–LUMO gap. From our DFT calculations of structures (a and b), it was found that the new state, created corresponding to 3.47 eV, could be explicitly assigned to DV 5 8 5 defect state only.30 Further it was conrmed that the zigzag edge termination with H-atoms failed to create any new intersystem band within the energy gap. However in the case of model ‘c’ (DV 5 8 5 defect with amino group edge termination) two new interbands was formed. It was evident that the density of states increased in amino-capped GQS with DV 5 8 5 defect in comparison to structures (a) and (b). For model ‘c’ the bands formed 3.80 eV could be assigned to the DV 5 8 5 defect state and the band observed 3.28 eV can be assigned to amino-edge termination. Here, the considerable suppression of DV 5 8 5 defect band towards the conduction band can be due to more strain generated from enhanced rippled surface curvature as edge amino group substitution led to more energetically favorable optimization. So, we conclude that primary amino-group termination (NH2) at the edge states of GQS has created an additional intersystem band within the energy gap. This is consistent with our experimentally observed results. Fig. 5(e) shows the optical absorption spectra corresponding to our simulated GQS model consisting of DV 5 8 5 defect with amino-group edge termination using rst principles analysis. We could see that this simulated spectrum has tted well with our experimentally obtained results and within the acceptable range. Moreover, the partial density of states (PDOS) results as demonstrated in Fig. 5(f) and (g) suggested that p orbital hybridization had occurred in between C–N atoms at the edge sites which led to creation of a new interstate energy level at 3.28 eV due to amino group substitution. This new energy state was not observable corresponding to the p orbital PDOS with inner C atom. The modied TDOS due to functionalization in our work conrm the conclusion of earlier results reported by Galande et al.31 It is worth mentioning that for our model we chose two types of functional groups to decorate while simulating the edges for our prototype amino functionalized graphene quantum structure model. It was our expectation that amino group insertion will lead to the formation of Lewis acids through nucleophilic substitution with ammonia by a ring opening reaction, which it did. From our FTIR and XPS analysis we observed secondary out of plane attachments of O]C–NH2 groups along with that of the primary amine (NH2). As a result, these two types of amino functional groups were primarily chosen in the theoretical model. Although most of the attachment is supposed to be at zigzag edge sites, in some cases out of plane attachment at armchair sites can be a realistic probability. Hence in our model both zigzag and armchair edge terminations were taken into consideration by fully capping the
Nanoscale, 2014, 6, 3384–3391 | 3387
View Article Online
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
Nanoscale
Paper
(a) Emission peak positions (peak 1 / blue emission; peak 2 / green emission) of GQDs sample as a function of excitation wavelength; (b) intensity comparison of two peaks with respect to excitation wavelength; and (c) the ratio of peak intensities corresponding to the two distinct emissive regions.
Fig. 4
structure with amino functional groups. However, by comparing the PDOS's of p orbital edge C and N atoms with the inner p orbital C atom, we found that no new energy interstate was created for secondary amino group (O]C–NH2). We believe that both functional groups behave as acceptors in our system. However the molecular electronegativity plays a crucial role in the attachment of carbon and nitrogen atoms to the graphene quantum dot. The direct interaction of N with C (edge graphene atom) in case of NH2 causes higher chemical polarity with respect to the secondary amine (amide) group and hence manages to form an interband within the energy gap. This only reiterated the fact that the secondary O]C–NH2 group acted as a neutral component much like the situation in case of an edge
terminated H capped system and no energy transfer took place from them. This newly formed interband due to primary amine edge attachment contributed to interstate to band induced transitions and can be thought of as a manifestation of shallow defect level bands being present within the structure as a result of functionalization. Our rst principle results explicitly suggest that the presence of an interband is imperative for heterogeneous systems like functionalized graphene quantum dots and some energy transfer from direct to indirect transitions can be expected. It is our understanding that the position of the interband plays a pivotal role and its location may not always be overlapping completely with the conduction band minima, as previously presumed denoting a change only in the direct band
Fig. 5 Optimized structural configurations for graphene quantum structures (GQS) (a) pristine periodic graphene with DV 5 8 5 defect (indicated by an ellipsoid), (b) DV 5 8 5 defect with H-termination at the edge sites and (c) DV 5 8 5 defect with amino-group termination at the edge sites (d) TDOS of all configurations. Black line indicates the TDOS of structure (a), red line indicates the TDOS of structure (b) and blue line indicates the TDOS of structure (c), (e) simulated optical spectra corresponding to the graphene quantum structure with DV 5 8 5 defect and amino-group termination at the edge sites fitted with experimentally obtained UV-vis absorption spectra for GQDs and QGFs. (f) PDOS of inner carbon, edge C and N atoms from NH2. (g) PDOS of inner carbon, edge C and N atoms from O]C–NH2.
3388 | Nanoscale, 2014, 6, 3384–3391
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
Paper
gap of functionalized systems. Inbound formation of the shallow defect interband within the energy gap more importantly explained the origin of cyan like emission formed due to equi-coupling of p*–p and n–p transitions, as mentioned earlier. Thus, our modeled graphene quantum structure with amino edge capping and DV 5 8 5 defect tted well with the experimental results.
PL quantum yield and time decay analysis Subsequently, the PL quantum yield was calculated using an analytical approach following the equation given below.32 The formula gives an approximated total percentage of radiative recombination from the system and hence forms an integral part of this analysis: ð lem 2 ½I ðl x em Þ Ib ðlem Þ lem dlem sðlem Þ Nem lem1 Ff ¼ ð lex þDl ¼ ; Nabs ½Ib ðlex Þ Ix ðlex Þ lex dlex sðlex Þ lex Dl where the total number of emitted photons (Nem) is obtained upon integration of the blank-corrected (Ib(lem)) and spectrally corrected (Ix(lem)) emission spectrum of the sample. The number of absorbed photons (Nabs) follows from the integrated difference between the excitation light resulting from measurements with the blank (Ib(lex)) and the sample (Ix(lex)). s(lem) and s(lex) are spectral responsivity of the emission and excitation channel respectively. Using the above equation we calculated the overall PL quantum yield to be 29% for GQDs and 16% for QGFs, assuming Gaussian t of the total emission spectra at excitation 310 nm. The difference in PL QY among both quantum structures can be attributed to the greater number of non-radiative recombination species present in the QGFs system. Contributing to the overall PL quantum yield, the direct (band to band) transition quantum yield was found to be 75% and for interstate to band transition 25%, for both structures. These values were obtained from the de-convolution of the emission spectra at their respective regions as shown in Fig. 6(a) and (b). Similarly, the time decay proles
Nanoscale
of the experimental data for band to band and interstate to band transitions along with the tted curves for GQDs and QGFs are all illustrated in Fig. 6(c). In GQDs, the time decay proles were recorded for band to band transition at 429 nm (corresponding to excitation at 310 nm) and for interstate to band transitions 497 nm (corresponding to excitation 410 nm). Similarly for QGFs, time decay was recorded at 431 nm (corresponding excitation 310 nm) for band to band transition and interstate to band emission 498 nm (corresponding to excitation 410 nm). These data were best tted to a tri-exponential function t t t IðtÞ ¼ A1 exp þ A2 exp þ A3 exp where s1, s1 s2 s3 s2 and s3 are the decay lifetimes of the luminescence, and A1, A2 and A3 are the weighting parameters. As for the average lifetime sav, tri-exponential decay is generally ascertained by substituting various components of the materials responsible for showing luminescence.33 sav ¼
A1 s1 2 þ A2 s2 2 þ A3 s3 2 A1 s1 þ A2 s2 þ A3 s3
s1, s2 and s3 for GQDs and QGFs were found to be 2.95, 9.09 and 0.65 ns, and 2.31, 8.74 and 0.28 ns, respectively, while the average time decay sav was 8.1 ns and 8.21 ns respectively due to interstate to band transitions. For band to band transitions, s1, s2 and s3 for GQDs and QGFs were found to be 2.89, 8.82 and 0.71 ns; and 3.06, 8.84 and 0.68 ns respectively. The average time decay sav due to band-to-band transitions in GQDs and QGFs was 6.68 ns and 6.92 ns, respectively. The average time decays observed for GQDs and QGFs are within a comparable range which again indicated that size variation had a negligible effect on the time decay response. Additionally, the higher time decay for interstate to band induced transitions directly implied an effect of the intersystem crossing through shallow defect bands and has been reported recently for graphene quantum dots.34 Moreover, by using the time decay prole data, the energy transfer rate from direct to interstate emission could be calcusDd 35 lated: rate of energy transfer ¼ 1 , where sDd, is the s Di time decay prole for the band to band transitions and sDi is
Fig. 6 (a and b) Gaussian fit of the overall emission spectra (brown line) along with individual deconvoluted spectra for blue (cyan line) and green emission (green line) regions corresponding to emission spectra (blue line). The red line specifies the excitation spectra. (c) Time decay profiles for GQDs and QGFs corresponding to band to band and interstate to band transitions. The black line specifies the fitted curve while the scatter plots specify the experimental time decay profiles of both samples at different excitations.
This journal is © The Royal Society of Chemistry 2014
Nanoscale, 2014, 6, 3384–3391 | 3389
View Article Online
Nanoscale
Paper
Calculated photoluminescence quantum yield, time decay and energy transfer efficiency due to band to band and interstate to band transitions
Table 1
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
Time decay savg (ns)
PL quantum yield (QY)%
Sample
Size (nm)
Band to band
Interstate
Overall
Band to band
Interstate
Rate of energy transfer (%)
GQDs QGFs
5–7 20–30
6.68 6.92
8.10 8.21
29 16
75 75
25 25
17 15
the time decay prole for the interstate to band transitions. It was found that energy transfer efficiencies for GQDs and QGFs were found to be 17% and 15% respectively. All results corresponding to time decay and PL quantum yield parameters are shown in Table 1.
Conclusion In summary, we traced the origin of the excitation dependent heterogeneous PL modulation leading to tunable emission in amino-functionalized graphene quantum dots. The modulated PL pattern showed a negligible inuence on the size dependence which was conrmed by the synthesis of amino functionalized QGFs. First principles calculations suggested that primary amine edge termination (NH2) resulted in formation of an additional interband 3.28 eV within the energy gap due to p orbital hybridization of C–N atoms at the edge sites. Furthermore, the existence of an interband due to functionalization was primarily responsible for exhibiting tunable heterogeneous PL emission with higher time decay. Conrmation that functionalization does not always change the direct band gap was established, as newly created energy levels can also exist as interbands within the energy gap in graphene quantum dot based heterogeneous systems. Due to the inbound positioning of the shallow defect interband higher time decay and cyan like emission was possible, which was comprehensively explained through an energy band model. Finally, extensive computation of PL quantum yield, time decay and energy transfer efficiency suggested that this material could be a future candidate for prospective graphene based tunable luminescent material in opto-electronic applications.
Acknowledgements One of the authors (R.R.) would like to thank and acknowledge the Council of Scientic and Industrial Research (CSIR), Government of India, for awarding him a SRF fellowship during the tenure of this work. The authors would also like to thank the UGC for ‘University with potential for excellence scheme (UPEII)’ and the DST, the Government of India for nancial support. The authors do acknowledge and express their thanks to Mr Hitesh Mamgain (Witec-Application Scientist) for helping to carry out Raman measurements. In this paper G. S. Kumar and R. Roy contributed equally to this work.
3390 | Nanoscale, 2014, 6, 3384–3391
References 1 A. K. Geim, Graphene: Status and Prospects, Science, 2009, 19, 1530–1534. 2 A. K. Geim and K. S. Novoselov, The rise of graphene, Nat. Mater., 2007, 6, 183–191. 3 K. P. Loh, B. Qiaoliang, G. Eda and M. Chhowalla, Graphene oxide as a chemically tunable platform for optical applications, Nat. Chem., 2010, 2, 1015–1024. 4 Q. Mei and Z. Zhang, Photoluminescent Graphene Oxide Ink to Print Sensors onto Microporous Membranes for Versatile Visualization Bioassays, Angew. Chem., Int. Ed., 2012, 51, 5602–5606. 5 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, Tunable Photoluminescence from Graphene Oxide, Angew. Chem., Int. Ed., 2012, 51, 6662–6666. 6 J. Peng, W. Gao, B. K. Gupta, Z. Liu, R. R. Aburto, L. Ge, L. Song, L. B. Alemany, X. Zhan, G. Gao, S. A. Vithayathil, B. A. Kaipparettu, A. A. Marti, T. Hayashi, J. J. Zhu and P. M. Ajayan, Graphene Quantum Dots Derived from Carbon Fibers, Nano Lett., 2012, 12, 844–849. 7 S. Zhu, L. Zhang, C. Qiao, S. Tang, Y. Li, W. Yuan, B. Li, L. Tian, F. Liu, R. Hu, H. Gao, H. Wei, H. Zhang, H. Sunb and B. Yang, Strongly green photoluminescent graphene quantum dots for bioimaging applications, Chem. Commun., 2011, 47, 6858–6860. 8 S. Kim, S. W. Hwang, M. K. Kim, D. Y. Shin, D. H. Shin, C. O. Kim, S. B. Yang, J. H. Park, E. Hwang, S. H. Choi, G. Ko, S. Sim, C. Sone, H. J. Choi, S. Bae and B. H. Hong, Anomalous Behaviors of Visible Luminescence from Graphene Quantum Dots: Interplay between Size and Shape, ACS Nano, 2012, 6(9), 8203–8208. 9 S. H. Jin, D. H. Kim, G. H. Jun, S. H. Hong and S. Jeon, Tuning the Photoluminescence of Graphene Quantum Dots through the Charge Transfer Effect of Functional Groups, ACS Nano, 2012, 7(2), 1239–1245. 10 L. Li, G. Wu, G. Yang, J. Peng, J. Zhao and J. J. Zhu, Focusing on luminescent graphene quantum dots: current status and future perspectives, Nanoscale, 2013, 5, 4015–4039. 11 D. Pan, J. Zhang, Z. Li and M. Wu, Hydrothermal Route for Cutting Graphene Sheets into Blue-Luminescent Graphene Quantum Dots, Adv. Mater., 2010, 22, 734–738. 12 G. Eda, Y. Y. Lin, C. Mattevi, H. Yamaguchi, H. A. Chen, I. S. Chen, C. W. Chen and M. Chhowalla, Blue
This journal is © The Royal Society of Chemistry 2014
View Article Online
Paper
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
13
14
15
16
17
18
19
20
21
22
23 24
25
Photoluminescence from Chemically Derived Graphene Oxide, Adv. Mater., 2010, 22, 505–509. L. Tang, R. Ji, X. Cao, J. Lin, H. Jiang, X. Li, K. S. Teng, C. M. Luk, S. Zeng, J. Hao and S. P. Lau, Deep Ultraviolet Photoluminescence of Water-Soluble Self-Passivated Graphene Quantum Dots, ACS Nano, 2012, 6(6), 5102–5110. J. Shen, Y. Zhu, X. Yang, J. Zong, J. Zhang and C. Li, One-pot hydrothermal synthesis of graphene quantum dots surfacepassivated by polyethylene glycol and their photoelectric conversion under near-infrared light, New J. Chem., 2012, 36, 97–101. M. Zhang, L. Bai, W. Shang, W. Xie, H. Ma, Y. Fu, D. Fang, H. Sun, L. Fan, M. Han, C. Liub and S. Yang, Facile synthesis of water-soluble, highly uorescent graphene quantum dots as a robust biological label for stem cells, J. Mater. Chem., 2012, 22, 7461–7467. H. Tetsuka, R. Asah, A. Nagoy, K. Okamot, I. Tajima, R. Ohta and A. Okamoto, Optically Tunable Amino-Functionalized Graphene Quantum Dots, Adv. Mater., 2012, 24, 5333–5338. Q. Tang, Z. Zhou and Z. Chen, Graphene-related nanomaterials: tuning properties by functionalization, Nanoscale, 2013, 5, 4541–4583. M. Li, S. K. Cushing, X. Zhou, S. Guo and N. Wu, Fingerprinting photoluminescence of functional groups in graphene oxide, J. Mater. Chem., 2012, 10, 35417–354122. L. Irimpan, B. Krishnan, A. Deepthy, V. P. N. Nampoori and P. Radhakrishnan, Excitation wavelength dependent uorescence behaviour of nano colloids of ZnO, J. Phys. D: Appl. Phys., 2007, 40, 5670–5674. K. Ozasa, N. Shigeyuki, M. Mizuo and H. Masahiko, Excitation-wavelength-dependent photoluminescence evolution of CdSe/ZnS nanoparticles, J. Appl. Phys., 2007, 101, 103503–103508. Y. M. Huang, B. G. Zhai and F. F. Zhou, Correlation of excitation-wavelength dependent photoluminescence with the fractal microstructures of porous silicon, Appl. Surf. Sci., 2008, 254, 4139–4143. G. H. Li, Y. Zhang, Y. C. Wu and L. D. Zhang, Wavelength dependent photoluminescence of anodic alumina membranes, J. Phys.: Condens. Matter, 2003, 15, 8663–8671. W. S. Hummers and R. E. Offeman, Preparation of Graphitic Oxide, J. Am. Chem. Soc., 1958, 80, 1339. L. J. Cote, F. Kim and J. Huang, Langmuir–Blodgett Assembly of Graphite Oxide Single Layers, J. Am. Chem. Soc., 2009, 131(3), 1043–1049. Z. Zhang and P. Wu, Hydrothermal aggregation induced crystallization: a facial route towards polycrystalline
This journal is © The Royal Society of Chemistry 2014
Nanoscale
26
27
28
29
30
31
32
33
34
35
graphite quantum dots with blue photoluminescence, CrystEngComm, 2012, 14, 7149–7152. Z. Ji, X. Shen, G. Zhu, H. Zhou and A. Yuan, Reduced graphene oxide/nickel nanocomposites: facile synthesis, magnetic and catalytic properties, J. Mater. Chem., 2012, 22, 3471–3477. A. C. Ferrari, Raman spectroscopy of graphene and graphite: disorder, electron–phonon coupling, doping and nonadiabatic effects, Solid State Commun., 2007, 143, 47–57. S. Zhu, J. Zhang, S. Tang, C. Qiao, L. Wang, H. Wang, X. Liu, B. Li, Y. Li, W. Yu, X. Wang, H. Sun and B. Yang, Surface Chemistry Routes to Modulate the Photoluminescence of Graphene Quantum Dots: From Fluorescence Mechanism to Up-Conversion Bioimaging Applications, Adv. Funct. Mater., 2012, 22(22), 4732–4740. J. Shen, Y. Zhu, C. Chen, X. Yang and C. Li, Facile preparation and upconversion luminescence of graphene quantum dots, Chem. Commun., 2011, 47, 2580–2582. A. Lherbier, S. M. M. Dubois, X. Declerck, Y. M. Niquet, S. Roche and J. C. Charlier, Transport properties of 2D graphene containing structuraldefects, Phys. Rev. B: Condens. Matter Mater. Phys., 2012, 86, 075402–075419. C. Galande, A. D. Mohite, A. V. Naumov, W. Gao, L. Ci, A. Ajayan, H. Gao, A. Srivastava, R. B. Weisman, P. M. Ajayan, Quasi-Molecular Fluorescence from Graphene Oxide, 2011, DOI: 10.1038/srep00085. C. Wurth, M. Grabolle, J. Pauli, M. Spieles and U. R. Genger, Comparison of Methods and Achievable Uncertainties for the Relative and Absolute Measurement of Photoluminescence Quantum Yields, Anal. Chem., 2011, 83, 3431–3439. B. K. Gupta, P. Thanikaivelan, T. N. Narayanan, L. Song, W. Gao, T. Hayashi, A. L. M. Reddy, A. Saha, V. Shanker, M. Endo, A. A. Mart´ı and P. M. Ajayan, Optical Bifunctionality of Europium-Complexed Luminescent Graphene Nanosheets, Nano Lett., 2011, 11, 5227– 5233. F. Liu, M. H. Jang, H. D. Ha, J. H. Kim, Y. H. Cho and T. S. Seo, Facile Synthetic Method for Pristine Graphene Quantum Dots and Graphene Oxide Quantum Dots: Origin of Blue and Green Luminescence, Adv. Mater., 2013, 25(27), 3657–3662. S. Sadhu, K. K. Haldar and A. Patra, Size Dependent Resonance Energy Transfer between Semiconductor Quantum Dots and Dye Using FRET and Kinetic Model, J. Phys. Chem. C, 2010, 114(9), 3891–3897.
Nanoscale, 2014, 6, 3384–3391 | 3391
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
PAPER
View Journal | View Issue
Cite this: Nanoscale, 2014, 6, 3384
Amino-functionalized graphene quantum dots: origin of tunable heterogeneous photoluminescence† G. Sandeep Kumar,‡b Rajarshi Roy,‡a Dipayan Sen,a Uttam Kumar Ghorai,b Ranjit Thapa,c Nilesh Mazumder,a Subhajit Sahab and Kalyan K. Chattopadhyay*ab Graphene quantum dots are known to exhibit tunable photoluminescence (PL) through manipulation of edge functionality under various synthesis conditions. Here, we report observation of excitation dependent anomalous m–n type fingerprint PL transition in synthesized amino functionalized graphene quantum dots (5–7 nm). The effect of band-to-band p*–p and interstate to band n–p induced transitions led to effective multicolor emission under changeable excitation wavelength in the functionalized system. A reasonable assertion that equi-coupling of p*–p and n–p transitions activated the heterogeneous dual mode cyan emission was made upon observation of the PL spectra. Furthermore, investigation of incremented dimensional scaling through facile synthesis of amino functionalized quantum graphene flakes (20–30 nm) revealed it had negligible effect on the modulated PL pattern. Moreover, an effort was made to trace the origin of excitation dependent tunable heterogeneous photoluminescence through the framework of energy band diagram hypothesis and first principles analysis. Ab initio results suggested formation of an interband state as a manifestation of p
Received 9th October 2013 Accepted 2nd December 2013
orbital hybridization between C–N atoms at the edge sites. Therefore comprehensive theoretical and experimental analysis revealed that newly created energy levels can exist as an interband within the
DOI: 10.1039/c3nr05376h
energy gap in functionalized graphene quantum structures yielding excitation dependent tunable PL for
www.rsc.org/nanoscale
optoelectronic applications.
Introduction Graphene and its chemical derivatives have been the center of attention in nanotechnology during the latter half of the last decade exhibiting superlative electronic and mechanical properties.1–3 On the other hand, the development of uorescent graphene oxide based materials has been in demand for usage in diverse opto-electronic applications.4,5 But slowly and steadily coming into the fray PL applications based on chemically synthesized graphene quantum dots (GQDs) have gathered prominence in recent years.6–10 Due to quantization effect, one can expect interesting PL behavior in graphene based quantum dots.11–13 In an expanding technological scenario, several techniques like surface passivation methods and edge termination with functional groups inside these structures are signicantly
a
Thin Films and Nanoscience Laboratory, Dept. of Physics, Jadavpur University, Kolkata-700032, India. E-mail: [email protected]
b
School of Materials Science and Nanotechnology, Jadavpur University, Kolkata700032, India
c
SRM Research Institute, SRM University, Kattankulathur, 603203, Tamil Nadu, India
† Electronic supplementary 10.1039/c3nr05376h
information
(ESI)
available.
‡ These authors contributed equally to the present work.
3384 | Nanoscale, 2014, 6, 3384–3391
See
DOI:
contributing towards broad tunable PL by varying the excitation wavelength only.14–18 Several mechanisms have been proposed to explain this type of PL behavior, such as incomplete solvation of the excited state of the species,19,20 size distribution for nanocrystallites,21 level splitting by doped ions.22 However, the origin of tunable PL in edge functionalized graphene quantum dots is yet to be well understood. As mentioned earlier, previously reported results showed that due to excitation redshi, continual reduction of PL intensity and tunable emission was observed when the amount of functionalization was varied at different temperature conditions in graphene quantum dots.13,16 But in most of these cases, emissions due to band to band transitions were considered, whereas the roles of interstate or defect induced emissions have not been extensively probed. For functionalized systems there will be a presence of certain interbands in the structures depending on the synthesis conditions and the degree of functionalization while their positions in the optical band gap will determine the nature of tunable PL. Considering a generic case, heterogeneous emissions are expected for heterogeneous systems; however previous reports have shown that due to variation in the degree of amino functionalization the direct band gap has been manipulated in functionalized graphene quantum dots. It was inferred that more the degree of amino edge functionality shrinkage of the
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
Paper
band gap took place with a redshi of PL9,16 and in most of the cases heterogeneous emissions were not at all considered. Our current motivation for this work emphasizes probing how the presence of direct and intersystem bands will affect the nature of excitation dependent PL in functionalized graphene based quantum structures. Hence with this view in mind, in this article we report excitation wavelength dependent tunable heterogeneous PL in amino functionalized graphene quantum dots (5–7 nm) and detailed explanation is given in connection with energy band diagram analogy and ab initio approach. To understand whether this PL conversion had any size dependence, we synthesized amino edge terminated quantum graphene akes (QGFs) (20–30 nm). From rst principles calculations, we have investigated the possible origin of newly created energy levels and the governing nature of hybridization as an outcome of amino edge functionalization using a prototype structural model. Furthermore, calculation of PL quantum yield using an analytical approach, energy transfer efficiency and time decay were carried out to ascertain the impact of functionalized quantum dots as diverse carbon based luminescent material.
Experimental Sample preparation For the synthesis of GO, a modied Hummer's method approach was employed from graphite powder. As synthesized, the resultant product was then vacuum dried. The obtained GO with a loading of (1 mg ml1) was further dissolved in DI water and exfoliated for 30 min in a high power sonicator. The resultant solution was processed subsequently to lter out larger bulk particles using a high speed centrifuge for 15 min at 25 000 rpm and the supernatant was collected to obtain a yellow colored stable hydrophilic suspension of exfoliated uniform graphene oxide layers.23,24 For GQDs synthesis, 10 ml of GO solution (1 mg ml1 loading), 10 ml of DI water and 6 ml of ammonia solution were added together. This mixture was stirred for 30 min, following which ultrasonication was carried out in a 250 W high power sonicator for another 30 min. Next, the resultant solution was transferred into a Teon lined autoclave and heated at 150 C for 5 h by hydrothermal treatment. Furthermore, this solution was naturally brought down to room temperature and subsequently the GQDs were collected by ltration through a 0.22 mm Teon membrane. The resultant supernatant was heated at 100 C for another 1 h to vapor out excess ammonia. The nal resultant solution was centrifuged at 25 000 rpm before further characterizations were carried out.15,16 For QGFs synthesis, 10 ml of GO solution, 10 ml of DI water and 12 ml of ammonia solution was added together in a beaker. This mixture was stirred for 30 min, followed by ultrasonication in a 250 W high power sonicator for another 30 min. Next, this resultant solution was heated at 100 C for 4 h in open atmosphere, following which the solution was brought down to room temperature and ltered through a 0.22 mm Teon membrane and the supernatant was collected. The resultant supernatant was heated at 100 C for another 1 h to vapor out excess
This journal is © The Royal Society of Chemistry 2014
Nanoscale
ammonia. The nal resultant solution was centrifuged at 25 000 rpm to precipitate the large akes before further characterizations were carried out. Characterizations For synthesis of the graphene quantum structures (GQDs and QGFs), a facile chemical route was utilized following the standard protocols. Extensive structural and compositional analysis were performed for our samples using high resolution transmission electron microscopy (HRTEM), Fourier transform infrared spectroscopy (FTIR), X-ray photoelectron spectroscopy (XPS) and Raman spectroscopy. Additionally, UV-vis characterizations were conducted to observe the optical absorption edge in the quantum structures. Photoluminescence and time decay analysis were carried out to investigate the excitation wavelength dependent tunable heterogeneous photoluminescence conversion along with computation of time decay proles for corresponding band to band and interstate to band transitions. Ab initio calculations were performed using the CASTEP computational algorithm. For more information on the computational modeling details, ess the ESI.†
Results and discussions TEM and Raman analysis In Fig. 1(a)–(h), HRTEM images for GQDs and QGFs are shown. We observed that well distributed uniform circular GQDs were formed with an average diameter 5–7 nm. For the amino edge terminated quantum graphene akes (QGFs), it was found that anisotropic akes were formed due to open atmosphere synthesis conditions with dimensions ranging in between 20 and 30 nm. From the high-resolution lattice images for both quantum structures (GQDs, QGFs), showed that the hexagonal lattice constant ‘a’ was 0.25 nm.25 Fig. 1(i) demonstrates Raman spectra for GQDs and QGFs respectively. The spectra for the Raman bands (D and G) represent the distinctive feature encountered for sp2 carbon nanostructures. The peak near 1590 cm1 can be assigned to G band is generally attributed to the E2g phonon of C sp2 atoms, while the peak near 1350 cm1 is generally assigned to D band which originates from a breathing k-point phonon with A1g symmetry which is related to local defects and disorders. It is quite evident that the relative D/ G ratio is greater in QGFs than GQDs along with considerable broadening of D band FWHM. This can be due to more structural disorders being present26,27 in QGFs due to open atmosphere synthesis conditions. The relative D/G intensity for GQDs and QGFs were found to be around 0.81 and 0.85 respectively which is quite consistent for these structures. UV-vis and PL data analysis Fig. 2(a) and (b) represent the combined UV-vis absorption spectrum and PL spectra under variable excitation wavelengths for amino-functionalized GQDs and QGFs respectively. From UV-vis spectrum, a broad absorption region was observed for both structures. Moreover, in QGFs we observed a similar sort of PL modulation to that of the functionalized GQDs. The
Nanoscale, 2014, 6, 3384–3391 | 3385
View Article Online
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
Nanoscale
Paper
(a–c) represent HRTEM images for GQDs and (d–f) for QGFs. (b and e) represent high resolution lattice images for GQDs and QGFs. The surface scan histogram of the high resolution lattice images for both quantum structures shown in (c and f) the surface line scans of the structures where the distance in between two consecutive peak to peak points represent the hexagonal lattice constant ‘a’ 0.25 nm. (g and h) Corresponding particle size distribution fractions for GQDs and QGFs (i) Raman spectra for GQDs and QGFs.
Fig. 1
(a and b) UV-vis absorption (black line) and excitation dependent tunable heterogeneous photoluminescence modulation spectra for amino-functionalized GQDs and QGFs (blue, cyan and green line).
Fig. 2
emission peak intensity in the green region was normalized to show how peak intensity in the blue region got suppressed with
3386 | Nanoscale, 2014, 6, 3384–3391
redshi of excitation wavelength for both structures. Nonetheless, a similar trend in PL emission pattern and UV absorption in both structures can be attributed to amino functionalization with negligible dependence on the size of the quantum structures. We believe that due to the presence of amino edge functional groups and C]C sp2 carbon domains different electronic transition states will appear due to bonding and antibonding of molecular orbitals. We attribute this tunable PL due to zigzag effect (band to band p*–p transition 429 nm) and interstate to band transitions (n–p transitions 497 nm) resulting from amino edge functionalization. Primarily p*–p transitions dominated at lower wavelength regimes. As excitation wavelength incremented to higher regimes, n–p transitions dominated while p*–p transitions got suppressed. However, equi-coupling of p*–p and n–p induced transitions resulted in dual mode cyan emission at excitation 365 nm. Although excitation dependent tunable PL is not uncommon, instead in
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
Paper
this scenario a broad range tunable PL could be observed without modifying the structural attributes through variation of pH or the solvent medium. Non-radiative oxygenated moieties were replaced by amino functional groups which are acting as new radiative recombination centers and enabling tunable PL response.28,29 Hence, effective emission from blue to cyan to green under variable excitations was observed. The excitation wavelength dependent heterogeneous PL modulation resulting from amino edge functionalization is in accordance with our proposed energy band diagram model (Fig. 3). Further investigation of the consistence of emission wavelength for both emissions (blue and green) with increasing excitation for the GQDs is shown in Fig. 4(a). It was found that the position of the emission peak wavelength remained constant as the excitation varied from 310 to 450 nm for both blue and green regions. From the derivative plot of intensity with respective emission peaks versus excitation wavelength as shown in Fig. 4(b), it could be suggested that continual suppression of blue emission with increment of excitation occurred while the peak corresponding to green emission remained unaltered throughout the excitation range. The intersection point of the two Y-axis intensities corresponding to two distinct emissive regions demarcates the origin of cyan emission further corroborating the obtained photoluminescence spectra for GQDs. Fig. 4(c) shows a ratio of the normalized intensities of peak 1 (blue) to peak 2 (green) and peak 2 to peak 1 with respect to excitation varying between 310 and 450 nm.
First principles calculations For verication of the proposed energy band model, we systematically carried out rst principles calculations and observed the creation of a new energy level due to amino group substitution at the edges of the graphene quantum structure (GQS) within the theoretical framework to validate our hypothesis. All technical computational details relating to the analysis using rst principles calculations have been provided in the characterization section of the ESI page.† To ensure a step by step identication technique towards formation of the interband, primarily three different structures were considered in our theoretical calculations as shown in Fig. 5 to model our analysis. For realistic modeling of our structures, a
Fig. 3 Schematic representation of the proposed energy band diagram model to explain p*–p and n–p transitions.
This journal is © The Royal Society of Chemistry 2014
Nanoscale
typical DV 5 8 5 lattice defect was introduced in the middle of the GQS basal plane. The combined total density of states (TDOS) plot for all structures is shown in Fig. 5(d). The dashed vertical line near zero energy designates the reference position for the Fermi level (Ef). All calculated results using rst principles calculations were subjected to a scissor correction factor of 2.5 and found to be in good agreement with the experimental results. From the TDOS results, it was clear that two separate states are created within the HOMO–LUMO gap. From our DFT calculations of structures (a and b), it was found that the new state, created corresponding to 3.47 eV, could be explicitly assigned to DV 5 8 5 defect state only.30 Further it was conrmed that the zigzag edge termination with H-atoms failed to create any new intersystem band within the energy gap. However in the case of model ‘c’ (DV 5 8 5 defect with amino group edge termination) two new interbands was formed. It was evident that the density of states increased in amino-capped GQS with DV 5 8 5 defect in comparison to structures (a) and (b). For model ‘c’ the bands formed 3.80 eV could be assigned to the DV 5 8 5 defect state and the band observed 3.28 eV can be assigned to amino-edge termination. Here, the considerable suppression of DV 5 8 5 defect band towards the conduction band can be due to more strain generated from enhanced rippled surface curvature as edge amino group substitution led to more energetically favorable optimization. So, we conclude that primary amino-group termination (NH2) at the edge states of GQS has created an additional intersystem band within the energy gap. This is consistent with our experimentally observed results. Fig. 5(e) shows the optical absorption spectra corresponding to our simulated GQS model consisting of DV 5 8 5 defect with amino-group edge termination using rst principles analysis. We could see that this simulated spectrum has tted well with our experimentally obtained results and within the acceptable range. Moreover, the partial density of states (PDOS) results as demonstrated in Fig. 5(f) and (g) suggested that p orbital hybridization had occurred in between C–N atoms at the edge sites which led to creation of a new interstate energy level at 3.28 eV due to amino group substitution. This new energy state was not observable corresponding to the p orbital PDOS with inner C atom. The modied TDOS due to functionalization in our work conrm the conclusion of earlier results reported by Galande et al.31 It is worth mentioning that for our model we chose two types of functional groups to decorate while simulating the edges for our prototype amino functionalized graphene quantum structure model. It was our expectation that amino group insertion will lead to the formation of Lewis acids through nucleophilic substitution with ammonia by a ring opening reaction, which it did. From our FTIR and XPS analysis we observed secondary out of plane attachments of O]C–NH2 groups along with that of the primary amine (NH2). As a result, these two types of amino functional groups were primarily chosen in the theoretical model. Although most of the attachment is supposed to be at zigzag edge sites, in some cases out of plane attachment at armchair sites can be a realistic probability. Hence in our model both zigzag and armchair edge terminations were taken into consideration by fully capping the
Nanoscale, 2014, 6, 3384–3391 | 3387
View Article Online
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
Nanoscale
Paper
(a) Emission peak positions (peak 1 / blue emission; peak 2 / green emission) of GQDs sample as a function of excitation wavelength; (b) intensity comparison of two peaks with respect to excitation wavelength; and (c) the ratio of peak intensities corresponding to the two distinct emissive regions.
Fig. 4
structure with amino functional groups. However, by comparing the PDOS's of p orbital edge C and N atoms with the inner p orbital C atom, we found that no new energy interstate was created for secondary amino group (O]C–NH2). We believe that both functional groups behave as acceptors in our system. However the molecular electronegativity plays a crucial role in the attachment of carbon and nitrogen atoms to the graphene quantum dot. The direct interaction of N with C (edge graphene atom) in case of NH2 causes higher chemical polarity with respect to the secondary amine (amide) group and hence manages to form an interband within the energy gap. This only reiterated the fact that the secondary O]C–NH2 group acted as a neutral component much like the situation in case of an edge
terminated H capped system and no energy transfer took place from them. This newly formed interband due to primary amine edge attachment contributed to interstate to band induced transitions and can be thought of as a manifestation of shallow defect level bands being present within the structure as a result of functionalization. Our rst principle results explicitly suggest that the presence of an interband is imperative for heterogeneous systems like functionalized graphene quantum dots and some energy transfer from direct to indirect transitions can be expected. It is our understanding that the position of the interband plays a pivotal role and its location may not always be overlapping completely with the conduction band minima, as previously presumed denoting a change only in the direct band
Fig. 5 Optimized structural configurations for graphene quantum structures (GQS) (a) pristine periodic graphene with DV 5 8 5 defect (indicated by an ellipsoid), (b) DV 5 8 5 defect with H-termination at the edge sites and (c) DV 5 8 5 defect with amino-group termination at the edge sites (d) TDOS of all configurations. Black line indicates the TDOS of structure (a), red line indicates the TDOS of structure (b) and blue line indicates the TDOS of structure (c), (e) simulated optical spectra corresponding to the graphene quantum structure with DV 5 8 5 defect and amino-group termination at the edge sites fitted with experimentally obtained UV-vis absorption spectra for GQDs and QGFs. (f) PDOS of inner carbon, edge C and N atoms from NH2. (g) PDOS of inner carbon, edge C and N atoms from O]C–NH2.
3388 | Nanoscale, 2014, 6, 3384–3391
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
Paper
gap of functionalized systems. Inbound formation of the shallow defect interband within the energy gap more importantly explained the origin of cyan like emission formed due to equi-coupling of p*–p and n–p transitions, as mentioned earlier. Thus, our modeled graphene quantum structure with amino edge capping and DV 5 8 5 defect tted well with the experimental results.
PL quantum yield and time decay analysis Subsequently, the PL quantum yield was calculated using an analytical approach following the equation given below.32 The formula gives an approximated total percentage of radiative recombination from the system and hence forms an integral part of this analysis: ð lem 2 ½I ðl x em Þ Ib ðlem Þ lem dlem sðlem Þ Nem lem1 Ff ¼ ð lex þDl ¼ ; Nabs ½Ib ðlex Þ Ix ðlex Þ lex dlex sðlex Þ lex Dl where the total number of emitted photons (Nem) is obtained upon integration of the blank-corrected (Ib(lem)) and spectrally corrected (Ix(lem)) emission spectrum of the sample. The number of absorbed photons (Nabs) follows from the integrated difference between the excitation light resulting from measurements with the blank (Ib(lex)) and the sample (Ix(lex)). s(lem) and s(lex) are spectral responsivity of the emission and excitation channel respectively. Using the above equation we calculated the overall PL quantum yield to be 29% for GQDs and 16% for QGFs, assuming Gaussian t of the total emission spectra at excitation 310 nm. The difference in PL QY among both quantum structures can be attributed to the greater number of non-radiative recombination species present in the QGFs system. Contributing to the overall PL quantum yield, the direct (band to band) transition quantum yield was found to be 75% and for interstate to band transition 25%, for both structures. These values were obtained from the de-convolution of the emission spectra at their respective regions as shown in Fig. 6(a) and (b). Similarly, the time decay proles
Nanoscale
of the experimental data for band to band and interstate to band transitions along with the tted curves for GQDs and QGFs are all illustrated in Fig. 6(c). In GQDs, the time decay proles were recorded for band to band transition at 429 nm (corresponding to excitation at 310 nm) and for interstate to band transitions 497 nm (corresponding to excitation 410 nm). Similarly for QGFs, time decay was recorded at 431 nm (corresponding excitation 310 nm) for band to band transition and interstate to band emission 498 nm (corresponding to excitation 410 nm). These data were best tted to a tri-exponential function t t t IðtÞ ¼ A1 exp þ A2 exp þ A3 exp where s1, s1 s2 s3 s2 and s3 are the decay lifetimes of the luminescence, and A1, A2 and A3 are the weighting parameters. As for the average lifetime sav, tri-exponential decay is generally ascertained by substituting various components of the materials responsible for showing luminescence.33 sav ¼
A1 s1 2 þ A2 s2 2 þ A3 s3 2 A1 s1 þ A2 s2 þ A3 s3
s1, s2 and s3 for GQDs and QGFs were found to be 2.95, 9.09 and 0.65 ns, and 2.31, 8.74 and 0.28 ns, respectively, while the average time decay sav was 8.1 ns and 8.21 ns respectively due to interstate to band transitions. For band to band transitions, s1, s2 and s3 for GQDs and QGFs were found to be 2.89, 8.82 and 0.71 ns; and 3.06, 8.84 and 0.68 ns respectively. The average time decay sav due to band-to-band transitions in GQDs and QGFs was 6.68 ns and 6.92 ns, respectively. The average time decays observed for GQDs and QGFs are within a comparable range which again indicated that size variation had a negligible effect on the time decay response. Additionally, the higher time decay for interstate to band induced transitions directly implied an effect of the intersystem crossing through shallow defect bands and has been reported recently for graphene quantum dots.34 Moreover, by using the time decay prole data, the energy transfer rate from direct to interstate emission could be calcusDd 35 lated: rate of energy transfer ¼ 1 , where sDd, is the s Di time decay prole for the band to band transitions and sDi is
Fig. 6 (a and b) Gaussian fit of the overall emission spectra (brown line) along with individual deconvoluted spectra for blue (cyan line) and green emission (green line) regions corresponding to emission spectra (blue line). The red line specifies the excitation spectra. (c) Time decay profiles for GQDs and QGFs corresponding to band to band and interstate to band transitions. The black line specifies the fitted curve while the scatter plots specify the experimental time decay profiles of both samples at different excitations.
This journal is © The Royal Society of Chemistry 2014
Nanoscale, 2014, 6, 3384–3391 | 3389
View Article Online
Nanoscale
Paper
Calculated photoluminescence quantum yield, time decay and energy transfer efficiency due to band to band and interstate to band transitions
Table 1
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
Time decay savg (ns)
PL quantum yield (QY)%
Sample
Size (nm)
Band to band
Interstate
Overall
Band to band
Interstate
Rate of energy transfer (%)
GQDs QGFs
5–7 20–30
6.68 6.92
8.10 8.21
29 16
75 75
25 25
17 15
the time decay prole for the interstate to band transitions. It was found that energy transfer efficiencies for GQDs and QGFs were found to be 17% and 15% respectively. All results corresponding to time decay and PL quantum yield parameters are shown in Table 1.
Conclusion In summary, we traced the origin of the excitation dependent heterogeneous PL modulation leading to tunable emission in amino-functionalized graphene quantum dots. The modulated PL pattern showed a negligible inuence on the size dependence which was conrmed by the synthesis of amino functionalized QGFs. First principles calculations suggested that primary amine edge termination (NH2) resulted in formation of an additional interband 3.28 eV within the energy gap due to p orbital hybridization of C–N atoms at the edge sites. Furthermore, the existence of an interband due to functionalization was primarily responsible for exhibiting tunable heterogeneous PL emission with higher time decay. Conrmation that functionalization does not always change the direct band gap was established, as newly created energy levels can also exist as interbands within the energy gap in graphene quantum dot based heterogeneous systems. Due to the inbound positioning of the shallow defect interband higher time decay and cyan like emission was possible, which was comprehensively explained through an energy band model. Finally, extensive computation of PL quantum yield, time decay and energy transfer efficiency suggested that this material could be a future candidate for prospective graphene based tunable luminescent material in opto-electronic applications.
Acknowledgements One of the authors (R.R.) would like to thank and acknowledge the Council of Scientic and Industrial Research (CSIR), Government of India, for awarding him a SRF fellowship during the tenure of this work. The authors would also like to thank the UGC for ‘University with potential for excellence scheme (UPEII)’ and the DST, the Government of India for nancial support. The authors do acknowledge and express their thanks to Mr Hitesh Mamgain (Witec-Application Scientist) for helping to carry out Raman measurements. In this paper G. S. Kumar and R. Roy contributed equally to this work.
3390 | Nanoscale, 2014, 6, 3384–3391
References 1 A. K. Geim, Graphene: Status and Prospects, Science, 2009, 19, 1530–1534. 2 A. K. Geim and K. S. Novoselov, The rise of graphene, Nat. Mater., 2007, 6, 183–191. 3 K. P. Loh, B. Qiaoliang, G. Eda and M. Chhowalla, Graphene oxide as a chemically tunable platform for optical applications, Nat. Chem., 2010, 2, 1015–1024. 4 Q. Mei and Z. Zhang, Photoluminescent Graphene Oxide Ink to Print Sensors onto Microporous Membranes for Versatile Visualization Bioassays, Angew. Chem., Int. Ed., 2012, 51, 5602–5606. 5 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, Tunable Photoluminescence from Graphene Oxide, Angew. Chem., Int. Ed., 2012, 51, 6662–6666. 6 J. Peng, W. Gao, B. K. Gupta, Z. Liu, R. R. Aburto, L. Ge, L. Song, L. B. Alemany, X. Zhan, G. Gao, S. A. Vithayathil, B. A. Kaipparettu, A. A. Marti, T. Hayashi, J. J. Zhu and P. M. Ajayan, Graphene Quantum Dots Derived from Carbon Fibers, Nano Lett., 2012, 12, 844–849. 7 S. Zhu, L. Zhang, C. Qiao, S. Tang, Y. Li, W. Yuan, B. Li, L. Tian, F. Liu, R. Hu, H. Gao, H. Wei, H. Zhang, H. Sunb and B. Yang, Strongly green photoluminescent graphene quantum dots for bioimaging applications, Chem. Commun., 2011, 47, 6858–6860. 8 S. Kim, S. W. Hwang, M. K. Kim, D. Y. Shin, D. H. Shin, C. O. Kim, S. B. Yang, J. H. Park, E. Hwang, S. H. Choi, G. Ko, S. Sim, C. Sone, H. J. Choi, S. Bae and B. H. Hong, Anomalous Behaviors of Visible Luminescence from Graphene Quantum Dots: Interplay between Size and Shape, ACS Nano, 2012, 6(9), 8203–8208. 9 S. H. Jin, D. H. Kim, G. H. Jun, S. H. Hong and S. Jeon, Tuning the Photoluminescence of Graphene Quantum Dots through the Charge Transfer Effect of Functional Groups, ACS Nano, 2012, 7(2), 1239–1245. 10 L. Li, G. Wu, G. Yang, J. Peng, J. Zhao and J. J. Zhu, Focusing on luminescent graphene quantum dots: current status and future perspectives, Nanoscale, 2013, 5, 4015–4039. 11 D. Pan, J. Zhang, Z. Li and M. Wu, Hydrothermal Route for Cutting Graphene Sheets into Blue-Luminescent Graphene Quantum Dots, Adv. Mater., 2010, 22, 734–738. 12 G. Eda, Y. Y. Lin, C. Mattevi, H. Yamaguchi, H. A. Chen, I. S. Chen, C. W. Chen and M. Chhowalla, Blue
This journal is © The Royal Society of Chemistry 2014
View Article Online
Paper
Published on 05 December 2013. Downloaded by University of Southern California on 04/04/2014 08:44:45.
13
14
15
16
17
18
19
20
21
22
23 24
25
Photoluminescence from Chemically Derived Graphene Oxide, Adv. Mater., 2010, 22, 505–509. L. Tang, R. Ji, X. Cao, J. Lin, H. Jiang, X. Li, K. S. Teng, C. M. Luk, S. Zeng, J. Hao and S. P. Lau, Deep Ultraviolet Photoluminescence of Water-Soluble Self-Passivated Graphene Quantum Dots, ACS Nano, 2012, 6(6), 5102–5110. J. Shen, Y. Zhu, X. Yang, J. Zong, J. Zhang and C. Li, One-pot hydrothermal synthesis of graphene quantum dots surfacepassivated by polyethylene glycol and their photoelectric conversion under near-infrared light, New J. Chem., 2012, 36, 97–101. M. Zhang, L. Bai, W. Shang, W. Xie, H. Ma, Y. Fu, D. Fang, H. Sun, L. Fan, M. Han, C. Liub and S. Yang, Facile synthesis of water-soluble, highly uorescent graphene quantum dots as a robust biological label for stem cells, J. Mater. Chem., 2012, 22, 7461–7467. H. Tetsuka, R. Asah, A. Nagoy, K. Okamot, I. Tajima, R. Ohta and A. Okamoto, Optically Tunable Amino-Functionalized Graphene Quantum Dots, Adv. Mater., 2012, 24, 5333–5338. Q. Tang, Z. Zhou and Z. Chen, Graphene-related nanomaterials: tuning properties by functionalization, Nanoscale, 2013, 5, 4541–4583. M. Li, S. K. Cushing, X. Zhou, S. Guo and N. Wu, Fingerprinting photoluminescence of functional groups in graphene oxide, J. Mater. Chem., 2012, 10, 35417–354122. L. Irimpan, B. Krishnan, A. Deepthy, V. P. N. Nampoori and P. Radhakrishnan, Excitation wavelength dependent uorescence behaviour of nano colloids of ZnO, J. Phys. D: Appl. Phys., 2007, 40, 5670–5674. K. Ozasa, N. Shigeyuki, M. Mizuo and H. Masahiko, Excitation-wavelength-dependent photoluminescence evolution of CdSe/ZnS nanoparticles, J. Appl. Phys., 2007, 101, 103503–103508. Y. M. Huang, B. G. Zhai and F. F. Zhou, Correlation of excitation-wavelength dependent photoluminescence with the fractal microstructures of porous silicon, Appl. Surf. Sci., 2008, 254, 4139–4143. G. H. Li, Y. Zhang, Y. C. Wu and L. D. Zhang, Wavelength dependent photoluminescence of anodic alumina membranes, J. Phys.: Condens. Matter, 2003, 15, 8663–8671. W. S. Hummers and R. E. Offeman, Preparation of Graphitic Oxide, J. Am. Chem. Soc., 1958, 80, 1339. L. J. Cote, F. Kim and J. Huang, Langmuir–Blodgett Assembly of Graphite Oxide Single Layers, J. Am. Chem. Soc., 2009, 131(3), 1043–1049. Z. Zhang and P. Wu, Hydrothermal aggregation induced crystallization: a facial route towards polycrystalline
This journal is © The Royal Society of Chemistry 2014
Nanoscale
26
27
28
29
30
31
32
33
34
35
graphite quantum dots with blue photoluminescence, CrystEngComm, 2012, 14, 7149–7152. Z. Ji, X. Shen, G. Zhu, H. Zhou and A. Yuan, Reduced graphene oxide/nickel nanocomposites: facile synthesis, magnetic and catalytic properties, J. Mater. Chem., 2012, 22, 3471–3477. A. C. Ferrari, Raman spectroscopy of graphene and graphite: disorder, electron–phonon coupling, doping and nonadiabatic effects, Solid State Commun., 2007, 143, 47–57. S. Zhu, J. Zhang, S. Tang, C. Qiao, L. Wang, H. Wang, X. Liu, B. Li, Y. Li, W. Yu, X. Wang, H. Sun and B. Yang, Surface Chemistry Routes to Modulate the Photoluminescence of Graphene Quantum Dots: From Fluorescence Mechanism to Up-Conversion Bioimaging Applications, Adv. Funct. Mater., 2012, 22(22), 4732–4740. J. Shen, Y. Zhu, C. Chen, X. Yang and C. Li, Facile preparation and upconversion luminescence of graphene quantum dots, Chem. Commun., 2011, 47, 2580–2582. A. Lherbier, S. M. M. Dubois, X. Declerck, Y. M. Niquet, S. Roche and J. C. Charlier, Transport properties of 2D graphene containing structuraldefects, Phys. Rev. B: Condens. Matter Mater. Phys., 2012, 86, 075402–075419. C. Galande, A. D. Mohite, A. V. Naumov, W. Gao, L. Ci, A. Ajayan, H. Gao, A. Srivastava, R. B. Weisman, P. M. Ajayan, Quasi-Molecular Fluorescence from Graphene Oxide, 2011, DOI: 10.1038/srep00085. C. Wurth, M. Grabolle, J. Pauli, M. Spieles and U. R. Genger, Comparison of Methods and Achievable Uncertainties for the Relative and Absolute Measurement of Photoluminescence Quantum Yields, Anal. Chem., 2011, 83, 3431–3439. B. K. Gupta, P. Thanikaivelan, T. N. Narayanan, L. Song, W. Gao, T. Hayashi, A. L. M. Reddy, A. Saha, V. Shanker, M. Endo, A. A. Mart´ı and P. M. Ajayan, Optical Bifunctionality of Europium-Complexed Luminescent Graphene Nanosheets, Nano Lett., 2011, 11, 5227– 5233. F. Liu, M. H. Jang, H. D. Ha, J. H. Kim, Y. H. Cho and T. S. Seo, Facile Synthetic Method for Pristine Graphene Quantum Dots and Graphene Oxide Quantum Dots: Origin of Blue and Green Luminescence, Adv. Mater., 2013, 25(27), 3657–3662. S. Sadhu, K. K. Haldar and A. Patra, Size Dependent Resonance Energy Transfer between Semiconductor Quantum Dots and Dye Using FRET and Kinetic Model, J. Phys. Chem. C, 2010, 114(9), 3891–3897.
Nanoscale, 2014, 6, 3384–3391 | 3391
