Fluorescence Dynamics in BSA-Protected Au25 Nanoclusters - The

Aug 3, 2012 - In this work we investigate the fluorescence dynamics in BSA-protected Au25 nanoclusters by time-resolved photoluminescence and transien...
0 downloads 9 Views 1MB Size
Article pubs.acs.org/JPCC

Fluorescence Dynamics in BSA-Protected Au25 Nanoclusters Xiaoming Wen,* Pyng Yu, Yon-Rui Toh, An-Chia Hsu, Yu-Chieh Lee, and Jau Tang* Research Center for Applied Sciences, Academia Sinica, Taipei, Taiwan ABSTRACT: Fluorescent Au25 nanoclusters recently have drawn considerable research interest due to their unique properties and potential applications. Despite significant advances in their synthesis methods and application development, the origin of the fluorescence and underlying mechanism still remain unclear. In this work we investigate the fluorescence dynamics in BSA-protected Au25 nanoclusters by time-resolved photoluminescence and transient absorption techniques covering picosecond to microsecond time scales. We demonstrate here that the red fluorescence consists of both prompt fluorescence and thermally activated delayed fluorescence, and the latter is more dominant. A small energy gap of 30 meV between the triplet and the singlet states was determined from our temperature-dependent time-resolved fluorescence measurement. Moreover, we elucidate that the absorption band at 2.34 eV corresponds to the HOMO− LUMO transition in this system due to the interaction between Au25 NCs and BSA. We also show that an effective relaxation pathway exists from the higher excited state to the LUMO. confinement.21 However, slightly different PL was observed in Au25 NCs with various ligands in the red or near-infrared,16,21,22 such as BSA at 640 nm9 and 674 nm,23 phenylethanethiol at 750 nm,5 glutathione at 700 nm,16 MSA at 700 nm,18 pepsin at 670 nm,24 and DHLA at 684 nm.25 Moreover, our experiment reveals that the PL wavelength of Au25 NCs can evidently be influenced by the pH even with the same BSA ligand. These observations confirm that the red PL is determined not only by the cluster itself but also by the ligands. The electronic structure of Au25 NCs has been extensively studied.7,26−28 It has been shown that Au25 NCs have a core− shell structure in which 13 Au(0) atoms form an icosahedral core surrounded by 6 Au2(SR)3 staples.3,5 Such a structure has been proven independent of the types of surface ligands.5 Au25 NCs exhibit molecule-like electronic structure with an energy gap, from the highest occupied molecular orbital (HOMO) to lowest unoccupied molecular orbital (LUMO) in the infrared.7,16 Link et al. observed two luminescence bands at 1.5 and 1.15 eV in glutathione-protected Au28 NCs and assigned them to fluorescence and phosphorescence, respectively.17 Controversially, other researchers suggested that the red band originates from interband transition between the 6sp conduction band and the filled 5d band.5,16,22 The timeresolved luminescence composed of a fast nanosecond component followed by a slow microsecond component has been observed.5,17,22 So far, obviously separated PL bands in the red have not been found in Au25 NCs although nanosecond fast and microsecond slow components were observed.5,22 Furthermore, Wu and colleagues suggested that charge transfer

1. INTRODUCTION Gold nanoclusters (NCs) have attracted great interest for both fundamental research and a wide range of applications, including catalysis, biosensing, photonics, and molecular electronics.1−3 As a consequence of enhanced quantum confinement effects, these ultrasmall NCs exhibit distinct features, such as unusual optical, electronic, and charging properties when their sizes become comparable to the Fermi wavelength.1,4 The NCs possess discrete energy levels and molecule-like properties in the absorption and fluorescence features.5−8 Moreover, the nanoclusters are too small to support surface plasmon resonance and thus are significantly differentiated from their counterpart plasmonic nanoparticles.5 Recently, gold NCs with a precisely controlled metal core have been synthesized with a specific number of metal atoms according to the requested emission wavelength and a monolayer ligand shell. In particular, Au25 NCs have been synthesized with various ligands, such as thiols, amine groups, and proteins (e.g., glutathione and bovine serum albumin (BSA)).9−11 Such Au25 NCs are highly luminescent with excellent photostability, offering great potential applications in biology and photonics.1,3,11 Photoluminescence (PL) is one of the most important properties of Au NCs because it is very relevant to most of their applications, such as biolabeling, bioimaging, and photonics. Physical understanding of the PL is necessary to significantly facilitate the progress of fundamental research and practical applications. The optical properties of Au25 NCs have been investigated in the past decade with various techniques,12−14 including an ultrafast pump−probe,15 time-resolved and ultrafast fluorescence,16,17 and temperature dependence.8,18−20 Essentially, the emission wavelength is determined by the number of atoms with the N1/3 law according to quantum © 2012 American Chemical Society

Received: June 15, 2012 Revised: July 30, 2012 Published: August 3, 2012 19032

dx.doi.org/10.1021/jp305902w | J. Phys. Chem. C 2012, 116, 19032−19038

The Journal of Physical Chemistry C

Article

correlated single-photon counting (TCSPC) technique on a Microtime-200 system (Picoquant) with laser excitation at 467 nm. The direct detection technique30 is applied to measure microsecond lifetimes using a photomultiplier tube (PMT, R928). An ultrashort 100 fs pulse is used as excitation source output from an optical parametric amplifier (OPA, TOPAS) pumped by a Ti:sapphire laser-seeded regenerative amplifier at 1 kHz repetition. The excitation wavelength is tuned at 400 nm to match the absorption, and the evolution of fluorescence was recorded by an oscilloscope (Tektronix MSO4032). Ultrafast femtosecond to picosecond transient absorption was measured with a pump (Spitfire) at 400 nm and a tunable probe from OPA. The time-gated PL and nanosecond to microsecond transient absorption spectra were acquired using an ICCD in a laser flash photolysis system (LP920-S) excited at 355 nm from a Q-switched Nd:YAG laser.

from the surface ligands to the core plays a major role in the red fluorescence.5 Despite considerable theoretical and experimental work on the PL of Au25 NCs, to date, fundamental understanding is far from complete. In particular, the PL mechanism and origin still remain unclear; the role of the triplet state and correlation with its electronic structure as well as electronic dynamics still remain unclear and urgently require clarification.3,529 Herein we investigate PL dynamics in BSA-protected Au25 NCs by timeresolved PL and transient absorption in the time scales of picoseconds to microseconds to obtain detailed insight for physical understanding. Our experiments indicate that the PL of BSA-protected Au25 NCs consists of prompt fluorescence (PF) in a nanosecond time scale and delayed fluorescence (DF) in a microsecond time scale (Scheme 1). The thermally activated DF dominates the red band due to efficient intersystem crossings (ISC) as a consequence of a very small energy gap between the singlet and the triplet states. Moreover, the experiments reveal that the transition of HOMO−LUMO is located at 2.34 eV and an effective relaxation occurs from a higher excited state to the LUMO.

3. RESULTS AND DISCUSSION The absorption and fluorescence spectra of pure BSA and BSAprotected Au25 NCs were observed previously.8 The absorption peak appears around 530 nm (2.34 eV) with no structure on the red side, and fluorescence appears at 690 nm, consistent with the observation by Xie et al.9 It should be noted that the absorption is evidently different from the Au25 NCs protected by glutathione16 and SRSC6H13 or SC12H255 in which an absorption peak appears around 680 nm that is attributed to the transition of LUMO−HOMO.15,17,31 The weak blue band was attributed to BSA emission because pure BSA exhibits similar fluorescence, and this blue band decreases upon formation of Au25 NCs.32 It is found that the PL spectra are almost the same at 355, 400, and 467 nm excitations. Figure 1 shows the PL evolutions (a) for nanosecond time scale and (b) for microsecond time scale, measured by TCSPC. At each wavelength, the evolution includes a fast nanosecond component followed by a slow microsecond component. The fast component exhibits a two-exponential decay, and the decay times are determined by two-exponential fit. The fitting parameters are summarized in Table 1. At each wavelength, the fast decay time τ1 is very similar, ∼1.2 ns, whereas the slow decay time τ2 slightly increases from 5.27 ns at 565 nm to 7.88 ns at 690 nm. In the microsecond time scale, the lifetime of the slow component is wavelength-independent between 600 and 800 nm and the evolutions can be well fit by a single exponential function. Figure 2 shows the temperature-dependent PL evolution of 650 nm from 77 K to 300 K. The lifetime decreases with increasing temperature, 4.85 μs at 77 K and 1.55 μs at 300 K. The intensity ratio between the fast and slow components is wavelength-dependent. At 550 nm, the short component is dominant. A significantly higher sharp peak and a very low slow component are observed. At 600 nm, the sharp peak decreases and the slow component obviously increases. A similar small sharp peak followed by a strong slow component is observed between 650 and 800 nm. Figure 3a shows time-gated fluorescence spectra with a fixed gate of 10 ns and different time delays. The spectrum of a 2 ns delay should mainly arise from the fast components, and the more delayed spectra solely consist of a slow component according to the lifetime measurement. At a 20 ns delay, the blue band almost disappears because of the short lifetime. It is clear that there is no evident difference in the spectrum between the fast and slow components. The transient absorption in the nanosecond

Scheme 1. Proposed Luminescence Mechanism in Au25@ BSA NCsa

a

S and T represent singlet and triplet states; ISC and RISC are intersystem crossing and reverse intersystem crossing, respectively.

2. EXPERIMENTAL SECTION a. Synthesis of Au25/BSA and Sample Fabrication. The Au25 nanoclusters used in this study were synthesized using a biomineralized approach developed by Ying et al.,9 similar to the previous study.8 The clusters comprising 25 gold atoms were formed and stabilized by the thiol group of cysteine in BSA. The film samples were fabricated by the conventional drop-casting technique for spectroscopic measurements. Experiments demonstrate the excellent chemical and photostability of Au25 NCs because the samples exhibit the same fluorescence spectrum after low temperature and spectroscopy experiments, consistent with other observations.23 b. Spectroscopic Measurements. Steady-state fluorescence was measured in a MicroHR spectrometer (Horiba Jobin Yvon) with excitation at 406 nm. The spectrum was recorded by a cooled CCD (SynapseTM CCD). For the temperaturedependent experiment, the sample was placed in a cryostat (ST500) with controllable temperature between 77 K and 450 K. The nanosecond lifetimes were measured using the time19033

dx.doi.org/10.1021/jp305902w | J. Phys. Chem. C 2012, 116, 19032−19038

The Journal of Physical Chemistry C

Article

Figure 3. (a) Time-gated fluorescence and (b) transient absorption with a fixed gate of 10 ns at different delay times excited at 355 nm.

Figure 1. Time evolution of luminescence measured by TCSPC with excitation at 467 nm and detection at various wavelengths (a) in the prompt regime and (b) in the delayed regime. Horizontal and vertical offsets are applied in panel b to avoid overlap.

lifetime, which suggests that the carriers are likely relevant to the triplet state. We propose that the fluorescence originates from the spectral overlap of prompt fluorescence and delayed fluorescence, namely, the fast component in nanoseconds is PF and the slow component in microseconds is DF. It is clear that the DF exhibits the same spectrum as the PF based on the gated fluorescence. We can approximately estimate the intensity ratio, IDF/IPF, from the TCSPC measurement by integrating the intensity in the prompt and delayed time scales. The ratios of 0.521 at 550 nm, 5.75 at 600 nm, and 10.76 at 700 nm are acquired. This result suggests that the red band predominantly originates from the slow component, DF. Essentially, delayed fluorescence can arise from thermal activation from the excited triplet state (E-type) or can be a consequence of triplet−triplet annihilation (P-type).33−36 Although known for many years, DF continues to be a rare phenomenon observed in organic semiconductors and metal complexes. For a given fluorophore, DF is usually much weaker than its PF.35,37,38 A DF has a spectrum similar to that of a PF, but it has a much longer lifetime than the PF because the population of the excited singlet state originates from the triplet state. E-type DF is generated by the process in which the excited singlet state populates by thermal activation from the triplet state for a small energy gap EST.39 In this case, the DF and the concomitant phosphorescence have an equal lifetime because the populations of the singlet and the triplet states are

Figure 2. PL evolutions of Au25 NCs as a function of temperature at 650 nm with excitation at 400 nm.

to microsecond time scale can arise from the excited state absorption of singlet and/or triplet states (Figure 3b). It reveals that the carriers at excited states exhibit an evidentally long 19034

dx.doi.org/10.1021/jp305902w | J. Phys. Chem. C 2012, 116, 19032−19038

The Journal of Physical Chemistry C

Article

We measure the decay rate of DF as a function of temperature, as shown in Figure 4. The red curve is the fit

in thermal equilibrium. In contrast, P-type DF arises from triplet−triplet annihilation, thus producing one electron in the excited singlet state and the other electron in the ground state. The lifetime of DF is evidently shorter than that of the concomitant phosphorescence because of the biphotonic process.33 To generate effective DF, reasonably high probability of ISC (S1→T1) and subsequent reverse ISC (T1→S1) are essential. In other words, high quantum yields of triplet formation ηS→T and singlet formation ηT→S should occur. The intensity ratio of DF and PF can be expressed as:37,40 η IDF 1 = DF = IPF 1/ηS → TηT → S − 1 ηPF

(1)

The ratio can be estimated based on the dynamic measurement. The short lifetime component exhibits twoexponential decay that should arise from ISC (KS→T ISC ) and PF (KPF). We suppose KS→T > K because DF is dominant. ISC PF Therefore, we ascribed the shorter decay time to KS→T ISC and the longer decay time to KPF. The quantum efficiency of triplet S→T formation can be estimated as ηS→T = KS→T ISC /(KPF + KISC ) = 86.6% at 690 nm. The lifetime of phosphorescence is usually in the range of hundreds of microseconds to seconds, much longer than the observed lifetime of DF, within a few microseconds. It is reasonably assured that the nonradiative trapping of triplet is negligible because the DF lifetime of 1.5 μs is significantly shorter than the normal lifetime of triplet states (>hundreds of microseconds).35,39 The efficiency of triplet to singlet ηT→S is approximately equal to 1 because both nonradiative and radiative decay rates of the triplet state are much smaller than the reverse ISC rate KT→S ISC . Therefore, IDF/ IPF is estimated to be 6.46. This estimation is roughly consistent with the estimation based on TCSPC measurement showing a ratio equal to 5.75 at 600 nm and 10.76 at 690 nm. The standard method of analysis of thermally activated DF and determination of the energy gap between the triplet and singlet state ΔEST is to measure the steady-state intensity ratio IDF/IPF as a function of temperature.37 Alternatively, we can analyze the DF by temperature-dependent dynamics. In the delayed time scale, e.g., >50 ns, the ISC from singlet to triplet and PF are already completed. The population of the triplet state, nT, can be described as dn T T→S = −KPn T − KISC nT dt

Figure 4. The decay rates of fluorescence as a function of temperature and the fit based on eq 4.

according to eq 4. The energy gap between the singlet and triplet states EST is determined as 30.6 ± 1.9 meV. It is the very small energy gap that results in efficient ISC and reverse ISC and thus the efficient thermally activated DF.41 The similar degeneracy of the singlet and triplet states in Au nanoclusters/nanoparticles was also studied theoretically42 and experimentally.43 Zheng et al. observed the spectral overlapped nanosecond and microsecond decayed components in glutathione-protected ∼2 nm Au nanoparticles. They interpreted that the singlet and triplet transitions are responsible for the fast and slow components, respectively. Differently, in their system the 400 nm excitation results in the triplet transition while the 520 nm excitation leads to the singlet transition, which may suggest that an effective intersystem crossing occurs only in a high energy region and then the excited carriers relax into the lower excited triplet state. The difference most likely arises from the different structures between glutathione and BSA. The lifetime of fluorescence was measured as a function of excitation fluence using the TCSPC technique to identify thermal activation as the mechanism of the DF. Figure 5 shows the lifetimes of the long component measured by PMT as a function of fluence. It is evident that the lifetime approximately remains as a constant with increasing fluence. Moreover, the fluorescence intensity of Au25 is detected as a function of fluence. The peak intensity is linearly proportional to fluence, and the slope is obtained as a = 0.98 ± 0.04 by fit using equation IPL = kIaex, as shown in the inset of Figure 5. This confirms that the DF mechanism is thermal activation. In the DF of triplet−triplet annihilation, lifetimes will vary with fluence, and the intensity of fluorescence is quadratically proportional to fluence because of its biphotonic process.33 On the basis of the above analysis, we propose that the fluorescence of Au25@BSA arises from spectrally overlapped prompt fluorescence and thermally activated delayed fluorescence. The previous study demonstrated that the red band consists of two PL bands, band-I at 710 nm and band-II at 640 nm, arising from two structure-correlated transitions.8 The spectrum of the red fluorescence band can be well fit by two

(2)

where KP is total rate away from the triplet state; KT→S ISC is reverse ISC rate from the triplet to the singlet state. In this case, KP ≪ KT→S ISC ≪ τS (lifetime of singlet state); thus, the lifetime of DF is determined by reverse ISC KT→S ISC . Equation 2 can be simplified as

dn T T→S ≈ −KISC nT dt

(3)

It is known that high values of KT→S ISC are favored by a small energy gap EST because it is approximately given as37,40 T→S T→S KISC = K̅ISC exp( −EST/KBT )

(4)

K̅ T→S ISC

where is the average rate constant for the adiabatic ISC, and KB is Boltzmann’s constant. Therefore, the DF will exhibit a single exponential decay and the decay rate will exponentially increase with increasing temperature. 19035

dx.doi.org/10.1021/jp305902w | J. Phys. Chem. C 2012, 116, 19032−19038

The Journal of Physical Chemistry C

Article

Figure 5. Lifetimes of luminescence of Au25 NCs as a function of excitation fluence measured by a PMT. The inset is the intensity as a function of excitation fluence, and a slope of a = 0.98 ± 0.04 is determined.

Gaussians, and the fit parameters are summarized in Table 2. The red fluorescence of Au25 NCs has been observed with a Table 1. Decay Time Constants of the Fast Component at Various Wavelengths τ1 (ns) τ2 (ns)

565 nm

605 nm

650 nm

690 nm

1.18 ± 0.02 5.27 ± 0.07

1.22 ± 0.03 5.60 ± 0.12

1.20 ± 0.05 6.79 ± 0.29

1.22 ± 0.05 7.88 ± 0.49

Table 2. Fitted Parameters of the Red Band at 300 K Using Two Gaussian Functions band-I band-II

wavelength (nm)

bandwidth (nm)

704.1 ± 0.29 (1.76 eV) 639.1 ± 0.27 (1.94 eV)

255.1 ± 1.0 (253.5 meV) 164.9 ± 2.4 (164.8 meV)

Figure 6. (a) PL excitation spectra with emission at 600 and 700 nm; (b) femtosecond to picosecond transient absorption of Au25 NCs at 535 nm pumped at 400 nm.

Taking into account the similar electronic structure of Au25 NCs,7,28 the two transitions most likely correspond to intraband transitions that were assigned based on Au25(SR)18,7,28 i.e., band-I: LUMO+1 to HOMO and bandII: LUMO+2 to HOMO (6sp, delocalized in core).7 It is worth noting that the LUMO+1 contains three degenerate states and the LUMO+2 is a single degenerate state.7 Band-I at 1.76 eV exhibits bandwidth of 253.5 meV, and band-II at 1.94 eV has bandwidth of 164.8 meV. The larger bandwidth and higher intensity of band-I is due to its 3-fold degeneracy (LUMO+1) and thermal population through reverse ISC. The separation between the two bands of 179 meV reasonably matches the separation between the states of LUMO+1 and LUMO+2 of 210 meV.7 Moreover, the absorption peak of 480−540 nm and the peak of PLE at 2.34 eV match well the calculation value of LUMO+1 to HOMO at 2.34 eV.7 On the other hand, the electronic structure and thus the optical transitions in Au25@BSA have been obviously different from that of Au25(SR)18, such as absorption and fluorescence spectra. The HOMO−LUMO energy gap of 1.37 eV in Au25 predicted by the density function theory (DFT)6,7 is located at the infrared region and is not consistent with the observations, and we believe that it is most likely due to the effect of BSA. Each BSA is constituted by 585 amine acid residues, including 35 cysteines.44,45 Pradeep et al. studied the evolution of Au25 NCs formation in BSA and confirmed that Au25 NCs form through a protein-bound Au(I) intermediate and subsequent emergence of free protein and Au(I)-S staple.46 Likely due to the enhanced conjugation between Au(I) and cysteines in BSA, no sharp peaks are observed in the absorption spectrum of

peak located at 640−750 nm for various ligands, which means that the fluorescence cannot solely be determined by the Au25 NC itself but that the ligands will also evidently influence the fluorescence. To date, it is difficult to identify which transitions exactly correspond to these two bands because the detailed electronic structure of BSA-protected Au25 NCs is unclear and the correlation to the optical transitions is presently not available. It should be emphasized that the absorption of BSAprotected Au25 NCs does not exhibit an absorption peak around 670 nm that was regarded as the fingerprint of Au25(SR)18.5 Figure 6a shows the PL excitation (PLE) spectrum. The first absorption band appears around 2.34 eV and the peak of 400 nm (3.1 eV) should correspond to the higher excited state absorption. A small difference of 10 nm is observed in PLE between 600 and 700 nm, and very similar PL spectra are acquired with excitation at 2.34 and 3.1 eV (not shown here). This implies an effective relaxation from the band of 3.1 eV into the band of 2.34 eV. We performed a femtosecond to picosecond transient absorption pump at 3.1 eV and probe at 2.34 eV, as shown in Figure 6b. A fast rise of 1.2 ps represents the effective relaxation from the band of 3.1 eV into the band of 2.34 eV consistent with the steady-state result and with the other observation by ultrafast techniques.16,31 A slow decay of 350 ps corresponds to the depopulation of the 2.34 eV band, ISC for triplet formation, and prompt fluorescence. The fast decay of 2.0 ps may be due to a relaxation to the lower band or trapping by surface/defect states.16 19036

dx.doi.org/10.1021/jp305902w | J. Phys. Chem. C 2012, 116, 19032−19038

The Journal of Physical Chemistry C



Au25@BSA. In contrast, fine absorption peaks have been observed in Au25(SR)18 and assigned to the electronic transitions.7,29 It should be emphasized that there is no peak around 670 nm; instead, the first peak appears around 2.34 eV in the absorption of Au25@BSA NCs. It is expected that the interaction between Au25 NC and BSA may modify the electronic structure so that the DFT-predicted LUMO+1/ LUMO+2 to HOMO in Au25 (2.34 eV) actually becomes the HOMO−LUMO and is responsible for the red band. The PLE spectrum also reveals a consistent result. The ultrafast transient absorption data exhibit two decay components. The slow component with hundreds of picoseconds lifetime is ascribed to the triplet formation and prompt fluorescence, and the fast component of 2 ps can be due to trapping by the surface/defect states. The Au(I) in the Au(I)-S semirings can form a Au(I) complex and triplet states.47 Then the excited electrons in this band can intersystem cross to the triplet state and finally contribute to the red band. It has been shown that BSA can influence the electronic structure of Au NCs or Au nanoparticles when conjugating with them.45,48 The other possibility is that the interaction between BSA and Au25 can suppress the relaxation from LUMO+1/LUMO+2 to LUMO. However, no evidence supports such a claim. Up to now, the detailed influence of BSA has not yet been fully understood, and further theoretical and experimental investigations are required. It is expected that the triplet state in Au25 nanoclusters is very relevant to the Au(I)-S semirings. It is found that the long lifetime component does not exist in BSA-protected Au8 NCs most likely because there is no Au(I) and Au(I)-S semiring structure. On the other hand, the long lifetime component has been observed in Au25 clusters with other ligands,5,17,22 and Au(I)-S semirings exist in every Au25 NC.7,49 It has been shown that Au(I) has a d10 electron structure, and there exists intramolecular and intermolecular Au(I)−Au(I) interaction for Au(I) complexes. The strong metal-induced spin−orbit coupling leads to efficient singlet−triplet state mixing which eliminates the spin-forbidden nature of the radiative relaxation of the triplet, thus enabling the enhancement of intersystem crossing efficiency.50 The Au(I)−thiolate complex with d10 electron structure can form ligand-localized triplet states and ligand-to-metal charge transfer triplet states.47 The strong goldinduced spin−orbit coupling leads to efficient intersystem crossing.50,51 Further investigation is nevertheless required for detailed understanding for the origin of the triplet state and the electronic structure influence by ligands.

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (X.W.); [email protected]. edu.tw (J.T.). Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS Professor Kenneth P. Ghigginos is gratefully acknowledged for allowing us to use his time-gated fluorescence and transient absorption experiments. Financial support from Academia Sinica (AS) Nano Program and National Science Council (NSC) of Taiwan under the program 99-2221-E-001-002-MY3 and No.99-2113-M-001-023-MY3 is also acknowledged.



REFERENCES

(1) Zheng, J.; Nicovich, P. R.; Dickson, R. M. Annu. Rev. Phys. Chem. 2007, 58, 409−431. (2) Jin, R. Nanoscale 2010, 2, 343−362. (3) Parker, J. F.; Fields-Zinna, C. A.; Murray, R. W. Acc. Chem. Res. 2010, 43, 1289−1296. (4) Nishida, N.; Shibu, E. S.; Yao, H.; Oonishi, T.; Kimura, K.; Pradeep, T. Adv. Mater. 2008, 20, 4719−4723. (5) Wu, Z.; Jin, R. Nano Lett. 2010, 10, 2568−2573. (6) Aikens, C. M. J. Phys. Chem. C 2008, 112, 19797−19800. (7) Zhu, M.; Aikens, C. M.; Hollander, F. J.; Schatz, G. C.; Jin, R. J. Am. Chem. Soc. 2008, 130, 5883−5885. (8) Wen, X.; Yu, P.; Toh, Y.-R.; Tang, J. J. Phys. Chem. C 2012, 116, 11830−11836. (9) Xie, J.; Zheng, Y.; Ying, J. Y. J. Am. Chem. Soc. 2009, 131, 888− 889. (10) Parker, J. F.; Weaver, J. E. F.; McCallum, F.; Fields-Zinna, C. A.; Murray, R. W. Langmuir 2010, 26, 13650−13654. (11) Li, S.; Shaojun, D.; Nienhaus, G. U. Nano Today 2011, 6, 401− 418. (12) Bigioni, T.; Whetten, R.; Dag, Ö . J. Phys. Chem. B 2000, 104, 6983−6986. (13) Wilcoxon, J.; Martin, J.; Parsapour, F.; Wiedenman, B.; Kelley, D. J. Chem. Phys. 1998, 108, 9137. (14) Huang, T.; Murray, R. W. J. Phys. Chem. B 2001, 105, 12498− 12502. (15) Miller, S. A.; Womick, J. M.; Parker, J. F.; Murray, R. W.; Moran, A. M. J. Phys. Chem. C 2009, 113, 9440−9444. (16) Devadas, M. S.; Kim, J.; Sinn, E.; Lee, D.; Goodson, T., III; Ramakrishna, G. J. Phys. Chem. C 2010, 114, 22417−22423. (17) Link, S.; Beeby, A.; FitzGerald, S.; El-Sayed, M. A.; Schaaff, T. G.; Whetten, R. L. J. Phys. Chem. B 2002, 106, 3410−3415. (18) Shibu, E. S.; Muhammed, M. A. H.; Tsukuda, T.; Pradeep, T. J. Phys. Chem. C 2008, 112, 12168−12176. (19) van Wijngaarden, J. T.; Toikkanen, O.; Liljeroth, P.; Quinn, B. M.; Meijerink, A. J. Phys. Chem. C 2010, 114, 16025−16028. (20) Yu, P.; Wen, X.; Toh, Y. R.; Tang, J. J. Phys. Chem. C 2012, 116, 6567−6571. (21) Zheng, J.; Zhang, C. W.; Dickson, R. M. Phys. Rev. Lett. 2004, 93, 077402. (22) Sakanaga, I.; Inada, M.; Saitoh, T.; Kawasaki, H.; Iwasaki, Y.; Yamada, T.; Umezu, I.; Sugimura, A. Appl. Phys. Expr. 2011, 4, 95001. (23) Retnakumari, A.; Setua, S.; Menon, D.; Ravindran, P.; Muhammed, H.; Pradeep, T.; Nair, S.; Koyakutty, M. Nanotechnology 2010, 21, 005103. (24) Kawasaki, H.; Hamaguchi, K.; Osaka, I.; Arakawa, R. Adv. Funct. Mater. 2011, 21, 3508−3515. (25) Shang, L.; Brandholt, S.; Stockmar, F.; Trouillet, V.; Bruns, M.; Nienhaus, G. U. Small 2012, 5, 661−665. (26) Wu, Z.; Jin, R. ACS Nano 2009, 3, 2036−2042. (27) Heaven, M. W.; Dass, A.; White, P. S.; Holt, K. M.; Murray, R. W. J. Am. Chem. Soc. 2008, 130, 3754−3755.

4. CONCLUSIONS We have elucidated the fluorescence mechanism in BSAprotected Au25 NCs. The fluorescence of Au25 NCs consists of both fast and slow components in nanosecond and microsecond time scales; which are attributed to PF and thermally activated DF, respectively. We determined a small energy gap of 30.6 meV between the singlet and the triplet states from the temperature-dependent time-resolved fluorescence measurements. The major finding of this work is that the unusual efficient intersystem crossing occurs as a sequence of such a small energy gap, resulting in the efficient delayed fluorescence as observed from this study. The conjugation of BSA can influence the electronic structure of Au25 NCs and the dynamics in the fluorescence. The triplet states are very relevant to the Au(I)-S semiring and Au(I) complex. 19037

dx.doi.org/10.1021/jp305902w | J. Phys. Chem. C 2012, 116, 19032−19038

The Journal of Physical Chemistry C

Article

(28) Aikens, C. M. J. Phys. Chem. Lett. 2010, 2, 99−104. (29) Devadas, M. S.; Bairu, S.; Qian, H. F.; Sinn, E.; Jin, R. C.; Ramakrishna, G. J. Phys. Chem. Lett. 2011, 2, 2752−2758. (30) Wen, X.; Van Dao, L.; Hannaford, P. J. Phys. D: Appl. Phys. 2007, 40, 3573−3578. (31) Qian, H.; Sfeir, M. Y.; Jin, R. J. Phys. Chem. C 2010, 114, 19935−19940. (32) Muhammed, M. A. H.; Shaw, A. K.; Pal, S. K.; Pradeep, T. J. Phys. Chem. C 2008, 112, 14324−14330. (33) Baldo, M. A.; Adachi, C.; Forrest, S. R. Phys. Rev. B 2000, 62, 10967. (34) Romanovskii, Y. V.; Gerhard, A.; Schweitzer, B.; Scherf, U.; Personov, R. I.; Bassler, H. Phys. Rev. Lett. 2000, 84, 1027−1030. (35) Pope, M.; Swenberg, C. E. Electronic processes in organic crystals and polymers; Oxford University Press: Oxford, 1999. (36) Fan, M.; Yao, J.; Tong, Z. Molecule Photochemistry and Photofunction Materials Science; Science Press: Beijing, 2009. (37) Berberan-Santos, M. N.; Garcia, J. M. M. J. Am. Chem. Soc. 1996, 118, 9391−9394. (38) Deaton, J. C.; Switalski, S. C.; Kondakov, D. Y.; Young, R. H.; Pawlik, T. D.; Giesen, D. J.; Harkins, S. B.; Miller, A. J. M.; Mickenberg, S. F.; Peters, J. C. J. Am. Chem. Soc. 2010, 132, 9499− 9508. (39) Baleizão, C.; Berberan-Santos, M. N. J. Chem. Phys. 2007, 126, 204510. (40) Parker, C. A. Photoluminescence of Solutions; Elsevier: Amsterdam, 1968. (41) Goushi, K.; Yoshida, K.; Sato, K.; Adachi, C. Nat. Photonics 2012, 6, 253−258. (42) Magyar, R.; Mujica, V.; Marquez, M.; Gonzalez, C. Phys. Rev. B 2007, 75, 144421. (43) Zhou, C.; Sun, C.; Yu, M. X.; Qin, Y. P.; Wang, J. G.; Kim, M.; Zheng, J. J. Phys. Chem. C 2010, 114, 7727−7732. (44) Barbosa, L. R. S.; Ortore, M. G.; Spinozzi, F.; Mariani, P.; Bernstorff, S.; Itri, R. Biophys. J. 2010, 98, 147−157. (45) Tsai, D. H.; DelRio, F. W.; Keene, A. M.; Tyner, K. M.; MacCuspie, R. I.; Cho, T. J.; Zachariah, M. R.; Hackley, V. A. Langmuir 2011, 27, 2464−2477. (46) Chaudhari, K.; Lourdu Xavier, P.; Pradeep, T. ACS Nano 2011, 5, 8816−8827. (47) Wang, X.; Del Guerzo, A.; Schmehl, R. H. J. Photochem. Photobiol. C 2004, 5, 55−77. (48) Khullar, P.; Singh, V.; Mahal, A.; Dave, P. N.; Thakur, S.; Kaur, G.; Singh, J.; Kamboj, S. S.; Bakshi, M. S. J. Phys. Chem. C 2012, 116, 8834−8843. (49) Aikens, C. M. J. Phys. Chem. Lett. 2010, 1, 2594−2599. (50) Zhao, Q.; Li, F.; Huang, C. Chem. Soc. Rev. 2010, 39, 3007− 3030. (51) Forward, J. M.; Bohmann, D.; Fackler, J. P., Jr.; Staples, R. J. Inorg. Chem. 1995, 34, 6330−6336.

19038

dx.doi.org/10.1021/jp305902w | J. Phys. Chem. C 2012, 116, 19032−19038