Nanoscale View Article Online
Published on 25 November 2013. Downloaded by University of Stellenbosch on 06/10/2014 11:01:56.
PAPER
Cite this: Nanoscale, 2014, 6, 1716
View Journal | View Issue
Near-infrared room temperature luminescence of few-atom Au aggregates in silica: a path for the energy-transfer to Er3+ ions Tiziana Cesca,*a Boris Kalinic,a Chiara Maurizio,a Carlo Scian,a Giancarlo Battaglin,b Paolo Mazzoldia and Giovanni Mattei*a Ultra-small molecule-like AuN nanoclusters made by a number of atoms N less than 30 were produced by ion implantation in silica substrates. Their room temperature photoluminescence properties in the visible and near-infrared range have been investigated and correlated with the Er sensitization effects observed in Er–Au co-implanted samples. The intense photoluminescence emission under 488 nm laser excitation occurs in three different spectral regions around 750 nm (band A), 980 nm (band B) and 1150 nm (band C) as a consequence of the formation of discrete energy levels in the electronic structure of the molecule-like AuN nanoclusters. Indeed, energy maxima of bands A and C scale with N1/3 as expected for quantum confined systems. Conversely, the energy maximum of band B appears to be almost independent of size, suggesting a contribution of electronic surface states. A clear correlation between the formation of band B in the samples and Er-related photoemission is demonstrated: the band at
Received 5th August 2013 Accepted 22nd November 2013
980 nm related to AuN nanoclusters resonant with the corresponding Er3+ absorption level, is suggested as an effective de-excitation channel through which the Au-related photon energy may be transferred
DOI: 10.1039/c3nr04108e
from Au nanoclusters to Er ions (either directly or mediated by photon absorption), eventually producing
www.rsc.org/nanoscale
the Er-related infrared emission at 1540 nm.
1
Introduction
Owing to their plasmonic properties as efficient scatterers or absorbers of light in the visible and near-infrared (NIR) range, noble metal nanoparticles (NPs) have been widely investigated in recent years for a variety of applications in different elds such as sensing,1–3 photovoltaics,4–6 catalysis,7–9 uorescence enhancement,10 nonlinear optics11–13 and medical therapies.14–16 All these applications are driven by the capability of exciting localized surface plasmon resonances (LSPR) in the metallic nanoparticles and to produce intense local-eld enhancements that can be further increased and tailored by controlling nanoparticle shapes and composition, and the interaction among the plasmonic nanostructures.17 On the other hand, novel and peculiar properties emerge when the dimensions of the nanoparticles are reduced down to sub-nanometer sizes, that is, in a range in which the clusters are too small to have the continuous density of states necessary for plasmon excitations to be supported.18–21 In particular, preplasmonic noble metal NPs made by 100 down to 5 atoms are of special interest when dealing with luminescence properties as a
Physics and Astronomy Department and CNISM, University of Padova, via Marzolo 8, I-35131 Padova, Italy. E-mail:
[email protected];
[email protected] b
Department of Molecular Sciences and Nanosystems, Ca' Foscari University of Venice, Dorsoduro 2137, I-30123 Venice, Italy
1716 | Nanoscale, 2014, 6, 1716–1724
they encompass the transition between bulk and molecular regimes.22,23 For such ultra-small, molecule-like nanoclusters the electronic conguration is characterized by discrete energy levels with an energy gap between the highest occupied molecular orbital and the lowest unoccupied molecular orbital (HOMO–LUMO gap) that exceeds the phonon energy even at room temperature. Therefore, efficient radiative relaxation can be in principle achieved under direct HOMO–LUMO excitation or under interband photo-stimulation.24–26 The production of highly uorescent molecular noble metal NPs with stable sizetunable emission properties in the visible–NIR range was demonstrated in these years18,23–30 opening interesting possibilities for the development of nanophotonic devices in many different elds.31 Recently, our group demonstrated another very interesting optical functionality in nanophotonics from molecule-like Au NPs obtained by ion implantation in silica. They act indeed as very efficient nanoantennae for the luminescent enhancement of photoemitting systems like Er ions.32,33 This sensitization effect was demonstrated to be a broad-band mechanism and works for both in-resonance (with the 4I15/2 / 4F7/2 Er absorption transition at 488 nm) and out-of-resonance (at 476.5 nm) Er excitation conditions with the same efficiency. This has important technological entailments in optoelectronics where many research efforts are being made in these years to increase the small cross-section for Er excitation (typically of the order of
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 25 November 2013. Downloaded by University of Stellenbosch on 06/10/2014 11:01:56.
Paper
1021–1019 cm2, depending on the matrix) and enhance the Er3+ luminescent emission at 1.54 mm.34 The use of ion implantation as the incorporation technique for the dopant species is of fundamental importance for this purpose since it provides a well-controlled way to produce local super-saturation in the silica matrix. This in turn allows for efficient control of the early stages of nucleation and growth of the implanted species by working very close to the critical size of nucleation (i.e., 2–3 Au atoms in silica) inducing the formation of ultrasmall Au nanoclusters. Furthermore due to the pre-plasmonic character of such ultra-small Au aggregates, the sensitization process was demonstrated to be due to a transfer of electromagnetic energy between the Au clusters and the rare-earth ions, and not to a plasmonic eld enhancement effect.35 In spite of all the obtained results, nonetheless, up to now the physical mechanisms controlling the Er sensitization process by molecule-like Au NPs are still to be claried. On the other hand, several data indicate that the photophysical properties of such ultra-small Au nanoclusters are of paramount importance to understand the basic processes triggering the energy-transfer (ET) to the Er ions. For this reason, in this work we tried to address the fundamental question still le open by our previous work32 about the physical mechanism through which ultra-small Au nanoclusters are able to “talk” to nearby Er3+ ions. To this aim, we performed a detailed investigation of the room temperature (RT) photoluminescence properties in the NIR range of molecule-like Au nanoclusters, studying in particular the correlation of the Au-related luminescence features with the energy-transfer process to Er ions in Er–Au coimplanted silica. Besides this goal of denite technological relevance this study allowed us address a problem of a more fundamental character, that is, the luminescence properties of ultra-small Au aggregates. Even if this aspect has been widely investigated up to now most of the works are devoted to uorescent ultra-small noble metal nanoclusters (gold in particular) produced by chemical synthesis in aqueous solutions with various capping layers, used to improve their stability or to obtain bio- or eco-compatibility, which in turn may affect the optical properties of the nanoclusters, e.g., through the formation of covalent Au–S bonds at the surface for thiol-based capping agents.18 The physical technique used in this work to synthesize the ultra-small Au nanoclusters (ion implantation) aimed instead to produce nominally non-capped NPs, whose properties can be, in principle, better correlated with their intrinsic structure, allowing decoupling of the intrinsic size effect from the extrinsic ligand effect.36
2 Experimental section Two classes of samples have been produced by performing sequential Au implantations on undoped and Er-doped silica slabs (Herasil I by Haereus). A triple energy scheme (60, 110, 190 keV) was used for the Au implantations and total Au uences in the range 3.9 1014 to 7.2 1015 Au+ cm2 were implanted, as estimated by Rutherford Backscattering Spectrometry (RBS). The Er-doped silica substrates were obtained by sequential Er implantations at three energies (50, 100 and 190 keV) with an Er
This journal is © The Royal Society of Chemistry 2014
Nanoscale
total uence of 7.2 1014 cm2. Before performing Au implantations the Er-implanted slabs were annealed for 1 hour at 800 C in a N2 atmosphere to completely recover the implantation damage. Both the Au and the Er implantation schemes have been chosen in order to obtain almost at concentration proles of the implanted elements (about 70 nm thick) and their full overlap in the coimplanted samples. Aer Au implantations the samples were heated for 1 hour at temperatures in the range 300–800 C in the N2 atmosphere to induce Au clustering and damage recovery of the silica matrix. In the following, the Er–Au co-implanted samples will be labeled depending on the uence ratio [Au]/[Er] ¼ X as Er1AuX. Correspondingly, the Au-implanted ones will be named as AuX. The reference Er-implanted sample (without gold) will be indicated as Er800. Photoluminescence (PL) measurements were performed at room temperature using the 488 nm line of a mechanically chopped (6 Hz) cw Ar laser as an excitation source and detecting the spectral emission in the range 600–1700 nm by a N2-cooled photomultiplier tube (Hamamatsu R5509-72). A single-grating monochromator provided the spectral discrimination of the luminescence signal and the spectra were recorded with a lockin amplier using the chopper frequency as a reference. Timeresolved PL analyses were carried out by xing the detected wavelength and collecting the PL intensity evolution as a function of the time with a transient digitizer (Tektronix TDS7104). Structural characterizations have been performed by Extended X-ray Absorption Fine Structure (EXAFS) measurements at the Au L3-edge in uorescence mode at the Italian beamline GILDA of the European Synchrotron Radiation Facility (Grenoble, France). The Au uorescence signal was collected by a 13-element high-purity Ge detector. The samples were cooled down at 80 K to avoid radiation damage and to reduce as much as possible thermal vibrations. Details on the measurement set-up as well as on the EXAFS analyses procedure can be found in ref. 32.
3 Results and discussion The room temperature PL spectra of silica samples implanted with Au at a uence of 7.2 1014 Au+ cm2 (Au1 series) and annealed in the N2 atmosphere at different temperatures in the range 300–800 C are reported in Fig. 1a. The samples were excited by cw pumping at 488 nm (i.e., resonantly with the 4I15/2 / 4F7/2 Er absorption transition). Measurements were performed also with excitation at 476.5 nm (corresponding to outof-resonance excitation conditions for the Er ions) obtaining analogous results. Nonetheless due to the much lower power of the laser available at this wavelength the spectra had a lower signal-to-noise ratio, so in the following the results obtained with 488 nm pumping are reported and discussed. The spectra are vertically shied by an arbitrary amount to improve their visualization. The samples exhibit different spectral features whose relative intensity depends strongly on the temperature of the annealing process. In particular, three main bands are evident, indicated as A, B and C. For thermal treatments at low temperatures (300 and 400 C, blue and green curves) the photoluminescent emission is mainly characterized by the
Nanoscale, 2014, 6, 1716–1724 | 1717
View Article Online
Published on 25 November 2013. Downloaded by University of Stellenbosch on 06/10/2014 11:01:56.
Nanoscale
PL spectra of (a) Au-implanted (Au1 series) and (b) Er–Au coimplanted (Er1Au1 series) silica samples annealed for 1 hour in the N2 atmosphere at the different indicated temperatures. The measurements have been performed with cw excitation at 488 nm and a pump power of 22 mW. In panel (a) the spectra are vertically shifted for clarity and the dashed horizontal lines indicate the corresponding reference level. The arrows in panel (b) indicate that different vertical scales have to be considered for the two spectral regions corresponding to wavelengths shorter (left-hand scale) and longer (right-hand scale) than 1400 nm. Bands A, B and C are related to the presence of Au, whereas band D is the Er3+ emission. Fig. 1
presence of a visible band centered around 750 nm (band A), and a second band in the near-infrared region, at 1150 nm (band C). At higher annealing temperatures a third band (band B) peaked at about 980 nm starts to be visible, being the dominant feature for annealing at 600 C. The further increase of the temperature of the thermal process up to 800 C induces quenching of the luminescent emission at wavelengths shorter than 1000 nm; only the band at 1150 nm (band C) survives, spectrally broadened and damped in intensity. A markedly different behavior as a function of the annealing temperature is observed in the spectral emission of Auimplanted samples if gold implantation is performed on silica substrates previously implanted with Er. Fig. 1b shows the PL spectra of Er–Au co-implanted samples annealed in the N2 atmosphere at the different indicated temperatures. The Au implantation conditions are the same as for samples of the Au1 series, but in this case the silica substrate was previously implanted with Er with a total uence of 7.2 1014 cm2 and annealed in the N2 atmosphere at 800 C (Er1Au1 series). The PL spectrum of the Er800 reference sample is also reported (black
1718 | Nanoscale, 2014, 6, 1716–1724
Paper
line). In order to make all the spectral features clearly visible different vertical scales are used for the wavelength region 600– 1400 nm (le-hand scale) and 1400–1700 nm (right-hand scale), as indicated by the arrows. As in Fig. 1a, the spectra are vertically shied to improve the visualization. However, a quantitative comparison among the relative intensities of the peaks in the different samples is still possible. In particular it is interesting to compare the intensity of the Er luminescent emission at 1540 nm (band D). This peak is present in the Er800 reference sample as a consequence of the resonant excitation of the Er ions at 488 nm through the 4I15/2 / 4F7/2 Er absorption transition. Nonetheless, in the Er1Au1 samples the intensity of this band depends on the annealing temperature: it becomes visible at T > 500 C and reaches a maximum in the sample annealed at 720 C (purple line). In this case the PL signal at 1540 nm is increased by more than a factor of 7 with respect to the Er800 reference sample. As demonstrated in ref. [32 and 33] this behavior is related to the presence in the Er–Au co-implanted systems of ultra-small Au nanoclusters that act as sensitizers for the Er emission, strongly enhancing the Er3+ luminescence through an energy-transfer process from the Au nanoclusters to the Er ions. It was also demonstrated that such a sensitization effect is mediated by the broad-band interband absorption of the sub-nanometer Au clusters, therefore working with the same efficiency (taking into account also the differences in the absorption cross-section) both for resonant Er excitation and for out-of-resonance conditions. Furthermore, the sensitization effect was strongly dependent on the size of the Au nanoaggregates: molecule-like Au aggregates are very efficient sensitizers, but if their size is increased their effect can be detrimental inducing quenching of the Er emission. This may be related to the progressive development of the metallic (plasmonic) character of the Au nanoclusters, which can promote the quenching of the luminescence of nearby emitters (see for instance ref. 10). In our systems, the transition from molecular to metallic character of the Au nanoclusters typically occurs for annealing temperatures above 600–700 C thus explaining the further decrease of the Er PL emission in the sample annealed at 800 C. Further interesting observations arise comparing the luminescent behavior of the Er1Au1 samples (Fig. 1b) with the one of the samples of the Au1 series (Fig. 1a). To provide the direct quantitative comparison of the PL spectra the same vertical scale as in Fig. 1a has been used in Fig. 1b for the le-hand scale. The spectral emission in the range 600–1400 nm is still characterized by the presence of the two bands in the visible and NIR regions (bands A and C), while the central band (band B) is now only faintly visible at low annealing temperatures. By increasing the temperature of the thermal process all the features become less evident and at T > 600 C the band B is completely quenched. Furthermore it is worth noting that in spite of the fact that the same acquisition and visualization parameters have been used for all the spectra reported in Fig. 1, no luminescence feature is visible in the 600– 1400 nm spectral range in the Er800 reference sample. In particular, we were not able to detect from this sample any appreciable signal related to the intrinsic Er emission at 980 nm.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 25 November 2013. Downloaded by University of Stellenbosch on 06/10/2014 11:01:56.
Paper
To better highlight the differences in the observed spectral emission due to the co-presence of Er in the silica matrix, we have plotted in Fig. 2a the direct comparison of the spectra of the Au1 and Er1Au1 samples annealed at 600 C in the N2 atmosphere and the spectrum of the Au1 sample annealed at 600 C in the Ar + H2 atmosphere. The graph clearly shows how the presence of Er in the silica matrix largely suppresses the emission of the 980 nm component (band B) which is instead strongly visible in the samples not containing Er. In Fig. 2b and 2c we have plotted the PL intensity measured at 980 nm and 1540 nm as a function of the annealing temperature for the samples of the two series Au1 and Er1Au1. The data are taken from the spectra in Fig. 1 evaluating the total luminescent emission at the two wavelengths. To better visualize the trends, different vertical scales are used in Fig. 2c, as indicated by the arrows: the intensity of the luminescent emission at 980 nm has to be referred to the le-hand scale in Fig. 2b, whereas for the 1540 nm emission the right-hand scale has to be considered. For the samples of the Au1 series (Fig. 2b) the trend of the luminescence intensity measured at 980 nm is characterized by a pronounced peak for temperatures around 600–700 C. Moreover, the temperature evolution of this emission is consistent with the trend exhibited by the Er-related luminescence (at 1540 nm) detected in the Er1Au1 samples (Fig. 2c). In these samples, in contrast, the emission at 980 nm is highly quenched and its trend is characterized by a monotonic decrease in intensity by increasing the annealing temperature. These results provide a denite indication of a strong interplay between the 980 nm luminescence and the Er emission at 1540 nm. Further evidence of this correlation is provided by the results of PL measurements performed on samples with a higher content of gold. To conrm this, in Fig. 3a we have reported the PL spectra of the samples of the Er1Au5 series annealed at different temperatures. A total Au uence of 3.7
Nanoscale
1015 Au+ cm2, that is ve times higher than in the Er1Au1 samples, was implanted in this case. As in Fig. 1b, different vertical scales are used to plot the spectra in the wavelength ranges 600–1400 nm (le-hand scale) and 1400–1700 nm (righthand scale). Fig. 3b shows the comparison of the trend as a function of the annealing temperature of the luminescence intensity detected at 980 nm and the Er-related emission measured at 1540 nm. Owing to the higher Au content in these samples, the differences in the temperature evolution of the two bands, already observed in the samples of the Er1Au1 series (Fig. 2c), are now amplied and an anti-correlated trend between the two emission features clearly emerges for annealing temperatures below 600 C. Time-resolved PL measurements were performed by monitoring the luminescence decay with time at different wavelengths corresponding to the maximum of each band. For band D (related to Er emission), as a function of the annealing temperature, lifetime values (s) of the order of few ms up to 11 ms (in the Er800 reference sample) were found,37 indicating a progressive recovery of the full octahedral conguration of oxygen atoms around the Er3+ ions. Nonetheless, in the wavelength region 600–1400 nm, the used experimental set-up allowed us to measure only lifetimes longer than a threshold of 500 ms. For this reason it was not possible to obtain a reliable estimation of s values for the bands in this region, but the results indicate that in all the measured samples, irrespective of the band (A, B or C), the lifetimes are shorter than the experimental threshold. On the other hand, ultra-small capped AuN NPs (with N ¼ 5–31 atoms) were reported to have PL lifetimes in the ns regime,38 i.e., well below our experimental threshold. Therefore, to get insights on the physical origin of the observed RT luminescence and to rule out the possibility that the detected features could be related to the formation of damageinduced luminescent centers due to Er and/or Au implantation,
Fig. 2 Left (a): comparison among the PL spectra of the Au1 and Er1Au1 samples annealed in the N2 atmosphere at 600 C and the spectrum of the Au1 sample annealed at 600 C in Ar + H2 mixture. Right: (b) PL intensity measured at 980 nm as a function of the annealing temperature in the Au1 samples annealed in the N2 atmosphere. (c) PL emission at 980 nm (left-hand scale) and 1540 nm (right-hand scale) versus annealing temperature of the Er1Au1 samples.
This journal is © The Royal Society of Chemistry 2014
Nanoscale, 2014, 6, 1716–1724 | 1719
View Article Online
Published on 25 November 2013. Downloaded by University of Stellenbosch on 06/10/2014 11:01:56.
Nanoscale
Paper
Fig. 4 PL spectra of Er–Xe co-implanted silica samples annealed in the N2 atmosphere at the different indicated temperatures. The spectra of the Er800 reference sample and the Er1Au5 sample annealed at 350 C are reported for comparison.
Fig. 3 (a) PL spectra of Er–Au co-implanted silica samples (Er1Au5 series) annealed in the N2 atmosphere at the different indicated temperatures. The measurements have been performed with cw excitation at 488 nm with a pump power of 22 mW. The spectra have to be referred to the left-hand scale for the emission at wavelengths in the range 600–1400 nm and to the right-hand scale for the range 1400–1700 nm. (b) PL intensity measured at 980 nm and 1540 nm as a function of the annealing temperature.
we have performed analogous PL measurements on Er–Xe coimplanted samples. The idea was motivated by a recent study performed by our group on the effects of implantation damage on the Er3+ luminescence in silica (see ref. 37). In that work, Xe was chosen as a co-dopant of Er to release to the silica matrix an implantation damage as similar as possible to that induced by Au implantation and Er–Xe co-implantations were performed on silica substrates following the same approach (in terms of ion implantations and thermal treatments) used to produce the Er–Au co-implanted samples. A structural and morphological similarity between Er–Xe and Er–Au co-implanted samples was denitely demonstrated by transmission electron microscopy (TEM) characterizations that revealed upon the same annealing treatments the formation of Xe bubbles analogous in size to the Au nanoclusters.37 From the point of view of the luminescence properties investigated in the present work, on the other hand, the Er–Xe co-implanted samples exhibit a markedly different behavior with respect to analogous Er–Au co-implanted samples. In Fig. 4 we have reported the room temperature PL spectra of Er–Xe co-implanted samples annealed in the N2
1720 | Nanoscale, 2014, 6, 1716–1724
atmosphere at temperatures in the range 300–800 C. In this case Xe was implanted at a uence of 7.8 1015 Xe+/cm2, corresponding to a [Xe]/[Er] ratio of about 12. The PL spectra of the Er800 reference sample and the Er1Au5 sample annealed at 350 C are also reported for comparison. As for the Er800 reference sample, the PL spectra of the Er–Xe co-implanted samples exhibit no luminescent feature in the whole wavelength range 600–1700 nm but the Er-related emission peak at 1540 nm, whose intensity progressively increases by increasing the annealing temperature. This is consistent with a progressive recovery of implant-induced defects in the silica matrix and the restoration of the full octahedral coordination of oxygen atoms around the Er3+ ions, as demonstrated in ref. 37. Indeed, differently from the Au nanoclusters, Xe clusters (bubbles) do not act as sensitizers for the Er emission, and since they do not have metallic character they do not produce the luminescence quenching of the plasmonic Au nanoclusters. Therefore, from the point of view of the Er luminescence, pure Er and Er–Xe systems exhibit a similar temperature evolution, since the Er emission occurs due to the resonant excitation at 488 nm and progressively increases by increasing the annealing temperature as a consequence of the progressive recovery of damage induced by the Xe implantation on the silica matrix and the restoration of the Er–O octahedral coordination. Therefore, the formation of damage-related luminescent centers upon either Er implantation alone or subsequent Xe or Au implantation can be denitely ruled out. Furthermore, there is no evidence of the Er-related luminescent emission at 980 nm from both the pure Er-implanted and the Er–Xe co-implanted samples at all the annealing temperatures. Such luminescence, if present, is negligible with respect to the signal exhibited by the Au-containing samples. As a consequence, the spectral features visible in the wavelength range 600–1400 nm in the PL spectra of the
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 25 November 2013. Downloaded by University of Stellenbosch on 06/10/2014 11:01:56.
Paper
Nanoscale
Er–Au co-implanted samples (as for example the Er1Au5 sample in Fig. 4) can be ascribed unequivocally to the presence of gold. Structural information about the gold-related luminescent centers has been obtained by EXAFS characterizations of Er1Au5 samples. The EXAFS technique has been employed for these characterizations for its unique capability of revealing the presence of clusters made of just few atoms (not visible by transmission electron microscopy), thus allowing us to follow the early stages of Au nucleation and cluster growth upon thermal treatments. Details of the measurement conditions and the analysis procedure can be found in ref. 32. The results, in terms of number of Au atoms per cluster (N) and cluster diameter (D) as a function of the annealing temperature, are reported in Table 1. The cluster diameter has been determined from the number N of Au atoms per cluster according to the pffiffiffiffi relation dened in ref. 32: D ¼ 2Ra 3 N , where Ra ¼ 0.1505 nm represents the average radius of an Au atom in the cluster. For the different annealing conditions explored, the EXAFS results reveal the formation of ultra-small Au aggregates with a mean diameter smaller than about 1 nm (measured on the sample Er1Au5 annealed at 800 C), corresponding to Au aggregates made of less than 30 atoms. The gold-related luminescence bands detected from the samples investigated in this work therefore originated from the ultra-small, molecule-like, nature of the AuN aggregates formed upon the thermal treatments. Indeed, molecule-like nanoclusters are characterized by discrete electronic states and the opening of a HOMO–LUMO gap in the electronic conguration that increases with decrease in the size of the nanoclusters. HOMO–LUMO gaps of 0.5 eV, 0.9 eV and 1.3 eV have been reported for Au clusters of 145, 38 and 28 atoms, respectively.24,25,39 In this size regime therefore the energy gap greatly exceeds the Au bulk phonon energy (which is of the order of 10 meV) making radiative transitions competitive with phonon-mediated nonradiative recombination processes, so strongly enhancing the luminescence quantum yield. As an example, a quantum yield of 4.4 1.5 105 has been estimated for the NIR luminescence of 1.7 nm gold nanocrystals,24 that is more than 5 orders of magnitude greater than that of bulk gold (of about 101).40 Moreover, since the quantum size effects are of fundamental importance in controlling the radiative properties of ultra-small aggregates, quenching of the luminescence is expected for clusters of increasing size.41 To highlight the quantum size effects in the luminescent emission detected from the samples investigated
in this work we have studied the correlation between the emission energy of the three main bands identied in the spectra (A, B and C) and the average size of the Au clusters formed in the samples. To conrm this we have performed multiple Gaussian peak deconvolution of the spectra in the energy domain. As an example in Fig. 5a we have plotted the result of this tting procedure for the Er1Au5 sample annealed at 350 C. The Gaussian peaks corresponding to the three main bands obtained from the t are also reported. In Fig. 5b the emission energy at the maximum of each band, extracted from the ts, is shown as a function of the number of Au atoms per cluster (N). The data have been obtained by the analysis of the PL spectra of the Er1Au5 samples annealed at the different temperatures (Fig. 3) that have the best signal-to-noise ratio to provide stable results of the tting procedure. It is worth underlining that the same behavior was obtained also for the purely Au-implanted samples, thus conrming that the trends observed in the Er–Au co-implanted samples represent indeed an intrinsic property of the Au nanoclusters. The values of N are
Table 1 EXAFS results on the AuN cluster size evolution as a function of annealing temperature for Er1Au5 samples: N is the average number of Au atoms per cluster and D is the corresponding average diameter calculated according to the relation reported in the text
Sample
Annealing T ( C)
N (atoms)
D (nm)
As-imp 435 620 720 800
4 4 16 20 30
0.47 0.12 0.49 0.11 0.75 0.05 0.87 0.04 0.94 0.03
(a) PL intensity versus emission energy of the Er1Au5 sample annealed at 350 C. The solid line is the result of a multiple Gaussian peak fitting. The three Gaussian bands obtained from the fit are also shown. (b) Emission energy corresponding to the maximum of each band as a function of the number of Au atoms per cluster (N) for the samples of the Er1Au5 series, obtained by EXAFS analysis.
Fig. 5
Er1Au5
1 1 6 10 10
This journal is © The Royal Society of Chemistry 2014
Nanoscale, 2014, 6, 1716–1724 | 1721
View Article Online
Published on 25 November 2013. Downloaded by University of Stellenbosch on 06/10/2014 11:01:56.
Nanoscale
extrapolated from the EXAFS results given in Table 1. By tting the data (solid lines in Fig. 5b) we could determine for bands A and C a denite dependence of the maximum emission energy pffiffiffiffi on the inverse of the cluster radius, or equivalently EA;C f1= 3 N . This scaling law is consistent with the relation demonstrated by Zheng and co-workers by applying the spherical jellium model that describes the size evolution of the uorescent emission in the visible range of few-atom gold quantum-dots.23 In our case such energy scaling provides the experimental evidence that bands A and C evolve according to the quantum size effect. Conversely, no dependence on the Au cluster sizes has been revealed for band B: its results peaked at 1.27 eV (977 nm) almost independently of N. Further insights on the physical origin of these bands can be found in the study by Whetten and co-workers reported in ref. 25. In that work the authors proposed a solid-state model to explain the radiative properties in the visible and infrared range of 28-atoms molecule-like Au nanoclusters capped by a monolayer of 16 glutathione molecules (Au28(SG)16). With visible excitation at 500 nm they observed two large bands peaked at 1.55 eV (800 nm) and 1.15 eV (1080 nm). According to their model, the higher energy luminescence band was to be ascribed to the radiative recombination of an electron in a higher excited state of the sp-conduction band with a hole in the lower lying d-band (interband transition), while the lower energy band was due to a relaxed recombination across the HOMO–LUMO gap within the sp-band (intraband transition). The same model could be tentatively invoked in our case to give an assignment to the visible band at 750 nm (band A) and the NIR band at 1150 nm (band C). Such description is consistent with the size-dependent energy scaling that we determined for these two bands considering that it involves intraband and interband electronic transitions of molecule-like Au nanoclusters. A different discussion has to be reserved to band B. No luminescent emission was detected by Whetten et al. in this wavelength region. In contrast Murray and co-workers reported a study on the luminescence properties of monolayer-protected ultra-small gold clusters with very different Au core sizes ranging from Au11 to Au201.42 Independently of core sizes and capping monolayers they showed visible–NIR luminescent emission over the same range of energies, from 700 nm (1.8 eV) to 1300 nm (0.9 eV). In particular, a quite intense luminescent band centered at about 960 nm was detected for the Au11 clusters. According to the authors, the core-size independence of the emission energy observed from their samples (irrespective also of markedly different absorption spectra) suggests that the emission process is not simply controlled by excited state HOMO–LUMO transitions, but involves localized core surface states with size-independent energetics. The band B at 1.27 eV (977 nm) detected in our case could be interpreted in this framework as related to the formation of electronic surface states at the Au clusters. On the other hand, the evolution of the PL emission intensity at 980 nm as a function of the annealing temperature for the Au-implanted samples (see Fig. 2b) indicates that the Au clusters have to grow to a denite size for such states to appear. This also suggests that band B cannot be related to dispersed Au atoms chemically bonded to oxygen in
1722 | Nanoscale, 2014, 6, 1716–1724
Paper
the silica network which would be found mostly for low annealing temperatures. From a structural point of view the physical origin of this band could be instead related to the local coordination with oxygen of the gold atoms in the outer shell of the Au clusters. The role of Au cluster surface states in the band B luminescence is further conrmed by the effect of the annealing atmosphere on the luminescent emission at 980 nm. As shown in Fig. 2a, the thermal treatment in a reducing atmosphere (Ar + H2) results in a strong reduction of the luminescence intensity at 980 nm which is consistent with the hypothesis that the emission in this wavelength region is related to the formation of surface states that can be more affected by alterations of the local chemical environment than core states (band A and band C). Finally, by taking into account all the experimental data a plausible mechanism for the Er sensitization process demonstrated by ultra-small Au nanoclusters can also be proposed, as depicted in Fig. 6. The main idea is that the PL emission at 980 nm observed in the spectra of the Au-implanted samples is resonant with the corresponding absorption transition (4I15/2–4I11/2) of Er3+ ions in silica, thus being a preferential channel for the transfer of energy from the ultra-small Au clusters to the Er ions. Therefore, two distinct paths can be envisaged for this process, indicated in Fig. 6 by the blue (a) and the red (b) arrows, respectively. In the rst case (a), the energytransfer from the gold clusters to the Er ions is a direct nonradiative process (i.e., not involving emission of photons); it may occur by multipole–multipole interaction and/or by exchange mechanisms owing to the overlap of the wavefunctions associated with the 980 nm band of the ultra-small Au clusters and the 4I11/2 level of the Er3+ ions.43 The second proposed mechanism (b) is a two steps process, in which the
Possible mechanisms for the Er sensitization process by ultrasmall Au clusters. The transfer of energy can occur directly through the overlap of the wavefunctions associated with the 980 nm band of the Au clusters and the 4I11/2 level of Er3+ ions (a), or it can be mediated by the Au-related luminescent emission at 980 nm followed by its absorption by the Er ions (b). The two distinct paths are depicted in the figure by the blue and red arrows, respectively. The direct excitation path of the Er3+ ions through the resonant 4I15/2 / 4F7/2 absorption transition at 488 nm is also indicated (c). Fig. 6
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 25 November 2013. Downloaded by University of Stellenbosch on 06/10/2014 11:01:56.
Paper
transfer of energy is mediated rst by the de-excitation of the gold clusters through their luminescent emission at 980 nm and then by the subsequent absorption of the emitted photons by the Er ions. Following this path the energy-transfer process may occur at long range and does not require an optimal distance between the gold clusters and the Er ions to be realized. It is worth noting that at present it is still not possible to denitely rule out one of the two mechanisms. Indeed, in ref. 32 it was demonstrated that both the size of the AuN clusters and the Er–AuN distance have necessarily to be optimized for the energy-transfer process to efficiently occur, but it was not possible to decouple the two contributions. With the results of the present work, we can conclude that, with regard to the AuN cluster dimensions, a proper thermal annealing is needed to promote the growth of the Au aggregates up to their optimal size for the 980 nm band (band B) to efficiently appear, by opening the channel through which the transfer of energy to the Er ions may occur either directly or photon-mediated. At the same time, the Er ions have to be restored in the octahedral conguration which is necessary for their efficient luminescent emission at 1540 nm but also to be ready for the energy-transfer. Nevertheless, further experiments are needed to better clarify the photophysical nature (i.e., radiative or nonradiative) of the energy transfer process and if the existence of a denite Er–AuN distance is needed to maximize the efficiency of the Er sensitization process.
4 Conclusions The room temperature photoluminescence properties in the visible and near-infrared range of ultra-small molecule-like AuN nanoclusters obtained by ion implantation on silica substrates have been investigated and correlated with the Er sensitization effects observed in Er–Au co-implanted samples. Under visible laser excitation intense RT luminescent emission in three different spectral regions around 750 nm (band A), 980 nm (band B) and 1150 nm (band C) was observed as a consequence of the formation of discrete energy levels in the electronic structure of the molecule-like Au nanoclusters. Indeed, the energy maxima of band A and C scale with N1/3 as expected for quantum conned systems. Conversely, the energy maximum of band B appears to be almost independent of size, suggesting a contribution of electronic surface states. A denite correlation between the formation of band B in the samples and Er-related emission was demonstrated that allowed us to elucidate the mechanism through which few-atoms Au nanoclusters act as very efficient Er sensitizers. In particular it was found that the formation of a luminescent band at 980 nm resonant with the corresponding Er3+ absorption level represents an effective deexcitation channel through which the Au-related photon energy may be transferred from Au nanoclusters to Er ions (either directly or mediated by photon absorption), eventually producing the Er-related infrared emission at 1540 nm.
Acknowledgements This work has been partially supported by the Progetto di Ateneo CPDA101587 of the University of Padova, Italy.
This journal is © The Royal Society of Chemistry 2014
Nanoscale
References 1 J. N. Anker, W. P. Hall, O. Lyandres, N. C. Shah, J. Zhao and R. P. Van Duyne, Nat. Mater., 2008, 7, 442–453. 2 K. M. Mayer and J. H. Hafner, Chem. Rev., 2011, 111, 3828– 3857. 3 G. Mattei, P. Mazzoldi, M. L. Post, D. Buso, M. Guglielmi and A. Martucci, Adv. Mater., 2007, 19, 561–564. 4 H. A. Atwater and A. Polman, Nat. Mater., 2010, 9, 205–213. 5 M. Gu, Z. Ouyang, B. Jia, N. Stokes, X. Chen, N. Fahim, X. Li, M. J. Ventura and Z. Shi, Nanophotonics, 2012, 1, 235–248. 6 K. R. Catchpole and A. Polman, Appl. Phys. Lett., 2008, 93, 191113. 7 M. S. Chen and D. W. Goodman, Science, 2004, 306, 252–255. 8 R. Narayanan and M. A. El-Sayed, J. Phys. Chem. B, 2005, 109, 12663–12676. 9 M. Haruta, Gold Bull., 2004, 37, 27–36. 10 P. Anger, P. Bharadwaj and L. Novotny, Phys. Rev. Lett., 2006, 96, 113002. 11 P. Genevet, J. Tetienne, E. Gatzogiannis, R. Blanchard, M. A. Kats, M. O. Scully and F. Capasso, Nano Lett., 2010, 10, 4880–4883. 12 A. L. Stepanov, Rev. Adv. Mater. Sci., 2011, 27, 115–145. 13 T. Cesca, P. Calvelli, G. Battaglin, P. Mazzoldi and G. Mattei, Opt. Express, 2012, 20, 4537–4547. 14 X. Wu, X. He, K. Wang, C. Xie, B. Zhou and Z. Qing, Nanoscale, 2010, 2, 2244–2249. 15 X. Wu, T. Ming, X. Wang, P. Wang, J. Wang and J. Chen, ACS Nano, 2010, 4, 113–120. 16 J. Chen, D. Wang, J. Xi, L. Au, A. Siekkinen, A. Warsen, Z.-Y. Li, H. Zhang, Y. Xia and X. Li, Nano Lett., 2007, 7, 1318–1322. 17 S. Eustis and M. A. El-Sayed, Chem. Soc. Rev., 2006, 35, 209. 18 R. Jin, Nanoscale, 2010, 2, 343–362. 19 N. Durante, A. Fortunelli, M. Broyer and M. Stener, J. Phys. Chem. C, 2011, 115, 6277–6282. 20 Z. Y. Li, J. P. Wilcoxon, F. Yin, Y. Chen, R. E. Palmer and R. L. Johnston, Faraday Discuss., 2008, 138, 363–373. 21 U. Kreibig and M. Vollmer, Optical Properties of Metal Clusters, Springer-Verlag, Berlin Heidelberg, 1995. 22 O. D. H¨ aberlen, S.-C. Chung, M. Stener and N. R¨ osch, J. Chem. Phys., 1997, 106, 5189–5201. 23 J. Zheng, P. R. Nicovich and R. M. Dickson, Annu. Rev. Phys. Chem., 2007, 58, 409–431. 24 T. P. Bigioni, R. L. Whetten and O. Dag, J. Phys. Chem. B, 2000, 104, 6983–6986. 25 S. Link, A. Beeby, S. FitzGerald, M. A. El-Sayed, T. G. Schaaff and R. L. Whetten, J. Phys. Chem. B, 2002, 106, 3410–3415. 26 M. Zhu, C. M. Aikens, F. J. Hollander, G. C. Schatz and R. Jin, J. Am. Chem. Soc., 2008, 130, 5883–5885. 27 M. Eichelbaum, K. Rademann, A. Hoell, D. M. Tatchev, W. Weigel, R. St¨ osser and G. Pacchioni, Nanotechnology, 2008, 19, 135701. 28 H. Kawasaki, K. Hamaguchi, I. Osaka and R. Arakawa, Adv. Funct. Mater., 2011, 21, 3508–3515.
Nanoscale, 2014, 6, 1716–1724 | 1723
View Article Online
Published on 25 November 2013. Downloaded by University of Stellenbosch on 06/10/2014 11:01:56.
Nanoscale
29 J. Xie, Y. Zheng and J. Y. Ying, J. Am. Chem. Soc., 2009, 131, 888–889. 30 L. A. Peyser, A. E. Vinson, A. P. Bartko and R. M. Dickson, Science, 2001, 291, 103–106. 31 L. Prodi, G. Battistini, L. Dolci, M. Montalti and N. Zaccheroni, in Frontiers in Surface Nanophotonics, Springer, New York, 2007, vol. 133, pp. 99–128. 32 C. Maurizio, E. Trave, G. Perotto, V. Bello, D. Pasqualini, P. Mazzoldi, G. Battaglin, T. Cesca, C. Scian and G. Mattei, Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 83, 195430. 33 E. Trave, G. Mattei, P. Mazzoldi, G. Pellegrini, C. Scian, C. Maurizio and G. Battaglin, Appl. Phys. Lett., 2006, 89, 151121. 34 A. Polman, J. Appl. Phys., 1997, 82, 1. 35 M. Eichelbaum and K. Rademann, Adv. Funct. Mater., 2009, 19, 2045–2052. 36 P. Crespo, R. Litr´ an, T. C. Rojas, M. Multigner, J. M. de la Fuente, J. C. S´ anchez-L´ opez, M. A. Garca, A. Hernando,
1724 | Nanoscale, 2014, 6, 1716–1724
Paper
37
38 39
40 41 42 43
S. Penad´ es and A. Fern´ andez, Phys. Rev. Lett., 2004, 93, 087204. T. Cesca, C. Maurizio, B. Kalinic, G. Perotto, P. Mazzoldi, E. Trave, G. Battaglin and G. Mattei, Opt. Express, 2012, 20, 16639–16649. J. Zheng, C. Zhang and R. M. Dickson, Phys. Rev. Lett., 2004, 93, 077402. S. Chen, R. S. Ingram, M. J. Hostetler, J. J. Pietron, R. W. Murray, T. G. Schaaff, J. T. Khoury, M. M. Alvarez and R. L. Whetten, Science, 1998, 280, 2098–2101. G. T. Boyd, Z. H. Yu and Y. R. Shen, Phys. Rev. B: Condens. Matter Mater. Phys., 1986, 33, 7923–7936. J. P. Wilcoxon, J. E. Martin, F. Parsapour, B. Wiedenman and D. F. Kelley, J. Chem. Phys., 1998, 108, 9137–9143. G. Wang, T. Huang, R. W. Murray, L. Menard and R. G. Nuzzo, J. Am. Chem. Soc., 2004, 127, 812–813. D. L. Dexter, J. Chem. Phys., 1953, 21, 836–850.
This journal is © The Royal Society of Chemistry 2014