Photoluminescence from Gold and Silver Nanoparticles

This review is an attempt to highlight some of the significant results of the work carried out on the photoluminescence from nanoparticles of the noble metals, particularly gold and silver, over the past two decades. Although quite an immense amount of reports can be found, those that have contributed in throwing some light on the underlying mechanism behind photoluminescence have been considered here. Interband radiative recombination of electrons in metals or photoluminescence (PL), though very weak, was first reported in Au, Cu and Au-Cu alloys. A simple model attributes the PL to the radiative recombination of conduction band electrons below the Fermi energy with d-band holes. Most of the mechanisms are based on this concept. Only small sized clusters are known to exhibit luminescence, with the appearance of additional features which changed with the surfactants suggesting ligand to metal charge transfer. Further, the observation that more polar ligands do indeed enhance the luminescence intensity supports ligand to metal charge transfer. A non-radiative decay of excited electrons from 6sp-band to interface electron energy levels or bands (IEEB), that could be created due to charge transfer from the ligand to the metal core, followed by radiative recombination of electrons from these levels with the hole in the d-band could be another possible mechanism, which is supported by the size independence of the PL emission peak position. However, it is possible that these mechanisms operate independently or even simultaneously depending on various factors like size, ligands, dispersion medium, particle surface topography and so on.


Introduction
The unique optical and electronic properties of metal nanoparticles have opened a range of applications, these include SERS based biosensing [1], cancer diagnostics and therapy [2], vapor sensing [3], catalysis [4] to cite a few. The quantum size effect, observed when the particle size approaches the electron Fermi wavelength, is predicted to cause a discretization of the energy levels of the conduction electron [5]. The appearance of the discrete energy levels can also be expected to give rise to other optical effects, not seen in the bulk. Relatively intense, and to a certain extent size dependent photoluminescence from metal nanoparticles, first reported by Wilcoxon et al. [6], is an example of a property exhibited by a metal only in nano size regime but not expected in bulk.
An examination of the basic requirements for observing this phenomenon would help in understanding why a bulk metal is not expected to show photoluminescence. Photoluminescence is a consequence of the interaction of light with matter. When electromagnetic radiation is incident on a material, it absorbs the energy and a fraction of the absorbed energy may be re-emitted in the form of electromagnetic radiation in the visible or near visible region. This is referred to as photoluminescence. Luminescence is a process which involves at least two steps: the excitation of the electronic system of the solid and the subsequent emission of photons. These steps may or may not be separated by intermediate processes. An intermediate process is implied if the emitted radiation is of a lower energy, which is usually the case. Therefore, only a part of the absorbed energy is transformed into luminescent light. The rest of the energy is lost by non-radiative means, such as the excitation of phonons in a solid, or molecular vibrations, or heat. Photoluminescence can be classified into fluorescence and phosphorescence. In fluorescence the time delay between the absorption and emission of energy is short, whereas in phosphorescence this delay can be relatively long. A decay time of ~10 -8 second can be taken as a demarcation line between fluorescence and phosphorescence.

2
Nano Hybrids Vol. 6 In a simple model of a semiconductor, photoluminescence results due to the relaxation of an electron excited from the valence to the conduction band. This relaxation would be radiative in nature, resulting in the emission of photons of energy at least equal to the band gap of the material.
Non-radiative decay, in this case, is less probable as the energy is too high to be dissipated via a phonon. However, the absence of a finite band gap between occupied and unoccupied electronic states in a metal makes photoluminescence exceedingly improbable. This is so because a photoexcited electron can relax non-radiatively back to the ground state by energy transfer through the closely spaced energy levels. The discretization of the energy levels in metal nanoparticles, arising from the quantum confinement effects, can lead to changes in the optical properties. The fluorescence reported from metal nanoparticles in the last two decades is associated with the changes in the fundamental properties of the electronic energy levels in metals. Therefore, an in-depth study of photoluminescence/fluorescence from metal nanoparticles would provide means for understanding the properties of matter in the nanoscale.
This article is an attempt to highlight some of the significant results of the work carried out in this field over the past two decades. Although quite an immense amount of reports on the photoluminescence from nanoparticles of the noble metals have been reported, those that have contributed in throwing some light on the underlying mechanism behind this phenomena have been considered here. Luminescence spectroscopy is among the simplest and versatile technique for many applications, such as molecular recognition and labeling, for medical, biological and general analytical purposes [7]. Research on fluorescent nanoparticles has mainly been concentrated on semiconductor nanoparticles such as CdSe, in which the fluorescence can easily be tuned to span over the visible region of the electromagnetic spectrum. But the possibility of use of these nanoparticles in-vivo appears very limited due to their toxicity [8]. Nanoparticles of noble metals are better suited for biological applications because of their reduced toxicity and the innumerable easy synthetic methods available yielding highly stable colloids. Among the noble metals, gold (AuNP) Nano Hybrids Vol. 6 3 and silver nanoparticles (AgNP) are the ones that have been investigated the most. Hence, we first look into the results of luminescence from AuNP followed by AgNPs.

Photoluminescence
From Au Nanoparticles. Interband radiative recombination of electrons in metals or photoluminescence (PL), though very weak, was first reported by Mooradian [9] in Au, Cu and Au-Cu alloys when excited by an argon-ion laser emitting in the range 457.9 nm to 514.5 or from Hg arc lamp emitting in the 300-400 nm range. Irrespective of the excitation wavelength the Au emission was found to be at a higher energy of 2.2 eV (~ 564 nm) compared to the Cu emission at 2.0 eV (~ 620 nm) and the emission peaks of the alloys were found to lie in between these two. The emission energies are close to the energy gap between the upper d-band and the Fermi energy and appeared as a broad peak centered near their respective interband absorption edge.
Mooradian [9] proposed a simple model to explain the PL from noble metals and it is noteworthy that many of the mechanisms proposed for PL from noble metal nanoparticles are to a large extent, based on this model. A few aspects of the band structure of noble metals, such as the interaction between the conduction band and the d-band at the region of crossover and the band being made up of a number of closely spaced bands in k space, are not explicitly detailed in the representation.
However, the simple model suffices to understand the underlying mechanism for the PL emission. A photoexcited electron from upper states in the d-band makes a transition to levels at and above the smooth films of Au and Cu could be attributed to direct radiative recombination of electrons below the Fermi level with holes in the d-bands around the X and L symmetry points, those from Ag appeared to be predominantly due to radiative emission from the surface and volume plasmons.
The work by Boyd was further investigated in more detail by Apell et al. [11] for bulk noble metals. Who developed a theory based on the optical constants and density of states of the metal being probed. According to this theory, the process of luminescence emission proceeds with the initial excitation of an electron in the d-band to states above the Fermi level. The hole left behind is scattered to a different state in the d-band by a phonon. Another electron from the filled sp-band, with energy less than the Fermi energy, transits directly into this hole emitting a photon of lower energy in the process.
It is interesting to note that a detailed study of luminescence from Au nanoparticles was carried out and the anticipated mechanism proposed for this phenomenon in one of the first known reports of photoluminescence from metal nanoparticles by Wilcoxon et al. [6]. Their work encompasses studies on the PL from Au nanoparticles synthesized by different methods, its behavior with change of particle size, the effect of different ligands and solvents. They have resorted to a rather different technique to purify and identify the size of Au clusters formed in water by the citrate method. Size exclusion liquid chromatography, used for determining the size distribution of water soluble organic polymers, was used to separate and purify the clusters. Clusters of different sizes eluted from the column at different times (larger clusters eluted at shorter times) and their absorption and PL spectra were recorded. Since this study was the first report on PL from metal nanoparticles, precautions were taken by the authors to confirm that the observed emission is indeed PL from Au clusters. It was concluded that the peak in the PL spectra obtained at ~ 440 nm is indeed that due to small Au clusters because (i) clusters of different sizes showed different PL intensities although the concentration of the reagents was the same and therefore, that of the by-products are the same, (ii) water containing the precursor ionic chemicals did not exhibit any photoluminescence and (iii) small Au clusters synthesized by the inverse micellar method in organic solvents too exhibited PL.

Nano Hybrids Vol. 6 5
Wilcoxon et al. [6] observed that the 15 nm clusters elute from the column after time t = 8.9 min, whereas clusters of 5 nm dimension elute at t = 13.2 min. A highly intense PL signal at 440 nm (excitation wavelength 230 nm) is obtained during the passage of the 5 nm fraction through the column whereas no PL is observed from the fraction eluting at 8.9 min. ie. the 15 nm clusters. But this fraction exhibited an absorption peak at 520 nm. The quantum yield from the 5 nm clusters had a very modest value of ~10 -4 -10 -5 , yet 10 5 more than that reported from bulk surfaces. Based on these observations it was concluded that (i) The emission occurring near 440 nm is due to the radiative recombination of Fermi level electrons with sp-or d-band holes (ii) the surface ligands attached to the Au core play a role in determining the non-radiative energy loss paths and therefore the shape and position of the final emitted signal and (iii) an efficient coupling between the incident radiation and the surface plasmon enhances the strength of the emitted radiation.
Varnavski and co-workers [12] reported ultrafast time resolved PL from Au and Ag nanoparticles-polyaminoamine (PAMAM) nanocomposites. The fluorescence efficiency was estimated to be three orders higher than that from bulk metals. The emission maximum was found to be close to the absorption peak at around 550 nm. The formation of elongated non-spherical clusters is reported to be the cause for the high enhancement in efficiency.  with that of the dendrimer to be one more possible mechanism for the ultrafast luminescence emission observed in dendrimer encapsulated non-spherical Au particles is also suggested.
Zheng and co-workers [13] synthesized poly(amidoamine) dendrimer (PAMAM) encapsulated Au nanodots having high quantum yield and whose absorption and fluorescence could be tuned from the UV to the NIR region by controlling their size, achieved by varying the relative Au:PAMAM concentration. These nanodots are anticipated to have good potential for application in biolabelling as their synthesis requires neither complicated high temperature procedures with toxic precursors nor difficult overcoating with surface passivation. The photophysical properties such as the excitation and emission wavelengths and the quantum yields varied with size of the clusters.
Emission wavelengths varied from 385 nm to 879 nm, with the varying excitation wavelengths from 330 nm to 765 nm for the cluster size variation from Au 5 to Au 31 . The quantum yield decreased from 70% to 10% for the corresponding variation in cluster size [13].
The quantum yields of these clusters, especially the smaller ones, are claimed to be comparable to the best water soluble emitters available. The observed decrease in the quantum yield with increase in cluster size has been attributed to the increased radiation pathways at lower energies. The authors observed that the emission energy scaled inversely with the cluster radius as N -1/3 , as per the jellium model, which indicates that not only are the electronic structure of these clusters determined by the metal cores but also that they are approximately spherical. Hence, these clusters have been called "multielectron artificial atoms" with the difference that in a true atom the electron density changes with atomic number whereas it does not in these clusters. The discrete excitation and emission spectra exhibited by these artificial atoms is therefore argued to be due to the conduction electrons making intraband transitions between nanocluster electronic states of well defined angular momenta.
Blue emission from second and fourth generation OH terminated PAMAM stabilized Au 8 nanodot had earlier been reported by Zheng and co-workers [14], but no emission was observed from the clusters stabilized with zeroeth generation dendrimer. An increase in the quantum yield in methanol had also been reported. Fluorescence lifetime measurements showed the presence of a   A similar result has also been observed by Link et al. [19] in a Au 28 (Fig. 3b). The possibility that the low energy emission could also originate as a result of ligand to metal charge transfer is also briefly discussed.
Wang et al. [20] have shown that MPCs with different core sizes and attached with different ligands all emit over a broad range of energies extending from ~ 700 nm to ~ 1.3 µm depending on the ligand attached. The emission intensities depend on the core metal as well as the polar nature of the attached ligands with more polar ligands emitting with higher intensity. Absorption studies gave values of optical "band gap" that increased with decreasing size of the Au clusters but no such size dependence was observed for the luminescence emission energy. These make the authors propose that localized core surface states may probably play a role in the emission from the Au nanoparticles.
The surmise that the ligands attached to the Au core affects the luminescence from these particles was studied in more detail by Wang et al. [21]. In this, luminescence from clusters of Au with 38 and could be made and conclusions arrived at from the detailed studies. It was shown clearly that irrespective of the number of atoms in the cluster, the intensity of luminescence emission increased linearly with the number of exchanged ligands when more polar ligands than the original were incorporated on to the cluster surface. Fig. 4 shows unambiguously this result for the Au 38 cluster and confirms the initial findings that more polar ligands do indeed enhance the luminescence intensity. NPs are said to provide confirmatory evidence for their argument. NIR luminescence from small dodecanethiolate stabilized AuNPs of size 1.1 nm and 1.7 nm was reported by Bigioni et al. [24].
Since an electronic energy gap of 0.9 eV and 0.5 eV respectively was observed in the absorption spectra of these particles the authors expected and did find a broad luminescence emission in the NIR region of the electromagnetic spectrum. Since the excitation wavelength used was 1064 nm the authors conclude that the luminescence emission is not associated with the ligand shell since the electronic transition in alkanethiols is in the deep ultraviolet region. Metal-ligand charge transfer or vice-versa is also ruled out due to the inaccessibility of the states in the ligand molecule. Since the emission range is of low energy, 6sp-5d transition is also ruled out. Similarly a radiative transition across the HOMO-LUMO gap again cannot explain the broad emission observed. It is therefore 16 Nano Hybrids Vol. 6 concluded by the authors that the photoluminescence could be due to intraband gold sp to sp-like transition.
PL from thiosalicylic acid stabilized AuNPs of average size 5, 8 and 11 nm were studied by Pattabi and Pattabi [25]. When colloidal solutions of the samples in water were excited at 300 nm, emission maximum was observed at ~ 400 nm, irrespective of sample size. The size independence of the PL emission peak is propounded to arise due to non-radiative decay of excited electrons from 6sp-band to interface electron energy levels or bands (IEEB) that could be created due to charge transfer from the ligand to the metal core followed by radiative recombination of electrons from these levels with the hole in the d-band. The position of the interface electron energy levels depend only on the ligand and the metal and therefore, can be expected to be unaffected by the particle size.
The above argument can explain the independence of the PL emission peak position with size of the particle. The schematic diagram of the proposed mechanism for the observed photoluminescence is shown in Fig. 5.  AuNPs is responsible for the observed emission [31]. Emission from Au colloids with diameter 50 nm, spin coated on to glass substrates was reported by Geddes et al. [32]. When excited at 532 nm, the emission intensity gradually increased up to 5 minutes after which it remained a constant. But no change in the PL line shape was observed with constant illumination. BSA coated monodispersed Au nanoparticles, 5.1 nm in diameter, was reported to fluoresce at 404 nm when excited at 320 nm [33].

20
Nano Hybrids Vol. 6 The quantum yield was observed to be around eight orders of magnitude higher than that from Au films. The PL intensity was found to be pH and solvent dependent. PL intensity was maximum for pH of 11 and 12, decreased sharply below this value and showed an increase again at 8. The intensity decreased again for lower values of pH. The PL intensity was almost the same when the nanoparticles were dispersed in water, methanol and ethanol but there was a sharp increase in the emission intensity in acetonitrile. It was also observed that in acetonitrile, there was a shift in the maximum excitation wavelength from 320 to 308 nm and emission maximum from 404 to 413 nm, compared to the other dispersing media studied. Philip reported PL from Au nanoparticles of different sizes prepared by different methods [34]. While the 3 and 4 nm sized particles obtained by borohydride reduction in the presence of polyethylene glycol (PEG) and the 9 nm particles synthesized by citrate reduction exhibited PL emission maxima at 408 nm, the 6 nm tannic acid/citrate reduced particles luminesced at 456 nm. The emission intensity was observed to increase upon addition of KCl upto an optimum value after which it decreased. The initial increase has been attributed to the electrostatic stabilization of the Au nanoparticles by Cland the decrease to the aggregation of the particles. Although a size dependent shift in the PL is observed it could also be attributed to the different synthesis method / stabilizers used.
Luminescence from nanorods of Au has been reported [35][36][37][38]. Highly enhanced two photon luminescence with a broad emission in the 400-650 nm region was observed from Au nanorods excited at 830 nm [35]. From the overlap of the PL excitation spectrum and the longitudinal plasmon band in the 700-900 nm region it has been concluded that the local field enhancement from the plasmon resonance has a role in dictating the intensity of the emission. Very large enhancement in fluorescence quantum yield from Au nanorods, compared to that from Au metal, was reported by Mohammed et al. [36].The emission maximum was found to red shift from 548 nm to 588 nm as the aspect ratio of the rods change from 2 to 5.4 for rods having more or less the same width of ~ 20 nm.
Although the longitudinal plasmon resonance peak also red shifts with increasing aspect ratio, the fact that the resonance peak occur at much higher wavelengths (~ 630 nm for rods with aspect ratio Nano Hybrids Vol. 6 21 2) suggests that the emission is not associated with surface plasmon radiative relaxation. The local field theory developed by Boyd et. al. [10] was applied to explain the data. Using this it was shown that the excitation at 480 nm caused both the excitation of surface plasmon oscillation and the delectrons. The emission is a result of slow relaxation of the plasmon oscillation followed by radiative recombination of the electrons in the sp band with holes in the d-band. The same theory also supported the argument that the high intensity was due to large enhancement of the exciting field and the emitted field through the excitation of the surface plasmon resonance which is made possible by the rod shape.
From the applications point of view, nanoparticles embedded in solid matrices have the advantage that they are mechanically stable. In addition, such matrices allow for flexible control of the optical properties of the particles by proper choice of the embedding material. Photoluminescence have been reported from Au nanoparticles embedded in several solid matrices [39][40][41][42]. PL spectra of AuNPs sandwiched between oxide layers of SiO 2 , ZnO and TiO 2 were studied by Liao et. al. [39]. It was observed that the PL emission intensity is appreciable only in the composite multilayers in which the Au particles have sizes greater than 20 nm. The PL peak position was always found to be red shifted with respect to the SPR peaks for all samples except the Au:TiO 2 composite. The PL and SPR from the Au:ZnO multilayers were studied for different thickness of the Au layer. The AuNPs formed after thermal annealing were of different sizes. The PL peak position was always found to lie on the longer wavelength side with respect to the SPR peak for the Au:ZnO films with different Au particle sizes. However, the separation between the SPR and PL peak positions was found to increase for larger values of SPR peak position. The PL emission energy from the composite multilayers seem to be influenced by the surface plasmons, but a definite correlation between the SPR and PL emission peak energies, among the different composites studied, is lacking.

22
Nano Hybrids Vol. 6 PL from AuNPs embedded in germane-silicate glass matrix [40] was observed to exhibit a broadband visible emission at ~ 833 nm as well as a narrowband emission at 1536 nm when excited at 488 nm from an Argon-ion laser. The broadband emission in the visible region has been attributed to interband transitions and the emission centered at 1536 nm to a relaxed radiative recombination across the HOMO-LUMO gap. Eichelbaum et al. [41] observed multiphoton induced white luminescence from Au nanoparticles generated inside thin glassy silicate-titanate films. From the slope of the logarithmic plot of the integrated emission intensity versus the integrated laser intensity it was deduced that at high gold concentration the excitation is a three photon process. The observed PL is again attributed to a radiative recombination of electrons excited to the 6 sp band of Au with the hole in the d band.
Silica beads labeled with luminescent tiopronin protected Au nanoparticles (stabilized with succinimidyl ester terminated ligands) and with size 1.6 nm were shown to have potential to be developed into molecule imaging agents because of their good photostability and bright signals [42]. well as the emission intensity remained stable after photoactivation, thus the "cage effect" of the dendrimers also protects the nanodots from quenchers in solution. Ag n clusters with n < 8 only exhibited emission thereby proving the necessity of the dendrimer to produce fluorescent clusters.
The work thus demonstrated the feasibility of synthesizing stable, highly fluorescent clusters by simple methods without resorting to expensive laser sources.
Huang and Murray [17] had observed emission appearing as a broad band centered at ~ 770 nm from tiopronin coated Au nanoparticles. The same researchers observed size independent luminescence emission at 500 nm from tiopronin coated Ag nanoparticles [44]. Their work was also an elegant demonstration of the PL emission from tiopronin coated Ag shifting gradually to that of Au when Ag was replaced with Au by reacting the Ag monolayer protected cluster with a thiolate complex of Au. When Ag was thus exchanged with Au, the surface plasmon resonance peak of Ag at 380 nm diminished in intensity. As expected, no Au peak was observed due to their size being less than 2 nm. As for the PL spectra, the emission peak of Ag at 500 nm diminished with increasing Au atoms, accompanied by the growth of an emission peak at 700 nm corresponding to that of Au. A conclusion drawn from their study is that the emission is controlled by surface atoms rather than the core atoms. The Ag NP emission has been attributed to interband recombination of d band holes with sp band electrons, further intensified due to a strong coupling with SPR excitation. Gao et al. [45]  Annapurni [47] synthesized Ag nanoparticles dispersed in water by bringing a 0.2 mm thick Ag wire into contact with a silver plate while a potential difference of 30 V was maintained between the wire 24 Nano Hybrids Vol. 6 and the plate. When these nanoparticles were excited at 390 nm, two fluorescence emission peaks were obtained, one at 450 nm and another at 490 nm. The intensity as well as the peak position was found to depend on the medium in which the nanoparticles were dispersed and the stabilizer used.
When bovine serum albumin was attached to the particle surface only the 490 nm peak blue shifted by about 20 nm, but in cyclohexane both the peaks were observed to blue shift by ~ 30 nm. The appearance of two emission signals is accounted for by the argument that the large interface area presented by small metal particles in solution causes chemical modification of the surface due to the decreased density of states and increased concentration of electrons. This leads to the formation of interface electron energy bands (IEEB) characteristic of the core material. Electrons photoexcited into the lowest unoccupied energy levels of the IEEB may undergo non-radiative decay into two different lower energy levels. A radiative transition from these two levels is proposed to be the reason for the two emission peaks observed. In an earlier work, fluorescence occurring as a broad peak extending from 500 to 600 nm was observed for excitation at 390 nm from Ag nanoparticles synthesized by the same electro-exploding wire technique [48] which has been assigned to emission from various small sized clusters in the solution. The gemini surfactant 2-hydroxy-1, 3bis(octadecyldimethylammonium) propane dibromide stabilized Ag nanoparticles [49] exhibited luminescence at 448 nm when excited at 309 nm. This has been attributed to metal-ligand charge transfer absorption.
PL emission wavelength shifting linearly to higher wavelengths with increase in the excitation wavelength was observed by Basak and co-workers [50]. This shift in the emission peak was observed only when the excitation wavelength was above 370 nm and when the excitation wavelength was increased from 380 nm to 500 nm, the emission peak red shifted from ~ 590 nm to ~ 750 nm. Since the size of the particles was ~ 8 nm, it is argued that the observed PL is not due to quantum confinement effects. At the same time, an enhancement in the peak intensity up to an excitation wavelength of 410 nm which lies in the region of the SPR peak of Ag is also observed.

Nano Hybrids Vol. 6
Based on these two observations, a significant role of the surface plasmon resonance on PL emission is inferred.
Solid phase synthesis of Ag nanoparticles by thermal reduction of silver ions in glycine matrix yielded nanoparticles of different sizes ranging from 2 to 30 nm which could be separated by centrifugation [51]. Luminescence could be observed at the single particle level. The absorption spectra of red emitting 18 nm Ag particles were broader than those from similar sized non- Luminescence at 380 and 510 nm was observed from PVA stabilized Ag nanoparticles when excited at 325 nm [52]. The peak intensity was enhanced with increase in the particle size. While the peak in the UV range is thought to be due to radiative recombination of Femi level electrons and spor d-band holes, the probable origin of the 510 nm peak has been reasoned either be due to the formation of irregular shaped and sized clusters during their deposition on glass slides or due to coupling of the surface plasmon with the Fermi level electrons. The luminescence at two wavelengths, 330 nm and 520 nm, from Ag sol, observed by Zhang et al. [53] and its shift with changes in the particle size has been interpreted as arising due to the formation of hybrid surface energy states due to the mixing, splitting and shifting of the primary plasmon energies.
Maretti et al. [8] synthesized Ag nanoparticles having size ~ 3.4 nm by reducing silver trifluoroacetate in cyclohexylamine with ketyl radicals. The ketyl radicals were photogenerated by irradiating I-2959 with UV light from four lamps. Luminescence was observed at ~ 530 nm when the particles were dispersed in toluene which red shifted to 580 nm in tetrahydrofuron (THF).

26
Nano Hybrids Vol. 6 Absorbance peak from the fluorescent Ag nanoparticles was close to 450 nm which is close to that reported for Ag dimers. The fluorescence decay could be fitted to a monoexponential function with a lifetime of 2.6 ± 0.1 ns. The emission from Ag 2 , Ag 3 and Ag 4 clusters have relatively long lifetimes and the emission and absorption are found to have a mirror-image relationship. The reported lifetime of Ag 2 clusters in Ar matrices is 4.6 ns [54]. These factors support the claim that the emission observed is likely to be from the Ag 2 clusters supported on the nanoparticle surface. The Ag nanoparticles generated by the electro-exploding wire technique resulted in the formation of Ag nanoparticles in the size range 8-20 nm [55]. In water as the dispersing medium, fluorescence emission was observed at 300 nm for excitation wavelengths ranging from 215-235 nm. A decrease in the fluorescence intensity with increasing excitation wavelengths was also observed.

Application
Several researchers have succeeded in synthesizing luminescent particles and demonstrated their applications in various fields. The GSH stabilized nanoparticle's capacity to sense Cu 2+ ions was demonstrated [22] by observing the PL quenching in the presence of these ions and has been attributed to the formation GSH-Cu 2+ complex which induces aggregation of the particles making them less luminescent. The applicabity of the Au nanorods synthesized by Wang et al. [35] as two photon luminescence imaging agents was demonstrated in vivo by imaging the flow of single gold nanorods through the blood vessels in mouse ear. Au 15 clusters protected with glutathione and further stabilized in cyclodextrin cavities were incorporated into chitosan films to obtain free standing films of Au 15 clusters [56]. The luminescence from these films were quenched on exposure to Cu 2+ ions thus making them useful as a metal ion sensor. Liu et. al. have synthesized NIR emitting Au nanoclusters capable of detecting Cu 2+ ions at concentrations as low 0.3 nM [57]. The emission at ~ 657 nm from very small lyzosome stabilized Au nanoclusters of diameter ~ 1 nm, was shown to be effectively quenched by Hg 2+ , demonstrating their applicability for the detection of Hg 2+ Nano Hybrids Vol. 6 27 ions with a detection limit of 10 nM [58]. Comparatively large sized, fluorescent (>16 nm) AuNPs, synthesized by citrate reduction technique were investigated for their viability for the imaging of cervical cancer HeLa cells [59]. The particles all fluoresced at 610 nm when excited at 514.5 nm and the emission peak was invariant with size. AuNPs of size 38 nm were bio-conjugated and were then targeted to the cancer cells. The cancer cells could be imaged by recording the fluorescence from the AuNPs. Au 8 clusters encapsulated with fourth generation polyamidoamine dendrimers with amine terminated groups was also found suitable for intra-cellular imaging of cancer cells by monitoring their fluorescence emission [60]. Glutathione stabilized Au nanoparticles, 2 nm in size, showed extremely high renal clearance while at the same time exhibiting good stability in the blood stream.
They appear to be promising candidates for biological imaging [61]. The feasibility of detecting calf thymus-DNA (ct-DNA) by spectrofluorometric method was demonstrated by Liu et al. [62] using AgNPs synthesized by an aqueous gas phased reaction of AgNO 3 and ammonia gas. The fluorescence from these particles quenched upon addition of calf thymus-DNA. Since the fluorescence intensity depended on the ct-DNA concentration, these particles are suggested to be useful in their detection.

Conclusion
Numerous studies have been carried out over the past two decades with an aim to elucidate the mechanism for photoluminescence from metal nanoparticles, especially those of the noble metals.
The results have been at variance, in terms of the size effect on the luminescence, the ligand effect on this phenomena and the effect of the core metal as well as the luminescence efficiency. A few theories have been proposed to explain PL based on either an interband sp to d transitions of the photoexcited electrons or on the formation of an interface electron energy band and the transitions through these IEEB. The concept of metal-ligand charge transfer and the formation of IEEB is contained in many of the results including the first report of Wilcoxon et al. [6], much earlier to the explicit mention of that. It is possible that these mechanisms operate independently or even 28 Nano Hybrids Vol. 6 simultaneously depending on various factors like size, ligands, dispersion medium, particle surface topography and so on. A consensus has still not been arrived at as to the mechanisms involved in PL from metal nanoparticles. Therefore, still more concerted effort, with good control over the growth process, is required to achieve a fundamental understanding of photoluminescence from metal nanoparticles. Only then, can it be hoped that PL from metal particles can find in vivo applications.