Low Threshold, Amplified Spontaneous Emission from Core‐Seeded Semiconductor Nanotetrapods Incorporated into a Sol–Gel Matrix

Abstract

Wet-chemically synthesized CdSe seeded CdS nanotetrapods incorporated into silica matrices have been explored as high performance nanocrystal-based optical amplifiers. Varying the physical dimensions of the tetrapods allowed for room temperature biexcitonic amplified spontaneous emission (ASE) at wavelengths corresponding to either the CdSe core or CdS arms, while higher pump intensities resulted in ASE from a higher CdSe excited state. These results collectively indicate that such heterostructured tetrapods can facilitate tunable, multiple-wavelength lasing across a wide spectral range from a single nanostructure. Wet-chemically synthesized colloidal semiconductor nanocrystals (NCs) are desirable as optical gain media due to their size-dependent emission wavelengths, ease of fabrication and flexible surface chemistry which facilitates incorporation into an optical feedback device of various configurations.1-3 Early attempts to achieve stimulated emission via close-packed films of NCs required low temperature and high pump fluences since such films generally suffer from low optical quality, poor heat dissipation, and consequently short operational device lifetimes.4, 5 On the other hand, the incorporation of spherical core-shell colloidal NCs such as CdS/ZnS and CdSe/CdZnS into sol-gel derived matrices facilitated room temperature amplified spontaneous emission (ASE) and lasing with continuous operability over a time scale of hours,1, 6 thus providing a more pragmatic route to NC-based optical amplifiers. Despite the enhanced thermal resilience afforded by the sol-gel derived host matrix, however, the excitation pump intensities and volume fraction of NCs required to attain ASE remained expectedly high since the thresholds are dominated by the Auger recombination processes stemming from the intrinsic photophysical properties of the NCs.5, 7, 8 In order to further enhance the performance of NC-based sol-gel derived composites for optical gain applications it is clear that NCs with more salient optical properties and reduced Auger recombination rates need to be employed. Recent studies on CdSe seeded CdS nanotetrapods synthesized via seeded growth techniques have shown that they can possess extremely large linear absorption cross-sections (ϵ ∼ 1.5 × 108 M−1 cm−1 at 350 nm) and relatively high quantum yields of up to ∼60%.9 Indeed, due to their unique tetrahedral-like arrangement, the tetrapod arms can capture photons from a wide range of incident angles and polarization more effectively than nanorods of the same volume, thus leading to a very large absorption cross section of the entire nanostructure. Radiative recombination typically takes place at the CdSe core at the centre of four rod-like CdS arms, and extensive leakage of the CdSe electron wavefunction into the CdS shell results in a dramatic redshift in its optical spectra relative to the original CdSe NC seed.9 The ensuing reduced spatial overlap between the CdSe electron and hole wavefunctions is expected to yield suppressed Auger recombination rates, as was previously observed in a related system of CdSe seeded CdS nanorods exhibiting relatively long biexciton lifetimes of ∼300 ps.10 While it has been suggested that CdSe seeded CdS nanotetrapods hold promise as good NC-based optical gain materials,9, 11 however, ASE from such structures has to the best of our knowledge not yet been demonstrated. In this work we synthesize and incorporate ZnS overcoated CdSe seeded CdS nanotetrapods into a solvent resistant and thermally stable sol-gel derived silica matrix, and demonstrate room temperature ASE from the resulting nanotetrapod-silica composites. Photoluminescence (PL) lifetime measurements on tetrapod solutions suggested remarkably long CdSe biexciton lifetimes of ∼1.6 ns while Variable Stripe Length (VSL) measurements on tetrapod-silica thin films revealed modal gains of ∼151 cm−1, along with extremely low thresholds of ∼68 μJ cm−2 that was achieved using relatively small NC volume fractions of ∼6%. The biexcitonic ASE from the core seeded tetrapods was found to originate from either the CdSe core or CdS arms depending on the CdSe core size while higher excitation intensities facilitated the exhibition of multiexcitonic ASE from the CdSe core under ambient conditions, thus affording a possible basis for multi-wavelength lasing within the same nanostructure over a large spectral range. These results collectively indicate the utility and potential of core-seeded heterostructured semiconductor nanotetrapods in NC-silica composites as high performance NC-based optical amplifiers. The CdSe seeded CdS tetrapods used in this study were synthesized using the seeded growth approach of Manna et. al.,12 with the distinction that a mixture of oleic acid (OA) and hexylphosphonic acid (HPA) was used as surfactants instead of octadecylphosphonic acid (ODPA) and HPA. As illustrated in the Transmission Electron Microscope (TEM) images in Figure 1(a), this allowed for monodisperse tetrapods of various arm dimensions with yields as high as 90% (with respect to nanorod, bipod and tripod by-products) to be obtained, which may be attributed to the stabilization of the zinc blende phase of CdSe by OA.13 Initial attempts to directly incorporate the as-synthesized CdSe seeded CdS tetrapods into sol-gel derived silica films proved futile as the tetrapods could not be ligand-exchanged at elevated temperatures with hydrophilic molecules such as 5-amino-1-pentanol and be rendered dispersible in ethanol. This may be understood from the high binding affinity of alkylphosphonic acids to Cd in cadmium chalcogenide NCs,14 making the surface intractable to ligand-exchange processes under mild conditions. In order to circumvent this issue, the tetrapods were overcoated with ZnS in the presence of trioctylphosphine (TOP) and trioctylphosphine oxide (TOPO) at a relatively low temperature of ∼ 130 °C since it was often observed that a significant fraction of tetrapod arms had broken off at temperatures in excess of ∼ 180 °C. Subsequent incorporation of the processed ZnS overcoated CdSe seeded CdS tetrapods into sol-gel derived silica via slight modifications of a published method6 was successful, attesting to the need for the surface chemical modification afforded by the ZnS overcoating process. Atomic Force Microscope (AFM) measurements on the resulting tetrapod-silica composite film, as exemplified in Figure 1(b), yielded a smooth morphology with a typical surface roughness (RMS) of about ∼7.5 nm, over a 1 μm2 scan area, suggesting a uniform incorporation of tetrapods into the sol-gel derived silica matrix. Absorbance measurements, profilometry analysis of the film thickness and the molar absorptivity of CdSe seeded CdS tetrapods determined via a previously reported method9 provided an estimate of the volume fraction of tetrapods in the silica film, which was about ∼6-10% for all the films presented in this work. It is important, however, to note that the volume fraction with respect to the CdSe core is less than 1% given the dominant contribution of CdS to the overall volume of the CdSe seeded CdS tetrapod structure. (a) Representative TEM image of monodispersed ZnS overcoated CdSe seeded CdS tetrapods featuring a CdSe core diameter of ∼3.1nm with CdS arms of ∼21 nm in length and ∼5 nm in diameter. (b) Representative AFM image of a typical tetrapod-silica film spin-coated onto a glass substrate. Optical excitation of the tetrapod-silica films proceeded via a frequency-doubled Ti:Sapph with a pulse duration of 150 fs at a repetition rate of 1 kHz. The laser pulses were focused into a ∼5.5 mm × 0.5 mm stripe using a cylindrical lens and the resulting emission was collected via an optical fiber coupled to a spectrometer and detected by a charge coupled device (CCD) camera. Figure 2(a) illustrates the room temperature photoluminescence (PL) spectra at different excitation pump intensities of tetrapod-silica composite films comprising of 5.4 nm CdSe cores with CdS arms approximately 21 nm long and 5 nm in diameter. A narrowing of the emission profile from ∼38 nm to ∼11 nm is readily observed at higher pump powers, signifying the onset of ASE. A low threshold of ∼ 68 μJ cm−2 was derived from a plot of the integrated PL intensity against pump power as depicted in Figure 2(b), which is 5 and 50 times lower than the best ASE thresholds previously reported in CdSe/CdS/ZnS based zirconia films,15 and CdS/ZnS based silica films6 respectively. It is possible that further reduction in the threshold may be achieved using state-resolved optical pumping, which was demonstrated recently in CdSe NC-based films.16, 17 The photostability of the ASE in CdSe/CdS tetrapod-silica film was evaluated over 75 min of constant excitation, or 4.5 × 106 laser shots, as shown in the inset of Figure 2(b). No obvious decrease in the ASE intensity was observed within the time window of our measurments, indicative of the optical robustness of the the tetrapod-silica composites. A slight increase in the ASE intensity may be seen during the first 25 min of excitation, which may either be due to a thermally-mediated densification of the silica matrix or photo-annealing of the incorporated semiconductor tetrapods. The exact nature of the photo-brightening of the tetrapods in such sol-gel derived films is presently unclear, although such a process has also previously been observed in spherical CdSe nanocrystals under above-gap illumination.18, 19 When CdSe cores of smaller diameter (∼3.1 nm) but with similar CdS arm dimensions were employed, excitation at moderate intensities yielded dual fluorescence emission from the CdS arms and CdSe core, as shown in Figure 2(c). A similar observation was made by Lutich et. al, except that in their case the CdS arm length was varied instead of the size of the CdSe core. Their measurements did not show any discernible manifestation of ASE despite excitation pump fluences of up to ∼600 μJ cm−2,20 however, presumably because of the low volume fractions associated with particles dispersed in solution. In the case of tetrapod-silica films, higher pump intensities resulted in ASE that was not from the CdSe core but rather from the CdS arms, yielding a ASE threshold of ∼120 μJ cm−2, as seen in Figure 2(d). In contrast with the ASE from the CdSe core, the ASE intensity from the CdS shell as a function of time, as depicted in the inset of Figure 2(d), showed considerable photo-brightening and decay that may be attributed to a less passivated surface and a resulting higher tendency to trap photoexcited carriers. Room temperature ASE spectra from CdSe seeded CdS tetrapod-silica films with approximately similar CdS arm dimensions of ∼21 nm in length and ∼5 nm in diameter but with a CdSe core size of (a) ∼5.4 nm diameter and (c) ∼3.1 nm diameter respectively. The insets in (a) and (c) illustrate the PL spectra at low (black) and high excitation intensities (red), where a narrowing of the PL is observed. (b) and (d) are the corresponding integrated PL intensity versus pump power spectra of the ASE transition, with the insets depicting the intensity of the ASE transition as a function of the number of laser shots from the pump source. The initial decrease in intensity in the inset of (d) is attributed to excitation pump instability. Additionally, the red dotted line in (d), which reflects the CdSe core emission in (c) shows a linear increase in PL as a function of pump power. The intriguing omission of ASE in small CdSe core seeded CdS tetrapods cannot be ascribed to a volume-based, defect-mediated photoinduced absorption process that was previously used to explicate the absence and presence of ASE in short and long CdSe seeded CdS nanorods respectively,21 given that in our case the dimensions of the CdS arms in both small and large core tetrapods as well as their volume fractions in the film are comparable. One plausible explanation would be to consider an exciton blocking effect in which the small CdSe core is populated with a high density of excitons at sufficiently large excitation intensities such that the electrons and holes photogenerated in CdS do not experience a strong driving force for relaxing into the CdSe core, thus recombining radiatively.20 This effect is further accentuated by the quasi-Type II band alignment between the small CdSe core and the CdS arms which results in slower single exciton CdSe recombination rates,9 therefore affording less opportunity for excitons in CdS to relax into the CdSe core. It is also likely that although the extensive leakage of the CdSe electron wavefunction into the CdS shell should result in reduced CdSe Auger recombination rates and thus lower ASE thresholds, a concurrent reduction in the oscillator strength can also diminish the optical gain cross section in such quasi-Type II structures.22 Conversely, for larger CdSe cores possessing a Type I band alignment with the CdS arms, stimulated emission in the CdSe core rapidly depletes the population of excitons photogenerated in CdS since both stimulated emission and CdS carrier relaxation into the CdSe core are extremely fast processes.23 Given that the achievement of stimulated emission in the CdSe core occurs at the expense of exciton buildup in the CdS shell, simultaneous ASE from both moieties is decidedly difficult, and our many attempts to achieve concurrent dual wavelength ASE from both the core and arms were without success. While an independent detailed study of the complex exciton dynamics of many-body interactions between multiple excitons generated in the CdSe core and CdS arms is clearly warranted here, it is foreseeable that large-core CdSe seeded CdS tetrapods with sufficiently long arms such that exciton relaxation to the CdSe core does not extend across a significant fraction of the arm length may exhibit dual wavelength ASE from both the core and its arms, provided other requirements on parameters such as quantum yield and volume fraction are satisfied. (a) Room temperature Variable Stripe Length (VSL) measurements of the same CdSe seeded CdS tetrapod-silica film from Figure 2(a). The dotted lines are a fit using a 1D amplifier model which takes into account saturation due to Auger recombination processes. (b) Lifetime measurements of a dilute solution of CdSe seeded CdS tetrapods with a CdSe core diameter of ∼5.4 nm and a CdS arm length of ∼21 nm under excitation intensities corresponding to an average of 1 e-h pair (black), 2 e-h pairs (red) and 3 e-h pairs (green) per particle. A single exciton and biexciton lifetime of ∼15.7 ns and ∼1.6 ns were extracted from the lifetime measurements respectively. where I, g, and z are the ASE intensity, modal gain coefficient and excitation stripe length respectively and z0 accommodates for possible pump beam inhomogeneity and delayed ASE onset. The parameter lα is defined as the modal gain lifetime of the NCs in the film multiplied by the speed of light within the gain medium and accounts for saturation in the ASE intensity.6 Extraction of the parameters g and lα from the fit yielded values of ∼151 cm−1 and 0.05 cm respectively, the latter of which translates to a modal gain lifetime of ∼3 ps. While the VSL obtained gain value is not appreciably higher than previous reports on NC-based composites,15, 24 presumably due to the extremely low volume fraction of CdSe cores in the film, the fitted gain lifetime is unlikely due to fast Auger recombination processes as ascribed to by previous studies on spherical CdS/ZnS NCs6 since it is expected that the CdSe core electron wavefunction can delocalize extensively into the four CdS arms, thereby reducing the Auger recombination rate. This supposition was experimentally verified by biexciton lifetime measurements as illustrated in Figure 3(b), where a remarkably long average biexciton lifetime of ∼1.6 ns was determined, though comparatively smaller than the 10 ns biexciton lifetime of spherical “giant” shell CdSe/CdS NCs previously reported by Klimov et. al.25 The observed short modal gain lifetimes is therefore attributed to depletion of electron-hole pair density due to stimulated emission, which is a well-known mechanism in VSL measurements on semiconductors. Since the threshold condition for ASE in NCs is inversely proportional to their Auger-dominated biexciton lifetime,5 the long-lived biexcitonic lifetimes in tetrapods, indicative of a suppressed Auger recombination rate, should allow ASE to be more readily achieved. Indeed, this is consistent with our observed low thresholds for the biexcitonic ASE transitions in the tetrapod-silica composite films. Given the structural robustness of the nanotetrapod-silica composites, we examined their capacity to exhibit multiexcitonic ASE, which can only take place under sufficiently intense optical excitation such that the excited energy states higher than 1Se are simultaneously populated.26, 27 As shown in Figure 4(a), a second narrow emission peak at 619 nm emerges at larger pump intensities, which may be contrasted with the lower energy biexcitonic ASE transition centered at 661 nm. Nearly identical energy spacings between the two ASE transitions were found in films of the same tetrapods but with different thicknesses, thus excluding the possibility that the higher energy ASE transition is due to a broad gain profile coupled to different waveguide modes.24 An analysis of its PL intensity as a function of pump power revealed that the threshold for the higher energy ASE transition was approximately 2 times that of the biexcitonic ASE. A similar observation was previously reported for CdSe/ZnS NCs in titania matrices where a higher energy ASE peak at large pump intensities was ascribed to stimulated emission originating from a non-degenerate CdSe 1Pe-1Ph transition.27 Taken together, these findings suggest that the higher energy ASE transition is due to multiexcitonic states in the CdSe core of the tetrapod with a threshold sufficiently low (∼171 μJ cm−2) to warrant dual wavelength optical in a As a with our nanotetrapod-silica we evaluated the optical gain performance of CdSe seeded CdS nanorods with core and rod-like shell dimensions similar to of the tetrapods and their The nanorods were incorporated into a thin film silica matrix using the same for tetrapods, and it was that parameters such as NC volume film thickness and roughness were between both and tetrapod-silica composites. As shown in Figure both biexcitonic and multiexcitonic ASE were observed at higher pump intensities, with an energy of and an in thresholds between the two ASE consistent with of the tetrapods in Figure The room temperature ASE thresholds of the two heterostructured composite films are in Figure It is readily seen that the tetrapods have a lower threshold than of their which is likely attributed to a larger absorption and a multiexcitonic lifetime due to the more extensive leakage of the electron wavefunction to four CdS arms as with a single rod-like CdS shell in the case of (a) PL intensity of a CdSe seeded CdS tetrapod-silica film under relatively high excitation pump in on the spectral featuring a higher energy ASE (b) PL intensity of a CdSe seeded CdS film under high excitation pump (c) of the thresholds obtained for ASE from the biexcitonic and multiexcitonic transitions in tetrapods and respectively. we have presented the fabrication of heterostructured nanotetrapod-silica composite films via the incorporation of monodisperse colloidal CdSe seeded CdS tetrapods into a sol-gel derived silica matrix using a relatively process. These tetrapod-silica composites were found to exhibit very low stable room temperature biexcitonic and multiexcitonic ASE from the CdSe core for Type I tetrapods while ASE at wavelengths from the CdS arms were for tetrapods of Type II with CdSe seeded CdS nanorods in silica films suggest that within the device parameters the tetrapod-silica composite films are more as NC optical amplifiers. of the NC volume fraction and film optical may well allow for the of low heterostructured NC of simultaneously at multiple wavelengths over a large spectral range from a single thus a unique of optical gain tetrapod Nearly monodisperse NCs were synthesized to a previously reported and used as for tetrapod via the seeded growth with a mixture comprising of cadmium oxide oxide (TOPO) oleic acid (OA) and hexylphosphonic acid (HPA) were and to °C under × NCs in was to a mixture of and trioctylphosphine (TOP) and was at 50 °C for in order to off the was to °C and the temperature was allowed to to °C the mixture of and CdSe was The temperature was allowed to to °C and the CdS arms were at this temperature for between 5 and min depending on the arm The as-synthesized tetrapods were processed by of in and in The resulting tetrapods were overcoated with ZnS using an and processed by of in and in film The route as silica was from a previously reported method6 with slight ZnS overcoated CdSe seeded CdS tetrapods were with 5-amino-1-pentanol in This afforded a clear solution of NCs into which and was and for The was allowed to a and spin-coated onto a glass yielding a smooth The film was on a for 5 min at 150 which or obvious of the film was observed. Optical The excitation was a amplifier nm) that was seeded by a oscillator nm wavelength laser pulses were obtained with a and focused by a cylindrical lens (with length to a stripe dimensions ∼5.5 mm × 0.5 The were to the excitation The emission from the film was collected in a by a pair of focused onto an optical fiber coupled to a spectrometer and detected by a charge coupled device The photoluminescence was detected by an system which has an of at the time window of ps. The measurements were typically with a time window of ns and a time of ∼300 ps. gain Room temperature modal gain measurements were using a VSL The excitation stripe and emission was the same as The excitation stripe length was varied an by a which was at the line of the cylindrical This work was by and a