We experimentally investigate PbS nanocrystal (NC) photoluminescence (PL) coupled to all-integrated Si-based ring resonators and waveguides at telecom wavelengths. Dissolving the NCs into Novolak polymer significantly improves their stability in ambient atmosphere. Polymer-NC blends of various NC concentrations can be applied to and removed from the same device. For NC concentrations up to 4vol%, the spontaneous emission rate into ring-resonator modes is enhanced by a factor of ~13 with respect to that into a straight waveguide. The PL intensity shows a linear dependence on the excitation intensity up to 1.64kW/cm2 and stable quality factors of ~2500.
©2013 Optical Society of America
For the near-infrared spectral range, lead-based colloidal nanocrystal quantum dots (NCs) are of considerable interest due to their unique optical and electronic properties [1–3]. Their absorption and emission wavelengths depend on the dot size, which can be easily controlled during chemical synthesis. Therefore, they are very attractive as building blocks for nanophotonic applications at telecommunication wavelengths. Especially Lead Sulfide (PbS) NCs have found a wide range of potential applications such as laser modelocking , surface emitting lasers , polymer strip-loaded plasmonic waveguides  and solar cells .
A convenient way to manufacture stable and reproducible quantum dot based devices is to introduce the NCs into polymers like PMMA, SU-8 or PFCB . In fact, by the choice of proper ligands, the quantum dots can be mixed into various solvents of organic polymers where, after solvent evaporation, the polymer acts as a matrix in which the NCs are homogeneously distributed [9, 10]. The choice of a polymer structurable by photo-, electron-beam- or soft-lithography opens the possibility to fabricate photonic devices from this compound material. Furthermore, inorganic surface capping of colloidal nanocrystals was achieved recently, allowing the encapsulation of NCs into an amorphous As2S3 matrix, which resulted in an all-inorganic thin film with stable infrared luminescence in the telecom region .
Previous studies of PbS NCs coupled to silicon based photonic crystal microcavities have either relied on free-space or fiber coupled micro-photoluminescence methods to pump and collect the emission from high-Q cavities [12–15]. While these types of experimental setups are very well suited to enable optical experiments on single NCs, for technical applications stable and robust integrated optical systems are preferable.
Therefore, in our work we concentrate on the implementation of colloidal PbS NCs emitting in the telecom wavelength regime into all integrated Si based photonic building blocks. PbS NCs dissolved into Novolak polymer host material are coupled to all-integrated Si-based ring resonator (RR) - bus waveguide (WVG) devices. We perform studies on the stability of the emission and investigate for the same device the cavity induced spontaneous emission rate (SER) enhancement relative to the SER of a straight WVG as a function of the NC concentration of the polymer-NC blends. We also demonstrate the linear dependence of the device performance on the excitation intensity up to a continuous wave pumping intensity of 1.64kW/cm2 without degradation of the cavity Q factor or the polymer host material.
2. Experimental characterization
Commercially available PbS colloidal nanocrystals (NCs) from Evident Technologies (EviDot core LNR03LPB) with a diameter of 5.1nm and a photoluminescence (PL) peak emission at 1385nm, dissolved in toluene (40mg/ml) are used as optically active material for the polymer-NC blend. The inhomogeneously broadened PL spectrum of a spin cast pure NC film after toluene evaporation is shown in Fig. 1(a) by the dotted line. As polymer host material Allresist AR-N 7700.18 (Novolak based negative electron beam resist) is used and NC-polymer blends containing different nanocrystal concentrations varying from 0.5vol% to 10vol% are prepared and analyzed.
For testing the degradation of the NC luminescence, the polymer-NC blend is drop cast onto a silicon substrate and cured on a hotplate at 75°C for 1min. The PL spectra of the NC-polymer films are measured by exciting the blends with a frequency doubled Nd:YAG laser at 532nm wavelength. In Fig. 1(a), the normalized PL spectra of the pure NCs (dotted line) and the polymer-NC blend (full lines with superimposed symbols) are shown for different aging steps of the blend, indicating that the emission spectrum of the NCs remains unaffected after their introduction into the polymer host and at least over a period of 63 days. The two dips in the PL spectra at around 1365nm and 1385nm are due to atmospheric absorption. For a quantitative monitoring of the PL intensity's temporal evolution, it is essential to keep both the number of NCs within the detection volume of the PL setup and the excitation intensity constant for measurements at subsequent observation intervals. In an integrated optical setup as sketched in Fig. 1(c), in which a drop of polymer-NC solution is deposited into the opening of an SU-8 passivation layer, only a fixed number of NCs in close vicinity to the WVG and RR can emit their PL radiation into them. In Fig. 1(b), the averaged PL intensity of the resonance peaks measured at the end of the WVG is plotted for several observation intervals. Figure 1(b) shows the excellent stability of the composite material's PL emission intensity over a period of 37 days. Because of this temporal and spectral stability shown in Figs. 1(a) and 1(b), the Novolak-NC composite material was used as active medium in the experiments discussed in the following.
For defining RRs coupled to bus WVGs, the 220nm thick Silicon on insulator (SOI) wafer was processed into 600nm wide and 220nm high Si monomode waveguides by Amo Ltd. using electron beam lithography and ICP-RIE etching. The 20µm diameter RRs are separated from the bus WVGs by a 150 nm wide gap.
The samples are first passivated by spin casting a 500nm thick MicroChem SU-8 2000.5 resist. Next, UV lithography was used to open L = 60µm≈2πr and W = 30µm wide windows centered around the RRs, where r is the radius of the ring [see Fig. 1(c) for a sketch]. This relation between r and L was chosen to ensure equal interaction surfaces of bus WVGs and RRs with the Novolak-NC blend. The windows were filled by drop casting the blend and subsequent solvent evaporation on a hotplate at 75°C for 1 minute.
The refractive indices of the Novolak AR-N 7700.18 (n = 1.575) and the SU-8 2000.5 (n = 1.572) differ only slightly [16,17] around 1.5µm wavelength, ensuring no significant mode distortion in the underlying bus waveguide at their interface. Additionally, the Novolak film is easily removed by Allresist AR 300-70 remover, whereas the hardened SU-8 remains unaffected. Compared to previously reported approaches , this high selectivity of the remover allows us to completely replace Novolak-NC films with different NC concentration in a simple way without affecting the optical properties of RRs and bus WVGs. Therefore, the same resonator can be used for analyzing the spontaneous emission rate (SER) enhancement with respect to the NC concentration of the Novolak film.
Transmission measurements are performed using a fiber-coupled continuous wave InGaAsP laser source tunable from 1460nm to 1580nm with an output power adjusted to 2mW. To inject and collect light, polarization maintaining single mode tapered fibers are used. Light detection is performed by a fiber-coupled InGaAs detector.
For investigating the PL emitted into the resonator and waveguide modes, the excitation lasers (532nm or 657nm wavelength) are focused onto the devices. PL emission is collected at the end of the bus WVG and dispersed by a fiber-coupled monochromator (resolution 0.2nm). A homogeneous distribution of the NCs in the host material was verified by removing and reapplying each mixture several times on the device with virtually identical results in transmission and PL.
Figure 2(a) shows a TM polarized photoluminescence spectrum of the RR – bus WVG device covered with a Novolak film containing 4vol% PbS NCs .For obtaining the results shown in Fig. 2(a), an excitation laser (532nm wavelength) with 80 µm spot diameter (larger than the SU-8 window dimensions) was used. Thus, NCs all over the window area are excited and can emit PL radiation into the continuous bus-WVG or discrete RR mode spectrum, provided that they are located within the respective region of the evanescent electric field surrounding the Si WVGs. A contour plot of the in-plane component of the waveguide mode's electric field obtained by a 3D Beam Propagation Method (BPM) simulation is shown in Fig. 2(c). Accordingly, the PL spectrum measured at the end facet of the bus WVG consists of a broadband PL signal superimposed by narrow peaks of intensity [see Fig. 2(a)]. Figure 2(b) shows that in transmission and emission experiments the resonances are observed at identical wavelengths and with similar peak widths.
3. Results and analysis
In Fig. 3(a), the ratio between the out-coupled RR peak PL Ires and bus WVG PL intensity Iwvg [Fig. 2(a)] from the PL spectrum measured at the output facet of the bus WVG is labeled as intensity ratio R = Ires/Iwvg (averaged over the wavelength region containing the 5 most intense resonances) and shown for NC concentrations in the range between 0.5 and 10vol%. For these measurements a constant excitation intensity of 32W/cm2 was used.
As the circumference of the RR and the section of the bus WVG coupled to NCs are of equal length and since the RR and the bus WVG have the same cross section, statistically an equal number of NCs are coupled to RR and bus WVG. Due to RR parasitic losses, only a fraction η of the photons emitted per time into the RR is coupled out via the bus WVG, where according to Ref . the collection efficiency η can be calculated by η = Qt/Qc. Here, Qt denotes the total Q factor of a resonator coupled to a read out waveguide, 1/Qt = 1/Qc + 1/Qi, Qc describes the coupling between the bus WVG and the RR and Qi is related to the intrinsic and parasitic losses of the RR [20, 21]. Thus, in the presence of losses, the ratio of SERs into the RR at resonance and into the bus WVG is R/η.
From transmission (emission) spectra as shown in Fig. 2(b), Qt can be determined from the resonance lines as ratio between resonance wavelength and full line width at half drop (maximum) after fitting a Lorentzian profile to the resonance line. As shown in Fig. 3(c), similar total Qt factors are observed in TM polarized transmission and emission spectra for all NC concentrations investigated in this work. Assuming therefore that also Qc is equal for transmission and emission, η can be calculated from the Q values obtained from the transmission spectra, from which Qc can be determined using Qc = 2Qt/(1 ± √Tmin) [20,21]. Here, Tmin is the minimum in the normalized transmission spectrum at resonance and the ± corresponds to the over- and under- coupled loading condition . Assuming under-coupled loading, a decomposition of Qt into Qc and Qt as shown in Fig. 3(c) is obtained.
In Fig. 3(b), the SER enhancement factor R/η is shown as a function of the NC concentration in the active material. For NC concentration up to 4vol%, R/η is constant (R/η≈13) within experimental uncertainties, whereas R increases in this concentration region. From Fig. 3(c) it is evident, that the increase of R with the NC concentration is a consequence of increasing the coupling of bus WVG and RR more towards critical coupling, as indicated by the decrease of Qc and a constant Qi. We ascribe the increased coupling to an increase of the composite material's index of refraction with the NC concentration. In the NC concentration range from 4vol% to 10vol%, the coupling is further enhanced; however, the scattering losses due to NC degrade Qi in this concentration range. While these two effects appear to cancel for R between 4vol% and 6vol% and degrade R only between 6vol% and 8vol%, R/η degrades already above 4vol%.
It is well known that the SER of an excited optical emitter coupled to an optical resonator can be enhanced by the Purcell factor [23, 24]. According to its definition, the ratio between the SERs into a RR and a bus WVG mode equals Fres/Fwvg where Fres and Fwvg denote the Purcell factors of RR and bus WVG, respectively, which can be calculated following Ref . For a RR and a bus WVG with equal cross section and equal circumference and length coupled to NCs (as investigated in our work), Fres/Fwvg = (1/π)(λ/neff)(Qi/2πr) results , where λ = 1.45µm, neff = 2.14 is obtained from 3D BPM simulations and Qi is the RR's intrinsic quality factor taken from our experimental analysis. In the calculation, both Fres and Fwvg are spatially averaged over the active emitting regions, however, since these regions have the same dimensions for RR and bus WVG, the corresponding factors cancel in the ratio Fres/Fwvg.
In Fig. 3(b), the ratio Fres/Fwvg is plotted together with the experimentally determined SER ratio R/η vs. the NC concentration. From the similarity of these two traces, we tentatively ascribe the observed enhanced SER into the RR with respect to that into the bus WVG to the Purcell effect. For such an assignment, we additionally have to assume that the PL emission linewidth of single NCs is much smaller than the width of the RR resonance line so that a large number of NCs can emit into the resonance. In this case, the spectral detection bandwidth restricts the number of NCs that contribute to the experimentally observed PL intensity and thus the same number of NCs can contribute to Iwvg and Ires at a given wavelength. If, however, the resonance linewidth becomes smaller than the spectral detection bandwidth, less NCs contribute to the detected RR PL than to the bus WVG PL and, therefore, R/η will be smaller than the ratio of the respective Purcell factors. In our experiments, such a situation is encountered for TE polarization, for which in transmission a ~15 times larger Qt was found than for TM polarization, however, with approximately the same values for η. Therefore, for TE polarization a quantification of the RR and bus WVG SER ratio could not be performed and for an unambiguous identification of the Purcell effect, time-resolved PL emission experiments are required, the results of which are beyond the scope of this paper and will be published elsewhere.
We assign the deviation at concentrations above 4vol% in Fig. 3(b) to the coupling of NC emission from the bus WVG into the RR at the resonance wavelengths, leading to a slight underestimation of Ires in the evaluation of the experimental data, as indicated in Fig. 2(a).
Finally, the PL intensity measured at the output of the bus waveguide was studied as a function of the continuous wave pump laser intensity for the devices containing NC concentrations of 4vol% and 6vol%. In Fig. 4(a), the PL intensity of the resonance peaks (black squares) and the bus WVG (open red squares) are plotted for devices covered by a Novolak film with the optimal NC concentration of 4vol%. Both curves show a linear behavior where the ratio of the respective slopes kres/kwvg equals the reported intensity ratio R of ~2.1. Additionally a high power AlGaInP laser operating at 657nm was focused to a spot size slightly larger than the ring resonator diameter (~20µm), resulting in a negligible fraction of the bus WVG being illuminated. Therefore, Fig. 4(c) contains just the ring resonator peaks and an approximately linear increase of the PL output power with the pumping intensity was also observed [Fig. 4(b)]. For the total range of excitation intensities accessible in this work, a constant slope was found for the PL intensity of the resonance peaks. As neither a degradation of the Q factors nor a saturation of the output power with increasing excitation intensity is observed, we conclude from this measurement that the obtained SER enhancement is stable for the investigated excitation intensities.
Saturation of the spontaneous emission sets in if non-linear processes limit the population of the excited states of the optical transition. For colloidal NCs it is well known that non-radiative Auger recombination is such a process. It virtually limits the number of electrons in excited NC states to one per NC . The corresponding saturation pump intensity Isat required to establish this level of excitation can be estimated by Isat = ħω/(στx). Here, ħω denotes the pump photon energy, σ the absorption cross-section of a single NC (σ~3·10−16cm2 obtained from absorption experiments), and τx the average excited NC state lifetime at small pump intensities, i.e. in the absence of Auger recombination. Thus, an experimental determination of Isat could provide further insight into the observed SER enhancement. However, assuming τx values of several hundreds of picoseconds , Isat results in a value larger than 1MW/cm2. Such pump intensities are more than 500 times larger than the values shown in Fig. 4 and currently beyond our experimental capabilities.
In conclusion, we demonstrate coupling of the emission of PbS NCs into the resonating modes of SOI ring resonators on an all-integrated optoelectronic chip at telecommunication wavelengths. Our device parameters and method of fabrication allow the investigation of the Q factors and SER enhancement of the same device for different NC concentrations. For NC concentrations up to 4vol% an optimal SER enhancement ratio of ~13 was found experimentally in TM polarization, with no degradation of the intrinsic resonator Q factor and in good agreement with theoretically predicted values. As the raw intensity increases with NC concentration, we thus find optimal device performance for a NC concentration of 4vol%. At concentrations of 6vol% and higher, scattering losses start to degrade the Q factor and the SER enhancement ratio of the device. The device also shows a linear dependence on the excitation intensity up to 1.64kW/cm2 without degradation of the Q factor or the polymer host material. Polymer-NC blends show unaltered performance weeks after preparation, indicating the high stability of this composite material. As the emission wavelength of chemically synthesized NCs can be precisely controlled, the working bandwidth of NC based optoelectronic chips can be easily tuned to other wavelength regions within the transparency region of the Si WVGs.
This work was supported by the Austrian Nanoinitiative within the PLATON project cluster under project numbers 819654, 819655, 819656, 834913, and 834924. We also acknowledge W. Heiss and G. Bauer for fruitful discussions.
References and links
1. D. V. Talapin, J. S. Lee, M. V. Kovalenko, and E. V. Shevchenko, “Prospects of colloidal nanocrystals for electronic and optoelectronic applications,” Chem. Rev. 110(1), 389–458 (2010). [CrossRef] [PubMed]
2. P. Guyot-Sionnest, “Electrical transport in colloidal quantum dot films,” J. Phys. Chem. Lett. 3(9), 1169–1175 (2012). [CrossRef]
3. A. Omari, I. Moreels, F. Masia, W. Langbein, P. Borri, D. Van Thourhout, P. Kockaert, and Z. Hens, “Role of interband and photoinduced absorption in the nonlinear refraction and absorption of resonantly excited PbS quantum dots around 1550nm,” Phys. Rev. B 85(11), 115318 (2012). [CrossRef]
4. K. Wundke, S. Pötting, J. Auxier, A. Schülzgen, N. Peyghambarian, and N. F. Borrelli, “PbS quantum-dot-doped glasses for ultrashort-pulse generation,” Appl. Phys. Lett. 76(1), 10–12 (2000). [CrossRef]
5. A. Ishida, Y. Sugiyama, Y. Isaji, K. Kodama, Y. Takano, H. Sakata, M. Rahim, A. Khiar, M. Fill, F. Felder, and H. Zogg, “2W high efficiency PbS mid-infrared surface emitting laser,” Appl. Phys. Lett. 99(12), 121109 (2011). [CrossRef]
6. J. Grandidier, G. C. des Francs, S. Massenot, A. Bouhelier, L. Markey, J. C. Weeber, C. Finot, and A. Dereux, “Gain-assisted propagation in a plasmonic waveguide at telecom wavelength,” Nano Lett. 9(8), 2935–2939 (2009). [CrossRef] [PubMed]
7. K. Szendrei, W. Gomulya, M. Yarema, W. Heiss, and M. A. Loi, “PbS nanocrystal solar cells with high efficiency and fill factor,” Appl. Phys. Lett. 97(20), 203501 (2010). [CrossRef]
8. Y. Olsson, G. Chen, R. Rapaport, D. Fuchs, V. C. Sundar, J. S. Steckel, M. G. Bawendi, A. Aharoni, and U. Banin, “Fabrication and optical properties of polymeric waveguides containing nanocrystalline quantum dots,” Appl. Phys. Lett. 85(19), 4469–4471 (2004). [CrossRef]
10. L. Bakueva, S. Musikhin, M. A. Hines, T.-W. F. Chang, M. Tzolov, G. D. Scholes, and E. H. Sargent, “Size-tunable infrared (1000–1600 nm) electroluminescence from PbS quantum-dot nanocrystals in a semiconducting polymer,” Appl. Phys. Lett. 82(17), 2895–2897 (2003). [CrossRef]
11. M. V. Kovalenko, R. D. Schaller, D. Jarzab, M. A. Loi, and D. V. Talapin, “Inorganically functionalized PbS-CdS colloidal nanocrystals: integration into amorphous chalcogenide glass and luminescent properties,” J. Am. Chem. Soc. 134(5), 2457–2460 (2012). [CrossRef] [PubMed]
12. D. F. Dorfner, T. Hürlimann, G. Abstreiter, and J. J. Finley, “Silicon photonic crystal nanostructures for refractive index sensing,” Appl. Phys. Lett. 91, 233111 (2007). [CrossRef]
13. R. Bose, X. Yang, R. Chatterjee, J. Gao, and C. W. Wong, “Weak coupling interactions of colloidal lead sulphide nanocrystals with silicon photonic crystal nanocavities near 1.55µm at room temperature,” Appl. Phys. Lett. 90(11), 111117 (2007). [CrossRef]
14. M. T. Rakher, R. Bose, C. W. Wong, and K. Srinivasan, “Spectroscopy of 1.55 μm PbS quantum dots on Si photonic crystal cavities with a fiber taper waveguide,” Appl. Phys. Lett. 96(16), 161108 (2010). [CrossRef]
15. Z. Wu, Z. Mi, P. Bhattacharya, T. Zhu, and J. Xu, “Enhanced spontaneous emission at 1.55 μm from colloidal PbSe quantum dots in a Si photonic crystal microcavity,” Appl. Phys. Lett. 90(17), 171105 (2007). [CrossRef]
16. W. H. Wong, J. Zhou, and E. Y. B. Pun, “Low-loss polymeric optical waveguides using electron-beam direct writing,” Appl. Phys. Lett. 78(15), 2110–2112 (2001). [CrossRef]
18. I. Moreels, B. De Geyter, D. Van Thourhout, and Z. Hens, “Transmission of a quantum-dot-silicon-on-insulator hybrid notch filter,” J. Opt. Soc. Am. B 26(6), 1243–1247 (2009). [CrossRef]
19. C. Grillet, C. Monat, C. L. Smith, B. J. Eggleton, D. J. Moss, S. Frédérick, D. Dalacu, P. J. Poole, J. Lapointe, G. Aers, and R. L. Williams, “Nanowire coupling to photonic crystal nanocavities for single photon sources,” Opt. Express 15(3), 1267–1276 (2007). [CrossRef] [PubMed]
21. R. Guider, N. Daldosso, A. Pitanti, E. Jordana, J. M. Fedeli, and L. Pavesi, “NanoSi low loss horizontal slot waveguides coupled to high Q ring resonators,” Opt. Express 17(23), 20762–20770 (2009). [CrossRef] [PubMed]
22. P. Barclay, K. Srinivasan, and O. Painter, “Nonlinear response of silicon photonic crystal microresonators excited via an integrated waveguide and fiber taper,” Opt. Express 13(3), 801–820 (2005). [CrossRef] [PubMed]
23. E. M. Purcell, “Spontaneous emission probabilities at radio frequencies,” Phys. Rev. 69, 681 (1946).
24. L. C. Andreani, J. M. Gérard, and G. Panzarini, “Strong-coupling regime for quantum boxes in pillar microcavities: Theory,” Phys. Rev. B 60(19), 13276–13279 (1999). [CrossRef]
25. R. S. Tummidi, R. M. Pafchek, K. Kim, and T. L. Koch, “Modification of spontaneous emission rates in shallow ridge 8.3 nm erbium doped silica slot waveguides,” in Proceedings of 6th IEEE International Conference on Group IV Photonics (San Francisco, 2009), pp226–228. [CrossRef]
26. V. I. Klimov, S. A. Ivanov, J. Nanda, M. Achermann, I. Bezel, J. A. McGuire, and A. Piryatinski, “Single-exciton optical gain in semiconductor nanocrystals,” Nature 447(7143), 441–446 (2007). [CrossRef] [PubMed]
27. V. Rinnerbauer, H. J. Egelhaaf, K. Hingerl, S. Werner, T. Warming, A. Hoﬀmann, M. Kovalenko, W. Heiss, G. Hesser, and F. Schaﬄer, “Energy transfer in close-packed PbS nanocrystal films,” Phys. Rev. B 77(8), 085322 (2008). [CrossRef]