*Corresponding author(s)
I. Introduction
Semiconductor nanocrystals exhibit peculiar properties, which are totally different from their bulk counterparts [1-3]. The size-dependent characteristics make nanocrystalline semiconductors promising materials for optoelectronic applications [5]. Despite of the fact that doped semiconductor nanocrystals were studied early in 1980’s [6], they actually did not attract attention until the publication of ref. [7] in 1994. This paper reported on Mn2+ doped ZnS nanocrystals with remarkable luminescent efficiency of 18 % and short lifetime of the order of nanoseconds. As proposed in this work, the d – electrons of Mn2+ acts as a luminescent center, and the strong interaction between these d – electrons and s-p states of the host matrix provides a highly effective energy transfer path that results in the lifetime shortening and emission efficiency increasing. This publication has inspired many other papers that studied various doped nanocrystals [8-15]. The report additionally has been written in the recent luminescent spectroscopy textbooks and is considered as the start of the research on doped semiconductor nanocrystals [16, 17]. Besides ref. [7], some other groups also observed the life time shortening of the Mn2+ emission due to strong quantum confinement in nano-scaled crystals, for example, ref. [15] and ref. [9] reported ns lifetime in ZnS:Mn2+ nanoclusters.
On the opposite viewpoint, ref. [18, 19] showed that the luminescent lifetime in ZnS:Mn2+ nanocrystals was around 1.9 ms, which is comparable to that in bulk ZnS. It also raised some questions about the work of ref [7]. For example, the hybridization of sp-d states should result in the shift of Mn2+ energy levels, which was not observed [4, 7]. Moreover, the assumption about the hybridization as well as the reduction of forbidden transitions in such doped nanocrystals needs more concrete evidence from a quantitative theoretical model. More importantly, the details about lifetime measurement were still lacking. Ref. [18] came to the conclusion that the lifetime shortening of Mn2+ emission with respect to crystal size reduction is incorrect. The as-reported nanosecond decay time most likely originated from the tail of broad defect-related (blue) emission that overlaps the Mn2+ (orange) emission at around 590 nm. From the lifetime measurement and time-resolved spectroscopy, ref. [18] showed that the Mn2+ emission possesses decay time in ms range, comparable to that in bulk ZnS:Mn2+.
In this work, we synthesized Mn2+ doped ZnS nanocrystals with average dimensions 2 – 4 nm by a co-precipitation method. From both decay time and time-resolved measurements we conclude that there is no lifetime shortening of orange emission from Mn2+ incorporated in ZnS nanocrystals in comparison with the bulk counterpart and the previously reported ns lifetime comes from the tail of ZnS related photoluminescence. Furthermore, we report on the dependence of Mn2+ emission lifetime on excitation wavelength. It is shorter for photon energies smaller than the ZnS bandgap.
II. Experimental
Zn(CH3COO)2 and MnCl2 were used as precursors while Na2S was the sulfur source. Initially, 10 ml Zn(CH3COO)2 1M was dissolved with different amounts of MnCl2 0.1M (0.5 – 10 ml) and 10 g Na(PO3)n (sol stabilizer). Deionized water was then added to the reaction flask and the mixture was stirred for 30 minutes at room temperature. After that 10 ml Na2S 1M was injected into the flask and the solution was continuously stirred for 30 minutes. We received white suspensions, which subsequently were separated out of the solution by extensive centrifugation (8000 – 12000 rpm), and then rinsed many times with deionized water. The washed powder was finally dried and sintered in inert gas (Argon 99.99%) environment.
The emission spectra and the excitation spectra of the samples were measured on a Jobin-Yvon Nanolog spectrofluorometer equipped with a photomultiplier detector. A Xenon (Xe) lamp was used as an excitation source. The decay of phosphorescence of the samples was also recorded using either 370 nm excitation with a NanoLED at 0.5 MHz (Jobin Yvon Nanolog system at Hanoi University of Science and Technology, Vietnam) or 1 MHz (Jobin Yvon Nanolog system at Beijing University, China), 337 nm excitation with a pulsed N2-laser (1-2 Hz) (Ghent University, Belgium), or a flash Xenon lamp (30 Hz) (Beijing University). For the measurements at low temperature, the 325 nm line of a helium-cadmium laser was used as excitation source for recording PL spectra. The laser beam was focused through a quartz lens into a spot of about 100 µm diameter on the sample, fixed in the cold-finger of a cryostat. The excitation density was kept lower than 1 W/cm2 to avoid heating and possible saturation effects [20].
The temperature dependence of PL intensities was investigated at various temperatures in the range from 10 to 300 K using a closed-cycle helium pumping system. The cooled Roper Scientific CCD combined with Acton Research Spectrometer allowed instant recording of the spectra at any moment.
III. Results and Discussion
In order to examine the phase composition and crystallography of samples, an X-ray Diffraction (XRD) pattern was taken. As shown in Fig. 1a, all peaks are well indexed as cubic zinc blende modification of ZnS structure with three typical diffracted planes of (111), (220) and (311). No peaks of any other phases or impurities were detected. Weak XRD peaks indicate that ZnS nanoparticles are not well crystallized. The dimension of the nanocrystals calculated quantitatively by the full-width at half-maximum (FWHM) of diffraction peaks and Scherrer equation [21] is about 2-4 nm.
The nanocrystals’ size was verified again by direct measurement from TEM images, Fig. 1b. Before measuring, the samples were dispersed in ethanol and brought to the copper grid of the TEM. The TEM results show that the crystals exhibit nearly spherical shape and quite uniform dimension of 4 – 10 nm on average. The discrepancy in size of ZnS nanoparticle between XRD and TEM measurements can be attributed to the fact that the dimension measured in TEM images is for agglomerates of several crystals.
The optical properties of ZnS:Mn2+ were examined under excitation with a monochromated Xe lamp. Figure 2a shows the emission (PL) and excitation (PLE) spectra of ZnS nanocrystals doped with 0.5 at% Mn2+. The emission spectrum is characterized by two peaks at 417 (blue) and 589 nm (orange). While the blue emission is known to originate from intrinsic defects in ZnS such as sulfur vacancies on the nanocrystal surface or zinc vacancies inside the structure [22, 23], the orange emission is ascribed to the 4T1 – 6A1 transition of Mn2+ ions in the ZnS host matrix [24-26]. Despite of the fact that the concentration of surface defects is very large due to the high surface to volume ratio in nanoscale structures, the intensity of such orange emission is dominant, even at rather low concentration of Mn2+, which suggests the effective transfer of electron-hole pairs into Mn2+ states, and subsequently the fast radiative recombination of d – electrons. According to previous studies [27, 28], if the Mn2+ ions are located on the surface of the ZnS crystal, ultraviolet emission at around 350 nm is observed. Therefore, the strong orange emission at 589 nm is indicative of the successful incorporation of Mn2+ into the ZnS matrix. The PLE spectrum recorded when monitoring the 589 nm emission revealed two peaks at 341 nm and 282 nm whereas very similar peaks at 340 and 282 nm were observed for the 417 nm emission. Since the position of PLE peak indicates the band edge of the host (in direct bandgap semiconductor), the PLE peaks at 342 and 341 nm are in the order of the room temperature band gap of cubic ZnS crystal (~335 nm). The PLE peak at 282 nm presents a strong blue shift to higher energy and can be explained due to the effect of strong quantum confinement in small ZnS nanocrystals [29]. The fact that the orange emission can be excited through host excitation indicates that energy transfer from ZnS to Mn2+ is taking place and confirms that Mn2+ is located within the ZnS nanoparticles.
Additionally, when the dopant concentration increases, the intensity of the Mn2+ emission increases dramatically at the same time with the extensive reduction of the blue emission, see Figure 2b. As mentioned earlier, since the blue emission relates to the intrinsic defects of ZnS, one may conclude that the Mn2+ doping eliminates these defects and makes ZnS a better matrix for luminescent materials. Moreover, the PLE peaks experience blue shifts from 341 to 322 nm and from 282 to 276 nm that indicate a band gap enlargement of the ZnS particles. When Mn2+ ions are doped into the ZnS matrix, they occupy Zn2+ lattice sites [30, 31]. Because of the difference in radius between Mn2+ (0.8 Å) and Zn2+ (0.74 Å) ions, the incorporation creates large lattice misfit (or high strain energy) which favors the formation of small grains. Therefore it is expected that the ZnS crystal size is smaller for higher doping dose. In other words, the quantum confinement effect becomes stronger for higher Mn2+ concentration, which results in a larger band gap of ZnS nanocrystals, or a blue shift of the PLE peaks [7, 32]. The band gap of bulk MnS is slightly smaller than that of ZnS, about 3.1 eV and 3.7 eV respectively [33, 34]. Therefore, a slight decrease in band gap would be expected upon Mn-doping in case the crystal size remains the same [35] In contrast, the orange emission experiences a red shift from 589 to 594 nm as seen in Fig. 2 as a result of the enhancement of electron-phonon coupling in smaller crystals [36].
The dependence of luminescent intensity on doping concentration was investigated in more detail. Figure 3 shows the PL spectra of ZnS:Mn2+ nanocrystals with different Mn2+ doping concentrations (from 0.5 to 12 at%). For ZnS without doping, we observe only one broad emission peak at 428 nm that is ascribed to intrinsic defects of ZnS. When Mn2+ ions were doped into the crystal, even with a small amount (e.g. 0.5 at%), there appears a new orange emission at 589 nm with considerable intensity and the blue emission intensity gradually decreases simultaneously. Probably, these Mn2+ ions eliminate Zn vacancies, which are the origin of blue emission. The intensity of the orange peak continuously increases with the increasing of doping concentration up to 5 at%, and then quenches out at higher concentrations [9, 37]. This luminescent quenching phenomenon is assigned to the formation of Mn2+ clusters at high dopant concentration since the excitation energy will be transferred non-radiatively among these clusters [38, 39].
We investigated the luminescent lifetime of ZnS:Mn2+ nanocrystals by recording the luminescent decay spectra at room temperature for 417 and 589 nm emissions from two samples, S1 and S2, with different excitation sources at different modulation frequencies. While sample S1 showed both blue and orange bands, sample S2 showed strong orange emission but only a very weak blue band (figure 4a). The decay curve for the blue emission from S1 under 370 nm and 0.5 MHz excitation (figure 4b) showed a multi-exponential decay with a fast initial decay (τ ≈ 18 ns) and a slower tail (τ ≈ 155 ns). This luminescent lifetime is similar to that found in previous studies [40, 41] and does not change significantly for S2 (i.e. 9 ns and 140 ns, respectively).
The decay time measurement for 589 nm emission with the same excitation source (figure 4c) also showed a multi-exponential decay. After the fast initial decay (τ ≈ tens of ns), the slower tail (τ ≈ few hundred ns) was present at high frequency excitations, 0.5 MHz and 1 MHz. This nanoscale lifetime is similar to what has been reported in ref. [7] and ref. [42]. Nonetheless, with low frequency modulation source, 1 Hz (figure 4d) and 30 Hz, the measurements presented a decay time of about 3.6 – 4.3 ms and a faster decay of 500 µs for both samples. It is obvious that such long decay constants can only be measured upon excitation at (very) low frequency, and not with an excitation in the MHz range. The decay times, i.e both fast and slow components which were measured with the different systems, are tabulated in table 1. It is worth noting that in our experiments (measurements) no fitting was performed to remove other over overlapping contributions.
Because of the similarity of the lifetime of the fast component of Mn2+ emission and the lifetime of the ZnS-related emission, time-resolved spectral measurements were carried out in order to verify if the ns decay time comes from the tail of ZnS emission which overlaps with the orange band. The measurement also helps to investigate if ms lifetime is from some large particles in the samples which still possess bulk properties. The time-resolved emission spectra of sample S1 under 337 nm excitation for different time delays but with the same integration time of 50 ns are shown in Fig. 5a, b and c. Shortly after the excitation pulse, only the blue peak is distinguishable but it does not completely drop to zero at longer wavelengths where the orange emission is expected. It is also seen that only after 4 microseconds the orange band can be clearly differentiated from the tail of the blue band. A very similar result has previously been reported in ref. [41]. The negligible fraction of orange emission intensity at the beginning suggests its long decay time compared to ns lifetime of the blue band. This conclusively indicates that the ns component determined from the decay curve measured at around 589 nm comes from the tail of the blue band and the orange band has a lifetime of the order of ms.
Remarkably, the lifetime of the orange emission of S1 sample shows a dependence on the excitation wavelength as presented in Fig. 5d. The intensity of the Mn2+ band drops faster when the wavelength of the excitation source increases from 337 nm to 370 nm for the first 100 microseconds. Three decay time constants of 70, 550, and 3650 µs were determined. In comparison with the excitation at an energy about the ZnS bandgap (337 nm), excitation at an energy smaller than the bandgap (370 nm) causes the appearance of the faster decay component of tens of µs. This can be understood when we consider the excitation mechanism in the ZnS:Mn system. The absorbed energy can excite d electrons of Mn2+ ions to their own excited state (direct excitation) or create free electrons in the conduction band of the host (indirect excitation), which then can be trapped by Mn2+ ions, when the photon energy is smaller or about ZnS bandgap, respectively. In both cases, the excited electrons relax to the ground state of Mn2+ ion after all, resulting in the orange emission. As the transfer process is typically very fast (on the picoseconds time scale, or at least < ns), the longer decay time at lower excitation energy (370 nm) observed in our experiment can be interpreted as due to the trap of electrons at some intermediate trapping states that available in the ZnS nanocrystals.
IV. Conclusion
In summary, in this work we report on the synthesis of ZnS:Mn2+ nanocrystals by the co-precipitation method and the excitation energy dependence of the lifetime of Mn2+ luminescence. The strong orange emission is indicative of successful incorporation of Mn2+ ions into the ZnS matrix. The decay spectra of 589 nm emissions reveal both very fast luminescent lifetimes of tens to few hundred ns and slow decays of few tens of microsecond to few ms. The fast lifetime is attributed to the tail of the blue emission as suggested from time-resolved measurements. This explanation, which can only be measured by probing the changes of the emission spectrum during the decay, clarifies a number of seemingly contradictory results published previously. The lifetime of orange emission can be one order of magnitude shorter when the excitation wavelength increases from 337 nm to 370 nm, changing the excitation from indirect to direct (to Mn2+ ions) mechanism.
Acknowledgements
This work was financially supported by the VLIR’s Own Initiatives’ Programmer, project ZEIN2010RIP07 and the National Application Oriented Basic Research Program, Project No. DTDL.052011-NCCB.
References
[1] R. Rosetti, R. Hull, J. M. Gibson and L. E. Brus, J. Chem. Phys. 82 (1985) 552-159.
[2] L. Brus, J. Phys. Chem. 90 (1986) 2555-2560.
[3] A. Henglein, Chem. Rev. 89 (1989) 1861-1873.
[4] R. N. Bhargava, J. Lumin. 70 (1996) 85-94.
[5] R. K. Chandrakar, R.N. Baghel, V.K. Chandra, B.P. Chandra, Superlattices Microstruct. 86 (2015) 256-269.
[6] W. G. Becker and A. J. Bard, J. Phys. Chem. 87 (1983) 4888-4893.
[7] R.N. Bhargava, D. Gallagher, X. Hong and A. Nurmikko, Phys. Rev. Lett. 72 (1994) 416-419.
[8] I. Yu, T. Isobe and M. Senna, J. Phys. Chem. Solids 57 (1996) 373-379.
[9] K. Sooklal, B. Cullum, S. Angel and C. Murphy, J. Phys. Chem. 100 (1996) 4551-4555.
[10] J. Huang, Y. Yang, S. Xue, B. Yang, S. Liu and J. Shen, Appl. Phys. Lett. 70 (1997) 2335-2337.
[11] M. A. Chamrro, V. Voliotis, R. Grousson, P. Lavallard, T. Gacoin and R. Cases, J. Cryst. Growth 159 (1996) 853-856.
[12] E. T. Goldburt, B. Kuljarni, R. Bhargava and M. Libera, Mat. Res. Soc. Symp. Proc. 424 (1997) 441-445.
[13] Y. L. Soo, Z. H. Ming, S. W. Hunag, Y.H.Kao, R. N. Bhargava and D. Gallagher, Phys. Rev. B 50 (1994) 7602-7607.
[14] H. Ito, T. Takano, T. Kuroda, F. Minami and H. Akinaga, J. Lumin. 72-74 (1997) 342-343.
[15] J. Yu, H. Liu, Y. Wang, F. E. Fernandez, W. Jia, L.Sun, C. Jin, D. Li, J. Liu and S. Huang, Opt. Lett. 22 (1997) 913-915.
[16] K. Yan, C. Duan, Y. Ma, S. Xia and J. C. Krupa, Phys. Rev. B 58 (1998) 13585-13589.
[17] F. Jain and W. Huang, J. Appl. Phys. 85 (1999) 2706-2712.
[18] A. A. Bol and A. Meijerink, J. Lumin. 87-89 (2000) 315-318.
[19] S. G. Emma, R. Laura, G. G. Santiago, T. F. A. Maria, M. C. F. Jose and S. M. Alfredo, Nanoscale 5 (2013) 9156-9161.
[20] H.F. Sijbom, J.J. Joos, L.I.D.J. Martin, K. V. Eeckhout, D. Poelman, P.F. Smet, ECS J. Solid State Sci. Technol. 5 (2016) R3040-R3048.
[21] J. Manam, V. Chatterjee, S. Das, A. Choubey and S.K. Sharma, J. Lumin. 130 (2010) 292-297.
[22] L. Chen, J. Zhang, Y. Luo, S. Lu and X. Wang, Appl. Phys. Lett. 84 (2004) 112-114.
[23] S. Sapra, A. Prakash, A. Ghangrekar, N. Periasamy and D. D. Sarma, J. Phys. Chem. B, 109 (2005) 1663-1668.
[24] L. Brus, IEEE J. Quantum Electron. 22 (1986) 1909-1914.
[25] S. Horoz, Q. Dai, F. S. Maloney, B. Yakami, J. M. Pikal, X. Zhang, J. Wang, W. Wang, and J. Tang, Phys. Rev. Applied 3 (2015) 024011-024017.
[26] M. Stefan, S. V. Nistor, D. Ghica, C. D. Mateescu, M. Nikl, and R. Kucerkova, Phys. Rev. B 83 (2011) 045301-0453011.
[27] M. Ihara, T. Igarashi, T. Kusunoki and
K. Ohno, J. Electrochem. Soc. 147 (2000) 2355-2357.
[28] M. Morita, D. Rau, H. Fujii, Y. Minani, M. Baba, M. Yoshita and H. Akiyama, J. Lumin. 87-89 (2000) 478-481.
[29] Houcine Labiadh, Badreddine Sellami, Abdelhafidh Khazri, Wiem Saidani, Said Khemais, Opt. Mater. 64 (2017) 179-186.
[30] H. E. Gumlich, J. Lumin. 23 (1981) 73-99.
[31] Y. Wang, N. Herron, K. Moller and T. Bein, Solid State Commu. 77 (1991) 33-38.
[32] H. Yang, P. H. Holloway and B. B. Ratna, J. Appl. Phys. 93 (2003) 586-592.
[33] C.D. Lokhande, A. Ennaoui, P.S. Patil, M. Giersig, M. Muller, K. Diesner, H. Tributsch, Thin Solid Films 330 (1998) 70-75.
[34] B.Y. Geng, X.W. Liu, Q.B. Du, X.W. Wei, Appl. Phys. Lett. 88 (2006) 163104-163106.
[35] V.D. Mote, Y. Purushotham, B.N. Dole, Ceramica 59 (2013) 395-400.
[36] A. K. Kole, C. S. Tiwary, and P. Kumbhakar, J. Appl. Phys. 113 (2013) 114308.
[37] RD. Son, D. R. Jung, J. Kim, T. Moon, C. Kim and B. Park, Appl. Phys. Lett. 90 (2007) 101910-101912.
[38] W.Q. Peng, G.W. Cong, S.C. Qu and Z.G. Wang, Opt. Mater. 29 (2006) 313-317.
[39] P. De Visschere and K. Neyts, J. Lumin. 52 (1992) 313-323.
[40] H. S. Bhatti, R. Sharma, N. K. Verma, N. Kumar, S. R. Vadera and K. Manzoor, J. Phys. D: Appl. Phys. 39 (2006) 1754–1760.
[41] A. A. Bol and A. Meijerink, Phys. Rev. B 58 (1998) R15997-R16000.
[42] W. Chen, V. F. Aguekian, N. Vassiliev, A. Y. Serov and N. G. Filosofov, J. Chem. Phys. 123 (2005) 124707-124711.