Issue
Eur. Phys. J. Appl. Phys.
Volume 89, Number 1, January 2020
Disordered Semiconductors: Physics and Applications
Article Number 10304
Number of page(s) 5
Section Thin Films
DOI https://doi.org/10.1051/epjap/2020190258
Published online 09 April 2020

© EDP Sciences, 2020

1 Introduction

As a potential high-efficient luminescent material in silicon-based photoelectric integration, the a-SiNxOy systems have been greatly attracted the interests of researchers [121]. The early research works were mainly focused on the luminescent properties of the a-SiNxOy films prepared under higher substrate temperature [111], and a detailed study on the related PL and electroluminescence (EL) properties have been carried out [38]. However, the luminescent efficiency of such a-SiOxNy films prepared under higher substrate temperature is very low, which only give out strong PL and EL after being annealed above 800 °C [16]. Meanwhile, a few works have been reported the PL mechanisms of a-SiOxNy films with larger oxygen doping amount (>10%) [1318]. In our previous work, we consciously doped small amounts of oxygen (2–5%) into a-SiNx prepared under low temperature (250 °C), and successfully realized highly efficient PL and EL with the modulable peak wavelengths (420–600 nm) of the visible light band through changing the Si/N ratio [2224]. Thereby, we further verified the luminescent N–Si–O defects are responsible for this type of highly efficient tunable luminescence in a-SiNxOy systems [2426].

However, light induced degradation is a commonly seen phenomenon in Si based luminescent materials [2732], and the a-SiNxOy films are no exception. Not only the deposition parameters and the film structure have certain influence, but also the post treatment conditions play an important role for the application of this kind of materials. We found that the as-deposited a-SiNxOy films have much more serious light induced degradation phenomenon. After being continuously irradiated with 325 nm He-Cd cw-laser (30 mW) for two hours, the PL integrated intensity of this kind of a-SiNxOy films dropped to nearly half of the initial intensity before irradiation in average. It is almost impossible to get high gain factor to realize stimulated emission under such poor luminescence stabilities. Therefore, it is necessary and urgent to explore suitable post-treatment processes in the purpose of finally eliminating the light induced degradation and thus, to obtain a-SiNxOy films with high efficiency and stable luminescence. In this work, we intensively investigated the light induced degradation properties of a-SiNxOy films, and further significantly improved the PL stability. After the furnace annealing combing with the continuously UV irradiation, the light induced degradation has been obviously improved. However, unfortunately, we found that when we exposed the post-treated samples in situ in the air after a period, the a-SiNxOy films exhibit a PL self-recovery property. In order to eliminate the light induced degradation and the meta-stable PL characteristics, we cautiously performed thermal annealing combined with pulsed laser annealing processes with the purpose of improving film density and eliminate weak bond angle, and finally access to the highly efficient and stable luminescent a-SiNxOy films in the visible range.

2 Experimental details

The a-SiNxOy films were deposited on Si and quartz substrates in a PECVD system, which processes were described in detail in our previous works elsewhere [23]. SiH4, NH3 and N2 were adopted as the precursor reactive gas, and the N/Si atom ratios were realized through precisely controlling flow rate R (R = NH3/SiH4) with mass flow meter by changing ammonia flow. After the growth, in-situ plasma oxidation was also performed. The thicknesses of a-SiNxOy films with different R ratios were controlled about 200 nm. The annealing processes were carried out by using a KrF ns-pulsed laser (λ exc = 248 nm, pulse width ∼20 ns, repetition rate= 1 Hz), and a JetFirst 300C vacuum thermal furnace, respectively. The PL properties were recorded by a Fluorolo-3 system (Jobin Yvon), using a He–Cd continue wave (CW) laser (λ exc = 325 nm, power 30 mW) as excitation source. The thickness and chemical composition of the a-SiNxOy layers, and related atomic scale defect states were checked by performing the XPS (Thermo ESCALAB 250) and the EPR (Bruker EMXplus X-band spectrometer) measurements, respectively.

3 Results and discussion

With the purpose of improving luminescent stability in a-SiNxOy systems, we carried out thermal furnace post-annealing treatments through changing the annealing conditions, such as annealing protection atmosphere [N2, highly pure N2(6N), Ar, O2, Ar (95%) + O2 (5%)], annealing time (0.5–2 h), annealing temperature (800–1150 °C), and the flow rate of the protection atmosphere (0.5–4 L/min) to the as-deposited a-SiNxOy films. Figure 1 shows the time evolutional PL spectra of light induced degradation of the annealed a-SiNxOy films through annealing treatment in N2 protection atmosphere (40 min, 2.5 L/min) in the furnace oven. After irradiating continuously for 2 h with 325 nm He-Cd CW laser, we found that the light induced degradation was improved obviously, and the PL integrated intensity of the annealed a-SiNxOy samples has dropped to ∼85% of the original value.

Since the furnace annealing cannot eliminate the light induced degradation completely, thus we further use 325 nm He-Cd CW UV-laser to irradiate the annealed a-SiNxOy samples continuously. We found that light induced degradation was further improved after being irradiated by He-Cd laser. Figure 2 is the attenuation characteristics of the PL integrated intensity with the increase of irradiation time of the annealed samples. After 12-hour continuous irradiation in the air, the PL integrated intensity firstly decreased rapidly and then became saturated, as shown in Figure 2. Based on the foresaid two ways to improve the light induced degradation, we further measured the luminescent stability of the annealed-irradiated a-SiNxOy samples. However, we found that the results were different from what we had supposed. After placing the annealed-irradiated samples in situ for 7 days, the PL initial integrated intensity had basically restored to the level before irradiation, while with the evolution of time, the PL showed self-restoring and metastable characteristics. It was reported that the PL intensity could be restored through annealing [24], while there was never a report on the self-restoring characteristic of PL after exposing in the air. This should be caused by the change and reconstruction of internal structure of the sample after laser irradiation [27,28,33,34].

To further identify the origin of the metastable PL properties mentioned above, we checked the bonding configuration and atom scale dangling bonds defects of a-SiNxOy samples by performing XPS and EPR measurements, respectively. The chemical composition and bonding configuration of both a-SiNxOy films and the controlled a-SiN x films were investigated by employing the XPS, including Si 2p, N 1s, and O 1s spectra. Figure 3 and Table 1 shows the change of Si 2p and N 1s peak positions of a-SiNxOy under different R and controlled a-SiNx films, which surfaces were all etched for several minutes by using Ar+ beam with Ar+ energy of 2000 eV (etch speed ∼1 Å/s). Si 2p peak positions of a-SiNxOy films have a blue-shift from 102.47 eV to 102.23 eV with the rise of R, and are between Si 2p peak position of standard proportion Si3N4 (101.9 eV) and SiO2 (103.4 eV), respectively in correspondence to binding energies: three kinds of bonding configuration [2324], Si-ON3 (102.3 eV), Si-O2N2 (102.6 eV) and Si-O3N (102.95 eV). We can obtain the similar rule from N 1s spectra. Therefore, from variation characteristics of Si 2p and N 1s bonding energies, there are indeed that oxygen atom had been diffused into a-SiNx films and the N-Si-O bonding configuration has been formed in a-SiNxOy samples after in-situ plasma oxidation [23,24].

Through the measured energy spectrum peaks, we calculated out all the atomic concentration ratios of Si, N and O within a-SiNxOy films by the integration of each peak area. In order to detect the in-situ oxide layer thicknesses and related oxygen concentration, we etched surface of a-SiNxOy samples by changing the sputtering times which ranged from 0 to 30 minutes (∼180 nm) before XPS measurements (the total thickness of the a-SiOxNy films is about 200 nm). From the measurement results, as shown in Figure 4, we found that the influence by higher content of oxygen (10–40%) is almost on sample surface (0–15 nm), and the oxygen concentration drops rapidly after etching for certain thicknesses (>15 nm). After surface etching for 30–180 nm, with rise of flow rate R, nitrogen atom content is increasing while Si atom content is decreasing, as summarized in Table 1. However, oxygen concentration has little variation and the value remains steady within 2.2–5.1%, which is different from concentration range of oxygen content within 10–50% reported previously [511], indicating that O atom just functions as an impurity doped into a-SiNx films.

Then we intensively investigated the properties of the related dangling bonds defects. Figure 5 exhibits the measured first derivative EPR spectra of annealed a-SiNxOy samples before and after irradiation with UV-light for 12 hours in the air, respectively. The related g values and spin densities are shown in Figure 5 insert. In our previous works, we have demonstrated that the trivalent Si DBs (such as Si≡SiN2, g = 2.0036) and luminescent NX defects (Si≡N*-O, g ≈ 2.0011) are coexisted in our a-SiNxOy systems; the PL efficiency of Si DBs is pretty low, even if the density of Si DBs defect states is higher than that of Nx defect states. Therefore, the PL is major originated from luminescent N–Si–O bonds related Nx defect states [24]. Meanwhile, we noticed that irradiation is affecting the g values. The g values in irradiated a-SiNxOy samples (g = 2.0034) were smaller than those without irradiation, indicating the luminescent N-Si-O bonding states slightly decrease in the irradiation processes.

The a-SiNxOy matrix deposited by PECVD system under low temperature is a porous “fractal-like network” structure, which is probably inherent in low-temperature deposition. As soon as the samples are exposed in the atmosphere, moisture H2O vapor in the air can percolate through numerous micro voids into these porous films. Liao et al. found that, for a-SiNx films by PECVD at temperatures ranging from 50 to 250 °C, the as-deposited films start to oxide as soon as the samples are taken out from the reaction chamber and exposed to the atmosphere [33]. From the XPS results above, we found that O atom just functions as an impurity doped into a-SiNx films. Since the sample is exposed in the air, it is easy to be further oxidized, and N is replaced by O, which makes the unstable Si–N bond be converted into the much more stable Si–O bond [33,34], as shown in Figure 2 insert.

In order to ensure the influence of oxygen absorption of the samples to the PL self-restoring characteristics, we put the annealed a-SiNxOy samples which being irradiated with UV-light for 12 hours in a vacuum container to carry out the contrast test for the time-evolving PL stability. After being irradiated with UV-light, the light induced degradation is almost saturated, and the PL integrated intensity is almost not restored in vacuum. Thus, we can conclude that, after forming the defect luminescence center N–Si–O bonding states, under the conditions of standing in the air and being irradiated with 325 nm He-Cd laser, the state density of different N–Si–O and Si–N bonds will transform mutually [27,2931], and thus lead to the luminescent metastable characteristics, as shown in Figure 2 insert. In the process of irradiating with 325 nm He-Cd laser, weak bonds (such as Si–H) will break off and the Si, N and O will bond again after diffusing, thus forming the relevant defect modes of N–Si–O bond, inducing the decrease of the total spin densities [24].

To eliminate the light induced degradation and the meta-stable PL characteristics, we cautiously performed furnace thermal annealing combined with a KrF ns-pulsed laser (λ exc = 248 nm, pulse width ∼20 ns, repetition rate= 1 Hz) annealing processes. The KrF laser annealing processes were described in detail in our previous works [35,36]. The laser beam is firstly focused and then homogenized, by a fused quartz lens and a beam homogenizer, to increase and uniform its energy density, respectively. The exposure areas in the surfaces of thin films can be varied from 10 × 20 to 10 × 5 mm2 in our experimental system. The ultrafast and strong pulsed laser beam can improve film density and eliminate weak bond angle in samples under certain conditions (by changing shot times, shot frequency, or laser energy, etc.) without any structure variation, thus accessing the highly efficient and stable luminescent a-SiNxOy films. The structure of the before and after laser-annealed samples under various laser beam energies was inspected by Raman spectroscopy (Jobin Yvon HR-800) with the backscattering geometry, using the 488 nm line Ar+ laser. The a-SiNxOy films under 40 mJ/cm2 have no 480 cm−1 broad band peak of amorphous Si, neither 520 cm−1 peaks in crystalline silicon standard. Combining with the XPS results in which no low energy component around 99.6 eV for elemental Si has been found, it shows that there is no Si quantum dot within a-SiNxOy films, which turn out to be the structure of amorphous matrix. After being post-treated, the PL stability of the samples is obviously improved. As shown in Figure 6, the PL integrated intensity has only dropped to ∼90% of original value under 325 nm He-Cd laser light irradiate continuously for 6 hours, and the average gradient of IPL attenuation is about 0.02, which is becoming saturated with the irradiation time, and has no self-restoring and metastable characteristics after placing in situ for 6 days.

thumbnail Fig. 1

Time evolutional PL spectra of the annealed a-SiNxOy films (R = 1.5) under UV irradiation. The inset shows the integrated PL intensity vs. irradiating times.

thumbnail Fig. 2

Time-evolving PL stability of annealed a-SiNxOy samples with 325 nm He-Cd laser irradiated continuously in the air and under vacuum condition, respectively. Insert shows the autoxidation and irradiation processes of the dangling bonds.

thumbnail Fig. 3

XPS spectra of Si 2p and N 1s of both a-SiNxOy and the controlled a-SiNx films which is measured after the Ar ion beam sputtering for 30 nm depth.

Table 1

Binding energy of N 1s and Si 2p and Si, N, O chemical compositions of a-SiNxOy films from XPS measurements.

thumbnail Fig. 4

The atom abundance of the a-SiOxNy films vs. etch depth in the XPS measurements.

thumbnail Fig. 5

EPR spectra of a-SiNxOy samples before and after 325 nm He-Cd laser irradiated continuously in the air, respectively. Insert shows the related measured and simulated constants of the dangling bonds.

thumbnail Fig. 6

Time-evolving PL stability of post treated a-SiNxOy samples placing in situ for 6 days.

4 Conclusions

In this work, we intensively investigated the light induced degradation and PL stability characteristics of a-SiNxOy films. We found that the irradiated a-SiNxOy films showed as metastable characteristics with the evolution of time after placed in the air. To eliminate the light induced degradation and the PL self-restoring characteristics, post-treatments including improving density and eliminating the weak bond were carried out to the a-SiNxOy films by combining thermal annealing with pulsed laser annealing, thus obtaining the high performance of luminescent stability in a-SiNxOy systems.

Author contribution statement

Pengzhan Zhang prepared the samples and performed the experiments; Pengzhan Zhang and Sake Wang analyzed all the data related to the manuscript; all the authors contributed to abundantly fruitful discussions.

Acknowledgments

The authors would like to thank the supports National Science Foundation for Post-doctoral Scientists of China (No. 2017M621711), and Science Foundation of Jinling Institute of Technology (No. 40620062).

References

  1. A.P. Li, L.D. Zhang, Y.X. Zhang, G.G. Qin, X. Wang, X.W. Hu, Appl. Phys. Lett. 69 , 4 (1996) [Google Scholar]
  2. G.F. Bai, Y.P. Qiao, Z.C. Ma, W.H. Zong, G.G. Qin, Appl. Phys. Lett. 72 , 3408 (1998) [Google Scholar]
  3. B.H. Augustine, E.A. Irene, Y.J. He, K.J. Price, L.E. McNeil, K.N. Christensen, D.M. Maher, J. Appl. Phys. 78 , 4020 (1995) [Google Scholar]
  4. B.H. Augustine, Y.Z. Hu, E.A. Irene, L.E. McNeil, Appl. Phys. Lett. 67 , 3694 (1995) [Google Scholar]
  5. K.J. Price, L.E. McNeil, A. Suvkanov, E.A. Irene, P.J. MacFarlane, M.E. Zvanut, J. Appl. Phys. 86, 2628 (1999) [Google Scholar]
  6. K.J. Price, L.R. Sharpe, L.E. McNeil, E.A. Irene, J. Appl. Phys. 86, 2638 (1999) [Google Scholar]
  7. T. Noma et al., Appl. Phys. Lett. 79, 1995 (2001) [Google Scholar]
  8. H. Kato et al., J. Appl. Phys. 93 , 239 (2003) [Google Scholar]
  9. A. Tewary, R.D. Kekatpure, M.L. Brongersma, Appl. Phys. Lett. 88 , 093114 (2006) [Google Scholar]
  10. M. Xu, S. Xu, J.W. Chai, J.D. Long, Y.C. Ee, Appl. Phys. Lett. 89 , 251904 (2006) [Google Scholar]
  11. M. Xu, S. Xu, J.W. Chai, J.D. Long, Q.J. Cheng, Y.C. Ee, K. Ostrikov, J. Appl. Phys. 103 , 053512 (2008) [Google Scholar]
  12. T. Qiu, X.L. Wu, G.J. Wan, Y.F. Mei, G.G. Siu, P.K. Chu, Appl. Phys. Lett. 86 , 193111 (2005) [Google Scholar]
  13. X.T. Zhang, Z. Liu, Z. Zheng, S.K. Hark, Y.B. Fu, G.B. Zhang, Appl. Phys. Lett. 90 , 183110 (2007) [Google Scholar]
  14. H.L. Hao, W.Z. Shen, Nanotechnology 19 , 455704 (2008) [CrossRef] [PubMed] [Google Scholar]
  15. P.D. Nguyen, D.M. Kepaptsoglou, Q.M. Ramasse, M.F. Sunding, L.O. Vestland, T.G. Finstad, A. Olsen, J. Appl. Phys. 112 , 073514 (2012) [Google Scholar]
  16. R. Ruggeri, F. Neri, A. Sciuto, V. Privitera, C. Spinella, G. Mannino, Appl. Phys. Lett. 100 , 042104 (2012) [Google Scholar]
  17. S. Jou, I. Liaw, Y. Cheng, C. Li, J. Luminescence 134 , 853 (2013) [Google Scholar]
  18. A.M. Hartel, S. Gutsch, D. Hiller, M. Zacharias, Phys. Rev. B 87 , 035428 (2013) [Google Scholar]
  19. G.R. Lin, Ch. J. Lin, C.K. Lin, L.J. Chou, Y.L. Chueh, J. Appl. Phys. 97 , 094306 (2005) [Google Scholar]
  20. G.R. Lin, Y.H. Pai, C.T. Lin, C.C. Chen, Appl. Phys. Lett. 96 , 263514 (2010) [Google Scholar]
  21. G.R. Lin, S.P. Su, C.L. Wu, Y.H. Lin, B.J. Huang, H.Y. Wang, C.T. Tsai, C.I. Wu, Y.C. Chi, Sci. Rep. 5 , 9611 (2015) [CrossRef] [PubMed] [Google Scholar]
  22. R. Huang, K. Chen, B. Qian, S. Chen, W. Li, J. Xu, Z. Ma, X. Huang, Appl. Phys. Lett. 89 , 221120 (2006) [Google Scholar]
  23. P. Zhang, K. Chen, H. Dong, P. Zhang, Z. Fang, W. Li, J. Xu, X. Huang, Appl. Phys. Lett. 105 , 011113 (2014) [Google Scholar]
  24. P. Zhang, K. Chen, Z. Lin, H. Dong, W. Li, J. Xu, X. Huang, Appl. Phys. Lett. 106 , 231103 (2015) [Google Scholar]
  25. P. Zhang, K. Chen, Z. Lin, D. Tan, H. Dong, W. Li, J. Xu, X. Huang, Appl. Phys. Lett. 108 , 111103 (2016) [Google Scholar]
  26. Z. Lin, K. Chen, P. Zhang, J. Xu, W. Li, H. Yang, X. Huang, Appl. Phys. Lett. 108 , 111103 (2017) [Google Scholar]
  27. W.L. Warren, C.H. Seager, J. Kanicki, M.S. Crowder, E. Sigari, J. Appl. Phys. 77 , 5730 (1995) [Google Scholar]
  28. S.V. Deshpande, E. Gulari, S.W. Brown, S.C. Rand, J. Appl. Phys. 77 , 6534 (1995) [Google Scholar]
  29. J.T. Yount, G.T. Kraus, P.M. Lenahan, D.T. Krick, J. Appl. Phys. 70 , 4969 (1991) [Google Scholar]
  30. W.L. Warren, J. Robertson, J. Kanicki, Appl. Phys. Lett. 63, 2685 (1993) [Google Scholar]
  31. V.A. Gritsenko, J.B. Xu, R.W.M. Kwok, Y.H. Ng, I.H. Wilson, Phys. Rev. Lett. 81, 1054 (1998) [Google Scholar]
  32. S. Naskar, S.D. Wolter, C.A. Bower, B.R. Stoner, J.T. Glass, Appl. Phys. Lett. 87 , 261907 (2005) [Google Scholar]
  33. W.S. Liao, C.H. Lin, S.C. Lee, Appl. Phys. Lett. 65, 2229 (1994) [Google Scholar]
  34. M. Martyniuk, J. Antoszewski, C.A. Musca, J.M. Dell, L. Faraone, J. Appl. Phys. 99 , 053519 (2006) [Google Scholar]
  35. X.F. Huang, Z.F. Li, Z.G. Liu, K.J. Chen, J. Non-Cryst. Solids 821, 198 (1996) [Google Scholar]
  36. X.F. Huang, K.J. Chen, J. Non-Cryst. Solids 1015, 266 (2000) [Google Scholar]

Cite this article as: Pengzhan Zhang, Sake Wang, Kunji Chen, Xinglong Wu, Investigation on the luminescent stability in amorphous silicon oxynitride systems, Eur. Phys. J. Appl. Phys. 89, 10304 (2020)

All Tables

Table 1

Binding energy of N 1s and Si 2p and Si, N, O chemical compositions of a-SiNxOy films from XPS measurements.

All Figures

thumbnail Fig. 1

Time evolutional PL spectra of the annealed a-SiNxOy films (R = 1.5) under UV irradiation. The inset shows the integrated PL intensity vs. irradiating times.

In the text
thumbnail Fig. 2

Time-evolving PL stability of annealed a-SiNxOy samples with 325 nm He-Cd laser irradiated continuously in the air and under vacuum condition, respectively. Insert shows the autoxidation and irradiation processes of the dangling bonds.

In the text
thumbnail Fig. 3

XPS spectra of Si 2p and N 1s of both a-SiNxOy and the controlled a-SiNx films which is measured after the Ar ion beam sputtering for 30 nm depth.

In the text
thumbnail Fig. 4

The atom abundance of the a-SiOxNy films vs. etch depth in the XPS measurements.

In the text
thumbnail Fig. 5

EPR spectra of a-SiNxOy samples before and after 325 nm He-Cd laser irradiated continuously in the air, respectively. Insert shows the related measured and simulated constants of the dangling bonds.

In the text
thumbnail Fig. 6

Time-evolving PL stability of post treated a-SiNxOy samples placing in situ for 6 days.

In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.