• Sonuç bulunamadı

Strong white light emission from a processed porous silicon and its photoluminescence mechanism

N/A
N/A
Protected

Academic year: 2021

Share "Strong white light emission from a processed porous silicon and its photoluminescence mechanism"

Copied!
6
0
0

Yükleniyor.... (view fulltext now)

Tam metin

(1)

Strong white light emission from a processed porous silicon and its

photoluminescence mechanism

T. Karacali

a,n

, K. Cicek

b a

Ataturk University, Engineering Faculty, Department of Electrical and Electronics, 25240 Erzurum, Turkey

bArdahan University, Engineering Faculty, Department of Electrical and Electronics, 75100 Ardahan, Turkey

a r t i c l e

i n f o

Article history: Received 7 March 2011 Received in revised form 10 May 2011

Accepted 12 May 2011 Available online 23 May 2011 Keywords: Porous silicon Temperature dependent photoluminescence Oxidation Carbonization Thermal annealing

a b s t r a c t

We have prepared various porous silicon (PS) structures with different surface conditions (any combination of oxidation, carbonization as well as thermal annealing) to increase the intensity of photoluminescence (PL) spectrum in the visible range. Strong white light (similar to day-light) emission was achieved by carrying out thermal annealing at 1100 1C after surface modification with 1-decene of anodic oxidized PS structures. Temperature-dependent PL measurements were first performed by gradually increasing the sample temperature from 10 to 300 K inside a cryostat. Then, we analyzed the measured spectrum of all prepared samples. After the analysis, we note that throughout entire measured spectrum, only two main peaks corresponding to blue and green-orange emission lines (which can be interpreted by quantum size effect and/or configuration coordinate model) were seem to be predominant for all temperature range. To further reveal and analysis these peaks, finally, measured data were inputted into the formula of activation energy of thermal excitation. We found that activation energies of blue and green-orange lines were approximately 49.3 and 44.6 meV, respectively. &2011 Elsevier B.V. All rights reserved.

1. Introduction

Light emitting devices based on silicon technology are of great interest in illumination and display applications. Thus, studies on the light emitting feature of porous silicon (PS) structures have tremendously intensified since the discovery of visible lumines-cence at room temperature[1]. The luminescence spreading from near infrared throughout the visible region is known to originate from the monocrystalline skeleton[2,3] and is affected by aging of dangling bonds on the PS surface [4,5]. Obtaining a stable photoluminescence (PL) spectrum is crucial in optoelectronic applications. Studies have been done to obtain a stable radiation of unprocessed PS structures, because PL spectrum shifts by time

[5–9]. To obtain a similar radiation property for processed PS structures, electrochemical or thermal oxidation was carried out on PS surface in some studies [10,11], while carbonization was applied in the others[12–15]. However, they achieved a stable radiation only for a limited spectrum in the visible range such as red color. Nonetheless, their findings lead to an increase in potential of such studies because white light illumination can be utilized in almost every aspect of industry applications. To accomplish a stable radiation in whole visible range (white light),

a thin film created by a chemical vapor deposition technique was prepared[16,17]. Although good results were achieved by this technique, the technique itself is expensive and may not be appropriate for mass production of such films. To produce relatively inexpensive PS structures in the whole visible range, the combination of thermal carbonization with oxidation surface preparation techniques were applied[18–20]. The common part of all these studies[16–20] in producing white lights stems from the usage of nanosize silicon, SiO2and carbon. Besides, white light

was also produced by after depositing either ZnS[21]or ZnO[22]

on a PS structure. However, the intensity of emitted white light by such an approach was not as strong as that was achieved by the combination of nanosize silicon, SiO2and carbon.

It was known that temperature variations during PL measure-ments can seriously affect PL peak energy and intensity data. Considering this important fact, researchers have investigated the temperature dependence of PL spectrum to understand the emission mechanism of PS structures. While some researchers noted that the PL peak position shifts towards only shorter wavelengths with an increase in temperature [23], others reported that a similar peak position shift can be towards not only shorter wavelengths (blue shift) but also longer wavelengths (red shift)[24]. The red shift due to a temperature rise can be explained by the quantum size effect, because the radiative recombination of the excitons confined in a quantum wire. However, blue shift can only be interpreted by the configurational Contents lists available atScienceDirect

journal homepage:www.elsevier.com/locate/jlumin

Journal of Luminescence

0022-2313/$ - see front matter & 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jlumin.2011.05.031

n

Corresponding author. Tel.: þ90 442 231 4878; fax: þ 90 442 2360957. E-mail address: tevhit@atauni.edu.tr (T. Karacali).

(2)

by our technique, we fabricated PS structures using electrochemical anodic etching from Pþþsilicon wafers. To improve their surface

conditions, we later applied anodic oxidation, carbonization and thermal annealing. Next, in Section 3, PL measurements are carried out to examine emission characteristics of the samples from 10 to 300 K. Then, in Section 4, both scanning electron microscopy (SEM) measurements and energy dispersive X-ray (EDX) analyses are carried out for the sample emitting the strongest white light. Here, we also calculated the activation energies of samples emitting white light. Finally, in Section 5, we summarize all the findings and present our conclusions.

2. Fabrication of the porous silicon structures

Some PS structures (7 samples), as shown in Table 1, were fabricated by electrochemical anodic etching of a highly doped silicon wafer (

r

¼0.01

O

cm) one sample at a time in anodization cell. The electrolyte solution was prepared using a 1:2 mixture of 48% HF and ethanol, and then etching process was carried out at room temperatures. Current densities from 100 to 20 mA/cm2

were gradually applied to each sample to produce an index matching layer during 10 s, before their essential PS layer was created by applying 20 mA/cm2during 300 s. Among the samples

PS processed, two of them were taken out from the anodization cell while one of them was left as it is (denoted as fabricated in

Table 1), the other was left in ambient temperature and then processed with carbonization to be discussed shortly (designated as only carbonized in Table 1). The remaining each sample in anodization cell was put in oxidization process. In this process, after washing them by ethanol without removing it from the anodization cell, these samples were anodical oxidized in 1 M H2SO4 by applying 3 mA/cm2for 5 min. Then, all samples were

extracted from anodization cell and one of them was not addi-tional processed (denoted as only oxidized in Table 1). Of the samples oxidized, one of them was left for annealing process, whereas the remaining three together with the one left for carbonization process were carbonized. In the carbonization process, prepared PS samples were immersed in a deoxygenated

of the PS samples, which are presented in Table 1. These measurements were performed using 325 nm line of HeCd laser with a 27 mW excitation power. Stanford SR830 (Lock-In ampli-ficator), ORIEL MS257 (monochromator with 25 cm focal length) and Hamamatsu PMT-R6356-6 (detector) were used for spectral measurement. Temperatures of the samples were controlled by Lakeshore 331 temperature controller in closed-loop helium cryostat. Nova 600 NanoLab focused ion beam (FIB) with EDX accessory equipment was used profiling elemental analysis.

4. Results and discussion

4.1. Structural characterization of sample P6

Among the samples emitting white light, only the one with strongest emission (sample P6) was analyzed for its structural characterization. Its cross-section SEM image is given in

Fig. 1(a) for better demonstrating its sponge-like structure. While, in the figure, upper left black color region corresponds to cross-sectional view of the sample, the remaining white color region presents its surface view. It is seen fromFig. 1(a) that the effect of sample carbonization on the interface of cross-sectional and surface regions is not discernable.

EDX measurements were employed for elemental analysis. Before measurements, we made a small hole on the surface of the sample by means of FIB technique (Fig. 1(b)). For example,

Fig. 2(a) illustrates the EDX measurements at near the surface of the sample. It is seen from this figure that there are five peak regions corresponding to platinum, silicon, gallium, oxygen and carbon (from right to left in the figure). Among these peaks, while those of platinum and gallium result from platinum barrier and gallium ions used to milling the hole on surface of the sample in FIB technique, those of silicon, oxygen and carbon correspond to the processed sample properties and thereafter concern us in the elemental analysis in the remainder of the study.Fig. 2(b) and (c), respectively, demonstrate the EDX measurements of the sample obtained from the middle and bottom points inside the hole.

After some analysis, we note that the peak values of carbon and oxygen obtained from middle and bottom points (Fig. 2(b) and (c)) are smaller than those obtained from the surface point (Fig. 2(a)). We think that the reason why carbon is absent deeper inside the PS layer can be explained by capillary effect. Since 1-decene is confined to the PS layer close to top surface due to capillary effect[26], it can form a Si–C bound only in this region as reported by Boukherroub et al.[12]and Gelloz et al.[13].

Broadened peaks of PS from samples P1 and P6 are evident in XRD spectrums, as given inFig. 3. The sample P6 (first oxidized-carbonized and then annealed at 1100 1C) have a much broader peak located at 68.971 when compared with the sample P1 (as fabricated). We think that this broadening indicates much thinner

Table 1

PS samples prepared for characterization. Sample name Properties P1 As fabricated P2 Only carbonized P3 Only oxidized

P4 Oxidized, annealed(1100 1C) P5 Oxidized, carbonized

P6 Oxidized, carbonized, annealed (11001 C) P7 Oxidized, carbonized, annealed (12001 C)

(3)

Si nanowires. The absence of any peak related to SiC may be attributed to the mono-layer formation on PS where its total diffraction was not detected by XRD system. However, literature data show that the diffraction peak of SiC is located at near 41.51

[27,28].

4.2. Photoluminescence characterization of samples

Light emission characteristic of the samples between 350 and 650 nm wavelengths at 300 K temperature are givenFig. 4. It is seen from Fig. 4 that whereas strong white light emission is

radiated from samples P6 and P7, a weak luminescence is radiated from sample P4 only while no emission is observed for the rest of the samples (samples P1–P3, P5).

The PL spectrum of sample P6 taken at 300 K is presented in

Fig. 5. This spectrum spreads throughout the entire visible region (350–650 nm).

Peak position of the spectrum is approximately at 500 nm and the spectrum looks like day-light spectrum. When both PL spectrums of samples P4 and P6 are analyzed, we deduce that the convolution of two Gaussian peaks at 400 and 500 nm results this spectrum. The peak at 500 nm is named as Green-Orange Gauss Peak and the peak at 400 nm is called as Blue Gauss Peak in

Fig. 1. SEM (a) and FIB (b) pictures of P6 sample. Gallium ions made a hole bottom of platinum barrier to examine elemental profile of P6 sample (b).

Fig. 2. Three point of EDX analysis of P6 sample in depth. Near surface of the sample (a), middle (b) and bottom of milled hole (c).

Fig. 3. XRD spectra of samples as fabricated PS (P1), processed (P6) and bulk Si.

(4)

the remainder of the text. The behavior of these two gausses can also clearly be observed in the temperature dependent PL mea-surements (Fig. 6). Temperature dependencies of peak positions and intensity of these two Gaussian-peaks can be used in explaining emission mechanism.

Temperature dependent PL measurements of the sample P6 have been recorded from 10 to 300 K with 20 K intervals (Fig. 6(b)). While PL signal decreased rapidly between 300 and 200 K, go to intensified from 150 to 10 K. Similar behavior was observed in the sample P4, which was oxidized and annealed at 1100 1C (Fig. 6(a)).

When temperature dependent PL peak positions of the sample P4 is analyzed, it is observed that both Blue Gauss and Green-Orange Peaks shift to lower energies, that is, longer wavelengths with increase in temperature (Fig. 7(a)). This situation is explained by quantum size effect in the literature [24]. We conclude that thinner silicon nanowires are formed by annealing the PS sample at 1100 1C following the oxidation. Hence, excitons are confined in a quantum wire.

the equation of exciton recombination model as IðTÞ

Ið0Þ1 ¼ C expðEa=ðkTÞÞ, ð1Þ

here, I(0) is the intensity of PL at absolute zero temperature, I(T) is the PL intensity at temperature T, C is the coupling constant, k is the Boltzmann constant and Ea is the activation energy of thermal

excitation from the radiative excited state to the nonradiative state. It is important to analyze the functional behavior of Eq. (1) in interpretation of the measured and fitted temperature dependent PL intensities of samples, which will be discussed shortly. A function g:H-K (from mapping H to the mapping K) is called ‘‘one-to-one’’ if and only if its functional dependency is either increasing or decreasing for every possible value of its independent variable

[29]. In addition, A function g:H-K is called ‘‘onto’’ if and only if for every possible variable in H, there is a correspondence in K[29]. In our case, T corresponds to the independent variable, while I(T) presents the dependent variable and Ea and C are parameters

independent upon T. To demonstrate the one-to-one and onto properties of Eq. (1), we define the following new variables as

k

1¼ Ið0Þ IðTÞ1, ð2Þ 1 C

k

k

2, ð3Þ klnð

k

2Þ Ea ¼

k

3, ð4Þ

k

3¼  1 T: ð5Þ

There is no specific reason for the selection of new variables except for simplification of the analytical analysis. It is seen from (2) to (5) that

k

1,

k

2and

k

3demonstrate the one-to-one property

since

k

1is a function linearly varying with I(T) and I(0), and

k

3is a

function of real natural logarithm. Furthermore, from (2) to (5), we also note that while

k

1and

k

3possess onto property for every

value of T,

k

3does not show this property because when

k

1¼0,

k

3

becomes indefinite. The possibility of

k

1¼0 can arise only when

I(T)¼I(0), which happens at T¼0. In our analysis, we restricted our measurements to above T¼ 0, thus meaning that for our analysis

k

1,

k

2and

k

3can be assumed having onto property. As a

result, we expect that both measurements and fitted data (expo-nential fit) should follow the one-to-one and onto properties.

After the functional study, we analyze the measured PL intensity data of samples. For example,Fig. 7shows the measured (dotted lines) and fitted (dashed lines) the temperature depen-dent PL intensities of the sample P4 (Fig. 8(a) and (b)) and the sample P6 (Fig. 8(c) and (d)). In the fitting process, we input the measured data of the two deconvoluted Gauss peaks into the polyfit function of MATLABs

.

Fig. 5. Photoluminescence (PL) of sample P6 at 300 K.

Fig. 6. (a) PL spectra of the sample P4 and (b) the sample P6 at 20 K intervals from 300 to 10 K.

(5)

Fig. 7. Temperature dependence of PL peak energies (wavelengths) of the sample P4 (a) and the sample P6 (b).

(6)

Acoustic (TA) phonon energies of polytype SiC (3C  , 4H  and 6H  ) were reported to be between 46.3 and 46.7 meV [31]. Therefore, considering these reported TA phonon energies, we think that, after carbonization of samples[31], Si–C bounds would take a role for quenching of blue and green-orange luminescence. Weaker temperature dependence inFig. 8(c) and (d) is a result of small coupling coefficients as calculated from Eq. (1).

At the beginning of this subsection, we concluded that mea-sured and fitted PL intensities would follow one-to-one and onto properties. It is seen from Figs. (8)(a)–(d) that while the fitted data possess these properties (because we applied exponential curve fitting), the measured ones do not hold these properties especially near values when the intensity data start to become constant. For example, in Fig. 8(a) near T ¼65 K, the intensity value first oscillates and then becomes constant. However, this oscillatory behavior contradicts our conclusion that Eq. (1) should have both one-to-one and onto properties. This contradiction arises from accuracy of the measurement equipment, and needs further analysis which is beyond the scope of this study.

5. Conclusion

In this study we reported that a strong white light emission was observed from the PS structures that are concurrently oxidized, carbonized and annealed at 1100 1C. Temperature depended PL measurements were carried out for optical emission mechanism. All PL spectrums are deconvoluted to two Gaussian peaks and analysis show that blue emission appears due to quantum size effect, but green-orange one from carbonized PS sample shifts to blue spectrum as temperature increases, which can be attributed to the configuration coordinate model. The most important impact of carbonization on PS samples and post annealing is the enhancement of room temperature PL data. This finding is expected to opening prospective new technologies and

[4] M.V. Wolkin, J. Jarne, P.M. Fauchet, G. Allen, C. Delerue, Phys. Rev. Lett. 82 (1999) 197.

[5] T. Karacali, B. Cakmak, H. Efeoglu, Opt. Express (2003) 123711 (2003) 1237. [6] J.L. Gole, F.P. Dudel, D. Grantier, D.A. Dixon, Phys. Rev. B 56 (1997) 2137. [7] D.A. Dixon, J.L. Gole, Phys. Rev. B (1998) 1200257 (1998) 12002. [8] D.A. Dixon, J.L. Gole, J. Phys. Chem. B (2005) 14830109 (2005) 14830. [9] J.L. Gole, E. Veje, R.G. Egeberg, A. Ferreira da Silva, I. Pepe, D.A. Dixon, J. Phys.

Chem., B 110 (2006) 2064.

[10] B. Gelloz, N. Koshida, J. Appl. Phys. (2000) 431988 (2000) 4319. [11] J. Salonen, V.P. Letto, E. Laine, Appl. Surf. Sci. 120 (1997) 191.

[12] R. Boukherroub, D.D.M. Wayner, D.J. Luckword, Appl. Phys. Lett. 81 (2002) 601.

[13] B. Gelloz, H. Sano, R. Boukherroub, D.D.M. Wayner, D.J. Luckword, N. Koshida, Appl. Phys. Lett. 83 (2003) 2342.

[14] R. Guerrero-Lemos, J.D. Moreno, R.J. Martin-Palma, F. Ben-Hander, J.M. Martinez-Duart, J.L.G. Fierro, P. Gomez-Garrido, Thin Solid Films 354 (1999) 34.

[15] J. Salonen, M. Bjorkqvist, E. Laine, L. Niinisto, Appl. Surf. Sci. 225 (2004) 389. [16] S.Y. Seo, K.S. Cho, J.H. Shin, Appl. Phys. Lett. 84 (2004) 717.

[17] A. Benami, G. Santana, A. Ortiz, A. Ponce, J. Aguliar-Hernandez, G. Contreras-Puente, J.C. Alonso, Nanotechnology (2007) 1557018 (2007) 15570. [18] J.A. Lara, P. Kathirgamanathan, IEE Proc. Optoelectron. 147 (2000) 363. [19] A.V. Vasin, Y. Ishikawa, N. Shibata, J. Salonen, V.P. Lehto, Jpn. J. Appl. Phys. 46

(2007) L465–467.

[20] T. Karacali, White light emission like day-light from porous silicon, in: Proceedings of International Conference on Nanomaterials and Nanosystems (NanoMats2009), Istanbul, Turkey, 10–13 August, 2009.

[21] K.W. Cheah, L. Xu, X. Huang, Nanotechnology 13 (2002) 238. [22] E. Kayahan, J. Lumin. (2010) 1295130 (2010) 1295.

[23] S. Gardelish, J.S. Rimmer, P. Dawson, B. Hamilton, R.A. Kubiak, T.E. Whall, E.H.C. Parker, Appl. Phys. Lett. 59 (1991) 2118.

[24] A. Shimizu, Y. Yamada, G. Izitsu, K. Yano, M. Kasuga, Jpn. J. Appl. Phys. 35 (1996) L276.

[25] Z.C. Feng, R. TSU, Porous Silicon, World Scientific Publishing Co. Pvt. Ltd., Singapore, ISBN 981 02 1634 3, 1994.

[26] T. Karacali, M. Alanyalioglu, H. Efeoglu, IEEE Sens. J. 9 (12) (2009) 1667. [27] J. Komiyama, Y. Abe, S. Suzuki, H. Nakanishi, Preparation of cubic silicon

carbide with smooth surface on carbonized porous silicon, in: Proceedings of Fall Meeting of The Electrochemical Society, USA, 2004.

[28] S. Madapura, A.J. Steckl, M. Loboda, J. Electrochem. Soc. 146 (3) (1999) 1197. [29] G.B. Afken, H.J. Weber, Mathematical Methods for Physicists, sixth ed.,

Elsevier, MA, USA.

[30] U. Serincan, M. Kulakci, R. Turan, S. Foss, T.G. Finstad, Nucl. Instrum. Methods Phys. Res. B 254 (2007) 87.

[31] G.L. Harris, (Ed.), Properties of Silicon Carbide, INSPEC, London, United Kingdom, 1995, ISBN: 0 85296 870 1.

Şekil

Table 1 ), the other was left in ambient temperature and then processed with carbonization to be discussed shortly (designated as only carbonized in Table 1 )
Fig. 2. Three point of EDX analysis of P6 sample in depth. Near surface of the sample (a), middle (b) and bottom of milled hole (c).
Fig. 6. (a) PL spectra of the sample P4 and (b) the sample P6 at 20 K intervals from 300 to 10 K.
Fig. 8. The temperature dependence of PL intensity of the sample P4 (a, b) and the sample P6 (c, d).

Referanslar

Benzer Belgeler

For this reason, there is a need for science and social science that will reveal the laws of how societies are organized and how minds are shaped.. Societies have gone through

Probability of bit error performances of this system are analyzed for various values of signal to interference ratios (SIR) (0 to 6 dB) and a constant signal to noise ratio (SNR)

Different from other studies, this study was studied parallel to the various criteria (topography, activity areas, privacy...) in the development of the residences in Lapta town and

As a result of long studies dealing with gases, a number of laws have been developed to explain their behavior.. Unaware of these laws or the equations

• Natural  radioactivity:  Unstable  isotopes  in  nature  cause  this  radioactivity.  The  half-lives  of  these  isotopes  are  very  long  and  they  are 

 Note, however, that the electron that left the P870 reaction centre was a high-energy electron, boosted by the absorption of a photon of light, while the electron that returns

 Note, however, that the electron that left the P870 reaction centre was a high-energy electron, boosted by the absorption of a photon of light, while the electron that returns

The device consists of a compartment into which a suppository is placed and a thermostated water tank which circulates the water in this compartment.. The