Optik - International Journal for Light and Electron Optics 174 (2018) 347–353

Contents lists available at ScienceDirect

Optik
journal homepage: www.elsevier.com/locate/ijleo

Original research article

Raman spectroscopy of porous silicon substrates

T
⁎
Magdaléna Kadlečíkováa, Juraj Brezaa, , Ľubomír Vančob, Miroslav Mikolášeka,
Michal Hubeňákc, Juraj Rackoa, Ján Gregušd
a
Slovak University of Technology, Faculty of Electrical Engineering and Information Technology, Ilkovičova 3, 812 19 Bratislava, Slovakia
b
Slovak University of Technology, University Science Park Bratislava Centre, Vazovova 5, 812 43 Bratislava, Slovakia
c
Comenius University, Faculty of Natural Sciences, Department of Inorganic Chemistry, Mlynská dolina, Ilkovičova 6, 842 15 Bratislava, Slovakia
d
Comenius University, Faculty of Mathematics, Physics and Informatics, Department of Experimental Physics, Mlynská dolina, Ilkovičova 6, 842 15
Bratislava, Slovakia

A R T IC LE I N F O ABS TRA CT

Keywords: We have investigated the effect of the etching time on the Raman spectra of porous silicon
Porous silicon prepared by anodic etching. Electrochemical destruction of the substrate increasing with the
Raman spectroscopy etching time and the correlation between the microstructure of the silicon wafer and the shape
Scanning electron microscopy (SEM) and position of their Raman spectra have been observed. Raman analysis has shown that the
intensity of the Raman dominant silicon band decreases and the bandwidth is shifted to lower
frequencies, depending on the morphology of the sample. Therefore we believe that the elec-
trochemical destruction of the surface of Si substrates leads to surface amorphization.

1. Introduction

Porous silicon and nanostructured silicon are attractive materials for many applications such as sensing [1,2], energy storage
[3,4] and photovoltaic conversion [5,6]. The ability to prepare porous silicon (PS) and nanostructured silicon with cheap techniques
with good control and flexibility of the resulting features is a key issue for such applications. Electrochemical etching of silicon has
already been successfully used for preparation of nanowires [7,8], nanopores [9] and porous silicon surface [10]. Good scalability to
large areas makes this technique highly attractive for industrial applications. There are two main electrochemical etching techniques,
both of them are based on hydrofluoric acid solutions for structuring of silicon: (i) metal assisted chemical etching, and (ii) anodic
etching in an electrochemical cell without a metal catalyst. The first technique utilizes metal catalysts such as Au, Ag and Pt to
localize etching of silicon. This localization is due to the reduction of the oxidant (typically H2O2) by the catalyst and to the migration
of the produced holes to silicon, which results into oxidation and etching of silicon [11]. The main disadvantage of this technique is
the requirement to use expensive noble catalysts and the possibility of their incorporation into the layer during etching. The in-
corporated metal particles in the layer can be a source of recombination and can significantly deteriorate the performance of pho-
tovoltaic devices with metal assisted etched silicon. On the other hand, the anodic etch technique does not require any catalyst and
the etch process is supported only by an external bias. Because of this, the anodic etch is highly appealing for preparation of a cheap
porous silicon surface. Despite the many research activities connected with the anodic etching of silicon [10,12,13], there is still a
lack of understanding of the etching mechanism and of the influence of the etching parameters on the obtained silicon structure.
Raman spectroscopy, when utilized to characterize porous silicon structures [14], is an ideal technique for this application
because it is simple, fast and mainly contactless. Many of the physical properties of porous silicon depend on the porosity. Around

⁎
Corresponding author.
E-mail address: juraj.breza@stuba.sk (J. Breza).

https://doi.org/10.1016/j.ijleo.2018.08.084
Received 19 June 2018; Accepted 23 August 2018
0030-4026/ © 2018 Elsevier GmbH. All rights reserved.
M. Kadlečíková et al. Optik - International Journal for Light and Electron Optics 174 (2018) 347–353

fifteen techniques have been identified to quantify porosity [15]. In this experimental work we report the results of Raman spec-
troscopy of porous silicon with different etching times and confirm the results with SEM micrographs of the samples.

2. Theoretical background for evaluation of porous silicon by Raman spectroscopy

The strong Raman signal at about 520 cm–1 can be used to monitor porous silicon. This is the dominant Raman band arising from
the interaction of monochromatic light with the crystal lattice. It is described as the first order scattering on the phonon in the middle
of the Brillouin zone (Γ) obeying the selection rule |k| ≈ 0. The frequency of the Raman band corresponds to the frequency of
phonons in Γ, which is 15.56 THz (520.7 cm–1). The band is triple degenerate because three optical phonons take part in the
scattering [16,17]. Straight explanation of the mechanism of scattering is provided by Feynman diagrams [18]. Two processes are
most likely:

• the exciting photon generates an electron (and a hole) that subsequently loses or gains energy in the interaction with the lattice
and then recombines in the form of a radiative transition in the Stokes or anti-Stokes range,
• the exciting photon generates (annihilates) a phonon and is radiated as a Stokes or anti-Stokes photon.
The Raman peaks of crystalline semiconductors are sharp because of the well-defined wave vectors of the phonons. Disorders in
the crystal (defects) or a small crystallite size – down to amorphousness – result in asymmetrical peak broadening and a shift to lower
frequencies. As a result, the Raman spectrum offers direct information on the crystallinity of the samples.
The Raman spectra of silicon for crystalline Si (c-Si), polycrystalline Si (poly-Si) and amorphous Si (a-Si) are clearly different. The
c-Si Raman peak is sharp and symmetric, poly-Si Raman peak is shifted towards a lower frequency and is asymmetrically broadened
towards the low frequency side. The intensity of this peak is also lower than the one of c-Si. This is due to a shorter penetration depth
of the exciting laser light in poly-Si compared with c-Si, because of a higher absorption coefficient. The smaller the crystal size, the
larger the downshift of the peak, the decrease in intensity, and the increase in asymmetry. This effect is extreme for a-Si, which shows
a broad weak peak [19].

3. Experiment

Porous silicon was obtained by electrochemical etching of single-crystalline n-type antimony-doped silicon substrates with (100)
crystal orientation and resistivity of approx. 0.02 Ωcm. The etching process was realized in solutions of H2O2 and HF while using an
auxiliary catalytic platinum electrode (the positive potential was on silicon). The thickness of the Si substrates was 487–563 μm. Four
samples with varied etch times were prepared at an external bias of 2.3 V and measured current of 50 mA. Etch times of 1, 2.5, 5 and
10 min were used for samples denoted as No 2/3/1, 2/3/2, 2/3/3 and 2/3/4, respectively. All etch experiments were carried out at
room temperature. Etch conditions are summarized in Table 1.
Samples No 2/3/1, 2/3/2, 2/3/3 and 2/3/4 were examined by a Raman spectrometer JobinYvon Labram 300 equipped with a
confocal microscope Olympus BX-40, He-Ne excitation source (632.8 nm) and a grid monochromator (1800 grooves/mm). The lateral
resolution is given by the diameter of the laser beam. The diameter of the laser spot was approx. 1 μm with a 100× objective and the
laser power was approx. 5 mW. The depth resolution depends on the penetration depth of the exciting light into the material, thus on
its wavelength, on the optics used and on the optical properties of the material under study. Raman spectroscopy was conducted at
room temperature. The apparatus was calibrated to the band at 520.7 cm–1 of the reference single crystalline silicon. This band is
symmetrical and the FWHM value was approx. 3 cm–1.
The morphology was visualized by a scanning electron microscope Jeol JSM 7500F equipped with a field emission electron source
using 5 kV and 15 kV accelerating voltages and a current beam of 20 μA.
The porosity can be estimated from SEM images. Table 1 shows the experimental details and porosity measurements of each
sample by analyzing the SEM images by the open source image processing program ImageJ.
Finally, reflectivity of the samples of porous silicon was measured in the visible range from 300 to 800 nm. The source of light was
a halogen lamp with a power of 100 W. Light was transmitted by optical fibre R200-7-SR and its intensity was measured by Ocean
Optics spectrometer USB650UV.

Table 1
Samples with different formation conditions of the surface of Si and their properties.
Sample No Electrochemical formation Porosity from SEM image (%) Raman peak positions FWHM
of porous silicon (cm–1) of Raman peak (cm–1)

Time (min) Voltage (V) Current (mA)

2/3/1 1 2.3 50 2 519.8 4

2/3/2 2.5 2.3 50 4 518.8 5
2/3/3 5 2.3 50 7 517.7 7
2/3/4 10 2.3 50 11 515.6 13

348
M. Kadlečíková et al. Optik - International Journal for Light and Electron Optics 174 (2018) 347–353

Fig. 1. Raman spectra of porous silicon, samples No 2/3/1, 2/3/2, 2/3/3 and 2/3/4.

4. Results

Fig. 1 shows the Raman spectra of four Si substrates prepared for characterization of porous silicon. Micrographs in Figs. 2–5
taken in combined secondary and backscattered electron modes with sample-to-detector distance 4 mm–6 mm and with magnifi-
cation 100.000× show the surface of porous silicon. Due to etching and (100) crystal wafer orientation, the pores are perpendicular
to the surface. SEM shows the change of the morphology of the samples after chemical etching. On the basis of SEM visualization,
electrochemical destruction of the substrate increasing with the etching time as well as the correlation between the microstructure of
the silicon wafer and the shape and position of their Raman spectra are obvious. The intensity of the Raman dominant silicon band
decreases and the bandwidth is shifted to lower frequencies, depending on the morphology of the sample. FWHMs calculated from
the Gaussian approximation of single bands are in Table 1.
We assessed the porosity of substrates based on SEM images. For this purpose, we defined the porosity as the ratio of the area of
the displayed pores in the image to the total area in percentage. Pore areas, histograms, and porosity were evaluated in ImageJ
software under identical conditions for all four SEM micrographs. We maximized the contrast between the pores and the backgrounds
and neglected images with a size of about 1 nm, which we consider to be non-evaluable noise. Based on the method used, the
histograms do not report the pore depth distribution. The calculated porosity of the samples accompanied by a trend line, depending
on the preparation time, is shown in Fig. 6. Approximately 11% porosity was achieved after 10 min of etching. Circular pore dis-
tribution histograms are shown in Fig. 7. From the estimation of the porosity from the SEM images, the following findings arise.
Analysis of the Feret diameters of the pores did not reveal their preferred directional dimension. In the case of substrate 2/3/1, the
distribution of the pores is different from the other cases. We attribute this to the short formation time when the pores are still only
being formed as germs on the surface of the substrate. On substrates 2/3/2–2/3/4, the increase in the number of germs and the
increase in pore size are visible, depending on the etching time. In the Raman spectrum of PS substrates, this tendency was manifested
by a change in the intensity of the dominant Raman band, by its shift to lower frequencies and by band asymmetry. The decrease in
the intensity of the dominant Raman band PS may also be related to the change in the penetration depth of the exciting radiation and
its absorption. Raman band broadening points at a change in the silicon structure. Band broadening has been observed in amorphous
silicon [19,20] but also in nanocrystalline silicon [21].
According to [22], the Raman band of silicon gets broader, shifts and is centred at 470–495 cm–1 in the case of the amorphous
phase. Amorphous silicon has random states in the atomic arrangement with a wider array of bond angles, bond energies, and bond

Fig. 2. SEM image of sample No 2/3/1.

349
M. Kadlečíková et al. Optik - International Journal for Light and Electron Optics 174 (2018) 347–353

Fig. 3. SEM image of sample No 2/3/2.

Fig. 4. SEM image of sample No 2/3/3.

Fig. 5. SEM image of sample No 2/3/4.

lengths as well as dangling bonds. The distribution of possible states leads to a abroad Raman band. According to [23], the Gaussian-
shaped band at 470 cm–1 and FWHM of 55 cm–1 in the Raman spectrum represents the amorphous phase in the structure on reaching
74% porosity of the sample. In our case, with respect to the time of destruction, the porosity of the substrate is lower than in the
above study. Deformation of the Raman spectrum of substrates 2/3/1–2/3/4, however, indicates the tendency that the proportion of
crystalline regions decreases with the surface formation time. Thus, in accordance with [19,20,24] we believe in the mechanism of
amorphization of Si substrates by chemical destruction.
Fig. 8 presents an inverse correlation between porosity and optical reflection for three wavelengths. Reflectivity is defined as a
ratio of the reflected light intensity to the intensity of impinging light. Light reflection takes place on two interfaces, thus between air

350
M. Kadlečíková et al. Optik - International Journal for Light and Electron Optics 174 (2018) 347–353

Fig. 6. Porosity of samples versus the time of preparation, and the trend line.

Fig. 7. Pore size distribution as evaluated from SEM micrographs in software ImageJ.

and the porous layer, and between the porous layer and c-Si. This simplification means that we do not consider the existence of native
oxide and c-Si is for the used wavelengths considered non-reflecting. The experiment proves that the ratio of reflected and impinging
light intensity drops down with a growing porosity of the samples. The higher porosity means also a growing surface roughness
reducing the light reflection. In view of the light wavelength that is comparable with the estimated sizes of pores, diffuse light
reflection takes place.

5. Conclusion

Porous silicon was obtained by electrochemical etching of single-crystalline n-type antimony-doped silicon substrates with (100)
crystal orientation. We used four methods to confirm the relation between the porosity of silicon and the time of etching. SEM
visualization revealed electrochemical destruction of the substrate surface that increased with the etching time. In dependence on the
etching time, a rising number of germs as well as an increasing size of pores were observed. In the Raman spectra of PS substrates,
these changes were manifested as varying intensities of the dominant Raman band. We recorded also a correlation between the

351
M. Kadlečíková et al. Optik - International Journal for Light and Electron Optics 174 (2018) 347–353

Fig. 8. Reflectivity of porous silicon samples versus porosity.

porosity (in percentage) and the intensity of the dominant Raman band, its shift to lower wavenumbers and band asymmetry. These
observations give evidence that the fraction of crystalline areas decreases with the time of surface formation. Unlike in Refs. [25,26],
where porosity was evaluated from gravimetric measurements, we used a simple automatic method provided by software ImageJ.
In general, porous silicon raises quite few very interesting theoretical and practical questions. Our analysis of Raman scattering on
optical phonons close to the wavenumber of 520.7 cm–1 proves the feasibility and contribution of Raman spectroscopy to creating a
qualitative model of the structure and morphology of the porous layer. We verified the relation between the porosity of silicon and
the changing Raman spectra of the material. The result is non-homogeneous broadening of the Raman band and its shift to lower
frequencies. It is well known that if the size of nanocrystallites is between 2 and 6 nm, the spectrum approaches that of amorphous
silicon. This is why we conclude that the mechanism of amorphization of Si substrates is in chemical destruction. Undoubtedly, even
in the case of low porosity, in our case 11%, non-destructive Raman spectroscopy is a fast tool for revealing the morphology of the
porous layer.

Acknowledgements

We are grateful to the Scientific grant agency of the Ministry of Education, Science, Research and Sport of the Slovak Republic for
financial support of project VEGA No. 1/0947/16 and the Slovak Research and Development Agency for financial support of projects
APVV-15-0152, APVV-16-0319 and APVV-17-0169.

References

[1] A. Mirzaei, Sung Yong Kang, Sun-Woo Choi, Yong Jung Kwon, Myung Sik Choi, Jae Hoon Bang, Sang Sub Kim, Hyoun Woo Kim, Fabrication and gas sensing
properties of vertically aligned Si nanowires, Appl. Surf. Sci. 427 (2018) 215–226.
[2] H.J. In, C.R. Field, P.E. Pehrsson, Periodically porous top electrodes on vertical nanowire arrays for highly sensitive gas detection, Nanotechnology 22 (2011)
355501.
[3] K. Peng, J. Jie, W. Zhang, S.T. Lee, Silicon nanowires for rechargeable lithium-ion battery anodes, Appl. Phys. Lett. 93 (2008) 033105.
[4] B.M. Bang, H. Kim, H.K. Song, J. Cho, S. Park, Scalable approach to multi-dimensional bulk Si anodes via metal-assisted chemical etching, Energy Environ. Sci. 4
(2011) 5013–5019.
[5] Kentaro Imamura, Takaaki Nonaka, Yuya Onitsuka, Daichi Irishika, Hikaru Kobayashi, Light trapping of crystalline Si solar cells by use of nanocrystalline Si layer
plus pyramidal texture, Appl. Surf. Sci. 395 (2017) 50–55.
[6] A.A. Abdul-Hameed, M.A. Mahdi, Basil Ali, Abbas M. Selman, H.F. Al-Taay, P. Jennings, Wen-Jen Lee, Fabrication of a high sensitivity and fast response self-
powered photosensor based on a core-shell silicon nanowire homojunction, Superlattices Microstruct. 116 (2018) 27–35.
[7] S.-W. Chang, V.P. Chuang, S.T. Boles, C.A. Ross, C.V. Thompson, Densely packed arrays of ultra-high-aspect-ratio silicon nanowires fabricated using block-
copolymer lithography and metal-assisted etching, Adv. Funct. Mater. 19 (2009) 2495–2500.
[8] M.L. Zhang, K.Q. Peng, X. Fan, J.S. Jie, R.Q. Zhang, S.T. Lee, N.B. Wong, Preparation of large-area uniform silicon nanowires arrays through metal-assisted
chemical etching, J. Phys. Chem. C 112 (2008) 4444.
[9] X. Li, Y. Xiao, C. Yan, K. Zhou, S.L. Schweizer, A. Sprafke, J.H. Lee, R.B. Wehrspohn, Influence of the mobility of Pt nanoparticles on the anisotropic etching
properties of silicon, ECS Solid State Lett. 2 (2013) P22–P24.
[10] K. Rajkumar, R. Pandian, A. Sankarakumar, R. Thangavelu, R. Kumar, Engineering silicon to porous silicon and silicon nanowires by metal-assisted chemical
etching: role of Ag size and electron-scavenging rate on morphology control and mechanism, ACS Omega 2 (2017) 4540–4547.
[11] Chang Quan Lai, Wen Zheng, W.K. Choi, C.V. Thompson, Metal assisted anodic etching of silicon, Nanoscale 7 (2015) 11123–11134.
[12] W.B. Hu, W. Zhao, J.L. Fan, S.L. Wu, J.T. Zhang, Emission properties of porous silicon electron emitters formed by pulsed anodic etching, J. Electron. Mater. 46
(2017) 895–902.
[13] S. Nöhren, E. Quiroga-González, J. Carstensen, H. Föll, Fabrication and characterization of silicon microwire anodes by electrochemical etching techniques, ECS
Trans. 66 (2015) 27–38.
[14] M. Ivanda, Raman spectroscopy of porous silicon, in: L. Canham (Ed.), Handbook of Porous Silicon, Springer, Cham, 2016.
[15] J. Peckham, G.T. Andrews, Simple optical method to determine the porosity of porous silicon films, Thin Solid Films 520 (2012) 2526–2531.
[16] L.P. Temple, C.E. Hathaway, Multiphonon Raman spectrum of silicon, Phys. Rev. B 7 (1973) 3685–3697.
[17] P. Giannozzi, S. de Gironcoli, Ab initio calculations of phonon dispersions in semiconductors, Phys. Rev. B 43 (1991) 7231–7242.
[18] W.H. Weber, R. Merlin, Raman Scattering in Materials Science, Springer-Verlag, Berlin, 2000.
[19] M.J. Pelletier, Analytical Applications of Raman Spectroscopy, Wiley-Blackwell, 1999.
[20] R.S. Dariani, Z. Ahmadi, Study of porous silicon structure by Raman scattering, Electron. Opt. Control. 124 (2013) 5353–5356.

352
M. Kadlečíková et al. Optik - International Journal for Light and Electron Optics 174 (2018) 347–353

[21] F. Zhong, Z.H. Jia, Raman scattering study on pristine and oxidized n-type porous silicon, Physica B Condens. Matter 411 (2013) 77–80.
[22] W.E. Hong, J.S. Ro, Kinetics of solid phase crystallization of amorphous silicon analyzed by Raman spectroscopy, J. Appl. Phys. 114 (2013) 073511.
[23] O. Bisi, S. Ossicini, L. Pavesi, Porous silicon: a quantum sponge structure for silicon based optoelectronics, Surf. Sci. Rep. 38 (2000) 1–126.
[24] H. Kim, N. Cho, Morphological and nanostructural features of porous silicon prepared by electrochemical etching, Nanoscale Res. Lett. 7 (2012) 408.
[25] C. Lai, X. Li, C. Liu, X. Gao, Z. Xiang, L. Zou, Improvement in gravimetric measurement for determining the porosity and thickness of porous silicon using an
optimized solution, Mater. Sci. Semicond. Process. 26 (2014) 501–505.
[26] Y. Al-Douri, N. Badi, C.H. Voon, Etching time effect on optical properties of porous silicon for solar cells fabrication, Optik 147 (2017) 343–349.

353