- Email: [email protected]
Porosity estimation of alumina samples based on resonant backscattering spectrometry
Nuclear Instruments and Methods in Physics Research B 373 (2016) 80–84
Contents lists available at ScienceDirect
Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb
Porosity estimation of alumina samples based on resonant backscattering spectrometry F. Mokhles Gerami a,⇑, O. Kakuee b, S. Mohammadi a a b
Department of Physics, Payame Noor University, P. O. Box 19395-3697, Tehran, Iran Physics & Accelerator Research School, Nuclear Science and Technology Research Institute, P. O. Box 14395-836, Tehran, Iran
a r t i c l e
i n f o
Article history: Received 23 November 2015 Received in revised form 29 February 2016 Accepted 8 March 2016 Available online 17 March 2016 Keywords: Porosity Porous alumina Structure Characterization Resonance Backscattering
a b s t r a c t In this work, columnar porous alumina samples were investigated using the 16O(a,a)16O resonance scattering at 3.045 MeV. If the incident energy is slightly above the resonance energy, a resonance peak appears in the energy spectra of the backscattered ions. The position and width of this peak for nonporous samples are mainly determined by the experimental setup, whilst for porous materials, the peak position shifts towards higher energies under certain conditions. This effect can be explained by the lower amount of material which the ions encounter along the backscattered trajectories. The energy shift of the resonance peak towards higher energies was revealed experimentally and discussed theoretically. The estimated porosities of the samples based on this energy shift were compared with those evaluated from the graphical analysis of the images obtained by field emission scanning electron microscopy. Ó 2016 Elsevier B.V. All rights reserved.
1. Introduction In recent years, porous alumina samples have attracted great attention due to their potential applications in a wide range of fields, such as micro- and optoelectronics, bionanotechnology, chemical sensing, gas separation, catalysis, purifications, etc. [1,2]. The most common techniques for pore characterization include: mercury porosimetry [3], adsorption and BET analysis [4], Ellipsometric Porosimetry (EP) [5], Transmission Neutron and X-ray Scattering (TNS and TXS) [6], and visualization techniques such as Transmission Electron Microscopy (TEM), Scanning Tunneling Microscopy (STM) and Atomic Force Microscopy (AFM) [7]. Ion beam analysis (IBA) is a fast, non-destructive, and standardless technique, which can be used for characterization of porous materials by means of collecting energy spectra of charged particles [8–13]. Since ions neither loose energy nor scatter in hollow cavities of porous materials, the backscattering spectrometry cannot be sensitive to gain information on the porous structure [14]. However, it has been shown that in the energy spectra of the backscattered particles, an extra ‘‘structure induced energy spread” appears due to fluctuating amount of material crossed by the individual ions on their way in and out . This extra ‘‘energy spread” which is easily observed in resonant backscattering measurements can be used to gain information on the structure and morphologi⇑ Corresponding author. E-mail address: [email protected] (F. Mokhles Gerami). http://dx.doi.org/10.1016/j.nimb.2016.03.016 0168-583X/Ó 2016 Elsevier B.V. All rights reserved.
cal details of porous samples. In other words, any peak width of the resonant backscattering spectrum depends on the structure of the porous materials. Therefore, it is possible to characterize the porestructure by performing resonant backscattering measurements [14–18]. In this study, samples of anodic porous alumina having a selfordered honeycomb structure with cylindrical pores were analyzed. Anodizing of aluminum generates a porous alumina layer containing cylindrical parallel pores extending essentially perpendicular to the substrate [19]. In order to investigate porosity, the resonant elastic scattering peak of the 16O(a,a)16O at the energy of 3.045 MeV was examined. The 4He+ backscattering spectra collected in porous and non-porous alumina samples under the same experimental conditions were compared. It is well known that columnar structure and extra ‘‘structure induced energy spread” could affect the resonance backscattering spectra. Using this fact, the porosity of alumina samples was estimated based on resonant backscattering spectrometry. Finally, the results are compared to the surface porosities estimated from the images obtained by Field Emission Scanning Electron Microscope (FESEM).
2. Theoretical approach To approach the subject, we consider 4He+ beam irradiation of porous and non-porous alumina samples under the same
F. Mokhles Gerami et al. / Nuclear Instruments and Methods in Physics Research B 373 (2016) 80–84
experimental conditions. In this method, the narrow resonance (10 keV) of a scattering from 16O at 3.045 MeV is used [20]. In fact, the incident beam with an excess energy of DE above the resonance energy (Er) crosses inward the sample, slows down to the energy Er and then backscatters with high probability. The observed differences between the resonant peak in the spectra of porous and non-porous samples will be used to obtain information on the porosity of the porous sample. The following theoretical consideration can help to estimate the porosity of alumina samples based on the resonant backscattering spectrometry: The ions irradiated porous sample with cylindrical pores, perpendicular to the sample surface. The individual ions, however, can cross various amount of material travelling in pores and pore walls to reach a given geometrical depth; or while they lose a well defined energy, as necessary in this resonance case, the ion will reach various geometric depths, as indicated in Fig. 1. In this case, the amount of crossed material will depend on the angle between the ion direction and pore direction. Let us consider if the ions reach the sample in perpendicular incidence (a = 0°, parallel to the pores) and the ions impinging to the pore walls. In this case the ions will travel only pore wall in the inward path while they reach the resonance energy, therefore the resonance event take place at the same depth for porous and non-porous materials. They reach the resonance energy, Er at the geometry depth X:
X¼
ðE0 Er Þ cos a S1
XS2 ð1 PÞ cos b
XS2 cos b
ð3Þ
In an ideal columnar porous sample, where the columns are thoroughly straight and parallel and the pore walls are smooth; if the beam is irradiated parallel to the columns, the ions in the pores will be travelling deeply without interactions, so they will not contribute to the resonance peak. Of course, the columns should be long enough, so the scattered ions at the end of the pores will not be able to reach the detector at the resonant peak energies. On the other hand, the incoming ions in the walls slow down to the resonance energy without crossing the pores. As a result, the depth ions travel and slow down to the resonance energy, X, for both porous and non-porous samples (in Eqs. (2) and (3)) are the same. By combining the Eqs. (2) and (3), the following expression can be easily derived:
P¼
Edp Ednon kEr Ednon
ð4Þ
Based on this equation, the porosity can be estimated from the difference between detected peak energies in porous and nonporous samples. Therefore, this theoretical approach can be followed for estimation of the porosity of a columnar porous sample when the incident ions remain in the walls in the inward path, reaching Er at a well-defined depth, and passing several pores and walls at outgoing the material.
ð1Þ
where E0 is the incident ion energy, a is the incident angle, S1 is the stopping power of He in alumina at energy of E0. The detected energy for the porous sample, Edp is:
Edp ¼ kEr
Ednon ¼ kEr
81
ð2Þ
where k is the kinematic factor of scattering, b is the exit angle, S2 is the stopping power of He in alumina at energy of kEr and P is the sample porosity (the volume fraction of the pores). In the latter equation, the homogeneous approximation is used, because the scattered ions cross several pores and walls in their outgoing way toward the detector placed to a scattering angle of 165°. For the non-porous sample, P is equal to zero. Therefore, the detected energy for the non-porous sample, Ednon is:
Fig. 1. A schematic representation of columnar porous material, with individual ions at incident angle a, exit angle b, and the average resonance depth X.
3. Experiments, results and discussions In this work, to demonstrate our methodology, three of the investigated samples are illustrated including: two porous alumina films prepared by DC anodization of pure aluminum and a nonporous alumina sample. The non-porous alumina sample was prepared by pulsed unipolar plasma electrolytic oxidation (PEO) in an electrolyte containing 10 gL1 NaAlO2 and 1gL1 KOH in deionized water. A current density of 0.1 A cm2 was applied for 20 min. The temperature of the bath was kept at 30 °C during the electrooxidation. The frequency and duty cycle of the pulses were 50 Hz and 80%, respectively. The low porosity anodized alumina (sample 1) was grown on high purity (99.999%) aluminum foil via a typical ‘‘mild anodization” process. First, a 0.3 mm-thick Al foil was degreased by multiple washing with acetone/ethanol. It was then electropolished in a stirring mixture of perchloric acid and ethanol (1:4) at 0 °C. The electropolishing step was carried out at 20 V constant potential and was continued for 60 s until a clean shiny surface was obtained. Finally, the anodization was performed in a 0.3 M oxalic acid (H2C2O4) solution at 40 V for 10 h. The higher porosity sample (sample 2) was prepared by a two-step anodization process [21]. The aluminum foil was electropolished (as described above) prior to the first anodization which was performed in 0.3 M oxalic acid at 40 V and 17 °C for 5 h. The oxide layer grown in the first step was stripped off in a solution containing 0.2 M H2Cr2O4 and 0.5 M H3PO4 by mild heating at 60 °C for 10 h. The second anodization was then performed under the same conditions as the first for 3 h. Finally, the sample was dipped into a 0.3 M phosphoric acid (H3PO4) at 30 °C for 40 min to widen the pores. The porous aluminum oxide samples were studied by Elastic Backscattering Spectroscopy (EBS) technique using 4He+ ions accelerated by the 3 MV Van de Graaff accelerator of Nuclear Science and Technology Research Institute (NSTRI) in Tehran. All the three samples were irradiated by 3.195 MeV 4He+ ions at normal incidence with a beam current of about 10 nA under the same experimental conditions. The experimental chamber with an inner diameter of 440 mm and height of 300 mm was evacuated to the order of 106 torr using the combination of a rotary backing pump
82
F. Mokhles Gerami et al. / Nuclear Instruments and Methods in Physics Research B 373 (2016) 80–84
and a turbomolecular pump. The chamber is equipped with a sample manipulator which can rotate the sample around three independent axes by means of stepping motors outside the vacuum system. The precision and range of the three rotational movements are: (h), a complete cycle (360° ± 0.018°) rotation around the vertical axis of the sample manipulator R1; (u), a half cycle (180° ± 0.0125°) rotation around the axis of the incident ion beam R2 and (c), tilting around the axis perpendicular to both R1 and R2 in the range of ±3.5° with a precision of 0.0071°. In addition, a 3axis goniometer outside the experimental chamber and attached to its bottom plate is provisioned for translating the sample (±10 mm) along the 3 axes of x, y and z with the precision of 0.01 mm. The backscattered alpha particles are detected by a surface barrier detector with a 14 keV energy resolution placed at a scattering angle of 165°. The experiments were carried out in an
advanced channeling chamber introduced earlier. However, by adjusting the widths of the two collimating slits in the beam path, the beam divergence was intentionally increased to 0.2° so that the ions hitting the sample belong to one of the following categories: those flying in the pores parallel to the columns –aligned ions-, those passing through the walls parallel to the columns –aligned ions-, and finally those randomly passing through both pores and walls –nonaligned ions-. The ions of the first category are more likely to be lost inside the pores while the two latters could contribute in the resonant backscattering peak. In Figs. 2 and 3, the 16O(a,a)16O resonance backscattering spectra of non-porous alumina is compared with those of samples 1 and 2, respectively.As can be seen in these figures, the backscattering resonant peaks in the porous samples are broader than that in the non-porous sample. In addition, comparing Figs. 2 and 3, the
Fig. 2. Experimental backscattering spectra of the columnar porous alumina 1 in comparison to the non-porous alumina. The peak in porous alumina is shifted towards higher energies. The inset shows the resonant peak spectra of porous alumina were fitted with two Gaussian peaks.
Fig. 3. Experimental backscattering spectra of the columnar porous alumina 2 in comparison to the non-porous alumina. The peak in porous alumina is shifted towards higher energies. The inset shows the resonant peak spectra of porous alumina were fitted with two Gaussian peaks.
83
F. Mokhles Gerami et al. / Nuclear Instruments and Methods in Physics Research B 373 (2016) 80–84 Table 1 Characteristics of the samples (The uncertainty in the measured porosity is 5%). Sample
Non-porous alumina Porous alumina 1 Porous alumina 2
Surface porosity from FESEM images
17% 41%
peak width in the higher porosity sample (sample 2 in Fig. 3) is broader than that in the lower porosity sample (Sample 1 in Fig. 2). This broadening effect is related to path fluctuations of individual ions. In fact, in a porous sample in comparison to a nonporous sample, each ion crosses different amounts of matter along its path and this is more significant in a higher porosity sample [17]. Another effect, which can be seen in Figs. 2 and 3, is the resonance peak shift of the porous sample in comparison to the nonporous sample. This effect has already been explained in the ‘‘theoretical approach” section and can be seen in Fig. 6 of the Ref. [12], but Pesiri et al., did not pay attention to it in their work. Besides, the displacement of the peak towards the surface when the beam is parallel to the columns at a porous sample tilt of 0° was reported by Paszti et al., [13,15]. It may be noticed that two peaks contribute in the formation of the porous alumina spectrum (Figs. 2 and 3). When the incoming ions travel parallel to the columns, one resonance peak will appear in the backscattering spectra [15]. If there is a slight tilt angle between the pore direction and the incident beam, two peaks will appear in the backscattering spectra. The peak at higher energies is related to the ions moving in the walls and slowing down to Er at a well determined depth without passing the pores, while the peak at lower energies belongs to the nonaligned ions passing through both pores and walls slowing down to Er in a deeper depth [13,15]. It should be noted that when the contribution of the nonaligned ions increases, the area of the lower energy peak could be scaled up. The amplified resonance backscattering yield is due to the interactions of greater number of nonaligned ions having high cross section values at resonance energy with the oxygen atoms in the sample. Moreover, the nonaligned ions see the surface of the porous sample as a laterally nonuniform layer leading to an increased broadening of the lower energy peak [22]. Since at the above-mentioned conditions, incoming ions reach the resonance energy at the same depth in the porous and nonporous samples but outgoing ions in the porous samples cross less amount of matter than that in the non-porous sample, the peak in
Peak centers (keV) resulting from fitting At lower energies
At higher energies
770 ± 13 846 ± 21
848 ± 1 898 ± 1 974 ± 5
Porosity from Eq. (4)
19% 47%
the porous alumina shifts towards higher energies. This energy difference is related to the sample porosity. We use Eq. (4) to determine the porosity of the samples through the procedure mentioned in section 2 (theoretical approach). The Eq. (4) can be applied for an ideal columnar porous sample where the incoming ions travel parallel to the columns. However, as can be seen in Figs. 2 and 3, two peaks appeared in the porous alumina spectrum. This is obvious, since the sample does not have an ideal structure or there is a slight tilt angle between the pore direction and the incident beam. In fact, the main requirement for estimation of porosity through resonant backscattering is to be able to identify the peak corresponding to the ions travelling in the walls and parallel to the columns, slowing down to Er. It is assumed that the natural resonance width is very small compared to the energy spread of the outgoing ions. In this case, each spectrum is fitted with the Gaussian curve after subtracting a linear background from it. Thus, the resonance peak in the spectra of porous samples was fitted with two Gaussian peaks by the program ORIGIN (Microcal software). The results of fitting are shown in Table 1. Since the peak at higher energies was produced by the incoming ions crossing the walls, we used the information of only this peak in Eq. (4). In fact, the peak at lower energies was not applied (this is related to the ions passing both the walls and the pores). The calculated porosities from Eq. (4) are shown in Table 1. FESEM plan-view images of the two porous samples are presented in Fig. 4. Surface porosity is defined as the area fraction of the pores in a plane cross section of the porous sample which can be estimated from FESEM images using ImageJ, a public domain Java image processing and analysis program [23]. The results are shown in Table 1. The uncertainty in the measured porosity through FESEM imaging is around 5% and the estimated uncertainty through resonant backscattering is about the same value. It should be noted that the uncertainty in the measured porosity by resonant backscattering method is subject to a number of factors [24] including beam energy (0.1%), calibration of detection electronics (0.1%), the Guas-
Fig. 4. FESEM plan-view images of (a) porous alumina 1(b) porous alumina 2.
84
F. Mokhles Gerami et al. / Nuclear Instruments and Methods in Physics Research B 373 (2016) 80–84
sian peak fitting (<3%), stopping power (4%) [25], and geometric error. The latter factor can be described as a combination of pore direction misalignment with respect to both the beam incident and the exit direction. In this work, a one degree variation in the incident beam angle, leads to a 0.01% resonance depth variation and a two degree deviation (due to acceptance angle of the detector) in the scattering angle will result 0.07% deviation in the outgoing path length which is negligible in the total uncertainty. It should be noted that in the preparation of high porosity samples, a further process known as ‘‘pore widening” is applied on the samples, leading to the curved edges of the columnar pores. Then the edges of the pores would not be sharp enough to be easily determined and hence the porosity measured through FESEM imaging will be accompanied with a higher uncertainty. 4. Conclusions The results of porosity estimation of the self-ordered anodic alumina films with cylindrical pores based on resonant backscattering spectrometry are reported. The broadening in the resonance peak of 16O(a,a)16O reaction in the porous alumina samples due to structure induced energy spread as well as an energy shift in the resonance peak were clearly observed. The experimental results indicate that the energy shift in the resonance peak appears when the beam is incident normal to the sample surface. In this geometry where the beam is partially parallel to the columns, the porosity can be estimated using the resonance peak shift observed for the porous sample with no need to have information on the pore size and the interpore distance. The measured porosities for samples of different porosities are consistent with the surface porosities evaluated based on the FESEM images. The slight difference between the measured and evaluated porosities becomes greater when the sample porosity increases and could be due to the error occurred in precise determination of the edge of the columns in the image analysis method. References [1] Y. Piao, H. Kim, Fabrication of nanostructured materials using porous alumina template and their applications for sensing and electrocatalysis, J. Nanosci. Nanotechnol. 9 (2009) 2215–2233. [2] A. Santos, T. Kumeria, D. Losic, Nanoporous anodic aluminum oxide for chemical sensing and biosensors, Trends Anal. Chem. 44 (2013) 25–37. [3] A.B. Abell, K.L. Willis, D.A. Lange, Mercury intrusion porosimetry and image analysis of cement-based materials, J. Colloid Interface Sci. 211 (1999) 39–44. [4] S. Brunauer, P.H. Emmett, E. Teller, The use of low temperature Van der Waals adsorption isotherm in determining surface area, J. Am. Chem. Soc. 60 (1938) 309.
[5] M.R. Baklanov, K.P. Mogilnikov, V.G. Polovinkin, F.N. Dultsev, Determination of pore size distribution in thin films by ellipsometric porosimetry, J. Vac. Sci. Technol., B 18 (2000) 1385–1391. [6] W.L. Wu, W.E. Wallace, E.K. Lin, G.W. Lynn, C.J. Glinka, E.T. Ryan, H.M. Ho, Properties of nanoporous silica thin films determined by high-resolution X-ray reflectivity and small-angle neutron scattering, J. Appl. Phys. 87 (2000) 1193– 1200. [7] B. Bhushan (Ed.), Scanning probe microscopy in nanoscience and nanotechnology 2, Springer Science & Business Media (2010). [8] S. Kumar, J.V. Ramana, C. David, V.S. Raju, Ion beam analysis of porous silicon layers, Nucl. Instrum. Methods Phys. Res. B 179 (2001) 113–120. [9] V. Torres-Costa, F. Pászti, A. Climent-Font, R.J. Martín-Palma, J.M. MartínezDuart, Porosity profile determination of porous silicon interference filters by RBS, Phys. Status Solidi (c) 2 (2005) 3208–3212. [10] V. Torres-Costa, R.J. Martín-Palma, F. Paszti, A. Climent-Font, J.M. MartínezDuart, In-depth RBS study of optical layers based on nanostructured silicon, J. Non-Cryst. Solids 352 (2006) 2521–2525. _ _ [11] H. Krzyzanowska, M. Kulik, Hydrogen and oxygen A.P. Kobzev, J. Zuk, concentration analysis of porous silicon, J. Non-Cryst. Solids 354 (2008) 4367–4374. [12] D.R. Pesiri, R.C. Snow, N. Elliott, C. Maggiore, R.C. Dye, The characterization of asymmetric alumina membranes by Rutherford backscattering spectrometry, J. Membr. Sci. 176 (2000) 209–221. [13] F. Pászti, E. Szilágyi, Pore structure investigations in porous silicon by ion beam analytical methods, Vacuum 50 (1998) 451–462. [14] F. Pászti, E. Szilágyi, A. Manuaba, G. Battistig, Application of resonant backscattering spectrometry for determination of pore structure changes, Nucl. Instrum. Methods Phys. Res. B 161 (2000) 963–968. [15] F. Pászti, E. Szilágyi, Z.E. Horváth, A. Manuaba, G. Battistig, Z. Hajnal, E. Vázsonyi, Morphological investigation of porous samples by resonant backscattering spectrometry, Nucl. Instrum. Methods Phys. Res. B 136 (1998) 533–539. [16] F. Pászti, G. Battistig, Ion beam characterisation and modification of porous silicon, Phys. Status Solidi (a) 182 (2000) 271–278. [17] Z. Hajnal, E. Szilágyi, F. Pászti, G. Battistig, Channeling-like effects due to the macroscopic structure of porous silicon, Nucl. Instrum. Methods Phys. Res. B 118 (1996) 617–621. [18] G. Amsel, E. d’Artemare, G. Battistig, V. Morazzani, C. Ortega, A characterisation of the morphology of porous silicon films by proton energy loss fluctuation measurements with a narrow resonance in the 15N (p, ac) 12C reaction, Nucl. Instrum. Methods Phys. Res. B 122 (1997) 99–112. [19] D. Kuang, R. Charrière, N. Matsapey, M. Flury, J. Faucheu, P. Chavel1, Modeling the specular spectral reflectance of partially ordered alumina nanopores on an aluminum substrate, in: Opt. Express 23 (4) (2015) 4506–4516. [20] J.R. Cameron, Elastic scattering of alpha-particles by oxygen, Phys. Rev. 90 (1953) 839. [21] H. Masuda, K. Fukuda, Ordered metal nanohole arrays made by a two-step replication of honeycomb structures of anodic alumina, Science 268 (1995) 1466–1468. [22] M. Mayer, Computer simulation of ion beam analysis of laterally inhomogeneous materials, Nucl. Instrum. Methods Phys. Res. B 371 (2016) 90–96. [23] T. Ferreira, W. Rasb, ImageJ user guide. (2012) from http://imagej.nih.gov/ ij/download.html, September 7, 2015. [24] C. Jeynes, N.P. Barradas, E. Szilágyi, Accurate determination of quantity of material in thin films by Rutherford backscattering spectrometry, Anal. Chem. 84 (2012) 6061–6069. [25] J.F. Ziegler, M.D. Ziegler, J.P. Biersack, SRIM – the stopping and range of ions in matter, Nucl. Instrum. Methods Phys. Res. B 268 (2010) 1818.
Contents lists available at ScienceDirect
Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb
Porosity estimation of alumina samples based on resonant backscattering spectrometry F. Mokhles Gerami a,⇑, O. Kakuee b, S. Mohammadi a a b
Department of Physics, Payame Noor University, P. O. Box 19395-3697, Tehran, Iran Physics & Accelerator Research School, Nuclear Science and Technology Research Institute, P. O. Box 14395-836, Tehran, Iran
a r t i c l e
i n f o
Article history: Received 23 November 2015 Received in revised form 29 February 2016 Accepted 8 March 2016 Available online 17 March 2016 Keywords: Porosity Porous alumina Structure Characterization Resonance Backscattering
a b s t r a c t In this work, columnar porous alumina samples were investigated using the 16O(a,a)16O resonance scattering at 3.045 MeV. If the incident energy is slightly above the resonance energy, a resonance peak appears in the energy spectra of the backscattered ions. The position and width of this peak for nonporous samples are mainly determined by the experimental setup, whilst for porous materials, the peak position shifts towards higher energies under certain conditions. This effect can be explained by the lower amount of material which the ions encounter along the backscattered trajectories. The energy shift of the resonance peak towards higher energies was revealed experimentally and discussed theoretically. The estimated porosities of the samples based on this energy shift were compared with those evaluated from the graphical analysis of the images obtained by field emission scanning electron microscopy. Ó 2016 Elsevier B.V. All rights reserved.
1. Introduction In recent years, porous alumina samples have attracted great attention due to their potential applications in a wide range of fields, such as micro- and optoelectronics, bionanotechnology, chemical sensing, gas separation, catalysis, purifications, etc. [1,2]. The most common techniques for pore characterization include: mercury porosimetry [3], adsorption and BET analysis [4], Ellipsometric Porosimetry (EP) [5], Transmission Neutron and X-ray Scattering (TNS and TXS) [6], and visualization techniques such as Transmission Electron Microscopy (TEM), Scanning Tunneling Microscopy (STM) and Atomic Force Microscopy (AFM) [7]. Ion beam analysis (IBA) is a fast, non-destructive, and standardless technique, which can be used for characterization of porous materials by means of collecting energy spectra of charged particles [8–13]. Since ions neither loose energy nor scatter in hollow cavities of porous materials, the backscattering spectrometry cannot be sensitive to gain information on the porous structure [14]. However, it has been shown that in the energy spectra of the backscattered particles, an extra ‘‘structure induced energy spread” appears due to fluctuating amount of material crossed by the individual ions on their way in and out . This extra ‘‘energy spread” which is easily observed in resonant backscattering measurements can be used to gain information on the structure and morphologi⇑ Corresponding author. E-mail address: [email protected] (F. Mokhles Gerami). http://dx.doi.org/10.1016/j.nimb.2016.03.016 0168-583X/Ó 2016 Elsevier B.V. All rights reserved.
cal details of porous samples. In other words, any peak width of the resonant backscattering spectrum depends on the structure of the porous materials. Therefore, it is possible to characterize the porestructure by performing resonant backscattering measurements [14–18]. In this study, samples of anodic porous alumina having a selfordered honeycomb structure with cylindrical pores were analyzed. Anodizing of aluminum generates a porous alumina layer containing cylindrical parallel pores extending essentially perpendicular to the substrate [19]. In order to investigate porosity, the resonant elastic scattering peak of the 16O(a,a)16O at the energy of 3.045 MeV was examined. The 4He+ backscattering spectra collected in porous and non-porous alumina samples under the same experimental conditions were compared. It is well known that columnar structure and extra ‘‘structure induced energy spread” could affect the resonance backscattering spectra. Using this fact, the porosity of alumina samples was estimated based on resonant backscattering spectrometry. Finally, the results are compared to the surface porosities estimated from the images obtained by Field Emission Scanning Electron Microscope (FESEM).
2. Theoretical approach To approach the subject, we consider 4He+ beam irradiation of porous and non-porous alumina samples under the same
F. Mokhles Gerami et al. / Nuclear Instruments and Methods in Physics Research B 373 (2016) 80–84
experimental conditions. In this method, the narrow resonance (10 keV) of a scattering from 16O at 3.045 MeV is used [20]. In fact, the incident beam with an excess energy of DE above the resonance energy (Er) crosses inward the sample, slows down to the energy Er and then backscatters with high probability. The observed differences between the resonant peak in the spectra of porous and non-porous samples will be used to obtain information on the porosity of the porous sample. The following theoretical consideration can help to estimate the porosity of alumina samples based on the resonant backscattering spectrometry: The ions irradiated porous sample with cylindrical pores, perpendicular to the sample surface. The individual ions, however, can cross various amount of material travelling in pores and pore walls to reach a given geometrical depth; or while they lose a well defined energy, as necessary in this resonance case, the ion will reach various geometric depths, as indicated in Fig. 1. In this case, the amount of crossed material will depend on the angle between the ion direction and pore direction. Let us consider if the ions reach the sample in perpendicular incidence (a = 0°, parallel to the pores) and the ions impinging to the pore walls. In this case the ions will travel only pore wall in the inward path while they reach the resonance energy, therefore the resonance event take place at the same depth for porous and non-porous materials. They reach the resonance energy, Er at the geometry depth X:
X¼
ðE0 Er Þ cos a S1
XS2 ð1 PÞ cos b
XS2 cos b
ð3Þ
In an ideal columnar porous sample, where the columns are thoroughly straight and parallel and the pore walls are smooth; if the beam is irradiated parallel to the columns, the ions in the pores will be travelling deeply without interactions, so they will not contribute to the resonance peak. Of course, the columns should be long enough, so the scattered ions at the end of the pores will not be able to reach the detector at the resonant peak energies. On the other hand, the incoming ions in the walls slow down to the resonance energy without crossing the pores. As a result, the depth ions travel and slow down to the resonance energy, X, for both porous and non-porous samples (in Eqs. (2) and (3)) are the same. By combining the Eqs. (2) and (3), the following expression can be easily derived:
P¼
Edp Ednon kEr Ednon
ð4Þ
Based on this equation, the porosity can be estimated from the difference between detected peak energies in porous and nonporous samples. Therefore, this theoretical approach can be followed for estimation of the porosity of a columnar porous sample when the incident ions remain in the walls in the inward path, reaching Er at a well-defined depth, and passing several pores and walls at outgoing the material.
ð1Þ
where E0 is the incident ion energy, a is the incident angle, S1 is the stopping power of He in alumina at energy of E0. The detected energy for the porous sample, Edp is:
Edp ¼ kEr
Ednon ¼ kEr
81
ð2Þ
where k is the kinematic factor of scattering, b is the exit angle, S2 is the stopping power of He in alumina at energy of kEr and P is the sample porosity (the volume fraction of the pores). In the latter equation, the homogeneous approximation is used, because the scattered ions cross several pores and walls in their outgoing way toward the detector placed to a scattering angle of 165°. For the non-porous sample, P is equal to zero. Therefore, the detected energy for the non-porous sample, Ednon is:
Fig. 1. A schematic representation of columnar porous material, with individual ions at incident angle a, exit angle b, and the average resonance depth X.
3. Experiments, results and discussions In this work, to demonstrate our methodology, three of the investigated samples are illustrated including: two porous alumina films prepared by DC anodization of pure aluminum and a nonporous alumina sample. The non-porous alumina sample was prepared by pulsed unipolar plasma electrolytic oxidation (PEO) in an electrolyte containing 10 gL1 NaAlO2 and 1gL1 KOH in deionized water. A current density of 0.1 A cm2 was applied for 20 min. The temperature of the bath was kept at 30 °C during the electrooxidation. The frequency and duty cycle of the pulses were 50 Hz and 80%, respectively. The low porosity anodized alumina (sample 1) was grown on high purity (99.999%) aluminum foil via a typical ‘‘mild anodization” process. First, a 0.3 mm-thick Al foil was degreased by multiple washing with acetone/ethanol. It was then electropolished in a stirring mixture of perchloric acid and ethanol (1:4) at 0 °C. The electropolishing step was carried out at 20 V constant potential and was continued for 60 s until a clean shiny surface was obtained. Finally, the anodization was performed in a 0.3 M oxalic acid (H2C2O4) solution at 40 V for 10 h. The higher porosity sample (sample 2) was prepared by a two-step anodization process [21]. The aluminum foil was electropolished (as described above) prior to the first anodization which was performed in 0.3 M oxalic acid at 40 V and 17 °C for 5 h. The oxide layer grown in the first step was stripped off in a solution containing 0.2 M H2Cr2O4 and 0.5 M H3PO4 by mild heating at 60 °C for 10 h. The second anodization was then performed under the same conditions as the first for 3 h. Finally, the sample was dipped into a 0.3 M phosphoric acid (H3PO4) at 30 °C for 40 min to widen the pores. The porous aluminum oxide samples were studied by Elastic Backscattering Spectroscopy (EBS) technique using 4He+ ions accelerated by the 3 MV Van de Graaff accelerator of Nuclear Science and Technology Research Institute (NSTRI) in Tehran. All the three samples were irradiated by 3.195 MeV 4He+ ions at normal incidence with a beam current of about 10 nA under the same experimental conditions. The experimental chamber with an inner diameter of 440 mm and height of 300 mm was evacuated to the order of 106 torr using the combination of a rotary backing pump
82
F. Mokhles Gerami et al. / Nuclear Instruments and Methods in Physics Research B 373 (2016) 80–84
and a turbomolecular pump. The chamber is equipped with a sample manipulator which can rotate the sample around three independent axes by means of stepping motors outside the vacuum system. The precision and range of the three rotational movements are: (h), a complete cycle (360° ± 0.018°) rotation around the vertical axis of the sample manipulator R1; (u), a half cycle (180° ± 0.0125°) rotation around the axis of the incident ion beam R2 and (c), tilting around the axis perpendicular to both R1 and R2 in the range of ±3.5° with a precision of 0.0071°. In addition, a 3axis goniometer outside the experimental chamber and attached to its bottom plate is provisioned for translating the sample (±10 mm) along the 3 axes of x, y and z with the precision of 0.01 mm. The backscattered alpha particles are detected by a surface barrier detector with a 14 keV energy resolution placed at a scattering angle of 165°. The experiments were carried out in an
advanced channeling chamber introduced earlier. However, by adjusting the widths of the two collimating slits in the beam path, the beam divergence was intentionally increased to 0.2° so that the ions hitting the sample belong to one of the following categories: those flying in the pores parallel to the columns –aligned ions-, those passing through the walls parallel to the columns –aligned ions-, and finally those randomly passing through both pores and walls –nonaligned ions-. The ions of the first category are more likely to be lost inside the pores while the two latters could contribute in the resonant backscattering peak. In Figs. 2 and 3, the 16O(a,a)16O resonance backscattering spectra of non-porous alumina is compared with those of samples 1 and 2, respectively.As can be seen in these figures, the backscattering resonant peaks in the porous samples are broader than that in the non-porous sample. In addition, comparing Figs. 2 and 3, the
Fig. 2. Experimental backscattering spectra of the columnar porous alumina 1 in comparison to the non-porous alumina. The peak in porous alumina is shifted towards higher energies. The inset shows the resonant peak spectra of porous alumina were fitted with two Gaussian peaks.
Fig. 3. Experimental backscattering spectra of the columnar porous alumina 2 in comparison to the non-porous alumina. The peak in porous alumina is shifted towards higher energies. The inset shows the resonant peak spectra of porous alumina were fitted with two Gaussian peaks.
83
F. Mokhles Gerami et al. / Nuclear Instruments and Methods in Physics Research B 373 (2016) 80–84 Table 1 Characteristics of the samples (The uncertainty in the measured porosity is 5%). Sample
Non-porous alumina Porous alumina 1 Porous alumina 2
Surface porosity from FESEM images
17% 41%
peak width in the higher porosity sample (sample 2 in Fig. 3) is broader than that in the lower porosity sample (Sample 1 in Fig. 2). This broadening effect is related to path fluctuations of individual ions. In fact, in a porous sample in comparison to a nonporous sample, each ion crosses different amounts of matter along its path and this is more significant in a higher porosity sample [17]. Another effect, which can be seen in Figs. 2 and 3, is the resonance peak shift of the porous sample in comparison to the nonporous sample. This effect has already been explained in the ‘‘theoretical approach” section and can be seen in Fig. 6 of the Ref. [12], but Pesiri et al., did not pay attention to it in their work. Besides, the displacement of the peak towards the surface when the beam is parallel to the columns at a porous sample tilt of 0° was reported by Paszti et al., [13,15]. It may be noticed that two peaks contribute in the formation of the porous alumina spectrum (Figs. 2 and 3). When the incoming ions travel parallel to the columns, one resonance peak will appear in the backscattering spectra [15]. If there is a slight tilt angle between the pore direction and the incident beam, two peaks will appear in the backscattering spectra. The peak at higher energies is related to the ions moving in the walls and slowing down to Er at a well determined depth without passing the pores, while the peak at lower energies belongs to the nonaligned ions passing through both pores and walls slowing down to Er in a deeper depth [13,15]. It should be noted that when the contribution of the nonaligned ions increases, the area of the lower energy peak could be scaled up. The amplified resonance backscattering yield is due to the interactions of greater number of nonaligned ions having high cross section values at resonance energy with the oxygen atoms in the sample. Moreover, the nonaligned ions see the surface of the porous sample as a laterally nonuniform layer leading to an increased broadening of the lower energy peak [22]. Since at the above-mentioned conditions, incoming ions reach the resonance energy at the same depth in the porous and nonporous samples but outgoing ions in the porous samples cross less amount of matter than that in the non-porous sample, the peak in
Peak centers (keV) resulting from fitting At lower energies
At higher energies
770 ± 13 846 ± 21
848 ± 1 898 ± 1 974 ± 5
Porosity from Eq. (4)
19% 47%
the porous alumina shifts towards higher energies. This energy difference is related to the sample porosity. We use Eq. (4) to determine the porosity of the samples through the procedure mentioned in section 2 (theoretical approach). The Eq. (4) can be applied for an ideal columnar porous sample where the incoming ions travel parallel to the columns. However, as can be seen in Figs. 2 and 3, two peaks appeared in the porous alumina spectrum. This is obvious, since the sample does not have an ideal structure or there is a slight tilt angle between the pore direction and the incident beam. In fact, the main requirement for estimation of porosity through resonant backscattering is to be able to identify the peak corresponding to the ions travelling in the walls and parallel to the columns, slowing down to Er. It is assumed that the natural resonance width is very small compared to the energy spread of the outgoing ions. In this case, each spectrum is fitted with the Gaussian curve after subtracting a linear background from it. Thus, the resonance peak in the spectra of porous samples was fitted with two Gaussian peaks by the program ORIGIN (Microcal software). The results of fitting are shown in Table 1. Since the peak at higher energies was produced by the incoming ions crossing the walls, we used the information of only this peak in Eq. (4). In fact, the peak at lower energies was not applied (this is related to the ions passing both the walls and the pores). The calculated porosities from Eq. (4) are shown in Table 1. FESEM plan-view images of the two porous samples are presented in Fig. 4. Surface porosity is defined as the area fraction of the pores in a plane cross section of the porous sample which can be estimated from FESEM images using ImageJ, a public domain Java image processing and analysis program [23]. The results are shown in Table 1. The uncertainty in the measured porosity through FESEM imaging is around 5% and the estimated uncertainty through resonant backscattering is about the same value. It should be noted that the uncertainty in the measured porosity by resonant backscattering method is subject to a number of factors [24] including beam energy (0.1%), calibration of detection electronics (0.1%), the Guas-
Fig. 4. FESEM plan-view images of (a) porous alumina 1(b) porous alumina 2.
84
F. Mokhles Gerami et al. / Nuclear Instruments and Methods in Physics Research B 373 (2016) 80–84
sian peak fitting (<3%), stopping power (4%) [25], and geometric error. The latter factor can be described as a combination of pore direction misalignment with respect to both the beam incident and the exit direction. In this work, a one degree variation in the incident beam angle, leads to a 0.01% resonance depth variation and a two degree deviation (due to acceptance angle of the detector) in the scattering angle will result 0.07% deviation in the outgoing path length which is negligible in the total uncertainty. It should be noted that in the preparation of high porosity samples, a further process known as ‘‘pore widening” is applied on the samples, leading to the curved edges of the columnar pores. Then the edges of the pores would not be sharp enough to be easily determined and hence the porosity measured through FESEM imaging will be accompanied with a higher uncertainty. 4. Conclusions The results of porosity estimation of the self-ordered anodic alumina films with cylindrical pores based on resonant backscattering spectrometry are reported. The broadening in the resonance peak of 16O(a,a)16O reaction in the porous alumina samples due to structure induced energy spread as well as an energy shift in the resonance peak were clearly observed. The experimental results indicate that the energy shift in the resonance peak appears when the beam is incident normal to the sample surface. In this geometry where the beam is partially parallel to the columns, the porosity can be estimated using the resonance peak shift observed for the porous sample with no need to have information on the pore size and the interpore distance. The measured porosities for samples of different porosities are consistent with the surface porosities evaluated based on the FESEM images. The slight difference between the measured and evaluated porosities becomes greater when the sample porosity increases and could be due to the error occurred in precise determination of the edge of the columns in the image analysis method. References [1] Y. Piao, H. Kim, Fabrication of nanostructured materials using porous alumina template and their applications for sensing and electrocatalysis, J. Nanosci. Nanotechnol. 9 (2009) 2215–2233. [2] A. Santos, T. Kumeria, D. Losic, Nanoporous anodic aluminum oxide for chemical sensing and biosensors, Trends Anal. Chem. 44 (2013) 25–37. [3] A.B. Abell, K.L. Willis, D.A. Lange, Mercury intrusion porosimetry and image analysis of cement-based materials, J. Colloid Interface Sci. 211 (1999) 39–44. [4] S. Brunauer, P.H. Emmett, E. Teller, The use of low temperature Van der Waals adsorption isotherm in determining surface area, J. Am. Chem. Soc. 60 (1938) 309.
[5] M.R. Baklanov, K.P. Mogilnikov, V.G. Polovinkin, F.N. Dultsev, Determination of pore size distribution in thin films by ellipsometric porosimetry, J. Vac. Sci. Technol., B 18 (2000) 1385–1391. [6] W.L. Wu, W.E. Wallace, E.K. Lin, G.W. Lynn, C.J. Glinka, E.T. Ryan, H.M. Ho, Properties of nanoporous silica thin films determined by high-resolution X-ray reflectivity and small-angle neutron scattering, J. Appl. Phys. 87 (2000) 1193– 1200. [7] B. Bhushan (Ed.), Scanning probe microscopy in nanoscience and nanotechnology 2, Springer Science & Business Media (2010). [8] S. Kumar, J.V. Ramana, C. David, V.S. Raju, Ion beam analysis of porous silicon layers, Nucl. Instrum. Methods Phys. Res. B 179 (2001) 113–120. [9] V. Torres-Costa, F. Pászti, A. Climent-Font, R.J. Martín-Palma, J.M. MartínezDuart, Porosity profile determination of porous silicon interference filters by RBS, Phys. Status Solidi (c) 2 (2005) 3208–3212. [10] V. Torres-Costa, R.J. Martín-Palma, F. Paszti, A. Climent-Font, J.M. MartínezDuart, In-depth RBS study of optical layers based on nanostructured silicon, J. Non-Cryst. Solids 352 (2006) 2521–2525. _ _ [11] H. Krzyzanowska, M. Kulik, Hydrogen and oxygen A.P. Kobzev, J. Zuk, concentration analysis of porous silicon, J. Non-Cryst. Solids 354 (2008) 4367–4374. [12] D.R. Pesiri, R.C. Snow, N. Elliott, C. Maggiore, R.C. Dye, The characterization of asymmetric alumina membranes by Rutherford backscattering spectrometry, J. Membr. Sci. 176 (2000) 209–221. [13] F. Pászti, E. Szilágyi, Pore structure investigations in porous silicon by ion beam analytical methods, Vacuum 50 (1998) 451–462. [14] F. Pászti, E. Szilágyi, A. Manuaba, G. Battistig, Application of resonant backscattering spectrometry for determination of pore structure changes, Nucl. Instrum. Methods Phys. Res. B 161 (2000) 963–968. [15] F. Pászti, E. Szilágyi, Z.E. Horváth, A. Manuaba, G. Battistig, Z. Hajnal, E. Vázsonyi, Morphological investigation of porous samples by resonant backscattering spectrometry, Nucl. Instrum. Methods Phys. Res. B 136 (1998) 533–539. [16] F. Pászti, G. Battistig, Ion beam characterisation and modification of porous silicon, Phys. Status Solidi (a) 182 (2000) 271–278. [17] Z. Hajnal, E. Szilágyi, F. Pászti, G. Battistig, Channeling-like effects due to the macroscopic structure of porous silicon, Nucl. Instrum. Methods Phys. Res. B 118 (1996) 617–621. [18] G. Amsel, E. d’Artemare, G. Battistig, V. Morazzani, C. Ortega, A characterisation of the morphology of porous silicon films by proton energy loss fluctuation measurements with a narrow resonance in the 15N (p, ac) 12C reaction, Nucl. Instrum. Methods Phys. Res. B 122 (1997) 99–112. [19] D. Kuang, R. Charrière, N. Matsapey, M. Flury, J. Faucheu, P. Chavel1, Modeling the specular spectral reflectance of partially ordered alumina nanopores on an aluminum substrate, in: Opt. Express 23 (4) (2015) 4506–4516. [20] J.R. Cameron, Elastic scattering of alpha-particles by oxygen, Phys. Rev. 90 (1953) 839. [21] H. Masuda, K. Fukuda, Ordered metal nanohole arrays made by a two-step replication of honeycomb structures of anodic alumina, Science 268 (1995) 1466–1468. [22] M. Mayer, Computer simulation of ion beam analysis of laterally inhomogeneous materials, Nucl. Instrum. Methods Phys. Res. B 371 (2016) 90–96. [23] T. Ferreira, W. Rasb, ImageJ user guide. (2012) from http://imagej.nih.gov/ ij/download.html, September 7, 2015. [24] C. Jeynes, N.P. Barradas, E. Szilágyi, Accurate determination of quantity of material in thin films by Rutherford backscattering spectrometry, Anal. Chem. 84 (2012) 6061–6069. [25] J.F. Ziegler, M.D. Ziegler, J.P. Biersack, SRIM – the stopping and range of ions in matter, Nucl. Instrum. Methods Phys. Res. B 268 (2010) 1818.