- Email: [email protected]
CdTe surface roughness by Raman spectroscopy using the 830 nm wavelength
Spectrochimica Acta Part A 65 (2006) 51–55
CdTe surface roughness by Raman spectroscopy using the 830 nm wavelength C. Frausto-Reyes a,∗ , J. Rafael Molina-Contreras b , C. Medina-Guti´errez c , Sergio Calixto d ´ Centro de Investigaciones en Optica A.C., Unidad Aguascalientes, Prolong., Constituci´on 607, Fracc. Reserva Loma Bonita, C.P. 20200, Apartado Postal 507, Ags., M´exico b Departamento de Ingenier´ıa El´ ectrica y Electr´onica, Instituto Tecnol´ogico de Aguascalientes, Av. L´opez Mateos 1081 Oriente, Fracc. Ojocaliente, C.P. 20256, Aguascalientes, Ags., M´exico c Universidad de Guadalajara Centro Universitario de Los Lagos, Av. Enrique D´ıaz de Le´ on s/n, Fracc. Paseo de la Monta˜na., C.P. 47460, Lagos de Moreno, Jal., M´exico d Centro de Investigaciones en Optica A.C., Loma del Bosque 115, Colonia Lomas del Campestre, C.P. 37150 Le´ on, Guanajuato, M´exico a
Received 14 July 2005; accepted 28 July 2005
Abstract A Raman spectroscopic study was performed to detect the surface roughness of a cadmium telluride (CdTe) wafer sample, using the 514.5, 632.8 and 830.0 nm excitations wavelengths. To verify the relation between the roughness and the structure of Raman spectra, in certain zones of the sample, we measured their roughness with an atomic force microscopy. It was found that, using the 830 nm wavelength there is a direct correspondence between the spectrum structure and the surface roughness. For the others wavelengths it was found, however, that there is not a clearly correspondence between them. Our results suggest that, using the excitation wavelength of 830 nm the Raman spectroscopy can be used as an on-line roughness monitor on the CdTe growth. © 2005 Elsevier B.V. All rights reserved. Keywords: CdTe; Roughness; Raman spectroscopy; AFM
1. Introduction Due to its potential use for short-wavelength optical storage, and its applicability in light emitting devices, solar cells, and infrared and gamma rays detection, cadmium telluride (CdTe) has been widely studied. Techniques as photoelectron spectroscopy using X-ray [1], photoreflectance [2,3], reflectancedifference [4], reflectivity and picosecond time-resolved photoluminescence [5], transient subpicosecond/picosecond Raman spectroscopy [6], transmission, modulated transmission and Raman spectroscopies [7], resonant spin-flip Raman scattering [8], have been used to investigate either surface/interface scattering effects as the intrinsic optical properties of the negatively charged excitons in modulation doped CdTe quantum wells. It is an actual fact, however, that despite all this studies, there is ∗
Corresponding author. Tel.: +52 449 442 8124/910 5002; fax: +52 449 442 8127. E-mail addresses: [email protected] (C. Frausto-Reyes), [email protected] (J.R. Molina-Contreras), wmedina [email protected] (C. Medina-Guti´errez). 1386-1425/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.saa.2005.07.082
no a complete picture of the optical and electrical characteristics of the cited compound which allow to fabricate efficient and durable optoelectronic devices. Roughness, on the other hand, that can strongly affect the electrical conductivity of a thin film [9], that could be one of the responsible mechanisms for the enhancement of the Raman signal [10], or even that can be adjusted by appropriate deposition parameters in some cases [11], now a day is still an unknown subject. There is no an exact answer to the question of what happens on an irregular interface with periodic or random variation of height onto an electromagnetic wave is incident [12]. Due to this, much work has to be done in this sense, in order for the problem of scattering from a rough surface be completed. Raman spectroscopy, a sensitive, non-destructive and noncontact technique that has been used to infer the presence of structural defects and structural inhomogeneities in (CdTe)x (In2 Te3 )1−x thin films [7] and to obtain information about the composition at the interface of a sample of CdTe after growth [13], is also a technique that has been used on-line to obtain information on layer thickness, crystalline quality and interdifusion effects in the interface region of CdTe grown on InSb
52
C. Frausto-Reyes et al. / Spectrochimica Acta Part A 65 (2006) 51–55
(1 1 0) by molecular beam epitaxy [14]. Due to this, it is a technique that can be used successfully to investigate the roughness phenomena. In this work, we report on a systematic study of the surface roughness of CdTe using Raman spectroscopy. In our study, we have investigated the morphology of our sample by means of an atomic force microscopy (AFM), and the Raman spectra behavior, using two Raman systems with the 514.5, 632.8 and 830.0 nm excitations wavelengths. Our results clearly suggest that using an excitation wavelength of 830 nm, the Raman spectroscopy can be used as an on-line roughness monitor on the CdTe growth. 2. Experimental CdTe is a cubic semiconducting compound with a zincblende-type symmetry and an energy gap of 1.51 eV at room temperature. The sample used in our experiment was a (0 0 1) oriented commercial wafer, on which we induced the reported roughness by softly carving the sample on a diamond paste. The Raman spectra, was measured using two micro-Raman system with a back scattering geometry [15]. One of them (Ranishaw system 1000B) has a 600 lines/mm grating, a CCD camera (Rem Cam 1024 × 256 pixels), and focus its 830 nm wavelength laser beam with a spot-size of about 2 m onto the sample, with a 50× objective of a Leica (DMLM) microscope. The other one, which uses the interchangeable 632.8 He–Ne (Melles Griot) and 514.5 nm argon (Spectra Physics) lasers, focus its laser beam with a spot-size of about 2.5 m onto the sample surface with a 40× objective of a Zeiss (Axioskop 2) microscope. In this case, the Raman spectra are recorded using a monochromator (Jobin Yvon HR 460) equipped with an air-cooled CCD (256 × 1024 pixels) camera and a 1200 lines/mm grating. To reject the Rayleigh emission light and the plasma frequencies of both lasers, the system uses holographic super Notch-plus filters (Kaiser Optical systems) and interference filters (Melles Griot), respectively. The calibration of the instruments was done using the 520 cm−1 Raman line of a silicon wafer and for the data acquisition it was used the Grams software. The experimental set up used to study the surface morphology and to obtain the mean roughness of the reported on zones was an atomic force microscope (Digital Instruments, Dimension 3100).
Fig. 1. AFM image of CdTe with a measured mean roughness of 6.0 nm.
Fig. 1, basically corresponds to a smooth plane zone, and that the observed periodic lines, are due to the AFM sweeping. Fig. 3 shows the AFM image of another area of the smooth zone of the sample. The inlets labeled “a”, “b” and “c” in this image, show three clearly defined zones because their roughness. The “b” band between the “a” and “c” zones, is practically a groove of 1.6 m deep and has a measured mean roughness of 279.8 nm, while the “a” and “c” zones have a mean roughness of 13.8 and 10.1 nm, respectively. In Fig. 4a and b, we show the Raman spectra of the zones labeled “a” and “c” of Fig. 3. The spectra were also measured using the 830 nm wavelength with an exposure time of 1.0 s and 0.6 mW of power at the sample. In these spectra clearly appear a Raman peak around 331 cm−1 and it is more intense for the 13.8 nm mean roughness than for the 10.1 nm taking as a reference the 165 cm−1 Raman peak intensity. In order to confirm the observed spectra behavior, we also measured the back of the sample. Fig. 5 shows the AFM image of this back zone, which is more roughness than its opposite face. This is confirmed with the great value of its measured mean roughness (in the inlet), which in this case is of 430.0 nm. Such value compared with the mean roughness of the groove of “b” zone of Fig. 3, let us to compare under the same conditions,
3. Results and discussion Fig. 1 shows the AFM image of a smooth zone of the studied sample surface. Such a zone with a measured mean roughness of 6.0 nm has, as it can be seen in Fig. 2, a wide Raman spectrum with a right line wing that do not show any observed change due to the measured roughness. However, the Raman spectrum, obtained using the 830 nm wavelength with an exposure time of 0.5 s and a 0.1 mW of power at the sample, clearly shows a Raman peak at 165 cm−1 , which according to the Refs. [13,16], corresponds to the longitudinal mode (LO) of the CdTe. For the sake of clarity, it is worth to note, that the AFM image of
Fig. 2. Raman spectra of CdTe with a mean roughness of 6.0 nm, using the excitation wavelength of 830 nm.
C. Frausto-Reyes et al. / Spectrochimica Acta Part A 65 (2006) 51–55
53
Fig. 5. AFM image of CdTe with a mean roughness in the enclosed zone of 430.0 nm. Fig. 3. AFM image of CdTe with a mean roughness in the enclosed zones of (a) 13.8 nm, (b) 279.8 nm and (c) 10.1 nm.
the Raman behavior of both “b” zone and back area. Fig. 6a and b show the Raman spectra of both “b” and back areas of the sample, respectively. In this case, the Raman spectra were measured using the 830 nm wavelength with an exposure time of 2.0 s and 6.0 mW of power at the sample. The Raman peaks around 331 cm−1 of both spectra besides to have a better definition, they are also more intense and keep the same relation than those of Fig. 4a and b. That is, the 331 cm−1 Raman peak of the back roughness which has a measured mean roughness of 430.0 nm is more intensive than the one of the groove, which has a measured mean roughness of 279.8 nm. We used again, as reference, the Raman peak intensity at 165 cm−1 . From the inlets of Fig. 6a and b, it is worth to remark that for this mean roughness, it begins to appear another Raman peak at 122 cm−1 , which is more intense for the measured mean roughness of 430.0 nm than for 279.8 nm, although this peak was attenuated by the cut-off
frequency of the notch filter. According to the reference [13], this peak could be related to the TeO2 Raman peak, but it is remarkable that its amplitude is a function of the mean roughness as it is evident from our experimental results. Results of the AFM image of Fig. 5 and the Raman spectra of Fig. 6a and b clearly confirm the suggested evidence that the observed intensity changes of 331 cm−1 Raman peak in Fig. 4a and b is due to the roughness of the sample, and that such phenomenon can be observed using the 830 nm wavelength. To observe the structural changes in the Raman spectra we analyzed the ratio of the intensities between the Raman peaks of 165 and 331 cm−1 (see Table 1), this ratio decrease when the mean roughness increase. To compare the results, we also used the 632.8 and 514.5 nm excitation wavelengths to examine some of the above-mentioned zones of the sample. The Raman spectra shown in Fig. 7a–c correspond to such zones with mean roughness of 6.0, 279.8 and 430.0 nm, respectively. These spectra were measured using the 632.8 nm wavelength with an exposure time of 60 s and
Fig. 4. Raman spectra of CdTe with a mean roughness of (a) 10.1 nm and (b) 13.8 nm using the excitation wavelength of 830 nm.
Fig. 6. Raman spectra of CdTe with a mean roughness of (a) 279.8 nm and (b) 430.0 nm using the excitation wavelength of 830 nm.
54
C. Frausto-Reyes et al. / Spectrochimica Acta Part A 65 (2006) 51–55
Table 1 Intensity ratios between the Raman peaks of 165 and 331 cm−1 of the CdTe spectrum for the different mean surface roughness, using the 830 nm wavelength Mean roughness (nm)
Intensity ratio
10.1 13.8 279.8 430
4.3 3.5 2.8 2.1
10.0 mW of power at the sample. In this case, as it can be seen from Fig. 7, in all the Raman spectra, appear the peaks 122, 140 and 165 cm−1 clearly defined. The Raman peak at 140 cm−1 corresponds to the transversal mode (TO) of the CdTe [13,16]. Fig. 8a–c, correspond to the Raman spectra of the zones with mean roughness of 6.0, 279.8 and 430.0 nm, respectively. These spectra were measured using the 514.5 nm wavelength with an exposure time of 5.0 s and 16.0 mW of power at the sample. It can be seen from Fig. 8, that the 331 cm−1 Raman peak disappears in all the spectra, and a Raman peak around 102 cm−1 , appears. The results obtained in this study clearly show that there are not significative changes among the spectra structure of Fig. 7a–c, although they were obtained from zones with different roughness. The same behavior is observed in the spectra of Fig. 8a–c. For this reason the excitation wavelengths of 632.8 and 514.5 nm cannot be used to identify roughness heights by Raman spectroscopy.
Fig. 8. Raman spectra of CdTe with a mean roughness of (a) 6.057 nm, (b) 279.74 nm and (c) 429.96 nm using the excitation wavelength of 514.5 nm.
4. Conclusions We have studied experimentally the utility of Raman spectroscopy as a non-contact technique to detect the surface roughness heights on a (0 0 1) CdTe wafer sample. We have obtained Raman spectra, using the 632.8, 514.5 and 830.0 nm excitation wavelengths, of zones with different mean roughness previously measured with an atomic force microscopy. Results clearly suggest that the ratio between the intensities of the 165 and 331 cm−1 Raman peaks, using the 830.0 nm wavelength, is related to the surface roughness of the sample. Raman spectra structure using the 632.8 or 514.5 nm wavelengths, do not change even when surface roughness is changed. Raman spectroscopy with 632.8 or 514.5 nm wavelengths cannot be used to detect changes in the surface roughness, but it can be used as an on-line surface roughness monitor on the CdTe growth with an excitation wavelength of 830.0 nm. Acknowledgments The authors express their gratitude to Martin Ortiz, Marian Poterasu, Gil Arturo P´erez and Juan M. Sarabia for all their invaluable technical help provided during the development of this study. References
Fig. 7. Raman spectra of CdTe with a mean roughness of (a) 6.057 nm, (b) 279.74 nm and (c) 429.96 nm using the excitation wavelength of 632.8 nm.
[1] J. Fritsche, T. Schulmeyer, A. Thißen, A. Klein, W. Jaegermann, Thin Solid Films 431–432 (2003) 267. [2] A. Kaneta, S. Adachi, J. Appl. Phys. 87 (2000) 7360.
C. Frausto-Reyes et al. / Spectrochimica Acta Part A 65 (2006) 51–55 [3] A. Lastras-Mart´ınez, R.E. Balderas-Navarro, P. Cant´u-Alejandro, L.F. Lastras-Mart´ınez, J. Appl. Phys. 86 (1999) 2062. [4] J.R. Molina C, R. Espinosa-Luna, J. Phys. D: Appl. Phys. 38 (2005) 12. [5] V. Ciulin, P. Kossacki, S. Haacke, J.D. Gani´ere, B. Deveaud, A. Esser, M. Kutrowski, T. Wojtowicz, Phys. Rev. B 62 (2000) 16310. [6] E. Grann, Y. Chen, K.T. Tsen, D.K. Ferry, T. Almeida, Y.P. Chen, J.P. Faurie, S. Sivananthan, J. Appl. Phys. 80 (1996) 3840. [7] M.A. Santana-Aranda, J. Luyo-Alvarado, M. Mel´endez-Lira, M. ZapataTorres, A. Zapata-Navarro, S. Jim´enez-Sandoval, R. Castro-Rodriguez, J.L. Pe˜na, Superficies y Vacio 8 (1999) 69. [8] A.A. Sirenko, T. Ruf, M. Cardona, D.R. Yakovlev, W. Ossau, A. Waag, G. Landwehr, Phys. Rev. B 56 (1997) 2114. [9] Palasantzas, J.Th.M. Hosson, J. Appl. Phys. 93 (2003) 320.
55
[10] J.A. S´anchez-Gil, J.V. Garc´ıa-Ramos, E.R. M´endez, Phys. Rev. B 62 (2000) 10515. [11] M. Rusu, I.I. Nicolaescu, G.G. Rusu, Appl. Phys. 70 (2000) 565. [12] P. Beckmann, A. Spizzichino, Scattering of Electromagnetic Waves from Rough Surfaces, Pergamon Press, Oxford, 1963. ´ [13] J. Huerta-Ruelas, M. L´opez-L´opez, O. Zelaya-Angel, Jpn. J. Appl. Phys. 39 (2000) 1701. [14] D. Drews, J. Sham, W. Richter, J. Appl. Phys. 78 (1995) 4060. [15] G. Turrel, M. Delhaye, P. Dhamelincourt, in: G. Turrel, J. Corset (Eds.), Raman Microscopy, Developments and Applications, Academic Press, London, 1996, pp. 27–49. [16] J. Huerta, M. L´opez, O. Zelaya, Superficies y Vacio 8 (1999) 125.
CdTe surface roughness by Raman spectroscopy using the 830 nm wavelength C. Frausto-Reyes a,∗ , J. Rafael Molina-Contreras b , C. Medina-Guti´errez c , Sergio Calixto d ´ Centro de Investigaciones en Optica A.C., Unidad Aguascalientes, Prolong., Constituci´on 607, Fracc. Reserva Loma Bonita, C.P. 20200, Apartado Postal 507, Ags., M´exico b Departamento de Ingenier´ıa El´ ectrica y Electr´onica, Instituto Tecnol´ogico de Aguascalientes, Av. L´opez Mateos 1081 Oriente, Fracc. Ojocaliente, C.P. 20256, Aguascalientes, Ags., M´exico c Universidad de Guadalajara Centro Universitario de Los Lagos, Av. Enrique D´ıaz de Le´ on s/n, Fracc. Paseo de la Monta˜na., C.P. 47460, Lagos de Moreno, Jal., M´exico d Centro de Investigaciones en Optica A.C., Loma del Bosque 115, Colonia Lomas del Campestre, C.P. 37150 Le´ on, Guanajuato, M´exico a
Received 14 July 2005; accepted 28 July 2005
Abstract A Raman spectroscopic study was performed to detect the surface roughness of a cadmium telluride (CdTe) wafer sample, using the 514.5, 632.8 and 830.0 nm excitations wavelengths. To verify the relation between the roughness and the structure of Raman spectra, in certain zones of the sample, we measured their roughness with an atomic force microscopy. It was found that, using the 830 nm wavelength there is a direct correspondence between the spectrum structure and the surface roughness. For the others wavelengths it was found, however, that there is not a clearly correspondence between them. Our results suggest that, using the excitation wavelength of 830 nm the Raman spectroscopy can be used as an on-line roughness monitor on the CdTe growth. © 2005 Elsevier B.V. All rights reserved. Keywords: CdTe; Roughness; Raman spectroscopy; AFM
1. Introduction Due to its potential use for short-wavelength optical storage, and its applicability in light emitting devices, solar cells, and infrared and gamma rays detection, cadmium telluride (CdTe) has been widely studied. Techniques as photoelectron spectroscopy using X-ray [1], photoreflectance [2,3], reflectancedifference [4], reflectivity and picosecond time-resolved photoluminescence [5], transient subpicosecond/picosecond Raman spectroscopy [6], transmission, modulated transmission and Raman spectroscopies [7], resonant spin-flip Raman scattering [8], have been used to investigate either surface/interface scattering effects as the intrinsic optical properties of the negatively charged excitons in modulation doped CdTe quantum wells. It is an actual fact, however, that despite all this studies, there is ∗
Corresponding author. Tel.: +52 449 442 8124/910 5002; fax: +52 449 442 8127. E-mail addresses: [email protected] (C. Frausto-Reyes), [email protected] (J.R. Molina-Contreras), wmedina [email protected] (C. Medina-Guti´errez). 1386-1425/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.saa.2005.07.082
no a complete picture of the optical and electrical characteristics of the cited compound which allow to fabricate efficient and durable optoelectronic devices. Roughness, on the other hand, that can strongly affect the electrical conductivity of a thin film [9], that could be one of the responsible mechanisms for the enhancement of the Raman signal [10], or even that can be adjusted by appropriate deposition parameters in some cases [11], now a day is still an unknown subject. There is no an exact answer to the question of what happens on an irregular interface with periodic or random variation of height onto an electromagnetic wave is incident [12]. Due to this, much work has to be done in this sense, in order for the problem of scattering from a rough surface be completed. Raman spectroscopy, a sensitive, non-destructive and noncontact technique that has been used to infer the presence of structural defects and structural inhomogeneities in (CdTe)x (In2 Te3 )1−x thin films [7] and to obtain information about the composition at the interface of a sample of CdTe after growth [13], is also a technique that has been used on-line to obtain information on layer thickness, crystalline quality and interdifusion effects in the interface region of CdTe grown on InSb
52
C. Frausto-Reyes et al. / Spectrochimica Acta Part A 65 (2006) 51–55
(1 1 0) by molecular beam epitaxy [14]. Due to this, it is a technique that can be used successfully to investigate the roughness phenomena. In this work, we report on a systematic study of the surface roughness of CdTe using Raman spectroscopy. In our study, we have investigated the morphology of our sample by means of an atomic force microscopy (AFM), and the Raman spectra behavior, using two Raman systems with the 514.5, 632.8 and 830.0 nm excitations wavelengths. Our results clearly suggest that using an excitation wavelength of 830 nm, the Raman spectroscopy can be used as an on-line roughness monitor on the CdTe growth. 2. Experimental CdTe is a cubic semiconducting compound with a zincblende-type symmetry and an energy gap of 1.51 eV at room temperature. The sample used in our experiment was a (0 0 1) oriented commercial wafer, on which we induced the reported roughness by softly carving the sample on a diamond paste. The Raman spectra, was measured using two micro-Raman system with a back scattering geometry [15]. One of them (Ranishaw system 1000B) has a 600 lines/mm grating, a CCD camera (Rem Cam 1024 × 256 pixels), and focus its 830 nm wavelength laser beam with a spot-size of about 2 m onto the sample, with a 50× objective of a Leica (DMLM) microscope. The other one, which uses the interchangeable 632.8 He–Ne (Melles Griot) and 514.5 nm argon (Spectra Physics) lasers, focus its laser beam with a spot-size of about 2.5 m onto the sample surface with a 40× objective of a Zeiss (Axioskop 2) microscope. In this case, the Raman spectra are recorded using a monochromator (Jobin Yvon HR 460) equipped with an air-cooled CCD (256 × 1024 pixels) camera and a 1200 lines/mm grating. To reject the Rayleigh emission light and the plasma frequencies of both lasers, the system uses holographic super Notch-plus filters (Kaiser Optical systems) and interference filters (Melles Griot), respectively. The calibration of the instruments was done using the 520 cm−1 Raman line of a silicon wafer and for the data acquisition it was used the Grams software. The experimental set up used to study the surface morphology and to obtain the mean roughness of the reported on zones was an atomic force microscope (Digital Instruments, Dimension 3100).
Fig. 1. AFM image of CdTe with a measured mean roughness of 6.0 nm.
Fig. 1, basically corresponds to a smooth plane zone, and that the observed periodic lines, are due to the AFM sweeping. Fig. 3 shows the AFM image of another area of the smooth zone of the sample. The inlets labeled “a”, “b” and “c” in this image, show three clearly defined zones because their roughness. The “b” band between the “a” and “c” zones, is practically a groove of 1.6 m deep and has a measured mean roughness of 279.8 nm, while the “a” and “c” zones have a mean roughness of 13.8 and 10.1 nm, respectively. In Fig. 4a and b, we show the Raman spectra of the zones labeled “a” and “c” of Fig. 3. The spectra were also measured using the 830 nm wavelength with an exposure time of 1.0 s and 0.6 mW of power at the sample. In these spectra clearly appear a Raman peak around 331 cm−1 and it is more intense for the 13.8 nm mean roughness than for the 10.1 nm taking as a reference the 165 cm−1 Raman peak intensity. In order to confirm the observed spectra behavior, we also measured the back of the sample. Fig. 5 shows the AFM image of this back zone, which is more roughness than its opposite face. This is confirmed with the great value of its measured mean roughness (in the inlet), which in this case is of 430.0 nm. Such value compared with the mean roughness of the groove of “b” zone of Fig. 3, let us to compare under the same conditions,
3. Results and discussion Fig. 1 shows the AFM image of a smooth zone of the studied sample surface. Such a zone with a measured mean roughness of 6.0 nm has, as it can be seen in Fig. 2, a wide Raman spectrum with a right line wing that do not show any observed change due to the measured roughness. However, the Raman spectrum, obtained using the 830 nm wavelength with an exposure time of 0.5 s and a 0.1 mW of power at the sample, clearly shows a Raman peak at 165 cm−1 , which according to the Refs. [13,16], corresponds to the longitudinal mode (LO) of the CdTe. For the sake of clarity, it is worth to note, that the AFM image of
Fig. 2. Raman spectra of CdTe with a mean roughness of 6.0 nm, using the excitation wavelength of 830 nm.
C. Frausto-Reyes et al. / Spectrochimica Acta Part A 65 (2006) 51–55
53
Fig. 5. AFM image of CdTe with a mean roughness in the enclosed zone of 430.0 nm. Fig. 3. AFM image of CdTe with a mean roughness in the enclosed zones of (a) 13.8 nm, (b) 279.8 nm and (c) 10.1 nm.
the Raman behavior of both “b” zone and back area. Fig. 6a and b show the Raman spectra of both “b” and back areas of the sample, respectively. In this case, the Raman spectra were measured using the 830 nm wavelength with an exposure time of 2.0 s and 6.0 mW of power at the sample. The Raman peaks around 331 cm−1 of both spectra besides to have a better definition, they are also more intense and keep the same relation than those of Fig. 4a and b. That is, the 331 cm−1 Raman peak of the back roughness which has a measured mean roughness of 430.0 nm is more intensive than the one of the groove, which has a measured mean roughness of 279.8 nm. We used again, as reference, the Raman peak intensity at 165 cm−1 . From the inlets of Fig. 6a and b, it is worth to remark that for this mean roughness, it begins to appear another Raman peak at 122 cm−1 , which is more intense for the measured mean roughness of 430.0 nm than for 279.8 nm, although this peak was attenuated by the cut-off
frequency of the notch filter. According to the reference [13], this peak could be related to the TeO2 Raman peak, but it is remarkable that its amplitude is a function of the mean roughness as it is evident from our experimental results. Results of the AFM image of Fig. 5 and the Raman spectra of Fig. 6a and b clearly confirm the suggested evidence that the observed intensity changes of 331 cm−1 Raman peak in Fig. 4a and b is due to the roughness of the sample, and that such phenomenon can be observed using the 830 nm wavelength. To observe the structural changes in the Raman spectra we analyzed the ratio of the intensities between the Raman peaks of 165 and 331 cm−1 (see Table 1), this ratio decrease when the mean roughness increase. To compare the results, we also used the 632.8 and 514.5 nm excitation wavelengths to examine some of the above-mentioned zones of the sample. The Raman spectra shown in Fig. 7a–c correspond to such zones with mean roughness of 6.0, 279.8 and 430.0 nm, respectively. These spectra were measured using the 632.8 nm wavelength with an exposure time of 60 s and
Fig. 4. Raman spectra of CdTe with a mean roughness of (a) 10.1 nm and (b) 13.8 nm using the excitation wavelength of 830 nm.
Fig. 6. Raman spectra of CdTe with a mean roughness of (a) 279.8 nm and (b) 430.0 nm using the excitation wavelength of 830 nm.
54
C. Frausto-Reyes et al. / Spectrochimica Acta Part A 65 (2006) 51–55
Table 1 Intensity ratios between the Raman peaks of 165 and 331 cm−1 of the CdTe spectrum for the different mean surface roughness, using the 830 nm wavelength Mean roughness (nm)
Intensity ratio
10.1 13.8 279.8 430
4.3 3.5 2.8 2.1
10.0 mW of power at the sample. In this case, as it can be seen from Fig. 7, in all the Raman spectra, appear the peaks 122, 140 and 165 cm−1 clearly defined. The Raman peak at 140 cm−1 corresponds to the transversal mode (TO) of the CdTe [13,16]. Fig. 8a–c, correspond to the Raman spectra of the zones with mean roughness of 6.0, 279.8 and 430.0 nm, respectively. These spectra were measured using the 514.5 nm wavelength with an exposure time of 5.0 s and 16.0 mW of power at the sample. It can be seen from Fig. 8, that the 331 cm−1 Raman peak disappears in all the spectra, and a Raman peak around 102 cm−1 , appears. The results obtained in this study clearly show that there are not significative changes among the spectra structure of Fig. 7a–c, although they were obtained from zones with different roughness. The same behavior is observed in the spectra of Fig. 8a–c. For this reason the excitation wavelengths of 632.8 and 514.5 nm cannot be used to identify roughness heights by Raman spectroscopy.
Fig. 8. Raman spectra of CdTe with a mean roughness of (a) 6.057 nm, (b) 279.74 nm and (c) 429.96 nm using the excitation wavelength of 514.5 nm.
4. Conclusions We have studied experimentally the utility of Raman spectroscopy as a non-contact technique to detect the surface roughness heights on a (0 0 1) CdTe wafer sample. We have obtained Raman spectra, using the 632.8, 514.5 and 830.0 nm excitation wavelengths, of zones with different mean roughness previously measured with an atomic force microscopy. Results clearly suggest that the ratio between the intensities of the 165 and 331 cm−1 Raman peaks, using the 830.0 nm wavelength, is related to the surface roughness of the sample. Raman spectra structure using the 632.8 or 514.5 nm wavelengths, do not change even when surface roughness is changed. Raman spectroscopy with 632.8 or 514.5 nm wavelengths cannot be used to detect changes in the surface roughness, but it can be used as an on-line surface roughness monitor on the CdTe growth with an excitation wavelength of 830.0 nm. Acknowledgments The authors express their gratitude to Martin Ortiz, Marian Poterasu, Gil Arturo P´erez and Juan M. Sarabia for all their invaluable technical help provided during the development of this study. References
Fig. 7. Raman spectra of CdTe with a mean roughness of (a) 6.057 nm, (b) 279.74 nm and (c) 429.96 nm using the excitation wavelength of 632.8 nm.
[1] J. Fritsche, T. Schulmeyer, A. Thißen, A. Klein, W. Jaegermann, Thin Solid Films 431–432 (2003) 267. [2] A. Kaneta, S. Adachi, J. Appl. Phys. 87 (2000) 7360.
C. Frausto-Reyes et al. / Spectrochimica Acta Part A 65 (2006) 51–55 [3] A. Lastras-Mart´ınez, R.E. Balderas-Navarro, P. Cant´u-Alejandro, L.F. Lastras-Mart´ınez, J. Appl. Phys. 86 (1999) 2062. [4] J.R. Molina C, R. Espinosa-Luna, J. Phys. D: Appl. Phys. 38 (2005) 12. [5] V. Ciulin, P. Kossacki, S. Haacke, J.D. Gani´ere, B. Deveaud, A. Esser, M. Kutrowski, T. Wojtowicz, Phys. Rev. B 62 (2000) 16310. [6] E. Grann, Y. Chen, K.T. Tsen, D.K. Ferry, T. Almeida, Y.P. Chen, J.P. Faurie, S. Sivananthan, J. Appl. Phys. 80 (1996) 3840. [7] M.A. Santana-Aranda, J. Luyo-Alvarado, M. Mel´endez-Lira, M. ZapataTorres, A. Zapata-Navarro, S. Jim´enez-Sandoval, R. Castro-Rodriguez, J.L. Pe˜na, Superficies y Vacio 8 (1999) 69. [8] A.A. Sirenko, T. Ruf, M. Cardona, D.R. Yakovlev, W. Ossau, A. Waag, G. Landwehr, Phys. Rev. B 56 (1997) 2114. [9] Palasantzas, J.Th.M. Hosson, J. Appl. Phys. 93 (2003) 320.
55
[10] J.A. S´anchez-Gil, J.V. Garc´ıa-Ramos, E.R. M´endez, Phys. Rev. B 62 (2000) 10515. [11] M. Rusu, I.I. Nicolaescu, G.G. Rusu, Appl. Phys. 70 (2000) 565. [12] P. Beckmann, A. Spizzichino, Scattering of Electromagnetic Waves from Rough Surfaces, Pergamon Press, Oxford, 1963. ´ [13] J. Huerta-Ruelas, M. L´opez-L´opez, O. Zelaya-Angel, Jpn. J. Appl. Phys. 39 (2000) 1701. [14] D. Drews, J. Sham, W. Richter, J. Appl. Phys. 78 (1995) 4060. [15] G. Turrel, M. Delhaye, P. Dhamelincourt, in: G. Turrel, J. Corset (Eds.), Raman Microscopy, Developments and Applications, Academic Press, London, 1996, pp. 27–49. [16] J. Huerta, M. L´opez, O. Zelaya, Superficies y Vacio 8 (1999) 125.