REVIEW OF SCIENTIFIC INSTRUMENTS 86, 016102 (2015)
Note: A stand on the basis of atomic force microscope to study substrates for imaging optics N. I. Chkhalo, N. N. Salashchenko, and M. V. Zorina Department of Multilayer Optics, Institute for Physics of Microstructures of the Russian Academy of Sciences, GSP-105, 603950 Nizhny Novgorod, Russia
(Received 25 November 2014; accepted 19 December 2014; published online 8 January 2015) A description of a stand based on atomic force microscopy (AFM) for roughness measurements of large optical components with arbitrary surfaces is given. The sample under study is mounted on a uniaxial goniometer which allows the sample to be tilted in the range of ±30◦. The inclination enables the local normal along the axis of the probe to be established at any point of the surface under study. A comparison of the results of the measurement of noise and roughness of a flat quartz sample, in the range of spatial frequencies 0.025–70 µm−1, obtained from “standard” AFM and developed versions is given. Within the experimental error, the measurement results were equivalent. Examples of applications of the stand for the study of substrates for X-ray optics are presented. C 2015 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4905336]
Recently, there has been an urgent need to revise the old and develop new methods for measuring the roughness of super-smooth optical surfaces used in projection nanolithography, X-ray microscopy, and astronomy. A distinctive feature of the elements used in these systems is their curvature. The radius of the curvature may vary from a few millimetres to a few metres, and the deflection is of up to 20 or more millimetres. To ensure diffraction quality imaging (spatial resolution is determined only by the radiation wavelength and numerical aperture of the optics) by such systems, substrates with surface irregularities (roughness) in the root-mean-square (r.m.s.) parameter, integrated over the range of spatial frequencies from 10−3–102 µm−1 (lateral dimensions 1 nm–1 mm) to 0.1–0.2 nm, are necessary, together with adequate roughness metrology. Atomic force microscopy (AFM) is the main method used to study roughness in the high and medium spatial frequency range of 10−1–102 µm−1.1 In some works, such as Refs. 2–5, using as a reference, the “ab initio” method of diffuse scattering of the hard and soft X-rays, it is shown that AFM is quite an adequate method of measuring lateral dimension surface roughness in the range of nanometres to 40–60 µm. However, there is a problem that limits the applicability of AFM in studying super-smooth optical surfaces attended for the high-resolution imaging. The thing is that, for the subatomic resolution of the roughness heights to be preserved, the range of vertical probe displacement should not exceed ∆z = 1 µm. In the case when the surface is inclined to the probe axis, the range of possible lateral scanning decreases yielding a limitation on the range of recorded spatial frequencies of roughness. For example, of a 3◦ inclination, the frame cannot exceed 19 µm.6 The local slopes of the surface to the axis in real optics can be one order of magnitude larger. Therefore, for measurements of sized and curved surfaces, researchers are turning to nonstandard measuring schemes, e.g., the mounting of an AFM head onto the sample, which leads to the significant influence of vibrations on the measurement results and causes the risk of damaging unique and expensive substrates. 0034-6748/2015/86(1)/016102/3/$30.00
This paper describes a stand based on AFM, which allows the problem of the measurement of large curved surfaces to be solved with precision, equal to that of the measurement of flat surfaces by conventional AFM. To solve these problems, we have developed a stand based on the AFM of the NTEGRA Prima Company NT-MDT, Zelenograd,7 and a uniaxial goniometer which allows the sample to be tilted in the range of ±30◦. The inclination enables the local normal along the axis of the probe to be established at any point of the surface under study. A diagram and a photo of the stand are shown in Figs. 1 and 2. The main elements of the stand are a honeycomb table top with a pneumatic vibration isolation system (1), a trapeze (2) which, by means of springs (3), by analogy with the factory configuration of the device, is a suspended measuring system (4). The measurement system consists of a base plate (5), a frame (6), and a cross-beam (7). The AFM head (8) is suspended on the cross-beam. Test sample (9) is mounted on a uniaxial goniometer (10). For the “rough” sample inlet to AFM probe (11), a lifting mechanism (12) mounted on plate (5) with screws (13) is used. The fine adjustment of the sample surface to the probe is performed by means of three micro-screws (14) and springs (15). The lifting mechanism provides vertical movement of the sample to 100 mm and the micro-screws to 25 mm. Thus, the stand allows one to study samples in a wide range of thicknesses, i.e., it is fairly universal. Studying the main characteristics of the stand was undertaken by using quartz samples differing in both surface form and dimensions and roughness values. The main parameters of the samples under study are given in Table I. Effective roughness σeff values are defined as 2 σeff
ν max = PSD(ν)dν,
(1)
ν min
where νmin and νmax are the minimal and maximal spatial frequencies, where the PSD (power spectral density) functions of the roughness are determined.
86, 016102-1
© 2015 AIP Publishing LLC
016102-2
Chkhalo, Salashchenko, and Zorina
Rev. Sci. Instrum. 86, 016102 (2015)
FIG. 2. Photograph of the stand with tested curved sample.
FIG. 1. General scheme of the stand: (a) 1—honeycomb table top with pneumatic vibration isolation system, 2—trapeze, 3—spring, 4—measuring system; (b) 5—base plate, 6—frame, 7—cross-beam, 8—AFM head, 9—test sample, 10—uniaxial goniometer, 11—AFM probe, 12—lifting mechanism, 13—screw, 14—micro-screws, and 15—spring.
Initially, the intrinsic AFM noise on the stand was compared with the noise of the standard (factory) configuration. As the experiment shows, AFM noise in both configurations coincides within the measurement error in thousandths of a nanometre and is about 0.13 nm in the range of spatial frequencies 0.025–60 µm−1. Thus, it was found that the AFM noise in the stand did not increase compared with a standard device. Moreover, there is a tendency for it to reduce. The next step was to compare the measured values of the roughness of the flat sample #1, which could be measured both on the stand and in the standard AFM configuration. Figure 3 shows the measured PSD functions. The solid line shows the data obtained in the standard configuration of the instrument, and the lines with symbols the configuration on the stand. As in the case of the noise, the results of roughness measurements in both configurations of the instrument coincided with high accuracy. Since the noise, characterized by σnois, and the surface roughness, σsurf , are statistically independent random variables, the effective (measured experimentally) roughness σeff can be written as the sum of the squares of the surface roughness and noise 2 2 2 σeff = σsurf + σnois .
(2)
TABLE I. Description of the investigated samples of fused silica: D— diameter; h—thickness; R—the radius of curvature; σ eff —effective roughness measured in the range of spatial frequencies 0.025–60 µm−1; and σ surf —the true roughness value obtained by subtracting the AFM noise.
#
D (mm)
h (mm)
R (mm)
σ eff (nm)
σ surf (nm)
1 2 3
25 250 100
6 31 30
∞ 3981.0 137.0
0.21 0.29 0.71
0.16 0.26 0.70
FIG. 3. PSD functions of the roughness of sample #1 measured in the standard configuration of the instrument (solid lines) and on the stand (lines with symbols).
016102-3
Chkhalo, Salashchenko, and Zorina
Rev. Sci. Instrum. 86, 016102 (2015)
center to the periphery, it is possible to establish the local normal to the surface along the axis of the probe. Thus, first, it is possible to bring the sample to the probe at any point, and second, the minimum inclination of the surface under study with respect to the horizon is reached, thus, one can make measurements at maximal frame, as in the study of flat surfaces. The maximum numerical aperture of the substrates can be NA = 0.3 and a maximum diameter up to 300 mm. The experiments show that in terms of the intrinsic noise, the operating spectral range of roughness and sensitivity of this device is not inferior to the standard AFM, at the same time, allowing the study of sized and curved samples. It is considered appropriate to equip the standard instruments with such an option, as it is sought by optical industry enterprises and groups producing X-ray optics. FIG. 4. PSD functions of the roughness of sample #2 measured on the stand.
This relation shows that when measuring the substrate with an effective roughness approaching 0.2–0.3 nm, the intrinsic AFM noise must be taken into account. In particular, for sample #1, the measured effective roughness was 0.21 nm. In view of the AFM noise, true surface roughness is 0.16 nm; see the last column of Table I. Figure 4 illustrates the application of the developed stand for the study of curved and sized substrate #2 that is designed for X-ray astronomy of the Sun. Sample #2 passed all stages of polishing, except for ion-beam etching. As the conclusion, this paper describes a stand designed on the basis of standard AFM. Unlike standard AFM, the stand, thanks to the integrated uniaxial goniometer, allows the optical elements with arbitrary surface shapes and large dimensions to be studied. This is achieved by the fact that due to the uniaxial goniometer at any point of the surface under study, from the
This work was supported by the Russian Fund for Basic Research, by the Ministry of Science and Education of Russia, and by the programme “Physics and technology of micro- and nanostructures” of the Public Centre of Use at the Institute for Physics of Microstructures of the Russian Academy of Sciences. 1J.
E. Griffith and D. A. Grigg, J. Appl. Phys. 74, R83–R109 (1993). E. Asadchikov, I. V. Kozhevnikov, Yu. S. Krivonosov, R. Mercier, T. H. Metzger, C. Morawe, and E. Ziegler, Nucl. Instrum. Methods Phys. Res., Sect. A 530, 575–595 (2004). 3V. V. Azarova, V. G. Dmitriev, Yu. N. Lokhov, and K. N. Malitskii, J. Opt. Technol. 69(2), 125–129 (2002). 4M. M. Barysheva, Yu. A. Vainer, B. A. Gribkov, M. V. Zorina, A. E. Pestov, N. N. Salashchenko, N. I. Chkhalo, and A. V. Shcherbakov, Tech. Phys. 58(9), 1371–1379 (2013). 5N. I. Chkhalo, S. A. Churin, A. E. Pestov, N. N. Salashchenko, Yu. A. Vainer, and M. V. Zorina, Opt. Express 22(17), 20094–20106 (2014). 6M. M. Barysheva, B. A. Gribkov, M. V. Zorina, N. N. Salashchenko, and N. I. Chkhalo, J. Surf. Invest.: X-Ray, Synchrotron Neutron Tech. 7(4), 797–801 (2013). 7See http://www.ntmdt.ru/modular-afm/prima for NT-MDT. 2V.
Review of Scientific Instruments is copyrighted by AIP Publishing LLC (AIP). Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. For more information, see http://publishing.aip.org/authors/rights-and-permissions.
Note: A stand on the basis of atomic force microscope to study substrates for imaging optics N. I. Chkhalo, N. N. Salashchenko, and M. V. Zorina Department of Multilayer Optics, Institute for Physics of Microstructures of the Russian Academy of Sciences, GSP-105, 603950 Nizhny Novgorod, Russia
(Received 25 November 2014; accepted 19 December 2014; published online 8 January 2015) A description of a stand based on atomic force microscopy (AFM) for roughness measurements of large optical components with arbitrary surfaces is given. The sample under study is mounted on a uniaxial goniometer which allows the sample to be tilted in the range of ±30◦. The inclination enables the local normal along the axis of the probe to be established at any point of the surface under study. A comparison of the results of the measurement of noise and roughness of a flat quartz sample, in the range of spatial frequencies 0.025–70 µm−1, obtained from “standard” AFM and developed versions is given. Within the experimental error, the measurement results were equivalent. Examples of applications of the stand for the study of substrates for X-ray optics are presented. C 2015 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4905336]
Recently, there has been an urgent need to revise the old and develop new methods for measuring the roughness of super-smooth optical surfaces used in projection nanolithography, X-ray microscopy, and astronomy. A distinctive feature of the elements used in these systems is their curvature. The radius of the curvature may vary from a few millimetres to a few metres, and the deflection is of up to 20 or more millimetres. To ensure diffraction quality imaging (spatial resolution is determined only by the radiation wavelength and numerical aperture of the optics) by such systems, substrates with surface irregularities (roughness) in the root-mean-square (r.m.s.) parameter, integrated over the range of spatial frequencies from 10−3–102 µm−1 (lateral dimensions 1 nm–1 mm) to 0.1–0.2 nm, are necessary, together with adequate roughness metrology. Atomic force microscopy (AFM) is the main method used to study roughness in the high and medium spatial frequency range of 10−1–102 µm−1.1 In some works, such as Refs. 2–5, using as a reference, the “ab initio” method of diffuse scattering of the hard and soft X-rays, it is shown that AFM is quite an adequate method of measuring lateral dimension surface roughness in the range of nanometres to 40–60 µm. However, there is a problem that limits the applicability of AFM in studying super-smooth optical surfaces attended for the high-resolution imaging. The thing is that, for the subatomic resolution of the roughness heights to be preserved, the range of vertical probe displacement should not exceed ∆z = 1 µm. In the case when the surface is inclined to the probe axis, the range of possible lateral scanning decreases yielding a limitation on the range of recorded spatial frequencies of roughness. For example, of a 3◦ inclination, the frame cannot exceed 19 µm.6 The local slopes of the surface to the axis in real optics can be one order of magnitude larger. Therefore, for measurements of sized and curved surfaces, researchers are turning to nonstandard measuring schemes, e.g., the mounting of an AFM head onto the sample, which leads to the significant influence of vibrations on the measurement results and causes the risk of damaging unique and expensive substrates. 0034-6748/2015/86(1)/016102/3/$30.00
This paper describes a stand based on AFM, which allows the problem of the measurement of large curved surfaces to be solved with precision, equal to that of the measurement of flat surfaces by conventional AFM. To solve these problems, we have developed a stand based on the AFM of the NTEGRA Prima Company NT-MDT, Zelenograd,7 and a uniaxial goniometer which allows the sample to be tilted in the range of ±30◦. The inclination enables the local normal along the axis of the probe to be established at any point of the surface under study. A diagram and a photo of the stand are shown in Figs. 1 and 2. The main elements of the stand are a honeycomb table top with a pneumatic vibration isolation system (1), a trapeze (2) which, by means of springs (3), by analogy with the factory configuration of the device, is a suspended measuring system (4). The measurement system consists of a base plate (5), a frame (6), and a cross-beam (7). The AFM head (8) is suspended on the cross-beam. Test sample (9) is mounted on a uniaxial goniometer (10). For the “rough” sample inlet to AFM probe (11), a lifting mechanism (12) mounted on plate (5) with screws (13) is used. The fine adjustment of the sample surface to the probe is performed by means of three micro-screws (14) and springs (15). The lifting mechanism provides vertical movement of the sample to 100 mm and the micro-screws to 25 mm. Thus, the stand allows one to study samples in a wide range of thicknesses, i.e., it is fairly universal. Studying the main characteristics of the stand was undertaken by using quartz samples differing in both surface form and dimensions and roughness values. The main parameters of the samples under study are given in Table I. Effective roughness σeff values are defined as 2 σeff
ν max = PSD(ν)dν,
(1)
ν min
where νmin and νmax are the minimal and maximal spatial frequencies, where the PSD (power spectral density) functions of the roughness are determined.
86, 016102-1
© 2015 AIP Publishing LLC
016102-2
Chkhalo, Salashchenko, and Zorina
Rev. Sci. Instrum. 86, 016102 (2015)
FIG. 2. Photograph of the stand with tested curved sample.
FIG. 1. General scheme of the stand: (a) 1—honeycomb table top with pneumatic vibration isolation system, 2—trapeze, 3—spring, 4—measuring system; (b) 5—base plate, 6—frame, 7—cross-beam, 8—AFM head, 9—test sample, 10—uniaxial goniometer, 11—AFM probe, 12—lifting mechanism, 13—screw, 14—micro-screws, and 15—spring.
Initially, the intrinsic AFM noise on the stand was compared with the noise of the standard (factory) configuration. As the experiment shows, AFM noise in both configurations coincides within the measurement error in thousandths of a nanometre and is about 0.13 nm in the range of spatial frequencies 0.025–60 µm−1. Thus, it was found that the AFM noise in the stand did not increase compared with a standard device. Moreover, there is a tendency for it to reduce. The next step was to compare the measured values of the roughness of the flat sample #1, which could be measured both on the stand and in the standard AFM configuration. Figure 3 shows the measured PSD functions. The solid line shows the data obtained in the standard configuration of the instrument, and the lines with symbols the configuration on the stand. As in the case of the noise, the results of roughness measurements in both configurations of the instrument coincided with high accuracy. Since the noise, characterized by σnois, and the surface roughness, σsurf , are statistically independent random variables, the effective (measured experimentally) roughness σeff can be written as the sum of the squares of the surface roughness and noise 2 2 2 σeff = σsurf + σnois .
(2)
TABLE I. Description of the investigated samples of fused silica: D— diameter; h—thickness; R—the radius of curvature; σ eff —effective roughness measured in the range of spatial frequencies 0.025–60 µm−1; and σ surf —the true roughness value obtained by subtracting the AFM noise.
#
D (mm)
h (mm)
R (mm)
σ eff (nm)
σ surf (nm)
1 2 3
25 250 100
6 31 30
∞ 3981.0 137.0
0.21 0.29 0.71
0.16 0.26 0.70
FIG. 3. PSD functions of the roughness of sample #1 measured in the standard configuration of the instrument (solid lines) and on the stand (lines with symbols).
016102-3
Chkhalo, Salashchenko, and Zorina
Rev. Sci. Instrum. 86, 016102 (2015)
center to the periphery, it is possible to establish the local normal to the surface along the axis of the probe. Thus, first, it is possible to bring the sample to the probe at any point, and second, the minimum inclination of the surface under study with respect to the horizon is reached, thus, one can make measurements at maximal frame, as in the study of flat surfaces. The maximum numerical aperture of the substrates can be NA = 0.3 and a maximum diameter up to 300 mm. The experiments show that in terms of the intrinsic noise, the operating spectral range of roughness and sensitivity of this device is not inferior to the standard AFM, at the same time, allowing the study of sized and curved samples. It is considered appropriate to equip the standard instruments with such an option, as it is sought by optical industry enterprises and groups producing X-ray optics. FIG. 4. PSD functions of the roughness of sample #2 measured on the stand.
This relation shows that when measuring the substrate with an effective roughness approaching 0.2–0.3 nm, the intrinsic AFM noise must be taken into account. In particular, for sample #1, the measured effective roughness was 0.21 nm. In view of the AFM noise, true surface roughness is 0.16 nm; see the last column of Table I. Figure 4 illustrates the application of the developed stand for the study of curved and sized substrate #2 that is designed for X-ray astronomy of the Sun. Sample #2 passed all stages of polishing, except for ion-beam etching. As the conclusion, this paper describes a stand designed on the basis of standard AFM. Unlike standard AFM, the stand, thanks to the integrated uniaxial goniometer, allows the optical elements with arbitrary surface shapes and large dimensions to be studied. This is achieved by the fact that due to the uniaxial goniometer at any point of the surface under study, from the
This work was supported by the Russian Fund for Basic Research, by the Ministry of Science and Education of Russia, and by the programme “Physics and technology of micro- and nanostructures” of the Public Centre of Use at the Institute for Physics of Microstructures of the Russian Academy of Sciences. 1J.
E. Griffith and D. A. Grigg, J. Appl. Phys. 74, R83–R109 (1993). E. Asadchikov, I. V. Kozhevnikov, Yu. S. Krivonosov, R. Mercier, T. H. Metzger, C. Morawe, and E. Ziegler, Nucl. Instrum. Methods Phys. Res., Sect. A 530, 575–595 (2004). 3V. V. Azarova, V. G. Dmitriev, Yu. N. Lokhov, and K. N. Malitskii, J. Opt. Technol. 69(2), 125–129 (2002). 4M. M. Barysheva, Yu. A. Vainer, B. A. Gribkov, M. V. Zorina, A. E. Pestov, N. N. Salashchenko, N. I. Chkhalo, and A. V. Shcherbakov, Tech. Phys. 58(9), 1371–1379 (2013). 5N. I. Chkhalo, S. A. Churin, A. E. Pestov, N. N. Salashchenko, Yu. A. Vainer, and M. V. Zorina, Opt. Express 22(17), 20094–20106 (2014). 6M. M. Barysheva, B. A. Gribkov, M. V. Zorina, N. N. Salashchenko, and N. I. Chkhalo, J. Surf. Invest.: X-Ray, Synchrotron Neutron Tech. 7(4), 797–801 (2013). 7See http://www.ntmdt.ru/modular-afm/prima for NT-MDT. 2V.
Review of Scientific Instruments is copyrighted by AIP Publishing LLC (AIP). Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. For more information, see http://publishing.aip.org/authors/rights-and-permissions.
