Evaluation of complex gonioapparent samples using a bidirectional spectrometer

Transcription

1 Evaluation of complex gonioapparent samples using a bidirectional spectrometer Nina Rogelj, 1, Niko Penttinen, 1 and Marta Klanjšek Gunde 2 1 Institute of Photonics, University of Eastern Finland, P.O.Box 111, FI80101 Joensuu, Finland 2 National Institute of Chemistry, Hajdrihova 19, Ljubljana SI-1000, Slovenia nina.rogelj@uef.fi Abstract: Many applications use gonioapparent targets whose appearance depends on irradiation and viewing angles; the strongest effects are provided by light diffraction. These targets, optically variable devices (OVDs), are used in both security and authentication applications. This study introduces a bidirectional spectrometer, which enables to analyze samples with most complex angular and spectral properties. In our work, the spectrometer is evaluated with samples having very different types of reflection, concerning spectral and angular distributions. Furthermore, an OVD containing several different grating patches is evaluated. The device uses automatically adjusting exposure time to provide maximum signal dynamics and is capable of doing steps as small as However, even 2 steps for the detector movement showed that this device is more than capable of characterizing even the most complex reflecting surfaces. This study presents srgb visualizations, discussion of bidirectional reflection, and accurate grating period calculations for all of the grating samples used Optical Society of America OCIS codes: ( ) Equipment and techniques; ( ) Diffraction and gratings; ( ) Scattering measurements; ( ) Optical devices; ( ) Spectroscopy, surface; ( ) Scattering; ( ) BSDF, BRDF, and BTDF; ( ) Spectrometers. References and links 1. G. Obein, R. Bousquet, and M. E. Nadal, New nist reference goniospectrometer, Proc. SPIE 5880, 58800T (2005). 2. V. B. Podobedov, M. E. Nadal, and C. C. Miller, Improving the performance of the nist five axis goniospectrometer for measurements of bidirectional reflectance distribution function, Proc. SPIE 8065, 80651I (2011). 3. A. Höpe and K.-O. Hauer, Three-dimensional appearance characterization of diffuse standard reflection materials, Metrologia 47(3), 295 (2010). 4. A. Ferrero, A. M. Rabal, J. Campos, A. Pons, and M. L. Hernanz, Spectral and geometrical variation of the bidirectional reflectance distribution function of diffuse reflectance standards, Appl. Opt. 51(36), (2012). 5. G. Obein, T. R. Leroux, and F. Vienot, Bidirectional reflectance distribution factor and gloss scales, Proc. SPIE 4299, (2001). 6. L. Simonot and G. Obein, Geometrical considerations in analyzing isotropic or anisotropic surface reflections, Appl. Opt. 46(14), (2007). 7. E. Perales, E. Chorro, W. R. Cramer, and F. M. Martínez-Verdú, Analysis of the colorimetric properties of goniochromatic colors using the MacAdam limits under different light sources, Appl. Opt. 50, (2011). 8. L. Simonot, M. Hébert, and D. Dupraz, Goniocolorimetry: From measurement to representation in the CIELAB color space, Color Res. Appl. 36(3), (2011). (C) 2015 OSA 24 Aug 2015 Vol. 23, No. 17 DOI: /OE OPTICS EXPRESS 22004

2 9. A. Ferrero, E. Perales, A. M. Rabal, J. Campos, F. M. Martínez-Verdú, E. Chorro, and A. Pons, Color representation and interpretation of special effect coatings, J. Opt. Soc. Am. A 31(2), (2014). 10. A. Takagi, S. Sato, and G. Baba, Prediction of spectral reflectance factor distribution of color-shift paint finishes, Color Res. Appl. 32(5), (2007). 11. E. Kirchner and W. Cramer, Making sense of measurement geometries for multi-angle spectrophotometers, Color Res. Appl. 37(3), (2012). 12. A. Ferrero, A. M. Rabal, J. Campos, A. Pons, and M. L. Hernanz, Variables separation of the spectral brdf for better understanding color variation in special effect pigment coatings, J. Opt. Soc. Am. A 29(6), (2012). 13. R. L. van Renesse, Optical Document Security, 3rd. ed. (Artech House, 2005). 14. L. Kotacka, T. Tethal, and V. Kolarik, Top-quality security optical elements: from holography towards dpi, Proc. SPIE 5954, 59540K (2005). 15. A. Argoitia and S. Chu, Diffractive pigments help document security, Eur. Coat. J 32, (2004). 16. S. Sumriddetchkajorn and Y. Intaravanne, Hyperspectral imaging-based credit card verifier structure with adaptive learning, Appl. Opt. 47(35), (2008). 17. S. Sumriddetchkajorn and Y. Intaravanne, Data-nonintrusive photonics-based credit card verifier with a low false rejection rate, Appl. Opt. 49(5), (2010). 18. N. Rogelj, I. Poberaj, and M. Klanjšek Gunde, Goniospectrophotometric space curves of diffraction gratings and their applicability as appearance fingerprints, Appl. Opt. 52(34), (2013). 19. IEC , Multimedia systems and equipment color measurement and management: Part 2-1. color management default RGB color space srgb, (1999). 20. E. Hecht, Optics (Addison-Wesley, 2002). 1. Introduction A bidirectional spectrometer with an independently movable light source and detector is used to measure the angle-resolved reflection of samples. Here, the ratio of the radiance reflected by sample to that by the reference sample (perfectly reflecting diffuser) at the same conditions is usually used. Such spectrometers are currently used to provide data for standards and reference materials [1 4]. Furthermore, this angular spectrometry is suitable for measuring the appearance of samples with angle-dependent optical properties. The spectral and spatial distribution of reflected light has been used to analyze the properties of gloss [5], surface reflection of complex materials [6], goniochromaticity, pearlescence and interference colors [7 9]. Many industrial applications within the printing, automotive, cosmetic and clothing sectors require controlling the sophisticated appearance of their products as one of the most valuable acceptability criteria [10,11]. Building and testing experimental devices and facilities for such purposes is, therefore, highly desired. Most applications of bidirectional spectrometry concentrate on effect coatings in which thin platelet pigments are uniformly distributed within the transparent medium, causing a strong dependence of appearance on illumination and viewing directions [7, 9 12]. However, the most complex gonioapparent paint finishes have still a smooth angular and spectral distribution in comparison to diffractive optically variable image devices (DOVIDs, also known as optically variable devices, OVDs) traditionally used in security printing for brand protection and document security [13]. A typical OVD contains a collection of custom-organized diffraction gratings, which are commonly formed as a logo or some artwork. An OVD is typically designed to produce identifiable first-order diffraction best observable with point-source white light in nonspecular conditions [13, 14]. Several attempts have been made to evaluate the optical properties of OVDs thus far; for example, identification with simple handheld optical microscopes [15]. In addition, genuine and counterfeit samples have been reported to be distinguishable using hyperspectral imaging [16, 17]. Moreover, diffraction theory has been applied to predict the bidirectional reflection distribution function (BRDF) of diffraction gratings for some selected in-plane illuminationviewing directions [18]. To the best of our knowledge, a BRDF with sufficiently high angular density to measure a real OVD, such as presented in this study, has not yet been published. Such (C) 2015 OSA 24 Aug 2015 Vol. 23, No. 17 DOI: /OE OPTICS EXPRESS 22005

3 an optically complex sample is also perfectly suitable for checking the operational capabilities of the bidirectional spectrometer that was constructed and tested in this study. 2. Device schematic A schematic representation of the bidirectional spectrometer is shown in Fig. 1. In the device, a sample is placed at the center of a rotational table, which also has an attached arm for the light source, a halogen lamp, at an azimuthal angle of 0. The incident angle can be manually adjusted and is selected to be 45 for all samples in this study. The detector is attached to a larger arc, which can also be rotated, and can cover, in polar coordinates, angles from 0 to 60. Similarly, the sample stage with the light source can be rotated fully, which means that it can cover azimuthal angles of 0 to 360. Both rotational axes can be driven so that the angle change may be as small as However, because of sufficient data collection in larger angle steps, the step for both rotating stages is set to 2 for all samples. Fig. 1. Schematic representation of the bidirectional spectrometer. Sample holder (red) is on the center of a rotational table. Light source (blue) is set to 45 incident angle and rotates with the sample table, covering all azimuthal angles. Detector (green) is mounted on a rotatable arc that covers polar angles between 0 and 60. The detector is an optical lens system that couples the light reflected from the sample to the fiber. This allowed us to accurately check the alignment of the collected light, which was one of the most important alignment procedures for the system. A light source was connected to the detector fiber and shone onto the sample in different arc angles. A collimated beam, that was seen in all arc angles, indicated that the light was collected only from the sample surface. We also observed that the light in the detection regions was homogeneous. The lens in the detector could be adjusted, giving us a circular integration diameter range of 1-60 mm. During the measurements the diameter was set to 35 mm. This measure, along with the distance between sample and detector (80.5 cm) allowed us to calculate the detector s opening angle to be 2.5. Choosing a smaller integration diameter, for example 5 mm, would result in detector s opening angle of 0.4 but the light amount collected by the detector would be immensely smaller. Generally, the choice of appropriate detector opening angle is very important as it defines the unique measurement positions in BRDF evaluations, and with this, the angular accuracy. With our device it was possible to use relatively large integration area for the detector, but due to the large arc diameter, we could still have good angular accuracy. A spectrometer Avantes AvaSpec-ULS2048XL with a wavelength range of nm and resolution of 0.6 nm was used as the detector. It was also tested alongside a Konica Minolta CS-2000 spectroradiometer using a white Spectralon R sample. In this test, the light source irradiance was adjusted so that the measured radiance is always in the range of 1-40 (C) 2015 OSA 24 Aug 2015 Vol. 23, No. 17 DOI: /OE OPTICS EXPRESS 22006

4 mw/sr m 2 nm, and both devices recorded simultaneously. After the measurements, the signal of both devices at 550 nm was plotted as the function of each other. This plot showed linear correlation coefficient of 0.99, which indicates that the Avantes spectrometer is fairly linear in the function of incident irradiance. Therefore, the exposure of the spectrometer used was adjusted, with a custom program, to be automatic. This makes maximum detector dynamics possible for all of the bidirectional configurations, which immensely facilitates the measurement process. It is worth pointing out that bidirectional spectral measurements, in general, require a wide range of different exposure times, which can be acquired if the detector is effective at a broad range of spectral radiances. Measurements were post-processed by subtracting detector and environmental noise and normalized with 99% reflecting Spectralon, which served as a white reference. Normalization was done according to angle and wavelength. We are aware that Spectralon might have some non- Lambertian behavior at large angles [3, 4], but the Spectralon was still considered to be the best diffuse sample available. It was observed that the device was capable of measuring samples having higher reflectance than a 5% reflecting Spectralon sample, which corresponds, for this device, to 9% radiance factor with the used light source. 3. Results and discussion Three types of samples were measured: a mirror, an effect coating, and diffraction gratings. This choice assured that samples with very different gonioapparent reflections were evaluated. Measurement of a front surface mirror ensures the testing of a clean specular reflection. The effect coating sample used contains platelet pigments, Iriodin 9502 WR (Merck), consisting of mica-based flakes with thin layer of iron (III) oxide (Fe 2 O 3 ), spread in resinous medium, which causes an interference effect on each pigment and a scattering of light from the pigments inside the coating. Two diffraction gratings, G1 (etched), G2 (electroformed), and an OVD sample were used. Grating G1 had 1000 nm and G2 766 nm grating period [Fig. 2]. The OVD sample was obtained from OVD Kinegram R, Switzerland. It had a 2 2 cm ornament, O-OVD [Fig. 3], and a cm orientation mark, M-OVD. The latter serves for automatic optical positioning of the O-OVD on the production line for exact positioning. The used OVD was an embossed transmissive hologram foil, where the grating period was created by altering between a high refractive index material and polymer. The structure of the measured OVD was evaluated with optical microscope and the inspection showed that the O-OVD has patches with 4 different grating periods; 666 nm, 714 nm, 870 nm, 3080 nm [Fig. 3] and the M-OVD is a uniform grating with period of 900 nm. Displaying the obtained spectra of BRDF is, in general, a challenge. Because the targets are also visually interesting, an srgb representation was generated, under D65 illumination, from the spectra measured on each sample [19]. Figure 4(a) shows this srgb representation for the mirror surface. Azimuthal angles are marked in degrees on the outer edge of the image, and the polar coordinates are displayed from the center of the image to its edge, covering the polar angle from 0 to 60. The majority of Fig. 4(a) is black, which indicates a low directional signal. Only one bright spot, the specular reflection, occurs at the 180 azimuthal angle and around the 45 polar angle. In contrast, the applied effect coating [Fig. 4(b)] shows both specular and diffuse reflection. The latter is shown as a smooth transition of color from light to dark bronze. No part of the corresponding srgb image is black, which shows that at least some light is observed in all directions. The srgb representations of diffraction gratings are much more complex. The diffraction orders are seen in reasonable narrow patches showing a full rainbow of colors. Because the diffraction effects are best seen if the incident plane is perpendicular to the grating s grooves, our samples were measured in that direction. Most of the bidirectional radiance of G1 and G2 (C) 2015 OSA 24 Aug 2015 Vol. 23, No. 17 DOI: /OE OPTICS EXPRESS 22007

5 G1 G2 y [μm] x [μm] Fig. 2. AFM images from the gratings G1 and G2 with measurement area of 1.5μm 5.0μm (left inset), and the corresponding profiles extracted from the images, that were used to calculate the grating periods (right inset). See also Table 1. Fig. 3. Schematic structure of the measured O-OVD sample. The patches are diffraction gratings with a single period: white 666 nm, light gray 714 nm, gray 870 nm, dark gray 3080 nm, black no grating structure. were observed [Figs. 5(a) and 5(b)] in the incident plane, along azimuthal angles 0 and 180. The M-OVD [Fig. 5(c)] shows a similar srgb image to that of G2 since both have a single grating with similar size of a period. In contrast, the O-OVD shows a much more complicated image [Fig. 5(d)]. The diffractions are not observed only at azimuth angles 0 and 180 but also at 30 and 330, indicating that some patches have grating grooves rotated by 30 in respect to the others. The smudges in the incident plane (azimuth angles 0 and 180 ) indicate the sample has several different grating periods, because the overlapping grating orders produce mixed reflection over the wavelengths. The angular and spectral capability of the built bidirectional spectrometer was further evaluated by calculating grating periods using the measured diffractions. For this purpose, the diffraction grating equation was used [20]: d = mλ/(sinθ i + sinθ m ), (1) where d is the grating period, θ i the incident angle, θ m the diffraction angle, m is the order of (C) 2015 OSA 24 Aug 2015 Vol. 23, No. 17 DOI: /OE OPTICS EXPRESS 22008

6 (a) (b) Fig. 4. srgb image of (a) front mirror surface and (b) coating with special effect pigments. the diffraction, and λ is the wavelength. Depending on the surface, different areas of the data were selected for the grating period calculations. To avoid specular reflection, only the spectra measured at polar angles between 0 and 40 were used. The G1, G2 and M-OVD samples have diffraction patterns only along the azimuthal directions 0 and 180 ; therefore a stripe of ±6 width was selected around that direction. Similarly, in case of O-OVD, an additional stripe near 30 of azimuth is used with the same width to account for diffractions observed outside the 0 and 180 directions. At each wavelength, measurements were averaged along the azimuthal spreading of ±6, creating a vector of data. After this, the diffraction angle θ m was obtained from the polar angle, where the maximum radiance was obtained. Interpolation along the polar angle was done for reduced deviance in the search for these maxima. Because each measured wavelength follows the grating equation (1), all wavelengths from a single sample create a statistical evaluation of the grating period of a single sample. Table 1 shows and compares the visual and AFM defined grating periods with the calculated ones. It further lists the standard deviation of the calculated grating periods. Finally, the calculated grating periods match very well with the real grating periods, and the calculated grating periods show very small deviation. Table 1. Grating periods, defined by AFM (G1 and G2, see Fig. 2) and optical microscope (M-OVD and O-OVD, micrographs not shown here), and calculated with standard deviation. The difference between the periods is expressed by relative difference. sample grating period (nm) calculated grating period (nm) relative difference G1 766 ± ± 10 4% G2 996 ± ± 9 0.6% M-OVD ± 7 3% ± 8 0.2% O-OVD ± 5 3% ± 7 1% ± 87 3% The spectral capability of the bidirectional spectrometer was analysed also using normalized reflectance spectra for azimuthal angle 0 and polar angles 0, 8 and 16 for G1 and O-OVD [Fig. 6]. Each individual spectrum of G1 [Fig. 6(a)] contains only one peak - first order diffrac- (C) 2015 OSA 24 Aug 2015 Vol. 23, No. 17 DOI: /OE OPTICS EXPRESS 22009

7 (a) G1 (b) G2 (c) M-OVD (d) O-OVD Fig. 5. srgb images of diffraction gratings (a) G1, (b) G2, and both parts of holographic foil (c) M-OVD, (d) O-OVD. tion with m = 1 in equation (1), while higher order diffractions are at shorter wavelengths outside the visible range. At larger polar angles, the diffractions shift towards longer wavelength. These properties result in color shown at the corresponding positions in the srgb image in Fig. 5(a). Each reflectance spectrum of O-OVD [Fig. 6(b)] shows several diffraction peaks, coming from patches with four different grating periods. The strongest maxima correspond to first order diffractions from gratings with periods 666, 714 and 870 nm, as marked on the figure. In addition, the second order diffraction of grating with period 870 nm is also visible. The patches with the largest grating period (3080 nm) is rotated by 30 respect to the incident plane and has its first order peaks at wavelengths in the near infrared. The higher order diffractions are visible as small peaks or even shoulders, overlapping the described first order diffractions. The positions of all peaks in reflectance follow equation (1) for corresponding grating period. The number of diffraction peaks and their position is additional proof that O-OVD sample consists of patches with four different grating periods with one of them rotated by 30. All reflectance spectra correspond to the colors shown on the srgb image of Fig. 5(d). The study confirms that the bidirectional spectrometer has sufficient spatial and spectral capability to resolve all the diffractions obtained even on the complex surface of O-OVD. (C) 2015 OSA 24 Aug 2015 Vol. 23, No. 17 DOI: /OE OPTICS EXPRESS 22010

8 (a) G1 (b) O-OVD Fig. 6. Normalized reflectance spectra for (a) G1 and (b) O-OVD, for azimuthal angle 0 and polar angles specified in the legend. In case of O-OVD, the grating period (in nm) associated with the peak is shown next to it, along with the diffraction order. 4. Conclusions In this study, we show a bidirectional spectrometer and evaluate its performance using samples with very different angular- and spectral reflections. The device has adjustable exposure time and angular accuracy that are sufficiently high for accurate angular and spectral measurements of the most optically complex samples into the entire hemisphere. The measured diffraction gratings and OVD indicate that the high accuracy angular and spectral reflection could store a vast amount of information, as seen from the polar srgb images seen in this study. These data were capable to provide also the period of measured diffraction grating and an OVD sample with up to 3% relative difference to the real one. Moreover, individual gratings used in OVD design could be resolved, including the direction of the grooves and the corresponding grating period. The obtained results also show that the build bidirectional spectrometer could efficiently analyze authenticity of diffractive gonioapparent products of higher complexity, which could be used in very wide range of security or authentication applications. Therefore, designing and testing accurate bidirectional spectrometers, such as the one in this study, could be considered beneficial in the future development of these complex and interesting surface types. (C) 2015 OSA 24 Aug 2015 Vol. 23, No. 17 DOI: /OE OPTICS EXPRESS 22011