In recent years, forgeries have become highly specialized because of the powerful printing techniques and advanced scanning technologies and are of sufficient quality that can easily outsmart money counters. Thus, banknotes are encoded with several security systems. In addition, many researchers have developed different forgery detection techniques for different currencies. However, only a few researchers have reported forgery detection and money recognition for Korean banknotes. For instance, studies have reported banknote classification using color and UV information, feature extraction for banknote classification using wavelet transform, and counterfeit banknote detection using full-frame optical coherence tomography[1-3]. These studies rely on visible light, UV, and image processing. However, to the best of our knowledge, no studies have reported on the detection of banknote forgeries based on the microwave near-field imaging method. When two materials interact with each other, a microwave near-field appears around the materials, and its spatial structure is dependent on the dielectric and conductive property of the material. Banknotes are made from different materials having unique electromagnetic properties. Therefore, when a banknote is excited by a microwave, its microwave near-field structure is determined by the spatial structure of materials that make up the banknote. Hence, counterfeit banknotes can be determined by analyzing the microwave near-field distribution of the banknote.

Thermoelastic optical indicator microscopy (TEOIM) visualizes microwave near-field distribution without scanning. It is a polarized light microscopic system comprising a CCD camera that has high spatial resolution of a few micrometers, wide field of view, and increased measurement throughput. Owing to these advantages, the TEOIM method has been used by different studies to monitor electronic device during operations, detect cracks in thin films of conductive metals, and inspect conductive particles embedded in opaque dielectric materials. In addition, studies have also used the TEOIM method to detect air bubbles and water droplets in a dielectric tube, determine the concentration of biological samples, and investigate microwave heating at different thickness values of aluminum thin films[4-11].

In this paper, we used microwave near-field imaging method and the TEOIM technique to detect forgeries on banknotes. We visualized the microwave near-field distribution of real and copied banknotes by using TEOIM system and analyzed their microwave near-field distribution. The real banknote was found to exhibit a strong microwave near-field distribution in the hologram, while the manuscript showed only a standing wave pattern. In this study, we showed that microwave near-field imaging method could be used for forgery detection by analyzing the differences in microwave near-field distribution between real banknotes and copied banknotes.

Figure 1(a) illustrates the measurement setup of TEOIM for forgery detection on banknotes. In the measurement setup, a signal generator (Mini-Circuits, SSG-15G-RC) first generated a microwave signal, which was then amplified by a power amplifier (Mini-Circuits, ZVE-3W-183+) and radiated through the waveguide onto the device under test. A light emitting diode (LED; λ = 530 nm) was used as the light source. The LED was circularly polarized and incident on the optical indicator through the sheet linear polarizer and quarter wave plate. The optical indicator then reflected the light, which passed through the second sheet linear polarizer (analyzer) and then proceeded to the CCD. Figure 1(b) shows a photograph of the measurement setup used in this study. The banknote to be tested is placed on the alumina plate (0.1 cm × 10 cm × 10 cm) located on the back of the optical indicator, which measured 10 cm × 10 cm. Then, the microwave near-field distribution was measured without changing the position of the banknote. The dimensions of the Korean banknotes are as follows: 1000 won: 13.6 cm × 6.8, 5000 won: 14.2 cm × 6.8, and 10,000 won: 14.8 cm × 6.8. Therefore, the measurement for all banknote samples was performed in an area of 6.8 cm × 10 cm.

Figure 1. (Color online) (a) Illustration of the measurement setup for forgery detection of banknotes. The microwave signal is generated by the signal generator and falls through the waveguide onto the device under test (DUT) after being amplified by the power amplifier. The light emitted from the light emitting diode (LED) is left circularly polarized by a sheet polarizer and recorded by the camera after being reflected and elliptically polarized by the optical indicator (OI). The OI is composed of a glass substrate coated with indium tin oxide (ITO) thin film. (b) The photo of the measurement setup.

When the banknote was irradiated by microwave, the microwave near-field was excited around the optical indicator. The excited microwave near-field interacted with the optical indicator, and this interaction changed the polarization state of the reflected light. In this study, we used a glass substrate coated with a thin film of conductive indium thin oxide (ITO) that acted as an optical indicator. The magnetic field component of the microwave near-field generated an electric current in the ITO thin film when the near-field interacted with the films. This induced electric current generated heat through the Joule heating process, and this generated heat was diffused into the glass substrate. This caused a thermal stress on the glass substrate that can be visualized from the photo-elastic effect. When circularly polarized light passed through a stressed medium, its polarization state changes to the elliptically polarized state depending on the stress direction and strength. The thermal stress direction and strength can be measured by measuring the intensity distribution of the probing light passing through a second polarizer (analyzer). The light reflected from the optical indicator was captured by a CCD camera. The normal and sheared stress distribution images (β₁ and β₂) can be measured by rotating the analyzer at 0° and 45°. The β₁ and β₂ distributions can be calculated based on the measured intensity using the following equations[4]:

where I and φ are the measured intensity and the polarizer angle, respectively. The heat source distribution causing thermal stress on the glass substrate was calculated from stress distribution images by following equation[4]:

where q denotes the heat source distribution, C is a constant related to the properties of the optical indicator, and x and y are the spatial position of pixel. The microwave near-field distribution from the heat source distribution can be visualized because the heat source distribution and the microwave near-field distribution are similar.

The microwave near-field distribution that appeared on an optical indicator without a sample was first visualized to investigate the change in microwave near-field distribution by banknotes. Figure 2 shows the microwave near-field distribution of an optical indicator when 11 GHz, 12 GHz, and 13 GHz microwaves were applied. Our findings showed that intense lines appeared along the vertical direction in common. Since TEOIM visualized the time-averaged intensity of microwave near-field, these microwave near-field patterns indicate that standing waves of the microwave near-field are generated in the optical indicator. The measurement result did not show any special pattern of the microwave near-field. Therefore, our findings indicated that the ITO thin film of the optical indicator has a spatially uniform electrical conductivity.

Figure 2. (Color online) The microwave near-field distribution of the optical indicator. (a) at 11 GHz (b) at 12 GHz (c) at13 GHz.

We also visualized the microwave near-field distribution of real banknotes and their copies to investigate whether certain microwave near-field patterns, which could be counterfeit fingerprints, appear on real banknotes. For our measurements, we placed the real and copied banknotes between the optical indicator and the waveguide antenna and then increased the microwave frequency from 7 GHz to 15 GHz in 0.5 GHz increments. Figure 3 shows the microwave near-field distribution of a real and a copy of 10,000 won bill measured at 11 GHz, 12 GHz, and 13 GHz. We considered only three frequencies since the microwave near-field distribution of the real and copied bank notes for other frequencies showed standing wave patterns as observed in experiments conducted without samples.

Figure 3. (Color online) The photographs and microwave near field distributions of the real copied 10000 won bill. (a) Photograph of the real 10000 won bill. (b–d) Microwave near-field distributions of the real 10000 won bill measured at 11 GHz, 12 GHz, and 13 GHz, respectively. (e) Photograph of the copied 10000 won bill. (f–g) Microwave near-field distributions of the copied 10000 won bill measured at 11 GHz, 12 GHz, and 13 GHz, respectively. The black dashed rectangle indicates the region where the microwave near-field was visualized. The red and white dashed circles indicate positions of hologram marks in optical and microwave near-field images.

As shown in the measurement results in Fig. 3, the real banknote showed a region (indicated by a dashed white circle) with a strong local near-field strength in addition to the standing wave pattern, while the microwave near-field distribution of the copied banknote showed only the standing wave pattern. The position of intense regions observed in three different microwave frequencies were identical to that of the hologram mark of the 10,000 won bill. Since the increase and localization of microwave near-intensity indicated a strong interaction between the material and the excitation microwave, our findings indicate that the holographic mark of the 10,000 won bill is strongly coupled with the excitation microwave radiated from the waveguide antenna. Here, the strongly localized and intense microwave near-field region appeared only for these three frequencies. These results indicate that real banknotes can be determined by monitoring the presence of strong microwave near-field regions only at these frequencies.

The measurement results showed the presence of a strong microwave near-field region in the holographic mark and the slightly different standing wave pattern for the real and the copied bank notes. This result indicates that the overall permittivity of the real and copied bank notes was different from each other because the standing wave pattern depends on the effective permittivity and conductivity of the sample. This is consistent with the fact that materials composing real and copied bank notes are different from each other. For instance, while the real bank note is printed on cotton papers, the copied bank note used in the present study was printed on a pulp paper. Thus, whether the sample is made of the same material of construction as the real banknote can be determined by monitoring the change in the standing wave pattern.

To verify whether the present method can be applied to other bank notes, we visualized microwave near-field distributions of 5000 won and 1000 won. Figure 4 shows the microwave near-field distributions of real and copied 5000 won and 1000 won bills measured at 11 GHz, 12 GHz, and 13 GHz. For the 5000 won bill case at 11 GHz and 12 GHz, a strong and localized microwave near-field appeared on the hologram mark, and the standing wave pattern of the real and copied bank note were different from each other. Interestingly, unlike the 10,000 won bill, the 5,000 won bill exhibited a strong microwave near the center of the bill. This result implies that a particular material that strongly interacts with excitation microwave in the 5000 won bill is embedded in the center of the bank note. Although we cannot determine the embedded material, the fake note of 5,000 won bill can be determined from this near-field pattern.

Figure 4. (Color online) The photographs and microwave near field distributions of the real copied 1000 won and 5000 won bills. (a) Photograph of the real 1000 won bill. (b–d) Microwave near-field distributions of the real 1000 won bill measured at 11 GHz, 12 GHz, and 13 GHz, respectively. (e) Photograph of the copied 1000 won bill. (f–h) Microwave near-field distributions of the copied 1000 won bill measured at 11 GHz, 12 GHz, and 13 GHz, respectively. (i) Photograph of the real 5000 won bill. (j–l) Microwave near-field distributions of the real 5000 won bill measured at 11 GHz, 12 GHz, and 13 GHz, respectively. (m) Photograph of the copied 5000 won bill. (n–p) Microwave near-field distributions of the copied 5000 won bill measured at 11 GHz, 12 GHz, and 13 GHz, respectively. The dashed red and white circles indicate locations of hologram marks.

While the microwave near-field of 10000 won and 5000 won bills showed a particular pattern that can be a fingerprint for determination of real bank note, the 1000 won bill showed only a standing wave pattern. The material difference constituting the real and copied banknotes can be confirmed from the near-field distribution because the standing wave patterns of the real and copied bills were different from each other. It should be noted that the microwave near-field did not show a particular pattern as observed in 5000 won and 10000 won bills even though a highly reflective hologram line existed in the 1000 won bill. This indicates that the hologram line in 1000 won bill is made of a different material than other bills. Indeed, the electrical conductivity of the hologram line was confirmed to be more than MΩ, while that of the hologram mark was confirmed to be several Ω by measuring the electrical conductivity of the hologram line of the 1,000 won bill and the hologram mark of the 5,000 won and 10,000 won bills. This result is consistent with the measurement result that excitation microwaves strongly interact with conductive materials, resulting in strong localization and enhancement of near-field strength. This result indicates that a highly reflective hologram line forged with a metal film can be detected by visualizing the microwave near-field distribution of the bank note.

In this study, we investigated the application of microwaves in forgery detection by imaging the distribution of microwave near-field. For this, we selected Korean banknotes of 1000, 5000, and 10000 won and simulated forged banknotes by scanning each of the real banknotes. By using TEOIM, we visualized a strong magnetic microwave near-field distribution on the hologram feature for 5000 and 10,000 won at 12 GHz. Our results distinguished the real banknote from its scanned one. In the case of the 1000 won banknote, which does not have a hologram feature, we found a different near-field distribution than that of the other two banknotes. Therefore, our study suggests that microwave technology based on microwave near-field distribution imaging can be used for detecting counterfeit money and identifying the value of money by including optically opaque microwave sensitive features.

This research was supported by the 2022 scientific promotion program funded by Jeju National University.