Response of thick and thin film λ/2 microstrip rejection filter to leaf moisture

Authors


Abstract

[1] Thick and thin film λ/2 microstrip L section rejection filters in the X and Ku bands have been used to investigate the effect of leaf moisture on their response. The results reported are in the form of a comparative study of the X and Ku band response of the rejection filters. The investigations are aimed at using the overlay technique for studying the effect of leaf moisture changes on the rejection property of the filter. The leaves investigated were Tradescantia (magenta), Pothas Scandens (green), and Acalypha (bicolored red and green). These leaves were chosen because of their differences in thickness, texture, and color or chlorophyll content. The effects in the Ku band are more dramatic than in the X band for high moisture content (fresh leaves). The results show differences in the response of the thick and thin film filters to condition of leaf overlay. The response to moisture content in the dried leaves is almost similar irrespective of type of metallization of the filter, though the thick film filter seems to be more useful in the X band since some differences between 24 hour dried and 48 hour dried leaf overlay are seen.

1. Introduction

[2] Leafy vegetation forms a major constituent of the agricultural sector. The microwave part of the electromagnetic spectrum can be used to study the moisture-related changes occurring in the leaves.

[3] The resonant cavity perturbation technique has been used to study nuts, seeds, grains, etc. [Kraszewski and Nelson, 1993, 1994, 1996]. The antenna has been used to study tea leaves [Okamura and Ma, 1998] and waveguide for corn leaves [Trablesi et al., 1994]. The microstrip component being in planar form can offer an alternative compact device for biomaterial studies. The use of the overlay technique [Karekar and Pande, 1976; Puri, 1994] offers further planarization. There are very few reports [Yogi et al., 1998, 1999; Abegaonkar et al., 1999a, 1999b] on the use of microstripline components to study biomaterials.

[4] Our lab [Rane and Puri, 1999, 2001a] has been investigating the feasibility of Ag thick film as an alternative for thin film microstripline components up to the 18 GHz frequency range. It has been found that the response of the thin film and thick film microstrip components is different for different dielectric overlays [Rane and Puri, 1998, 2001b, 2002]. The λ/2 L section microstrip rejection filter has a high rejection at the resonant frequency, and this is very sensitive to the medium above the filter. This aspect can be used to study the material placed as in-touch overlay on the filter.

[5] In this paper the changes in the rejection properties in the X and Ku bands of thick and thin film λ/2 L section microstrip rejection filters due to leaf overlay is reported. Three leaves Tradescantia (magenta), Pothas Scandens (green), and Acalypha (bicolored red and green) have been used as in-touch overlay on the filter. The changes in the filter response due to leaf surface, orientation, and change in moisture content as the leaf dries are also reported here.

2. Experiment

2.1. Details of Microstrip Rejection Filter

[6] The λ/2 L section microstrip rejection filters in the X and Ku bands were fabricated using thick film and thin film technology. The photo mask used for both these metallizations was the same. The thick film microstrip rejection filter was fabricated using screen-printing technology with a peak firing temperature of 700°C in a three-zone furnace. A firing schedule of 45 min was maintained. The metallization used was silver. The silver paste used was SBR4 [Rane and Puri, 1999], formulated in the lab itself. The thickness of the thick film was ∼10 μm. The thin film filter was delineated by a photolithographic process on copper thin film of thickness ∼5 μm deposited by vacuum evaporation and electroplating. The rejection characteristics of both the thick and thin film filters were measured point by point in the frequency range 8–18 GHz with the help of a microwave bench consisting of source, isolator, attenuator, and detector.

2.2. Leaf Information

[7] For the in-touch overlay, the leaves were cut to a size of 1.5 × 1.5 cm from the center so that the central vein was part of the overlay. The experiments were conducted for fresh leaves, after 24 hours of drying of leaves in air, and after 48 hours of drying of leaves in air. The three leaves chosen for the study were all ornamental plants. The size of Tradescantia plant was ∼0.62 m, Pothas Scandens was ∼1 m, and Acalypha plant had a height of ∼1.54 m. The leaves were taken from full-grown plants from the central region of the plant. The choices of these leaves were made on the basis of differences in thickness, texture, and color (chlorophyll content). The thickness of the fresh leaves was 0.176 cm for Tradescantia, 0.129 cm for Pothas Scandens, and 0.086 cm for Acalypha. The exact chlorophyll content was not measured.

2.3. Leaf Study

[8] Three identical microstrip rejection filters, each of thick film and thin film both in the X and Ku bands, were investigated. The rejection filter circuit to circuit variations were ∼−0.8 dB. Five leaf samples from each of the three leaves were used as overlay. A pressure block of thermocol and glass block was used to hold the leaf in place and to avoid curling of the leaf when dried. Once the leaf was placed as in-touch overlay on the filter, the system was not disturbed for 48 hours. The pressure block did not change the characteristics of the rejection filters when placed over them.

[9] The leaf to leaf variations when leaves were fresh were of the order of ∼−1 dB for Tradescantia and ∼−0.6 dB for the other two. Four types of measurements were done: (1) upper surface of leaf in contact, with central vein parallel to the direction of propagation (USP); (2) lower surface of leaf in contact, with central vein parallel to the direction of propagation (LSP); (3) upper surface of leaf in contact perpendicular to the direction of propagation (USPR); and (4) lower surface of leaf in contact perpendicular to the direction of propagation (LSPR).

3. Results and Discussion

3.1. The λ/2 Microstrip Rejection Filter Without Overlay

[10] Figure 1 shows the frequency response of the thick and thin film microstrip rejection filter in the X (8–12 GHz) and Ku (13.4–18 GHz) bands. The data of the resonance frequency rejection and Q are tabulated in Table 1.

Figure 1.

Thick and thin film microstrip rejection filter characteristics without overlay in the X and Ku bands.

Table 1. Data of λ/2 Microstrip Rejection Filter Without Overlay in the X and Ku Banda
PropertyX BandKu Band
IIIIII
  • a

    I, thick film filter; II, thin film filter.

Designed resonance frequency, GHz10.010.016.016.0
Experimental resonance frequency, GHz9.610.016.816.2
Rejection at resonance, dB−14.3−20.0−7.4−16.7
Q39.533.345.032.5
Off-resonance rejection, dB∼−5.0∼−2.0 to −5.0∼−3.0∼−5.0

[11] In the X band, the thick film filter shows center frequency lower than the designed value, and in the Ku band, it is higher; here the thin film filter also gives higher center frequency. The Q values are not very different in both frequency bands and also for both metallizations. It is expected that the thickness and skin depth effects due to metallization will be negligible. The notch frequency depends more on the length of the resonator used. The size of the L section decreases as the frequency increases. The fabrication tolerances in the thick film process are smaller than those in the thin film process, whereby the width, length, and gap of the resonant section becomes different from the designed value. The designed width of the various straight sections was 625 μm. The actual width was 693 μm for the thin film and 788 μm for the thick film. The edge definition was 19 and 62 μm for the thin and thick film filters, respectively. The edge definition of the thin film filter is superior to the thick film filter. In spite of the width being larger and the edge definition being inferior, the thick film filter shows comparable Q and rejection with the thin film filter, both in the X and Ku band, indicating that the specific problems inherent with thick film technology will not be a major problem in the microwave frequency up to 18 GHz. Leaf overlay studies were done on these filters in order to ascertain the differences in response of the filter fabricated using different technology to the properties of the overlay.

3.2. Thick and Thin Film λ/2 Rejection Filter With Leaf Overlay

[12] The changes in the rejection characteristics of the rejection filter due to in-touch overlay of Tradescantia, Pothas Scandens, and Acalypha kept with the vein parallel to input direction are given in Figures 2, 3, and 4, respectively. The shape of the curves due to overlay kept perpendicular to the input direction were similar to parallel direction, with slight change in the amplitude only due to the fresh leaf. To avoid repetitiveness, only one orientation is plotted.

Figure 2.

Effect of Tradescantia overlay: fresh leaf (solid triangles), 24 hour dried leaf (solid circles), and 48 hour dried leaf (open circles).

Figure 3.

Effect of Pothas Scandens overlay on the thick and thin film microstrip rejection filter: fresh leaf (solid triangles), 24 hour dried leaf (solid circles), and 48 hour dried leaf (open circles).

Figure 4.

Effect of Acalypha overlay on the thick and thin film microstrip rejection filter: fresh leaf (solid triangles), 24 hour dried leaf (solid circles), and 48 hour dried leaf (open circles).

3.3. Effect of Tradescantia (TR)

[13] From Figure 2 it is seen that the fresh leaf has the effect of increasing the attenuation in the Ku band more dramatically than in the X band of the thick film filter and in the X and Ku band of the thin film filter. Multiple resonances are observed in the thin film filter. In the Ku band some surface dependent effects are also observed in both filters. As the leaf dries, the attenuation in the X band decreases, and the thin film filter regains the notch at 10 GHz after 24 hours, whereas the thick film filter regains the notch at 9.6 GHz because of the 48 hour dried leaf overlay. In the Ku band, the thick film filter shows a rejection of ∼−5 dB throughout the frequency range due to 24 hour dried Tradescantia overlay; a broad minimum starts developing at 16.6 GHz because of the 48 hour dried leaf overlay. The thin film filter in the Ku band shows two well-defined notches due to 24 hour dried leaf overlay, which becomes one broad minimum after the 48 hour dried leaf is put as overlay.

3.4. Effect of Pothas Scandens (PS)

[14] Figure 3 depicts the effects due to Pothas Scandens overlay. The response of the thick film filter to the PS overlay is almost similar in the X and Ku bands for all the orientations and both surfaces, with the filter losing its notch property. The attenuation is more due to USP position than due to LSP. After the leaf dries, the filter starts showing resonance at 9.4, 11, and 16.4 GHz. An interesting result is that at 8 GHz, the rejection is high.

[15] The thin film filter shows dramatic increase in rejection in the Ku band because of the fresh leaf, whereas in the X band, though there is an increase in rejection, the filter still shows resonance characteristics. Because of the dried leaf overlay, the filter shows a notch with decrease in rejection in the X band, whereas in the Ku band, the filter shows almost flat response.

3.5. Effect of Acalypha (AC)

[16] As seen from Figure 4, the fresh AC overlay has the effect of making the thick film filter response flat in both X and Ku band frequencies. However, the thin film filter shows a resonance notch at 10.1 GHz, and two minima are observed in the Ku band. Because of dry leaf overlay, both thick and thin film filters regain their no-overlay characteristics in the X band, whereas in the Ku band, the thin film filter attains a broad minimum, and the thick film filter shows a notch due to 48 hour dried Acalypha.

[17] The data of rejection of the thick film and thin film rejection filters in the X and Ku bands for perpendicular (PR) orientation of leaf are tabulated in Table 2. The resonance frequency is the experimentally obtained value without overlay. The off-resonance frequencies in the lower and higher range have also been selected, since in these regions, changes due to leaf overlay are also observed.

Table 2. Data of Rejection of λ/2 Microstrip Rejection Filter in the X and Ku Band When Leaf Is Kept in Perpendicular (PR) Positiona
Filter ConditionRejection, dB
X BandKu Band
8 GHzResonance Frequency11.5 GHz13.4 GHzResonance Frequency18 GHz
IIIIIIIIIIIIIIIIII
  • a

    I, thick film filter; II, thin film filter; TR, Tradescantia, PS; Pothas Scandens; AC, Acalypha.

Without overlay−6.0−2.5−14.3−20.0−6.0−1.0−1.8−5.0−7.4−16.7−2.5−4.8
 
Tradescantia Overlay
USPR fresh−8.2−10.9−10.9−30.0−12.2−34.0−43.0−8.4−46.0−57.0−53.0−18.9
After 24 hours−4.2−4.9−11.1−20.3−3.8−1.9−4.3−5.2−4.6−13.7−5.0−5.1
After 48 hours−5.2−4.0−25.5−18.1−7.0−3.5−1.6−5.4−4.4−15.9−1.0−7.2
LSPR fresh−4.1−12.2−9.4−35.2−9.3−34.5−34.7−5.7−45.2−29.8−42.0−16.3
After 24 hours−5.0−4.8−14.0−20.3−4.5−2.2−4.7−5.6−5.0−12.1−6.0−3.0
After 48 hours−3.5−3.4−19.5−18.1−1.0−3.7−1.6−5.4−5.7−17.2−1.0−7.2
 
Pothas Scandens Overlay
USPR fresh−24.0−5.9−9.8−26.4−21.5−13.5−14.3−19.6−14.8−22.1−18.0−25.3
After 24 hours−7.1−8.9−3.3−28.0−8.6−7.6−1.0−7.9−5.2−15.4−2.0−10.4
After 48 hours−10.2−3.5−4.4−19.6−7.9−4.9−1.6−6.8−2.9−14.9−1.0−8.7
LSPR fresh−21.2−3.2−11.7−26.4−20.5−7.6−8.2−21.6−8.5−26.2−11.5−25.5
After 24 hours−5.4−8.0−6.1−21.9−12.3−8.4−1.0−9.3−4.3−15.6−1.0−11.6
After 48 hours−12.1−3.1−1.0−16.0−10.2−4.7−1.3−7.3−4.1−14.4−1.0−9.0
 
Acalypha Overlay
USPR fresh−10.8−13.5−10.8−9.1−12.2−16.6−7.3−10.9−7.5−16.9−8.0−12.2
After 24 hours−3.2−9.3−15.6−1.9−10.1−12.1−5.2−5.3−6.2−17.2−2.3−7.6
After 48 hours−3.8−7.8−2.1−3.5−6.5−12.7−1.5−5.2−2.9−17.6−1.0−7.3
LSPR fresh−6.7−3.1−11.2−11.7−12.9−14.8−4.3−11.5−8.7−15.6−9.0−12.7
After 24 hours−2.9−1.5−19.0−2.2−10.1−12.9−4.3−5.2−7.0−16.9−5.4−7.2
After 48 hours−3.8−5.5−20.9−2.9−6.5−13.6−1.0−5.2−7.2−17.1−1.0−7.2

[18] The design parameters of the filter have been achieved with air as dielectric on top. Any change in the properties of the medium above changes the properties of the filter. Leaves are heterogeneous media consisting of components with different dielectric behaviors. The amount of water in the leaf is a dominant factor dictating the dielectric behavior of the leaf. The real part of permittivity of pure water is ∼60 at 10 GHz, but the imaginary part increases in the microwave range and attains maximum value at ∼20 GHz [Nyfors and Varnikainen, 1989]. The leaves have complex dielectric constant, with the real part depending on the phase and the imaginary part on the amplitude change. Since our experimental data gave only the amplitude change, the formula suggested by Gouker and Kushner [1994] was used to calculate the phase change due to overlay. The amount of phase trim (in degrees) for a particular overlay is given by

equation image

where Lover = equation image is the length of the overlay.

[19] From the value of phase trim ΔΦ for the overlay, ɛ′eff of microstripline with overlay was calculated. The value of ɛeff,over was found from the capacitance measurement using an HP4284A impedance meter. The values for 10 and 16 GHz were found from extrapolation. In order to obtain the approximate dielectric constant (both the real and imaginary part) of the leaf, the expression given by Kim et al. [1997] was used:

equation image

where d is the thickness of the leaf.

[20] The value of ɛ′eff and ɛ″eff of the leaf is tabulated in Table 3. The values have been obtained at 10 and 16 GHz, the designed resonance frequency, since, as previously indicated, the filter did not show resonance due to leaf overlay. As expected, ɛ′eff (real part) is higher for fresh leaves. Because the measured attenuation varied with the type of metallization and also with the frequency range, this effect is observed in ɛ″eff (imaginary part). The high value of ɛ′eff even after 48 hours drying indicates the presence of free and bound water. It has been reported [Trablesi et al., 1998] that a change of 8% moisture content produces ∼120% variations in ɛ″.

Table 3. Data of Effective Dielectric Constant equation imageeff and equation imageeff as Calculated Using Gouker and Kushner [1994] and Kim et al. [1997]a
Overlay Conditionequation imageeffΔΦ, degequation imageeff equation imageeff
10 GHz16 GHz
IIIIII
  • a

    I, thick film filter; II, thin film filter.

Tradescantia
Fresh88.1130.1137.7     
USP    54.952.3380.8201.9
LSP    126.452.3499.6374.4
USPR    90.7115.0483.0514.2
LSPR    109.9181.4472.8167.4
24 hours30.895.678.9     
USP    71.68.925.11.0
LSP    48.38.918.34.8
USPR    66.88.938.719.0
LSPR    38.78.931.035.0
48 hours15.761.937.3     
USP    79.29.319.35.9
LSP    62.513.319.315.3
USPR    49.925.319.35.3
LSPR    26.616.619.35.3
 
Pothas Scandens
Fresh75.1125.9129.7     
USP   7878.167.0142.6110.0
LSP    137.767.05.0331.0
USPR    107.767.081.967.0
LSPR    78.067.03.7117.0
24 hours32.295.979.6     
USP    99.08.782.521.3
LSP    160.28.718.434.9
USPR    53.28.729.112.6
LSPR    115.48.737.810.9
48 hours15.561.037.0     
USP    80.011.919.618.7
LSP    108.011.917.918.7
USPR    115.711.945.117.0
LSPR    119.111.934.919.6
 
Acalypha
Fresh68.0125.3124.8     
USP    107.1172.845.012.2
LSP    62.094.9116.023.1
USPR    87.6144.88.512.4
LSPR    82.7113.16.113.4
24 hours30.894.076.6     
USP    3.8179.219.116.2
LSP    3.8179.019.07.6
USPR    22.9182.119.14.7
LSPR    9.5179.216.21.9
48 hours21.679.156.7     
USP    10.7147.617.218.9
LSP    17.2146.816.41.0
USPR    30.4143.543.27.4
LSPR    23.8148.48.22.5

[21] All three leaves had high moisture content when fresh: 95% for Tradescantia, 88% for Pothas Scandens, and 65% for Acalypha. Tradescantia was magenta in color, Pothas Scandens was green, and Acalypha was a combination of green and red. After 48 hours, TR had 27% moisture content, PS 29%, and AC 21%. In the Ku band it appears that the moisture is the dominant factor responsible for the attenuation produced in both the thick film and thin film filters and also due to the three leaves, whereas in the X band, metallization and chlorophyll content of the leaf also seem to contribute.

[22] The leaf is kept on the L section of the filter covering the coupling region. The design of these circuits is made in such a way that the odd mode is used for coupling. Apart from the fringing fields present in air, additional modal-equivalent fringing field lengths Δle and Δlo [Kirschings et al., 1983] are also present. The changes in the equivalent fringing field lengths affect the odd and even mode propagation in the presence of overlay over it. The fringing fields in the even mode are mainly present in air. It is this field which interacts with the overlay to produce the changes observed in the filter. When a resonator is used to study the wave-matter interaction, the change in ɛ′ is reflected in the change in fr, and the change in ɛ″ is reflected mainly in the change in bandwidth. The filters not showing any resonance due to fresh leaf overlay in the frequency band studied might be due to a large shift toward the low-frequency side. Such large shifts have been observed by Yogi et al. [1998], using a thin film ring resonator in the X band. The leaves studied were different and thinner than the ones reported in this work. The lower side of the leaf mainly contains veins which are like channels filled with water, and the upper surface is more porous because of stomata present. At the coupling region, when the lower side of the leaf is touching the filter, it is expected to be more like a conducting layer (metallic), whereas the upper surface is expected to behave more like a dielectric. Since not much difference is observed in the behavior because of the two surfaces touching the filter, it appears that the moisture-dominant microwave absorption effects override all other effects.

[23] Another effect which might be occurring at the coupling region of the filter because of the presence of water in leaves is the dissociation of the physisorbed water of the leaves. Because of this, charge transport occurs, and the electrical condition at the overlay is changed. This might add to the unwanted fields generated at the edges of the coupling section, thereby decreasing the transmittance of the circuit. These effects appear to have been enhanced in the Ku band. Because of differences in the edge definition between the thick film and thin film filters, the fringing field structure might be different in the coupling region of the filter. This also causes differences in the response of the two filters to moisture in leaf due to metallization.

4. Conclusions

[24] The response of the leaf-overlaid λ/2 microstrip rejection filter changes because of the condition and type of the leaf overlay. In the X band, differences between the response of thick film and thin film filters have been obtained. It appears that both thick and thin film filters are sensitive to deficiency of chlorophyll (more changes due to Tradescantia), with the thin film filter showing more changes with moisture content.

[25] In the thick film filter, the resonance frequency might have shifted to the lower-frequency side because of fresh leaf overlay. In the Ku band, the response of both the thick film and thin film filters are almost similar. The Ku band seems to be more useful than the X band when overlay has very high moisture content. For lower moisture content, thick film filters are slightly more sensitive (difference between 24 and 48 hours) than thin film filters, and the X band seems to be more useful.

[26] By using the overlay technique, if the phase change due to overlay is measured, the permittivity of various leaves for both fresh and dried conditions at different frequencies can be detected accordingly. Efforts are in progress to use these types of measurements to predict the moisture content and chlorophyll content of the leaves.

Acknowledgments

[27] Vijaya Puri gratefully acknowledges University Grants Commission, India, for the award of “Research Scientist B.”

Ancillary