**4. Results and discussions**

248 Acoustic Waves – From Microdevices to Helioseismology

Piranha solution (one part of 30% H2O2 in three parts H2SO4). After 2 min exposure time, the sensors were rinsed with distilled water. The surface was dried in a stream of nitrogen gas. The SU 8 – 2002 sample was dispensed on MTSM sensor surface and sensors were spin coated for 40 seconds. The sensors were then soft baked for 1 min at 95 oC. The SU 8-2002 films were exposed to UV light for 4 seconds under 25 mJ/cm2. This was followed by 1 min

The reference measurements were taken for air and phosphate buffer saline (PBS). Next, the sensors were exposed to rabit-immunoglobulin G (IgG) (50 μg/ml) suspended in diwater (Fisher Scientific, pH: 5.34, Cat No: 25—555-CM) for 50 minutes to allow IgG coating of the

The thicknesses of the SU 8 – 2002 films were determined by using optical profilometer (Zygo Inc. Model #: NV6200). For the thickness measurements, a very small portion of MTSM sensor surface was not exposed to UV light. After the films were developed, the SU 8-2002 layer was removed from this portion. To obtained different thicknesses of film layer,

The surface topography of the film layer was measured using atomic force microscopy (AFM). The prepared samples were placed on a glass slide installed on the atomic force microscope (Bioscope; Veeco), that was mounted on the inverted fluorescence microscope (TE2000; Nikon, Melville, N.Y.). Measurements were made using contact mode with a scan

A 14 mm diameter, 0.33 mm thick, 5 MHz quartz crystal with deposited 7 mm gold electrodes was placed in a custom fabricated brass sensor holder (ICM). The sensor holder was connected to a Network Analyzer (NA) (HP4395A). A LabView program on a personal computer was used to control the network analyzer and collect the data at 5, 15, 25 and 35 MHz. The experiments were done in room temperature (24oC±1oC). Magnitude and phase responses of MTSM sensor were monitored during the experiments (figure 6). The sampling

1:1 solution of SU8-2002 and cyclopentanone (Acros Organics) was prepared.

**d. Measurement system and MTSM sensor data analysis technique** 

rate was 30 seconds. Each experiment was repeated three times.

Fig. 6. a) Magnitude and b) phase responses of MTSM sensor

hard baking on hot plate at 95 oC.

sensor surface by adsorption.

rate of 2 Hz.

**b. Antibody adsorption on MTSM sensor surface** 

**c. Characterization of geometrical properties of the thin film** 

Initially, two different thicknesses of SU8 2002 layers were spin coated on sensor surface and changes in the frequency and magnitude responses were monitored at 5, 15, 25 and 35 MHz. The thicknesses of the layers were measured by using optical profilometer (fig. 7a). The average thicknesses of the layers were 1920±25 nm and 770±50 nm respectively. Surface topography of the SU8 - 2002 layers was measured by using AFM (fig. 7b). The average roughness of the layer was 20 nm and no cracks on the surface were observed.

Fig. 7. A) Thickness measurements from optical profilometer sample a. SU8-2000 solution sample b. 1:1 dilution of SU8-2002 and cyclopentanone B) Surface topography of SU 8 layer

### **a. Determination of mechanical and geometrical properties of SU8 layer of 1.92 μm thickness**

First set of experiments were performed by spin coating 2 μm thick SU 8 - 2002 layer on sensor surface. The MTSM/GA determined properties are presented in table 2. The average thickness of the polymer layer determined to range from 2080 nm to 2140 nm among the harmonics. Although these values are slightly higher than the value (1920±25 nm) obtained in control experiments, they are still in less than 10% experimental errors. The variation between the frequencies for density value was also very small, ranging from 1240 to 1253 kg/m3. These numbers correlate well with the literature value of 1200 kg/m3 (Jiang et al., 2003) for SU8.


Table 2. Comparison density and thickness values of SU 8-2002 layer determined using MTSM/GA sensor at 5, 15, 25 and 35 MHz with profilometer and Jiang et al. (Jiang et al. 2003)

Modeling of Biological Interfacial Processes Using Thickness–Shear Mode Sensors 251

film thickness, the losses increase to -1.9 dB and -11.5 dB at 5 MHz and 35 MHz respectively,

The shear modulus values determined via the MTSM/GA technique are presented. Both loss and storage modulus were decreased compared to the values obtained when film thickness was 2 μm (figure 8). It has been shown that the scale effect on the mechanical properties of the polymers might be the reason for the decrease in the values (Liu et al, 2009,

Fig. 8. a) Storage and b) loss modulus as a function of harmonic frequency for 770 nm and 1920 nm thick SU8 layer at 5, 15, 25 and 35 MHz (error bars are smaller than symbols when

Third set of experiments were done by adsorbing an antibody layer on MTSM sensor surface under static conditions at 5, 15, 25 and 35 MHz. Antibodies play crucial importance in many applications such as biosensing (Hanbury et al. 1996) and drug delivery (Morrison et al., 1995). The sensor surface was saturated with antibody to form a uniform protein layer on the surface. Change in the frequency and magnitude responses at 15, 25 and 35 MHz are presented in figure 9. At the fundamental frequency (5 MHz), high fluctuations observed in sensor response are likely due to insufficient energy trapping as described by others (Li et

Fig. 9. Time response of A. resonant frequency and B. maximum magnitude responses of

MTSM sensor to antibody binding at 15, 25 and 35 MHz

**c. Determination of mechanical and geometrical properties of an antibody layer** 

while initial loses were similar to what observed for 770 nm film thickness.

Luo et al, 2003).

not visible)

al. 2004).

The frequency dependent shear modulus of SU 8-2002 layer obtained using the MTSM/GA is presented in table 3. Both loss and storage modulus varies with the operating frequency. These extracted values were compared with the values obtained by Jiang et al (2003) (table 3). Jiang et al calculated the shear modulus of SU8 layer by using the impedance-admittance characteristics of the equivalent circuit models of loaded and unperturbed TSM sensors operating at 9 MHz.


Table 3. Comparison of determined GI and GII values of SU8 layer using MTSM/GA at 5, 15, 25 and 35 MHz and Jiang et al (2003)

As seen in table 3, the values obtained by Jiang et al. fall between the values obtained using the MTSM/GA method for 5 and 15 MHz. The small variation in the GI and GII may be due to difference in the film preparations. Alig et al. (1996) has shown that variations in film preparation methods can affect the mechanical properties of the polymer layers.

### **b. Determination of mechanical and geometrical properties of SU8 layer of 0.770 μm thickness**

The second set of experiments was done with the ~770 nm thick SU 8-2002 layer on MTSM sensor. As seen from the table 4, the thickness of the layer determined using the MTSM/GA method correlates well with the expected thickness for each harmonic (less than 10% error). Furthermore the results vary only 10 nm between the harmonics. Similarly, determined values for density were consistent between the harmonics, which is around ~1200 kg/m3.


Table 4. Determined density and thickness values by MTSM/GA or 770 nm thick SU8 layer at 5, 15, 25 and 35 MHz

The initial losses before coating were -0.53 dB and -2.5 dB for 5 and 35 MHz respectively. The losses increase to -0.59 dB for 5 MHz and -4.18 dB for 35 MHz. As seen from these results, the losses remain relatively low when the thickness of the layer was decreased to 770 nm in contrast to the phenomenon observed when the film thickness was 2 μm. For 2 μm

The frequency dependent shear modulus of SU 8-2002 layer obtained using the MTSM/GA is presented in table 3. Both loss and storage modulus varies with the operating frequency. These extracted values were compared with the values obtained by Jiang et al (2003) (table 3). Jiang et al calculated the shear modulus of SU8 layer by using the impedance-admittance characteristics of the equivalent circuit models of loaded and unperturbed TSM sensors

Frequency MTSM/GA Results Jiang et al.[38]

5 (4.55±2.12) x107 (1.89±0.26) x105

15 (2.33±0.18)x108 (1.00±0.03) x106 25 (3.82±0.52)x108 (4.69±0.52) x106 35 (5.81±0.71) x108 (6.49±0.18) x106

5 820±45 1180±40

15 820±20 1190±30 25 810±35 1190±50 35 810±52 1213±35

(MHz) GI (N/m2) GII (N/m2) GI (N/m2) GII (N/m2)

Table 3. Comparison of determined GI and GII values of SU8 layer using MTSM/GA at 5, 15,

As seen in table 3, the values obtained by Jiang et al. fall between the values obtained using the MTSM/GA method for 5 and 15 MHz. The small variation in the GI and GII may be due to difference in the film preparations. Alig et al. (1996) has shown that variations in film

The second set of experiments was done with the ~770 nm thick SU 8-2002 layer on MTSM sensor. As seen from the table 4, the thickness of the layer determined using the MTSM/GA method correlates well with the expected thickness for each harmonic (less than 10% error). Furthermore the results vary only 10 nm between the harmonics. Similarly, determined values for density were consistent between the harmonics, which is around ~1200 kg/m3.

MTSM Frequency MTSM/GA Results Profilometer Jiang et al.[39]

Table 4. Determined density and thickness values by MTSM/GA or 770 nm thick SU8 layer

The initial losses before coating were -0.53 dB and -2.5 dB for 5 and 35 MHz respectively. The losses increase to -0.59 dB for 5 MHz and -4.18 dB for 35 MHz. As seen from these results, the losses remain relatively low when the thickness of the layer was decreased to 770 nm in contrast to the phenomenon observed when the film thickness was 2 μm. For 2 μm

(MHz) d(nm) ρ (kg/m3) d (nm) ρ (kg/m3)

preparation methods can affect the mechanical properties of the polymer layers.

**b. Determination of mechanical and geometrical properties of SU8 layer of 0.770 μm** 

( at 9 MHz)

7.80e7 2.00e5

770±50 1200

operating at 9 MHz.

MTSM

**thickness** 

at 5, 15, 25 and 35 MHz

25 and 35 MHz and Jiang et al (2003)

film thickness, the losses increase to -1.9 dB and -11.5 dB at 5 MHz and 35 MHz respectively, while initial loses were similar to what observed for 770 nm film thickness.

The shear modulus values determined via the MTSM/GA technique are presented. Both loss and storage modulus were decreased compared to the values obtained when film thickness was 2 μm (figure 8). It has been shown that the scale effect on the mechanical properties of the polymers might be the reason for the decrease in the values (Liu et al, 2009, Luo et al, 2003).

Fig. 8. a) Storage and b) loss modulus as a function of harmonic frequency for 770 nm and 1920 nm thick SU8 layer at 5, 15, 25 and 35 MHz (error bars are smaller than symbols when not visible)

### **c. Determination of mechanical and geometrical properties of an antibody layer**

Third set of experiments were done by adsorbing an antibody layer on MTSM sensor surface under static conditions at 5, 15, 25 and 35 MHz. Antibodies play crucial importance in many applications such as biosensing (Hanbury et al. 1996) and drug delivery (Morrison et al., 1995). The sensor surface was saturated with antibody to form a uniform protein layer on the surface. Change in the frequency and magnitude responses at 15, 25 and 35 MHz are presented in figure 9. At the fundamental frequency (5 MHz), high fluctuations observed in sensor response are likely due to insufficient energy trapping as described by others (Li et al. 2004).

Fig. 9. Time response of A. resonant frequency and B. maximum magnitude responses of MTSM sensor to antibody binding at 15, 25 and 35 MHz

Modeling of Biological Interfacial Processes Using Thickness–Shear Mode Sensors 253

The density of the antibody layer was also determined by the MTSM/GA to be 1030±14 kg/m3. This density value is close to the water density in which the antibodies were suspended. Hook et al. (2002) considered the density of antibody layer as 1050 kg/m3 when the antibodies were not attached to gold surface. After the cross-linking, the density value was1300 kg/m3, this is closer to the density value of dry protein. Voros (2004) also showed that the wet density of antibody layer is significantly different than the dry protein density value due to the solvent present in the adsorbed proteins. Therefore we believe that the

MTSM Frequency Thickness (nm) Density (kg/m3) GI (N/m2) GII (N/m2)

Table 6. Determined properties for antibody layer by MTSM/GA at 15, 25 and 35 MHz

As seen from the table 6, the adsorbed antibody layer has low storage modulus (<1e5 N/m2), and relatively higher loss modulus. While storage modulus was same for each harmonic, loss modulus changed with frequency. It has been experimentally shown that the adsorbed protein layers on TSM sensor, such as antibody, vesicles and cells do not behave like "rigid and thin" films (Voinova et al, 2002). Therefore the linear relationship between resonant frequency shift and mass deposition is not observed. Saluja et al. (2005) indicated low concentrations (less than 60 mg/ml) of antibody suspension behave like Newtonian medium. But it should not be expected that the properties of adsorbed layer will not be the same as the properties of antibody suspension. The effect of the binding between protein layer and gold layer should be considered. No literature value was found for direct comparison. Therefore we believe that MTSM/GA technique will lead to development of a

It was shown that MTSM sensor combined with genetic algorithm can be used to extract mechanical and geometrical properties of biological layers. The developed technique was first experimentally tested with SU8-2002 polymer layers with known properties having two different thicknesses. It was shown that the developed technique was successfully determined the mechanical and geometrical layers of thin polymer layers. MTSM/GA technique was then applied to extract the properties of antibody layer coated on MTSM sensor. The obtained data support our hypothesis about use of MTSM/GA technique can be a powerful tool for quantitative characterization of interfacial biological interfacial

We are thankful to Dr. Moses Noh for providing supplies and micro-fabrication facilities for

15 11±0.3 1050±10 (5.20±0.5) x104 (4.80±0.58) x105 25 10.4±0.6 1080±12 (5.00±0.13) x104 (9.50±1.40) x105 35 10.3±0.4 1040±14 (5.60±0.12) x104 (1.52±0.31) x106

determined value of the density is in a reasonable range.

quantitative tool for study of biological interfacial processes.

**5. Conclusions** 

processes.

**6. Acknowledgments** 

polymer coating.

The properties of the medium were determined at t1 = 10 and t2 = 70 minutes. At t1 = 10, the system is modeled as MTSM sensor loaded with semi-infinite Newtonian medium ( DIwater) (fig 10A). The height of the column (2 mm) was much higher than the penetration depth of the acoustic wave at 5 MHz (~250 nm in DI water).

At t2 = 10 min., the MTSM/GA determined properties of the layer at 15, 25 and 35 MHz are presented in table 5. The variations in the determined thickness values were very high (ranging from 300 nm to 5 μm due to the fact that column height was much larger than the penetration depth. Solution range for thickness values was set to be between 1 nm to 10 μm in genetic algorithm. Thus any thickness value larger than the penetration depth will satisfy the solution because the MTSM sensor is not sensitive to the changes beyond the penetration depth. However, the solutions were always higher than penetration depth as expected. Due to the high fluctuations in thickness values, it was not presented here. In contrast the solutions for ρ1, η1 and C1 match with the literature values very well. (Literature values are ρ1 = 1000 kg/m3, η1 = 0.001 kg/m.s and C1 = 0 N/m2 at room temperature (Greczylo and Deboswka 2005)).

Fig. 10. Physical model for MTSM sensor system at A) t=10 and B) t=70


Table 5. Determined properties for semi-infinite Newtonian medium layer by MTSM/GA at 15, 25 and 35Hz

At t = 70 min., the physical model is presented in fig 10b. A viscoelastic layer (protein layer) with finite thickness and semi-infinite Newtonian medium were loaded on MTSM sensor. The properties for diwater layer were entered into the algorithm as known variables and the unknown properties (ρp, ηp, C1 and dp) of viscoelastic layer were determined using the MTSM/GA method. The results are presented in table 6. The thickness of the layer was determined to range from 10.3 to 11 nm for the harmonics. This number correlates well with the values presented by the other researches. Westphal et al (Westphal and Bornmann 2002) calculated the height of antibody layer as 9.2 nm. Furthermore Liao et al (Liao et al 2004) measured the average height of the antibody layer as 10.1±3.3 nm.

The density of the antibody layer was also determined by the MTSM/GA to be 1030±14 kg/m3. This density value is close to the water density in which the antibodies were suspended. Hook et al. (2002) considered the density of antibody layer as 1050 kg/m3 when the antibodies were not attached to gold surface. After the cross-linking, the density value was1300 kg/m3, this is closer to the density value of dry protein. Voros (2004) also showed that the wet density of antibody layer is significantly different than the dry protein density value due to the solvent present in the adsorbed proteins. Therefore we believe that the determined value of the density is in a reasonable range.


Table 6. Determined properties for antibody layer by MTSM/GA at 15, 25 and 35 MHz

As seen from the table 6, the adsorbed antibody layer has low storage modulus (<1e5 N/m2), and relatively higher loss modulus. While storage modulus was same for each harmonic, loss modulus changed with frequency. It has been experimentally shown that the adsorbed protein layers on TSM sensor, such as antibody, vesicles and cells do not behave like "rigid and thin" films (Voinova et al, 2002). Therefore the linear relationship between resonant frequency shift and mass deposition is not observed. Saluja et al. (2005) indicated low concentrations (less than 60 mg/ml) of antibody suspension behave like Newtonian medium. But it should not be expected that the properties of adsorbed layer will not be the same as the properties of antibody suspension. The effect of the binding between protein layer and gold layer should be considered. No literature value was found for direct comparison. Therefore we believe that MTSM/GA technique will lead to development of a quantitative tool for study of biological interfacial processes.
