**7. Gas sensing characteristics of RSAW/STW resonant sensors coated with solid and semisolid chemo sensitive films**

In the following sections we present results from gas probing experiments on RSAW/STW sensors coated with solid HMDSO and semisolid ST and AA films. Four different chemical agents at different vapor concentrations are used for gas probing as follows:


intended for high-resolution measurements at low gas concentrations, then a thickness close to 100 nm should be chosen due to the highest Q and lowest loss. If measurements at higher gas concentrations are expected then a 300 nm thickness may be more appropriate since the

**6.1 Critical thickness in RSAW/STW based sensor resonators coated with solid and** 

As shown in the previous sections, the sensing layer does not only shift the resonant frequency down, increase the loss and decrease the loaded Q as a result of mass loading but it also influences the longitudinal modes supported by the resonator geometry that appear on the left side of the main resonance. In the uncoated resonator these modes are well enough suppressed (typically by 5 to 15 dB) and do not cause any problems when the resonator is operated in an oscillator circuit. As soon as a sensing layer is deposited on the surface, it will change the phase conditions along the device topology and this will cause the adjacent longitudinal modes to arise in magnitude at the expense of the main resonance. This situation gets worse at thick solid films for both, the STW and the RSAW mode. At a certain thickness which we call "critical thickness" the magnitude of the first adjacent lowfrequency longitudinal mode on the left becomes equal to the magnitude of the higherfrequency main resonance. This creates a potential for instability in the sensor oscillator stabilized with this sensor since it can easily jump from the main resonance onto the left longitudinal mode during gas probing which will ruin the measurement. The critical thickness situation is illustrated in Fig. 5 a) and b) for a STW and a RSAW device from Table 1, accordingly, in the process of Parylene C deposition as described in Section 6. As evident from Fig. 5 a), at thickness values above 185 nm the first longitudinal mode on the left starts rapidly growing until its magnitude becomes equal to the main resonance. The critical thickness at which this happens is about 350 nm. At this thickness also a strong Love mode excitation on the right is observed. The RSAW device in Fig. 5 b) reaches its critical thickness at about 650 nm. From these data we can draw the conclusion that the devices from Table 1 can be usable as Parylene C coated sensors as long the film thickness is lower than 300 nm and 600 nm for the STW and RSAW devices, respectively. Comparing the coating behavior of both modes in Fig. 5 we see that the STW mode retains a much better behaved resonance than its RSAW counterpart until the critical thickness is reached. At that thickness the STW device has a loss of 14 dB (Fig. 5 a)), versus 35 dB for the RSAW device (Fig. 5 b)). Therefore, the STW mode tolerates solid and semisolid sensing films much better than the RSAW one

thicker layer may adsorb larger amounts of gas without film saturation.

and is more appropriate for operation with such films in practical gas sensors.

agents at different vapor concentrations are used for gas probing as follows:

• Dichloroethane 6500 ppm • Ethylacetate 17600 ppm • Tetrachloroethylene 2650 ppm • Xylene 1400 ppm

**solid and semisolid chemo sensitive films** 

**7. Gas sensing characteristics of RSAW/STW resonant sensors coated with** 

In the following sections we present results from gas probing experiments on RSAW/STW sensors coated with solid HMDSO and semisolid ST and AA films. Four different chemical

**semisolid sensing layers** 

The purpose of the gas probing tests is to identify which acoustic wave mode provides better performance in real-world gas sensing conditions.

Fig. 5. Critical thickness in a) STW and b) RSAW devices in the process of Parylene C coating

### **7.1 Computer controlled automatic system for gas probing measurements**

The block diagram of the computer controlled system for measuring the gas sensing characteristics of the RSAW/STW polymer coated sensors is shown in Fig. 6. For correct comparison of the gas probing performance of both acoustic wave modes four pairs of devices (one RSAW and one STW sensor in each pair, coated with the same polymer to the same film thickness and in the same deposition process) are mounted in open TO 92 packages and placed in the sensor head which can accommodate a total of eight sensors. Each device is connected to one of the 8 sensor oscillator circuits in the head. During gas probing each of the 8 oscillators is turned on for a short period of time to take the measurement. The oscillators are operated one at a time and multiplexed consecutively to avoid possible injection locking. Their output frequency is down converted to an intermediate frequency in the 4-9 MHz range by means of a stable heterodyne reference oscillator to allow fast high-resolution measurements with a reciprocal frequency counter. The chemical compounds 1 through 4 used for gas probing are vapors from the 4 liquidphase analytes in the 4 containers. A permeation cell is placed on top of each container to allow a defined vapor pressure which is controlled by the rotation speed of the pump. By a switch block of valves the vapors of each analyte are then consecutively fed to the sensor head where they interact with the sensors. After the measurements at each analyte are completed the sensors are flushed with dry air passing through a silica gel integrator which provides also a homogenous air and gas flow. The entire system is controlled by a computer which performs measurements in probe-flush cycles over time and provides real-time sensor data on the computer screen. The data in Fig. 7 is the gas probing performance of a 700 MHz STW styrene coated sensor probed with dichloroethane vapors at 6500 ppm concentration in 100 s probe-flush cycles. It should be noted that prior to this measurement the sensor was probed with a different compound (xylene) for 62 hours and 40 minutes. Note the excellent reproducibility of the noise free sensor signal with a magnitude

Polymer Coated Rayleigh SAW and STW Resonators for Gas Sensor Applications 531

This study aims at finding out which of both acoustic wave modes provides better gas sensitivity when coated with solid HMDSO films and what is the optimum film thickness at which maximum sensitivity is achieved. For this purpose, 5 pairs of RSAW/STW devices according to Table 1 were coated at 5 different HMDSO thicknesses (50, 100, 190, 280 and 350 nm) each, in the same plasma deposition process for each pair. Figure 8 compares the gas sensing characteristics of both modes gas probed with tetrachloroethilene at 2650 ppm concentration. The results from all gas probing experiments on the 5 pairs of devices are summarized in Table 2. In this table the "sensitivity factor" is the ratio between the relative sensitivities (in ppm) for the two devices of each pair. It is given for each of the 5 film thicknesses and shows which mode is more sensitive and at which thickness. From the data

1. *The HMDSO coated sensors have very short response times and reach adsorption-desorbtion equilibrium just a few seconds after the gas flow is applied. We attribute this behavior to the* 

> **Ethylacetate 17600 ppm**

> > 2.8 KHz (4 ppm)

4 KHz (9.2 ppm)

4.8 KHz (6.9 ppm)

3.8 KHz (8.8 ppm)

8.5 KHz (12.1 ppm)

6.5 KHz (15 ppm)

14 KHz (20 ppm)

2.3 KHz (5.3 ppm)

20 KHz (29 ppm)

Table 2. Gas sensitivity comparison of RSAW vs. STW devices coated with solid HMDSO film of 5 film thicknesses. Data on the 100 nm coated SAW device are not available

**Tetrachloroethylene 2650 ppm**

> 3.5 KHz (5 ppm)

3 KHz (6.9 ppm)

7.5 KHz (10.7 ppm)

3.5 KHz (8.1 ppm)

7 KHz (10 ppm)

6 KHz (14 ppm)

15 KHz (21.4 ppm)

5.1 KHz (11.8 ppm)

**37 KHz (53 ppm)** 

0.82 0.43 0.72 0.68

**1.02** 0.78 **1.32** 0.91

**1.49** 0.81 0.71 0.57

**3.74 3.77 1.81 1.4** 

**Xylene 1400 ppm** 

2.4 KHz (3.4 ppm)

2.2 KHz (5 ppm)

3.7 KHz (5.3 ppm)

2.5 KHz (5.8 ppm)

4 KHz (5.7 ppm)

4.3 KHz (10 ppm)

9.5 KHz (13.6 ppm)

4.2 KHz (9.7 ppm)

9 KHz (13 ppm)

**7.2 Gas sensitivity comparison of RSAW/STW sensor resonators coated with solid** 

in Fig. 8 and Table 2 the following important practical conclusions can be drawn:

*surface sorption of the HMDSO which is typical for solid sensing polymers*.

**6500 ppm**

2 KHz (2.9 ppm)

1.5 KHz (3.5 ppm)

3 KHz (4.3 ppm)

1.8 KHz (4.2 ppm)

8.3 KHz (11.9 ppm)

> 3.5 KHz (8 ppm)

11 KHz (15.7 ppm)

1.8 KHz (4.2 ppm)

11 KHz (16 ppm)

**Sensor/Compound Dichloroethane**

**HMDSO films** 

700 MHz STW 50 nm HMDSO

433 MHz RSAW 50 nm HMDSO

**Sensitivity factor** (STW/ RSAW)

700 MHz STW 190 nm HMDSO

433 MHz SAW 190 nm HMDSO

**Sensitivity factor** (STW/RSAW)

700 MHz STW 280 nm HMDSO

433 MHz SAW 280 nm HMDSO

**Sensitivity factor** (НПАВ/RSAW)

700 MHz STW 350 nm HMDSO

433 MHz SAW 350 nm HMDSO

**Sensitivity factor** (STW/RSAW)

700 MHz STW 100 nm HMDSO

*Δf=*160KHz over time indicating that prolonged xylene treatment has not had any influence on the sensor performance. This is a clear indication that styrene has very good physical sorption properties to a variety of gas phase compounds.

Fig. 6. Block diagram of the automated system for simultaneous gas sensitivity measurements on eight sensor devices

Fig. 7. Gas sensing performance of a 700 MHz styrene coated STW sensor probed with dichloroethane at 6500 ppm concentration in 100 s probe-flush cycles

*Δf=*160KHz over time indicating that prolonged xylene treatment has not had any influence on the sensor performance. This is a clear indication that styrene has very good physical

Fig. 6. Block diagram of the automated system for simultaneous gas sensitivity

0 100 200 300 400 500 600 700 800 900

Time (s)

Fig. 7. Gas sensing performance of a 700 MHz styrene coated STW sensor probed with

dichloroethane at 6500 ppm concentration in 100 s probe-flush cycles

 STW Sensor J7 after Xylene treatment for 62 h. 40 min. Di-chloro ethane probing at 6500 ppm concentration

measurements on eight sensor devices

0 20k 40k 60k 80k 100k 120k 140k 160k

Frequency shift with probing (Hz)

sorption properties to a variety of gas phase compounds.

### **7.2 Gas sensitivity comparison of RSAW/STW sensor resonators coated with solid HMDSO films**

This study aims at finding out which of both acoustic wave modes provides better gas sensitivity when coated with solid HMDSO films and what is the optimum film thickness at which maximum sensitivity is achieved. For this purpose, 5 pairs of RSAW/STW devices according to Table 1 were coated at 5 different HMDSO thicknesses (50, 100, 190, 280 and 350 nm) each, in the same plasma deposition process for each pair. Figure 8 compares the gas sensing characteristics of both modes gas probed with tetrachloroethilene at 2650 ppm concentration. The results from all gas probing experiments on the 5 pairs of devices are summarized in Table 2. In this table the "sensitivity factor" is the ratio between the relative sensitivities (in ppm) for the two devices of each pair. It is given for each of the 5 film thicknesses and shows which mode is more sensitive and at which thickness. From the data in Fig. 8 and Table 2 the following important practical conclusions can be drawn:

1. *The HMDSO coated sensors have very short response times and reach adsorption-desorbtion equilibrium just a few seconds after the gas flow is applied. We attribute this behavior to the surface sorption of the HMDSO which is typical for solid sensing polymers*.


Table 2. Gas sensitivity comparison of RSAW vs. STW devices coated with solid HMDSO film of 5 film thicknesses. Data on the 100 nm coated SAW device are not available

Polymer Coated Rayleigh SAW and STW Resonators for Gas Sensor Applications 533

A similar comparative study was performed also when pairs of RSAW/STW devices from Table 1 were coated at three different thicknesses of the semisolid ST and AA polymer films in a glow discharge plasma reactor. Since no equipment was available to measure the thickness of semisolid layers directly we used the deposition time in seconds for each layer as a measure of the layer thickness. In this study the deposition times for all three thicknesses were 10, 15 and 20 s for the films with the lowest, medium and highest thicknesses, accordingly. The results from these gas probing tests are summarized in Tables 4 and 5 for the ST and AA coated sensors, accordingly. Note that in these experiments we had to reduce the concentrations of all four probing gases by a factor of 4 to avoid saturation of the sensing layers due to the much higher adsorption capacity of the ST and AA films compared to HDMSO. When comparing the data from Tables 4 and 5 an interesting behavior is observed. Styrene coated STW devices are up to 3 times more sensitive than their ST coated RSAW counterparts while with the AA coated sensors we see the opposite behavior - the AA coated RSAW devices are up to 3,6 times more sensitive than their AA coated STW counterparts. We attribute this behavior to the fact that AA is the softest of the three polymers we used in this work and this material behaves much more like a soft polymer than a semisolid one. This implies that the RSAW mode might be more suitable for

> **Dichloroethane (1550 ppm)**

> > 92 ppm

> > 171 ppm

> > 201 ppm

*STW/RSAW 1.09 / 1.2 / 1.09* 

**Dichloroethane (1550 ppm)** 

> *RSAW/STW 2.91 / 3.18*

66.4 ppm

83.1 ppm

Acoustic mode STW RSAW STW RSAW STW RSAW STW RSAW

22.8 ppm

23.1 ppm

Table 5. Gas sensitivity comparison of semisolid AA coated RSAW and STW devices

Acoustic mode STW RSAW STW RSAW STW RSAW STW RSAW

100 ppm

206 ppm

219 ppm

Table 4. Gas sensitivity comparison of semisolid ST coated RSAW and STW devices

**Ethylacetate (4190 ppm)** 

> 76 ppm

> 95 ppm

> 122 ppm

*STW/RSAW 1.24 / 2.0 / 1.69*

> **Ethylacetate (4190 ppm)**

*RSAW/STW 2.14 / 2.87*

69.9 ppm

90.5 ppm

32.6 ppm

31.5 ppm

94 ppm

190 ppm

206 ppm

**Xylene (330 ppm)** 

> 44 ppm

> 65 ppm

> 76 ppm

*STW/RSAW 1.61 / 2.15 / 1.46* 

> **Xylene (330 ppm)**

*RSAW/STW 3.55 / 3.42*

25.2 ppm

28.4 ppm

7.1 ppm

8.3 ppm

71 ppm

140 ppm

111 ppm

**7.3 Gas sensitivity comparison of RSAW/STW sensor resonators coated with** 

**semisolid styrene (ST) and allylalcohol (AA) films** 

operation with soft sensing films than the STW mode.

119 ppm

254 ppm

239 ppm

12.3 ppm

14.3 ppm

**Tetrachloroethylene (630 ppm)** 

> *STW/RSAW 1.61 / 3.0 / 2.21*

**Tetrachloroethylene (630 ppm)** 

> *RSAW/STW 1.69 / 1.81*

74 ppm

85 ppm

108 ppm

> 20.8 ppm

> 25.9 ppm

**Compound/ Concentration** 

Deposition time 10 s/styrene

Deposition time 15 s/styrene

Deposition time 20 s/styrene

*Sensitivity factor 10s / 15s / 20s*

**Compound/ Concentration** 

Deposition time 20 s/AA

Deposition time 25 s/AA

*Sensitivity factor* 

*20s / 25s*


Table 3. Gas sensitivity comparison of the RSAW and STW sensors coated at their optimum HMDSO thickness values

Fig. 8. Tetrachloroethilene probing data of HMDSO coated a) STW and b) RSAW sensors at 50, 100, 190, 280 and 350 nm film thicknesses


**Ethylacetate 17600 ppm** 

> 20 KHz (29 ppm)

> 6.5 KHz (15 ppm)

*STW/RSAW 1.93* 

> 0 1k 2k 3k 4k 5k 6k 7k

Frequency shift with probing (Hz)

Fig. 8. Tetrachloroethilene probing data of HMDSO coated a) STW and b) RSAW sensors at

2. *Starting from very thin films (50 nm in this case) and increasing the thickness, the gas sensitivity increases in both the RSAW and STW devices, accordingly, until a thickness value is reached at which maximum sensitivity is achieved. This optimum thickness value is different for both modes (100 nm for the STW and 280 nm for the RSAW mode at the wavelength of 7,22* 

*in this case). Further increase in film thickness beyond the optimum thickness value only reduces* 

3. *The optimum thickness values for both modes are far below critical thickness and are well within the thickness ranges in which the sensor devices demonstrate high mass sensitivity while retaining low insertion loss, high Q and a well behaved single-mode resonance in the Parylene C coating experiment from Fig. 5. Therefore, the practical film thickness optimization method described in Section 6 is well suited for identifying the optimum film thickness at which* 

4. *The relative sensitivities for both acoustic wave modes at their optimum thickness values, summarized in Table 3, demonstrate a 1,3 to 3,8 times higher sensitivity to all 4 gases of the STW mode versus its RSAW counterpart operating at the same acoustic wavelength. This suggests that the STW mode is much more appropriate for operation with solid chemo* 

*the relative frequency sensitivity and increases the loss of the gas sensor*.

Table 3. Gas sensitivity comparison of the RSAW and STW sensors coated at their optimum

 A: 50 nm B: 190 nm C: 280 nm D: 350 nm E: 100 nm **Tetrachloroethylene 2650 ppm** 

> 37 KHz (53 ppm)

> 6 KHz (14 ppm)

*STW/RSAW 3.79* 

HMDSO coated 433 MHz SAW resonators

0 50 100 150 200 250 300 350 400 450 500 550 600 650

Time (s) b)

Tetra-chloro ethylene probing at 2650 ppm concentration A: 190 nm

**Xylene 1400 ppm** 

9 KHz (13 ppm)

4.3 KHz (10 ppm)

*STW/RSAW 1.3* 

> B: 50 nm C: 280 nm D: 350 nm E: uncoated

> > μ*m*

**Dichloroethane 6500 ppm** 

> 11 KHz (16 ppm)

3.5 KHz (8 ppm)

*STW/RSAW 2.0* 

0 100 200 300 400 500 600 700 800

*maximum gas sensitivity should be expected*.

*sensitive polymers*.

Time (s) a)

50, 100, 190, 280 and 350 nm film thicknesses

**Compound Concentration** 

700 MHz STW, 100 nm HMDSO

433 MHz RSAW, 280 nm HMDSO

> *Sensitivity factor*

0 5k 10k 15k 20k 25k 30k 35k 40k

Frequency shift with probing (Hz)

HMDSO thickness values

HMDSO coated 700 MHz STW resonators Tetra-chloro ethylene probing at 2650 ppm concentration

### **7.3 Gas sensitivity comparison of RSAW/STW sensor resonators coated with semisolid styrene (ST) and allylalcohol (AA) films**

A similar comparative study was performed also when pairs of RSAW/STW devices from Table 1 were coated at three different thicknesses of the semisolid ST and AA polymer films in a glow discharge plasma reactor. Since no equipment was available to measure the thickness of semisolid layers directly we used the deposition time in seconds for each layer as a measure of the layer thickness. In this study the deposition times for all three thicknesses were 10, 15 and 20 s for the films with the lowest, medium and highest thicknesses, accordingly. The results from these gas probing tests are summarized in Tables 4 and 5 for the ST and AA coated sensors, accordingly. Note that in these experiments we had to reduce the concentrations of all four probing gases by a factor of 4 to avoid saturation of the sensing layers due to the much higher adsorption capacity of the ST and AA films compared to HDMSO. When comparing the data from Tables 4 and 5 an interesting behavior is observed. Styrene coated STW devices are up to 3 times more sensitive than their ST coated RSAW counterparts while with the AA coated sensors we see the opposite behavior - the AA coated RSAW devices are up to 3,6 times more sensitive than their AA coated STW counterparts. We attribute this behavior to the fact that AA is the softest of the three polymers we used in this work and this material behaves much more like a soft polymer than a semisolid one. This implies that the RSAW mode might be more suitable for operation with soft sensing films than the STW mode.




Table 5. Gas sensitivity comparison of semisolid AA coated RSAW and STW devices

Polymer Coated Rayleigh SAW and STW Resonators for Gas Sensor Applications 535

**Di-chloro ethane (1550 ppm)** 

144 KHz (206 ppm)

Resolution *R* 7,5 ppb 22,9 ppb 67 ppb 7,2 ppb

(STW/RSAW)-1 4,3 1,9 2,9 3,3 Table 7. Detection limits of the styrene coated STW sensors at the 4 gas-phase analytes

> σ*y(*τ

**8.1 External factors that may degrade measurement resolution in practical** 

*)* limit, the better it has been designed.

Table 7 this resolution improvement is by a factor of 1,9 to 4,3. We attribute this advantage

• *The styrene coated STW sensors feature lower flicker noise values than their RSAW* 

• *The STW mode features better relative gas sensitivity than its RSAW counterpart at the same acoustic wave length and type of semisolid polymer film (styrene) as evident from Table 4.* 

The data in Tables 6 and 7 represent the physical detection limits that can be achieved with practical RSAW/STW based sensor systems. If all other factors that may have a negative effect on the measurements are excluded then the only limiting quantity to the measurement resolution remains the electrical flicker phase noise of the sensor oscillator represented by its

factors that may seriously degrade sensor resolution and therefore care should be taken to eliminate them or to reduce their influence to acceptable values. The closer the system is

The major factors that may degrade sensor resolution in practical sensor systems will be

This is one of the key disturbances that may seriously degrade sensor noise and should be eliminated first. If the gas flow applied to the sensor head from Fig. 6 is not homogeneous then the sensor devices will sense a variable gas concentration during the probe-flush cycles. Since the measurement cycle is much longer than the time over which inhomogeneities occur, amplitude fluctuation of the sensor signal at or close to equilibrium will occur. This situation is illustrated in Fig. 9 showing results from a tetrachloroethilene measurement with the setup from Fig. 6 when the integrators are removed. This causes a serious turbulence of the gas flow in the system which results in strong noise levels on top of the sensor signals. Even a periodicity on the noise signal is observed which is attributed to the pump rotation. Fortunately, gas flow

*).* Unfortunately, in practical sensor systems there are several external

*RSAW and STW sensors, respectively). This suggests that the STW devices tolerate the styrene* 

**Ethyl acetate (4190 ppm)** 

> 133 KHz (190 ppm)

*) measurement (5,6x10-9/s vs. 3,04x10-9/s for the* 

**Xylene (330 ppm)** 

98 KHz (140 ppm)

**Tetra-chloro ethylene (630 ppm)** 

Acoustic mode/*fo* STW/700 MHz Dep. time/polymer 15 s/styrene

(254 ppm)

*y(1s)* 3,04×10-9/s

**Compound/ Concentration** 

Resolution factor

σ

Sensor signal *Δf* 178 KHz

of the STW mode to the following factors:

*film better than the RSAW ones;* 

**RSAW/STW sensor systems** 

σ*y(*τ

briefly discussed next. a. *Gas flow homogeneity*.

σ*y(*τ

Allan's variation

brought to its

*counterparts. This is evident from the* 

For the semisolid film coated RSAW/STW sensors we can make the following conclusions:


### **8. Noise and measurement resolution (detection limit) of RSAW/STW resonant sensors operating with gas sensing polymer layers**

As shown in Section 3, if the sensor resonator is connected in the feedback loop of a sensor oscillator whose short-term stability over the time *τ* has been measured as σ*y(*τ*)*, then measuring the sensor signal *Δf* which is the response of the sensor oscillator to the gas with the concentration *C* , the measurement resolution *R*, also called detection limit of the sensor system, can readily be calculated with Equation (1). As an example, let us determine the measurement resolution of the sensor system with which the dichloroethane measurement at *C*=6500 ppm (6,5x10-3) from Fig. 7 was performed. With σ*y(*τ*)* measured as 1,17x10-9/s at the oscillator frequency *f0*=700 MHz (7x108 Hz), for the sensor resolution over the measurement time *τ=*1 s we obtain *R*=33,3 ppb (parts per billion). This means that this sensor system can detect changes in the dichloroethane concentration as small as 33 ppb. Such high sensor resolutions are extremely difficult to achieve with other sensor technologies. They are attributed to the fact that RSAW/STW resonant sensors retain excellent resonance characteristics, low loss, high Q and low flicker phase noise when coated with solid and semisolid chemo sensitive polymer films at optimum thickness.

The data in Tables 6 and 7 represent the detection limits of the styrene coated RSAW and STW sensors, respectively, during the measurements on the 4 gas-phase analytes used in this work. The σ*y(*τ*)* values for the sensor oscillators were measured as 5,6x10-9/s and 3,04x10-9/s, respectively. The "Resolution factor" in Table 7 indicates the resolution improvement of the STW mode versus its RSAW counterpart in these measurements. In


Table 6. Detection limits of the styrene coated RSAW sensors at the 4 gas-phase analytes


Table 7. Detection limits of the styrene coated STW sensors at the 4 gas-phase analytes

Table 7 this resolution improvement is by a factor of 1,9 to 4,3. We attribute this advantage of the STW mode to the following factors:


### **8.1 External factors that may degrade measurement resolution in practical RSAW/STW sensor systems**

The data in Tables 6 and 7 represent the physical detection limits that can be achieved with practical RSAW/STW based sensor systems. If all other factors that may have a negative effect on the measurements are excluded then the only limiting quantity to the measurement resolution remains the electrical flicker phase noise of the sensor oscillator represented by its Allan's variation σ*y(*τ*).* Unfortunately, in practical sensor systems there are several external factors that may seriously degrade sensor resolution and therefore care should be taken to eliminate them or to reduce their influence to acceptable values. The closer the system is brought to its σ*y(*τ*)* limit, the better it has been designed.

The major factors that may degrade sensor resolution in practical sensor systems will be briefly discussed next.

a. *Gas flow homogeneity*.

534 Acoustic Waves – From Microdevices to Helioseismology

For the semisolid film coated RSAW/STW sensors we can make the following conclusions: 1*. Semisolid sensing films improve gas sensitivity of the RSAW and STW modes dramatically compared to the solid films. ST coated devices demonstrate one to two orders of magnitude* 

2*. As observed with the solid film, also semisolid layers seem to have an optimum film thickness at which maximum gas sensitivity is achieved. Here these optimum thicknesses are achieved at 15 s for ST and 20 s for AA with the STW mode, while the RSAW mode needs somewhat higher* 

3*. There is a significant film type dependent difference in gas sensitivities for both modes. ST provides much better gas sensitivity compared to AA and this is attributed to both – the different sorption properties and different viscoelastic properties of the films which determine how the* 

As shown in Section 3, if the sensor resonator is connected in the feedback loop of a sensor

measuring the sensor signal *Δf* which is the response of the sensor oscillator to the gas with the concentration *C* , the measurement resolution *R*, also called detection limit of the sensor system, can readily be calculated with Equation (1). As an example, let us determine the measurement resolution of the sensor system with which the dichloroethane measurement

the oscillator frequency *f0*=700 MHz (7x108 Hz), for the sensor resolution over the measurement time *τ=*1 s we obtain *R*=33,3 ppb (parts per billion). This means that this sensor system can detect changes in the dichloroethane concentration as small as 33 ppb. Such high sensor resolutions are extremely difficult to achieve with other sensor technologies. They are attributed to the fact that RSAW/STW resonant sensors retain excellent resonance characteristics, low loss, high Q and low flicker phase noise when coated

The data in Tables 6 and 7 represent the detection limits of the styrene coated RSAW and STW sensors, respectively, during the measurements on the 4 gas-phase analytes used in

3,04x10-9/s, respectively. The "Resolution factor" in Table 7 indicates the resolution improvement of the STW mode versus its RSAW counterpart in these measurements. In

Resolution *R* 32,5 ppb 43,2 ppb 192 ppb 24 ppb Table 6. Detection limits of the styrene coated RSAW sensors at the 4 gas-phase analytes

σ*y(*τ

*)* values for the sensor oscillators were measured as 5,6x10-9/s and

**Ethylacetate (4190 ppm)** 

53 KHz (122 ppm)

**Dichloroethane (1550 ppm)** 

> 87 KHz (201 ppm)

σ*y(*τ*)*, then

**Xylene (330 ppm)** 

33 KHz (76 ppm)

*)* measured as 1,17x10-9/s at

4*. The RSAW mode operates better with the relatively soft AA than with the semisolid ST.* 

**8. Noise and measurement resolution (detection limit) of RSAW/STW** 

oscillator whose short-term stability over the time *τ* has been measured as

with solid and semisolid chemo sensitive polymer films at optimum thickness.

**Tetrachloroethylene (630 ppm)** 

Acoustic mode/*fo* RSAW/433 MHz Dep. time/polymer 20 s/styrene

(108 ppm)

*y(1s)* 5,6×10-9/s

**resonant sensors operating with gas sensing polymer layers** 

*higher relative gas sensitivities compared to HMDSO coated ones.* 

*optimum thicknesses – 20 s for ST and 25 s for the AA film.* 

*wave interacts with the film and the gas sorbed in it.* 

at *C*=6500 ppm (6,5x10-3) from Fig. 7 was performed. With

this work. The

σ

**Compound/ Concentration** 

σ*y(*τ

Sensor signal *Δf* 47 KHz

This is one of the key disturbances that may seriously degrade sensor noise and should be eliminated first. If the gas flow applied to the sensor head from Fig. 6 is not homogeneous then the sensor devices will sense a variable gas concentration during the probe-flush cycles. Since the measurement cycle is much longer than the time over which inhomogeneities occur, amplitude fluctuation of the sensor signal at or close to equilibrium will occur. This situation is illustrated in Fig. 9 showing results from a tetrachloroethilene measurement with the setup from Fig. 6 when the integrators are removed. This causes a serious turbulence of the gas flow in the system which results in strong noise levels on top of the sensor signals. Even a periodicity on the noise signal is observed which is attributed to the pump rotation. Fortunately, gas flow

Polymer Coated Rayleigh SAW and STW Resonators for Gas Sensor Applications 537

c. *Adsorption-desorbtion noise (ADN) in soft film coated RSAW sensors operated far below gas* 

but free of ADN could readily be obtained with ST coated sensors.

When RSAW sensors are coated with soft polymer films featuring profound bulk sorption these films can accommodate large amounts of gas without being driven into saturation. The larger the amount of adsorbed gas, the more turbulent the adsorptiondesorbtion process at equilibrium becomes even though the film is operated far below its saturation limit. This results in ADN evident in Fig. 11 which presents results from tetrachloroethilene probing on RSAW/STW sensors coated with the soft PIB film. ADN is visible on top of all sensor signals regardless of how strong they are. It should be noted that ADN levels depend entirely on the sorption characteristics of the soft polymer films and this makes elimination of this type of noise very difficult. In practical sensor systems one should either cope with ADN or, in critical situations, a different type of polymer film with lower ADN should be used. For example, the magnitude of the sensor signals in Fig. 11 is comparable with those in Fig. 7 where ST was used and no measurable ADN levels were observed. Therefore, results as the ones from the PIB film coated sensors from Fig. 11

0 5 10 15 20 25 30 35 40

**Time (min)**

Fig. 11. Tetrachloroethilene probing on RSAW/STW sensors coated with the soft PIB

**9. Corrosion proof RSAW resonant sensors using gold electrode structure**  Typically, RSAW/STW sensors are fabricated with Al electrode structure using a well established photolithographic process. Al metallization is cheap and provides excellent electrical performance in almost all SAW devices fabricated to date. However, a major problem occurs if SAW devices with Al metallization are coated with sensing layers and used as gas sensors. Very often, the chemical gas-phase compounds to be detected form corrosive bases and alkalis with the ambient humidity and attack the Al electrode structure

**Tetra-chloro ethylene probing on PIB coated SAW and STW devices at 433 MHz**

**STW 4x STW 5x STW 6x STW 7x SAW 7x SAW 6x** SAW 5x SAW 4x

process returns back to normal (see Fig. 10 b)).

0 10k 20k 30k 40k 50k 60k 70k 80k 90k 100k 110k 120k 130k 140k 150k 160k 170k

polymer film and operated far below gas saturation

**Frequency shift with probing (Hz)**

*saturation.*

very turbulent adsorption-desorbtion process takes place, the films get lighter and heavier in a stochastic way and this generates noise on top of the sensor signals. Once the gas concentration is reduced, equilibrium occurs and the adsorption-desorbtion

inhomogeneities are easily eliminated. After the integrators are placed back at the air inlet and outlet a noise free measurement similar to the one from Fig. 7 is obtained.

Fig. 9. Tetrachloroethilene probing with 5 HMDSO coated RSAW sensors with the integrators from the setup in Fig. 6 removed

b. *Gas saturation of the sensing films*.

Fig. 10. Ethylacetate probing on ST and AA coated STW sensors at a) 17600 and b) 4190 ppm vapour concentration

Saturation of the sensing films occurs when gas concentrations become so high that sorption limit of the layer is reached. A situation like this during an ethylacetate measurement at 17600 ppm concentration is illustrated in Fig. 10 a). In this case strong peaks of overpressure in the analyte container as well as noise and distortion on the sensor signals are observed as a result of film saturation. When the gas concentration is reduced by a factor of 4 to 4190 ppm, the sensor signals become much more uniform and noise fluctuations are greatly reduced. The behaviour in Fig. 10 a) has the following explanation. When the films become saturated dynamic equilibrium is disturbed and a

40k **Tetra-chloro etylene probing on HMDSO coated**

inhomogeneities are easily eliminated. After the integrators are placed back at the air inlet and outlet a noise free measurement similar to the one from Fig. 7 is obtained.

**433 MHz SAW devices with both integrators removed**

**190 nm 50 nm 350 nm 280 nm uncoated**

0 5 10 15 20 25 3

**Time (min)** b)

**Ethyl acetate measurement at 4190 ppm concentration (701 MHz STW)**

**ST/10s ST/15s ST/20s ST/25s AA/10s AA/15s AA/20s AA/25s**

0 100 200 300 400 500 600 700

**Time (s)**

0 10k 20k 30k 40k 50k 60k 70k 80k 90k 100k 110k 120k 130k 140k 150k

**Frequency shift with probing (Hz)**

Fig. 10. Ethylacetate probing on ST and AA coated STW sensors at a) 17600 and b) 4190 ppm

Saturation of the sensing films occurs when gas concentrations become so high that sorption limit of the layer is reached. A situation like this during an ethylacetate measurement at 17600 ppm concentration is illustrated in Fig. 10 a). In this case strong peaks of overpressure in the analyte container as well as noise and distortion on the sensor signals are observed as a result of film saturation. When the gas concentration is reduced by a factor of 4 to 4190 ppm, the sensor signals become much more uniform and noise fluctuations are greatly reduced. The behaviour in Fig. 10 a) has the following explanation. When the films become saturated dynamic equilibrium is disturbed and a

Fig. 9. Tetrachloroethilene probing with 5 HMDSO coated RSAW sensors with the

 **ST/10s ST/15s ST/20s ST/25s AA/10s AA/15s AA/20s AA/25s**

0 5k 10k 15k 20k 25k 30k 35k

integrators from the setup in Fig. 6 removed

 **Ethyl acetate measurement at 17600 ppm concentration (701 MHz STW)**

0 5 10 15 20 25 30

**Time (min)** a)

b. *Gas saturation of the sensing films*.

0 50k 100k 150k 200k 250k 300k 350k

vapour concentration

**Frequency shift with probing (Hz)**

**Frequenz shift (Hz)**

very turbulent adsorption-desorbtion process takes place, the films get lighter and heavier in a stochastic way and this generates noise on top of the sensor signals. Once the gas concentration is reduced, equilibrium occurs and the adsorption-desorbtion process returns back to normal (see Fig. 10 b)).

c. *Adsorption-desorbtion noise (ADN) in soft film coated RSAW sensors operated far below gas saturation.*

When RSAW sensors are coated with soft polymer films featuring profound bulk sorption these films can accommodate large amounts of gas without being driven into saturation. The larger the amount of adsorbed gas, the more turbulent the adsorptiondesorbtion process at equilibrium becomes even though the film is operated far below its saturation limit. This results in ADN evident in Fig. 11 which presents results from tetrachloroethilene probing on RSAW/STW sensors coated with the soft PIB film. ADN is visible on top of all sensor signals regardless of how strong they are. It should be noted that ADN levels depend entirely on the sorption characteristics of the soft polymer films and this makes elimination of this type of noise very difficult. In practical sensor systems one should either cope with ADN or, in critical situations, a different type of polymer film with lower ADN should be used. For example, the magnitude of the sensor signals in Fig. 11 is comparable with those in Fig. 7 where ST was used and no measurable ADN levels were observed. Therefore, results as the ones from the PIB film coated sensors from Fig. 11 but free of ADN could readily be obtained with ST coated sensors.

Fig. 11. Tetrachloroethilene probing on RSAW/STW sensors coated with the soft PIB polymer film and operated far below gas saturation

### **9. Corrosion proof RSAW resonant sensors using gold electrode structure**

Typically, RSAW/STW sensors are fabricated with Al electrode structure using a well established photolithographic process. Al metallization is cheap and provides excellent electrical performance in almost all SAW devices fabricated to date. However, a major problem occurs if SAW devices with Al metallization are coated with sensing layers and used as gas sensors. Very often, the chemical gas-phase compounds to be detected form corrosive bases and alkalis with the ambient humidity and attack the Al electrode structure

Polymer Coated Rayleigh SAW and STW Resonators for Gas Sensor Applications 539

a) c)

b) d)

This section compares the new Au vs. the old Al sensor devices in their sensor characteristics to find out if the replacement causes any performance degradation of the sensor system the Au devices were expected to operate in. To check the sensor performance we again used two types of polymer films: (A) solid Parylene C to simulate coating behaviour with solid and semisolid films as described in Section 6 and (B) a soft polymer called poly[chlorotrifluoroethylene-co-vinylidene fluoride] (PCFV) to test sensor performance at high gas concentrations. Since Parylene C coating was discussed in Section 6 already, here we will briefly discuss a relatively novel soft polymer deposition method, which is called electro spray method and is described in [24] in detail. We applied it successfully to all 4 devices from Fig. 12 to obtain very uniform high-quality soft PCFV films for reproducible sensor performance. According to this method, the holder with the SAW devices mounted on it, spins in a cloud of very small liquid-phase polymer droplets coming out from a narrow capillary tube and directed by an electrostatic field towards the sensor surface. The droplets settle onto the device surface, stick together and form a uniform film. Its thickness depends on the deposition time. Since the SAW device loss increases with film thickness, a certain insertion loss value, as necessary for optimum sensor performance, can be obtained simply by adjusting the deposition time. Except for excellent control over film thickness and uniformity [24], major advantage of this polymer coating method is that, even

Fig. 12. Frequency (upper curves) and phase/group delay (lower curves) responses of RSAW single mode two-port resonators (upper row) and coupled resonator filter (lower

row) using (a) and (b) Au and (c) and (d) Al electrode structure

**9.2 Chemo sensitive polymer films and deposition methods used** 

**Al - CRF device** MKR( 250): 432.566MHz

**Al - TPR device** MKR( 250): 432.436MHz

MAGTD ( ) **-6.52dB** 5dB/ DLY ( ) **3.94us** 2us/

CF: 432.436MHz SPAN: 10MHz

MAGTD ( ) **-6.45dB** 5dB/ DLY ( ) **4.01us** 2us/

CF: 432.566MHz SPAN: 5MHz

**TPR\_Au device** MKR( 250): 437.3758MHz

**CRF\_Au device** MKR( 299): 410.45MHz

MAGTD ( ) **-7.56dB** 5dB/ PHASE ( ) -7.7deg 100deg/

CF: 437.3758MHz SPAN: 10MHz

MAGTD ( ) **-10.45dB** 5dB/ PHASE ( ) -45.0deg 100deg/

CF: 409.96MHz SPAN: 5MHz

by entering into a chemical reaction with the Al film. The problem is further aggravated by the presence of the sensing layer which, by absorbing large amounts of gas, greatly increases the concentration of the aggressive analyte that comes in contact with the Al electrodes. As a result of corrosion, the sensors suffer irreversible performance degradation, provide inconsistent data and even dye within a limited number of measurement cycles. The solution to that problem is the implementation of SAW devices with corrosion proof gold (Au) metallization that can successfully stand severe corrosion attacks by chemically reactive substances. The design of RSAW sensor resonators with Au electrode structure is not so straight forward as with Al metallization. Due to the fact that Au has a 7 times higher density than Al and is much softer several side effects, such as excitation of strong SSBW modes and transverse waveguide modes occur that can cause serious loss and Q-degradation as well as distorted characteristics at the main resonance. However, by careful selection of the Au thickness and choosing proper device geometry, these side effects can be kept under control and very good resonance characteristics appropriate for gas sensor applications can be achieved [22], [23]. In the next sections we will discuss the performance of RSAW sensors with Au electrode structure intended for operation as gas sensors in highly reactive chemical environments. These sensors were designed to replace their predecessors using the problematic Al electrode structure in a practical sensor system operating at 433 MHz.

### **9.1 Electrical performance comparison of Au vs. Al RSAW sensor resonators**

Generally two types of resonator devices are used in a practical resonator system – two-port resonators (TPR) featuring a single-mode resonance and coupled resonator filters (CRF) that have a two-pole resonance achieved with a coupling grating in the centre of the resonant cavity. As evident from Fig. 12 a) and b), the CRF devices have twice the phase slope of a TPR in their filter pass bands and generally provide better stabilization of the sensor oscillator than the TPR, especially in measurements at high gas concentrations. In a realworld sensor system, the sensor oscillator is designed to operate on the right CRF resonant mode since the left one vanishes when the device is coated with a polymer film [6]. Typically, the sensor oscillator provides stable oscillation on the right mode and never jumps onto the left one since a 180 deg. phase reversal makes oscillation impossible at that mode (see Fig. 12 b)). The frequency and group delay responses of the Al RSAW devices previously used in the sensor system are shown in Fig. 12 c) and d) while a) and b) represent the electrical performance of their Au substitutes designed in [23]. The insertion loss, Q and group delay data from these devices at resonance are compared in Table 8. Es evident from that data and also from Fig. 12 the Al and Au devices feature very similar frequency responses, insertion loss and loaded Q values and the replacement of the Al devices with their Au successors was made without any changes or adjustments of the sensor circuitry. The Au devices are slightly inferior to the Al ones in terms of loss and loaded Q. This is attributed to the loss of energy as a result of the heavy Au loading on the quartz surface.


Table 8. Comparison of the insertion loss, Q and group delay data at resonance of the uncoated RSAW sensors with Al and Au metallization characterized in Fig. 12

by entering into a chemical reaction with the Al film. The problem is further aggravated by the presence of the sensing layer which, by absorbing large amounts of gas, greatly increases the concentration of the aggressive analyte that comes in contact with the Al electrodes. As a result of corrosion, the sensors suffer irreversible performance degradation, provide inconsistent data and even dye within a limited number of measurement cycles. The solution to that problem is the implementation of SAW devices with corrosion proof gold (Au) metallization that can successfully stand severe corrosion attacks by chemically reactive substances. The design of RSAW sensor resonators with Au electrode structure is not so straight forward as with Al metallization. Due to the fact that Au has a 7 times higher density than Al and is much softer several side effects, such as excitation of strong SSBW modes and transverse waveguide modes occur that can cause serious loss and Q-degradation as well as distorted characteristics at the main resonance. However, by careful selection of the Au thickness and choosing proper device geometry, these side effects can be kept under control and very good resonance characteristics appropriate for gas sensor applications can be achieved [22], [23]. In the next sections we will discuss the performance of RSAW sensors with Au electrode structure intended for operation as gas sensors in highly reactive chemical environments. These sensors were designed to replace their predecessors using the

problematic Al electrode structure in a practical sensor system operating at 433 MHz.

**9.1 Electrical performance comparison of Au vs. Al RSAW sensor resonators** 

Generally two types of resonator devices are used in a practical resonator system – two-port resonators (TPR) featuring a single-mode resonance and coupled resonator filters (CRF) that have a two-pole resonance achieved with a coupling grating in the centre of the resonant cavity. As evident from Fig. 12 a) and b), the CRF devices have twice the phase slope of a TPR in their filter pass bands and generally provide better stabilization of the sensor oscillator than the TPR, especially in measurements at high gas concentrations. In a realworld sensor system, the sensor oscillator is designed to operate on the right CRF resonant mode since the left one vanishes when the device is coated with a polymer film [6]. Typically, the sensor oscillator provides stable oscillation on the right mode and never jumps onto the left one since a 180 deg. phase reversal makes oscillation impossible at that mode (see Fig. 12 b)). The frequency and group delay responses of the Al RSAW devices previously used in the sensor system are shown in Fig. 12 c) and d) while a) and b) represent the electrical performance of their Au substitutes designed in [23]. The insertion loss, Q and group delay data from these devices at resonance are compared in Table 8. Es evident from that data and also from Fig. 12 the Al and Au devices feature very similar frequency responses, insertion loss and loaded Q values and the replacement of the Al devices with their Au successors was made without any changes or adjustments of the sensor circuitry. The Au devices are slightly inferior to the Al ones in terms of loss and loaded Q. This is attributed to the loss of energy as a result of the heavy Au loading on the quartz surface.

**Parameter / Device (433 MHz) Al-CRF Au-CRF Al-TPR Au-TPR**  Unmatched insertion loss [dB] 6.5 10.5 6.5 7.5 Group delay (50Ω load) [μs] 4.01 3.44 3.94 2.83 Loaded Q 5450 4430 5350 3890 Unloaded Q 10400 6190 10160 6730 Table 8. Comparison of the insertion loss, Q and group delay data at resonance of the uncoated RSAW sensors with Al and Au metallization characterized in Fig. 12

Fig. 12. Frequency (upper curves) and phase/group delay (lower curves) responses of RSAW single mode two-port resonators (upper row) and coupled resonator filter (lower row) using (a) and (b) Au and (c) and (d) Al electrode structure

### **9.2 Chemo sensitive polymer films and deposition methods used**

This section compares the new Au vs. the old Al sensor devices in their sensor characteristics to find out if the replacement causes any performance degradation of the sensor system the Au devices were expected to operate in. To check the sensor performance we again used two types of polymer films: (A) solid Parylene C to simulate coating behaviour with solid and semisolid films as described in Section 6 and (B) a soft polymer called poly[chlorotrifluoroethylene-co-vinylidene fluoride] (PCFV) to test sensor performance at high gas concentrations. Since Parylene C coating was discussed in Section 6 already, here we will briefly discuss a relatively novel soft polymer deposition method, which is called electro spray method and is described in [24] in detail. We applied it successfully to all 4 devices from Fig. 12 to obtain very uniform high-quality soft PCFV films for reproducible sensor performance. According to this method, the holder with the SAW devices mounted on it, spins in a cloud of very small liquid-phase polymer droplets coming out from a narrow capillary tube and directed by an electrostatic field towards the sensor surface. The droplets settle onto the device surface, stick together and form a uniform film. Its thickness depends on the deposition time. Since the SAW device loss increases with film thickness, a certain insertion loss value, as necessary for optimum sensor performance, can be obtained simply by adjusting the deposition time. Except for excellent control over film thickness and uniformity [24], major advantage of this polymer coating method is that, even

Polymer Coated Rayleigh SAW and STW Resonators for Gas Sensor Applications 541

40% thicker solid films than their Al counterparts for the same amount of loss increase. Finally, Fig. 16 compares the frequency sensitivities of the four tested devices with Parylene C film thickness which is also an indication of their gas probing sensitivity with solid films. Up to about 300 nm thickness, the sensitivity slope of the devices is nearly identical with a small advantage of the Al TPR device, followed by the Al CRF. Above 300 nm, the sensitivity slope of the Al TPR device increases but in view of its strong loss degradation, its sensitivity advantage gets lost. The other three devices keep a nearly constant sensitivity slope up to 500 nm thickness. Below the practical 370 nm thickness at which the critical for this particular system 20 dB of loss is reached for the Al devices, the sensitivity of all four devices differs by less than 20%. This difference is insignificant for practical sensor systems.

a)

b) Fig. 14. Parylene C coating behaviour of the RSAW CRF devices from Fig. 12 using (a) Au

and (b) Al metallization

at 433 MHz, the droplets are much smaller than the acoustic wavelength of about 7 μm at this frequency. Because of the small droplet size, the electro spray films cause much less propagation loss for the SAW, compared to films of the same type and thickness, deposited in an older airbrush coating technique. In contrast to electro spray films, airbrush coatings have a rough textured structure and the droplet size typically exceeds the acoustic wavelength. The two optical microscope pictures in Fig. 13 a) and b) compare the film structures of PCFV deposited with the two methods on identical 433 MHz RSAW devices. Fig. 13 a) shows part of the electrode structure and bus bars at the centre of the electro spray coated SAW device. For better visibility of the textured film structure obtained in an airbrush technique, the picture in Fig. 13 b) has been taken with a slightly higher magnification and shows part of the reflector with some free substrate area on which the large drops are clearly visible.

Fig. 13. Comparison of two identical RSAW devices PCFV coated with (a) the electrospray and (b) the airbrush method

### **9.3 Polymer coating behaviour of Au vs. Al sensor resonators**

The three dimensional data plots in Fig. 14 compare the Parylene C coating behaviour of the Au and Al CRF devices from Fig. 12 b) and d) accordingly with a film thickness varying from 0 to 700 nm. In both cases the magnitude of the left longitudinal mode decreases with thickness until it disappears completely at about 300 nm for the Au and 450 nm for the Al device. Above these thickness ranges both devices demonstrate a smooth and well behaved single-mode resonance. In the 300 to 500 nm range the Au CRF reaches its highest mass sensitivity, accompanied with a gradual increase in insertion loss while the loss of the Al device decreases more rapidly. The critical for the system operation loss value of 20 dB is reached at 370 nm vs. 450 nm for the Au and Al devices, respectively.

An identical comparative Parylene C coating test, (not shown here), was performed also with the Au vs. Al TPR devices from Fig. 12 a) and c). In these tests again the Au devices were found to tolerate the solid Parylene C better than their Al counterparts. The increase in device insertion loss with Parylene C thickness for all four tested devices from Fig. 12 is shown in Fig. 15. At film thicknesses up to about 180 nm, all devices yield the same loss increase. Above 200 nm the loss behaviour starts to diverge. The two Al devices keep the same insertion loss up to about 550 nm thickness. Above 200 nm thickness, the loss of the Au devices increases at a much lower rate indicating that these devices can tolerate up to

at 433 MHz, the droplets are much smaller than the acoustic wavelength of about 7 μm at this frequency. Because of the small droplet size, the electro spray films cause much less propagation loss for the SAW, compared to films of the same type and thickness, deposited in an older airbrush coating technique. In contrast to electro spray films, airbrush coatings have a rough textured structure and the droplet size typically exceeds the acoustic wavelength. The two optical microscope pictures in Fig. 13 a) and b) compare the film structures of PCFV deposited with the two methods on identical 433 MHz RSAW devices. Fig. 13 a) shows part of the electrode structure and bus bars at the centre of the electro spray coated SAW device. For better visibility of the textured film structure obtained in an airbrush technique, the picture in Fig. 13 b) has been taken with a slightly higher magnification and shows part of the reflector with some free substrate area on which the

a) b) Fig. 13. Comparison of two identical RSAW devices PCFV coated with (a) the electrospray

The three dimensional data plots in Fig. 14 compare the Parylene C coating behaviour of the Au and Al CRF devices from Fig. 12 b) and d) accordingly with a film thickness varying from 0 to 700 nm. In both cases the magnitude of the left longitudinal mode decreases with thickness until it disappears completely at about 300 nm for the Au and 450 nm for the Al device. Above these thickness ranges both devices demonstrate a smooth and well behaved single-mode resonance. In the 300 to 500 nm range the Au CRF reaches its highest mass sensitivity, accompanied with a gradual increase in insertion loss while the loss of the Al device decreases more rapidly. The critical for the system operation loss value of 20 dB is

An identical comparative Parylene C coating test, (not shown here), was performed also with the Au vs. Al TPR devices from Fig. 12 a) and c). In these tests again the Au devices were found to tolerate the solid Parylene C better than their Al counterparts. The increase in device insertion loss with Parylene C thickness for all four tested devices from Fig. 12 is shown in Fig. 15. At film thicknesses up to about 180 nm, all devices yield the same loss increase. Above 200 nm the loss behaviour starts to diverge. The two Al devices keep the same insertion loss up to about 550 nm thickness. Above 200 nm thickness, the loss of the Au devices increases at a much lower rate indicating that these devices can tolerate up to

**9.3 Polymer coating behaviour of Au vs. Al sensor resonators** 

reached at 370 nm vs. 450 nm for the Au and Al devices, respectively.

large drops are clearly visible.

and (b) the airbrush method

40% thicker solid films than their Al counterparts for the same amount of loss increase. Finally, Fig. 16 compares the frequency sensitivities of the four tested devices with Parylene C film thickness which is also an indication of their gas probing sensitivity with solid films. Up to about 300 nm thickness, the sensitivity slope of the devices is nearly identical with a small advantage of the Al TPR device, followed by the Al CRF. Above 300 nm, the sensitivity slope of the Al TPR device increases but in view of its strong loss degradation, its sensitivity advantage gets lost. The other three devices keep a nearly constant sensitivity slope up to 500 nm thickness. Below the practical 370 nm thickness at which the critical for this particular system 20 dB of loss is reached for the Al devices, the sensitivity of all four devices differs by less than 20%. This difference is insignificant for practical sensor systems.

Fig. 14. Parylene C coating behaviour of the RSAW CRF devices from Fig. 12 using (a) Au and (b) Al metallization

Polymer Coated Rayleigh SAW and STW Resonators for Gas Sensor Applications 543

better the polymer film which is much lighter than Au. Finally, Fig. 19 compares the frequency sensitivity of both devices to increased PCFV thickness. The curve for the Au

Fig. 17. Frequency responses (upper plots) and phase responses (lower plots) of an Au TPR device before (data on the right) and after (data on the left) 7.5 minutes of PCFV deposition

> **Au TPR data points Linear fit of the Au TPR data Al CRF data points Exponential fit of the Al CRF data**

0 2 4 6 8 10 12 14 16 18

**Coating time in minutes**

Fig. 18. Loss increase vs. deposition time for Au and Al devices electro spray coated with

using the electro spray method

soft PCFV film

0

5

10

**Loss increase in dB**

15

20

device is steeper meaning that its mass sensitivity is higher than its Al counterpart.

Fig. 15. Insertion loss behaviour of the Au and Al devices from Fig. 12 vs. Parylene C thickness

Fig. 16. Frequency (mass) sensitivity behaviour of the Au and Al devices from Fig. 12 vs. Parylene C thickness

The soft PCFV polymer coating experiments were performed on the two TPR devices from Fig. 12 a) and c) since they have about the same amount of loss prior to coating, as required by the sensor system. Since the devices are mounted on a spinning holder, monitoring of their electrical performance in the process of electro spray deposition is not possible. That is why we recorded the frequency and phase responses of each device prior to and after the deposition to obtain the frequency shift and loss increase as a function of the deposition time which we used as a measure for the thickness of the soft PCFV film. The data plots in Fig. 17 illustrate the PCFV coating behaviour of an Au TPR device in a 7,5 minutes deposition time. As a result of film loading the device loss increases by about 8 dB to 17,9 dB while its frequency shifts down by 3 MHz. The coated device on the left retains a well behaved single-mode resonance with a smooth phase response in the resonance region. The loss increase vs. thickness proportional deposition time for the Al and Au devices is shown in Fig. 18. For the Au device this dependence is linear while the Al device shows a small loss increase up to about 10 min. of deposition time and after that its loss degrades very rapidly. We attribute this behaviour again to the difference in Au vs. Al densities. Once the Au device has been optimized for operation under the heavy Au film load, it tolerates much

**TPR\_Au CRF\_Al CRF\_Au TPR\_Al**

0 50 100 150 200 250 300 350 400 450 500 550 600

**Parylene thickness [nm]**

0 50 100 150 200 250 300 350 400 450 500

**Parylene thickness [nm]**

Fig. 16. Frequency (mass) sensitivity behaviour of the Au and Al devices from Fig. 12 vs.

The soft PCFV polymer coating experiments were performed on the two TPR devices from Fig. 12 a) and c) since they have about the same amount of loss prior to coating, as required by the sensor system. Since the devices are mounted on a spinning holder, monitoring of their electrical performance in the process of electro spray deposition is not possible. That is why we recorded the frequency and phase responses of each device prior to and after the deposition to obtain the frequency shift and loss increase as a function of the deposition time which we used as a measure for the thickness of the soft PCFV film. The data plots in Fig. 17 illustrate the PCFV coating behaviour of an Au TPR device in a 7,5 minutes deposition time. As a result of film loading the device loss increases by about 8 dB to 17,9 dB while its frequency shifts down by 3 MHz. The coated device on the left retains a well behaved single-mode resonance with a smooth phase response in the resonance region. The loss increase vs. thickness proportional deposition time for the Al and Au devices is shown in Fig. 18. For the Au device this dependence is linear while the Al device shows a small loss increase up to about 10 min. of deposition time and after that its loss degrades very rapidly. We attribute this behaviour again to the difference in Au vs. Al densities. Once the Au device has been optimized for operation under the heavy Au film load, it tolerates much

Fig. 15. Insertion loss behaviour of the Au and Al devices from Fig. 12 vs. Parylene C

 **CRF\_Au TPR\_Au CRF\_Al TPR\_Al**


**Frequency shift [Hz]**

thickness

Parylene C thickness

**Increase in insertion loss [dB]**

better the polymer film which is much lighter than Au. Finally, Fig. 19 compares the frequency sensitivity of both devices to increased PCFV thickness. The curve for the Au device is steeper meaning that its mass sensitivity is higher than its Al counterpart.

Fig. 17. Frequency responses (upper plots) and phase responses (lower plots) of an Au TPR device before (data on the right) and after (data on the left) 7.5 minutes of PCFV deposition using the electro spray method

Fig. 18. Loss increase vs. deposition time for Au and Al devices electro spray coated with soft PCFV film

Polymer Coated Rayleigh SAW and STW Resonators for Gas Sensor Applications 545

down to a few ppb when coated with solid, semisolid and soft polymer sensing layers. Furthermore, the RSAW and STW modes do not only compete but rather complement each other in different measurement tasks. The STW mode operates better with solid and semisolid sensing films featuring surface sorption and is better suited for high-resolution measurements at low gas concentrations (<1%) while the RSAW mode tolerates much better thick soft sensing layers with profound bulk sorption that operate better at high gas concentrations (>1%). Carefully designed RSAW sensors with Au metallization provide an excellent corrosion proof substitute of their Al counterparts when operated in highly reactive gas-phase environments, thereby greatly increasing system reliability and measurement reproducibility over time and a large number of measurement cycles. All gas sensors, regardless of acoustic wave mode, design, metallization and type of sensing polymer requires a careful thickness optimization to provide highest gas sensitivity,

The author wishes to gratefully acknowledge Dr. E. Radeva from the Georgi Nadjakov Institute of Solid State Physics, Bulgarian Academy of Sciences in Sofia, Bulgaria for expert preparation of the HMDSO films used in this study as well Professor Shigeru Kurosawa and his research associates from the National Institute of Materials Chemistry in Tsukuba, Japan for the deposition of the semisolid ST and AA films. Special thanks are directed to Dr. Michael Rapp and his research team at the Research Centre Karlsruhe in Germany for the

[1] R. M. White, Acoustic sensors for physical, chemical and biochemical applications, *Proc. IEEE 1998 International Symposium on Frequency Control*, pp. 587-594. [2] R. M. White, Surface acoustic wave sensors, *Proc. IEEE 1985 Ultrasonics Symposium*, pp.

[3] H. Wohltjen, Mechanism of operation and design considerations for surface acoustic

[4] R. Chung, R. A. McGill and P. Matthews, "Phase noise characterization of polymer

[5] S. J. Martin, G. C. Frye, J. J. Spates, and M. A. Butler, Gas sensing with acoustic devices,

[6] M. Rapp, J. Reibel, S. Stier, A. Voigt, and J. Bahlo, SAGAS: Gas analyzing sensor systems

sensor technology, in *Proc. IEEE Int. Freq. Contr. Symp.*, 1997, pp. 129–132. [7] E. J. Staples, Dioxin/furan detection and analysis using a SAW based electronic nose,

[8] E. J. Staples, T. Matsuda, and S Viswanathan, Real Time Environmental Screening of Air,

coated SAW-gas sensors: Implications for the performance of an oscillator circuit",

based on surface acoustic wave devices—An issue of commercialization of SAW

Water and Soil Matrices Using a Novel Field Portable GC/SAW System, Environmental Strategies for for the 21st Century, *Asia Pacific Conference*, pp. 8-10

opportunity to perform a substantial part of this work at those laboratories.

wave device vapour sensors, *Sens. Act.*, vol. 5, p. 307, 1984.

*Proc. 1997 IEEE Int. Freq. Control Symp.,* pp. 169-174.

*Proc. IEEE 1996 Ultrasonics Symposium*, pp. 423-434.

*Proc. IEEE 1998 Ultrasonics Symposium*, pp. 521-524.

maximum dynamic range and lowest detection limit.

**11. Acknowledgments** 

**12. References** 

490-494.

April 1998.
