I don't have any recent direct application knowledge (my work in SAW is on the order of 25 years old) but there are certain fundamental concepts that hold, regardless. Mainly I'm thinking about the wavelength, which is larger at lower frequencies, meaning that the "skin depth" of the SAW is likely to be deeper as well, while the inter-atomic spacing of the crystal lattice remains the same. Thus, for the same physical (i.e. atomic) perturbation of the wave, such as a defect in the crystal lattice, I would expect less impact on a lower frequency SAW. However, you mention "if a sensing element meant to affect the acoustic wave...", which implies that the sensor likely scales with frequency, in which case the impact of the sensor likely scales as well. That's about as deep as I care to try to analyze, and someone will likely come along with better information, but hopefully that helps.

Hi Michael, thank you for your answer. Your analysis is helpful and is what I thought may be the case.

It is very critical for SAW to change frequency (from ready made component point of view), suppose that a SAW device has been designed for a certain frequency, main parameters (mainly insertion loss and cutt-off) changes dramatically with frequency.

Thanx@Michael.

Yasser makes a good point that I did not mention as I took it as assumed. Obviously a SAW device is narrow band. That's why they're used as filters, etc. Thus the physical spacing of the interdigitated electrodes, etc. controls the frequency response of the SAW device and anything outside that band will be severely attenuated (at least until you hit a harmonic). My assumption was that we were talking about designing a SAW device for a given frequency and what happens as the design frequency changes on the same substrate, etc.

Thank you Yasser you bring up a good point, as Michael said.

And you were right Michael, I am thinking of a SAW device designed for a given frequency, a delay line device. Between the delay lines is a material that acts as a sensor element, meant to influence the surface wave properties as it travels when exposed to what is meant to be measured. It could be mass loading, or what I am thinking of is change in Young's modulus of this sensing material.

If the sensing material size and thickness is fixed. But two devices are design for different frequencies a lower frequency 20 MHz and higher frequency 200 MHz lets say, wavelengths of 200 um and 20 um. There is a considerable difference in wavelength and penetration depth.

And my thinking that these differences is wavelength lead to more or less sensitivity.

For the change in Young's Modulus of the film this would be a change in acoustic velocity of the wave over this area leading to a change in the delay time. And if the lower frequency wave is much more confined to the substrate than the sensing layer the effect would be negligible.

Back to physics, just remember that (to the contrary of sound waves) surface waves are "dispersive" meaning their velocity is sharply frequency-dependant ! (so the shape of a multi-frequency pulse changes with distance because HF components travel faster). This can bias seriously any attempt of FFT analysis between the original pulse and the received one...

There are two parameters, which determine the response of the SAW sensor, which is constructed, using SAW delay line: central frequency and length of delay path. SAW delay line is the device which has two transducers and space between them. The space is an "active" part. The sensitive film can be deposited on it and the phase of propagated signal in such a system becomes the function of the density of the film. The first parameter of such a sensor is the central frequency of the SAW delay line. The sensitivity of the sensor depends on the central frequency of delay line as a square of this frequency. In other words the same chane in density of sensitive film in the delay line with central frequency 2F will cause the frequency shit 4 times bigger that in the delay line with frequency F. It is true for Raleigh type of waves always. It might be some specific changes for other types of wave, but basicaly: the bigger frequency - the bigger is sensitivity. Supprisinly the length of sensitive area does not improve sensitivity, but ist does improve the signal to noise ratio, which is a big deal in SAW sensors. Regards.

Thank you very much for this information Dr Balashov, particularly the bit about the length of the sensitive area. I had seen papers with 10-15 mm length sensitive areas and more recently 5 mm, and was curious whether this was a significant factor or not.

Hi, David. actually the question of the length is abit tricky. As I posted above the length gives you better signal/noise ratio, but if you start to increase it you will quickly have the situation when in your pass band there will be more than one points with the phase equal to zero (actually 2*pi*N). This will be very bad. Thus you will have to increase the length of your transducers to cut off some of these points simultaneoulsy. You can not increase the length of transducer too much as well. Thus if you want to analyze the concrete design you will have to take into account the length of transducers and active area simultaneously. It is better to think in terms of wavelengths, then in centimeters. Thus, for example,for quarts the total length of two transducers is about 100 - 120 lambda and the distance between them is 50 -80 lambda (lambda is the wave length of central frequency of your transducers). In this case the sensor will work. In this case the stability of generation will be about +/- 3*10^(-7). Regards.

Probably it is snot worth adding anything to this exchange but i found it by chance and I would simply add a comment. You need to clarify the term sensitivity. For instance, a large device will be much sensitive to acceleration than a small simply because of its mass... What kind of sensitivity is questioned here ? But as said above, my contribution comes probably too late. BR