How does a piezoelectric drop generator work?

In a piezoelectric drop generator used for generating a single droplet which is 0.5 mm in diameter, there is a piezo-actuated brass diaphragm to generate droplets (for example here or here). What is the main factor that pushes the water inside the generator in order to cause a droplet to be ejected from nozzle? (whether the main factor is diaphragm bending OR it is acoustic wave that piezo element may produce?)

Note: I use a piezo element which has a brass plate with 27 mm diameter and 0.3 mm thickness. However, when I installed in its place (Fig. 1) and fill the beneath reservoir with water (without any air bubble), I am not sure whether the brass diaphragm should bend rapidly downward to push a droplet out of the nozzle OR it should produce acoustic waves that push the water to eject a droplet. Which one is actually supposed to happen?)

Thanks

Figure 1. Schematic of a piezoelectric drop generator

• Dear Floris, 1- In the beginning of your answer,what do you mean by bending in "bending the diaphragm to generate acoustic waves?" 2- Based on your main paragraph, how are the pressure waves produced by piezoelectric diaphragm? (through bending or acoustic waves) 3- The first paper that I linked to in my question mentioned that each pulse that they sent to piezo had a 10 $\mu$s rise time and a 10 $\mu$s fall time and around 1500 $\mu$s pulse width. What do you think with this information regarding which one is the cause:"plate bending" or "acoustic waves"? Thanks for all your helpful comments – vorujak Nov 5 '14 at 1:12
• What I mean is - a sound wave is generated because "something" moves. In this case, the piezoelectric transducer's surface - what I call the diaphragm. If the pulse duration is 1500 $\mu s$ then this is a "plate bending" situation - the sound / pressure wave had plenty of time to bounce around and come to equilibrium. That explains why they get away with the strange shape of the interior cavity... – Floris Nov 5 '14 at 1:28