The atomization procedure is the first step in the spray pyrolysis deposition system. The idea is to generate droplets from a spray solution and send them, with some initial velocity, towards the substrate surface. Spray pyrolysis normally uses air blast, ultrasonic, or electrostatic techniques . The atomizers differ in resulting droplet size, rate of atomization, and the initial velocity of the droplets. It has been shown that the size of the generated droplet is not related to any fluid property of the precursor solution and depends solely on the fluid charge density level as shown in 
The ultrasonic nozzles apply a relatively small amount of energy and can produce droplets with a radius down to approximately 2m, but at the cost of a low atomization rate (2cmmin) . Due to ease of production, many companies chose to use pressure atomizers instead of the ultrasonic atomizers. Therefore, this section will mainly concern itself with the pressure and electrostatic atomizers, characterized in , , respectively.
An air blast atomizer uses high speed air in order to generate an aerosol from a precursor solution. Increasing the air pressure causes a direct decrease in the generated mean droplet diameter. Inversely, increasing the liquid pressure causes a direct increase in the mean droplet diameter . Increasing the distance between the spray nozzle and the surface to be coated reduces the heating effect, resulting in a reduced deposition rate, but an increased coating area. Another way to achieve the same effect is to increase the spray angle of the nozzle in use. Perednis  showed that all droplets sprayed from an air blast atomizer are contained within a 70 spray cone angle, while half are within a narrower 12 angle. It was also determined that the flow rate has a very small influence on the spray characteristics, which can be mostly ignored for modeling.
For Electrostatic Spray Deposition (ESD), the cone-jet mode is suitable for thin film deposition. There are two types of ESD nozzles: the cone-jet mode and the multi-jet mode. Their differences are shown in Figure 4.3.
For the cone jet mode nozzle, the convex shape of the liquid surface is distorted by an electric field to form a Taylor cone, which is extended at its apex by a permanent jet with a small diameter. In multi jet mode, the liquid is distorted at the tip of the tube nozzle into many different jets of small diameter. The flow rates which were achieved with the two methods are 2.8ml/h and 5.7ml/h for the multi-jet and single-jet modes, respectively .