D1, C1 and C2 along with the IR2110 form the bootstrap circuitry. When LIN = 1 and Q2 is on, C1 and C2 get charged to the level on VB, which is one diode drop below +VCC. When LIN = 0 and HIN = 1, this charge on the C1 and C2 is used to add the extra voltage – VB in this case – above the source level of Q1 to drive the Q1 in high-side configuration. A large enough capacitance must be chosen for C1 so that it can supply the charge required to keep Q1 on for all the time. C1 must also not be too large that charging is too slow and the voltage level does not rise sufficiently to keep the MOSFET on. The higher the on time, the higher the required capacitance. Thus, the lower the frequency, the higher the required capacitance for C1. The higher the duty cycle, the higher the required capacitance for C1. Yes, there are formulae available for calculating the capacitance. However, there are many parameters involved, some of which we may not know – for example, the capacitor leakage current. So, I just estimate the required capacitance. For low frequencies such as 50Hz, I use between 47µF and 68µF capacitance. For high frequencies like 30kHz to 50kHz, I use between 4.7µF and 22µF. Since we’re using an electrolytic capacitor, a ceramic capacitor should be used in parallel with this capacitor. The ceramic capacitor is not required if the bootstrap capacitor is tantalum.