A water droplet that impacts on a cold surface will start to freeze and in time ice will accumulate. To exemplify, effects of ice accretion is important in areas such as power generation e.g. wind power and vehicles located in a cold climate e.g. aircraft, cars, and boats. The common denominator for these examples is that ice accumulation can lead to a loss of efficiency and in some cases danger. Most studies have so far focused on investigating freezing water droplets visually in experiments or numerically in regards to how the freezing process behaves in terms of shape or freezing time for either a sessile or impacting droplet. It has been observed that the surface material and structures of the substrate is of importance. One part of the freezing process that has been less investigated is the internal flow and how it affects the freezing process. In this thesis, the internal flow in a freezing water droplet has been investigated experimentally. The internal flow inside a droplet is calculated by using Particle Image Velocimetry. A metal plate with a groove filled with ice was used to generate an area for the nucleation to start and to be able to control the shape of the droplet.  Previous work indicate that the substrate is of importance for the freezing process. The influence of the substrate material on the internal flow for similar shaped droplets is therefore investigated in Paper A, for a substrate temperature of -8°C. The results show that the substrate material, here in terms of metals such as aluminum, copper and steel, affect the magnitude of the internal velocity. In paper B it is investigated how the contact angle influence the internal flow. The vector field is examined at 9% of the total freezing time for water droplets at five different contact angles. A droplet with a higher contact angle will have a higher internal velocity in the center. A lower contact angle will barely show any movement in the center, however a higher velocity magnitude is observed close to the free surface compared to a droplet with a higher contact angle. Paper C studies the time until the directional change of the internal flow in a water droplet. Experiments at -8°C as in Paper B are used as well as experiments at -12°C for the five different contact angles. The time until the directional change is similar in time for both -8°C and -12°C while the total freezing time and also the time of the directional change varies with contact angles. A droplet with a lower contact angle will have a shorter time until the directional change occure while an increase in contact angle prolongs both freezing time and the time until the directional change.