Non-invasive treatment of brain disorders using ultrasound would require a transducer array that can propagate ultrasound through the skull and still produce sufficient acoustic pressure at a specific location within the brain. Additionally, the array must not cause excessive heating near the skull or in other regions of the brain. A hemisphere-shaped transducer is proposed which disperses the ultrasound over a large region of the skull. The large surface area covered allows maximum ultrasound gain while minimizing undesired heating. To test the feasibility of the transducer two virtual arrays are simulated by superposition of multiple measurements from an 11-element and a 40-element spherically concave test array. Each array is focused through an ex vivo human skull at four separate locations around the skull surface. The resultant ultrasound field is calculated by combining measurements taken with a polyvinylidene difluoride needle hydrophone providing the fields from a 44-element and a 160-element virtual array covering 88% and 33% of a hemisphere respectively. Measurements are repeated after the phase of each array element is adjusted to maximize the constructive interference at the transducer's geometric focus. An investigation of mechanical and electronic beam steering through the skull is also performed with the 160-element virtual array, phasing it such that the focus of the transducer is located 14 mm from the geometric centre. Results indicate the feasibility of focusing and beam steering through the skull using an array spread over a large surface area. Further, it is demonstrated that beam steering through the skull is plausible.