Reducing the number of array elements in ultrasonic imaging is crucial for lowering system cost and improving scalability. This study introduces a co-prime array configuration to achieve high-resolution imaging with fewer elements. A co-prime array synthesizes a beam pattern equivalent to that of an array with N×M elements by multiplying the beam patterns of two subarrays with co-prime numbers of elements. Thus, only N+M elements are required, enabling significant hardware reduction without degrading spatial resolution. Previous study applied the co-prime array concept to reflective transmit/receive systems in plane-wave coherent imaging to reduce transmission. There, transmit angles were treated analogously to array elements, allowing angular sampling to be minimized. In contrast, we target divergent wave synthetic aperture imaging and implement co-prime subarrays directly in the transducer layout. To replicate the beam pattern of a full N×M array, we employ both dense/dense and sparse/sparse Tx/Rx combinations. Experiments with N = M = 8 achieved azimuth and range FWHMs of 0.24 mm and 0.21 mm, CR of 16.50 dB, and SNR of 1.42 which are comparable to or better thana 64-elements conventional array. This confirms that our approach provides high image quality while significantly reducing the number of elements required.