Ultrasound-induced thermal strain imaging (TSI) has been used to successfully identify lipid and water-based tissues in atherosclerotic plaques in some research settings. However, TSI faces several challenges to be realized in clinics. These challenges include motion artifacts, displacement tracking accuracy as well as limited heating capability which contributes to low thermal strain signal-to-noise ratio and a limited field of view. The goal of this paper is to address the challenge in heating tissue in TSI. Current TSI systems use separate heating and imaging transducers, which require physically aligning the heating and imaging beams and result in a bulky setup that limits in vivo operation. This paper proposes and evaluates a new design for heating beams that can be implemented on a linear array imaging transducer and can provide an improved heating area and efficiency as compared to previous implementations. The designed heating beams were implemented with a clinical linear array imaging transducer connected to a research ultrasound platform. In-vitro experiments using tissue mimicking phantoms with no blood flow showed that the new design resulted in an effective heating area of approximately 0.85 cm2 and a 0.3°C temperature rise in 2 seconds of heating, which compared well with in- silico finite element simulations. With the new heating beams, TSI was shown to be able to detect a lipid-mimicking rubber inclusion with a diameter of 1 cm from the water-based gelatin background, with a strain contrast of 2.3 (+0.14% strain in the rubber inclusion and −0.06% strain in the gelatin background). Lastly, lipid-based tissue in a 1-cm diameter human carotid endarterectomy (CEA) sample was identified with good agreement to histology.