The increased use of adhesively bonded joints in aerospace structures has given rise to ample research opportunities on the topic of ensuring structural integrity. One such method of verifying bond strength is through the use of non-destructive testing methods. One particular defect type, kissing bonds or zero-volume bonds, is notoriously difficult to detect using traditional techniques such as ultrasonic C-Scan. This thesis explores the feasibility of using Acoustic Emission (AE) monitoring to detect defects and characterize failure modes for adhesively bonded composite joints. Numerous Double Cantilever Beam (DCB) specimens containing various defect types were manufactured and tested under static Mode I loading conditions. Defect types included were pristine specimens, inclusion defects and artificial kissing bonds replicated through contamination of the adherend surface using release agent. Mechanical testing revealed that bondline defects significantly reduce the effective fracture toughness Gef f of defective specimens and induced alternative failure modes instead of complete cohesive failure. Most notably, specimens contaminated with release agent showed adhesion failure consistent with kissing bond type defects, while remaining undetectable when using conventional ultrasonic C-Scan inspection. AE monitoring allowed for identification of signal characteristics corresponding to cohesive failure, delamination and adhesion failure caused by kissing bonds and other defect types. Clustering of acoustic signals revealed characteristic frequency bands for certain defect types and failure modes. In particular, a cluster of signals of 130 [kHz] and 170 [kHz] peak frequency respectively could be correlated to delamination and cohesive failure. Wavelet transforms of the measured signals further revealed the broad frequency spectrum present in all specimen types. However, inconsistencies in the characteristic failure modes attributed to certain frequency bands make this mode of identification insufficient in its current state to detect failure modes with complete certainty.