Nowadays counterfeit electronic devices are in wide circulation and cause huge financial losses to the industry. In this work, visually imperceptible random patterns on a printed circuit board (PCB) surface resulting from the PCB manufacturing process are investigated as the potential fingerprints for counterfeit detection of electronic devices. For device authentication, surface fingerprints are matched by computing the normalized cross-correlation. The experimental results show that the variations (e.g.small marks of random size, shape, orientation, shape distortions, texture, etc.) encountered in the interlayer connecting vias resulting from the PCB manufacturing process imperfections can be used as surface fingerprints for device authentication. For performance evaluation PCB surfaces with specially designed test patterns produced by industrial grade PCB manufacturing facilities are considered. Appropriate measures are suggested to address the challenges that are not yet addressed in this research work.