Characterization of III-V materials by optical interferometry

Digital interference microscopy is a new measuring technique with submicron horizontal resolution and nanometric vertical resolution, that can be used for the three-dimensional analysis of surface defects and device features in many microelectronics applications on bulk materials and epitaxial layers. In this paper we show how certain defects can be analysed on III-V materials and devices using two different interferometric techniques. The choice of the technique depends on the height and the slope of the surface features to be measured. We show that small defects less than λ/2 in height, or surfaces with shallow continuous slopes upto one or two microns high are best profiled with the phase stepping technique (PSM) because of the high vertical resolution of 1 nm and the higher speed and precision. This is illustrated by studies of the surface polish of InP wafers, defects after chemical etching of tin doped InP, defects on an epitaxial layer of GaAs on InP and quantum dot structures on GaAs. For measuring devices which contain mesas and grooves with step heights greater than λ/2, the peak fringe scanning (PFSM) method is the better choice. The vertical resolution is slightly less (4 nm), but the vertical range is higher (upto 15 μm) as demonstrated with the measurement of an etched groove in a laser/detector device on a quaternary layer on InP, and a MESFET device on GaAs. Compared with electron microscopy and the new near field scanning techniques, digital interference microscopy has the advantages of ease of use and speed of analysis and being able to resolve certain problems that are difficult or not possible by other means, such as profiling deep narrow etched grooves, or measuring the relief of a surface hidden under a transparent layer. The main disadvantages are that the horizontal resolution is limited to the resolving power of the objective and that errors due to variations in the optical properties of the sample need to be taken into account.