Large area flat displays are nowadays considered being a commodity. After the era of bulky CRT TV technology, LCD and OLED have taken over as the most prevalent technologies for high quality image display devices. An important factor underlying the success of these technologies has been the development of high performance photomask writers in combination with a precise photomask process. Photomask manufacturing can be regarded as an art, highly dependent on qualified and skilled workers in a few companies located in Asia. The manufacturing yield in the photomask process depends to a great extent on several steps of measurements and inspections. Metrology, which is the focus of this thesis, is the science of measurement and is a prerequisite for maintaining high quality in all manufacturing processes. The details and challenges of performing critical measurements over large area photomasks of square meter sizes will be discussed. In particular the development of methods and algorithms related to the metrology system MMS15000, the world standard for large area photomask metrology today, will be presented. The most important quality of a metrology system is repeatability. Achieving good repeatability requires a stable environment, carefully selected materials, sophisticated mechanical solutions, precise optics and capable software. Attributes of the air including humidity, CO2 level, pressure and turbulence are other factors that can impact repeatability and accuracy if not handled properly. Besides the former qualities, there is also the behavior of the photomask itself that needs to be carefully handled in order to achieve a good correspondence to the Cartesian coordinate system. An uncertainty specification below 100 nm (3σ) over an area measured in square meters cannot be fulfilled unless special care is taken to compensate for gravity-induced errors from the photomask itself when it is resting on the metrology tool stage. Calibration is therefore a considerable challenge over these large areas. A novel method for self-calibration will be presented and discussed in the thesis. This is a general method that has proven to be highly robust even in cases when the self-calibration problem is close to being underdetermined. A random sampling method based on massive averaging in the time domain will be presented as the solution for achieving precise spatial measurements of the photomask patterns. This method has been used for detection of the position of chrome or glass edges on the photomask with a repeatability of 1.5 nm (3σ), using a measurement time of 250 ms. The method has also been used for verification of large area measurement repeatability of approximately 10 nm (3σ) when measuring several hundred measurement marks covering an area of 0.8 x 0.8 m2. The measurement of linewidths, referred to in the photomask industry as critical dimension (CD) measurements, is another important task for the MMS15000 system. A threshold-based inverse convolution method will be presented that enhances resolution down to 0.5 µm without requiring a change to the numerical aperture of the system. As already mentioned, metrology is very important for maintaining high quality in a manufacturing environment. In the mask manufacturing industry in particular, the cost of poor quality (CoPQ) is extremely high. Besides the high materials cost, there are also the stringent requirements placed on CD and mask overlay, along with the need for zero defects that make the photomask industry unique. This topic is discussed further, and is shown to be a strong motivation for the development of the ultra-precision metrology built into the MMS15000 system. / <p>QC 20130515</p>