Positron annihilation age momentum correlation ͑AMOC͒ experiments were performed in linear poly͑methyl methacrylate͒, PMMA, at 420 K and 100 K. The lifetime and Doppler broadened line-shape S(t) data were analyzed with Byakov and Stepanov's blob model for the formation of positronium. It is shown that the analysis is consistent with a delayed formation of positronium and also offers an alternative explanation for the observed broadening of the annihilation radiation line shape at short times ͑a few tens of picuseconds͒ after positron injection, the so called young age broadening. DOI: 10.1103/PhysRevB.68.132202 PACS number͑s͒: 61.80.Fe, 36.10.Dr, 78.70.Bj, 82.35.Lr The interpretation of positron annihilation data and of positronium formation in molecular liquids and solids, particularly in polymers, present some challenging riddles. Common lifetime spectra ͑LT͒ in polymers are analyzed as a composition of three exponential decay modes or components, indicated by their lifetimes 1,2,3 and their intensities I 1,2,3 . The shortest component is interpreted as due to the annihilation of parapositronium (p-Ps), the second component is due to annihilation of free positrons, and the longest one is the orthopositronium (o-Ps) component, where the intensity of the p-Ps component should be equal to one-third of the intensity of the o-Ps component. Such an interpretation is satisfactory if one is only interested in the o-Ps component and in the relation between o-Ps and the mean size of the free-volume sites. Detailed analysis may lead to a number of discrepancies, given as follows.͑1͒ The lifetime of p-Ps, which is expected to be not very different from the lifetime in vacuum, i.e., 124 ps, has values as low as 100 ps or as high as 200 ps and its fitted intensity ratio to the long component is much higher than 1:3. Most authors dismiss these discrepancies as due to a complicated mixing of p-Ps with other unidentified states.͑2͒ The intensity of o-Ps, which has been believed to be related to the density of free-volume sites, is now known to depend on the chemical composition, the radiation 1,2 and mechanical 3 history of the sample, the electric field, 4 the polarity, 5 the content of electron-paramagnetic-resonanceactive species, 6 and on bleaching by light. 7 ͑3͒ In many cases the longest component splits up into two components. A number of possible explanations have been suggested, such as ͑i͒ the existence of two o-Ps states or of bound positron-ion states, ͑ii͒ the influence of slow trapping of the positronium into the free volume, and ͑iii͒ the existence of a broad distribution of the o-Ps lifetime. 8 ͑4͒ The available AMOC ͑age momentum correlation͒ data on poly͑methyl methacrylate͒ ͑PMMA͒ 9 and PE 10 have shown clearly that it is impossible to reproduce the timedependent line-shape parameter S(t) with the standard 3-or 4-component model.In order to solve the foregoing problems it is necessary to establish a detailed model of positronium formation, particularly its evolution in time at the sub-nanosecond scale an...