The formaldehyde (H2CO) observed in cometary atmospheres presents a so-called extended source, meaning that its distribution in the coma cannot be explained by sublimation from the nucleus alone; a production inside the coma has to be included. Polyoxymethylene (formaldehyde polymers: (-CH2-O-)n, also called POM) is sometimes evoked as a parent molecule for this extended source. This solid polymer on cometary grains could release gaseous formaldehyde through thermal and photolytic degradation. We have developed an experimental program in order to study the chemical reactions of degradation of POM by UV photolysis and heating. It provides identification of the degradation products and the determination of the photo degradation quantum yields or the thermal degradation kinetics. In this paper we present the improvements of our experimental setup and new measurements on the kinetics of gaseous formaldehyde production by thermal degradation of two types of commercial POM on a greater temperature range than the previous ones. These new data are then included in a model of the outer coma in order to interpret observations. This model takes into account the production of gaseous molecules from solid material present on grains and leads to a very good agreement with Giotto observations of comet 1P/Halley, if we assume that the cometary grains contain 1--16% of POM by mass. Thus, without being final evidence of this polymer presence on comets, we conclude that the degradation of solid POM-like polymers on cometary grains is to date the best explanation of the H2CO extended source.