Life expectancy is defined as the number of hours of operation until 50% of them fail. This means that it is possible for some lamps to fail after a short amount of time and for some to last significantly longer than the rated lamp life. This is an average (median) life expectancy. The life expectancy of filaments, new or refurbished, from any reasonable manufacturer is identical.
Two factors influence filament life expectancy - degree of vacuum and heating voltage. The heating voltage required for saturation is dependent on the bias current selected, so in a SEM/ microprobe a high bias current may be run to achieve something like 40µm emission - but the price paid may be only 40 hours of filament life. A poor vacuum would shorten filament life, as under those conditions the filament succumbs due to evaporation and not etching.
In a TEM the emission may be only 10µm (low bias setting, therefore low heating voltage), and then a filament may last 100 hours and more - if there is a good vacuum. A vacuum of less than 10-5 mbar will result in etching reducing filament life.
Dead filaments with the wire burned out away from the centre emission tip may have suffered from poor vacuum, and they are 'etched'. Filaments which show an excessive metal deposit on the insulator's base may have been operated at an excessive heating current.
Using an excessive bias setting and over-saturating the filament voltage (particularly early in the life of a filament) are the main reasons for early filament failure. Without extensive tests it is impossible for an individual to determine minor filament variations. Operator control affects filaments much more than would another batch of thoriated tungsten wire. The degree of vacuum measured in the gun chamber is important in regard to filament life.
All filaments are annealed. They are heated under vacuum prior to adjusting the centre.