There is a change in the surface.
The gas carburizing atmosphere is a reducing atmosphere for iron. This means that the oxides of iron will be converted to elemental iron. Unfortunately, the atmosphere cannot be introduced unless the furnace is above 1300F or so. Below that temperature, the atmosphere is explosive and you can literally blow the doors off the furnace (a rather unsettling occurrence). So, while the parts are heating to about 1400F, the surfaces will oxidize. Then, once the carburizing atmosphere is introduced, this oxide gets reduced back to iron. Of course, the reduced iron does not go back nicely to where it came from before it was oxidized; it ends up being a very poorly adhering plating, with a rough texture.
This effect is very slight, since not much oxidizing occurs in most cases. I've seen loads where the carburizing gas was lost during the carburizing cycle and considerable scaling occurred. When these are re-run without abrasive cleaning, the parts appear to have foil growing on them.
The extent of the effect also depends on the source of the carburizing gas, as it is much worse when using endothermic gas then when using nitrogen/methanol systems. With nitrogen/methanol, the heat treater can run pure nitrogen until the temperature is high enough to introduce the methanol. With endo systems, they have to have the gas turned off until the temperature is high enough. With vacuum carburizing, it effect may not even be noticeable (I don't know, I don't have any direct experience with vacuum carburizing).
rp