North American B-25 Mitchell XB-25F

By the time the United States entered the Second World War the B-25 was flying in every theatre, and the limitations of the inflatable leading-edge de-icer boots that protected its wings were becoming obvious. The Army decided a better system was needed and set up a dedicated facility, the Ice Research Base, to study the effects of ice and frost on aircraft and to develop effective solutions. Aircraft of almost every type in service took part in the programme (a B-17F, two B-24s, an XC-53A, an A-26, a C-82 and a B-26), and North American Aviation contributed its own test bomber, the XB-25E.

In parallel, the NACA had been working on thermal ice prevention for years under researcher Lewis Rodert at the Ames laboratory. His approach used the heat of the engine exhaust to warm air that was circulated through internal ducts in the wing, windshield and tail. That was the system fitted to the XB-25E.

The XB-25E was not a new-build aircraft but a conversion of a B-25C-10-NA (serial 42-32281). Work began in early 1943 and continued through February 1944: the inside of the wing was extensively reworked to pass heated air, the engine cowlings were modified to supply the air to be heated, a heat exchanger was fitted to the exhaust and control valves were added to regulate the flow. It made its first flight on 4 February 1944 with Joe Barton at the controls and, after minimal testing by the manufacturer, was delivered to the NACA research centre at Cleveland in July 1944.

The XB-25F answered a different question: could the same result be achieved electrically? To settle it, another B-25C was fitted with insulated electrical de-icing coils mounted inside the leading edges of the wing and empennage. A single example was produced, again by conversion, under the designation XB-25F-A. Instead of ducts, heat exchangers and valves, the aircraft carried resistance elements that heated the skin of the ice-prone surfaces directly.

The NACA knew that route well: at Ames it had tried internal electrical heating, running resistance wires along the inside surface, and it would later test airfoils with external electric heaters in the icing tunnel. Laboratory results showed that electricity could remove ice with modest energy inputs, but also that the system added weight, was susceptible to failure through damage to the heating circuits and had other disadvantages compared with hot air.

The sources on the XB-25F agree on the same conclusion: the hot-air system tested on the XB-25E was found to be more practical than the electrical coils. The XB-25F remained a single experimental airframe and led to no production version. The XB-25E, by contrast, enjoyed a long life as a flying laboratory: nicknamed "Flamin' Mamie" for its frequent engine fires, it was used at Cleveland to isolate the effect of icing on the propellers in February and March 1945, to develop anti-icing systems for almost every part of the aircraft through 1945 and 1946 and, later, to carry electrical heating in the front fuselage, propellers, inboard wing, cowls and antennae, until it was returned to the Air Force at Wright Field in February 1953.

That detail is worth stressing: although the purely electrical system of the XB-25F did not prosper, electrical heating did not vanish from the programme; it ended up coexisting with hot air on the XB-25E itself, in the places the ducts could not reach well (propellers, antennae, windshield). Exhaust-gas thermal de-icing proved its effectiveness, even though no production B-25 carried it before the end of the war because of its cost, and many later propeller aircraft adopted the XB-25E system.

Public documentation on the XB-25F is very thin. Its serial number, the dates of conversion and first flight, where it was tested and the final fate of the airframe have not been located, nor has the maker of the heating elements. The NACA reference works on ice prevention describe the XB-25E and electrical trials on other aircraft at length but do not mention the XB-25F by name. Any figure not given here should be treated as awaiting a source.