Theoretically it should be possible e.g. the filament in an incandescent lamp glows at about 6000 .C. In this case, the heat should be transported to the heat exchanger by means of radiation in a vacuum or innert environment, as materials get quite reactive with oxygen in air at these temperatures.
Considering heating elements my concern is that one needs to insulate the bare resistor wire from the heat exchanger, which invariably will be conductive, unless one uses the heat exchanger itself as the resistor. The latter is probably not feasible, because one will have to deisgn a resistive mechanical part, that is likely going to be requiring too much current. Furhtermore, electrical insulators are also thermal insulators, so one will be restricted to how much heat can be conducted into the heat exchanger.
I would propose you rather consider induction heating of the heat exchanger. 1000 degrees C is quite possible (an ex colleague of mine has buit an induction furnace for melting platinum) and the conductor can stay insulated from the hot parts. In the design, one should take into consideration that the material properties change with temperature to not heat up, and weaken, the heat exchanger to the point of failure.
You need about 12 kW of power to heat the said air flux to the desired temperature (not considering the heat losses). To transfer this amount of heat from the pipe wall to the air ( supposing that temperature difference between the wall and the air is not more than 200C) the pipe length must be of approximately 1 km (If we suppose that the pipe diameter =0.05m and the air flow rate=10 L/s). This estimation is based on the data taken from the book "T.Cebeci, P.Bradshaw. Physical and computational aspects of convective heat transfer" , according to the data presented in fig 7.4 of this book the Nu number under the above conditions is about 30, and hence the heat transfer coefficient is close to 28 W / (m 2 * K). From this follow above estimates. So by my opinion the more realistic ways of air heating in these conditions is either the wire mash heating or an electric arc in the gas stream.
You need about 12 kW of power to heat the said air flux to the desired temperature (not considering the heat losses). To transfer this amount of heat from the pipe wall to the air ( supposing that temperature difference between the wall and the air is not more than 200C) the pipe length must be of approximately 1 km (If we suppose that the pipe diameter =0.05m and the air flow rate=10 L/s). This estimation is based on the data taken from the book "T.Cebeci, P.Bradshaw. Physical and computational aspects of convective heat transfer" , according to the data presented in fig 7.4 of this book the Nu number under the above conditions is about 30, and hence the heat transfer coefficient is close to 28 W / (m 2 * K). From this follow above estimates. So by my opinion the more realistic ways of air heating in these conditions is either the wire mash heating or an electric arc in the gas stream.
Using the numbers from your post (28 W/m2/K, 50 mm tube, 200 C delta T), one can get tube length ~14 m, not 1 km. But, as Azd noticed, heating up gas through the tube wall is not the best approach - annular design is much more effective.
But your suggestion about using arc looks attractive.
Azd,
You can try to use plasma torches like this: http://www.westinghouse-plasma.com/wpc_plasma_torches/
Kanthal - A part of Sandvik is manufactering the Electric Heaters, Electric Heat Exchangers & furnaces which can operate to very high temperatures. The unit based in India( Hosur) can help you with any high temp & pr. heating solutions.