In spite of many remarkable prototype device performances reported usi
ng semiconductor multilayers that incorporate thin tunnel barriers, it
has never been established that the devices are capable of routine ma
nufacture. With reference to the simplest possible tunnel device struc
ture, a single similar to 3 nm thick AlAs barrier within an asymmetric
doping environment in GaAs (the so-called ASPAT microwave detector di
ode), we show here that is not yet possible to design, grow or qualify
the semiconductor multilayers with sufficient accuracy, precision, un
iformity or reproducibility that would allow reverse engineering, a pr
erequisite for manufacture, to be undertaken with confidence. We descr
ibe the improvements in design, growth and qualification that will be
required to achieve manufacturability, and comment on the feasibility
of attaining this goal. Our conclusions for thin tunnel barrier device
concepts apply a fortiori to any device ideas that seek to exploit me
soscopic phenomena in semiconductors.