Chapter 11: Aperture

The campus lab had the specific quality of all shared research spaces: the accumulated presence of multiple projects in various states of completion, the equipment arranged in the compromises that resulted from several people’s needs occupying the same benches, the ambient smell of electronics and optical coatings and the particular chemical undertone of cleaning solvents used on precision surfaces. I had been allocated bench space in the northeast corner, which had the advantages of the room’s most stable temperature differential and the least foot traffic, and which I had organized over the previous three weeks into a configuration that felt, if not quite like the garage, then at least like a space that was responding to my logic rather than resisting it.

The experiment I was running was a development of the phased array architecture I had been designing since the previous spring — an attempt to use quantum dots as repeater elements in a coherent array, phase-locked through the feedback system I had been refining since the laser diode assembly in Charlotte. The quantum dots were the new variable: semiconductor nanocrystals whose optical properties were tunable by size, which meant I could configure them to respond at specific wavelengths and use that specificity to create a more precise phase relationship across the array than conventional optical elements allowed. The theoretical advantage was significant. The practical implementation had been producing results that were interesting in the specific way that results were interesting when they were not quite what the theory predicted — close enough to confirm the underlying principle, different enough to suggest that something in the system was behaving in a way the model hadn’t fully captured.

I had been working on this for four hours when the noise started.

The renovation work had been ongoing in the adjacent wing of the building for two weeks — structural work of some kind, the specific details of which I had not investigated because they were not relevant to my experiment and the building management had assured the lab users that electromagnetic interference would be within acceptable limits. This assurance had been approximately correct for the first two weeks. What began at approximately two in the afternoon on a Tuesday in October was not within acceptable limits.

The spectrum analyzer registered it first — a broadband elevation across a frequency range that was wide enough to be immediately anomalous, with a spectral structure that was complex in a way that simple equipment would not produce. Not white noise, which would have been distributed evenly. Something with internal structure: multiple frequency components with specific relationships between them, varying in amplitude with a pattern that suggested a device operating under load rather than a device producing steady-state output. I noted it and looked at the lab’s other instruments. The oscilloscope showed the same structure in the time domain — a signal that was not periodic in any simple sense but was not random either. It had the quality of a system producing complex output under varying conditions, which was consistent with a welding generator operating at variable load, or possibly a combination of a generator and motor drive system interacting with each other and with the building’s power infrastructure.

I made a note in the lab notebook — time, frequency range, approximate amplitude, spectral character — and continued with the experiment, because the interference was in a frequency range that should not have directly affected the optical system. The quantum dot array operated at visible and near-infrared wavelengths. The electromagnetic noise was primarily in the radio frequency range. The coupling between them, if any, should have been negligible.

The array’s output changed thirty seconds later.

Not gradually. The change had the quality I had learned to associate with threshold transitions — the abrupt shift from one regime of behavior to another, the fringe pattern on the detection screen going from the complex but analyzable interference structure I had been recording to something that the pattern recognition part of my mind flagged before the analytical part had caught up.

I stopped.