[rawhide]

"It turns out that all explosives use groups of nitrogen molecules bonded to two oxygen molecules, which makes them very reactive — they can't help but seep out into the air around anything concealing them. When our green laser hits these molecules they photoluminesce at the 705-nm wavelength. Anything else that radiates at that frequency also radiates at nearby frequencies. Only explosives emit in this very narrow band," said Hummel.
"As a consequence, Hummel's detector could theoretically sense a single molecule from an explosive's "aroma" without fear of false positives, by using two sensors — one for explosives and one to prevent false positives. The primary sensor has a filter so it only senses exactly 705-nm wavelengths, thus "arming" the alarm. But the alarm is only set off if a second, "fail-safe" sensor verifies that there is no equally strong radiation at wavelengths adjacent to 705 nm, thus preventing false positives, according to Hummel."

I do not agree this will be fail safe. One way I see to outwit this test would be to have in addition to the explosives material present, another item which would give off radiation at the 705-nm wavelength and also at the additional nearby frequencies. This would trip off the alarm on the second 'failsafe' sensor, thereby indicating this item giving off radiation is not confined to just explosives frequency, when in fact explosives along with this second item are both giving off radiation, thereby giving the tester the false assumption that no explosives are present.

[SC Swamp Fox]

I do not agree this will be fail safe. One way I see to outwit this test would be to have in addition to the explosives material present, another item which would give off radiation at the 705-nm wavelength and also at the additional nearby frequencies.

Bingo!

You just rendered the device useless.