Chemical sensing with quantum-cascade lasers

Abstract

Why is it that only a diode laser can burn a movie on a DVD? Why does it take a laser and not a lightbulb to power a fiberoptic network? Here, a laser’s brightness—its ability to focus onto a tiny spot or to travel with minimal loss over long fiber distances—is key (see Figure 1). With their high brightness and small size, mass-producible semiconductor light sources have transformed the telecommunications, lighting, and entertainment industries. Today, quantum cascade lasers (QCLs), with similarly impressive potential, are attracting the interest of the traditional mid-IR spectroscopy community, which has been using blackbody lamps for over a century. However, acceptance of these new light sources has been slow. The main objections are that, compared with the reliability of 10W blackbody sources, QCLs are less powerful, may not work continuously, barely tune 1% of the IR, yet cost 100 times more. Although recent improvements in the power1, 2 and tuning range3 of QCLs are putting the lie to this conventional wisdom, applications have yet to be found that can do for the new sources what DVDs and fiber optic communication did for diode lasers. In this article, we describe research aimed at exploring such applications. IR spectroscopy has been used extensively in chemistry for both structural interpretation and quantitative analysis, in a way similar to mass spectroscopy (MS) but with a new twist: whereas MS smashes a big molecule apart and measures the mass of smaller pieces to identify the original molecule, IR spectral data indicate the shape (or ‘stereo’ structure) of the molecule. IR has typically been overshadowed by MS because in many chemical analytical instruments, its sensitivity is suboptimal. For example, in a traditional gas chromatography (GC)-IR setup, the coupling efficiency and transmission of IR radiation from a blackbody source are so low that the lightpipe or a hollow waveguide (HWG) must have a bore size over 1mm and length less than several inches to get even minimal light for signal Figure 1. Even though a one-dollar lightbulb emits 10W for >1kWh (left), its brightness (power emit from unit area d and angle d around the norm vector N ) is far less, and its etendue ( ) much greater, than a semiconductor laser (the tiny stripe at the fingertip: see inset). Accordingly, quantum-cascade lasers (QCLs) could be coupled efficiently into hollow waveguides, whose etendue (S S ) is much smaller, whereas blackbody lamps cannot.