Published: October 13, 2011 r2011 American Chemical Society 8492 dx.doi.org/10.1021/ac201775f | Anal. Chem. 2011, 83, 84928500 ARTICLE pubs.acs.org/ac Top Notch Design for Fiber-Loop Cavity Ring-Down Spectroscopy Cathy M. Rushworth, Dean James, Jason W. L. Lee, and Claire Vallance* Department of Chemistry, University of Oxford, Chemistry Research Laboratory, 12 Manseld Road, Oxford, OX1 3TA, U.K. INTRODUCTION Cavity ring-down spectroscopy (CRDS) is a highly sensitive spectroscopic technique most commonly used for measuring weak absorptions of trace gas species. 1 Light is trapped within an optical cavity usually formed from two highly reective super- mirrors, gaseous sample is introduced between the two mirrors, and the increase in cavity losses is measured by comparing the decay time constant for the cavity (the ring-down time) in the absence and presence of sample. 2 CRDS gains its sensitivity from two sources: rst, as the light is trapped within a high nesse cavity, the path length through the sample is vastly increased relative to a single-pass experiment; second, the decay time is independent of the initial light intensity, so the measured signal is essentially immune to uctuations in the intensity of the light source. The use of two-mirror cavities has reached maturity over the past few decades. However, interesting cavity innovations continue to emerge, particularly in the area of small volume liquid sample analysis. 3,4 Fiber-optic-based variants of CRDS are parti- cularly attractive for small volume liquid analysis since the small size of optical bers (typically <500 μm core diameter) means that picoliter to nanoliter sample volumes can easily be inter- rogated. There are two types of ber-optic-based cavities, employing linear and loop congurations, respectively. Linear ber CRDS measurements use a length of ber with either mirrors 5À7 or ber Bragg gratings 8 at either end to conne the light in the cavity, in a direct analogue to the two-mirror approach. In this case, a region of the ber cladding is often stripped and the evanescent eld from the surface of the ber core is absorbed by sample in contact with the ber surface. In a potentially simpler cavity arrangement, ber-loop CRDS, 9 the cavity comprises a loop of optical ber formed by join- ing together two ends of a length of optical ber. Light is side-coupled into and out of the ber-loop cavity, and sample can be introduced to the circulating light simply by separating the two ber ends by a short distance and injecting liquid sample between the ber ends. The rst ber-loop CRDS measurements were reported by Stewart and co-workers, 10 who constructed a cavity from standard ber-optical components and inserted a 5 cm microoptical gas sample cell. The sample region in this case (even without an absorbing species) introduced a loss of around 20% per pass, with further losses from the couplers. A ber amplier was required in order to oset the loop losses, and the sensitivity of the measurements thus became dependent on the ampliers stability. Passive (nonamplied) ber-loop CRDS has been developed by Loock and co-workers. 11,12 In the rst pulsed CRDS experiments, light from a Nd:YAG-pumped dye laser was coupled into and out of the ber loop by bending the optical ber at two dierent positions and exploiting the resulting macro- bending losses. Bend coupling is inherently inecient: Loocks group reported a coupling fraction of 10 À8 when coupling light into the loop and 10 À4 when coupling out using 3 cm bends in a loop of 50 μm core diameter optical ber, 11 and the coupling losses themselves were signicantly higher, since light was lost in all directions from the bend. Coupling eciency can be improved by using smaller bends, but at the cost of higher round-trip losses. An alternative coupling strategy is to use commercially available couplers, although these are also relatively inecient, and often occasion higher insertion losses than expected. Lehmann and co- workers used two 99%/1% split ratio input/output couplers and found round-trip losses of around 13%, attributed to losses from Received: July 10, 2011 Accepted: October 13, 2011 ABSTRACT: Fiber-loop cavity ring-down spectroscopy (CRDS) is a highly sensitive spectroscopic absorption technique which has shown considerable promise for the analysis of small-volume liquid samples. We have developed a new light coupling method for ber-loop CRDS, which overcomes two disadvantages of the technique: low eciency light coupling into the cavity and high loss per pass. The coupler is based on a 45° reective notch polished between 10 and 30 μm into the core of a large-core-diameter (365 μm) optical ber, and allows for nearly 100% light coupling into the cavity, with a low loss per pass (<4%). The coupler has the additional advantage that the input and output light is spatially separated on opposite sides of the ber. The detection sensitivity of a ber-loop CRD spectrometer employing the new coupling method is established from ring-down measurements on aqueous rhodamine 6G (Rh6G) at 532 nm. The results are compared with data obtained using the same light source and detector, but a conventional bend-coupled small-core-diameter (50 μm) optical ber loop. With our new coupler, a detection limit of 0.11 cm À1 is found, which corresponds to detection of 0.93 μM Rh6G in a volume of only 19 nL. This is an improvement of over an order of magnitude on our bend-coupled small-core optical ber results, in which a detection limit of 5.3 cm À1 was found, corresponding to a detection of 43 μM Rh6G in a volume of 20 pL.