Cavity Ring-Down Spectroscopy (CRDS)

Why cavity ring-down spectroscopy?

A wide variety of laser-based spectroscopic techniques are used to probe chemical dynamics, combustion and plasma environments, the Earth's atmosphere, and many other fascinating problems in modern Chemistry. Methods such as Laser Induced Fluorescence (LIF), Resonance Enhanced Multiphoton Ionisation (REMPI), Degenerate Four-Wave Mixing (DFWM) and Photoacoustic Spectroscopy rely on measuring a side effect of the laser excitation of the sample, and have proved to be sensitive tools for spectroscopic studies. An important place remains, however, for laser absorption spectroscopy since it offers the possibility of quantitative measurements and (unlike LIF for example) can be used to study molecules with very short-lived excited states. The major drawback has been that the sensitivity of direct absorption methods can be limited: the detection of weak absorptions is made difficult by large background fluctuations in absolute light intensity. Cavity ring-down spectroscopy (CRDS), by virtue of its insensitivity to fluctuations in laser output and the enormous (many km) pathlengths through the sample that can be achieved, has become the method of choice for ultra-sensitive, quantitative absorption measurements.

We are currently using pulsed and cw laser CRDS, as well as related techniques such as cavity enhanced absorption spectroscopy (CEAS) and long-path diode laser absorption spectroscopy in a variety of projects. See the bottom of this page for some representative references.

1. Atmospheric chemistry: we are using modern, tuneable cw diode lasers to develop compact and fast detectors for volatile organic compounds (VOCs) [1-3], NO₂ [4,5], various radicals, and atmospheric aerosol particles. We have demonstrated sensitivities to VOCs and NO₂ in the parts per billion (ppb) to parts per trillion (ppt) range and are currrently improving on these detection limits. In September 2006, we measured IO mixing ratios in the range 10 - 40 ppt in the marine boundary layer at Roscoff in France using a field-deployed CRD spectrometer [17]. Pre-concentration of ethene and acetylene from Bristol air samples, followed by CRDS detection, allowed rapid measurement of concentrations of these small organic molecules at levels of 1 - 6 ppb without any need for gas-chromatographic separation [3].

Other significant work concerns the first excited electronic state of molecular oxygen, for which we were able to determine the previously uncertain radiative decay rate [6] by measuring the Einstein A and B coefficients for the very weak a - X band from CRDS absorption spectra. The transition is strictly electric dipole forbidden, and occurs via a weak, spin and orbitally forbidded magnetic dipole transition. The radiative lifetime is crucial in infering ozone abundances from airglow data because UV photolysis of ozone produces molecular oxygen in its first electronically excited state. A very high quality simulation of the a - X spectrum (which appears to improve upon the current HITRAN recommendation) is now available using the new PGOPHER simulation program (see link in the left hand panel) and an input file available from Andrew Orr-Ewing on request.

Above: pulsed laser CRDS spectrum of the O₂ a - X absorption band.

We have constructed a pulsed laser CRDS experiment to measure simultaneously high resolution UV absorption cross sections for HCHO and quantum yields for H and HCO formation from the photodissociation of HCHO under tropospheric conditions of temperature (220 - 300 K), pressure, and wavelength of UV radiation (300 - 360 nm) [7,8]. Formaldehyde photochemistry plays a key role in production of HO_x (i.e., OH and HO₂) in the troposphere, but models of the chemistry of the atmosphere require much improved values of the HCHO absorption cross sections and quantum yields to account properly for this contribution. The wavelength of the UV laser is scanned, and its attenuation by a known pressure of HCHO measured and analysed using the Beer-Lambert law to obtain absorption cross sections. Simultaneously, the yield of HCO radicals is measured by CRDS using a red laser beam at wavelengths around 650 nm.

Left: A schematic diagram of the formaldehyde photochemistry experiment. Right: a sample absorption spectrum of HCHO (bottom) and of the CRDS measurement of the yield of HCO as the UV laser is scanned.

We are developing novel CRDS instruments based on cw diode lasers to measure optical properties of single aerosol particles and ensembles of multiple particles dispersed in air. References [15,16] describe the theoretical background and some first implementations of our methods. CRDS with a pulsed, widely tuneable laser system is also being used in collaboration with Dr J.P. Reid to measure optical extinction of atmospheric aerosol particles.

2. Probing of diamond deposition plasmas: both pulsed and cw laser CRDS are being used to study the gas phase chemistry of activated CH₄/H₂ gas mixtures used to grow thin diamond films by chemical vapour deposition (CVD), as part of the group's extensive activities in laser and mass spectrometry based diagnostics of plasmas and CVD processes. We have recorded very high quality spectra of the C₂ d-a Swan band and CH A-X band in a dc arc jet operating with an Ar / CH₄ / H₂ feedstock mixture. Analysis demonstrates a rotational temperature in the arc jet of 3300 K and absolute column densities of both radicals have been measured. Within 5 mm of the growing diamond film, we observe a boundary region in which there is a pronounced rise in the gas temperature and the column densities of both radical species. Comparison with the models of Mankelevich (Moscow State University) shows extraordinarily good agreement and provides a clear understanding of the chemical processes in the plasma plume [9-11]. CRDS measurements of H(n=2) (via the Balmer-beta absorption) in an Ar/H₂ arc discharge plume allowed us to derive local electron densities from Stark broadening of the spectral line. Again, comparisons with models (with allowance for radiation trapping, and production of H(n=2) by reaction of Ar⁺ with H₂, followed by dissociative electron attachment to ArH⁺) give quantitative agreement. A novel design of translatable cavity and input laser beam axis has enabled us to map the number densities of C₂ radicals and H(n=2) atoms as a function of radial distance from the centre of the arc jet plume. The CRDS work is currently being extended to the study of a microwave plasma used for diamond film deposition.

3. Fundamental studies of molecular spectroscopy, structure and dynamics: the focus of this research has been predissociation dynamics in the excited electronic states of the molecules HNO, SH, S2, BrO, IO, ClO and FCO. CRD spectra of the electronic transitions of these radicals show broadening of rotational lines because of the reduced lifetimes of the predissociated excited state levels. These lifetimes may be in the picosecond or femtosecond range. Spectral simulations and fits allow us to quantify the lifetime broadening effects, and thus deduce the rates of predissociation of different rotational and vibrational levels. In some cases, the upper state lifetimes are so short that the individual rotational lines are not resolved, and band contour fitting must be used to deduce lifetimes. Aided by high level ab initio calculations, together with Fermi Golden rule calculations of the dissociation rates for different vibrational levels, this work has resulted in a much clearer understanding of the electronic states involved in these predissociation processes. In the case of ClO, for example, we presented a quantitative model of the lifetimes of the A-state that accounted for the strong vibrational level dependence for the first time [12,13]. The motivation for the halogen monoxide studies comes from the atmospheric significance of these species, which are implicated in ozone destruction cycles in both the stratosphere and the marine boundary layer of the troposphere. These radicals cannot be studied by LIF (with the exception of a few transitions of IO) because the upper states dissociate on timescales faster than the fluorescence lifetime, but CRDS can be used to study their reaction kinetics, spectroscopy, and, we anticipate, their ppt level concentrations in the atmosphere.

4. Spectroscopy of thin films of nanoparticles: we have demonstrated that absorption spectra can be obtained of monolayer films of CdSe quantum dots (spherical particles with diameters of a few nanometres), using CRDS to provide the necessary sensitivity. The Q-dot absorption spectra look very like those in the solution phase, except for marked red shifts that are likely to be a consequence of interactions with near neighbour nanoparticles. The measurements are made possible by mounting the monolayer sample (on a transparent substrate) at Brewster's angle within the ring-down cavity to minimize scattering losses at the interfaces with air.

How it works:

The schematic diagram shows a pulsed laser cavity ring-down apparatus:

The basic experimental requirements are a tuneable laser source, two highly-reflective (> 99.9%) dielectric-coated concave mirrors, a photo-sensitive detector and data acquisition equipment. The mirrors form the windows of a vacuum cell inside which the species of interest may be placed. Although CRDS using continuous wave (cw) lasers is becoming widespread, the discussion here will concentrate on the use of pulsed lasers. The laser beam is aligned along the axis of the ring-down cavity defined by the two mirrors. Clearly most of the laser light is reflected straight back off the input mirror, but a small percentage (< 0.1%) is transmitted through. The mirrors are mounted so that their positions can be minutely adjusted, and with careful alignment the laser pulse may be trapped inside the cavity, being reflected backwards and forwards. In this way a pulse can be stored for microseconds (thousands of round trips, giving effective kilometres-long pathlengths between the mirrors) before decaying away due to cavity losses (such as mirror transmission or sample absorption.

A photo-sensitive detector is positioned behind the output mirror to record the (tiny) intensity of light transmitted through it; because the light intensity is reduced by a given percentage on each round trip, the detector sees an exponential decay of light ringing down. (if short laser pulses are used, a very fast detector will see a train of pulses within an exponential decay envelope, but the time response of detection electronics usually means the pulses are smoothed into a single exponential curve.) This is shown below:

View an animation of the exponential decay of light ringing down inside the cavity.

The signal from the detector is amplified and digitised, and fed to a computer, which fits the trace to a first-order exponential function to determine the decay time constant for each pulse. This constant is determined by two factors: the reflectivity of the mirrors and attenuation of the laser pulse by any absorbing medium inside the cavity (i.e. the pulse rings down faster if some intensity is lost to absorption on each round trip). The time dependence of the pulse intensity is then given by

where t₀ is the time the laser pulse takes to decay to 1/e of its initial intensity in an empty cavity, a is the absorption coefficient of the absorbing medium, and c is the speed of light. When there is no absorption inside the cavity (a = 0) the decay rate is simply 1/t₀ - on the other hand, when a sample inside the cavity absorbs some of the light, the decay rate is given by 1/t₀ + ac. Therefore by plotting the decay rate as a function of the laser frequency as it is scanned, an absorption spectrum is built up, and because the difference between on- and off-resonance features is simply ac the recorded spectrum is quantitative.

On the plus side:-

• The sensitivity of CRDS stems in part from the huge number of passes each laser pulse makes between the ultra-high reflectivity mirrors - for reflectivities of 0.9999, the number of round trips is about 5000, giving an effective pathlengh of 10 kilometres, far higher than for conventional multi-pass arrangements.

• By measuring the ring-down decay rate rather than absolute intensity of the laser pulse, shot-to-shot variations in laser output are removed from the final spectrum.

These two aspects allow the detection of absorptions smaller than one part in 10⁷ per pass.

• Whereas other techniques such as LIF may rival the sensitivity of CRDS, the ability to record quantitative absorption spectra makes the ring-down technique preferable in situations where absolute intensities are required, or where fluorescence yields are poor (e.g. in short-lived predissociated systems).

• CRDS is applicable at any wavelengths for which the requisite highly-relective mirrors are available, and thus encompasses a large spectrum extending from the mid-infrared to deep into the ultraviolet. As such, ring-down spectroscopy has been widely used to study vibrational and electronic spectra of molecules present at very low concentrations or having very low transition probabilities, making their detection difficult by conventional methods.

• The apparatus required for CRDS is relatively compact and inexpensive, consisting of bench-top apparatus and a single laser system. Recent developments with cw diode lasers have realised the potential for compact instruments to perform in-situ measurements of atmospheric trace gas constituents.

On the down side:-

• CRDS is non-species selective: all absorbing molecules (at the wavelength of the laser) in the cavity will contribute to the ring-down decay rate, causing spectral contamination in some instances.

• The technique requires constant mirror reflectivities over long timescales, something that is hard to achieve in "dirty" environments causing deposition on mirror surfaces. This may be overcome in some cases by directing inert gas streams across the mirror faces.

• Each set of mirrors has limited spectral range (typically 50 nanometres).

• The method samples along a column rather than at a point, which is disadvantageous when attempting accurate spatial profiling measurements - but this can be overcome as we have shown in our work on the DC arc jet reactor.

References:

[1] Trace detection of methane using continuous wave cavity ring-down spectroscopy at 1.65 mm, B.L. Fawcett, A.M. Parkes, D.E. Shallcross and A.J. Orr-Ewing, Phys. Chem. Chem. Phys. 4, 5960 (2002).

[2] Trace detection of volatile organic compounds by diode laser cavity ring-down spectroscopy, A.M. Parkes, B.L. Fawcett, R.E. Austin, S. Nakamichi, D.E. Shallcross and A.J. Orr-Ewing, The Analyst 128, 960-965 (2003).

[3] Combining pre-concentration of air samples with cavity ring-down spectroscopy for detection of trace volatile organic compounds in the atmosphere, A.M. Parkes, R.E. Lindley and A.J. Orr-Ewing, Anal. Chem. 76, 7329-7335 (2004).

[4] Fast Fourier transform analysis in cavity ring-down spectroscopy: application to an optical detector for atmospheric NO₂, M. Mazurenka, R. Wada, A.J.L. Shillings, T.J.A. Butler, J.M. Beames and A.J. Orr-Ewing, Appl. Phys. B, 81, 135-141 (2005).

[5] Continuous wave cavity ring-down spectroscopy measurement of NO₂ mixing ratios in ambient air, R. Wada and A.J. Orr-Ewing, The Analyst 130, 1595 - 1600 (2005).

[6] Integrated absorption intensities and Einstein coefficients for the O₂(a¹D_g-X³S_g^-) transition - a comparison of cavity ring-down and high resolution Fourier transform spectroscopy using a long-path absorption cell , S.M. Newman, I.C. Lane, A.J. Orr-Ewing, D.M. Newnham and J. Ballard, J. Chem. Phys. 110, 10749 (1999).

[12] The UV absorption of ClO 1: The A² P - X² P spectrum at wavelengths from 285 - 320 nm studied by cavity ring-down spectroscopy, W.H. Howie, I.C. Lane, S.M. Newman, D. Johnson and A.J. Orr-Ewing, Phys. Chem. Chem. Phys. 1, 3079 (1999).

[13] The UV absorption of ClO 2: Predissociation of the A² P state studied by ab initio and Fermi Golden rule calculations, I.C. Lane, W.H. Howie and A.J. Orr-Ewing, Phys. Chem. Chem. Phys. 1, 3087 (1999).