Article pubs.acs.org/ac
A UV−Vis Photoacoustic Spectrophotometer Joseph R. Wiegand, L. Dalila Mathews,† and Geoffrey D. Smith* Department of Chemistry, University of Georgia, Athens, Georgia 30602, United States S Supporting Information *
ABSTRACT: A novel photoacoustic spectrophotometer (PAS) for the measurement of gas-phase and aerosol absorption over the UV−visible region of the spectrum is described. Light from a broadband Hg arc lamp is filtered in eight separate bands from 300 to 700 nm using bandpass interference filters (centered at 301 nm, 314 nm, 364 nm, 405 nm, 436 nm, 546 nm, 578 and 687 nm) and modulated with an optical chopper before entering the photoacoustic cell. All wavelength bands feature a 20-s detection limit of better than 3.0 Mm−1 with the exception of the lower-intensity 687 nm band for which it is 10.2 Mm−1. Validation measurements of gas-phase acetone and nigrosin aerosol absorption cross sections at several wavelengths demonstrate agreement to within 10% with those measured previously (for acetone) and those predicted by Mie theory (for nigrosin). The PAS instrument is used to measure the UV−visible absorption spectrum of ambient aerosol demonstrating a dramatic increase in the UV region with absorption increasing by 300% from 405 to 301 nm. This type of measurement throughout the UV−visible region and free from artifacts associated with filter-based methods has not been possible previously, and we demonstrate its promise for classifying and quantifying different types of light-absorbing ambient particles.
A
UV it has the potential to alter significantly the photochemistry of the troposphere including the production of ozone9,10 and can potentially contribute half as much or more warming per mass as black carbon at UV wavelengths (see, e.g., Barnard et al.11 and Feng et al.8 and references therein). Clearly, then, a precise accounting of atmospheric aerosol optical properties requires an accurate measure of aerosol absorption throughout the UV−visible region of the spectrum. Methods for measuring aerosol absorption have been summarized in an excellent review by Moosmüller and coworkers7 and so will be outlined only briefly, here. There are two common approaches to measuring aerosol absorption: (1) filter-based methods in which particles are collected on a filter and analyzed directly or after extraction by a solvent, and (2) in situ methods of suspended aerosol particles in which absorption is measured from the difference of extinction and scattering or from sound created by modulated absorption (photoacoustic instruments). Among the filter-based instruments are the aethalometer, the particle soot absorption photometer (PSAP) and the multiangle absorption photometer (MAAP), all of which measure the change in transmission of light through the loaded filter. There are well-documented artifacts and issues with such methods, including multiple scattering within the filter, scattering of light by collected particles (though the MAAP does measure and
erosols in the Earth’s lower atmosphere are important for several reasons, including their abilities to facilitate chemical reactions, influence human health, alter cloud properties, and scatter and absorb sunlight. Aerosols significantly influence radiative forcing, the net perturbation in the Earth’s energy balance resulting from an imposed change, with a direct impact on climate change. In fact, the net cooling by aerosols, including both direct and indirect effects, is estimated to be approximately 50% of the warming by the most dominant greenhouse gas, CO2.1 This cooling is a result of both aerosol interaction with clouds and the direct scattering away from the Earth of some of the radiation incident on particles, but it is mitigated by warming resulting from the absorption of light by some types of particles.2 One such type of absorbing particle is black carbon, sometimes referred to as elemental carbon, which is a very strong absorber with weak wavelength dependence throughout the UV and visible regions of the spectrum.3 Another class of particle which has recently been recognized as also being a significant absorber is light-absorbing organic material, or “brown carbon.”4 Brown carbon is characterized by much stronger absorption at UV and near-UV wavelengths than in the visible region. Black and brown carbon are estimated to be emitted at approximately 8.0 Tg/year and 33.9 Tg/year, respectively, from sources such as fossil and biofuel combustion and biomass burning.5 While black carbon has been studied for decades, brown carbon is less extensively studied, and determining its origin and how it absorbs light is an active area of investigation.6−8 With its enhanced absorption in the © 2014 American Chemical Society
Received: April 2, 2014 Accepted: May 25, 2014 Published: June 6, 2014 6049
dx.doi.org/10.1021/ac501196u | Anal. Chem. 2014, 86, 6049−6056
Analytical Chemistry
Article
its ability to measure absorption throughout the spectrum using acetone vapor and nigrosin particles. With 20-s detection limits of better than 3 Mm−1 (3 × 10−8 cm−1) for all but one wavelength (10 Mm−1 at 687 nm), we demonstrate the instrument’s ability to measure the UV−visible absorption spectrum of ambient aerosol. Finally, from this spectrum we are able to infer the relative contributions of black and brown carbon, and we illustrate the necessity of being able to make measurements throughout the UV−visible region of the spectrum to do so.
correct for this), loading-dependent correction factors and potential adsorption and loss of volatile and semivolatile species.12,13 These instruments generally have excellent response times of seconds to minutes and can measure absorption at several wavelengths, though only the aethalometer measures into the near-UV (at 370 nm). Thus, important information about the UV region of the absorption spectrum is lost with these instruments. Alternatively, particulate matter collected on filters can be extracted using a solvent, such as water, methanol, or acetone, and then analyzed using a traditional spectrophotometer.14,15 This approach allows an entire UV−vis spectrum to be measured and makes it possible to measure very small particle concentrations since aerosols are generally collected for several hours. However, this collection also limits the response time to on the order of an hour to a day. It also has the disadvantages that it measures absorption by only the solvent-extractable fraction of the aerosol (i.e., black carbon is not detected) and is insensitive to particle mixing state (e.g., coated particles). In situ measurements of particle absorption have been made using a so-called “difference” approach in which particle scattering, measured using an integrating nephelometer, is subtracted from particle extinction, measured using a cavity ring-down spectrometer16 or long-path extinction cell.17 These instruments typically have fast time responses of seconds to minutes and avoid complications of filter-based methods. However, this approach can be challenging when absorption is a small fraction of extinction, as is often the case. Furthermore, it is limited to visible wavelengths since the integrating nephelometer operates at 450, 550, and 700 nm, though some laboratory instruments can measure scattering at 355 nm18 and 405 nm.19 Another type of instrument, the photoacoustic spectrophotometer (PAS), also measures absorption of suspended particles but does so by turning absorbed energy into sound.20 This approach has the advantages that it measures absorption more directly, is not sensitive to scattering, and has a fast time response of seconds to minutes. Typically, PAS instruments are limited to operating at visible wavelengths accessible with compact lasers, though two laboratory versions have demonstrated the use of UV wavelengths (266 nm/355 nm21 and 355 nm18). Also, PAS instruments are generally limited in the number of wavelengths and spectral coverage possible since separate PAS cells are typically used for each wavelength. Notable exceptions to this are the commercial PASS-3 instrument (Droplet Measurement Technologies; 405, 532, and 781 nm) and recently developed instruments based on a supercontinuum laser source22 (417 nm, 475 nm, 542 nm, 607 and 675 nm) and a pulsed OPO (optical parametric oscillator) laser source23 (410−710 nm cell and 710−2500 nm cell). Without the ability to measure broadband aerosol absorption, none of these instruments is capable of providing an accurate UV−visible spectrum of ambient aerosols or of determining the relative contributions of brown and black carbon. In this work, we describe the development of a lamp-based PAS instrument capable of measuring aerosol absorption throughout the UV−visible region (300−700 nm) of the spectrum. The use of a single light source coupled to a single PAS cell with just one microphone simplifies alignment and calibration and minimizes potential systematic differences between measurements at different wavelengths. We demonstrate calibration of the instrument using NO2 and validation of
■
EXPERIMENTAL SECTION Instrument Description. The instrument operates on the basis of the photoacoustic principle that has been described in detail previously.7,24 Briefly, when modulated light is absorbed by a gas-phase species or a particle, the energy is translated into a transient heating of the surrounding gas, creating pressure waves. These acoustic waves can be detected with a sensitive microphone, and the amplitude of the signal is proportional to the strength of absorption. When the light is modulated at a resonant frequency of the cavity, a standing wave is created, and the sound is amplified. A schematic of the UV−vis PAS instrument is shown in Figure 1. The light from a 100 W Hg arc lamp (Optical
Figure 1. Schematic of UV−vis PAS instrument. L1: lens 1 (focal length = 40 mm), C: chopper wheel, A: aperture (1 cm i.d.), L2: lens 2 (focal length = 40 mm), FW: filter wheel, M: microphone.
Building Blocks, LPS 100) is tightly focused by a 2 in. o.d., 40 mm focal length lens through an optical chopper (Thorlabs, MC2000) which modulates it. Care was taken to shield the light since it contains harmful UV radiation. The modulated light passes through a 1 cm aperture and then a second 2 in. o.d., 40 mm focal length lens which loosely focuses it before it passes through one of the bandpass filters (Semrock) housed in a motorized filter wheel (Edmund Optics, 84-889). The light then enters the PAS cell which is based on the design of Lack and co-workers.25,26 The main cell cavity is a 150 mm long, 25 mm diameter acoustic resonator and is capped at each end with a 1/4-λ acoustic filter and a quartz window for transmission of the light. We note that unlike many laser-based PAS instruments, this one does not employ a multipass mirror arrangement because the light from the lamp is too divergent. An optoacoustic microphone (Optoacoustics, 4110) is placed flush against the surface of the acoustic resonator at the acoustic antinode in the middle of the cell. The microphone signal is amplified by a lock-in amplifier (Stanford Research Systems, SR830) with a 0.1 s time constant and a 24 dB rolloff before being recorded by a PC using a custom LabVIEW (National Instruments) program for data acquisition and processing. The broadband light of the arc lamp is filtered by a series of high transmission, narrow bandwidth dichroic filters providing eight separate wavelength bands throughout the UV and visible regions of the spectrum. Since the filters do not block out-of6050
dx.doi.org/10.1021/ac501196u | Anal. Chem. 2014, 86, 6049−6056
Analytical Chemistry
Article
which lack distinct absorption features. What is more, these bandwidths are narrower than most other commercial aerosol instruments, such as the integrating nephelometer (TSI 3563, fwhm = 40 nm27), the aethalometer (Magee Scientific AE31, fwhm = 11−85 nm depending on wavelength28), the 3-λ PSAP (Radiance Research, fwhm = 20−42 nm depending on wavelength28), and the MAAP (Thermo Fisher Scientific, fwhm = 18 nm28). Finally, the detection limit (20 s), as calculated from measurements of NO2 absorption (described later), is listed for each wavelength band, with all but the 687 nm band falling in the range of 1.3−2.6 Mm−1; the much lower power available in the 687 nm band translates into a larger detection limit of 10.2 Mm−1. Chemicals Used and Particle Generation. Nitrogen (Airgas, 99.99%) and NO2 (Airgas, 10 ppm in N2) were used from the cylinders without further purification. Acetone vapor was generated by flowing N2 through a glass bubbler containing liquid acetone (Sigma-Aldrich, CAS # 67−64−1). Nigrosin (Sigma-Aldrich, CAS# 8005-03-6) particles were generated by an atomizer (TSI 3076) from a 1 mg/mL aqueous solution and a N2 flow rate of 2.7 SLPM. The particles were dried using a silica gel diffusion dryer (TSI 3062) to a relative humidity of
