Article pubs.acs.org/est
Fast Time Resolution Oxidized Mercury Measurements during the Reno Atmospheric Mercury Intercomparison Experiment (RAMIX) Part of the “RAMIX: Reno Atmospheric Mercury Intercomparison eXperiment” group Jesse L. Ambrose,*,† Seth N. Lyman,†,∥ Jiaoyan Huang,§ Mae S. Gustin,§ and Daniel A. Jaffe†,‡,∥ †
Science and Technology Program, University of Washington-Bothell, Bothell, Washington, 98011, United States Department of Atmospheric Sciences, University of Washington, Seattle, Washington, 98195, United States § Department of Natural Resources and Environmental Science, University of Nevada-Reno, Reno, Nevada, 89557, United States ‡
S Supporting Information *
ABSTRACT: The Reno Atmospheric Mercury Intercomparison Experiment (RAMIX) was carried out from 22 August to 16 September, 2011 in Reno, NV to evaluate the performance of new and existing methods to measure atmospheric mercury (Hg). Measurements were made using a common sampling manifold to which controlled concentrations of Hg species, including gaseous elemental mercury (GEM) and HgBr2 (a surrogate gaseous oxidized mercury (GOM) compound), and potential interferents were added. We present an analysis of Hg measurements made using the University of Washington’s Detector for Oxidized Hg Species (DOHGS), focusing on tests of GEM and HgBr2 spike recovery, the potential for interference from ozone (O3) and water vapor (WV), and temporal variability of ambient reactive mercury (RM). The mean GEM and HgBr2 spike recoveries measured with the DOHGS were 95% and 66%, respectively. The DOHGS responded linearly to HgBr2. We found no evidence that elevated O3 interfered in the DOHGS RM measurements. A reduction in RM collection and retention efficiencies at very high ambient WV mixing ratios is possible. Comparisons between the DOHGS and participating Hg instruments demonstrate good agreement for GEM and large discrepancies for RM. The results suggest that existing GOM measurements are biased low.
■
INTRODUCTION Mercury (Hg) is an environmental pollutant of concern globally due to its impact on ecological and human health.1−7 The atmosphere is a primary medium through which Hg is dispersed and introduced to aquatic ecosystems.5,8,9 Atmospheric deposition of reactive mercury (RM), which is thought to be composed mostly of divalent inorganic oxidized Hg species (HgII), is understood to be a key mechanism both for removing Hg from the atmosphere and for delivering chemically active, bioavailable HgII to the surface.5,6,9 Thus, accurately describing the atmospheric chemistry and distribution of RM is essential for predicting the fate of Hg in the atmosphere and Hg contamination of aquatic ecosystems.5,10−12 The chemical identity of atmospheric RM species and the mechanisms by which gaseous elemental mercury (GEM) is converted to RM in the atmosphere are largely unknown,11,13−15 although mounting evidence supports a role of halogen radicals in GEM oxidation in some environments.16−25 Reactive Hg compounds of potential atmospheric significance, e.g. HgBr2, HgCl2, BrHgCl, BrHgO, BrHgOH, ClHgOH, HgO, Hg(OH)2, and the HgI compound HgBr,16,25−33 are known (or expected) to have high water solubility and low vapor pressures, suggesting that atmospheric RM species are effectively © XXXX American Chemical Society
scavenged by suspended particulate matter and precipitation, and have high wet and dry deposition fluxes.9,26,34 Conversely, with a mean lifetime of ≈100 days to 1.5 years,9,23,35,36 GEM is distributed on hemispheric to global scales from its emission sources before being appreciably oxidized in the atmosphere and deposited as RM to the surface. Atmospheric measurements provide valuable constraints on the environmental cycling of Hg,5,11,12,23,37 yet they are limited, especially for RM species. All existing RM measurements are made by “operational” methods, meaning that they lack any means of calibration. The RM concentrations measured represent only an unknown fraction of atmospheric HgII (and HgI). At present, we have no direct information on the identity of atmospheric RM compounds;13 instead, the measured RM is characterized on the basis of its phase state using sampling schemes which separate species in the gas phase (gaseous oxidized mercury, GOM) from those bound to particles (particle bound mercury, PBM). This critical limitation, coupled with the experimental difficulties of working with Received: September 27, 2012 Revised: February 7, 2013 Accepted: February 20, 2013
A
dx.doi.org/10.1021/es303916v | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
sampled air from a common 2.22 cm (7/8 in.) i.d. PFA Teflon manifold maintained at 115 °C. The manifold was used to distribute ambient and spiked air containing unknown concentrations of GEM, HgBr2, O3, and water vapor (WV) (see Supporting Information (SI) for further details). The design and performance of the manifold and spiking system are described by Finley et al.56 The mean total flow through the manifold was ≈200 L min−1 at STP (273.15 K, 1013.25 mbar; hereafter all flows are reported at STP). A Teflon-coated aluminum cyclone inlet (4.5 m above ground level) gave a 50% particle size cut of ≈1 μm at the mean flow. From 22 August to 3 September the study focused on ambient Hg measurements; from 3 to 16 September spike and interference tests were performed. Our analysis considers mostly this latter time period. DOHGS. A schematic diagram of the DOHGS instrument is shown in SI Figure S1. The design and operation of the DOHGS were discussed previously by Lyman and Jaffe.50 We provide only a brief description here; however, a complete description is provided in the SI. During RAMIX, the instrument was housed in an airconditioned trailer (≈20 °C). Ambient air was drawn from the main sample manifold through a heated (110 °C) 1.6 m, 0.95 cm (3/8 in.) i.d. PFA Teflon sample line. Two cold vapor atomic fluorescence spectrometers (CVAFS) (Tekran Instruments, Inc.; model 2537B Hg vapor analyzer) subsampled air from the sample line. The first CVAFS sampled through a quartz pyrolyzer tube, which was packed with quartz wool and heated to 650 °C to convert all sampled Hg species to GEM, thereby providing measurement of THg. The second CVAFS sampled through a quartz wool RM trap at ambient temperature and provided measurement of GEM. The difference between the THg and GEM measurements yielded measurement of RM. Hereafter we will refer to the DOHGS CVAFS separately as the “THg analyzer” and the “GEM analyzer”. Instrument Modifications. Prior to RAMIX, in order to improve signal-to-noise and thereby lower the RM limit of detection (LOD), the standard sample cuvettes and bandpass filters in the CVAFS were replaced with mirrored cuvettes and improved bandpass filters (Tekran). As a result, the CVAFS precision improved by 20−37%. During RAMIX, the RM LOD was 39% lower than the value measured during the WAMO campaign, which verifies the efficacy of the modifications made to our analyzers (see SI for further details). Because RM concentrations in the lower to middle troposphere are typically 18.0 MΩ cm). A key difference between the instrument configurations during RAMIX and WAMO was the use of a single GEM sample train during RAMIX (SI Figure S1). The RM trap was swapped for a clean one typically every 2 days. Blanking and Calibration. The instrument was blanked 1 to 4 times daily for 10 to 20 min by back-flushing the sample line with zero air, which was generated by pumping ambient (room) air through an activated carbon trap backed by a quartz wool plug and glass fiber filter (Whatman, grade GF/D, 2.7 μm). Mean THg and GEM concentrations (±1σ) during all blanking periods (n = 32) were 0.065 ± 0.046 ng/m3 and 0.051 ± 0.044 ng/m3, respectively. The absolute mean (±95% CI) difference between blanks measured concurrently with both CVAFS was 0.014 ± 0.008 ng/m3. The response factors (RF) for the CVAFS (in units of GEM fluorescence peak height/pg Hg sampled) were calibrated by standard additions (performed 3−12 times daily for 10−20 min) on ambient or zero air of GEM in ultra high purity (UHP) N2 from an external permeation source (Figure S1, see SI for further details). THg, GEM, and RM Quantification. The concentrations of THg and GEM in the RAMIX manifold were determined from the measured Hg0 vapor fluorescence peak heights and the RFs of the CVAFS.45 The corresponding RM concentrations were calculated as the difference between the THg and GEM concentrations after scaling the GEM concentrations by the analyzer bias (see SI for further details). For the period 3− 16 September, the mean (±1σ) analyzer bias was −2.1 ± 1.7% (N = 79). The overall uncertainty of the THg and GEM measurements depends primarily on the uncertainties of the RFs (1.1 to 4.5%, 1σ) and the sample flow rates of the CVAFSs (which determine the uncertainty of the sample volume). The uncertainties of the sample flow rates were
