Environ. Sci. Technol. 2001, 35, 3207-3212
A New Instrument To Measure Gaseous Nitrous Acid (HONO) in the Atmosphere JO ¨ RG HELAND,# JO ¨ RG KLEFFMANN, RALF KURTENBACH, AND PETER WIESEN* Bergische Universita¨t-Gesamthochschule Wuppertal, FB 9/Physikalische Chemie, Gaussstrasse 20, 42097 Wuppertal, Germany
A new in situ instrument (LOPAP: long path absorption photometer) to measure gaseous nitrous acid (HONO) using wet chemical sampling and photometric detection has been developed. This instrument is aimed to overcome the known problems with current HONO measurement techniques and was designed to be a cheap, sensitive, compact, and continuously working HONO monitor for ambient air measurements in the troposphere or for measurements of higher concentrations e.g. in smog chambers, in exhaust gases, and in indoor environments. Laboratory investigations were carried out to characterize the instrument components with respect to collection efficiency, optimum dye formation, optimum detection, and interfering species. Detection limits ranging from approximately 3 to 50 pptV have been obtained with response times from 4 to 1.5 min, respectively, using different instrument parameters. The accuracy of the measurements is in the range between ((1015)%. The validation of the instrument was performed in the laboratory for HONO concentrations of 3 and 30 ppbV using ion chromatography and with a DOAS (differential optical absorption spectrometer) instrument in a large outdoor smog chamber in the range from 0.1 to 20 ppbV. The deviations were well within the errors of the measurements; however, when comparing the data with the DOAS instrument systematically higher values were found with the LOPAP instrument.
1. Introduction Nitrous acid plays an important role in atmospheric chemistry because it acts as a source of OH radicals via photolysis during the day, especially during sunrise (1). In addition, HONO is an important indoor pollutant, which can react with amines leading to nitrosamines, which are known to be carcinogenic (2, 3). The mechanisms by which HONO is formed in the atmosphere are not well established. Although it is accepted that HONO is formed by heterogeneous processes, i.e., the conversion of NO2 on wet surfaces (4), it is at present not clear under which conditions HONO production is dominated by the surface of aerosols or by the ground surface. While surface flux measurements point toward a ground surface source (5), combined HONO and aerosol observations * Corresponding author phone: +49-202-4392515; fax: +49-2024392757; e-mail:
[email protected]. # Present address: Deutsches Zentrum fu ¨ r Luft- und Raumfahrt (DLR), Oberpfaffenhofen, 82234 Wessling, Germany. 10.1021/es000303t CCC: $20.00 Published on Web 07/04/2001
2001 American Chemical Society
indicate that the aerosol could be the major HONO source (6-8). Simultaneous vertical profile measurements of HONO, NO2, and the aerosol surface area, which could answer this question are not available at present. However, measurements of gaseous HONO have been made for many years with various techniques (e.g. refs 9-16) of which, to our knowledge, none has been brought to the stage of a sensitive, compact, and cheap online measurement device. For instance, fast path-averaging DOAS instruments, although very sensitive with detection limits down to 50200 pptV (8, 9), suffer from great experimental efforts and expensive system components. Dry carbonate denuder techniques often suffer from a poor time-resolution at low HONO mixing ratios and possible unquantified interferences (10). In addition, for wet effluent diffusion denuders, which are fast and sensitive instruments, it was very recently shown that they also might suffer from unknown interferences (17). In this work a new two channel in situ HONO measurement instrument (LOPAP: long path absorption photometer) is described, which is designed to be a cheap, sensitive, compact, and continuously working HONO monitor for ambient air measurements in the troposphere, for aircraft and/or balloon-borne measurements, for exhaust gas measurements, or indoor pollution monitoring. In this paper the principle of the LOPAP instrument, the instrument components, possible interferences, and an intercomparison with a DOAS instrument in a smog chamber will be addressed.
2. Experimental Section In principle each channel of the LOPAP instrument consists of the following three separate units. (1) The collection unit in which the HONO from the sample gas flow is collected continuously in a liquid phase using a glass stripping coil of 10 cm length with an inner diameter of 2 mm. The stripping solution R1 contains 0.06 M sulfanilamide in 1 M HCl which immediately forms a diazonium salt with HONO (18, 19). After a debubbler in which the air is separated from the liquid the solution is pumped with a peristaltic pump (ISMATEC, ECOline VC-MS/CA8-6) into (2) the azo dye unit in which a 0.8 mM n-(1-naphthyl)ethylenediamine-dihydrochloride solution (R2) is injected with the peristaltic pump to form the final azo dye during an appropriate residence time in a mixing volume (19, 20). Leaving this mixing volume, the solution is pumped continuously with a HPLC pump (Knauer, K120) into (3) the detection unit, which consists of a long Teflon tubing acting as an absorption cell. Visible light is focused into the tubing via fiber optics and, with the refractive index of the tubing material being lower than that of the liquid, undergoessdependent on the angle of incidencesmultiple total reflections on the inner walls of the tubing and stays inside the liquid for absorption (21). On the opposite end of this so-called liquid core waveguide, LCW, the light is collected again by a glass fiber and detected with a minispectrometer. The absorption spectra are stored on a minicomputer for later data analysis. As shown in Figure 1, two stripping coils in series are used in the LOPAP instrument. The first coil (channel 1) takes out almost all the HONO but only a small fraction of any interfering species. The second coil (channel 2) takes out about the same amount of the interfering species but has almost no HONO in it. By subtracting the calibrated signal of the second channel from the calibrated signal in the first channel a better measure of the true HONO concentration is obtained than using the signal of the first VOL. 35, NO. 15, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3207
FIGURE 1. Schematic setup of the LOPAP instrument. channel only. This setup minimizes the occurrence of possible unknown interferences as e.g. observed for a wet effluent diffusion denuder (17). The calibration of the channels is performed with a liquid nitrite standard, which is injected into R1 at known amounts or with a pure HONO source (23, 24). Most of the measurements described in this paper were performed with a prototype instrument which was designed for first measurements in a smog chamber, i.e., for HONO concentrations in the ppbV range. For sub-ppbV measurements in ambient air an improved and optimized instrument with different well-defined instrument parameters, such as absorption path length and gas and liquid flow rates, was used. In addition, the volume of e.g. debubblers and tubings was minimized for this optimized instrument.
3. Results and Discussion 3.1. Test of the Instrument Components. Collection and Azo Dye Unit. To minimize possible interferences to e.g. PAN and NO2 + SO2 it was decided to use an acidic stripping solution instead of pure or alkaline water. However, since it is known that the solubility of HONO is low under moderate acidic conditions (22), the standard reagent, R1, see above, was used to collect HONO. The formation of the diazonium salt in R1 was found to be extremely fast, leading to an almost complete sampling of gaseous HONO. After the collection of HONO in R1 the solution was further processed with R2 (19). Since it was observed that the formation of the dye was complete within 1-2 min, the mixing volume or better the volume of the tubing in which the dye is formed was adequately chosen. A HONO collection and azo dye formation efficiency of >97% for gas flow rates 1.1 L/min and vice versa. The actual HONO mixing ratios were measured with a tunable diode laser system described by Becker et al. (25) or ion chromatography (HIC-6A and SPD-6A, Shimadzu) with a preconcentration column (TAC-LP1, Dionex). Absorption spectra of the azo dye with various nitrite concentrations are plotted in Figure 2. The absorption feature has an approximate width of 100 nm centered around 544 nm. Since water spectra were used as reference spectra, the structures in the lower wavelength region below 400 nm are due to absorption of the reagents themselves. Detection Unit. The detection of the azo dye absorption in the Teflon tubings via fiber optics is performed with a two 3208
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 15, 2001
FIGURE 2. Absorption spectra of the azo dye for different nitrite concentrations. A nitrite concentration of 3.62 × 10-7 mol/L corresponds to a mixing ratio of ≈9 ppbV when a gas flow rate of 1 L/min and a liquid flow rate of 1 mL/min are used. channel grating minispectrometer using a diode array detector (SD2000, Ocean Optics). To be independent from instrumental drifts or intensity fluctuations, which are caused by e.g. temperature changes and/or bubbles of air entering the tubing no so-called background (I0) spectra, i.e., spectra without absorption of the azo dye are used for data evaluation. It was observed that the logarithm of the ratio of two intensities taken at different wavelengths in the same measured spectrum is a more stable and robust measure of the azo dye concentration. These two intensities are taken (a) at the absorption maximum at 544 nm or a neighboring point, Iabs, and (b) at a spectral point where no azo dye absorption occurs, Iref, e.g. above 700 nm. Knowing that Iref is directly proportional to the background intensity at the absorption maximum, the logarithm of (Iref/Iabs) becomes a linear measure of the concentration c according to LambertBeers law
ABS ) log(Iref/Iabs) ) kλ × l × c + log(const)
(1)
where l denotes the absorption path length and kλ is the absorption coefficient of the azo dye which is ≈5 × 104 L mol-1 cm-1 at 544 nm (19). The intercept of eq 1 depends mainly on the chosen absorption and reference wavelengths and may show variations, which are due to different purities of fresh reagents. Accordingly, zero air measurements have to be performed approximately every day during the operation of the instrument and a calibration of the two channels with a liquid nitrite standard orssince the slope of the calibration is stable within the error bars for dayssat least a zero air measurement is mandatory when the reagents have been renewed. An example of the calibration of the two channels of the prototype instrument with Iabs measured at 560 nm and Iref at 710 nm with an absorption path length of approximately 1.3 m is shown in Figure 3. The measured gas flow rate at 20 °C (Tylan Flowmeter calibrated against a Sensidyne Gilibrator-2 Base) during this calibration was (1.17 ( 0.03) L/min. The corresponding liquid flow rates of R1 and R2 were ≈1 mL/min (( 3%, 1σ) each in the two channels. The liquid calibration standard containing 0.05 mg/L nitrite was injected into R1 with an independent peristaltic pump with varying pumping speeds. The standard itself was prepared shortly before use from a 1000 mg/L nitrite standard, which is stable in 0.02 molar NaOH and which is checked frequently by titration with KMnO4 standard solutions. In the case of these measurements the absorption of the calibration of the
TABLE 1. Summary of the Parameters of the LOPAP Instrument for Two Different Applications application instrument parameters: gas flow rate liquid flow rate (R1) absorption length range of λabs measurement range response time (10-90% of full signal) detection limits
FIGURE 3. Example of a calibration of the prototype LOPAP instrument with a liquid nitrite standard during the validation measurements with a DOAS instrument. The absolute values to be added to the total uncertainty given in brackets correspond to the uncertainty of the intercept in ppbV. two channels, as shown in Figure 3, can be described by the linear dependencies
ABSCH1 - log(const.) ) (9.197 ( 0.079) 10-2 cppbV +
(0.597 ( 0.298) 10-2 (2)
ABSCH2 - log(const.) ) (11.333 ( 0.058) 10-2 cppbV +
(0.705 ( 0.197) 10-2 (3)
which were typical for the measurements during the evaluation campaign with the DOAS instrument, see section 3.4. Different slopes result from different liquid flow rates and different absorption path lengths in the two channels of the instrument. The uncertainties in eqs 2 and 3 result from the weighted linear fit procedure (26). The relative overall uncertainties for each channel of ((10-11)% are calculated from the sum of the uncertainties of (a) the gas flow rate, (b) the liquid flow rates, and (c) the uncertainty of the slope of the fit. The absolute values to be added to the total uncertainty in Figure 3 correspond to the uncertainty of the intercept in ppbV. The detection limit was calculated as three times the uncertainty of the intercept of each channel and is typically in the order of 25-50 pptV for 1.3 m absorption tubings used in the prototype setup. 3.2. Instrument Parameters. The detection limit, the response time i.e., the time the instrument needs to rise from 10% to 90% of the full signal, and the measurement range of the LOPAP instrument are strongly dependent on (a) the sample gas flow rate, (b) the liquid flow rates, and (c) the length of the absorption tubings. The theoretical sensitivity of the instrument will increase linearly with the gas flow rate and/or the absorption lengths and/or decrease with the liquid flow rates. In practice, these parameters are limited e.g. by a decreasing HONO collection efficiency with increasing gas flow rate and intensity reductions with increasing path length. The response time depends on the volume of the absorption tubing and the liquid flow rates. Since the lower limit of the measurement range is, of course, set by appropriate values of the above-mentioned parameters, the upper limit of the measurement range can also be shifted during the data evaluation by shifting the wavelength of Iabs for the analysis in the stored spectra. The relevant slopes and intercepts for the use of various absorption and reference wavelengths can be calculated from the spectra of only one calibration measurement. The highest sensitivity, i.e., the largest slope of the calibration, will be obtained when the absorption maximum of the azo dye at
smog chamber measurements
ambient air measurements
(prototype) 1.1 L/min 1.0 mL/min 1.3 m 540-590 nm 0.10-25 ppbV ≈1.5 min
(optimized instrument) 1.0 L/min 0.4 mL/min 1.9 m 540-590 nm 0.02-12 ppbV ≈4.0 min
25-50 pptV
3-6 pptV
λ ) 544 nm is used. If the spectra are saturated at the absorption maximum, the sensitivity can be reduced by shifting Iabs to the wings of the absorption band toward higher wavelengths, e.g. to λ ) 590 nm, see Figure 2. Here, the detection limits are defined as the mixing ratios calculated from three times the standard deviation of the intercept of the calibration curve. The minimum detectable change of the mixing ratios or resolution of the instrument was found to be approximately 1% of the measured value. Table 1 summarizes the measurement range, the response time, and the detection limits, for two different sets of parameters used for different applications of the instrument. 3.3. Interferences. Most interferences in this section are given as the interference signal measured as HONO in pptV divided by the mixing ratio of the interfering compound in % () 100 × [pptV-signal HONO/pptV-compound]). All values given in this section correspond to a gas flow rate of 1.0 L/min and a liquid flow rate of R1 of 0.4 mL/min at 20 °C. It is worth repeating that the first channel measures the signal resulting from HONO plus interferences and the second channel only the interferences. The difference between the channels gives a measure of the true HONO concentration. This setup minimizes the occurrence of possible unknown interferences as e.g. observed for a wet effluent diffusion denuder (17). In this section the interferences are discussed either as the interference per channel or/and the real interference, the later being calculated from the difference of the two measurement channels. Thus, the real interference is always smaller than the interference per channel. However, when the interference per channel was found to be below the detection limit of the instrument, it was conservatively assumed to be equal to the total real interference. It was found thatsdependent on the experiments and mixing ratios usedsadditional HONO was formed on the walls of the PFA-tubings before the gas inlet caused by heterogeneous reactions. In these cases careful attention was drawn on eliminating the additional HONO (but not the species themselves) from the gas flow and keeping all tubings as short as possible. NO2. The NO2 interference was measured several times with diluted NO2 calibration gas (2.19 ppmV in N2, Messer Griesheim). Figure 4 shows the NO2-interference per channel as a function of the gas flow rate for the stripping coil used in the instrument. For a gas flow rate of ≈1 L/min the interference per channel was found to be (0.06 ( 0.02)%, whereas the real interference after the subtraction of channel 2 from channel 1 was only (0.011 ( 0.005)%. To correct for the NO2 interferencesas the fraction of NO2 which is measured as HONOsof ca. 0.1% per channel for gas flow rates ≈1 L/min, a two channel instrument has to be used. Worst case estimates with a [NO2]/[HONO] ratio of roughly up to 100 in the atmosphere would therefore imply an NO2-interference signal per channel which is approximately 10% of the true HONO signal itself. VOL. 35, NO. 15, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3209
FIGURE 4. NO2 interference of the LOPAP instrument as a function of the sample gas flow rate at 293 K and 1013 hPa for a glass stripping coil of 10 cm length using a liquid flow rate of 0.4 mL/min. The actual NO2 mixing ratio was 140 ppbV. NO. The pure NO interference was measured with a NO mixing ratio of 239 ppbV diluted from a calibration gas mixture containing 10 ppmV in N2 (Messer Griesheim). This lead to a real interference of
