Application for a Newly Developed High-Capacity NO_x Denuder: Low-NO_x Diesel Transformation Experiments

ABSTRACT

To conduct low oxides of nitrogen (NO_x) chamber experiments with modern diesel emissions (DE), a high-capacity NO_x denuder was developed and used at the European Photoreactor (EUPHORE) outdoor simulation chamber. The denuder displayed a sufficient NO_x storage capacity for use with DE, and efficient removal of NO_x during injections of DE was achieved (>98%). Degradation of the denuder performance after repeated regeneration by heating (400 °C) and flushing with an air/oxygen ratio of 2:1 was not observed for a total of nine experiments. Evaluation of dark (with chamber cover closed) experiments (four in total) with and without the denuder in-line revealed some reduction (22%) of diesel particulate matter (DPM) with use of the denuder, most likely a result of impaction or settling of DPM during DE transit. However, DPM reduction may have also been a result of reductions in effective load of the engine-dyno system during the DE injections. Extensive chemical characterization of DPM revealed no significant perturbation of major compound groups associated with denuder use, except for nitrated polyaromatic hydrocarbon (NPAH) concentrations. The implications of high-NO_x experiments without the use of a NO_x denuder are discussed.

IMPLICATIONS

Achieving DPM mass concentrations appropriate for chemical analysis and toxicity evaluations in DE transformation experiments requires injection of a relatively large volume of modern DE into a chamber of known volume, leading to an unrealistically high mixing ratio of NO_x (mostly as nitric oxide). The high NO_x levels result in chamber atmospheres that are photochemically inhibited and therefore not relevant to transformations that may occur in the ambient atmosphere during aging. Successful application of a newly developed high-capacity NO_x denuder for DE transformation experiments was achieved at the EUPHORE outdoor simulation chamber. Chemical characterization of experimental samples with and without the denuder in-line and the implication of high-NO_x chamber atmospheres for NPAH concentrations are discussed.

INTRODUCTION

Modern light-duty diesel engines emit relatively low vapor and particulate emissions in comparison to older engine models.1 At the same time, oxides of nitrogen (NO_x, or nitric oxide [NO] + nitrogen dioxide [NO₂]) emissions from these engines have not changed significantly (frequently in the range of 300–500 parts per million [ppm]), which presents a challenge when performing chamber experiments based on pseudoambient dilution ratios (≥300–400:1). To achieve diesel particulate matter (DPM) mass concentrations appropriate for chemical analysis and toxicity evaluations, a relatively large volume of modern diesel emissions (DE) must be introduced into a chamber of known volume, leading to an unrealistically high mixing ratio of NO_x. In such conditions, the formation of major atmospheric oxidants (hydroxyl [OH] radical, ozone ]O₃], nitrate [NO₃] radical) will be extremely low, with rapid removal from the system through reactions with NO to form NO₂ and nitric acid (HNO₃). The photochemical transformation of organic compounds resulting from primary emissions would be effectively shut down, making experiments carried out in sunlight inefficient.2 Therefore, a NO_x removal (denudation) technology is crucial to carry out low-NO_x experiments while achieving necessary in-chamber DPM concentrations.

This study of atmospheric transformations of DE3 pursued the development and optimization of a NO_x denuder that would be able to reduce high NO_x concentrations before chamber infusion. The study was carried out in 2005/2006 at the European Photoreactor (EUPHORE) outdoor simulation chamber in Valencia, Spain.4–6 The aim was to evaluate changes in chemistry and toxicity resulting from the aging of DE in different chamber atmospheres. To achieve in-chamber mass concentrations often observed in urban environments (30 μg m⁻³) and necessary for these analyses, the required direct injection of DE resulted in high-NO_x conditions (1–2 ppm).

The development of diffusion denuders for the capture of gases has become common practice in the atmospheric chemistry community.7–11 Despite differences in design, the fundamental principle of diffusivity is used to separate gases from particles. The relatively high diffusivity of gases (D _(NO2) = 10 cm²/min) versus particles (D _(1μm) = 1.6 × 10⁻⁵ cm²/min) allows for capture of gases on appropriately coated denuder walls during effluent (DE) transit. In recent years, researchers12–18 have used cobalt oxide as an absorption material for the capture of NO_x and HNO₃ from exhaust streams. Cobalt oxide coatings can be regenerated by heating and flushing with air or oxygen at 400 °C, resulting in the release of absorbed NO_x, thus allowing the material to be used again. Previous researchers have used cobalt oxide coatings on the walls of cylindrical or rectangular channel denuders or on ceramic granules that surround screen tubes. This paper focuses on the application of a newly developed high-storage-capacity NO_x denuder suitable for use in chamber experiments with a cobalt-oxide-coated absorbent material. Results from experiments at EUPHORE that are pertinent to NO_x denuder performance will be discussed. Comprehensive experimental descriptions and results of the project are provided elsewhere (see refs 3 and 19).

EXPERIMENTAL PROCEDURES

Initial experiments with a new modern diesel engine in January 2005 (Ford, 1.8-L Lynx V277 90PS Stage 3, Delphi Fuel System, Fixed Geometry Turbo used in Ford Focus and Transit Connect automobiles) revealed extremely low vapor/particulate emissions and the presence of relatively high NO_x (400–500 ppm) in the exhaust stream. Development experimentation from February 2005 through March 2006 led to the construction of a cordierite honeycomb denuder followed by a high-capacity annular denuder, which is the focus of this article. A brief description of the annular denuder used at EUPHORE in 2006 is provided below. Although the denuder development process is not the focus of this article, data and material logistics provided by initial miniature denuder experiments were important for successful application on larger scales during the May/June 2006 EUPHORE campaign. Detailed descriptions of the experimental development process and results are provided in Zielinska et al., 2010.3 A diagram describing the EUPHORE study components is provided in Figure S1 of the supplemental material (<published at http://secure.awma.org/onlinelibrary/samples/10.3155-1047-3289.61.3.319_supplmaterial.pdf).

A firebrick prerequisite material (with an industrial reference name of GROG) composed of silica (∼50%), alumina (∼40%), iron oxide (∼2%), titania (∼2%), and several other earth metals (sodium, potassium, etc.) was selected as an absorbent substrate (Smithe-Sharpe). The GROG was coated with cobalt oxide (Co/GROG) and used to surround the perforated channels of the denuder. A stainless steel (304) perforated tube with 24-gauge (∼1 mm) apertures was used for denuder channels (Perforated Tubes, Inc.). This spiral-seamed tubing is 167 cm in length (66 in.) and has a 2.54-cm (1-in.) outer diameter. The channels are straight cylinders and the configuration allowed for a minimum of 1.5 cm of Co/GROG space between all channels (descriptive figures provided in supplemental materials). A mixing chamber of 30.5 cm precedes the packed channel section. Dimensions of the effluent channel are expanded to approximately 9 cm before the connection to the main denuder body (Figure S2). The output section consists of a 30.5-cm hollow section. Several alignment plates (at least three) inserted within the channel section ensure proper distribution of channels (Figure S3). Because the Co/GROG has a relatively high density (1.28 g cm⁻³), embedded perforated tube channels must be aligned before filling. The total volume of the packed section (after subtraction of the hollow channel sections) is 132 L. This indicates a total weight of 169 kg for just the Co/GROG inside of the denuder body. The denuder dimensions are 167 cm (66 in.) in length for the packed section, with a 36.8-cm (14.5-in.) internal diameter and a total height of 2.3 m for the main body (includes mounting brackets). It contains 57 diffusion channels (1-in. outer diameter) spaced approximately 2.5 cm (1 in.) apart. To deal with the large mass, the denuder body was mounted in brackets with wheels for easy mobility and placement in the EUPHORE facility. Once the denuder assembly was completed, the nearest possible proximity was established to the chamber injection valve, which controls DE introduction to the chamber. Insulated copper tubing (2.5 cm) was used to connect the valve-denuder-chamber system. The NO_x denuder was regenerated between runs by heating to 400 °C with an established air/oxygen (2:1) mixture flow for 3–4 hr and was cooled overnight (12–14 hr) to 80 °C, which allowed for the next DE injection with efficient NO_x removal. Chamber NO_x and particle measurements were made via chemiluminescence (Eco Physics 700 AL series, Teledyne API 200AU), and scanning mobility particle sizer (SMPS; TSI-3936), respectively.20 See Zielinska et al.3 for more detail. For experiments without the NO_x denuder, assuming a conservation of NO_x and an approximate 204.5-m⁻³ chamber volume, an exhaust flow of 129 L/min (±4%, n = 4) was calculated by using the engine-out (Horiba Analyzer) and chamber NO_x mixing ratios immediately after injection. Direct mass flow measurements during several 2006 experiments indicated a reduction of injection flow to approximately 50 L/min (±10%) with the denuder in-line.

The nitrated polyaromatic hydrocarbon (NPAH) samples were collected from the chamber at the end of the exposures using an XAD-4 (Aldrich Chemical Company, Inc.) coated annular denuder followed by a 90-mm Teflon-impregnated glass fiber filter and an XAD-4 cartridge with a 100-standard L/min flow.21 The XAD denuder strips the gas-phase species by molecular diffusion out of a laminar flow stream before collection of the particles on the second stage (filter/cartridge).22The gas chromatography (GC)/mass spectrometry (MS) analytical method required a 30-m DB-17MS column (J&W Scientific). Isolation of the NPAH-rich fraction was achieved with the use of a solid-phase extraction (SPE) aminopropyl Sep-Pak cartridge (Waters) followed by normal-phase high-performance liquid chromatography fractionation (with a semipreparative Chromega-bond amino/cyano column). Negative ion chemical ionization (NICI) with methane as a reagent gas was utilized to enhance detection sensitivity. Limits of quantitation (signal/noise ratio [S/N] > 10) for NPAHs range from 1 to 10 pg with the NICI GC/MS technique.23 All analyses were performed with a Varian 1200 triple quadrupole GC/MS/MS system. MS/MS confirmation of compounds was achieved by evaluating molecular fragments (M – 30 amu) after collision-induced fragmentation (argon collision gas). Table S1 in the supplemental material provides a list of the speciated NPAH compounds. For extensive discussion of characterization, formation, and degradation of NPAHs in the EUPHORE experiments, see Samy.24

RESULTS AND DISCUSSION

Initial conditions for a subset of experimental runs performed at the EUPHORE chamber in May through June 2006 with and without the NO_x denuder can be viewed in Table 1. The runs are listed in chronological order (between May 26 and June 13, 2006) with engine-out NO_x, DE injection time, in-chamber NO_x, initial DPM concentrations (assuming 1-g cm⁻³ particle density using SMPS data), and median and mean particle diameters. The NO_x denuder was regenerated between runs. Degradation of the denuder performance after repeated regenerations was not apparent (total of nine experiments). It can be concluded based on the 2006 data that the newly developed denuder removes NO_x very efficiently with sufficient storage capacity. For example, the NO_x concentrations in the chamber were in the range of 9–50 parts per billion [ppb] for 20- to 27-min injection times (1000- to 1350-L injection volumes). Without the NO_x denuder (for equivalent times), these concentrations would be in the range of 1–3 ppm (Table 1). Thus, the denuder achieved greater than 98% removal efficiency when evaluated on a mass flow basis (see discussion below on injection flow estimates). Some perturbation of particle distributions with the denuder in-line was observed (Table 1). For example, the median particle diameter increased from approximately 60 to 90 nm with the denuder in-line. The increased injection and transit times may be partially responsible for this shift (i.e., additional connective plumbing allowing more time for the small particles to coagulate).

Table 1. Results from EUPHORE experiments with (May 30 and 31, 2006) and without denuder (May 26 and June 13, 2006) for DE injections

Experiment NOx Denuder Usage	Engine-Out NOx (ppm)	Injection Time (min)/Volume (L)	In-Chamber NOx (ppm)	In-Chamber DPM (μg m^−3)	Median Diameter (nm)	Mean Diameter (nm)	Injected DE Mass Concentrations (μg L^−3)
DE without denuder	430	6 (774) ^a	1.7	33	62	71	8.7 ^b
DE without denuder	410	10 (1290)	2.6	60	61	69	9.5
DE with denuder	390	20 (1000)	0.009	30	75	84	6.1
DE with denuder	415	27 (1350)	0.050	54	88	100	8.2
Notes:
^aValue in parentheses represents total injection volume using estimated flows of 129 (without) and 50 L/min (with);
^bCalculated using in-chamber DPM multiplied by total chamber volume (204.5 m^3) divided by injection volume.

The DPM concentration in the injected effluent stream was reduced because of NO_x denuder usage. Table 1 provides the injected DE mass concentration, which displays an average 22% reduction of DPM concentration with the denuder in-line. This loss may be due to particle deposition or impaction during effluent transit. However, the DPM concentration depends on the engine loading, and small differences in the loading over the course of the DE injection can make a significant difference for DPM values and size distributions.25–27 The EUPHORE dynamometer loading tends to decrease over the course of an injection, so a longer injection time results in a lower-than-predicted particle concentration. See Figure S4 of the supplemental material for an example of the decrease in torque (Nm) applied by the dynamometer during injections. The longer injection times in 2006 along with the reduced loading trend displayed during the injections may partially explain the lower DPM concentrations with the denuder in-line. In addition, the injection valve system at EUPHORE depends on a positive pressure split flow regime; addition of the denuder (including transfer lines) does reduce the injection flow rate into the chamber (i.e., creates additional back pressure at the split valve). This may result in some particle impaction loss because of increased turbulence. The lower injection flow rate resulted in longer injection times to achieve required chamber mass loadings and the increased injection times resulted in greater decreases in loading during injections, which in turn produced lower DPM concentration outputs by the dyno-engine system.

To determine if use of the NO_x denuder significantly perturbed the chemical composition of the emissions, concentrations for several compound groups (e.g., polyaromatic hydrocarbons [PAHs], NPAHs, alkanes, ho-panes, steranes) that were quantified for the EUPHORE study3 were examined and are available in Table S2 of the supplemental material. All compound group concentrations for experiments with (May 30 and 31, 2006) and without denuder (May 26 and June 13, 2006) did not display a significant difference, with the exception of gas-phase NPAHs (see Table S2 for a full list of compounds). For instance, Table 2 displays the concentrations of the most abundant NPAHs in primary DE—1-nitronaphthalene (1NN) and 1-nitropyrene (1NP).1 1NN was predominately found in the gas phase, and the concentrations with (0.30 ± 0.27 μg [mg elemental carbon{EC}⁻¹) versus without the NO_x denuder (1.70 ± 0.02 μg [mgEC]⁻¹) were significantly lower.

Table 2. Percent difference values for NPAHs in μg (mgEC) ⁻¹

	Low NOx			High NOx
Compound/Group	May 30, 2006, D-1a	May 31, 2006, D-1b	Average with Denuder ^a	May 26, 2006, D-1c	June 13, 2006 D-1d	Average without Denuder ^b	Percent Difference ^c
1NN ^d	0.10	0.50	0.30 (±0.27)	1.72	1.68	1.70 (±0.02)	82
1NP ^e	0.11	0.02	0.07 (±0.06)	0.05	0.16	0.11 (±0.08)	38
NPAH (particle)	0.36	0.13	0.25 (±0.17)	0.65	0.28	0.46 (±0.26)	46
NPAH (gas)	0.11	0.40	0.26 (±0.20)	2.32	2.26	2.29 (±0.04)	88
Notes: “Particle” implies particle-phase totals, “gas” implies gas-phase totals.
^aValues with denuder are averaged for May 30 and 31, 2006;
^bValues without denuder are averaged for May 26 and June 13, 2006;
^cPercent difference =[(without denuder – with denuder)/without denuder]× 100;
^dGas + particle phase;
^eGas + particle phase.

Implications of High NO_x

The high concentration of gas-phase NPAHs in experiments without the denuder (i.e., high-NO_x conditions) may be due to in-chamber formation reactions or sampling artifacts.28–30 The sum of gas-phase NPAH concentrations (Table 2) are approximately an order of magnitude higher for high-NO_x experiments (2.29 ± 0.04 μg [mgEC]⁻¹) versus low-NO_x (0.26 ± 0.20 μg [mgEC]⁻¹) experiments, mostly because of the difference in 1NN. However, in-chamber formation (gas-phase) is unlikely because of the lack of OH or NO₃ radicals within the chamber atmosphere. A surface-catalyzed electrophilic nitration formation pathway has been suggested from past studies.31,32 1NP has been shown to form on chamber walls and/or on sampling media in past experiments,33 and the impact of NO_x levels on heterogeneous chemistry needs to be considered when interpreting results.14 Thus, it is possible that nitration occurred on the XAD-denuder surface used for gas-phase sample collection.29,30,34

Although the average particle-phase NPAHs are nearly 2 times higher in high-NO_x experiments without the denuder, the variability between the replicate experiments excludes any significance (Table 2). The most abundant NPAHs in this study include 1NN and 1NP, which is consistent with past studies of primary DE1 and/or formation from electrophilic nitration. A simple evaluation of the 1NN-to-naphthalene (the parent PAH) ratio does also indicate a significant difference (0.27 ± 0.01 in high-NO_x experiments and 0.05 ± 0.03 in low-NO_x experiments), which rules out selective denudation of naphthalene in low-NO_x (with denuder) experiments. With the absence of photodecomposition (i.e., all dark experiments discussed above), which has been shown to be the main loss pathway in the atmosphere, primary NPAHs (and any NPAHs formed) would be expected to persist within the chamber atmosphere until sample collection.2 For a more extensive discussion of NPAH formation and degradation in the EUPHORE experiments, see Samy.24

CONCLUSIONS

Simulation chamber experiments conducted with a newly developed high-capacity NO_x denuder displayed no performance degradation, and the regenerative property of the Co/GROG absorbent material in this study was sufficient for multiple-use scenarios. The high-capacity NO_x denuder successfully captured and retained NO_x with a sufficient efficiency (>98%), which allowed low-NO_x experiments with modern DE at EUPHORE in 2006. However, evaluation of the saturation capacity of the NO_x denuder was not possible under these experimental conditions because of maximum injection times of 35 min, which resulted in only trace concentrations of NO_x in the chamber (25–30 ppb).

The evaluation of these dark experiments in high-and low-NO_x conditions revealed a significantly higher concentration of gas-phase NPAHs (mostly because of 1NN) in high-NO_x experiments. Electrophilic nitration on chamber surfaces or sampling media cannot be ruled out as a possible mechanism for the elevated NPAH concentrations. This newly developed high-capacity denuder is an important component in the successful experimentation with modern DE in atmospherically relevant conditions.

ACKNOWLEDGMENTS

The authors acknowledge Dr. Fred Rogers, Larry Sheetz, and Rick Purcell for their contribution throughout the experimentation and development stages. In addition, staff at the Organic Analytical Laboratory at the Desert Research Institute and the Fundación Centro de Estudios Ambientales del Mediterráneo were very helpful in providing needed support at crucial moments. The Health Effects Institute funded this work.