Article pubs.acs.org/est
Biofuel-Promoted Polychlorinated Dibenzodioxin/furan Formation in an Iron-Catalyzed Diesel Particle Filter Norbert V. Heeb,*,† Maria Dolores Rey,†,§ Markus Zennegg,† Regula Haag,† Adrian Wichser,† Peter Schmid,† Cornelia Seiler,† Peter Honegger,‡ Kerstin Zeyer,‡ Joachim Mohn,‡ Samuel Bürki,∥ Yan Zimmerli,∥ Jan Czerwinski,∥ and Andreas Mayer⊥ †
Laboratory for Advanced Analytical Technologies, and ‡Laboratory for Air Pollution/Environmental Technology, Swiss Federal Laboratories for Materials Testing and Research (EMPA), Ü berlandstrasse 129, CH-8600 Dübendorf, Switzerland § Chemical Engineering Department, University of Alicante, 03080 Alicante, Spain ∥ Laboratory for Exhaust Emission Control, University of Applied Sciences Biel (UASB), Gwerdtstrasse 5, CH-2560 Nidau, Switzerland ⊥ Technik Thermischer Maschinen (TTM), Fohrhölzlistrasse 14b, CH-5443 Niederrohrdorf, Switzerland S Supporting Information *
ABSTRACT: Iron-catalyzed diesel particle filters (DPFs) are widely used for particle abatement. Active catalyst particles, socalled fuel-borne catalysts (FBCs), are formed in situ, in the engine, when combusting precursors, which were premixed with the fuel. The obtained iron oxide particles catalyze soot oxidation in filters. Iron-catalyzed DPFs are considered as safe with respect to their potential to form polychlorinated dibenzodioxins/furans (PCDD/Fs). We reported that a bimetallic potassium/iron FBC supported an intense PCDD/F formation in a DPF. Here, we discuss the impact of fatty acid methyl ester (FAME) biofuel on PCDD/F emissions. The iron-catalyzed DPF indeed supported a PCDD/F formation with biofuel but remained inactive with petroleum-derived diesel fuel. PCDD/F emissions (I-TEQ) increased 23-fold when comparing biofuel and diesel data. Emissions of 2,3,7,8-TCDD, the most toxic congener [toxicity equivalence factor (TEF) = 1.0], increased 90-fold, and those of 2,3,7,8-TCDF (TEF = 0.1) increased 170-fold. Congener patterns also changed, indicating a preferential formation of tetra- and penta-chlorodibenzofurans. Thus, an inactive ironcatalyzed DPF becomes active, supporting a PCDD/F formation, when operated with biofuel containing impurities of potassium. Alkali metals are inherent constituents of biofuels. According to the current European Union (EU) legislation, levels of 5 μg/g are accepted. We conclude that risks for a secondary PCDD/F formation in iron-catalyzed DPFs increase when combusting potassium-containing biofuels.
■
INTRODUCTION Renewable Biofuels: Alternatives to PetroleumDerived Fuels? Our societies strongly depend upon abundant, affordable, and safe fuels to operate large fleets of on- and offroad vehicles, locomotives, ships, and construction and mining machinery. The majority of these fuels are derived from petroleum sources. Increasing shares of biofuels produced from biomass are now blended with petroleum-derived fuels or used as such. Gasoline is replaced by ethanol obtained from various plant materials, mainly sugar cane and corn. Bioethanol is widely used in the U.S.A., Brazil, and Europe. Diesel-like fuels are produced from different oil plants and to some degree from animal fat. Extracted triglycerides are combusted as such or transformed to fatty acid methyl esters (FAMEs). Sunflower and rapeseed oils and respective methyl esters replace diesel fuel in Europe, and palm, soybean, and jatropha oils are used in the tropics. The European Union (EU) launched a program to © XXXX American Chemical Society
substitute a minimum of 10% of its fuel consumption with renewable biofuels by 2020.1 In other words, substantial amounts of biofuels are consumed today, and infrastructure for their production, distribution, and handling has been developed. Engine parts in contact with biofuels must be compatible, and the long-term performance of engines and vehicles must be guaranteed. The impact of biofuels on emissions of regulated and non-regulated pollutants and exhaust toxicity is an important issue. Because current emission standards for engines and vehicles cannot be met anymore without efficient converter technologies, biofuels Received: March 3, 2015 Revised: May 29, 2015 Accepted: July 1, 2015
A
DOI: 10.1021/acs.est.5b01094 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
■
Article
EXPERIMENTAL SECTION Engine, Test Cycle, Fuels, and Lubricant. Filter testing was performed at the engine dynamometer (University of Applied Sciences Biel, Switzerland) under standardized conditions, which have been described in a Swiss Norm.12 Tests were performed on a EURO-III diesel engine (6.4 L, 116 kW, type 934 D, Liebherr, Bulle, Switzerland). The eight-mode ISO 8178/4 C1 cycle for construction engines was applied twice per test configuration (total sampling time of 200 min). Figure S1 of the Supporting Information displays the torque− revolutions per minute (rpm) diagram and exhaust temperatures before and after the DPF. The engine was operated with commercial, low-sulfur petroleum-derived diesel fuel (SN EN 590, class 0) and rapeseed FAME biofuel (EN 14214:2008). Tables S1 and S2 of the Supporting Information report respective fuel characteristics. Both fuels were doped with 1,6dichlorohexane (Fluka, Buchs, Switzerland) to chlorine levels of 10 μg/g. The lubricant (Lubrizol Blue, 15W40) had a chlorine content of 120 μg/g, as determined by wavelength-dispersed Xray fluorescence spectrometry. Particle Filters, Catalysts, and Test Configurations. Three identical, non-coated silicon carbide wall-flow filters (DiSiC A3, 100 cpsi, 22.8 L, DINEX, Denmark) were tested in different fuel and catalyst combinations and compared to engine-out emissions, which were chosen as reference points. Table S3 of the Supporting Information lists experimental conditions of all test configurations. Two blends of the ironbased FBC (Satacen-3, INNOSPEC, Cheshire, U.K.), one with diesel and the other with biofuel, were prepared together with a third blend of an iron/potassium FBC (ITN, Krakow, Poland) with diesel fuel. Iron concentrations of 85, 77, and 69 μg/g and potassium concentrations of 98% for particles in a size range of 20−300 nm in new and aged conditions (>1800 h field operation). If applied on construction machinery, they must not exceed particle number emissions of 1 × 1012 particles/kWh. In addition, filters must reduce emissions of carcinogenic compounds and must not support the formation of toxic secondary pollutants,10 an issue also addressed in the U.S. Clean Air Act.13 A critical issue to be assessed is the PCDD/F formation potential of filters. In principle, any efficient filter should remove PCDD/Fs as it does for other aromatic hydrocarbons. Filters also reduce emissions of carcinogenic PAHs by 31−87 and 75−98% for low- and high-oxidation potential filters, respectively.9 Thus far, four catalytic DPFs have been identified that supported a PCDD/F formation.14−17 Three of these active filters used copper-based fuel-borne catalysts (FBCs).14−16 The catalytic effect of copper has been recognized before in filters that accumulate municipal waste incinerator ash.18−27 The use of copper-catalyzed DPFs has been prohibited in Switzerland, in 1998 at tunnel construction sites and in 2009 on all construction machinery.10 Recently, we reported that a bimetallic iron/potassium FBC induced a substantial PCDD/F formation in a DPF, whereas the iron-only catalyst did not support this chemistry.17 The extent of PCDD/ F formation and the observed pattern changes were comparable to active copper-catalyzed filters.15,16 Apart from these four filters, all other tested and approved systems did not catalyze a PCDD/F formation. These investigations were performed with petroleum-derived diesel fuels following prescribed protocols.12 When realizing that biofuels may contain up to 5 μg/g of potassium in accordance with current fuel specifications, we hypothesized that the combustion of such biofuels may alter the PCDD/F formation potential of thus far inactive iron-catalyzed DPFs. Such filters are widely used in Europe in thousands of vehicles. Here, we report on effects of a FAME-based biofuel on PCDD/F emissions and patterns of an Fe-catalyzed DPF. Results are compared to those of an inactive Fe-catalyzed filter and an active Fe/K-catalyzed DPF, which were exposed to diesel fuel exhausts. B
DOI: 10.1021/acs.est.5b01094 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
blended with diesel fuel to levels of 69 and 46 μg/g, respectively. Effects of these catalysts on emissions of regulated pollutants were minor too (see Table S4 of the Supporting Information), but these catalysts also add to the ash deposits in DPFs. It has to be mentioned that the combustion of FBCs induces the formation and release of large numbers of potentially harmful metal oxide nanoparticles. Therefore, any application of FBCs without efficient filters is prohibited in Switzerland. In conclusion, the iron- and iron/potassium-catalyzed filters tested with diesel fuel or biofuel efficiently removed solid particles but had only minor effects on other regulated pollutant emissions. Fuel, Catalyst, and Filter Effects on PCDD/F Emissions. PCDDs and PCDFs are two classes of compounds that form during various combustion processes but are also contaminants in certain pesticides.30 At most, 75 PCDD and 135 PCDF congeners have to be expected, including the 17 toxic 2,3,7,8substituted isomers. Figure 1 displays their structures and
J&W, Folsom, CA) gas chromatography (Varian 3400, Orlando, FL) coupled with high-resolution mass spectrometry (MAT 95, Bremen, Germany). PCDD/F congeners could be detected in all samples. Mean (n = 5) engine-out PCDD/F emissions (diesel fuel, without metal dopant) were 3−10-fold higher than blank samples. These data were used as reference points but not corrected for blank values because of increased uncertainties of the latter. However, both data sets are reported in Table S6 of the Supporting Information, allowing for corrections if appropriate.
■
RESULTS AND DISCUSSION Impact of Fuels, Filters, and Catalysts on Regulated Pollutant Emissions. Diesel engines are increasingly operated with biofuels, which are applied as such or blended with diesel fuels. Tables S1 and S2 of the Supporting Information report characteristics of the FAME biofuel and the petroleum-derived diesel fuel used in this study. They differ in physicochemical properties, such as density, viscosity, and flash point, but were both within the EN specifications. The biofuel was produced from rapeseed triglycerides via transesterification reactions. Its chemical composition, especially its oxygen, sulfur, and phosphorus contents, and the levels of impurities, such as glycerol and various glycerides, deviate considerably from the diesel fuel. The oxygen content and levels of inorganic impurities, such as alkali and alkali earth metal ions, are affected by the biofuel production process. However, both fuels could be used without adaptations with the EURO-III engine (6.4 L, 116 kW, Liebherr, Switzerland). Figure S1 of the Supporting Information displays the eight-mode ISO 8178/4 C1 cycle, and Table S3 of the Supporting Information lists all of the fuel, filter, and catalyst configurations examined. Despite substantial differences in chemical composition of biofuel and diesel fuel, only minor effects were observed for the regulated pollutant emissions (see Table S4 of the Supporting Information). Engine-out THC and CO emissions were reduced from 0.39 and 1.03 g/kWh to 0.14 and 0.72 g/kWh when using biofuel instead of diesel fuel. NOx and CO2 emissions were not affected. They varied in narrow ranges of 6.12−6.22 and 792−799 g/kWh, respectively. Three identical, VERT-approved, silicon carbide wall-flow filters (DINEX, DiSiC A3) were tested, each in one specific catalyst/fuel combination (see Table S3 of the Supporting Information). These filters reached particle number filtration efficiencies of >98% in a size range of 20−300 nm. While soot and ash particle emissions were lowered substantially by the filters, gaseous pollutants were not affected, varying within 10%, with the exception of the iron-doped diesel fuel, which induced a 20% decrease of the THC emissions after filtration. On the basis of higher concentrations of ash-forming impurities of the biofuel, one expects, on a long-term, higher ash burdens accumulating in filters. Phosphorus and sodium levels were
