lable at ScienceDirect

Atmospheric Environment 139 (2016) 167e175

Contents lists avai

Atmospheric Environment

journal homepage: www.elsevier .com/locate/atmosenv

On-road measurement of NH3 and N2O emissions from a Euro Vheavy-duty vehicle

Ricardo Suarez-Bertoa a, *, Pablo Mendoza-Villafuerte a, **, Pierre Bonnel a,Velizara Lilova b, Leslie Hill b, Adolfo Perujo a, Covadonga Astorga a, ***

a European Commission Joint Research Centre Ispra, Institute for Energy and Transport, Sustainable Transport Unit, 21027 Ispra VA, Italyb HORIBA Europe GmbH, Emission Engineering, Automotive Test Systems, Hans-Mess-Str. 6, 61440 Oberursel, Germany

h i g h l i g h t s

� NH3 and N2O on-road emissions can be measured using a QCL-IR.� Substantial NH3 and N2O emissions were observed.� Current regulatory emissions evaluation for HD could be applied for NH3 and N2O.

a r t i c l e i n f o

Article history:Received 20 October 2015Received in revised form22 April 2016Accepted 25 April 2016Available online 26 April 2016

Keywords:AmmoniaNitrous oxideVehicle emissionsPEMSQCL-IRReal driving emissions

* Corresponding author.** Corresponding author.*** Corresponding author.

E-mail addresses: ricardo.suarez-bertoa@jrc.ec.eupablo.mendoza-villafuerte@jrc.ec.europa.eu (P. Mendastorga-llorens@jrc.ec.europa.eu (C. Astorga).

http://dx.doi.org/10.1016/j.atmosenv.2016.04.0351352-2310/© 2016 The Authors. Published by Elsevier

a b s t r a c t

The use of selective catalytic reduction systems (SCR) to abate NOx vehicular emissions brings newconcerns on the emissions of the byproducts NH3 and N2O. Therefore, NH3 and N2O on-road emissionsfrom a Euro V truck equipped with a SCR were measured in real time using a QCL-IR. Results bring tolight possibility to perform this kind of real time measurements for other pollutants besides, hydro-carbons, NOx, CO and CO2. The capability to measure NH3 and N2O in a second-by-second basis will allowapplying the currently agreed regulatory emissions evaluation for gaseous compounds. Average N2Oemission factors calculated applying the current PEMS-based data analysis to all available windows fromthe tests ranged from 0.063 g/kWh to 0.139 g/kWh. Average NH3 concentrations ranged from 0.9 ppm to5.7 ppm. Although calculated average N2O and NH3 emissions were within current limits, NOx emissionswere substantially higher than Euro V limits under the studied conditions.© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND

license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

According to the latest edition of the European EnvironmentAgency’s (EEA) annual report, European policies have led to anoverall improvement on air quality. However, the transport sectorstill generates harmful levels of air pollutants. Air pollution is animportant environmental health hazard, resulting in health prob-lems for the population and high costs for health care systems(European Environmental Agency, 2014). The population living in

ropa.eu (R. Suarez-Bertoa),oza-Villafuerte), covadonga.

Ltd. This is an open access article u

most of the European cities is exposed to pollution levels deemedunsafe by the World Health Organization (WHO). In fact, the recentGlobal Burden of Disease study indicates that in western, centraland eastern Europe (the WHO European Region), half million pre-mature deaths could be attributed to exposure to ambient airpollution in 2012 (World Health Organization, 2013). More than95% of the urban population is exposed to unsafe levels of pollut-ants such as: particle matter (PM), NOx and ground-level (tropo-spheric) ozone (O3). In addition to causing premature death, airpollution has both long- and short-term health effects, increasingthe incidence of a wide range of respiratory and cardiovasculardiseases. Alongside health, these pollutants also have a significantimpact on plant life and ecosystems. The European EnvironmentAgency has stated that, at present, PM and ground-level O3 areEurope’s most problematic pollutants in terms of harm to humanhealth, followed by nitrogen dioxide (NO2). In terms of damage to

nder the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Table 1Vehicle features.

N3 vehicle

Emissions category Euro VEngine type 11 ltMax. power 343 kWPayload 50%Fuel DieselETC ref work 55 kWhAfter treatment DPF þ SCR

R. Suarez-Bertoa et al. / Atmospheric Environment 139 (2016) 167e175168

ecosystems, the most harmful air pollutants are ground-level O3,ammonia (NH3) and NOx.

The transport sector is the largest contributor to NOx emissionsand an important source of volatile organic compounds VOCs. PM iseither directly emitted to the atmosphere (primary PM), or formedin the atmosphere (secondary PM). The main precursor gases forsecondary PM are SO2, NOx, NH3 and VOCs. These compounds reactin the atmosphere to form ammonium, sulphate compounds, andnitrate compounds. They account for up to 40% of the PM2.5 in someEuropean cities (Sillanp€a€a et al., 2006). PM2.5 emissions accountedfor up to 78% of the reported PM10 emissions from transport in 2012(European Environmental Agency, 2014). Numerical simulationsperformed by the Institut National de l’Environnement Industriel etdes Risques (INERIS) and measurements of the chemical composi-tion of PM showed that ammonium nitrate was a main contributorto an air pollution episode that affected France, southern UnitedKingdom, Belgium, the Netherlands and Germany in 2014(European Environmental Agency, 2014). Ammonium nitrateresulted from the chemical reaction between NH3, whose emis-sions were due to agricultural fertilizer spreading during thisperiod, and NOx emissions coming from road traffic. Besides beingone of themajor precursors of PM2.5, NH3 emissions are also relatedwith the hypertrophication of water and acidification of soils(Bouwman et al., 2002; Erisman et al., 2003; Sutton et al., 2000).Hence, an increase of NH3 emissions due to introduction and use ofsome emission abatement technologies (e.g., selective catalyticreduction) may lead to a harsh detriment of urban air quality.

N2O is a well-known powerful greenhouse gas. It has 298 timesthe global warming potential of CO2 over 100 years. Furthermore,N2O is nowadays the single most important ozone-depleting sub-stance (ODS) (Ravishankara et al., 2009).

PM, NO2 and O3 are considered to be the Europe’s most prob-lematic pollutants in terms of harm to human health, and the roadtransport sector is considered to amajor contributor to PM and NOxpollution in the European cities, where the density of population ishigher (80% of the European population are city dwellers) (Eurostat.Statistics, 2015). For that reason, the European legislation hasmainly aimed at reducing PM and NOx emissions from vehicles,with the introduction of Euro V/5 and Euro VI/6 emissions stan-dards. In order to comply with the new standards, new after-treatment systems have been developed and are now usedamong the cars and trucks fleets. While diesel particle filters (DPF)are commonly used to reduce direct PM emissions, selective cata-lytic reduction systems (SCR) are one of the preferred technologiesto abate NOx. The use of SCR brings new concerns on the emissionsof pollutants, and air pollution because diesel vehicles, which areequipped with this system, will emit unknown amounts of NH3 andN2O, besides NO, NO2, CO2, CO, hydrocarbons and PM.

As the development of new technologies shape the new realityfor the efficiency of internal combustion engines, clean diesel en-gines and vehicles remain a key to achieve the ambitious objectivesset by the EU for the reduction of greenhouse gas and pollutantemissions at the horizon 2020.

Changes in the perspective of the European citizen towards thepollution produced by their means of transportation, combinedwith the lack of significant changes on the measured concentra-tions of pollutants in the cities, have succeeded in the legislatorimposing more restrictive emissions limits. It also has encouragedthe introduction of Portable Emissions Measurement System(PEMS) and their use in the measurement of second-by-secondemissions during on-road operation for heavy-duty vehicles as ameans to ensure compliancewith the legislation over the useful lifeof the engines (European Commission, 2011).

The heavy-duty Euro VI Regulation (EC) No 595/2009 (EuropeanCommission, 2011) and its implementing Regulation (EC) 582/2011

introduced a procedure for testing using PEMS as a mandatory partof the type approval legislation in order to check the conformity ofheavy-duty engines with the applicable emissions certificationstandards during the normal life of those vehicles: this is the so-called “In-Service Conformity” (ISC) requirement. In addition,Euro VI vehicles also have to comply with “off-cycle” in-use emis-sions (i.e., emissions measured during tests different to the type-approval cycles, such as the World Harmonized Transient Cycle,WHTC, used for the Euro VI) requirements already at type approvalthrough the PEMS demonstration test. Regulation (EC) 582/2011also describes the provisions for vehicles type approved underdirective 2005/55/EC (Euro V) to be able to be tested using PEMS.

NH3 and N2O emissions are not regulated for Euro V vehicles inEurope. Nonetheless, NH3 has a regulatory limit of 10 ppmweighted average over the type approval cycle applicable for EuroVI (European Commission, 2011). Furthermore, U.S. EnvironmentalProtection Agency (EPA) has recently developed a GHG emissionsprogram, under the Clean Air Act, that includes N2O emissionstandards of 0.10 g/bhp-h (0.133 g/kWh) measured over the Heavy-Duty Engine FTP cycle (U.S. Environmental Protection Agency,2015). This limit is applicable in the US since 2014 for compres-sion ignition heavy-duty engines and from 2016 for spark ignitionengines.

The purpose of the present study was to demonstrate thecapability to measure NH3, N2O emissions second-by-second andduring on-road operation from HDVs, as currently done for theother gaseous pollutants using PEMS. The second-by-secondmeasurement during on-road operation allows analyzing emis-sions on different driving conditions with real on-road variables(e.g. traffic, different altitudes, weather, etc.). The possibility to doso opens the door to a more detailed observation of the differentfactors that affect the emissions while vehicles are being used indifferent urban, rural or motorway locations and conditions, whichis of major importance. Furthermore, being able to measure thesepollutants in a second-by-second basis will also allow applying toNH3 and N2O the currently agreed regulatory emissions evaluationmethods described in Reg (EC) 582/2011. Hence, we present acomprehensive analysis of the on-road obtained data, NOx, NH3and N2O, in different ways reflecting the effect of the boundarieswith which current legislation is limiting the PEMS based test. Thisseemed important as current boundaries have been proven to leavenon-negligible emissions out of the analysis (Perujo Mateos DelParque and Mendoza Villafuerte, 2015).

2. Experimental

2.1. Analytical instrumentation

In the present work, real time on-road exhaust emissions from aEuro V heavy-duty diesel vehicle equipped with a SCR, urea-basedDeNOx system, has been comprehensibly studied by combiningmeasurements performed using PEMS instrumentation, GPS sys-tem, ECU data acquisition tool and a quantum cascade laser infra-

R. Suarez-Bertoa et al. / Atmospheric Environment 139 (2016) 167e175 169

red spectrometer (QCL-IR). The vehicle was tested on-road usingthree different routes and two starting conditions, cold and hotstart. Details of the vehicle are summarized in Table 1 and thegeneral setup of the instrumentation in the vehicle is representedin Fig. 1.

2.1.1. Portable emissions measurements system (PEMS)instrumentation

The PEMS equipment used was the Semtech-DS manufacturedby Sensors, Inc. It consists of a tailpipe attachment, heated exhaustlines, an exhaust flow meter (EFM), exhaust gas analyzers, datalogger to vehicle network, a global positioning system (GPS) and aweather station (WS) for ambient temperature and humidity. Alldata were recorded at a frequency of 1 Hz. The whole systemfurther adds ~100 kg to the vehicle, besides the weight of the driver(~80 kg). The Semtech DS measures exhaust gas concentrations ofunburned hydrocarbons (THC) by heated flame ionization detector,carbon monoxide (CO) and carbon dioxide (CO2) by a non-dispersive infrared sensor, and nitrogen monoxide (NO) and ni-trogen dioxide (NO2) by a non-dispersive ultra-violet sensor. NOx iscalculated by the sum of the concentrations of NO and NO2. EFMuses the pitot tube principle in order to measure pressure differ-ential to be able to calculate mass flow. The measurement princi-ples and accuracy from the Semtech DS are in-line to thosedescribed by current legislation for this type of testing (EuropeanCommission, 2011). As a standard procedure, test runs prepara-tion included routine calibration of the gas analyzers (zero andspan of gases) prior to each test. For that purpose, two gas stan-dards bottles were used. One containing a mixture of CO2 (13 mol-%) CO (1300mol-ppm), NO (1800mol-ppm) and propane (250mol-ppm) in nitrogen, and the other containing NO2 (218 mol-ppm) inpurified dry air.

2.1.2. HORIBA MEXA 1400 QL-NX. Quantum cascade laser infra-redspectrometer (QCL-IR)

MEXA 1400QL-NX is an analyzer for simultaneous direct mea-surement of four nitrogen compounds (NO, NO2, N2O, NH3) inautomobile exhaust gas in real time by using Infrared AbsorptionSpectroscopy (IR Spectroscopy). The instrument combines aQuantum Cascade Laser (QCL) light source and a precisely adjustedlong dual-path optical cell. It shows high sensitivity with a limit ofdetection which complies with the current European legislativerequirements. Furthermore, the QCL has awide dynamic range (i.e.,0e50 ppm to 0e2000 ppm) for the measurement of ammoniaemissions in the exhaust gas. The MEXA 1400 QL-NX has a wave-length resolution close to 0.006 cm�1.

Fig. 1. Vehicle setup. OBD ¼ On-board diagnostics, GPS ¼ Global positioning system,WS ¼ weather station, DPF ¼ Diesel Particulate Filter, SCR ¼ Selective CatalyticReduction, EFM ¼ Exhaust flow meter.

The analyzer utilizes a high resolution spectrum and a highvacuum optical cell in order tominimize the interference offered bythe co-existing gases. Moreover, the ammonia response has beenimproved by using a vacuum sample transfer line maintained at atemperature of 113 �C, which ensures a shorter residence time andminimum adsorption and condensation of ammonia on the surfaceof the walls. The temperature is optimized to prevent the decom-position of urea or its by-products and ensures a selective mea-surement of pure ammonia.

The analyzer can be used for the measurement of exhaust gascomponents from various fuel and engine types at 10 Hz frequency.It consists of three main components e a Main Control Unit (MCU),an analyzer unit and a heated filter (F-01HN). The MCU serves as aninstrument for the calculation of emissions and for the display ofthe measured concentration. The analyzer unit contains the core ofthe MEXA 1400QL-NX e the sensor including the QCL element, agas cell and optics, as well as a sample handling system specificallydesigned for the measurement of NO, NO2, N2O, NH3. The heatedfilter is connected to the analyzer unit via a heated line. The sam-pling of the exhaust is conducted via a stainless steel samplingprobe covered with a heated jacket in order to avoid cold spots. TheF-01HN contains a quartz filter element specifically designed tominimize the desorption of ammonia molecules present in theexhaust gas. Further details, on the performance during the realtime measurement of NH3 emission on transient cycles can befound in Suarez-Bertoa et al., 2015 (Suarez-Bertoa et al., 2015). Toensure accurate measurements the QCL-IR was calibrated beforeeach test. The calibration sequence includes zero and span adjust-ment. The QCL-IR has 8 ranges, two measurement ranges percompound (low range and high range) which were calibratedseparately. The span gases selected for the calibration must have aminimum concentration of 90% of the measurement range. Themeasurement ranges and the span gases used for the calibration aredescribed in Table S1 of the supplementary material.

2.2. Route description

The current PEMS Procedure described in the legislation(European Commission, 2011) prescribes the minimum duration ofthe trip to be determined by the amount of work performed by theengine during operation. In the case of a Euro V vehicle, this isdefined as 3 times the work produced during the European Tran-sient Cycle (ETC) for heavy-duty vehicles. The shares of urban, ruraland motorway operation are specific to the different vehicle cate-gories (i.e., N1, N2, N3, etc.).

This approach was not strictly followed as the objectives of thecampaignweremainly: (i) the ability to measure second-by-secondconcentrations of NH3 and N2O during different driving conditions,with special emphasis on the urban operation and (ii) the study ofthe emissions behavior of the vehicle under different diving con-ditions, with an special focus on the urban operation (see section

Table 2Trips characteristics for Test (European Environmental Agency, 2014; World HealthOrganization, 2013; Sillanp€a€a et al., 2006; Bouwman et al., 2002).

Test 1 Test 2 Test 3 Test 4

Work (kWh) 103.7 106.4 103.2 97.0Urban share (%) 56 59 55 57Rural share (%) 22 20 29 29Motorway (%) 22 21 16 14Trip duration (s) 9022 9874 10,844 10,794Trip distance (km) 118.4 116.4 138.55 136.9Av. Amb. temperature (�C) 10.8 16.0 10.7 12.6Engine start temperature (�C) 7 12 9 70

R. Suarez-Bertoa et al. / Atmospheric Environment 139 (2016) 167e175170

2.3. PEMS based data analysis). Therefore, the different routescontain a larger percentage of urban operation compared to therural and motorway shares (Table 2) whereas the legislativerequirement for N3 vehicles is 20% urban, 25% rural and followed by55% motorway operation. Three different routes were used withinthe 4 tests included in this study, these routes include at least 50%of urban operation and the rural and motorway shares vary be-tween 17-29% and 15e23%, respectively. The different tests alwaysstarted in an urban environment, the subsequent share of operationvaried sometimes followed by the rural or by the motorway part.The same routewas followed during Test 3 and Test 4. Test 3 startedwith a cold engine (i.e., starting coolant temperature <70 �C) andTest 4 was a hot start (i.e., starting coolant temperature >70 �C).This route included a large amount of urban driving in themiddle ofthe test. Test 1 and Test 2 started with a cold engine.

Through the design of these routes, the intention was also tostudy the NOx emitted by the vehicle and the behavior of its Se-lective Catalytic Reduction (SCR) system to reduce these emissions.The operation of the SCR is directly linked with the exhaust tem-peratures attained by the engine operating in different power re-gimes, which is partly influenced by the road driving conditions(i.e., urbanemainly low power operation, motorwayemainly highpower operation). For example, placing an urban share of drivingafter a motorway share would allow seeing the impact on NOxemissions (in the urban share) due to a forced preconditioning ofthe SCR system (temperature of the system above 200 �C). Whereasin Test 1 or Test 2 the route is such that the SCR was graduallyincreasing its temperature up to a point on which the NOx abate-ment efficiency increases.

Although the routes characteristics divert from the currentlegislative requirements, the data analysis procedure follows thelegislative prescription. This provided a baseline against which toassess the impact of the boundary conditions described in section2.3.3.

2.3. PEMS based data analysis

2.3.1. Moving averaging window (MAW) method for data analysisThe averaging window method (European Commission, 2011;

Perujo Mateos Del Parque and Mendoza Villafuerte, 2015) is amoving averaging process, based on a reference quantity obtainedfrom the engine characteristics and its performance on the typeapproval transient cycle (i.e., for Euro V e the amount of workproduced over an ETC or the mass of CO2 emitted over an ETC). Thereference quantity sets the characteristics of the averaging process(i.e., the duration of the windows). Using the MAW method, theemissions are integrated over windows whose common charac-teristic is the reference engine work or CO2 mass emissions.

Using the engine work or CO2 mass over a fixed cycle as refer-ence quantity is an essential feature of the method, leading to thesame level of averaging and range of results for various engines.

The first window (i.e., averaged value) is obtained between thefirst data point and the data point for which the reference quantity(1 � CO2 or work achieved at the ETC) is reached. The calculation isthen moving, with a time increment equal to the data samplingfrequency (at least 1 Hz for the gaseous emissions).

The term “averaging” is related to the average power of thewindow, which by the current legislative requirements will defineits validity (MAWAvPower > 20%(15%); i.e., MAW whose averagepower is below 20% (or 15%) of the maximum indicated enginepower).

The result of this analysis concludes when emission factors arecalculated by dividing the integrated emissions by the work of thewindow, g/kWh (i.e., ETCwork). It is also possible to calculate the socalled Conformity Factors (CF) which are defined by dividing the

emission factor by the current legislated limit of the pollutant,these factors are important as they determine if the vehicle passesor fail a PEMS based test. In this study the results are represented bythe emissions factors (EFs) as the testing performed is not intendedto be a type approval or In-Service Conformity (ISC). Results are alsolimited to NOx and the discussed unregulated pollutants (i.e., NH3and N2O).

As expressed before, the ammonia emissions over the laboratorytype approval cycle are regulated (European Commission, 2011;United Nations, 2013) in terms of average concentration (ppm/test). To be consistent with this concept, the present study reportedammonia emissions as average concentration (ppm units) over theMAW, but also as EF (g/kWh), similarly to the other studiedcompounds.

2.3.2. EMROAD© data evaluation toolData from the PEMS based testing was analyzed using

EMROAD©. EMROAD© (Bonnel, 2015; Bonnel et al., 2011) is aMicrosoft Excel add-in used to analyze on-road emissions datacollected with Portable Emissions Measurement Systems (PEMS). Itwas developed as a reference calculation tool to support the dataanalysis in the frame of the European legislative programsincluding PEMS based testing (heavy-duty, non-road engines andlight-duty vehicles). EMROAD is used to support the developmentof new PEMS data evaluation methods for emissions legislationsuch as the In-Service Conformity (ISC, heavy-duty and non-roadengines) and real driving emissions (RDE, light-duty vehicles).

2.3.3. Boundary conditions for data evaluationThe current PEMS procedure for heavy-duty vehicles is ruled by

a series of boundary conditions that limit the amount of data to betaken into consideration for the final emissions analysis. They arethe following: (i) vehicle/engine conditioning: for cold start emis-sions the data is not taken into account if coolant temperature isbelow 70 �C, (ii) low engine power operation: MAWwhose averagepower is below 20% (or 15%) of the maximum indicated enginepower are not considered, (iii) the use of the 90th cumulativepercentile of the MAW emissions CF as the representative result ofthe vehicle instead of the maximum CF (which would represent theworst pollutant “window”).

The process followed to analyze the data from each test was thefollowing:

a) Obtain emissions factors following the methodology asdescribed in the legislation (considering all boundary con-ditions) in order to establish a baseline.

b) Assessment 1e Includes the analysis of data considered to beon cold start operation of those trips which have a percent-age of it.

c) Assessment 2 e Including cold start, all windows even thosebelow the 20% power threshold but still using the 90th cu-mulative percentile.

d) Assessment 3 e includes all possible analyzable data.e) Assessment 4 e Binning of the MAW emissions factors ac-

cording to the average speed of the windows into urban(<50 km/h), rural (50e75 km/h) and motorway (>75 km/h).The binning is performed using all MAWwithout boundariesand data are given of the maximum obtained values.

3. Results and discussion

3.1. Instrumentation inter-comparison

Real time NO and NO2 emissions from a Euro V heavy-dutyvehicle equipped with a SCR DeNOx system were measured using

R. Suarez-Bertoa et al. / Atmospheric Environment 139 (2016) 167e175 171

the Semtech-DS and a QCL-IR during 3 cold start (Test 1e3) and onehot-start (Test 4) on-road tests (see Fig. S1 of the supplementarymaterial). Furthermore, NH3 and N2O emissions were alsomeasured in real time using the QCL-IR during the 4 on-road tests.Fig. 2 shows that NO measurements obtained using the Semtech-DS and the QCL-IR presented a very good correlation. Fig. 3 andFig. S2 illustrate the NO emission profiles measured with the twoinstruments, which were nearly identical. The NO signals wereused for the measurements alignment. Although the NO2 emissionsprofiles obtained using the Semtech-DS and the QCL-IR followedthe same trend (Fig. 3 and Fig. S3), the correlation of the measuredvalues was poor (Fig. 2). This was most probably due to a negativeoffset on the NO2 concentrations measured by the Semtech (offsetof ~20 ppm), which resulted in zero NO2 emissions when theywerebelow 20 ppm and in NO2 concentrations ~20 ppm lower thanthose measured with the QCL-IR when the emissions were above20 ppm. However, since NO2 emissions were very low compared toNO emissions, NOx (NO þ NO2) emissions estimated for both sys-tems still presented a good correlation (R2 � 0.95) as shown inFig. 2. NOx emissions were significant during the whole length ofthe on-road tests, including urban (speed � 50 km/h), rural(speed > 50 � 75 km/h) and highway (speed > 75 � 90 km/h)shares, reaching NO concentration above 2000 ppm (see Fig. 3).

3.2. Engine NOx emissions

Table 3 shows the impact on NOx emissions of the differentcalculation scenarios described in section 2.3.3. The different stepsdefined by the assessments gradually show the increase of emis-sions factors by looking outside the current legislative boundarieson the process of analysis of PEMS data. NOx EFs for Test 1 and Test2 on baseline analysis and assessment 1 are identical and are theonly ones shown for these assessments. This can be explained bytwo important factors: (i) these were the tests that presented atleast 50% of valid windows as required by the terms of theregulatory-baseline data evaluation procedure at 20% powerthreshold, (ii) the analysis including the cold start data showed thatthe windows affected by the cold start were not reflected in thefinal result. This may be attributed either to the 90th cumulative

Fig. 2. Correlation of NO, NO2 and NOx obtained from the me

percentile boundary as second intrinsic boundary condition, or tothe fact that the highest NOx produced by the vehicle occurs over“warm” operation. In this vehicle the highest emissions factorswere observed under “warm” operation. Table 3 also shows theresults of the assessment 2 and 3. It can be observed that withoutthe boundary conditions the EFs increase by more than 40%. Underthe studied conditions the vehicle presented NOx emissions thatwere up to 5 times higher than the Euro V standards (2 g/kWh) forengine emissions.

3.3. Engine NH3 and N2O emissions

Figs. 4 and 5 show the on-road NH3 and N2O emission profilesfor this heavy-duty vehicle during Test 1 and Test 3. More than100 ppm of NH3 was measured during several minutes along thetests (see Figs. 4 and 5). N2O emissions were observed togetherwith those of NH3. N2O concentrations presented peaks reachingvalues as high as 130 ppm (Fig. 4). The emissions of these pollutantswere observed in correspondence with the signal of urea injectionrecorded using the vehicle’s OBD (see Figs. 4 and 5), suggesting thatthese emissions are linked to the use of the SCR system. Injection ofurea, as well as NH3 and N2O emissions, took place only during therural and motorway parts of the tests, when the vehicle exhausttemperature was at 250 �C or higher.

The data also allowed calculating NH3 emission concentrationand NH3 and N2O emissions per window of operation usingEMROAD© and the MAW methodology. The analyses of theseemissions were carried out without boundary conditions. Hence,NH3 emissions concentration calculated applying the MAWapproach to all available windows varied from 0.5 ppm (min MAWaverage concentration e Test 4) to 6.8 ppm (max MAW averageconcentration e Test 1). The average NH3 concentration corre-sponding to the length of MAW defined by the work produced atthe ETC during these tests ranged from 0.9 to 5.7 ppm (see Table 4).When expressed as mass and/or brake-specific emissions, averageNH3 emissions corresponding to the work performed through theETC varied from 0.019 g/kWh to 0.063 g/kWh. Total NH3 emissionduring the tests ranged from 1.331 g (Test 4: 10,794 s and 136.9 km)to 3.769 g (Test 1: 9022 s and 118.4 km), which corresponds to 0.014

asurements of Semtech-DS and the QCL-IR during Test 1.

Fig. 3. Real time NO and NO2 emission profiles obtained using the Semtech-DS and the Horiba QCL-IR at 1 Hz resolution during Test 1.

Table 3NOx emissions (emissions factors in g/kWh) impact through the various iterations.

Test 1 Test 2 Test 3 Test 4

Baseline (CS excluded, 90th%ile, 20% PT) 7.82 6.23 e e

Assessment 1 (CS included, 90th%ile, 20% PT) 7.82 6.23 e e

Assessment 2 (CS included, 90th%ile, No PT) 9.20 7.12 8.56 9.56Assessment 3 (CS included, 100th%ile, No PT) 11.48 9.90 10.46 9.73

1 CS: Cold Start; 2 PT: Power Threshold.

R. Suarez-Bertoa et al. / Atmospheric Environment 139 (2016) 167e175172

and 0.036 g/kWh, respectively. The obtained average NH3 emissionconcentrations were lower than the 10 ppm allowed in test benchfor Euro VI HDV (weighted average over the WHTC and the WHSC).Results demonstrate a lack of efficiency on the operation of this SCRsystem supported by the high NOx emitted and the presence of NH3and N2O emissions.

Average N2O EFs obtained for an ETC equivalent, using all theavailable windows (i.e., no boundary conditions applied to the dataanalysis), ranged from 0.063 g/kWh (Test 4) to 0.139 g/kWh (Test 1),with maximum EFs as high as 0.158 g/kWh. The average N2O EFsobtained using the described methodology for the studied truckon-road measurement were lower than the U.S. N2O emission limit(0.133 g/kWh) for engine testing (U.S. Environmental ProtectionAgency, 2015), with the exception of Test 1 which was slightlyhigher (0.139 g/kWh). The calculated average N2O and NH3 EFswere 1.4e3.2 and 2.7e9 times higher than those reported byTadano et al. for an Euro V engine equipped with an SCR system,tested over the European Steady Cycle (ESC) (Tadano et al., 2014).Furthermore, while Tadano et al. (2014) reported NOx EFs withinEuro V limits, the obtained NOx EF for the on-road tests rangedfrom 6.23 g/kWh, when 90th cumulative percentile and 20% powerthreshold were fixed and cold start excluded and up to 11.48 g/kWh, when all the available windows were used to the calculation,far from the 2.0 g/kWh Euro V limit (ETC).

3.4. Effect of driving conditions upon emissions

NOx, NH3 and N2O emissions were also analyzed as function ofthe vehicle share of operation (e.g., urban, rural, motorway) duringeach trip, in line with Assessment 4 as described in Section 2.3.3.Trip composition often has an effect on the production of specificpollutants. This is due to different variables that have specificimpact on its production (e.g., traffic, weather, altitude, idle or lowpower operation, etc.). The MAW analysis allows the binning ofwindows by the MAW vehicle average speed (e.g., speed bin: urban(<50 km/h), rural (50e75 km/h) andmotorway (>75 km/h)). It doesso using all the available windows (no boundaries applied). Hence,NOx, NH3 and N2O EFs (g/kWh) and NH3 emission concentration(ppm) for the urban, rural and motorway shares calculated for Test1e4 are shown in Table 5. There were no MAWs for the motorwayoperation. This is mainly because the construction of the MAWtakes the amount of seconds of operation required to complete thecomparable ETC work and reports the average speed of the win-dow, i.e., there were not enough seconds to complete the workrequired for a ETC window during the motorway operation only.This averaged speed is used to bin the windows. NH3 and N2Oemissions were slightly higher during the rural share for all fourtests, going from 0.019 to 0.060 g/kWh and 0.079e0.158 g/kWh,respectively. However, NOx emissions were substantially higherduring the urban operation, where a larger fraction of the popu-lation could be exposed to these toxic gases. In the urban operationNOx emissions were as high as 11.48 g/kWh. In order to reduce NOxemissions under urban operation only, the emissions controlstrategy would require injecting a larger amount of urea, whichmay result in even higher emissions of NH3 and N2O.

4. Conclusions

In service conformity requirements for heavy-duty vehicleshave been introduced to check the conformity of heavy-duty en-gines with the applicable emissions certification standards during

Fig. 4. aed. Real time emission profiles and vehicle parameters measured during Test 1 at 1 Hz resolution. a) speed profile, b) NH3 and N2O emission profiles, c) NH3 emission profileand Adblue injection, d) Engine exhaust temperature and Adblue injection.

R. Suarez-Bertoa et al. / Atmospheric Environment 139 (2016) 167e175 173

their normal life. The Euro VI Regulation (EC) No 595/2009 and theimplementing Regulation (EC) 582/2011 (European Commission,2011) introduced a procedure for Portable Emissions Measure-ment Systems (PEMS) testing as a mandatory part of the typeapproval legislation of those vehicles. NOx, CO, CO2 and THC are thepollutants monitored with PEMS systems during on-road in serviceconformity testing. Although NH3 and N2O are regulated com-pounds for engine type approval in Europe and US, respectively, on-road emissions for these two pollutants were never reported.

The results obtained in this study indicate that real time on-roademissions of N2O and NH3 can bemeasured using a QCL-IR from theroad exhaust of a heavy-duty vehicle. Furthermore, the obtainedresults can be included and analyzed by the moving average win-dow (MAW) method using EMROAD tool.

NO and NO2 emission concentrations were measured using theSemtech-DS and the QCL-IR. While NO measurements were verywell correlated, NO2 were more scattered. Nonetheless, NOx con-centrations, obtained from both instruments, presented a very good

Fig. 5. aed. Real time emission profiles and vehicle parameters measured during Test 3 at 1 Hz resolution. a) speed profile, b) NH3 and N2O emission profiles, c) NH3 emission profileand Adblue injection, d) Engine exhaust temperature and Adblue injection.

R. Suarez-Bertoa et al. / Atmospheric Environment 139 (2016) 167e175174

correlation due to the much lower contribution of NO2 on the NOxconcentration compared to NO. NOx emissions factors resultingfrom the MAW analysis show some relatively high emissions whencompared with the Euro V legislative (ETC cycle) limit of 2 g/kWh,ranging from 6.23 g/kWh (when considering 90th percentile and20% power threshold) to 11.48 g/kWh (when considering all theavailable windows). The highest NOx emissions were alwaysobserved in the urban share of operation of the trips. On the otherhand, the highest N2O and NH3 emissions were present in themotorway and/or rural share, reaching maximum peak concen-trations above 140 ppm. The average N2O EFs were lower than the

U.S. N2O emission limit (0.133 g/kWh) for engine testing, with theexception of Test 1 which was slightly higher (0.139 g/kWh). Theaverage NH3 concentrations calculated from the tests were alsolower than the Euro VI limit (10 ppm) going from 0.9 to 5.7 ppm.When expressed as average EFs (g/kWh) NH3 emissions varied from0.019 g/kWh to 0.063 g/kWh. Although calculated average N2O andNH3 emissions were within current limits, NOx emissions weresubstantially higher than Euro V limits. Since the SCR system(which is a urea-based DeNOx) is used to reduce NOx emissions,more urea solution would need to be injected in the SCR to reactand reduce NOx emissions. As a consequence, higher emissions of

Table 4NOx, NH3 and N2O emission factors (g/kWh) and NH3 emission concentration (ppm)analyzed applying assessment 3 (i.e., including all possible analyzable data).

Test 1 Test 2 Test 3 Test 4

NOx (g/kWh) 11.48 9.90 10.46 9.73NH3 [ppm] Max 6.8 3.9 2.6 1.7

Min 2.8 0.7 1.1 0.5Average 5.7 2.6 1.7 0.9

NH3 (g/kWh) Max 0.061 0.036 0.030 0.079Min 0.040 0.010 0.012 0.048Average 0.052 0.030 0.019 0.063NH3 Total trip [g] 3.769 2.080 2.283 1.331

N2O (g/kWh) Max 0.158 0.099 0.116 0.079Min 0.089 0.055 0.073 0.048Average 0.139 0.074 0.090 0.063

Table 5NOx, NH3 and N2O EFs (g/kWh) and NH3 emission concentration (ppm) for the ur-ban, rural and motorway shares calculated for Test 1e4 binning of all possibleanalyzable data.

Test 1 Test 2 Test 3 Test 4

NOx (g/kWh) Urban 11.48 9.90 10.46 9.73Rural 9.03 7.18 8.47 8.72MW e e e e

NH3 [ppm] Urban 4.9 2.7 2.0 1.5Rural 6.7 3.7 2.6 1.7MW e e e e

NH3 (g/kWh) Urban 0.057 0.035 0.025 0.017Rural 0.060 0.036 0.030 0.019MW e e e e

N2O (g/kWh) Urban 0.128 0.098 0.096 0.071Rural 0.158 0.099 0.115 0.079MW e e e e

R. Suarez-Bertoa et al. / Atmospheric Environment 139 (2016) 167e175 175

N2O and NH3 could be expected. Being NH3 a precursor in theformation of atmospheric PM2.5, which is widely recognised as ahuman health hazardous compound, higher emissions of thiscompound may result in higher PM levels in urban areas.

Disclaimer

The opinions expressed in this manuscript are those of the au-thors and should in no way be considered to represent an officialopinion of the European Commission.

Acknowledgments

The VELA staff is acknowledged for the skilful technical assis-tance, in particular M. Cadario, M. Carriero, R. Colombo, F. Forni, P.Le Lijour, D. Lesueur, F. Montigny and M. Otura-Garcia.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.atmosenv.2016.04.035.

References

Bonnel, P., 2015. Portable Emission Measurements (PEMS)e Data Evaluation andPost-processing Manual for the Data Evaluation Software EMROADeVersion5.80. Joint Research Centre.

Bonnel, P., Kubelt, J., Provenza, A., 2011. Heavy-duty Engines Conformity TestingBased on PEMSelessons Learned from the European Pilot Program. JointResearch Centre.

Bouwman, A.F., Van Vuuren, D.P., Derwent, R.G., Posch, M., 2002. A global analysis ofacidification and eutrophication of terrestrial ecosystems. Water, Air, Soil Pollut.141 (1e4), 349e382.

Erisman, J.W., Grennfelt, P., Sutton, M., 2003. The European perspective on nitrogenemission and deposition. Environ. Int. 29 (2e3), 311e325.

European Commission, 25.6.2011. Commission regulation (EU) No 582/2011 of 25May 2011 implementing and amending regulation (EC) No 595/2009 of theEuropean Parliament and of the council with respect to emissions from heavyduty vehicles (Euro VI) and amending annexes I and III to directive 2007/46/ECof the European Parliament and of the council. Official J. Eur. Union L 167,1e168.

European Environmental Agency, 2014. Air Quality in Europe.Eurostat. Statistics on European cities. 2015; Available from: http://ec.europa.eu/

eurostat/statistics-explained/index.php/Statistics_on_European_cities#Main_statistical_findings.

Perujo Mateos Del Parque, A., Mendoza Villafuerte, P., 2015. PEMS Emissions Testingof Heavy Duty Vehicles/engines: Assessment of PEMS Procedures in Fulfilmentof Article 14(3) to Regulation (EU) 582/2011.

Ravishankara, A.R., Daniel, J.S., Portmann, R.W., 2009. Nitrous oxide (N2O): thedominant ozone-depleting substance emitted in the 21st century. Science 326(5949), 123e125.

Sillanp€a€a, M., Hillamo, R., Saarikoski, S., Frey, A., Pennanen, A., Makkonen, U.,Spolnik, Z., Van Grieken, R., Brani�s, M., Brunekreef, B., Chalbot, M.C.,Kuhlbusch, T., Sunyer, J., Kerminen, V.-M., Kulmala, M., Salone, R.O., 2006.Chemical composition and mass closure of particulate matter at six urban sitesin Europe. Atmos. Environ. 40 (Suppl. 2), 212e223.

Suarez-Bertoa, R., Zardini, A.A., Lilova, V., Meyer, D., Nakatani, S., Hibel, F., Ewers, J.,Clairotte, M., Hill, L., Astorga, C., 2015. Intercomparison of real-time tailpipeammonia measurements from vehicles tested over the new world-harmonizedlight-duty vehicle test cycle (WLTC). Environ. Sci. Pollut. Res. 22 (10),7450e7460.

Sutton, M.A., Dragosits, U., Tang, Y.S., Fowler, D., 2000. Ammonia emissions fromnon-agricultural sources in the UK. Atmos. Environ. 34 (6), 855e869.

Tadano, Y.S., Borillo, G.C., Godoi, A.F., Cichon, A., Silva, T.O., Valebona, F.B.,Errera, M.R., Penteado Neto, R.A., Rempel, D., Martin, L., Yamamoto, C.I.,Godoi, R.H., 2014. Gaseous emissions from a heavy-duty engine equipped withSCR aftertreatment system and fuelled with diesel and biodiesel: assessment ofpollutant dispersion and health risk. Sci. Total Environ. 500e501 (0), 64e71.

United Nations, 2013. Regulation 49: Uniform Provisions Concerning the Measuresto Be Taken against the Emission of Gaseous and Particulate Pollutants fromCompressioneignition Engines and Positive Ignition Engines for Use in Vehi-cles. UN(ECE).

U.S. Environmental Protection Agency, 2015. Phase 2 Greenhouse Gas EmissionsStandards and Fuel Efficiency Standards for Medium- and Heavy-duty Enginesand Vehicles. Available from: http://www3.epa.gov/otaq/climate/regs-heavy-duty.htm.

World Health Organization, 2013. Review of Evidence on Health Aspects of AirPollution- REVIHAAP Project. WHO Technical Report. Regional Office forEurope, Copenhagen, Denmark.