Atmospheric Environment

Transcription

1 Atmospheric Environment 44 (2010) Contents lists ville t ScienceDirect Atmospheric Environment journl homepge: Experimentl investigtion of prticulte emissions from diesel engine fueled with ultrlow-sulfur diesel fuel lended with diglyme Yge Di,, *, C.S. Cheung, Zuohu Hung Stte Key Lortory of Multiphse Flow in Power Engineering, Xi n Jiotong University, Xi n , PR Chin Deprtment of Mechnicl Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon, PR Chin rticle info strct Article history: Received 22 April 2009 Received in revised form 15 Septemer 2009 Accepted 25 Septemer 2009 Keywords: Diglyme Ultrlow-sulfur diesel Prticulte mss concentrtion Prticle size distriution Trde-off Experiments re conducted on 4-cylinder direct-injection diesel engine using ultrlow-sulfur diesel s the se fuel nd diglyme s the oxygente component to investigte the prticulte emissions of the engine under five engine lods t two engine speeds of 1800 rev min 1 nd 2400 rev min 1. Blended fuels contining 5%, 10.1%, 15.2%, 20.4%, 25.7% nd 53% y volume of diglyme, corresponding to 2%, 4%, 6%, 8%, 10% nd 20% y mss of oxygen, re studied. The study shows tht with the increse of oxygen in the fuel lends, smoke opcity, prticulte mss concentrtion, NO x concentrtion nd rke specific prticulte emission re reduced t the two engine speeds. However, the proportion of solule orgnic frction is incresed. For ech lended fuel, the totl prticle numer concentrtion is higher while the geometric men dimeter is smller, compred with tht of ultrlow-sulfur diesel, though the prticle numer decreses with the oxygen content of the lended fuel. Furthermore, the lended fuels lso increse the numer concentrtions of prticles smller thn 100 nm. Ó 2009 Elsevier Ltd. All rights reserved. 1. Introduction Prticulte emissions from diesel engine hve received incresing concern due to their possile dverse effects on humn helth nd the environment. Moreover, the incresing crude oil price hs lso creted pressure on reducing diesel oil consumption. These hve led to serious considertion on seeking lterntives to replce diesel fuel for diesel engines, either in prt or in full, to reduce PM emission nd to reduce diesel oil consumption. In this regrd vrious oxygentes hve een considered sed on their vilility, price, toxicity, sfety nd comptiility with diesel fuel (Miymoto et l., 1998; Hnsen et l., 2005; Agrwl, 2007; Demirs, 2007; Rieiro et l., 2007; Lpuert et l., 2008). Among the oxygentes, iodiesel, ethnol nd methnol re the most widely investigted (Agrwl, 2007). Diethylene glycol dimethyl ether (DGM), which is lso known s diglyme, is n oxygente with high oxygen content. However there re fewer investigtions on DGM. DGM hs good intersoluility with diesel fuel, high cetne numer, lower sooting tendency nd suitle physicl chemicl properties for ppliction to commercil diesel engine (Bertoli et l., 2000). DGM is cler, colorless liquid which hs slight * Corresponding uthor t: Stte Key Lortory of Multiphse Flow in Power Engineering, Xi n Jiotong University, Xi n , PR Chin. Tel.: þ ; fx: þ E-mil ddress: diyge_2003@sin.com.cn (Y. Di). ether-like odor, reltively low toxicity nd low rectivity. It cn e produced in commercil quntities y vriety of methods ( chemindustry.ru/diglyme.php; Bimridge et l., 2004) nd is redily ville in Chin. Thus, DGM is potentil oxygented dditive or lterntive for diesel engine, t lest in Chin where significnt mount of diesel oil is imported, hs een investigted y some reserchers. The results cn e summrized s follows. In the comustion spect, with n increse of DGM in diesel fuel, the ignition dely, the diffusive comustion durtion, the totl comustion durtion nd the mount of het relese in the premixed phse re reduced, while the mximum men gs temperture nd the mount of het relese in the diffusive comustion phse increse (Znnis et l., 2004; Ren et l., 2007). In the emission spect, HC, CO, PM nd smoke re reduced when DGM is dded to diesel fuel (Bennethum nd Winsor, 1991; Miymoto et l., 1998; Shih, 1998; Betrice et l., 1999; Tmnouchi et l., 1999; Znnis et l., 2004; Ren et l., 2007) while NO x emissions cn either increse (Znnis et l., 2004), decrese (Miymoto et l., 1998; Shih, 1998; Betrice et l., 1999) or do not chnge (Ren et l., 2007). However, very few of them re relted to prticulte emissions, especilly in terms of numer concentrtion nd size distriution. It is known tht diesel vehicles re one of the mjor sources of prticles in the urn tmosphere nd most of the prticles emitted y diesel engines re in the rnge with dimeter less thn 100 nm, nd most of the mss of the prticles is in the ccumultion mode with dimeter in the rnge of nm (Kittelson, 1998). Due to /$ see front mtter Ó 2009 Elsevier Ltd. All rights reserved. doi: /j.tmosenv

2 56 Y. Di et l. / Atmospheric Environment 44 (2010) Tle 1 Engine specifictions. Model Isuzu 4HF1 Type In-line 4-cylinder Mximum power 88 kw/3200 rev min 1 Mximum torque 285 Nm/1800 rev min 1 Bore stroke 112 mm 110 mm Displcement 4334 cc Compression rtio 19.0:1 Fuel injection timing (BTDC) 8 Injection pump type Bosch in-line type Injection nozzle Hole type (with 5 orifices) the links etween PM emission nd helth effects (Peters et l., 1997; Somers et l., 2004; Pope nd Dockery, 2006), it is necessry to mesure the prticle chrcteristics such s numer size distriution nd mss concentrtion in order to hve etter understnding of the prticulte mtter. Thus, in this study, smoke opcity, prticulte mss concentrtion, prticle numer concentrtion, size distriutions nd the proportion of solule orgnic frction (SOF) in the prticulte re investigted. 2. Test engine nd fuel properties Experiments were crried out on nturlly-spirted, wtercooled, 4-cylinder direct-injection diesel engine. Specifictions of the engine re shown in Tle 1. The engine ws coupled with n eddy-current dynmometer. A control system ws ville for djusting the engine speed nd torque. The fuels used in this study include ultrlow-sulfur diesel fuel (ULSD) nd DGM. The mjor properties of ULSD, DGM nd their lends re shown in Tle 2. Six ULSD DGM lends were prepred for oxygen content y weight of 2%, 4%, 6%, 8%, 10% nd 20%, which re designted s UD-1, UD-2, UD-3, UD-4, UD-5 nd UD-6, respectively. The volumetric concentrtion of ech lended fuel ws determined for the desired oxygen content. The lower heting vlue of DGM is 24.5 MJ kg 1 (Ren et l., 2007) while the lower heting vlue of ULSD is 42.5 MJ kg 1. The lower heting vlue of ech lended fuel is clculted sed on the mss frction of ech component in the lended fuel. 3. Experimentl setup, equipments nd procedures Fig. 1 shows the experimentl system. T 1 nd T 2 re thermocouples for inlet ir temperture nd exhust gs temperture respectively. Prticulte mss concentrtion in the engine exhust ws mesured with tpered element oscillting microlnce (R&P TEOM 1105). A scnning moility prticle sizer (SMPS, TSI Model 3934) ws used to mesure the size distriution nd numer Fig. 1. Schemtic digrm of experimentl system. concentrtion. The exhust gs from the engine ws diluted with Dekti mini-diluter (Wong et l., 2003) efore pssing through the TEOM nd the SMPS. The dilution rtio ws determined from the mesured CO 2 concentrtions of ckground ir, undiluted exhust gs nd diluted exhust gs. The Soxhlet pprtus (Wheton Science Products Inc., USA) ws used to determine the SOF in the prticles. Pllflex T60A20 teflon-coted glss fier filter (Pllflex Inc., USA) with dimeter of 47 mm ws used to collect the prticulte smple. A solution contining toluene with purity of more thn 99.5% nd nhydrous ethnol with purity of more thn 99.7% ws used s n extrcting solution. Smoke opcity of the exhust gs ws mesured with Diesel tune smoke-meter (SPX DX.210). CO 2 nd NO x concentrtions were lso mesured. NO x ws mesured with heted chemiluminescent nlyzer (HCLA CAI Model 400). CO 2 ws mesured with non-dispersive infr-red nlyzer (NDIR CAI 300). The gs nlyzers were clirted with stndrd nd zero gses efore ech experiment. Experiments were crried out t stedy sttes t different engine lods nd two engine speeds of 1800 rev min 1 nd 2400 rev min 1. In ech test, the engine ws llowed to run for few minutes until the exhust gs temperture, the cooling wter temperture, the luricting oil temperture, s well s the CO 2 concentrtion hve ttined the stedy-stte vlues nd dt were then mesured. The gseous emissions nd the prticulte mss emission were continuously mesured for 5 min nd the verge results presented. The stedy stte tests were repeted twice to ensure tht the results re repetle, within the experimentl uncertinties of the mesurements. The experimentl uncertinty nd stndrd errors in the mesurements re determined sed on the method of Kline nd McClintock (1953). The mximum stndrd errors re 2.7% for mss consumption of fuels, 1.3% for prticulte mss concentrtion, 2.5% for prticle numer concentrtion, 2% for smoke opcity, 2.9% for NO x nd 0.37% for CO 2. Tle 2 Properties of ultrlow-sulfur diesel, DGM nd the lended fuels. Properties ULSD DGM UD-1 UD-2 UD-3 UD-4 UD-5 UD-6 Lower heting vlue/mj kg Het of evportion/kj kg Density@200C, kg m Viscosity@200C, mp s Cetne numer Sulfur content/ppm wt <50 0 Boiling point/ C Stoichiometric ir-fuel rtio/kg ir/kg fuel Chemicl formul (C 7.2 H 14 )n C 6 H 14 O 3 C wt. % H wt. % O wt. % ULSD/vol. % DGM/vol. %

3 Y. Di et l. / Atmospheric Environment 44 (2010) Results nd discussion Experiments were performed t the rted torque speed of 1800 rev min 1 nd t higher engine speed of 2400 rev min 1, nd t engine lods of 28 Nm, 70 Nm, 130 Nm, 190 Nm nd 230 Nm, corresponding to rke men effective pressures (BMEP) of 0.08 MP, 0.20 MP, 0.38 MP, 0.55 MP nd 0.67 MP, respectively. At ech engine lod, experiments were crried out using ULSD nd the lended fuels Engine performnce In ech test, the volumetric flow rte of fuel ws mesured nd then converted into the mss consumption rte sed on the density of the fuel. Bsed on the mss consumption rte of the fuel nd the rke power, the rke specific fuel consumption (BSFC) nd rke therml efficiency (BTE) cn e clculted. At ech engine lod, the BSFC increses with the ddition of DGM, especilly in the cse of UD-6. Blending DGM with the diesel fuel reduces the lower heting vlue of the lended fuels, thus more fuel is required to mintin the sme power output. For the engine lod of 0.55 MP nd engine speed of 1800 rev min 1, the BSFC increse re 1.9%, 4.0%, 6.1%, 8.3%, 10.9% nd 39.7%, respectively, for UD-1, UD-2, UD-3, UD-4, UD-5 nd UD-6, compred with tht of ULSD. For the engine lod of 0.55 MP nd engine speed of 2400 rev min 1, the corresponding increse of BSFC re 1.5%, 3.8%, 5.3%, 7.3%, 10.6% nd 45.1%. Similr trends were otined y Znnis et l. (2004) nd Ren et l. (2007). For the lended fuels UD-1, UD-2, UD-3, UD-4 nd UD-5, the BTE increse with the increse of DGM in the fuel, compred with tht of ULSD. However, UD-6 shows shrp decrese in BTE compred with tht of ULSD, indicting tht comustion hs deteriorted when the engine is fueled with UD-6. For the engine lod of 0.55 MP nd engine speed of 1800 rev min 1, the BTE increses y 0.5%, 0.9%, 1.2%, 1.6% nd 1.7% respectively, for UD-1, UD-2, UD-3, UD-4 nd UD-5; ut decreses y 3.6% for UD-6, compred with tht of ULSD. Miymoto et l. (1998) reported n increse in BTE even for pure DGM, which is slightly different with the results of this study. Ren et l. (2007) lso found increse of BTE with the ddition of DGM, however, in their investigtion the mximum oxygen content in the fuel is only 8% y weight. For specific engine lod, the BSFC is higher ut the BTE is lower t the engine speed of 2400 rev min 1 indicting tht the engine performnce is poorer, compred with tht t the engine speed of 1800 rev min 1. For the engine lod of 0.55 MP nd UD-6, the BSFC increses y 15.7% ut the BTE decreses y 3.7% when the engine speeds up from 1800 rev min 1 to 2400 rev min 1. For other lends, similr trend re lso oserved. Ren et l. (2007) lso showed similr trends. The increse in mechnicl frictionl loss, the increse in incomplete comustion nd the reduction in volumetric efficiency t the higher engine speed contriute to the results Emissions The ULSD DGM lends hve significnt influence on smoke nd prticulte emission. In this study, the effects on smoke opcity, prticulte mss concentrtion, SOF, prticle numer concentrtion nd size distriution re investigted Smoke opcity nd prticulte mss emission Fig. 2 () nd () compre the smoke emissions for the two engine speeds. The smoke opcity increses with increse of engine lod nd this is due to the increse of the mount of fuel urned in the diffusion mode. Ren et l. (2007) lso reported similr trends. The smoke opcity is higher t 2400 rev min 1 thn t 1800 rev min 1, which my result from the decrese of volumetric efficiency nd hence decrese of post-flme oxidtion of the soot prticles t the higher engine speed. There is lso decrese in smoke opcity with the increse of oxygen in the lended fuel, nd the reduction is more ovious t the higher engine lod. For the lended fuel contining 20% oxygen, the smoke is lmost zero t the engine lod of 0.20 MP for oth engine speeds. Ren et l. (2007) showed tht the smoke extinction coefficient ws reduced y 3.7% for every 1% increse of the oxygen mss frction in the lended fuel. Miymoto et l. (1998) lso found rpid decrese of smoke emission with the increse of oxygen content, nd the Fig. 2. Effect of oxygente, engine lod nd speed on smoke opcity.

4 58 Y. Di et l. / Atmospheric Environment 44 (2010) Bosch smoke (BS) numer reched zero t the oxygen content of 30 wt-percent. Severl fctors my contriute to the reduction of soot nd hence smoke opcity. Firstly, the oxygen in the lended fuel cn promote comustion during the stge of diffusion comustion. Secondly, the increse of oxygen content corresponds to decrese of cron content in the lended fuels, s shown in Tle 2, leding to the reduction of C C onds in the lended fuel which re source of soot formtion. Thirdly, the decrese of romtics compounds in the lended fuel contriutes to the decrese of soot precursors (Chng nd Gerpen, 1998; Lpuert et l., 2008). Finlly, the temperture increse in diffusion comustion phse with the ddition of DGM (Znnis et l., 2004; Ren et l., 2007) promotes the oxidtion of soot. As shown in Fig. 3, the prticulte mss concentrtion decreses with the increse of oxygen frction in the fuel ut increses with the increse of engine lod. At the engine speed of 1800 rev min 1 nd engine lod of 0.20 MP, the reductions in PM mss concentrtion for UD-1, UD-2, UD-3, UD-4, UD-5 nd UD-6 re 7.2%, 18.0%, 27.3%, 34.8%, 39.4% nd 53.0%, respectively, compred with tht of ULSD; while t the engine speed of 1800 rev min 1 nd engine lod of 0.67 MP, the corresponding reductions re 8.8%,16.2%, 23.7%, 36.9%, 48.3% nd 67.4%, respectively. Thus, the reduction in prticulte mss concentrtion increses with the increse of DGM in the lended fuel. Bennethum nd Winsor (1991) reported 6% nd 20% reductions in PM for 0.5% nd 5% ddition of DGM. Betrice et l. (1999) otined 57% decrese in soot for 30% volume frction of DGM. The prticulte reduction is dependent on the oxygen frction in the ULSD DGM lends. Prticulte mtter consists of soot nd other liquid- or solidphse mterils (Tree nd Svensson, 2007). Lpuert et l. (2007) suggested tht the prticulte emitted y diesel engine is formed minly y soot gglomertes. Tmnouchi et l. (1999) lso reported tht the PM reduction ws ttriuted to the reduction of soot component. Thus the fctors leding to reduction of soot emission lso contriute to the reduction of prticulte emissions, which include the increse of oxygen content nd reduction of the cron content of the lended fuels. To understnd the effect of oxygente ddition on soot formtion, Cheng et l. (2002) numericlly studied soot emission of ethnol nd dimethoxy methne (DMM). Their results showed tht the oxygente components reduce the formtion of soot precursors (nd susequent soot nd PM) through severl key mechnisms. Firstly, the pyrolysis nd decomposition of oxygente (ethnol nd DMM) could reduce the soot precursors. Secondly, high rdicl concentrtions (primrily OH) limits romtic ring growth nd soot prticle inception. Thirdly, high rdicl concentrtions s oxygente ddition promote cron oxidtion to CO nd CO 2, limiting cron vilility for soot precursor formtion. The mechnisms proposed in Cheng et l. (2002) re likely pplicle in this study. In ddition, ccording to Chng nd Gerpen (1998), Xio et l. (2000) nd Lpuert et l. (2008), romtics tend to increse soot emission nd prticulte emission. Thus, the romtics-free DGM dilutes the romtics content of the lended fuel nd hence contriute to the reduction of PM. Moreover, Lpuert et l. (2008) showed tht the reduction of PM could e relized y reducing the sulfur content of the lended fuel nd hence the sulfte in the PM. The sulfur-free DGM lso contriutes to the reduction of sulfte in the PM. Fig. 3 () nd () compre the prticulte mss concentrtions t two engine speeds. The prticulte mss concentrtions t the engine speed of 2400 rev min 1 re higher thn those t 1800 rev min 1. For ULSD, the prticulte mss concentrtions increse y 78.5%, 60.1%, 22.8, 42.4% nd 103.9%, respectively, for the engine lods of 0.08 MP, 0.20 MP, 0.38 MP, 0.55 MP nd 0.67 MP, when the engine speed is incresed from 1800 rev min 1 to 2400 rev min 1. This is consistent with the increse of smoke opcity t the higher engine speed due to the decrese of volumetric efficiency nd soot oxidtion. Besides, s discussed ove, the BSFC is higher ut the BTE is lower t the engine speed of 2400 rev min 1 compred with tht t 1800 rev min 1, which my lso contriute to the higher smoke opcity nd PM emission t the higher engine speed. Prticulte emission converted in rke specific prticulte emission (BSPM) is shown in Fig. 4. For ech fuel, the BSPM emission decreses with the increse of oxygen frction in the fuel. Miymoto et l. (1998) reported tht the BSPM of pure DGM reduced to one-tenth of tht of diesel fuel. In this study, the lended fuel UD-6 contins 53% DGM in volume, nd the BSPM t the engine lod of 0.67 MP decreses y 61.8% for 1800 rev min 1 nd 63.6% Fig. 3. Effect of oxygente, engine lod nd speed on PM mss concentrtion.

5 Y. Di et l. / Atmospheric Environment 44 (2010) Fig. 4. Effect of oxygente, engine lod nd speed on rke specific PM emission. for 2400 rev min 1, compred with tht of ULSD. The results show minimum BSPM t certin intermedite engine lod. This is consequence of the low rke therml efficiency t low engine lod nd high prticulte emissions t high engine lod Solule orgnic frction (SOF) in prticulte mtter Diesel prticulte mtter consists of oth solule orgnic frction (SOF) nd insolule frction (Tree nd Svensson, 2007). In this study, the proportion of SOF in the prticles is nlyzed using the Soxhlet extrction method. The SOF extrction followed the procedures descried in Bgley et l. (1998). Our results show tht the proportion of SOF increses with the increse of oxygen content in the lended fuel. At the engine lod of 0.55 MP nd t the engine speed of 1800 rev min 1, the proportion of SOF re 18.8%, 25.4%, 43.5%, 57.6%, 69.3%, 79.2% nd 88.5%, respectively, for ULSD, UD-1, UD-2, UD-3, UD-4, UD-5 nd UD-6. Tmnouchi et l. (1999) otined similr results when DGM ws lended with diesel fuel. With n increse of oxygente component in the lended fuel, the soot in the prticles is reduced nd the unurned HC dsored on the prticles my increse, contriuting to the increse of SOF proportion in the PM. The increse in SOF proportion is minly due to the increse of the unurned HC dsored on the prticles s well s decrese in the soot content. However, the reduction in prticulte mss emissions indictes tht there is more reduction in soot thn the increse of dsored HC Prticle numer concentrtion nd size distriution Influence of prticles to the environment nd humn helth depends not only on its mss concentrtion ut lso on its numer concentrtion nd size distriution. Smller prticles re considered to e more hrmful thn the lrger ones. In this study, mesurement nd comprison on numer concentrtion nd size distriutions mong different lends were mde. Prticle size distriutions of different lends t three engine lods nd two engine speeds re plotted in Figs. 5 7 while Fig. 8 gives totl Fig. 5. Effect of oxygente on prticulte numer concentrtion nd size distriution for 0.20 MP. Fig. 6. Effect of oxygente on prticulte numer concentrtion nd size distriution for 0.38 MP.

6 60 Y. Di et l. / Atmospheric Environment 44 (2010) Fig. 7. Effect of oxygente on prticulte numer concentrtion nd size distriution for 0.55 MP. numer concentrtion nd GMD of the prticles. In this study, the mesured prticles re in the rnge of nm, thus lrger or smller size prticles, which might exist in the exhust gs, were not mesured. As shown in Figs. 5 7, the size distriution curves re ll unimodl in shpe. For ech engine condition, the size distriution curves shift upwrds nd towrds smller size fter the ddition of DGM in ULSD, indicting n increse of the prticle concentrtion nd n increse of the smller-sized prticles. However, the increse of prticle concentrtion nd smller-sized prticles decrese with increse of DGM in the lended fuel. For exmple, t the engine speed of 1800 rev min 1, the totl numer concentrtions of UD-1 re 3.7 times, 2.1 times nd 2.3 times tht of ULSD, respectively, t the engine lods of 0.20 MP, 0.38 MP nd 0.55 MP. For UD-6, the corresponding vlues decrese to 2.2 times, 1.13 times nd 1.13 times. As shown in Fig. 8 (), for ech engine condition, the GMD of the prticles decreses with the increse of DGM in the lended fuel, this is in line with the shifting trend shown in Figs. 5 7, indicting tht finer prticles re formed fter the ddition of DGM, compred with tht of ULSD. Figs. 5 8 show n increse in numer of prticles in the size rnge of nm while Figs. 3 nd 4 show decrese of prticulte mss concentrtion with the lended fuels. It is possile tht, with the lended fuels, the prticles hve lower density due to the increse in the proportion of SOF nd the reduction of soot nd sulfte. Moreover, there is shift in the GMD indicting tht the prticles re finer. In fct, with the increse of DGM in the lended fuel, the corresponding reduction of cron content, s well s increse of oxygen content in the fuel, re fvorle cuses for reducing soot nuclei nd hence prticle numer concentrtion. However, the cogultion nd gglomertion of the smll prticles to form lrger prticles will lso e slowed down, leding to n increse of the smller-sized prticles nd the decrese of GMD. Moreover, the higher BSFC with the ddition of DGM will lso contriute to the increse of the prticles formed. The fctors mentioned ove my contriute to the increse in the prticle numer concentrtion fter the ddition of DGM into ULSD. However, with the further ddition of DGM, the oxygen enrichment, coupled with the reduction of cron content, my e positive fctor to promote the oxidiztion of primry prticles, leding to decrese of prticle numer concentrtion compred with tht of UD-1. The results otined from the diesel DGM lends re different from those otined y the sme uthors on the sme engine nd operting conditions for diesel iodiesel nd diesel ethnol lends (Di et l., 2009). The diesel iodiesel shows n increse of prticles with increse of iodiesel in the lended fuel, while the diesel ethnol shows decrese of prticles with increse of ethnol in the lended fuel. The results otined from the diesel DGM show n Fig. 8. Effect of oxygente, engine lod nd speed on prticle totl numer concentrtion nd GMD.

7 Y. Di et l. / Atmospheric Environment 44 (2010) increse of prticles over the diesel fuel ut the prticles decreses with increse of DGM in the lended fuel. Such differences etween the different fuels might e consequence of mny fctors, including fuel properties, which could not e clerly identified in this study. For ech engine speed, prticle numer concentrtions increse with the increse of engine lod, which is illustrted in Figs. 5 8 (). As shown in Fig. 8 () for UD-1, the totl numer concentrtion increses y 55.1% when the engine lod increses from 0.20 MP to 0.55 MP for the engine speed of 1800 rev min 1, nd the corresponding increse is 86.4% for the engine speed of 2400 rev min 1. The GMD lso increses with the increse of engine lod. With the increse of the numer of prticles, cogultion rte is incresed nd lrger prticles re formed, leding to the increse of GMD. Besides, Lpuert et l. (2007) suggested tht t higher engine lod, with the higher mount of urning fuel mss, comustion took plce with lower excess oxygen, which coupled with higher pressure nd temperture levels in the comustion chmer, contriuted to soot nucletion nd promoted the growth of the existing soot nuclei. In ddition, Tsolkis (2006) considered tht the increse of prticles is due to the decrese of soot oxidtion when the oxygen content is smll, nd further increse in the oxygen content in the lended fuel would increse the soot oxidtion. It is lso noticed from Fig. 8 () tht for ech fuel, the GMD increses with engine lod s well. In fct, t high engine lods, more fuel is consumed in the diffusion mode nd hence more prticles re formed. With n increse in the numer of prticles, cogultion rte increses nd hence lrger prticles re formed, leding to the increse of GMD. The distriution curves t the engine speed of 2400 rev min 1 re similr to those t 1800 rev min 1. However, higher numer concentrtions nd smller GMD re otined t the higher engine speed. As shown in Fig. 8 for UD-1, the totl numer concentrtions increse y 51.3%, 66.5% nd 81.8% respectively for the engine lods of 0.20 MP, 0.38 MP nd 0.55 MP when the engine speed increses from 1800 rev min 1 to 2400 rev min 1. Despite n increse in the totl numer of prticles t the higher engine speed, the shorter residence time of the prticles in the engine cylinder nd the exhust system results in lower cogultion nd gglomertion of the prticles, leding to lower GMD (Lpuert et l., 2007). The prticles cn e further clssified into two groups. Prticles with dimeters lrger thn 100 nm re referred to s PN >100nm, nd those with dimeters less thn 100 nm re designted s PN <100nm. Prticle toxicity could increse with decresing prticle size due to the higher specific surfce re of smller prticles for dsoring of hrmful chemicls (Peters et l., 1997; Somers et l., 2004; Pope nd Dockery, 2006). It is likely tht nnometer size prticles re more dngerous thn micron size prticles. So it is necessry to investigte the vrition of the PN <100nm prticles with the mount of DGM in the lended fuel. According to Fig. 9, the numer concentrtion of PN <100nm is much higher thn tht of PN >100nm, indicting tht most of the prticles in the engine exhust re less thn 100 nm in dimeter. This is consistent with the prticle size distriutions shown in Figs With the increse of oxygen content in the fuels, for ech engine condition, the numer concentrtion of PN <100nm increses from ULSD to UD-1 nd reches pek vlue when the engine is fueled with UD-1, nd then the numer concentrtion of PN <100nm decreses grdully from UD-1 to UD-6. However, for ech engine condition, the minimum numer concentrtion of PN <100nm, corresponding to UD-6, is still lrger thn tht of ULSD. The trend is similr to tht of the totl prticle numer concentrtion. Besides, the numer concentrtion of PN >100nm or PN <100nm increses with engine lod s well s engine speed, s illustrted in Fig. 9. The percentge of PN <100nm increses with the increse of oxygen content in the lended fuels. At the engine lod of 0.2 MP nd engine speed of 1800 rev min 1, the percentge of PN <100nm is 85.9% for ULSD, ut the corresponding vlues vry from 89% for UD-1 to 92% for UD-6. For the engine speed of 2400 rev min 1, the corresponding vlues vry from 86% for UD-1 to 92% UD-6, indicting the percentge of PN <100nm decreses slightly with the increse of engine speed. Menwhile, the percentge of PN <100nm decreses with n increse of engine lod. This is ecuse the fine prticles usully consist of voltile orgnic compounds nd re formed during dilution nd cooling of the exhust (Kittelson, 1998). Thus, the higher exhust gs temperture t higher engine lod or Fig. 9. Effect of oxygente, engine lod nd speed on numer concentrtion of PN >100nm nd PN <100nm.

8 62 Y. Di et l. / Atmospheric Environment 44 (2010) PN <100nm nd totl prticle numer concentrtions re incresed. While the percentges of PN <100nm in the totl numer of prticles nd the geometric men dimeter re decresed. The percentges of PN <100nm is decresed with the increse of engine lod. Smoke opcity, prticulte mss concentrtion, geometric men dimeter, PN <100nm, nd totl prticle numer concentrtions re incresed with the increse of engine lod. 3. ULSD DGM lends cn effectively reduce smoke, prticulte mss concentrtion nd simultneously NO x emission. Acknowledgement The uthors would like to thnk the Hong Kong Polytechnic University (Project Numer GU-319), Ntionl Nturl Science Fund of Chin ( , ), nd Xi n Jiotong University for supporting this project. Fig. 10. Trdeoff-curve etween NO x nd prticulte mss concentrtion. higher engine speed prevents, to certin extent, the voltile orgnic compounds from condenstion, leding to the decrese in percentge of PN <100nm prticles Trde-off etween PM nd NO x In conventionl diesel engine, there exists trdeoff-curve etween PM nd NO x emissions, s shown in Fig. 10. The reson is tht, for diesel engine, oxygen deficiency in the cylinder increses PM emission, while oxygen enrichment increses NO x emission, resulting in the trdeoff-curve etween PM nd NO x emissions. With minor engine modifiction it is difficult to chieve simultneous reduction of oth pollutnts. Ren et l. (2007) showed tht, in conventionl diesel engine, postponing fuel delivery ngle could reduce NO x emission ut increse the smoke. Bertoli et l. (1998), Cheng et l. (2002) nd Song et l. (2002) demonstrted tht engine modifictions coupled with oxygented lends could simultneously reduce oth pollutnts. Murym et l. (1995) reported the simultneous reduction of smoke nd NO x in DI diesel engine using EGR nd dimethyl cronte (DMC). Previous work showed tht diesel DGM lends could simultneously reduce NO x nd smoke (Ren et l., 2007). Blending DGM with the diesel fuel cn simultneously reduce PM nd NO x. The increse of oxygen content in the lended fuel reduces the PM. The increse of DGM shortens the ignition dely, the mount of premixed comustion, leding to reduction of NO x emissions. 5. Conclusions Experiments were conducted on diesel engine using ultrlowsulfur diesel lended with diglyme (DGM). The min results re summrized s follows. 1. With the increse of DGM, nd hence oxygen content, in the lended fuels, smoke opcity, prticulte mss concentrtion, rke specific prticulte emission nd the geometric men dimeter of prticles re reduced. The percentges of the PN <100nm in totl numer concentrtion, s well s the solule orgnic frction, re incresed. The PN <100nm nd totl prticle numer concentrtions re incresed in the cse of diesel DGM lend thn tht of ULSD, ut these concentrtions follow decresing trend with the increse of DGM in the lended fuel. 2. With the increse of engine speed, smoke opcity, prticulte mss concentrtion, rke specific prticulte emission, References Agrwl, A.K., Biofuels (lcohols nd iodiesel) pplictions s fuels for internl comustion engines. Progress in Energy nd Comustion Science 33, Bgley, S.T., Grtz, L.D., Johnson, J.H., Effect of n oxidtion ctlytic converter nd iodiesel fuel on the chemicl, mutgenic, nd prticle size chrcteristics of emissions from diesel engine. Environmentl Science nd Technology 32, Bimridge, C.L., Rouge, B., Bolomey, P.V., Love, J.D., Method of producing glycol ethers. United Sttes Ptent Appliction Puliction No. US 2004/ A1. Betrice, C., Bertoli, C., Gicomo, N.D., Migliccio, M.N., Potentility of oxygented synthetic fuel nd reformulted fuel on emissions from modern DI diesel engine. SAE Technicl Pper No Bennethum, J., Winsor, R., Towrd improved diesel fuel. SAE Technicl Pper No Bertoli, C., Alessio, J.D., Gicomo, N.D., Lzzro, M., Mssoli, P., Mocci, V., Running light-duty DI diesel engines with wood pyrolysis oil. SAE Technicl Pper No Bertoli, C., Gicomo, N.D., Betrice, C., Migliccio, M.N., Evlution of comustion ehvior nd pollutnts emission of dvnced fuel formultions y single cylinder engine experiment. SAE Technicl Pper No Chng, D.Y., Gerpen, J.H.V., Determintion of prticulte nd unurned hydrocron emissions from diesel engines fueled with iodiesel. SAE Technicl Pper No Cheng, A.S., Dile, R.W., Buchholz, B.A., The effect of oxygentes on diesel engine prticulte mtter. SAE Technicl Pper No Demirs, A., Progress nd recent trends in iofuels. Progress in Energy nd Comustion Science 33, Di, Y.G., Cheung, C.S., Hung, Z.H., Comprison of the effect of iodiesel diesel nd ethnol diesel on the prticulte emissions of direct injection diesel engine. Aerosol Science nd Technology 43, Hnsen, A.C., Zhng, Q., Lyne, P.W.L., Ethnol diesel fuel lends review. Bioresource Technology 96, Kittelson, D.B., Engine nd nnoprticles: review. Journl of Aerosol Science 29, Kline, S.J., McClintock, F.A., Descriing uncertinties in single smple experiments. Mechnicl Engineering 75, 3 8. Lpuert, M., Mrtos, F.J., Herreros, J.M., Effect of engine operting conditions on the size of primry prticles composing diesel soot gglomertes. Journl of Aerosol Science 38, Lpuert, M., Arms, O., Jose, R.F., Effect of iodiesel fuels on diesel engine emissions. Progress in Energy nd Comustion Science 34 (2), Lpuert, M., Arms, O., Herreros, J.M., Emissions from diesel ioethnol lend in n utomotive diesel engine. Fuel 87, Miymoto, N., Ogw, H., Nurun, N.M., Ot, K., Arim, T., Smokeless, low NO x, high therml efficiency, nd low noise diesel comustion with oxygented gents s min fuel. SAE Technicl Pper No Murym, T., Zheng, M., Chikhis, T., Simultneous reduction of smoke nd NO x from DI diesel engine with EGR nd dimethyl cronte. SAE Technicl Pper No Peters, A., Wichmnn, H.E., Tuch, T., Heinrich, J., Heyder, J., Respirtory effects re ssocited with the numer of ultr fine prticles. Americn Journl of Respirtory nd Criticl Cre Medicine 155, Pope III, C.A., Dockery, D.W., Helth effects of fine prticulte ir pollution: lines tht connect. Journl of the Air & Mngement Assocition 56, Ren, Y., Hung, Z.H., Mio, H.Y., J, D.M., Zeng, K., Liu, B., Wng, X.B., Effect of the ddition of diglyme in diesel fuel on comustion nd emissions in compression-ignition engine. Energy & Fuels 21,

9 Y. Di et l. / Atmospheric Environment 44 (2010) Rieiro, N.M., Pinto, A.C., Quintell, C.M., Roch, G.O.D., Teixeir, L.S.G., Gurieiro, L.L.N., Rngel, M.D.C., Veloso, M.C.C., Rezende, M.J.C., Cruz, R.S.D., Oliveir, A.M.D., Torres, E.A., Andrde, J.B.D., The role of dditives for diesel nd diesel lended (ethnol or iodiesel) fuels: review. Energy & Fuels 21, Shih, L.K., Comprison of the effects of vrious fuel dditives on the diesel engine emissions. SAE Technicl Pper No Somers, C.M., McCrry, B.E., Mlek, F., Quinn, J.S., Reduction of prticulte ir pollution lowers the risk of heritle muttions in mice. Science 304, Song, J., Cheenkchorn, K., Wng, J., Perez, J., Boehmn, A.L., Yong, P.J., Wller, F.J., Effect of oxygented fuel on comustion nd emissions in light-duty turo diesel engine. Energy & Fuels 16, Tmnouchi, M., Akimoto, T., Aihr, S., Morihis, H., Effects of DGM nd oxidtion ctlyst on diesel exhust emissions. SAE Technicl Pper No Tree, D.R., Svensson, K.I., Soot processes in compression ignition engines. Progress in Energy nd Comustion Science 33, Tsolkis, A., Effect on prticle size distriution from the diesel engine operting on RME Biodiesel with EGR. Energy & Fuels 20, Wong, C.P., Chn, T.L., Leung, C.W., Chrcteriztion of diesel exhusts prticle numer nd size distriutions using mini-dilution tunnel nd ejector-diluter mesurement techniques. Atmospheric Environment 37, Xio, Z., Ldommtos, N., Zho, H., The effect of romtic hydrocrons nd oxygentes on diesel engine emissions. Proceedings of the Institution of Mechnicl Engineers Prt D Journl of Automoile Engineering 214, Znnis, T.C., Hountls, D.T., Kouremenos, D.A., Experimentl investigtion to specify the effect of oxygented dditive content nd type on DI engine performnce nd emissions. SAE Technicl Pper No