Potential of PN Reduction in Passenger Cars with DPF and GPF

Abstract

Particle number concentration (PN) in vehicle exhaust and ambient air describes the number of ultrafine particles (UFPs) below 500 nm, which are recognized as a toxic and carcinogenic component of pollution and are regulated in several countries. Metal nuclei, ash, and organic matter contribute significantly to the ultrafine particle size fraction and, thus, to the particle number concentration. Exhaust gas filtration is increasingly being used worldwide to significantly reduce this pollution, both on diesel particulate filter (DPF) and gasoline particulate filter (GPF) engines. In recent years, the EU has also funded research projects dealing with the possibilities of retrofitting gasoline vehicles with GPFs. This paper presents the results and compares the PN emissions of different vehicles. An original equipment manufacturer (OEM) diesel car with a DPF is considered as a benchmark. The PN emissions of this car are compared with a CNG car without filtration and with gasoline cars equipped with GPFs. It can be concluded that the currently used GPFs still have some potential to improve their filtration efficiency and that a modern CNG car would still have remarkable possibilities to reduce PN emissions with an improved quality GPF.

1. Introduction

The invisible nanoparticles (NPs) from combustion processes, the so-called ultrafine particles (UFPs), easily penetrate into the human body through the respiratory and olfactory pathways and carry numerous health hazard potentials [1,2,3].

Several studies have shown that gasoline cars, even with MPI (PFI), have very high PN emissions in certain driving situations, comparable to diesel engines without exhaust filtration. The extension of the PN limits in the EU legislation to all gasoline vehicles, not only to gasoline direct injection (GDI), is proposed [4,5]. China introduced the PN limit (CN 6/Euro 6; 6 × 1011#/km) for all light-duty vehicles (LDVs) (including gasoline) on 7 January 2020.

Some recent studies by institutions concerned with health and environmental protection require a radical reduction in PN limits in workplaces and for the current particle mass PM 10 or PM 2.5 limits, as well as improved PN monitoring and PN limits in ambient air quality control [6,7].

In this situation, the extended use of exhaust gas filtration (DPF, GPF), which is actually a mature technology, is an important contribution to the improvement of air quality in highly frequented places.

The nanoaerosols in vehicle exhaust are a complex mixture of different volatile and non-volatile species. They often exhibit a bimodal particle size distribution with a nucleation mode that is smaller than 20 nm and a larger accumulation mode containing aggregates of primary particles.

The larger accumulation mode is usually composed of graphitic soot particles with an elemental carbon (EC) structure, whereas the nucleation mode particles are reported to be mainly volatile organics, especially in the absence of sulfur in fuel and lubricating oil [8,9,10,11,12]. However, in-depth studies have also detected low-volatility particle fractions in the ultrafine size range when sampling according to the Particle Measurement Program of the GRPE (PMP) protocol at 300 °C [13,14,15].

The particles in the nucleated mode are thought to be nucleated metal oxides originating from metal additives in lubricating oil or fuels [16,17,18,19,20,21]. The formation of this fraction of particles is often observed at low soot levels, such as in the idling condition of diesel vehicles. These particles are mainly in the ultrafine size range < 23 nm. While the mass contribution of these ultrafine particles in vehicle emissions is very small, their contribution to the number concentration is significant. In addition, these ultrafine particles can contribute to the surface composition of the aerosol and, therefore, have a significant impact on the health effects associated with pollution.

Studies of gasoline internal combustion engines have shown that this class of vehicles can also emit significant amounts of particles [22,23,24,25]. In particular, gasoline direct injection (GDI) technology has significantly higher particle number (PN) emissions than modern diesel vehicles equipped with the best available DPF technology. As the trend for gasoline vehicles with GDI technology is increasing, a significant increase in emissions is predicted in the near future.

In the second and third decade of this century, there was much discussion and research on ultrafine particles < 23 nm [26,27,28,29], which led to the introduction of sub 23 nm particles in the Euro 7 emissions legislation, extending the measured and restricted size range down to 10 nm [30,31]. The particular health hazards of these UFPs were taken into account.

The nanoparticle emissions are mainly produced during cold start and warm-up conditions and during dynamic engine operation [32,33,34]. The lubricating oil contributes to these emissions in terms of the nuclei mode concentration and composition [18,19,20,21].

Studies of the morphology of nanoparticles from direct-injection gasoline engines have revealed mainly graphitic structures, which can store some metal oxides under certain conditions and can overlap with condensates [35,36].

Car manufacturers and exhaust after-treatment technology suppliers offer several mature solutions of GPF for the efficient elimination of nanoparticles from direct-injection spark-ignition (DI SI) engines [37,38,39,40,41].

There are also nanoparticle emissions from gasoline vehicles with multipoint port injection (MPI). Some of them emit high amounts of PN and PM. In a study conducted by the Emission Control Laboratory FH Biel, Switzerland (AFHB), it was found that an older model with multipoint port injection (MPI) emitted up to 4 orders of magnitude more nanoparticles than a lower-emitting GDI car at stationary part load operation. The main reason for this increased PN emission was attributed to the increased lube oil consumption. Nevertheless, a lower-quality mixture preparation cannot be excluded.

MPI technology has a large share of the world market due to its lower cost and simplicity, and in several countries, this technology will remain the primary option for several years to come.

From this point of view, and taking into account the progressing emission legislation aimed at increased care for health and environmental protection, it is necessary to include vehicles with MPI in the efforts to reduce PN and PM.

The investigations in this paper have been carried out at AFHB.

After describing the vehicles tested, the test methods, and the instrumentation, this paper presents comparisons between the PN emissions of vehicles operating on different fuels, with different fuel/air mixture preparation systems, and with different exhaust gas filtration devices. The results presented relate to the following issues, which have not yet received sufficient attention from experts and emission legislation:

-

GPFs on MPI vehicles;

-

Stronger coating of 4WC (4-way catalytic converter);

-

GPFs on GDI vehicles;

-

PN emissions of CNG cars.

2. Experimental Section

2.1. Tested Vehicles

Table 1. Data of investigated vehicles.

	V1 Fiat Panda 4 × 4 TwinAir	V2 Renault 18 Break	V3 Renault Mégane Scénic RX4	V4 Volvo V60 T4F	V5 Opel Insignia 1.6 EcoFlex	V6 Peugeot 4008 1.6 HDi SST	V7 Audi A3 Sportback g-tron
Number and arrangement of cylinders	2/in line	4/in line	4/in line	4/in line	4/in line	4/in line	4/in line
Displacement cm3	875	2164	1998	1596	1598	1560	1395
Power kW	66.2 @ 5500 rpm	74 @ 5000 rpm	101.5 @ 5500 rpm	132 @ 5700 rpm	125 @ 6000 rpm	84 @ 3600 rpm	81 @ 6000 rpm
Torque Nm	145 @ 1900 rpm	162 @ 2000 rpm	188 @ 3750 rpm	240 @ 1600 rpm	260 @ 1650–3200 rpm	270 @ 1750 rpm	200 @ 1500 rpm
Injection type	MPI	MPI	MPI	DI	DI	DI	DI/MPI
Curb weight kg	1200	1110	1495	1554	1701	1462	1355
Gross vehicle weight kg	1585	1585	1990	2110	2120	2060	1820
Drive wheel	AWD	Front-wheel drive	AWD	Front-wheel drive	Front-wheel drive	Front-wheel drive	Front-wheel drive
Gearbox	m6	m5	m5	a6	m6	m6	m6
First registration	2017	1985	2001	2012	2014	2013	2015
Exh. after-treatment	Euro 6b	Euro 0	Euro 3	Euro 5a	Euro 5b+	Euro 5b	Euro 6b
Fuel	gasoline					Diesel	CNG

shows the main data of the vehicles tested. In different test series, the PN emissions of gasoline vehicles (MPI or GDI) retrofitted with different GPFs were analyzed and compared with the emissions of a modern CNG vehicle (without exhaust gas filtration) and with the emissions of a diesel vehicle with an OEM DPF. The gasoline vehicles of different ages and emission classes were selected in such a way that special effects and comparisons were possible. Particularly noteworthy are the following: V2, older vehicle (1985) with increased oil consumption; V1, modern (2017) vehicle with a downsized engine and a particularly high PN emission load; and V3, a much-used vehicle, which was available for the endurance runs. All three of these vehicles are equipped with multipoint port injection (MPI). The other two (V4 and V5), on the other hand, represent part of the fleet with direct injection (DI).

2.2. Fuels and Lube Oils

The fuels used were from the Swiss market: gasoline RON 95 according to SN EN228, [42] and diesel according to SN EN590, [43]. For all vehicles, the lubricating oils were not changed and not analyzed.

2.3. Test Methods and Instrumentation

The vehicles were tested on a chassis dynamometer at constant speeds (or idling) and in the dynamic driving cycles with cold and warm engine start.

Chassis dynamometer—the following test systems were used:

• Roller dynamometer: AFHB GSA 200;
• Driver conductor system: Tornado, version 3.3.;
• CVS dilution system: Horiba CVS-9500T with Roots blower;
• Air conditioning in the hall is automatic (intake and dilution air).

The driving resistances of the test bench were set according to the legal prescriptions, responding to the horizontal road (table values according to ECE R83).

The test equipment fulfills the requirements of the Swiss and European exhaust gas legislation. The test equipment for regulated exhaust gas emissions was as follows:

• Gaseous components:
• Exhaust gas measuring system Horiba MEXA-9400H [44];
• CO, CO₂—infrared analyzers (IR), HCIR, HCFID, NO/NOX, CLA.

The dilution factor (DF) in the constant volume sampling (CVS) dilution tunnel is variable and can be controlled by means of the CO₂ analysis.

Nanoparticle Analysis

The measurements of nanoparticle size distributions were conducted with different scanning mobility particle sizer (SMPS) systems, which enabled different ranges of size analysis at steady-state operations [45,46]:

SMPS: DMA TSI 3081 and CPC TSI 3772 (10–429 nm) (DMA, differential mobility analyzer).

nSMPS: nDMA TSI 3085 and CPC TSI 3776 (2–64 nm) (CPC, condensation particle counter).

For the dilution and sample preparation, an aerosol sampling and evaporation tube (ASET) system from Matter Aerosol was used (Figure 1) [47]. This system contains the following:

• Primary dilution—MD19 tunable rotating disc minidiluter (Matter Eng. MD19-2E).
• Secondary dilution—dilution of the primary diluted and thermally conditioned sample gas on the outlet of the evaporative tube.
• Thermoconditioner (TC)—sample heating at 300 °C.

This sample preparation system fulfills the requirements of PMP and was used for all measurements. At steady-state operation, this system worked with summary dilution factors DF = 65 to 765.

The estimated accuracy of the PN measurement in the size range of 80–120 nm with DF = 100 is ±6%.

For the measurements of summary PN at transient operation, a condensation particle counter (CPC) TSI 3790 (PMP conform) was used.

In the tests, the gas sample for nanoparticle analysis was taken from the undiluted exhaust gas at the tailpipe for stationary operations of the scanning mobility particle sizer (SMPS) or from the diluted exhaust gas in the CVS tunnel for transient operations (CPC). The schematic of the general sampling set-up is shown in Figure 2. The use of SMPS systems allows the recording of particle size distributions (PSDs) and, thus, an in-depth analysis of the measured aerosol, in addition to the summary measurement of all particle sizes with the CPC.

2.4. Driving Cycles and Test Routine

The worldwide light-duty test cycle (WLTC), as part of the worldwide light-duty test procedure (WLTP) [48], represents typical driving conditions around the world. This cycle, Figure 3, was also used in this study. It represents different driving conditions: urban, rural, highway, and extra highway.

In the test program with MPI vehicles, the ADAC high-speed cycle (ADAC 130) [49] was also used. The ADAC 130 cycle represents highway driving and requires some full-load accelerations and higher load operations for the vehicle class investigated. It was decided to use this drive cycle to achieve higher temperatures and higher gas flow rates. This also better matches the operating conditions of the exhaust after-treatment elements in gasoline vehicles.

Figure 3 shows the time courses, and

Table 2. Research conditions.

Cycle	Duration s	Distance m	Vmax km/h	amax m/s2	amin m/s2
WLTC	1800	23,262	131	1.58	−1.49
ADAC130	740	18,755	130	6.94	−5.00

summarizes the most important data of these driving cycles.

All tests were carried out by the same specialized driver. The audited repeatability of the standard exhaust emissions (CO, HC, NOx) in dynamic test cycles is ±2%.

The diluted exhaust gas components were measured online from the bags and evaluated after the cycle. For the cold start, the vehicle was conditioned the day before: the conditioning for WLTC cold was another WLTC warm. The conditioning for the warm cycle was 3 min 80 km/h and 1 min idling. This type of conditioning includes the legal requirement for the engine temperature but also takes into account the conditioning of the exhaust gas after-treatment system.

All tests were performed on a 4WD chassis dynamometer, and for front drive, the rear wheels were pulled at the same speed as the front wheels.

3. Results and Discussion

3.1. GPFs on MPI Vehicles

Figure 4 compares the PN emissions of vehicles V1, V2, and V3 in two driving cycles, WLTCc and ADACw. Vehicles V1 and V2 were equipped with coated GPFs (cGPFs) or uncoated GPFs added after the original TWC, the so-called “add-on” retrofit. In the V3 vehicle, the original TWC was replaced with a TWC-coated filter substrate. This solution is designed as a 4-way catalytic converter (4WC). The reference means were from the vehicle in its original condition without exhaust filtration.

It can be noted that in the WLTCc (real legal cycle), the unfiltered PN emissions are higher than the limit, and filtration can reduce the PN below the limit. On vehicle V1, the catalytic GPF (cGPF) is slightly more efficient than the uncoated one (GPF). This is due to the reduction in pore size by the presence of a washcoat and coating.

The catalytic activity of this cGPF is confirmed in Figure 5 by a greater reduction in CO and HC in both cycles (WLTCc and ADACw). In ADAC 130, there is no reduction in PN by filtration for vehicles V1 and V3. This is due to the driving cycle: in this high-speed cycle, the engines work at high or full load, and the filtration efficiency drops almost to zero.

The drop in filtration efficiency is due to the highest spatial velocities combined with the highest filter temperatures. Figure 6 shows an example of the instantaneous filtration efficiency and exhaust gas temperature before the GPF in this high-speed cycle for vehicle V1. The peak values of T_EXH > 700 °C are very high, and the loss of diffusion filtration during acceleration events is obvious. Over the duration of the cycle, the temperature increases while the average filtration efficiency decreases. (The filter was not damaged in this test, as further results with high filtration efficiencies were obtained afterward).

Why is there a significant reduction in PN for vehicle V2 in the ADAC 130? V2 is an older vehicle with higher lube oil consumption. Figure 7 compares the particle size distribution of this vehicle with that of a newer vehicle at part-load engine operation. The strongly increased particle numbers of the V2 in the size range below 50 nm are typical for a high presence of lubricating oil in the combustion [17,50,51]. This aerosol consists mainly of lubricating oil droplets, which, in the given example of the ADAC 130 cycle, are oxidized in the very hot GPF with sufficient oxygen supply during the deceleration phases. In this case, the retrofitted (add-on) GPF acts as an additional thermal reactor.

3.2. Stronger Coating of 4WC

In addition to the investigation with the base coating of the 4WC (cGPF1), another 4WC with a stronger coating (cGPF2) was tested on vehicle V3. This cGPF2, with approximately 30% more catalytic coating, was provided by the industrial partner for research purposes with the expectation that it would improve filtration efficiency by compensating for the effect of large pore sizes.

Figure 8 compares the average emission results of CO, HC, and particulate counts of filtration efficiencies (PCFE) for the two cGPFs investigated (4WCs).

It can be seen that the thicker coating of cGPF2 resulted in slightly lower emissions (better conversion of components) and slightly higher particulate counts of filtration efficiency (PCFE).

The definition of PCFE is: $\mathrm{P}\mathrm{C}\mathrm{F}\mathrm{E}={\displaystyle \frac{\mathrm{P}\mathrm{N}\mathrm{w}/\mathrm{o}-\mathrm{P}\mathrm{N}\mathrm{w}\mathrm{i}\mathrm{t}\mathrm{h}\mathrm{G}\mathrm{P}\mathrm{F} }{\mathrm{P}\mathrm{N}\mathrm{w}/\mathrm{o}}}$.

The case of “negative PCFE” can occur as a result of the following:

-

Amplitude scatter of both CPC traces in the calculation interval;

-

Scattering of the phase of both signals in the calculation interval;

-

Artifact: formation of nanoparticles from substances that pass through the cGPF as volatiles and nucleate after the GPF—an increase in PN after the cGPF.

The authors suggest that these effects are also present in the results of Figure 8.

After 4000 km, there are slightly higher PCFE values in WLTC compared to a new cGPF. Nevertheless, the filtration quality is rather poor compared with the experience with diesel exhaust filtration [52]. At the end of the WLTC and in ADAC 130, the exhaust gas temperature in front of the cGPF in this car exceeds 800 °C several times.

In summary, the thicker coating has a visible but not significant effect on the PCFE. A much more important parameter is the average pore size of the filter substrate.

With both 4WCs, a long-term test for GPF soot loading was performed as a real operation over 4000 km. The following figures show some results of the soot loading test.

Figure 9 shows the results of the filter weighing and the appearance of the filter inlet side after the long run.

During the tests, the GPF was weighed at the same temperature of 280 °C, and no increase (even a slight decrease) in weight was observed over 4000 km. Weighing (with an accuracy of 0.1 g) was always carried out at the same temperature in order to avoid the influence of moisture condensation.

For real road driving tests, a German highway was sometimes used, and the maximum temperature before the GPF rose above 850 °C. The entry surfaces of the GPFs were cleaned after long runs, and no traces of soot were visible. It should be noted that such high exhaust temperatures, which can occur at unlimited driving speeds on the highway, contribute to the aging of the catalytic converters.

Figure 10 shows examples of the data logger statistics of the pressure and temperature before the filter in the last parts of the soot loading test.

The data logger evaluation of p/T before 4WC shows that there are very high-temperature events that contribute to the self-regeneration of the GPF. The mean values of the back pressure and the standard deviations remain nearly constant during the whole test, up to 4000 km. In conclusion, the soot loading of the studied cGPFs on this vehicle is not present in these tests. Finally, it must be said about this attempt that the term “soot loading test” was incorrect because it suggested something that did not exist. As for the ash loading, it should be noted that the applied test distance is too short [37,38,39].

3.3. GPFs on GDI Vehicles

Extensive research on nanoaerosols from GDI vehicles has been conducted in the Swiss network with the Swiss Federal Material Testing Institute (EMPA), Paul Scherrer Institute (PSI), Verification of Emission Reduction Technologies, Industrial Association (VERT), AFHB, and Instytut Nafty i Gazu (INIG) [53,54]. In the following figures, the most important information on the GPF particle count filtration efficiency (PCFE) (state of the art 2016–2018) is given. Figure 11 compares the filtration efficiencies of different GPFs on two GDI vehicles (V4 and V5). The PCFE values are calculated according to the above formula, taking into account the integral average of the PN emissions (CPC) over the WLTC. GPF1 actually contains a serial DPF substrate and, of course, has an excellent filtration quality, as required for DPFs [52]. The other three substrates were serial GPFs. All GPFs in these test series were mounted on the tailpipe of the vehicles and were, therefore, exposed to less pressure pulsation and lower flow and temperature peaks during acceleration. These conditions are favorable for higher PCFEs.

Figure 12 shows an example of the particle size distributions (PSDs) measured with SMPS and with nSMPS on vehicle V4 at 95 km/h. There are variants without GPFs and with two GPFs of different filtration qualities.

In general, there is a very good agreement in PSDs measured with both systems SMPS and nSMPS in the common size range (10–64 nm). The difference in filtration efficiency of both GPFs is well demonstrated. It is important to note that as the filtration efficiency increases, the particles in the sub-23 nm size range are increasingly eliminated. This is related to the deep filtration (or diffusion filtration) mechanism, which is based on the molecular movement of the UFPs and the superficial Van der Waals forces that trap and attach the UFPs to the walls of the filtration material. Porosity and pore size are the defining parameters of PCFEs.

It can be said that from a health and environmental perspective, it makes more sense to improve filtration quality than to try to more accurately measure the sub-23 nm size range, which would still be eliminated by the increased filtration efficiency. This approach would also be much easier and quicker for the legislator.

3.4. PN Emissions of CNG Car

In this section, some PN results of a modern CNG (V7, no exhaust filtration) are compared with the emissions of a diesel car with original DPF (V6) and a vehicle V2 (gasoline MPI) retrofitted with an uncoated add-on GPF.

Figure 13 shows a comparison of the PN emissions on the WLTC cold test cycle. It is clear that the OEM DPF (V6) sets the benchmark for the best filtration and lowest PN emissions. The filtration quality of the GPF used (V2) is mediocre and has room for improvement. CNG (V7) has the highest integral average PN emissions in phases 2, 3, and 4 of the cycle, a drawback that could undoubtedly be significantly reduced with high-quality filtration. In phase 1 of the cycle, all three variants have the highest PN emissions, confirming the high contribution of cold start and the first part of the cold dynamic operation to PN.

Further illustrations, analyses, and confirmation of these relationships are given in Figure 14 and Figure 15 below.

Figure 14 shows the particle size-dependent PN emissions measured with SMPS and nSMPS at two steady-state operating conditions (95 km/h and idle).

With DPF (diesel), excellent filtration efficiency is confirmed with near-zero emissions. These PN concentrations are at or below ambient air levels (which are in the range of 10³#/cc).

In the SMPS size range, the PN concentrations of CNG (V7) and GPF (V2) are similar. In the sub-23 nm size range, there are no UFPs at all for diesel and CNG.

For GPF (V2), there are sporadic particle counts down to 10 nm at 95 km/h and a little more down to 7 nm at idle. These PN emissions from GPF (V2) are caused by the composition of the aerosol, which contains a high amount of heavy HC from the lube oil. In this case, nucleation and/or pyrolysis are possible even after GPF. Solid or semi-solid UFPs are formed and cannot be completely eliminated by the sample preparation of the measuring system. At high speeds, the GPF is hot enough to promote more intense oxidation of the heavy HC (SOF), and there are fewer UFPs. The V2 test vehicle provided a useful service in representing older vehicles with higher lube oil consumption.

Figure 15 summarizes the integral average emissions measured with SMPS and nSMPS at three steady-state operating conditions (95 km/h, 45 km/h, and idle). This plot confirms the following results:

-

PN emissions of diesel with DPF are zero or negligible, which sets a quality benchmark;

-

The PN emissions of the CNG vehicle (without filtration) are higher or equal to the PN emissions of the high-emitting gasoline vehicle with GPF (V2), with the exception of idling, where the effect of lube oil consumption and aerosol composition of V2 is predominant.

With regard to the possible introduction of GPF, the disadvantage of increased fuel consumption (fc) due to increased backpressure is often mentioned. There are some tendencies for increased fc, especially at higher constant-load operating points, but not at transient cycles [32]. There is no fuel penalty in the average driving conditions performed—mainly low and partial loads.

The fuel penalty can, of course, occur with soot or ash-loaded filters, especially at high-power operation. For DPFs on work machinery, the average fuel penalty is 2–3% [52].

According to the current state of knowledge of the research team, the soot loading of the GPF will occur very rarely or not at all due to much higher exhaust temperatures, much lower soot masses, and higher oxygen availability during operations. External regeneration (if any) will be very rare.

Ash loading is, logically, to be expected, but only after very long operating distances [37,38]. The fuel penalty to be expected over the lifetime of a GPF vehicle is likely to be significantly lower than for a DPF in heavy-duty applications.

For unexpected cases of filter plugging, e.g., due to increased lube oil consumption or injector damage, OEMs apply precautionary strategies in the OBD system.

4. Conclusions

The most important findings can be summarized as follows:

• The present work demonstrated that gasoline vehicles with MPI can emit a significant amount of PN emissions.
• With GPFs, it is possible to lower the emissions below the current European limit of 6.0 × 10¹¹#/km.
• The particle count filtration efficiency (PCFE) of the investigated GPFs (and 4WCs) is considerably lower than the PCFE of the DPFs studied. (This drawback has recently been addressed, and new filter substrates for GPFs are available that produce the same high-quality PCFEs as the DPFs [55]).
• The thicker coating of a 4WC has a visible but insignificant effect on PCFE; the more important parameter is the average pore size of the substrate.
• As filtration efficiency increases, particles in the sub-23 nm size range are increasingly eliminated; with high-quality filtration (such as DPF), there are no sub-23 nm particles at all.
• The CNG vehicle (without exhaust filtration) emits average particulate concentrations up to two orders of magnitude higher than diesel with DPF during some phases of the WLTC.
• In the real road tests of two 4WCs, no increase in weight or backpressure could be observed up to 4070 km. (After completion of the paper, operation with a GPF was continued, and positive results were confirmed until the GPF was removed at over 17,000 km).

The present research showed some trends of significantly increased PN emissions. With this knowledge and considering the immense multiplication factor of MPI vehicles and the upcoming gas propulsion worldwide, the legal PN limits for gasoline and gas engines should be advanced quickly. Filtration offers excellent potential for emission reduction.