The present section analyses the behavior of the two vehicles using the methodology included in the European Commission’s proposal for the future Euro 7 regulation.

Fig. 4, depicts the aggregate results of all the WHVC_Cold and WHVC_Hot experiments performed at different ambient temperatures, −7, 0, 22 and 35 °C. A summary with the emission factors can be found in Table S2 of the SM. It should be noted that Fig. 4 provides a general overview of the vehicles’ emission performance without claiming to be statistically representative, due to the limited number of tests and the numerous effects taken into account. It should be considered as an indication of the variability in emissions within broadly defined operative conditions boundaries. Furthermore, it is important to highlight that the tests conducted did not cover all the potential usage scenarios. The median emissions from the two buses (see Table S2 of the SM) show that both vehicles would comply with the Cold and Hot limits proposed for HCHO, CH₄ and NH₃. The emissions of HCHO were very low (<0.001 g/kWh) for both vehicles during all WHVC tests. The highest emissions, 0.002 g/kWh, were registered from the CNG vehicle at −7 °C. The emissions were in line with previous studies (Lähde and Giechaskiel, 2021, Suarez-Bertoa et al., 2021).

Although below the limits, the CNG presented substantially higher median emissions of CH₄ and NH₃ (0.388 g/kWh and 0.034 g/kWh for CH₄ and NH₃, respectively) during the WHVC_Cold tests as compared to the Diesel, for which the emissions were very low (less than 0.002 g/kWh), as expected. The CNG’s median emissions of CH₄, a short-lived greenhouse gas (atmospheric lifetime of 12 years) with a global warming potential (GWP) over 20-year timeframe 84 – 87 times higher than CO₂, were equivalent to 33 gCO₂eq/kWh or 3.8% of the CO₂ emitted by the vehicle. The CNG emissions decreased during the hot operation but NH₃, with 18 mg/kWh, remained far from negligible and higher than the Diesel.

NH₃ is an important precursor of PM2.5 and plays an important role in the eutrophication of waters (Suarez-Bertoa et al., 2019, Pusede, 2015). Different pathways lead to the emission of NH₃ from the technologies studied. On one hand, the NH₃ emissions from spark ignition engines equipped with TWC are related to fuel enrichment after the light-off of the TWC (Bouwman et al., 2002, Heeb et al., 2008). During fuel rich events, H₂ is formed from the reforming of hydrocarbons, which by reacting with NOx over the TWC forms NH₃. Indeed, it has been shown that NH₃ correlate well with CO emissions after the light-off of the TWC regardless of being heavy-duty or light-duty vehicles (Giechaskiel et al., 2022a, Selleri et al., 2021). As for CO emissions, NH₃ emissions from TWC equipped vehicles are often controlled by adjusting the air-to-fuel ratio. The presence of NH₃ in the exhaust of engines using SCR is linked to the use of DEF. There are several physical and chemical events, reactions and process occurring in the after-treatment system during the NOx conversion process that can lead to the formation and emissions of this pollutant (for more details please see Suarez-Bertoa et al., 2020). In order to maintain NH₃ emissions low, ammonia oxidation catalysts are applied in SCR-equipped Euro VI HD Diesel vehicles.

Both vehicles presented higher NOx emissions than those proposed in Euro 7. The Diesel bus exceeded the Cold and Hot NOx limits by 3 and 2.5 times, respectively, while the CNG exceeded them by 1.4 and 3.4 times, respectively. The results obtained during the hot tests highlight the importance of the contribution of NOx during cold start, before the catalyst reaches their operation temperature, but also after their light-off, when emissions should be very low. The CNG bus also exceeded the CO Hot Limit (by 2.4 times) and the Diesel exceeded the Hot SPN10 limit (by 1.7 times) and the two N₂O limits (by 2.3 and 6.4 times).

N₂O has been indicated as the most important anthropogenic Ozone Depleting Substance (ODS) (Ravishankara et al., 2009), and it is also a strong greenhouse gas with a GWP over 100 years 298 times higher than CO₂. These high N₂O emissions meant that the median values from the Diesel bus represented 108 gCO₂eq/kWh and 191 gCO₂eq/kWh for the WHVC_Cold and WHVC_Hot, respectively. For the CNG, median N₂O emissions during the WHVC_Hot were negligible and over the WHVC_Cold reached 0.044 gN₂O/kWh. Higher values over cold start cycles compared to hot start cycles is in line with what have been reported for vehicles equipped with TWC, including CNG (Clairotte et al., 2020, Lähde and Giechaskiel, 2021). Nonetheless, in previous studies N₂O accounted for less than 1% of the CO₂eq (Vojtíšek-Lom et al., 2018), while the CNG bus during the WHVC_Cold at 35 °C reached up to 0.101 gN₂O/kWh, or 3.5% of the CNG’s CO₂ emissions.

Like NH₃, N₂O also forms following different processes in the studied systems. While in the CNG it is formed on the TWC with a narrow temperature window during the catalyst light-off (Nevalainen et al., 2018), in the Diesel vehicle it can be formed through different pathways involving NOx and/or NH₃ on the DOC, the SCR and/or the ammonia oxidation catalyst (see Preble et al., 2019, Selleri et al., 2021 for further information). As illustrated in Fig. 5 the N₂O emissions from the Diesel bus studied were directly related to the DEF injection and SCR operation. Thus, judging the NOx emissions profiles, the DEF injection appeared to start later during the WHVC_Cold compared to the WHVC_Hot, resulting in lower N₂O emissions. This is in line with what suggested by the specialized literature (Colombo et al., 2012).

As expected, the scatter across the WHTC_Hot tests was very low. The trend shows that once the vehicles (engine + after-treatment/s) were at working operating temperature, the ambient temperature plays a limited role, resulting in similar emissions. During the WHVC_Cold, the emissions of NH₃, N₂O and SPN10 showed high dispersion across the temperatures. The maximum values obtained during the tests showed a slightly different picture when compared to the median values. As predictable, differences with the NOx limits increased (4.3x and 2.7x for the Diesel, and 1.8x and 4.1x for the CNG, compared to the Cold and Hot limits, respectively); the CNG failed the NH₃ Cold limit (4.9 times higher) and the SPN10 Cold limit (1.2 times higher). The Diesel bus further exceeded the N₂O limits (2.7 and 7.2 times higher) and alongside the Hot SPN10 limit (2.0 times higher than the limit) it also failed the Cold SPN10 limit (1.2 times higher).

The highest emissions of CO, CH₄ and NH₃ from the CNG were observed during the tests at sub-zero temperatures. Increase of emissions of CO, CH₄ and/or NH₃ from TWC-equipped vehicles at sub-zero temperatures during cold start tests have been widely discussed in the literature for light-duty vehicles (Giechaskiel et al., 2022b, Suarez-Bertoa and Astorga, 2018). The increase of emissions has been explained by longer cold start periods due a longer time to reach light-off of the catalyst combined with rich operating conditions. Hence, very much in line with the results obtained for this CNG bus.

SPN23 emissions were very low, especially for the CNG vehicle. Diesel SPN23 emissions were at the expected levels, with cold start cycles having higher emissions (Napolitano et al., 2020). When looking at the SPN10 emissions, it is interesting to notice that while during WHVC_Hot tests the Diesel presented substantially higher emissions (up to 2 times higher than the Hot limit) than the CNG bus (maximum 15% of the Hot limit), the emissions during the WHVC_Cold were comparable (up to 1.2 times the SNP10 Cold limit for both Diesel and CNG, and also similar medians, 66% and 72% the limit for Diesel and CNG, respectively). Sub-23 nm particle emissions from heavy-duty Diesel vehicles are linked to urea particles emissions linked to the SCR operation (Mamakos et al., 2022, Mamakos et al., 2023). CNG heavy-duty vehicles are well known to generate particles below 23 nm possibly linked to lubrication oil (Napolitano et al., 2020). The Diesel bus, equipped with a DPF, presented consistent emissions whether the vehicle was tested hot or cold and the SPN10 emissions from CNG bus, that was not equipped with particulate filter because this was unnecessary to meet the Euro VI standard, were more affected by the cold start cold ambient tests, with the highest emissions occurring during the −7 °C tests (∼6x 10¹¹ #/kWh – See Table S3 of the SM). In the case of light-duty vehicles, the impact of cold start and cold temperatures on SPN10 emissions from CNG vehicles has been linked to higher emissions (Giechaskiel et al., 2022). In case Euro 7 would have maintained the cut-off size of the measurement at 23 nm instead of reducing it to 10 nm and using the same values proposed for Euro 7 as limits, both vehicles would meet the SPN requirements.

Fig. 6 illustrates the results got from the on-road and ISC-like tests analyzed according to what described in the Section 2.4.2 for tests that accumulated over 3 times the reference work (i.e., 76.4 and 75.9 kWh, respectively for the Diesel and CNG). Similar to what obtained for the WHVC_Cold tests, the CNG’s worst window (i.e., the 100^th percentile window) presented higher emissions than Diesel for THC, CH₄ and CO. The emissions of the worst windows from Diesel were higher than the CNG for N₂O and SPN23, as they were for the WHVC tests (SPN10 measurement was not available for the on-road and ISC-like test).

The pollutants surpassing the limits on-road were the same as those exceeding the WHVC. The level of these exceedances, however, changed, increasing substantially – See Table S3 in the Supplementary Material. Hence, the CNG exceeded the 90^th percentile limit of CO by 4 times. The Diesel exceeded the 100^th and 90^th percentile limit of N₂O by 5.2 and 8 times, respectively. This means that the median of 100^th and 90^th percentile windows were 0.837 and 0.799 gN₂O/kWh, with a maximum at 1.061 gN₂O/kWh (i.e., up to 37% of the CO₂ emitted by this vehicle during the WHVC). High N₂O emissions, reaching up to 0.196 g/kWh, have been previously reported for Euro VI HDV (Giechaskiel et al., 2019, Mendoza-Villafuerte et al., 2017, Selleri et al., 2022). Note that US EPA requires that HD do not emit over 0.133 g/kWh of N₂O. As stated above, N₂O emissions from vehicles equipped with TWC have been considered negligible (Clairotte et al., 2020) and the emissions of this CNG bus during the WHVC were indeed close to zero during the hot tests and minor during the cold ones. However, unlike the WHVC_Cold tests, there were emissions of N₂O in the 100^th percentile windows during the on-road and ISC-like tests regardless the ambient temperature. The emissions ranged from 0.068 gN₂O/kWh to 0.080 gN₂O/kWh and, as for the WHVC_Cold tests, they were concentrated on the cold start, with low emissions in 90^th percentile windows (up to 0.003 gN₂O/kWh).

Median NOx emissions on-road for the worst window of the Diesel (1.155 g/kWh) bus were comparable to those obtained during the WHVC_Cold, showing a common effect of the emissions during cold start. However, in the CNG case the median emissions of the worst on-road windows (1.062 g/kWh) were 2.2 times higher than those got over the WHVC_Cold. Moreover, in the case of the 90^th percentile window the CNG vehicle presented similar emissions (0.932 g/kWh) to those of the worst window consistently with the high NOx emissions measured all along the test. Unlike during the WHVC tests, where NOx was 36% lower during WHVC_Hot as compared to WHVC_Cold, during the on-road and ISC-like tests the 100^th percentile window, which included the cold start, were 14% higher than the 90^th percentile window, which started well after the light-off of the TWC. Fig. 7 shows the presence of further NOx emissions peaks from the CNG vehicle even after the cold start phase. As consequence, the CNG exceeded the 90^th percentile limit by 10 times. After the cold start, the Diesel’s emissions improved (0.639 g/kWh), remaining however still high, i.e., 7.1 times the limit.

Comparing the emission factors obtained for the 90^th percentile window of the ISC_I and ISC_III tests using the Euro 7 methodology (e.g., NOx Diesel 0.763 g/kWh and CNG 0.605 g/kWh) and using the Euro VI-D, highlights that the emissions excluded by previous procedures led to 48% and 57% lower emission factors. Moreover, the NOx emission factors obtained using all the emissions, including cold start, are about 2 times higher than those reported by Zhu et al. (2024), using EPA’s and CARB’s 2/3 BIN MAW methodologies.

On-road and ISC-like NH₃ emissions of the CNG bus were about on order of magnitude lower compare to those reported by Zhu et al. (2024) and to Thiruvengadam et al. (2016), which reported 0.10–0.19 gNH₃/bhp-hr (0.134 – 0.255 gNH₃/kWh). They were similar in the case of Diesel. These emissions were below the Euro 7 limit for both vehicles and similar to those achieved in the laboratory during tests with comparable ambient conditions. All this suggests that, although the NH₃ emissions remain high, they are lower than in USA where it is not regulated.