Non-carbon greenhouse gas emissions for hybrid electric vehicles: three-way catalyst nitrous oxide and ammonia trade-off

Transport sector decarbonisation is leading to increased demand for electrified powertrains including hybrid vehicles. The presence of an internal combustion engine and electric motor offer multiple performance and efficiency advantages. However, changes in the conditions that catalytic aftertreatment systems are subjected to can present challenges in meeting forthcoming emissions standards. This work investigated the three-way catalyst performance to abate regulated and unregulated emissions from a gasoline direct injection engine working under conditions related to hybrid vehicle operation. The focus on unregulated emissions of NH3 and N2O is of interest due to limited literature on their formation in conventional, and particularly hybrid, vehicle aftertreatment systems. Furthermore, the likelihood of their regulation when the EURO 7 emissions standards are introduced increases the pertinence of this work. For this particular engine and aftertreatment setup, it was found that starting the engine whilst the three-way catalyst temperature was below 150 ℃ led to an increase in tailpipe regulated emissions and N2O. Whilst, starting the engine when three-way catalyst temperatures were above 350 ℃ lead to tailpipe NH3 emissions. This was due to the selectivity of NO to form N2O at lower temperatures and NH3 at higher temperatures. For the case of the studied catalyst, a vehicle energy management strategy opting to start the engine with the three-way catalyst within a targeted temperature range allowed for a trade-off between regulated emissions, N2O and NH3. These findings are significant since it can be used to optimise hybrid vehicle control strategies minimising both regulated and unregulated emissions.


Introduction
The transportation sector contributes up to a third of all anthropogenic CO 2 emissions (The Society of Motor Manufacturers and Traders 2019).In order to tackle this, CO 2 limits in road transport are becoming increasingly stringent worldwide and the EU aims to reduce emissions from 2021 levels by 15% in 2025 to 37.5% in 2030 (International Council on Clean Transportation 2019).To achieve such aggressive targets, vehicle fleets will need to incorporate increased numbers of alternatively fuelled and propelled vehicles.The plug-in hybrid electric vehicle (PHEV), as a medium-term road transport option, has the potential to provide the CO 2 reductions required whilst not compromising on practicality (Orecchini et al. 2018).A PHEV utilises a battery, electric motors and an internal combustion engine (ICE).This means that the ICE will operate differently when compared to ICEs in conventional vehicles.This presents new challenges and opportunities for the operation of the aftertreatment system.
The aftertreatment system of a gasoline direct injection (GDI) engine makes use of a three-way catalyst (TWC)-socalled for its ability to simultaneously convert emissions of CO, NO X and HC into CO 2 , N 2 and H 2 O.The performance of a TWC depends significantly upon its temperature.When the TWC is below the light-off temperature (250 ℃-300 ℃) there is reduced catalyst activity (Fernández-García et al. 2004).When there are multiple engine starts over a single journey-typical of PHEV operation-this problem is exacerbated due to the cooling of the aftertreatment system whilst the engine is off.Therefore, understanding how the TWC operates within the modern set up of a PHEV and improving TWC performance under start-up engine conditions will be vital in order to reduce future harmful emissions and meet future emissions regulations.Furthermore, it is looking increasingly likely that currently unregulated noncarbon species such as ammonia (NH 3 ) and nitrous oxide (N 2 O) will become regulated with the introduction of the EURO 7 emissions regulations.Farren et al. reported that in the UK, the total amount of NH 3 emitted from passenger cars within urban areas was 17 times higher than what had been previously reported by the UK National Atmospheric Emissions Inventory suggesting that NH 3 emissions could be of greater concern then first realised (Farren et al. 2020).However, there has been very little published research into the emissions of NH 3 and N 2 O from GDI engines with TWCs during conventional operation and even less so in hybrid vehicles.
NH 3 is a secondary pollutant which is typically formed within the TWC through reactions including NO, CO, H 2 O and H 2 as precursor molecules (Heeb et al. 2006;Wang et al. 2019;Liu et al. 2021a, b).The scarce literature agrees on the presence of tailpipe (post-TWC) NH 3 spikes during vehicle accelerations (Heeb et al. 2006;Suarez-Bertoa et al. 2014;Wang et al. 2019;Liu et al. 2021a, b) due to a reduction in engine lambda promoting TWC selectivity towards NH 3 formation (Heeb et al. 2006).Wang et al. (Wang et al. 2019) stated that with catalyst ageing there is a further increase in the selectivity towards NH 3 production.Catalyst temperature has been identified as a key parameter influencing NH 3 formation.A wide temperature range between 250 ℃ and 550 ℃ has been reported where NH 3 selectivity is the highest (Wang et al. 2019).N 2 O is also a secondary pollutant formed within the TWC requiring NO as a precursor (Mejía-Centeno and Fuentes 2009).N 2 O formation also varies depending upon temperature with different species contributing at different temperatures (Mejía-Centeno and Fuentes 2009).The maximum formation of N 2 O was reported to occur between 200 ℃ and 350 ℃ with the principal precursors being CO and NO.Nevalainen et al. acknowledged the reaction between CO and NO, but also stated the role of H 2 and HCs as important N 2 O precursor species (Nevalainen et al. 2018).
The start/stop nature of engine operation within PHEVs is likely to increase the frequency and duration of the TWC's operation through the temperature regions where NH 3 and N 2 O selectivity is high in comparison with a conventional ICE vehicle.The scarce number of studies that do report unregulated emissions of hybrid electric vehicles (HEVs) commonly take a holistic approach without investigating the performance of the TWC under such conditions.Suarez-Bertoa et al. reported regulated and NH 3 emissions that were measured from a range extender hybrid vehicle over the WLTC, finding the NH 3 emissions to be comparable to a conventional gasoline vehicle (Suarez-Bertoa 2016).Clarotte et al. also found tailpipe N 2 O emissions for light duty hybrid vehicles to be comparable and in some cases even higher than those of conventional gasoline vehicles (Clarotte et al. 2020).The catalyst light-off period has been identified as the source of NH 3 production from HEVs (Suarez-Bertoa et al. 2014;Suarez-Bertoa 2016).
Therefore, a research gap has been identified in the reviewed literature concluding that there is very little available information on the operation of a TWC in an HEV powertrain with respect to currently unregulated non-carbon emissions.This research aims to tackle this gap by investigating the emissions of NH 3 and N 2 O over an engine start procedure at multiple TWC temperatures.This is done to replicate the TWC conditions witnessed within a modern hybrid powertrain.NH 3 and N 2 O are of interest since it is expected that these will both be regulated with the introduction of the EURO 7 emissions standard (ACEA 2020).The research was undertaken at The University of Birmingham Future Fuels and Powertrains Laboratory.

Experimental setup
This work was carried out experimentally using a modern turbocharged GDI engine that also featured port fuel injection (details in Table 1).An electric dynamometer was used to control engine speed and load.The testing setup was instrumented to measure appropriate temperatures, pressures and flow rates.The model TWC was provided by Johnson Matthey PLC.The total PGM loading was 30 g/ ft 3 with 23 g/ft 3 of palladium and 7 g/ft 3 of rhodium.The monolith dimensions were a length of 109 mm and a diameter of 118.4 mm with a cell density of 600/2 and a single layer washcoat loading of 3.0 g/in 3 .The monolith was positioned after the turbocharger as shown in Fig. 1.The zoomed in section of the figure labelled 'TWC set up' shows how the TWC was placed with the locations to measure emissions and temperatures before and after the TWC monolith.Temperature measurements were taken before and after the TWC monolith, using k-type thermocouples and a Pico Technology TC-08 data logger.Two thermocouples were placed before the monolith-one measuring the gas temperature and one measuring the monolith wall temperature on the inlet and outlet faces of the catalyst monolith.A further two thermocouples were placed after the monolith measuring the gas and monolith wall temperatures.Emission probes were located before and after the TWC monolith.Real-time hydrogen emissions were measured using a HSense V&F electron impact ionisation mass spectrometer.All other Fig. 1 Engine and TWC setup including locations for emission probes and thermocouples gaseous emission species, including CO, CO 2 , NO X , NH 3 , N 2 O and HCs were measured using an MKS multi-gas Fourier transform infrared (FTIR) spectrometer (MKS Mul-tiGAS 2030 Series).Samples were collected as shown in Fig. 1 with heated lines to the analysers.The heated lines were maintained at 150 ℃ to prevent condensation of the water and hydrocarbon species.

Methodology
The engine start procedure was selected in terms of torque and speed.This featured a warmup period of the engine followed by changing to an engine condition with a sufficient power output to charge the battery whilst also operating at a high efficiency point of 2500 rpm and 90 Nm.The engine start cycle completed for each experiment is shown in Fig. 2a.In order to understand how the TWC performs at a variety of conditions the engine was started at various TWC monolith temperatures with three repeats for each temperature.This was achieved by running the engine to heat the TWC before switching the engine off and allowing the TWC to cool to the desired temperature before starting the engine.The monolith temperature at the inlet face was the temperature that was targeted.To heat the TWC, the engine was operated at 2500 rpm and 90 Nm until the oil, coolant and monolith temperatures had stabilised.Once the temperatures had stabilised the engine was brought to an idle condition of 900 rpm and 3 Nm for 60 s after which the engine was turned off to allow the TWC monolith to cool to the desired temperature.During this period, the engine would also cool with the aim of this to study also how the engine out emissions performance was affected.The cooling of both the catalyst and the engine was desirable as this provided a more realistic comparison to a complete PHEV.
Multiple engine and monolith starting temperatures were selected to cover a wide range of conditions the TWC would be subjected to depending upon the engine off time in a PHEV.The engine starts were chosen to provide TWC monolith start temperatures of ambient, 100 ℃, 150 ℃, 200 ℃, 250 ℃ and 350 ℃.These TWC temperatures were selected based upon previous studies (Zeng et al. 2018;Guille des Buttes et al. 2020) that evaluated a HEV throughout a WLTC and reported TWC temperature.There were multiple engine starts reported by the studies within the range of TWC temperatures selected for our study, particularly during the first part of the WLTC.Temperatures above 350 ℃ were not selected, as the TWC will be operated at maximum efficiency.Overall, these testing conditions allowed for study of TWC performance under different pre-TWC emissions and catalyst temperatures both induced from starting the engine at various oil and coolant temperatures.It is considered that the conditions selected are likely to occur within a variety of HEV powertrain architectures (e.g.series, parallel and power-split hybrid), vehicle energy management strategies, and vehicle mode of operation (e.g.charge sustaining and charge depleting).The exhaust gas flow rate was also monitored due to the transient warm up nature of the engine start cycle.This is shown in Fig. 2b as this will also affect TWC performance and total mass of emissions.
The starting TWC, engine oil and coolant temperatures for each of the three repeats are shown in Table 2.This shows how the TWC temperature was linked to the engine start temperature as the engine off period allowed for cooling of not only the TWC but also the engine too.Once the target catalyst temperature was reached the engine started and followed the engine start procedure in Fig. 2 starting with the temperatures shown in Table 2.

Engine output regulated emissions and TWC conversion performance
The TWC was exposed to different inlet concentrations of pollutants and temperatures depending upon the starting temperature of the engine.Regulated engine output emissions (pre-TWC) for each temperature (average of the three repetitions) are shown in Fig. 3 Since these are the emissions before the TWC these are affected only by the engine temperature.These have been reported as the conditions the TWC is exposed to under HEV operation is of interest in understanding how TWC performance is affected.
The most notable influence that the engine start temperature has on the pre-TWC concentrations is the peak in the CO concentration during the first 10 s of the engine start procedure.This is shown in the zoomed in section of Fig. 3a.There is not an appreciable reduction in the magnitude of the CO peak until the engine is started with an oil and coolant temperature of 67.2 ℃ and 67.1 ℃, respectively (the 250 ℃ test-data in Table 2).After this initial peak in CO during the first 10-15 s, the levels of engine out (pre-TWC) CO recorded were similar for each engine start at each temperature.There was increased CO over the first 50 s for the ambient start test, but other than this the levels of CO are comparable.The reason for the larger peaks over the first 50 s in pre-TWC CO emissions for the ambient start was the colder temperatures.Starting the engine at higher temperatures improves the mixture formation in the cylinder minimising CO formation (Bielaczyc et al. 2011).In addition to this, at lower engine start temperatures, the engine runs with a richer fuel mixture in order to improve combustion stability.This will also lead to increased CO emissions (Hu et al. 2007).To minimise the demand on the oxidative capacity of the TWC during the engine start, targeting oil and coolant temperatures of 67.2 ℃ and 67.1 ℃, respectively, should be done.This is because engine temperatures lower than this generates considerably higher concentrations in CO for the TWC to deal with.
The pre-TWC NO X emissions are plotted in Fig. 3b for each test temperature.By comparing Figs.2a and 3b, it is clear that the NO x emissions are affected more so by the engine condition than the other pollutants are.This explains the low concentrations of NO X during the first 60 s when the engine is idling.Each increase in NO X concentration matches with an increase in engine load.This is because as the load increases, so too does the in-cylinder temperature resulting in increased NO X formation via the Zeldovich mechanism (Anetor et al. 2014).The starting temperature of the engine also affects the NO X concentration that the TWC must deal with.Over the first 150 s, the NO X emissions are lower when starting the engine with an oil and coolant temperature less than 49.2 ℃ and 46.4 ℃, respectively (results for the ambient and 100 ℃ tests).This is due to the richer combustion at colder temperatures discussed above and the lower in-cylinder temperatures over the engine start.By keeping the engine load low, it is possible to keep the engine NO X emissions at or close to zero.Operating at low loads during the engine start could prevent NO X concentrations being present pre-TWC giving the catalyst time to warm up before load is increased.The pre-TWC concentrations of THCs are shown in Fig. 3c.The THCs are the most sensitive to the engine starting temperature with the concentration of THC still higher for the ambient start test even after 300 s.During the 100 s, the concentration of THCs is approximately double for the ambient test (starting oil and coolant temperatures of 18.4 °C) than for the other starting temperatures.This is a result of poorer mixture formation at these low temperatures as well as richer combustion leading to unburnt or partially burnt hydrocarbons (Li and Liu 2011).
Even by starting the engine at oil and coolant temperatures of 49.2 ℃ and 46.4 ℃, respectively, provides a considerable reduction in the pre-TWC concentration of THCs and will reduce the demand on the catalyst.There are smaller reductions in the pre-TWC THC concentration at higher start temperatures.It is unlikely that the engine in a PHEV will have to do an ambient start more than once per journey.This is because engine off periods in a PHEV rarely exceed 5 min.When reviewing the literature considering engine performance within a PHEV over different drive cycles, engine off periods varied between 1-5 min (Overington and Rajakaruna 2015;García et al. 2019;Zhao et al. 2020;Gezer et al. 2021).The time recorded for the engine to cool sufficiently to start the 100 ℃ experiment was 63 min.Whilst the cooling rate of oil and coolant may be quicker for an engine within a vehicle than those observed in these engine dyno tests due to outside colder temperatures and forced convective cooling as the vehicle continues to travel it is still unlikely that the oil and coolant temperatures will reach such a low level.These conditions though could occur after the vehicle has been left between journeys.The reason why this work opted to use TWC temperature instead of engine off time as the independent variable was to produce results more applicable to different set ups and conditions.The CO concentration post-TWC (tailpipe) for each of the different start temperature tests is shown in Fig. 4a.The large peaks in CO over the first 10-15 s are still present but now only for the lower catalyst start temperatures (< 150 ℃).This is because the catalyst activity when starting at 200 ℃ or higher is sufficient to oxidise all the CO that is shown in Fig. 3a.Starting the engine with a TWC temperature of 200 ℃ or greater removes all CO from the exhaust gas.With the Fig. 3 Pre-TWC concentrations for a CO, b NO X and c THC exception of the first peak in pre-TWC CO concentration, the inlet concentrations to the TWC were all similar for each test.This highlights that the CO concentrations measured post-TWC are resulting from the reduced TWC performance.This is also shown in Fig. 4b where the conversion efficiency of the TWC is plotted for each test temperature.Starting from a lower temperature not only increases tailpipe CO concentration through increased CO formation in the engine but also through reduced TWC activity.The light-off temperature is defined as the temperature at which 50% conversion is achieved.Using this definition, light-off for CO is achieved immediately for TWC temperatures starting at 200 ℃ or higher.Light-off is reached within the first 10 s for the 150 ℃ experiment, whilst light-off is not reached until 145 s and 165 s for the 100 ℃ and ambient tests, respectively.Whilst there is little change in the light-off time from ambient to 100 ℃, the reduction in light-off time from 100 ℃ to 150 ℃ is substantial.Keeping the TWC temperature at 200 ℃ or above would provide complete CO conversion.However, if this is not possible, a target of 150 ℃ should be set since after this CO conversion deteriorates considerably.
The NO X concentration post-TWC (engine out) for each of the different catalyst starting temperatures is shown in Fig. 5a.There is little to no NO X reported post-TWC during the first 60 s of engine operation.This is due to the lack of NO X pre-TWC (Fig. 3b).After 60 s, when the engine load is increased as per the start-up strategy shown in Fig. 2a, there is a breakthrough in NO X post-TWC when starting the engine with a TWC temperature at 200 ℃ or lower.On the other hand, starting the engine with a TWC temperature of 250 ℃ or higher completely removes NO X concentration post-TWC.The NO X concentration post-TWC follows the same trend as the concentration pre-TWC as each load increase yields an increase in NO X .The conversion efficiency of NO X is plotted for each of the different starting temperature tests in Fig. 5b.The performance of the TWC can be analysed here by comparing the time taken to reach light-off (50% conversion) once NO X emissions are present pre-TWC (after 60 s).Light-off is achieved for the ambient test after 174 s and after 154 s for the 100 ℃ test.However, the test starting with a catalyst temperature of 150 ℃ does not drop below 50% conversion.As per the results for CO conversion, there is little improvement when comparing an ambient start to one with a TWC temperature of 100 ℃, but when comparing 100 ℃ to 150 ℃ there is a considerable difference.This could once again be a secondary target if keeping the starting catalyst temperature above 250 ℃ (no NO X ) cannot be achieved.A further consideration from these results is the ability to minimise NO X by operating at a low load condition until the TWC has reached sufficient temperature.Further catalyst heating strategies could be employed such as retarded ignition to help increase catalyst temperature whilst NO X emissions are kept low.Once the catalyst has reached a sufficient temperature, the engine can then proceed to higher loads and any subsequent NO X emissions can be converted by the now active TWC.However, this will need to be balanced with the effect on other emissions, fuel economy and driveability.The THC concentration post-TWC for each starting temperature test is shown in Fig. 6a.When starting the engine with a TWC temperature of 250 ℃ or higher there is no THC concentration measured post-TWC.Therefore, total conversion of the THCs pre-TWC (Fig. 3c) is occurring within the TWC.There are low levels of THCs recorded post-TWC for the 200 ℃ catalyst start temperature test.However, for each temperature lower than this the concentrations increase as the temperature is reduced.There is a large difference between the 150 ℃ to the 100 ℃ tests and another large difference between the 100 ℃ and ambient tests.When comparing the different regulated emissions post-catalyst in Figs. 4, 5 and 6, HC species are the most demanding to remove within the TWC.This matched other results in the literature where the T 50% for THC is reported to be higher than that for CO regardless of TWC composition and higher than NO X for the majority of TWC compositions (Rood, et al. 2020).In the results presented here, HCs require the highest starting catalyst temperature to ensure there is no slip during the engine start.The HC conversion efficiency for each starting catalyst temperature is presented in Fig. 6b.HC conversion takes the longest time out of the regulated pollutants to consistently reach > 90% for the same starting TWC temperature.Light-off was reached after 179 s for the ambient start, after 159 s for the 100 ℃ start and 138 s after the 150 ℃ start.For 200 ℃ and above the conversion efficiency does not fall below 50%.When comparing the performance of the TWC between the 100 ℃ and 150 ℃ starting temperatures there is a large difference before the light-off.
The 100 ℃ starting temperature is at roughly 20% conversion whilst the 150 ℃ starting temperature is at roughly 80% conversion.This same magnitude in conversion efficiency increase was not witnessed for any of the other catalyst start temperatures.Therefore, aiming to start the engine with a monolith temperature greater than 150 ℃ will avoid the substantial drop off in TWC performance witnessed at 100 ℃ when compared to 150 ℃.When starting the engine with a monolith temperature of 250 ℃ or above, there is complete THC conversion.To maintain TWC performance within a HEV experiencing multiple engine starts it would be desirable to keep the monolith temperature at 250 ℃ or above with a second aim of not letting it drop below 150 ℃ if maintaining 250 ℃ is not possible.

Impact of exhaust gas composition and temperature on secondary pollutant formation in TWCs
The formation of NH 3 and N 2 O over the TWC has been reported by several studies (Behrentz et al. 2004;Adhikari and Fernando 2006;Anderson et al. 2012;Balas et al. 2014;Badshah and Khalek 2015;Bai et al. 2018;Wang et al. 2019;Awad et al. 2020;Bagheri et al. 2021).The results presented here find agreement with the literature.Whilst there were very little N 2 O or NH 3 emissions recorded pre-TWC (< 20 ppm), there were considerable levels of NH 3 and N 2 O recorded post-TWC over the first 300 s after the engine had started.These are shown in Fig. 7a, respectively.Despite each test following the same engine start procedure, these figures show there to be varying peaks in concentration at various times.As the flow rate was not constant over the first 300 s of engine operation (shown in Fig. 2b), the cumulative masses for both NH 3 and N 2 O are both shown in Table 3 to give a fair comparison.
The cumulative masses for NH 3 show there to be an optimum TWC temperature window to start the engine where NH 3 formation is minimised.This was between 150 ℃ and 250 ℃.The cumulative N 2 O mass was found to be highest when starting the engine with the TWC between 100 ℃ and 150 ℃.Avoiding the temperature regions where NH 3 and N 2 O are formed within the TWC is important due to their reported environmental effects.NH 3 emissions are undesirable due to the role of NH 3 as a secondary inorganic aerosol contributor through the formation of ammonium nitrate and ammonium sulphate which contributes to PM emissions (Huang et al. 2021).However, some studies have attempted to generate NH 3 within the TWC purposefully in order for it to later be used downstream as a reducing agent in selective catalytic reduction (SCR) systems to reduce NO x emissions (Oh and Triplett 2014).Emissions of N 2 O are also undesirable as N 2 O is rated as a powerful greenhouse gas with a global warming potential 300 times that of carbon dioxide whilst also being a threat to atmospheric ozone (Hensen et al. 2013).Whilst either NH3 or N 2 O are regulated under the current EURO VI emissions standard there are reports that when the EURO 7 regulations are launched these will become regulated (Mendoza-Villafuerte et al. 2017; ACEA 2020).
There are several different reaction pathways that have been proposed for the formation of NH 3 and N 2 O over a TWC with typical exhaust gas inlet conditions.However, there is good agreement over the temperature ranges over which these secondary pollutants are formed.NH 3 has been reported to form over palladium-based catalysts within a temperature window between 300 ℃ and 500 ℃ (Mejía- Centeno et al. 2012;Oh and Triplett 2014;Wang et al. 2019;Liu et al. 2021a, b;Bae et al. 2022).This helps to explain the different peaks in NH 3 concentration found in Fig. 7a.Whilst the TWC temperature is increasing, the tests with a higher TWC start temperature will transition into the NH 3 formation region sooner, hence the NH 3 peaks appear sooner after the engine start.This contrasts to the colder TWC start temperature tests where the TWC takes longer to reach the NH 3 formation region.The literature reports that N 2 O formation occurs over a lower temperature region of between 150 ℃ and 300 ℃ (Jobson et al. 1994;Mejía-Centeno and Fuentes 2009;Mejía-Centeno et al. 2012;Gong and Rutland 2013).When looking at Fig. 7b, the temperatures starting at 250 ℃ and higher do not produce any N 2 O.This is because these temperatures are higher than the formation region for N 2 O. Therefore, the TWC does not pass through this region and no N 2 O is formed.When looking at how the peak in N 2 O is shifted for each different starting temperature it can be seen that the lower starting temperatures reach the formation zone later after the engine start.The inlet catalyst temperature for each start is shown in Fig. 8 with the formation regions for NH 3 and N 2 O highlighted.
The formation of NH 3 and N 2 O require concentrations of NO in order to form (Jobson et al. 1994;Mejía-Centeno et al. 2012;Gong and Rutland 2013;Wang et al. 2019;Bae et al. 2022).By reviewing Figs.7 and 3b, it can be seen that the peaks in NH 3 and N 2 O coincide with the NO peaks.A further strategy to reduce the chances of NH 3 and N 2 O would be to manage the formation of NO within the engine.This could be achieved by maintaining the engine load at a low condition but this must be balanced against other emission, fuel economy and driveability.

Balancing unregulated emissions of ammonia and nitrous oxide
It is desirable to reduce the emissions of both NH 3 and N 2 O.However, it is challenging to tackle both simultaneously since selectivity for these vary with temperature.At  As discussed, CO, NO X and HC emissions reduce as the starting temperature of the TWC is increased.It is to be highlighted in Fig. 9 that there is a large reduction in the CO and NO X emissions from 100 to 150 ℃ which confirms the comments made in the regulated emissions section.NH 3 formation within the TWC has an optimum temperature window that appears between 150 and 250 ℃ catalyst temperature where the production is low.However, N 2 O peaks for the 150 ℃ catalyst start temperature experiment.These relationships between unregulated emissions (NH 3 and N 2 O) and TWC temperature change the convention that a higher TWC start temperature is always most beneficial in order to minimise emissions.In this case, the optimum starting catalyst temperature would be 250 ℃ if balancing of all emissions both regulated and unregulated is desired.
The initial temperature of the TWC varies depending upon the time between engine starts, ambient temperature, vehicle speed and catalyst insulation or heating.This will change depending upon the hybrid architecture, operating strategy and the driving style.Thus, it has to be considered that with the reduction in temperature for a given amount of time being different for each aftertreatment set up and environment it is not possible to link accurately the starting temperatures tested here to a time between engine starts.However, despite the different EMS employed by different engineers and manufacturers it is possible to predict engine off periods.Often in the case of blended or charge-sustaining EMS, times between engine starts can be between 1 and 5 min (Overington and Rajakaruna 2015;García et al. 2019;Wang et al. 2022).This would push the engine to starting with higher TWC temperatures where NH 3 production within the TWC would be a concern.However, chargedepleting strategies feature longer engine off periods where lower TWC temperatures will be a concern (Graver et al. 2011;Wang et al. 2022).It was decided that by targeting certain catalyst temperatures rather than engine off times that this work would be applicable to PHEVs regardless of their EMS.Since it is more likely based upon the available literature that HEVs would deal with increased N 2 O and NH 3 emission during each engine start, strategies should be defined to minimise these.This could involve a longer idling period to effectively limit the pre-catalyst concentration of NO X and therefore eliminate the precursor species for N 2 O and NH 3 formation.An alternative option could be to increase the engine load quickly to pass through the temperature zones with high N 2 O and NH 3 production to limit the total emission per engine start.However, this may incur a fuel consumption or other regulated emission penalty.

Conclusion
This investigation has shown how regulated (CO, NO X and HC) and unregulated emissions (NH 3 and N 2 O) are affected by engine restarts over a range of catalyst temperatures that are typical of HEV operation for this particular engine and aftertreatment system setup.Regulated emissions all reduced as the catalyst temperature was increased with no emissions slip post-TWC after starting at 250 ℃ TWC temperature.This is a significant finding to provide guidelines for the selection of energy and emissions optimisation management strategies for HEVs since it suggests that by keeping the TWC above a threshold temperature at each engine start, the regulated emissions can be completely abated.
This work has also demonstrated the emissions performance for unregulated emissions.NH 3 and N 2 O are expected to become regulated when the next emission regulations are released which makes these results significant for the future.NH 3 and N 2 O both require NO emissions pre-TWC to be catalytically formed in the TWC.However, the formation of N 2 O and NH 3 is dominant over different temperature ranges.NH 3 formation peaks at 350 ℃ whilst N 2 O emission peaks at 150 ℃.To minimise both ammonia and N 2 O emissions, a compromise must be struck between them if TWC temperature is to be optimised for each engine start.
The aim of this study was to identify how the TWC operates within a HEV with respect to the catalyst starting temperature for both regulated and unregulated emissions.This was selected as an aim due to the increased number of hybrid powertrains available on the market and the midterm direction of the automotive industry.This work has found consensus with current literature over the formation of N 2 O and NH 3 over the TWC but taken this further by linking this issue to modern HEV powertrains.This work also has provided a new direction for the optimisation of HEV energy management strategies to consider emissions of NH 3 and N 2 O.
The principal limitation of this study is that it was completed using an engine dynamometer and not actual hybrid electric vehicles.This provided additional control, repeatability and instrumentation but at the expense of not using the engine as part of an actual HEV powertrain.The assumptions made on the operation were supported by the available literature, and a range of catalyst starting temperatures were selected to cover a variety of different scenarios that could be witnessed by the TWC over the journey of a HEV.
The next steps would be confirming these findings with a study operating a full hybrid electric vehicle over a driving cycle or/and in real driving emission conditions in order to evaluate the N 2 O and NH 3 emissions.Further implications for this work include the evaluation of different engines and aftertreatment setups, various engine starting strategies in order to minimise the emissions of N 2 O and NH 3 or targeting these emissions using a control strategy to eliminate or at least minimise the formation of these species.

Fig. 2 a
Fig. 2 a Engine start procedure torque in Nm (solid line) and speed in rpm (dashed line) and b exhaust gas flow rate

Fig. 4 a
Fig. 4 a Post-TWC CO concentration and b TWC CO conversion efficiency

Fig. 6 a
Fig. 6 a Post-TWC THC concentration and b TWC THC conversion efficiency

Fig. 8
Fig. 8 TWC monolith inlet face temperature for each catalyst starting temperature.Note the highlighted temperature regions for NH 3 and N 2 O formation

Fig. 9
Fig. 9 Accumulated mass of emissions for the first 500 s of engine operation with respect to the TWC temperature at the time of engine start for NH 3 , N 2 O, CO, NO X and THC.Confidence intervals show the 95% confidence level calculated from 3 repeats of the start-up test

Table 1
Test engine specification

Table 3
Cumulative mass emissions for NH 3 and N 2 O recorded over the first 500 s of the engine start for each catalyst starting temperature test.The ± figures show the 95% confidence interval calculated from 3 repeats of the start-up test