Impact of a Yttria-Stabilized Zirconia Thermal Barrier Coating on HCCI Engine Combustion, Emissions, and Efficiency

Homogeneous charge compression ignition engines (HCCI) have distinct advantages in both emissions and efficiency over spark-ignited engines. Lean, unthrottled operation enables HCCI engines to simultaneously have higher cycle efficiency and lower NO_x emissions [1,2]. However, low temperature combustion (LTC) is driven by chemical kinetics, making HCCI engines highly sensitive to in-cylinder charge temperature history [3]. As an added difficulty, in-cylinder temperatures are subject to strong spatial variation. Gas temperatures measured via planar laser-induced florescence (PLIF) showed the development of late-compression thermal stratification in a chamber that was homogenous in temperature at bottom dead center [4–7]. The significant cycle-to-cycle variability in thermal stratification can be seen by comparing an ensemble-averaged PLIF image (Fig. 1(a)) with one of a single cycle (Fig. 1(b)) [4]. Volumes of gas influenced by wall temperatures were not limited to near wall regions by the end of compression, and these wall-affected regions exhibited temperatures 50 K below the bulk average temperature. Other parcels of gas mixture were as much as 50 K warmer than the bulk average temperature for a single cycle [4].

Chang et al. [8] showed a strong dependence of HCCI autoignition and combustion on in-cylinder wall temperatures. By increasing coolant temperature from 80 °C to 95 °C, combustion stability improved, autoignition advanced, and peak rates of heat release increased up to 35%. Thus, large changes in combustion duration and ignition timing were possible with small changes in combustion chamber surface temperatures, providing important evidence of the influence that wall-affected-gas temperatures have on HCCI combustion and autoignition.

Additional evidence of wall temperatures influencing HCCI combustion was found with the buildup of combustion chamber deposits (CCDs) [9–11]. Increasing deposit accumulation in an HCCI engine advanced autoignition and reduced combustion duration. It was found that deposits act as thermal insulation, having a low thermal diffusivity and increasing combustion chamber surface temperatures. The increased wall temperature swing decreases the temperature differential between the charge and the chamber walls, reducing heat transfer, increasing the bulk gas temperature during compression, and advancing the autoignition process. The cumulative effect was a stabilization of combustion and increased rates of heat release, ultimately enabling the engine to operate at leaner equivalence ratios and shifting the low load limit [9–11]. The buildup of the deposit layer reached equilibrium after 40 h of testing at a constant HCCI operational point [9–11]. This equilibrium thickness is reached when the CCD surface temperature increases to the point where condensation of new deposits equals their rate of burn-off. Thus, a shortcoming of relying on CCDs as thermal insulation is that their thickness depends on engine operational history.

Thermal barrier coatings (TBCs) can offer thermal insulation similar to combustion deposits with reduced drive cycle dependence. Hoffman et al. [11] utilized a magnesium zirconate coated piston in an HCCI engine. Combustion phasing advanced relative to a standard engine, resulting in increased combustion efficiency. The phasing advance makes it difficult to isolate the fundamental effects coatings have on LTC. Advanced combustion increases gas temperatures, promoting lower emissions and higher efficiency. Overall findings encourage further research efforts related to TBCs for HCCI.

A newly completed single-cylinder HCCI engine setup enables experimental measurements of in-cylinder heat flux and coating surface temperatures when used in conjunction with TBCs. In this study, yttria-stabilized zirconia (YSZ) was applied to the piston top. This coating has been proven both effective and durable in gas turbine applications. However, YSZ was not previously evaluated in an HCCI engine. Two experimental campaigns were conducted to isolate YSZ impacts on HCCI combustion, emissions and efficiency. The first group of experiments compared the metal-engine baseline to the engine operated with the YSZ coated piston over the same conditions. As expected, the use of the YSZ coated piston advanced autoignition and combustion phasing. The second set of experiments intentionally prevented the advancement of combustion, and isolated the purely thermal impacts of the coating on heat loss, as well as thermal and combustion efficiencies. Cooled exhaust gas recirculation (EGR) was added to maintain combustion phasing as the baseline when operating with the YSZ TBC. In both cases, coating surface temperatures were computed using a recently developed inverse heat conduction solver. By combining measurements of thermal boundary conditions, engine performance metrics, and combustion phasing control, a deeper understanding of the impact TBCs have on LTC was developed.

This paper is organized as follows: first, a brief background on the use of TBCs is presented, followed by a discussion of the experimental setup and TBC process. Next, the experimental results with the YSZ coated piston are presented and discussed, followed by concluding statements about the effects of TBCs in HCCI engines.

Thick “monolithic” TBCs were previously pursued as a potential solution for achieving a low heat rejection engine [12]. The coatings were thick to increase the thermal resistance and reduce the heat transferred to the coolant [12–14]. However, the high thermal mass of the coatings stored combustion heat and released it during gas exchange, causing substantial charge heating and losses in volumetric efficiency [12]. The surface temperature of the thick TBCs followed a similar behavior to the “traditional insulation” in Fig. 2.

In contrast, thin TBCs are suggested to avoid problematic charge heating while still reducing combustion heat transfer [16–18]. These coatings differ from previous TBCs in that the thermal resistance is provided by a low thermal conductivity material instead of by the coating's thickness. A thin coating with low conductivity reduces thermal energy storage and enables the coating to respond quickly to gas temperature changes, displaying behavior comparable to the “temperature swing insulation” in Fig. 2 [15–18]. As a result, coating surface temperature is closer to the burned gas temperatures during combustion and similar to the fresh air temperature during gas exchange. A finite element model of a thin TBC subjected to experimentally measured heat fluxes and temperatures confirmed the ability of thin, low conductivity coatings to both reduce combustion heat transfer and minimize charge heating in a LTC engine [15].

The engine used for testing is a modified single-cylinder Ricardo Hydra, seen in Fig. 3, with four valves per cylinder, direct fuel injection, and a rebreathe style exhaust cam. Additional information regarding the engine geometry and timing specifications are found in Table 1.

Exhaust gas is sampled by a Horiba MEXA 7100D-EGR to provide unburned hydrocarbon (HC), CO, CO₂, O₂, and NO_x emissions measurements. Emissions are normalized by the injected fuel quantity and presented as emissions index values [20]. Additionally these emission concentrations are used to calculate both combustion efficiency [20] and the exhaust air-to-fuel ratio. Lambda measurements are also taken in the exhaust plenum via an ETAS ES630.1 and a Bosch LSU4.9 wideband lambda sensor. The accuracy of the lambda sensor measurement is ± 3% of the reading between the equivalence ratios of 0.6 and 1. Additionally, crank angle resolved exhaust runner pressure is measured by a Kistler 4045 A pressure sensor.

Additional combustion phasing control is available through coolant and oil temperature control. However, in this study, coolant and oil temperatures are held at a constant 95 °C.

Fuel is delivered by means of direct injection at 10.4 MPa and is timed to inject at 333 deg before combustion top dead center, allowing sufficient time for near homogenous mixing. A piston accumulator and high pressure nitrogen are used to supply fuel pressure. The test fuel used in this experiment is Gage Products 91 RON test fuel (40665-55f). Fuel flow is measured to an accuracy of ±0.2% by a Max Machinery 213 piston type flow meter.

For this experimental testing regimen, combustion phasing was allowed to vary, showing the real-world coating effects. Exhaust backpressure, the difference between the average exhaust and intake pressures, was adjusted with the gate valve in the exhaust to obtain a pressure differential of 3.7 kPa at wide open throttle with the baseline metal engine. This provides 45% hot residual gas fraction (RGF) via the rebreathe cams at the 2000 RPM, 11 mg fuel/cycle operating point. Subsequent tests with the YSZ coated piston were completed with the same gate valve position, though the residual fraction varied slightly due to changes in the exhaust temperature. The intake temperature was set to 90 °C, while coolant and oil temperatures were held constant at 95 °C. Fuel injected per cycle and air to fuel ratio were kept constant for both pistons at each operating point. Additional operating conditions are found in Table 2.

Further experiments were conducted to remove the effect of combustion phasing. For this testing, CA50 was set to 7 deg after top dead center (ATDC) for the YSZ coated and uncoated cases. Fueling, AFR, airflow, exhaust gate valve position, coolant, and oil temperatures were identical to the fuel-matched (FM) points. Combustion phasing for the metal engine was achieved by increasing the intake temperature until CA50 was 7 deg ATDC. This intake temperature was also used with the YSZ coated piston at the same load points. However, CA50 for the YSZ coated piston was retarded to 7 deg by adding external EGR. Normally, combustion phasing is achieved through modification of the intake temperature. However, imposing different intake temperatures for the metal and YSZ coated cases would confound the effects of the coating on combustion through alteration of the cylinder temperature at IVC and subsequent heat transfer phenomena.

An overview of the FMPM operating conditions is presented in Table 2. Additionally, EGR is given as a percent of the incoming fresh air mass, and the RGF is presented as a fraction of the total charge mass in the cylinder during the closed portion of the cycle. Likewise, the fresh air equivalence ratio phi (ϕ) and equivalence ratio with dilution, phi prime (ϕ′), are also included in Table 2.

The data were collected with a 0.5 crank angle degree resolution. The first step in postprocessing is ensemble averaging 200 consecutive cycles. The cylinder pressure is pegged to intake pressure in a ±10 deg window at bottom dead center during the intake stroke. The heat release rate is calculated by performing a first law analysis on the ensemble-averaged cylinder pressure [2]. The heat transfer coefficient is computed using a correlation specifically designed for use with HCCI engines by Chang et al. [21].

For this experiment, a thin, 150 μm YSZ coating was applied to a piston. The piston was chosen because it occupies a large portion of the combustion chamber area and requires little preparation to receive the coating. In order to maintain compression after coating application, a total of 150 μm was first machined off the piston crown. Next, an air plasma spray (APS) technique was used to apply a base layer of NiCrAl, and then, a YSZ low conductivity topcoat. Typical properties for APS YSZ coatings are included in Table 3.

Coating thickness was measured with a Fischer MP-20 Dualscope. The measured thicknesses are provided in Table 4, and the corresponding locations for the APS-YSZ piston are shown in Fig. 4. The thickness measurements were accurate to 0.5% of the thickness measurement. The measurements were taken “N” times at each spatial location, and subsequently the mean and standard deviation were recorded. Some spatial variation of coating thickness was noted, owing to the complex piston geometry. In addition to the piston, one of two cylinder head heat flux probes was also coated with YSZ. The average coating thickness on the probe was 157 μm with a standard deviation of 10.4 μm (n = 21).

Yttria-stabilized zirconia is often chosen for TBCs due to its high fracture toughness and high coefficient of thermal expansion [23,24]. The ASP process uses a high temperature plasma arc to melt powdered coating components [25]. The melted coating droplets are then accelerated with a gas jet toward the workpiece. The droplets impact the work surface forming lamellae, and layered accumulation of these “splats” builds the coating thickness [25]. The final coating composition is comprised of 7% Y₂O₃ and 93% ZrO₂, often called “yttria-stabilized zirconia.”

The engine is equipped with two Medtherm fast-response heat flux probes capable of providing subcrank-angle resolved temperatures. The cylinder head is provisioned for mounting two of these heat flux probes, denoted H1 and H2 in Fig. 5. The probes have a co-axial construction, with a central wire of constantan, an electrically insulating dielectric separator, and an outer casing made of iron. The ends of the tube then covered with a 10 μm layer of vacuum-deposited metal that provides an electrical connection between the thermocouple [8,9]. The microsecond-order temporal response of these probes is sufficient to measure with resolution finer than the 0.5 CAD at engine speeds relevant to this study. Additionally, these probes contain a second coaxial J-type junction 4 mm into the probe body, enabling surface heat flux calculation through a Fourier analysis. Refer to Ref. [21] for more details on the heat flux computation.

The surface thermocouple junction is normally exposed to the combustion chamber in the baseline metal engine, allowing surface temperature and heat flux measurements. However, with the addition of the YSZ coating atop the probe, the surface thermocouple junction is then located at the probe/TBC interface and measures the subcoating temperature. The YSZ surface temperature and heat flux are estimated from the subsurface temperature and heat flux measurements using an inverse heat conduction solver. More details on this solver are provided in Ref. [26].

The engine was first operated in the baseline configuration (also denoted as “uncoated” and “metal”) at the operating points listed in Table 2. Next, the engine was run with the YSZ coated piston, while all major engine inputs such as fueling, AFR, and temperatures were kept constant between the two cases. The purpose of this testing methodology was to see the real-world effects of using TBCs in HCCI engines.

The YSZ coating advanced combustion, increasing peak cylinder pressures and peak heat release rates. The coating's effect on cylinder pressures is seen in Fig. 6(a) and rates of heat release in Fig. 6(b). The combustion advance exhibited by the YSZ results from decreased compression heat transfer, leading to higher gas temperatures near top dead center.

The combustion advance illustrated in Figs. 6(a) and 6(b) occurred over the entire range of tested speeds and load. Figure 7 quantifies the YSZ-induced combustion advance as between 2.2 and 4.9 CA deg. Additionally, Fig. 7 shows that the combustion duration is shortened by 2.6–5.0 deg when utilizing the YSZ-coated piston. This trend in combustion duration is directly influenced by the YSZ-induced ignition advance. It is noteworthy that the TBC-induced combustion advance decreases as engine speed increases, as the real time for heat transfer decreases with engine speed.

The YSZ coating advanced combustion phasing and caused increased peak heat release rates and cylinder temperatures. These effects, combined with the decreased heat transfer from the YSZ coating and increased temperature in the wall-affected regions, led to enhanced autoignition, and increased combustion efficiency [6].

The main pathway for decreased HC emissions is more complete combustion. The YSZ coating produces a 25–37% decrease in unburned hydrocarbons over metal engine operation, as shown in Fig. 8. Additionally, the magnitude of hydrocarbon emission improvement decreased with increasing engine speed. This trend aligned with the effect of engine speed on time available for heat transfer and diminishing combustion phasing advance from the YSZ coating.

Carbon monoxide emissions decreased at all engine speeds for the YSZ–TBC equipped engine, see Fig. 9. The decrease in CO emissions was between 29% and 37% across the operating range; however, the change was not monotonic. The trend in carbon monoxide emissions was not a strong function of engine speed. Instead, CO emissions follow the trend in combustion advance. The reduction in carbon monoxide was highest for 1600 RPM and then for 1200 RPM, corresponding to the largest and second largest combustion advances. Advanced combustion phasing led to increased gas temperatures in the wall-affected regions, which improved carbon monoxide oxidation.

The large decrease in hydrocarbon and carbon monoxide emissions with the YSZ coating increased combustion efficiency. Combustion efficiency improves at all engine speeds during YSZ operation, as shown in Fig. 10. The increase in efficiency is between 1.0% and 1.9% over the baseline metal HCCI engine. This can be attributed to reduced heat transfer, and consequently increased charge temperatures in the near-wall region. This, in turn, causes more of the fuel-air charge to react.

However, the improvement in combustion efficiency diminishes with increasing engine speed. This effect is a direct result from the trend in the combustion advance and decreasing time for heat transfer impacts. The advance in combustion phasing, increase in cylinder pressure, and increased combustion efficiency combine to improve the cyclic work output during YSZ operation. This increases the gross-indicated thermal efficiency between 1.9% and 4.1%, as shown in Fig. 11. However, because the combustion advance is intermingled with the increased YSZ combustion efficiency, the YSZ impact is difficult to decipher from this testing. In particular, the combustion phasing of the metal engine at 1200 RPM was retarded beyond optimal. Hence, the advanced phasing achieved with the YSZ coated piston produced an added benefit, i.e., part of the 4.1% improvement should be attributed to combustion phasing rather than direct impact of oxidation reactions and reduced heat loss. For a more rigorous examination, the YSZ impact must be isolated from combustion phasing.

The effect of the TBC on combustion provided a number of benefits, such as reduced emissions and increased combustion and thermal efficiencies. However, to isolate the thermal impact of the YSZ–TBC, combustion phasing must be held constant. A fixed combustion phasing, CA50, of 7 CA deg ATDC was chosen for this investigation. The intake charge temperature was adjusted with the metal engine until CA50 of 7 deg ATDC was achieved. During YSZ operation, the same intake temperature was utilized, and external EGR was added until CA50 retarded to the design point. The addition of EGR occurs prior to the intake heater, so the mixture of fresh air and recirculated exhaust gas are the same intake temperature. It should be noted that the 1200 RPM operating point is absent from this section due to the inability to add sufficient EGR from a combination of low backpressure and physical size limitations of the EGR circuit. By maintaining a constant intake temperature, the charge temperature at IVC is constant between tests, isolating the impact of the YSZ coating on combustion and heat transfer.

Figure 12(a) illustrates the phase-matched pressure curves at 2000 RPM, and the respective rates of heat release are presented in Fig. 12(b). The cylinder pressure trace for the YSZ coated case is similar to the metal one. The peak rate of heat release is slightly higher for the YSZ case than for the baseline engine due to a slight decrease in combustion duration with the coating.

The relative phasing and combustion durations for the phase-matched points are presented in Fig. 13. With combustion phasing tightly controlled, it is apparent that combustion duration decreases by 0.1–0.6 deg CA due to the YSZ coating. Notably, the addition of the coating advances the CA90 point for the same CA50, shortening the CA50-90 duration between 0.4 deg and 0.5 deg. The coating increases the temperatures of the wall-affected regions, increasing the combustion rate, and shortening its duration. However, additional testing is planned to verify this effect.

After the effects of combustion phasing are removed, the improvement in hydrocarbon emissions from the coating alone is still present, but lower in magnitude when compared to the fueling-matched case. Namely, the YSZ coating caused a decrease in HC emissions between 2.3% and 11%, as shown in Fig. 14.

The change in carbon monoxide emissions with the YSZ coating was consistent with the hydrocarbon emission trend. Reductions in CO emissions of 4.3% and 8.8% were observed, as seen in Fig. 15. This reduction is from increased gas temperatures in the wall-affected regions, improving CO oxidation. Similarly to the HC emissions trends, the magnitudes of YSZ induced CO emission reduction is less than the fuel matched results shown in Fig. 9.

The decrease in hydrocarbon and carbon monoxide emissions led to increased combustion efficiency for the YSZ–TBC. For the phase matched points, combustion efficiency increased between 0.1% and 0.4%, as shown in Fig. 16.

The same trend seen in the FM operating points is present where shorter timescales decrease the impact of the coating at higher speeds.

Finally, through the use of the YSZ coating, the HCCI engine experienced between a 1.9% and 3.0% increase in gross indicated cycle efficiency as depicted in Fig. 17. A decrease in heat transfer during combustion increased the expansion work and compounded with the previously described impact of the TBC on combustion efficiency. The increase in indicated efficiency was slightly higher for the FMPM points than for the FM points. At 1600 RPM, this was caused by retarding combustion phasing from 3.9 deg to 7.1 deg CA with the YSZ coating, resulting in a reduction in heat transfer [2]. Meanwhile, at 2000 RPM, FMPM combustion with the YSZ coating was only advanced 0.3 deg CA toward more optimal combustion phasing when compared to FM operation.

The mechanism responsible for the observed improvements can be confirmed with directly measured heat flux. The crank angle-resolved YSZ surface heat flux, Fig. 18(a), and YSZ surface temperature, Fig. 18(b), are calculated using an inverse conduction solver and the two junction temperatures of the probe. For more information on the calculation methodology, please consult O'Donnell et al. [26]. The YSZ coated probe, Fig. 18(b), demonstrated a surface temperature similar to an idealized thin TBC (see “temperature swing insulation” in Fig. 2).

The YSZ surface temperature profile is similar in magnitude to the metal surface temperature during gas exchange, thereby avoiding any adverse effect on volumetric efficiency. The YSZ surface then undergoes a 36 °C temperature swing during compression and combustion, much higher than in the case of the metal piston.

The heat transfer shown in Fig. 19(a) was calculated using the instantaneous surface heat flux from the cylinder head probe location and the exposed cylinder surface area. A cumulative sum of the heat transferred during combustion and expansion is included in Fig. 19(b).

The large coating temperature swing helps to reduce the difference between gas temperature and the wall temperature, reducing the driving force for combustion heat transfer and contributing to the increased thermal efficiency for the YSZ case. Eventually the instantaneous heat loss with the YSZ piston exceeds that of the baseline metal configuration, illustrating a well-known heat storage effect, see Fig. 19(a). Consequently the difference in cumulative heat loss increases early in the expansion process and eventually diminishes close to 450 CA deg ATDC of gas exchange. In short, the key to achieving the desired effect on the cycle efficiency is a dynamic change of the surface temperature, rather than excessive insulation.

Application of a YSZ–TBC to the piston of an LTC engine advanced combustion phasing between 2.2 deg CA and 4.9 deg CA. The combustion advance boosted peak cylinder pressures and temperatures, causing increased peak rates of heat release. The higher cylinder temperatures, especially in the wall-affected zones, caused more of the fuel-air charge to autoignite. The result was 25–37% less unburnt hydrocarbons in the exhaust and 1.0–1.9% higher combustion efficiency. Carbon monoxide emissions also decreased between 29% and 37%. The increase in the cylinder pressure combined with the additional energy released from higher combustion efficiency to increase gross-indicated thermal efficiency by 1.9–4.1% over the metal engine.

Subsequent experimentation with cooled EGR was designed to hold combustion phasing constant and isolate the thermal effect of the coating on cycle parameters. With constant combustion phasing, the magnitude of YSZ benefits in combustion efficiency and emissions decreased relative to FM operation, but were still present. With the effects of the coating isolated, the YSZ coating lowered hydrocarbon emissions by 2–11% and improved combustion efficiency by 0.1–0.4%. Carbon monoxide emissions decreased between 4.3% and 8.8%. Gross indicated efficiency improved by 1.9–3.0% due to the thermal effects of the coating on instantaneous heat loss during combustion.

The advance in combustion phasing with the YSZ coating has a large, positive impact on HC and CO emissions, combustion efficiency, and thermal efficiency. After removing the combustion advance, many of the key benefits of the TBC are preserved, but have a reduced magnitude. Further improvements in these parameters are likely with additional reductions in heat transfer. Use of engineered coatings with reduced thermal conductivities would likely offer enhanced benefits when used in LTC engine applications [28,29], this will be investigated in future efforts.

The authors want to thank Dr. Eric Jordan at the University of Connecticut for providing the YSZ coating and General Motors R&D for providing the single-cylinder engine used in this investigation.

Funding for this research was provided by the National Science Foundation (NSF)/Department of Energy (DOE) Partnership on Advanced Combustion Engines Grant No. 1258714, “Thermal Barrier Coatings for the LTC Engine—Heat Loss, Combustion, Thermal versus Catalytic Effect, Emissions, and Exhaust Heat”.

• National Science Foundation (NSF)/Department of Energy (DOE) Partnership on Advanced Combustion Engines (Grant No. 1258714).

ATDC =

after top dead center

BTDC =

before top dead center

CA50 =

crank angle at 50% mass fraction burned relative to top dead center, firing

CCD =

combustion chamber deposits

CO =

carbon monoxide

CO₂ =

carbon dioxide

EGR =

exhaust gas recirculation

FM =

fuel matched

FMPM =

fuel matched, phase matched

HC =

hydrocarbon

HCCI =

homogenous charge compression ignition

IVC =

intake valve closing

LTC =

low temperature combustion

NO_x =

nitrogen oxides

O₂ =

oxygen

PLIF =

planar laser-induced florescence

RGF =

residual gas fraction

TBC =

thermal barrier coating