Color ratio pyrometry tracks soot formation in DISI engine cylinders

Soot, the dark particulate matter that billows from the tailpipes of modern gasoline engines during cold starts, has long been measured only after it leaves the cylinder — by which point the story of how it formed is already lost. Now, researchers at the Karlsruhe Institute of Technology have peered directly inside a running engine and watched soot being born in real time, using nothing more exotic than a high-speed color camera. In a study published in Automotive and Engine Technology, Lukas Heinz, Uwe Wagner and Thomas Koch demonstrate that Color-Ratio Pyrometry, a single-camera optical technique, can map both the temperature and the concentration of soot inside a turbocharged direct-injection spark-ignition engine — and in doing so, they have revealed a counterintuitive truth about ethanol blends: adding alcohol to the wrong kind of gasoline can actually make particulate emissions worse, not better.

The technique the team refined builds on a physics problem that has occupied combustion scientists since the 1930s. Individual soot particles behave approximately like black bodies, radiating with an intensity that depends on wavelength and temperature. But a cloud of soot does not radiate like a perfect black body: its emissivity depends on how many particles lie along the optical path. Hottel and Broughton captured this empirically in a classic relation in which emissivity at a given wavelength depends on an exponential term containing the product KL — an absorption coefficient multiplied by the length of the flame along the optical axis. Because KL is proportional to the soot concentration along the line of sight, measuring radiation at multiple wavelengths allows researchers to solve simultaneously for both the apparent temperature and the soot loading. The catch, in the conventional two-color method, is that the absolute spectral power density must be measured directly, which demands elaborate optics: beam splitters, narrow-band filters, sometimes two cameras, and painstaking radiometric calibration that is easily thrown off by soot deposits on the optical windows.

Color-Ratio Pyrometry sidesteps nearly all of that. Instead of measuring absolute intensity, the method compares two color ratios — red to green, and red to blue — captured simultaneously by a single-chip color high-speed camera, in this case a Vision Research Phantom v1612 mounted on an 8-millimeter endoscope looking into the first cylinder of a BMW B48 four-cylinder engine. Because a ratio discards the absolute signal level, the technique is inherently robust against window fouling and against uncertainties in the distance between the glowing soot cloud and the lens. The red-green ratio, it turns out, is only weakly sensitive to soot concentration in this configuration, making it an excellent thermometer; the red-blue ratio grows increasingly sensitive to soot loading as temperature rises. The team exploited this asymmetry with a two-step calculation: temperature is first determined from the red-green ratio alone, and that temperature then unlocks an unambiguous value of KL from the red-blue ratio. All raw 12-bit Bayer data were kept linear to avoid nonlinear gamma correction, demosaiced in post-processing, and converted through a look-up table built by convolving Planck’s radiation law with the measured spectral sensitivity of each camera channel.

Validation came from two independent directions. First, the researchers compared the optically derived gas temperatures with cylinder-pressure-based calculations using single-zone modeling with variable isentropic exponents and an assumed residual gas fraction of five percent. The two approaches agree closely between roughly 10 and 30 degrees of crank angle after top dead center, with measured temperatures around 2500 kelvin — squarely within what is expected for the burnt zone of a gasoline combustion chamber. Outside that window, deviations are physically meaningful rather than errors: immediately after ignition, radical chemiluminescence contaminates the blue channel, while late in combustion the fading soot glow drowns in noise. Second, and crucially, the optical soot indicator was benchmarked against exhaust measurements from a Cambustion DMS-500 differential mobility spectrometer, mounted downstream of the turbocharger, which resolves particle mobility diameters from 5 nanometers to 1 micrometer at 10 hertz. The summed KL integral over the image tracked the exhaust particle number concentrations remarkably well, confirming that the camera was seeing the soot that actually survives to leave the engine.

With the diagnostic validated, the team turned it loose on a question with real-world stakes: how does fuel formulation shape soot formation in a modern gasoline direct-injection engine? Three fuel families were tested at steady-state load points from idle to 9.4 bar BMEP, under both cold-start conditions with coolant at 20 degrees Celsius and fully warm operation at 95 degrees. One reference fuel, dubbed ALK-G, was an aromatic-free alkylate gasoline with a low final boiling point of 183 degrees Celsius — the best-case scenario. The other, LV-G, was a low-volatility gasoline rich in heavy aromatics with a final boiling point of 230 degrees Celsius, chosen deliberately as a high-soot worst case. Each was then blended with ethanol or methanol in mass fractions of up to 30 percent, and the two reference fuels bracket the soot-forming potential of ordinary EN 228 pump gasoline.

The results for the pure fuels behaved as textbooks predict. The aromatic-free alkylate produced so little particulate that the optical system detected essentially no soot glow at all — a combination of low soot-forming chemistry and the strong homogenization that comes with a volatile, quickly evaporating fuel. The low-volatility fuel, by contrast, lit up the combustion chamber with substantial soot luminance and produced markedly higher particle concentrations in the exhaust. But the surprises began when alcohol entered the picture. When ethanol was blended into the low-volatility gasoline, soot formation increased rather than decreased, and raising the ethanol content from 20 to 30 percent pushed particle concentrations still higher. Yet when the identical ethanol fractions were blended into the volatile alkylate, no increase in soot or particle emissions appeared at any operating point.

The optical imaging explained why. Ethanol’s high evaporation enthalpy chills the mixture-formation process, robbing the injected spray of the heat needed to vaporize the high-boiling components of the base fuel. In the low-volatility blends, the temperature at the injector tip dropped below the threshold needed to fully evaporate LV-G’s heavy fractions, leaving liquid residue that burned as an oxygen-starved diffusion flame — a phenomenon the researchers call tip sooting. Early injection timing during warm operation also wetted the piston crown with liquid fuel in the ethanol blends, producing a second soot-formation mode the team terms a pool fire, a persistent burning fuel film that generates soot precisely in the late combustion phase, after roughly 20 degrees of crank angle past top dead center, when too little oxygen remains for the OH radicals in the flame front to oxidize it away. Because soot oxidation is largely finished by then, these late KL peaks translated directly into elevated exhaust particle numbers.

Methanol added yet another twist. The 15 percent methanol blend in the low-volatility fuel produced a striking, spatially homogeneous soot glow beginning unusually early in the combustion process — a pattern the authors attribute to the rapid, early conversion of methanol promoting fast growth of polycyclic aromatic hydrocarbons, the molecular precursors of soot. Similar observations have been made under flash-boiling conditions in earlier work, but the Karlsruhe team ruled out flash boiling here, since the fuel was conditioned to 20 degrees Celsius and illuminated images of the injection showed no signs of it. Notably, the methanol blend’s pronounced early glow was not matched by a proportional rise in exhaust soot, suggesting the nascent soot precursors were largely oxidized before they could grow into particles — although the blend did exhibit a persistent nucleation-mode particle signature across all operating points.

Beyond the fuel chemistry, the study carries a methodological message for engine developers. Because Color-Ratio Pyrometry requires no beam splitters, no narrow-band filters and no absolute radiometric calibration, it turns a standard color high-speed camera into a quantitative soot diagnostic that delivers temperature fields, soot volume fraction fields, and — through the KL integral — a cylinder-resolved soot mass indicator, all synchronized to crank angle. The spatial information is something no exhaust-line instrument can provide: knowing that a pool fire is blooming on the piston crown at a particular crank angle, or that soot is streaming from the injector tip, tells engineers exactly which lever to pull. The authors point to late adaptation of the main injection to prevent piston wetting at higher ethanol contents, and to injector design changes that reduce deposit formation and thereby suppress tip sooting, as concrete control strategies the imaging makes possible. Their formulation advice is equally specific: avoid combining low-boiling oxygenates with high-boiling gasoline components, because that pairing is what traps heavy fractions in liquid films that burn sootily.

The broader significance lies at the intersection of fuel diversification and emissions regulation. As renewable and synthetic gasoline components multiply, cold-start and low-temperature operation remain the moments when particle emissions are hardest to control, and the study shows that the oxygen content of a fuel alone is an unreliable predictor of its particulate behavior — what matters is the interplay of boiling range, evaporation enthalpy, aromatic content and injection strategy. By making the invisible chemistry of the combustion chamber visible, frame by frame at thousands of frames per second, the Karlsruhe team has given fuel designers and calibrators a tool to see not just how much soot an engine makes, but where, when and why. In an era when every milligram of particulate matters, that kind of sight may prove as valuable as the cleaner fuels it will help create.