Aristotle University Of Thessaloniki (AUTH)
Development, experimental investigation, and field applications of an optoacoustic sensor for Black Carbon measurement
Abstract
dc:descriptionBlack carbon (BC), the strongly light-absorbing carbonaceous product of incomplete combustion, is the second-largest anthropogenic contributor to near-term warming after CO₂ and among the most health-relevant components of fine particulate matter. It lives in the atmosphere for only days to weeks, so that emission reductions can deliver rapid climate and health benefits. This combination has made BC a priority "superpollutant" target. Due to its complex and evolving nature, BC measurement currently rests on an unresolved metrological foundation. The need for real-world, absorption-based measurement of BC emissions is the motivation for this thesis.The opening chapters establish the nature of the problem. BC is not a single, well-defined substance but a family of light-absorbing carbonaceous particles. Measurement is based on operationally defined quantities — equivalent BC (eBC), refractory BC (rBC) and Elemental Carbon (EC) — each tied to a particular principle and physical property, and none mutually traceable. The climate-relevant property is light absorption, but the filter-based photometers that dominate monitoring networks infer absorption indirectly and carry well-documented loading and multiple-scattering artefacts, while the absence of an agreed reference method leaves measurements structurally incomparable across sites and techniques. Fixed monitoring stations sample aged, transported, source-mixed aerosol, whereas mitigation is enacted at the source, where the governing quantity, the emission factor under real operating conditions, can only be measured directly. The field therefore needs an instrument that is physically grounded, field-deployable, and able to carry direct, filter-free absorption measurement to the source. Photothermal interferometry (PTI) is positioned as the emerging lab-grade absorption reference, and photoacoustic spectroscopy (PAS) as its field-deployable counterpart, sharing the same direct-absorption basis while trading some metrological rigor for portability and robustness.The sensor developed in this work is a quartz-enhanced optoacoustic instrument built around a novel ellipsoidal acoustic chamber. A continuous-wave 808 nm laser diode, a wavelength conventionally chosen for BC selectivity and strong absorption, is amplitude-modulated and focused at one focal point of the cavity. The ellipsoidal geometry refocuses the resulting acoustic signal onto a quartz tuning fork (QTF) resonating near 100 kHz at the second focal point, spatially separated from both the optical path and the sample flow. This decoupling of signal generation from detection yields a contamination-tolerant, conditioning-free, and inexpensive architecture. The sensitivity is carried by the high Q factor detector rather than by acoustic resonance. The ellipsoidal chamber also results in relaxed alignment tolerances enabling low-cost, 3D-printed construction. The mechanicalviidesign and geometry optimization were developed within the wider RSENSE group; the experimental characterization and all field demonstrations reported here are the contribution defended in this thesis.Laboratory characterization of the standalone sensor established its intrinsic metrological behavior over a campaign using an AVL Micro Soot Sensor (MSS) as a co-principle reference. The QTF resonance proved stable to within about 0.93 Hz around 99,886 Hz with no systematic drift, allowing a single fixed modulation frequency. The photoacoustic response was strictly linear with mass concentration when comparing mean plateau values across roughly three orders of magnitude, from the μg m⁻³ regime to about 11 mg m⁻³, with day-to-day repeatability of 7.5% (3σ) and a reproducible ~8% spread over hour-long stability runs. The one property not reaching laboratory-grade performance was the noise floor, attributed to the simpler mechanical construction; this doesn’t hinder the sensor's use, since target source concentrations lie well above it. Consistent with this scoping, the near-unity correlations are presented as evidence of linearity against a co-principle reference, not of traceable accuracy.A subsequent four-instrument intercomparison on a wide variety of laboratory-generated particles tested whether photoacoustic detection, and this miniaturized implementation in particular, delivers what the metrological case anticipates. Three optoacoustic instruments of differing architecture (the commercial MSS2, the research-grade BCT, and the present sensor) were compared against the filter-based AE33 aethalometer across miniCAST and spark-generated aerosols. The three optoacoustic instruments agreed in their response independently of organic-carbon content and particle morphology, indicating that the robustness resides in the optoacoustic principle itself rather than in any single implementation. The optoacoustic instruments tracked rapid concentration changes with rise and fall times of roughly 0.8–2.4 s, four- to fivefold faster than the AE33's ~6–7 s, and the filter-based instrument underestimated short transient peaks by 25–40%, a one-sided, structural bias that averaging cannot remove and that falls most heavily on the brief acceleration and cold-start events dominating transport emission budgets. The AE33 also ill-quantified the morphologically different PALAS aerosol, while the photoacoustic instruments did not. Miniaturization imposed no response-time penalty; the sensor's sole significant limitation was a drift-limited detection limit of about 2 μg m⁻³, above urban ambient levels but comfortably below the source concentrations it targets, where dilution can be tuned.The real-world applications form the thesis centerpiece. The main study reports the first on-road, real-driving BC emission measurements for L-category vehicles, achieved by integrating the sensor into a Smart Emission Measurement System light enough to mountviiieven on a 50 cc scooter. A dedicated conversion factor for the further-miniaturized configuration was derived against the MSS (R² = 0.97). On the chassis dynamometer the portable system reproduced the reference's emission dynamics second-by-second, and on the road it yielded the first measurement-based BC emission factors for Euro 5 L-category vehicles, ranging 0.05–0.20 mg/km. Crucially, it resolved the emission structure of this source: BC is emitted as short, intense peak events driven by acceleration and cold start, with the highest-emitting 10% of one-second intervals accounting for 73.5–90.1% of total trip BC mass — confirming that capturing this source demands precisely the compactness, sensitivity, and fast response the sensor provides. Two complementary demonstrations extend the application envelope: an earlier on-board passenger-car campaign establishing time-response fidelity against PEMS particle number (r = 0.91, signal in arbitrary units), and a heavy-duty DPF-failure-detection study in which the sensor cleanly separated functional from malfunctioning filters under two driving cycles on a chassis dynamometer, with a dilution-drift day-by-day artefact corrupting only the absolute scale and not the qualitative or temporal information.The synthesis draws these threads together. The agreement of differently configured optoacoustic instruments on real, transient exhaust extends the controlled-particle equivalence of the laboratory chapters into field conditions, establishing the robustness of the optoacoustic absorption principle for BC outside the laboratory. The peak-dominated emission structure carries direct implications for near-road exposure, for emission inventories that rely on average emission factors, and for fleet-screening methods that may miss the brief events carrying most BC mass. The principal limitation surfaced by deployment is the fixed mass absorption cross-section: because the conversion from absorption to mass depends on soot morphology, mixing state, and coating, the single calibration factor carries a morphology-dependent uncertainty wherever real soot departs from the calibration aerosol, a limitation of the conversion step, not of the absorption measurement itself. The sensor's contribution to the standardization problem is therefore framed honestly: by sharing the physical basis of the lab-grade absorption reference, it belongs to the instrument class that can extend physically-grounded absorption toward the source. The thesis demonstrates cross-configuration consistency on real exhaust, not a closed traceability loop that requires further steps. Future directions include anchoring calibration to a traceable absorption standard, lowering the detection limit toward ambient levels, adding wavelengths for black/brown-carbon apportionment, and exploiting the achieved miniaturization for airborne (UAV) plume measurement.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Raptis, Ioannis
- Ράπτης, Ιωάννης
Subjects
dc:subject × 21- Ατμοσφαιρική ρύπανση
- Σωματιδιακή ύλη ντίζελ
- Καυσαέρια, μετρήσεις
- Αιθάλη
- Μαύρος άνθρακας (BC)
- Φωτοακουστική μέθοδος
- Όργανο μέτρησης
- Αισθητήρας
- Black carbon (BC)
- Aerosol particles
- Air pollution
- Optoacoustics
- Photoacoustic method
- Sensor
- Exhaust measurement
- Επιστήμη Μηχανολόγου Μηχανικού
- Επιστήμες Μηχανικού και Τεχνολογία
- Θερμοδυναμική
- Mechanical Engineering
- Engineering and Technology
- Thermodynamics
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/63065