Inter-Laboratory Comparison Exercise - PowerPoint Presentation

1. Inter-Laboratory Comparison ExerciseCPC CALIBRATIONCALIBRATION AEROSOL AND LABORATORY SETUPPMP Meeting2018-11-07Alexander TerresKey Findings

2. Round Robin: StatusAll measurements and evaluation finishedReport circulated to all participants for final commentsTransfer of findings to CPC calibration sub-groupTerres: CPC Calibration Round Robin | Key Findings, 47th PMP Meeting | 2018-05-16Page 2

3. Key Findings: Aerosol GeneratorSoot aerosol provided lower counting efficiencies than emery oil at 23nm (29% vs 49%) and at 41nm (78% vs. 91%)Soot-like generators offered a low uncertainty, even without standardization Inter-laboratory standard deviation 2% - 3% for in-house CAST and Palas at 41nm and 70nmInherently, soot has more degrees of freedom than electrospray + emery oil Required: Proven setup, regular checks and standardization (e.g. fuel/air ratio, size distribution)Thermal treatment necessary for stability and universal application (mandatory for PCRF and PN-PEMS)Transfer function from soot-like to emery oil is possible (but was not part of the RR)To be checked: Can the CPC meet D50 and D90 with soot by changing temperatures?Using emery oil is a trade-off: Low uncertainty from a purpose-specific generator (only for CPC), but high uncertainty from overall procedure with various generators (Emery oil + soot + NaCl + …)Terres: CPC Calibration Round Robin | Key Findings, 47th PMP Meeting | 2018-05-16Page 3

4. Key Findings: CPC/Reference RepeatabilityWithin 1 LaboratorySame reference CPC vs. same electrometer1 year, 40 calibrationsStandard deviation/range of valid measurements:@41nm: 1.5 / 90%-96%@70nm: 1.3 / 90%-95%Without calibration: 0.7 – 1.2Blue/green: before/after calibration variability within 1 laboratoryTerres: CPC Calibration Round Robin | Key Findings, 47th PMP Meeting | 2018-05-16Page 4

5. Key Findings: Circulated CPCNotable deterioration between starting measurement and final measurement at manufacturer (w/ emery oil)At 23nm: 51%  46%At 41nm: 93%  88% At 55 nm: 96% 93%Sample flow was reduced but within tolerance (efficiency is flow corrected)In addition to laboratory facilities uncertainty,Circulated CPC seems to be a significant source of uncertainty between labsNormalization to cancel out reference counter did not improve comparability between labsCPC stability between labs hard to estimate (see above). Influence of transport, drying and contamination possibleEmery oil calibrated CPCs require recalibration to meet legal targets at 23/41nm for soot-aerosol  validate!(not part of RR) Comparability between CPC models better when KF is always applied to calibration at 23/41nm Some CPC models internally apply a KF all the time Standardization necessary!Page 5

6. Key Findings: ProcedurePlateau uncertainty: At 70 nm the CPC did not reach the plateau with soot (2-3 percentage points less)At 70 nm the std. deviation was 2-3 percentage points w/ miniCAST, PALAS (range: 90%-96%), Doubly charged correction at this size was 3% (Lab A) to 7% (Lab B) when using electrometerPlateau uncertainty is best indicator for improvement in calibration (biggest impact in the field – PMP use!)Implementation of ISO 27891 procedure + reference aerosolOpen questions at 23nm where uncertainty is largest (not investigated in this RR):DMA calibration Thermal stability, soot morphology (miniCAST) (measurements from one lab showed an influence)CPC drift, contaminationPage 6

7. Key Findings: Procedure/DocumentationCalibration results should be shown in the form: ηCPC = 95% ±3% documentation of the uncertainty budget Example in ISO 17025 for calibration labsCPCs can deteriorate during one calibration interval (1 year)Important: quality check of calibration setup with shorter frequency (1-2 months)WLTP regulation includes these checks for test CPCsNo calibration for the aerosol generator availableRegular documentation of performance + operation points necessaryNeeded for complete traceability of PN calibrationTerres: CPC Calibration Round Robin | Key Findings, 47th PMP Meeting | 2018-05-16Page 7

8. Key Findings: 10nmOnly few measurements with 10nm Round Robin CPC (due to lack of suitable reference and/or lack of time)Impression: uncertainty at 10nm/15nm similar or better than engine exhaust CPC at 23nmMaterial influence at 10nm/15nm: seems smaller than for 23nm; few data General points, calibration:Sub-10nm reference device needed: electrometer, ultrafine CPC (full flow!)Huge impact of diffusion losses: Residence times must be the same, mathematical correction significantly increases the uncertainty and can manipulate the resultsGenerator:miniCAST able to produce 10nm particles, composition not investigated  concentration 1000-3000 1/cm³Thermal stability at 10/15nm needs to be demonstratedDMA:Long-DMA suitable for 10/15nm ?Traceable DMA calibration unclear (typically extrapolation from 70nm/100nm PSL)Page 8

9. Key Findings: ConclusionsGenerator: soot-like provides low level of uncertainty; still more degrees of freedom than emery oilReference Counter: influence relatively small, traceability needed, flow correction a mustCPC: deterioration during 1 yearProcedure: Implement common procedure following ISO 27891 + generator for PMPDocumentation: for calibration lab – provide uncertainty budget + regular quality checks (e.g. monthly)10nm: Soot generator needs further investigation, e.g. properties of CAST aerosol, concentration with PALAS DMA calibration needs further investigation Aerosol influence might be different than for 23nm CPCsTerres: CPC Calibration Round Robin | Key Findings, 47th PMP Meeting | 2018-05-16Page 9

10. Key Findings: OutlookOverall PN measurements uncertainty will be improved significantly if:A single standard aerosol is definedIt is used for both stationary PMP-compliant devices and mobile PN-PEMSAn agreement is made on K-factor applicationTraceability and ISO 27891 are implemented by all usersThese topics should be covered by calibration procedures for future post Euro 6 EU legislation.Terres: CPC Calibration Round Robin | Key Findings, 47th PMP Meeting | 2018-05-16Page 10

11. Terres: CPC Calibration Round Robin | Key Findings, 47th PMP Meeting | 2018-05-16Page 11Thank you for your attention