Measuring VOCs
in the Atmosphere
From proton-transfer reaction mass spectrometry in the field to global chemistry-climate models — the complete toolchain for understanding volatile organic compounds.
Why VOC Measurement is Challenging
VOCs exist in the atmosphere at mixing ratios ranging from parts-per-trillion (pptv) to parts-per-billion (ppbv), across a staggering chemical diversity of >1000 species with vastly different physical properties. No single instrument can measure the full VOC spectrum. The field has therefore converged on a complementary suite of techniques, each optimised for specific VOC classes, time resolutions, and deployment environments.
The Measurement Challenge in Numbers
Biogenic isoprene over tropical forests reaches ~10 ppbv, while toxic urban benzene must be quantified at <1 ppbv against a complex matrix of thousands of co-eluting compounds. Remote marine DMS fluxes demand sub-pptv sensitivity with fast response to resolve eddy covariance fluxes. Reactive species like Criegee intermediates have lifetimes of milliseconds, demanding near-real-time detection. This dynamic range and chemical complexity demands a multi-instrument approach combined with state-of-the-science atmospheric models.
VOC Measurement Techniques — An Overview
Measurement approaches fall broadly into online/real-time (continuous, fast-response instruments) and offline/grab-sample methods (collected in field, analysed in lab). Each has trade-offs in sensitivity, chemical coverage, cost, and deployment logistics.
Proton-Transfer Reaction MS (PTR-MS)
Soft chemical ionisation using H₃O⁺ as reagent ion. Near-universal ionisation of most VOCs at their molecular ion (M+H)⁺. Real-time, high-sensitivity detection at Hz temporal resolution. PTR-TOF-MS variant resolves isobaric compounds by exact mass.
Online · Real-time · pptvGas Chromatography–Mass Spec (GC-MS)
Gold standard for speciated VOC identification. Separation by GC followed by MS detection provides unambiguous molecular identification. Typically requires pre-concentration on Tenax or DNPH sorbent tubes. Offline analysis; excellent for C2–C12 NMHCs and oxygenated VOCs.
Offline · Speciated · pptv–ppbvGC–Flame Ionisation Detection (GC-FID)
Workhorse for light hydrocarbons (C2–C6 NMHCs). FID provides near-universal carbon response; canister or bag sampling common. High precision for quantification of ethane, propane, isoprene, and alkene ratios. Used in many long-term monitoring networks.
Offline / Airborne · pptvDOAS (Diff. Optical Absorption Spectroscopy)
Absorption spectroscopy using sunlight (LP-DOAS, MAX-DOAS) or artificial light sources. Measures column-integrated or path-averaged concentrations of formaldehyde, glyoxal, and other oVOCs. MAX-DOAS enables vertical profiling of HCHO for satellite validation.
Remote sensing · Column / ProfileFTIR Spectroscopy
Fourier Transform Infrared spectrometry enables simultaneous detection of multiple trace gases including C₂H₂, C₂H₄, HCHO, and CH₃OH against the solar disc (solar FTIR / TCCON stations). Provides total column measurements consistent with satellite products like MOPITT and IASI.
Column · Remote sensing · ppbvChemical Ionisation MS (CIMS / CI-APiTOF)
Highly selective ionisation chemistry (I⁻, CF₃O⁻, Br⁻ reagent ions) targeting oxygenated VOCs, peroxy acids, and extremely low-volatility organics (ELVOC). Critical for SOA formation studies and HOM (Highly Oxygenated Molecules) detection in chamber and forest environments.
Online · Oxidised VOCs · fptvEddy Covariance (Flux Measurements)
Combines fast (≥10 Hz) VOC concentration measurements (PTR-TOF-MS or CIMS) with 3D wind data from a sonic anemometer. Derives net ecosystem-atmosphere VOC exchange fluxes. Essential for constraining biogenic emission inventories and evaluating the MEGAN emission model.
Surface fluxes · 10 Hz · EcosystemSatellite Retrievals (TROPOMI / OMI)
TROPOMI on Sentinel-5P retrieves global HCHO and CHOCHO columns at 3.5×5.5 km resolution, daily. Used to infer top-down VOC emission inventories, validate model fields, and detect biomass burning plumes. OMI has provided a 20-year HCHO record for trend analysis.
Global · Column · Daily coveragePTR-TOF-MS: Ionicon 8000 (Analytik GmbH, Austria)
The Proton-Transfer Reaction Time-of-Flight Mass Spectrometer is the central instrument in my field measurement work. The Ionicon 8000 represents the state-of-the-art in online VOC detection, combining the chemical selectivity of PTR ionisation with the mass resolution of TOF-MS to resolve hundreds of VOC species simultaneously at sub-second time resolution.
The Ionicon 8000 operates as a high-sensitivity online mass spectrometer using soft proton-transfer chemical ionisation. Ambient air is continuously sampled into a drift tube where H₃O⁺ reagent ions react with VOC molecules. Because the proton affinity of most VOCs exceeds that of water, proton transfer is thermodynamically favourable, producing (M+H)⁺ ions with minimal fragmentation. The TOF analyser then separates ions by mass-to-charge ratio at mass resolving power of >4000 m/Δm, allowing isobaric separation — e.g. distinguishing C₅H₈ (isoprene, m/z 69.070) from C₄H₅O (methylbutenol fragment, m/z 69.034).
In my field campaigns — spanning urban environments in India, marine boundary layer transects in the Indian Ocean, and forest canopy measurements — the instrument operated in both continuous ambient monitoring mode and eddy covariance flux mode (10 Hz). Typical deployment involved a Teflon-lined inlet with a Nafion dryer to remove humidity, which would otherwise suppress ionisation efficiency.
Key Operational Considerations
- Drift tube conditions: E/N ratio (ratio of electric field to gas number density) tuned between 100–130 Td. Higher E/N reduces clustering but increases fragmentation — critical for isoprene vs. MVK/MACR discrimination at m/z 69 and 71.
- Humidity sensitivity: H₃O⁺ signal suppression by water vapour requires careful humidity normalisation, especially in tropical environments (>90% RH). Nafion inlet drying or post-hoc humidity correction applied.
- Inlet residence time: Minimised by short (~1 m) heated Teflon tubing at 60°C to prevent wall loss of sticky oxygenated VOCs and monoterpenes.
- Calibration: Multi-point calibration using certified gravimetric standards (Apel-Riemer NCAR standards, 40-component gas mixtures). Sensitivity factors applied per compound class.
- Power & logistics: Instrument requires ~200W; deployed with portable generator on ship campaigns. Weighs ~35 kg; aircraft-deployable variant requires dedicated rack and vibration isolation.
- Data analysis: Raw TOF spectra analysed using PTRwid (MATLAB-based) or PTRAnalyser (Python). Peaks fitted with Gaussian profiles; mass axis calibrated to internal standards (acetone m/z 59, acetonitrile m/z 42).
discharge
impact
ion (PA=166 kcal)
+ analyte VOC
>PA(H₂O)
analyser
mass spectrum
factors
time series
Isoprene discrimination challenge: At m/z 69.070, PTR-MS detects both isoprene (C₅H₈·H⁺) and the sum of its primary oxidation products MVK + MACR + ISOPOOH. In high-E/N mode, fragmentation of monoterpene products also contributes. Accurate isoprene quantification in aged air masses requires auxiliary HCHO or MVK/MACR measurements to apply oxidation corrections.
GC-MS & GC-FID: Speciated VOC Analysis
While PTR-TOF-MS provides real-time data, Gas Chromatography–Mass Spectrometry is indispensable for unambiguous molecular identification and quantification of VOCs that share nominal masses. GC-MS forms the calibration backbone and speciation dataset against which PTR-MS identifications are validated.
Canister Sampling + GC-MS/FID
Electropolished 6L SUMMA canisters or 1L mini-canisters are used for grab sampling ambient air. Back in the laboratory, pre-concentration on a cryo-trap or Peltier-cooled adsorbent (Tenax-TA + Carbotrap + Carbosieve) followed by thermal desorption into the GC column enables sub-pptv detection. Used extensively in Indian Ocean ship cruise campaigns for C₂–C₁₂ NMHCs, monoterpenes, and DMS quantification.
Grab sample · Speciated · pptvDNPH Cartridge + HPLC-UV
Carbonyl compounds (aldehydes and ketones: formaldehyde, acetaldehyde, acetone, MEK) are sampled by active pumping through 2,4-dinitrophenylhydrazine (DNPH) coated silica cartridges. Hydrazones are eluted with acetonitrile and quantified by HPLC with UV detection at 360 nm. ISO/NIOSH-standard method. Used for indoor air quality assessments and urban carbonyl profiles.
Carbonyls · Indoor / Outdoor · ppbvSorbent Tube + TD-GC-MS
Passive or active diffusive sampling on Tenax-TA or multi-sorbent tubes (Tenax-Carbotrap) provides time-integrated VOC measurements. Thermal desorption (TD) at 250°C releases analytes into the GC column. Commonly used for site screening, workplace exposure assessment, and forested site characterisation. Complementary to PTR-MS for heavier sesquiterpene identification.
Time-integrated · >C₆ · pptv–ppbvOnline GC with Auto-Sampler
Automated online GC systems (e.g. Perkin-Elmer Clarus, Agilent 6890 with cryo-trap) provide hourly resolved speciated VOC data. Air is automatically drawn, cryotrapped, and injected into GC columns. Combined with PTR-TOF-MS continuous data, this setup allows robust mass-balance analysis and isoprene/terpene flux attribution. Used at the ATTO tower in Amazon and comparable forest sites.
Hourly automated · C₂–C₁₅ · FieldPTR-TOF-MS vs. GC-MS — complementary strengths: PTR-TOF-MS excels at temporal resolution (1 Hz vs. GC's 30–60 min cycle), making it essential for flux measurements and transient event capture (biomass burning plumes, ship stack contamination detection). GC-MS excels at definitive molecular identification — critical for distinguishing, e.g., isobaric monoterpene isomers (α-pinene vs. β-pinene at C₁₀H₁₆·H⁺, m/z 137) that PTR-MS cannot separate. In my campaigns, PTR-TOF-MS provided the time series backbone while canister GC-MS provided the speciation "ground truth."
From Raw Signal to Atmospheric Insight
VOC data analysis is a multi-step pipeline bridging instrument electronics, analytical chemistry, and atmospheric science. Each step introduces decisions that propagate through to the final science product.
Raw TOF Spectra
High-rate (1–10 Hz) time-of-flight spectra stored as HDF5 / binary. Mass axis calibrated to known internal ions (m/z 21, 59, 137). Spectral averaging to desired time resolution.
Peak Fitting
Gaussian/Lorentzian peak fitting to extract signal (cps) per identified m/z. Isobaric separation using exact mass. PTRwid or custom Python pipeline.
Humidity Correction
H₃O⁺ reagent ion suppression by H₂O corrected using measured RH or water cluster signals. Normalised to reagent ion signal (ncps). Essential in tropical deployments.
Calibration
Conversion from ncps to mixing ratio (pptv/ppbv) using compound-specific sensitivity factors from multi-point calibration with NCAR gravimetric standards. Diel and thermal drift corrections applied.
Quality Control
Data flagging: instrument background (zero air), inlet contamination spikes, rain events, local source exclusion. Uncertainty propagation from calibration and signal noise.
Analysis & Modelling
Clean dataset ingested for source apportionment (PMF), box model (CAABA-MECCA) initialisation, or boundary condition constraint for WRF-Chem / EMAC simulations.
Calibration & Measurement Uncertainty
Rigorous calibration and uncertainty quantification underpin the credibility of atmospheric VOC measurements, particularly when data are used to constrain or validate atmospheric models.
Primary Calibration Standards
NCAR (National Center for Atmospheric Research) certified gravimetric gas standards — typically 40-component NMHC mixtures at ~1 ppmv in N₂. Traceable to NIST SRM 2779 for C₁–C₁₂ hydrocarbons. For oxygenated VOCs (acetone, MEK, methanol), custom permeation-tube standards used with known permeation rates (ng/min).
Zero-Air Background Measurements
Instrument background measured every 30–60 minutes by sampling VOC-free zero air (catalytic converter at 450°C or zero-air generator). Background subtracted from ambient signal. Particularly critical for formaldehyde (m/z 31) which has a non-negligible instrument background from internal outgassing.
Intercomparison & Blind Samples
Participation in WMWG (WMO World Meteorological Organization VOC Working Group) intercomparison exercises. Blind canister sample analysis against reference laboratories (NOAA, FZJ-Jülich). PTR-TOF-MS cross-validated against co-deployed GC-MS for key species (isoprene, benzene, toluene) — typical agreement within ±10–15% for well-characterised species.
Propagated Measurement Uncertainty
Total measurement uncertainty (1σ) estimated from quadrature sum of: calibration standard uncertainty (~3–5%), sensitivity precision (~2%), signal counting noise (<1% at ppbv), humidity correction (~3–8% in tropics), and inlet wall loss estimates (~5% for sticky OVOCs). Typical total uncertainty: ±8–15% for most VOC species at the ppbv level.
Detection Limit Estimation
Method Detection Limit (MDL) calculated as 3σ of zero-air background measurement at the relevant time average. For PTR-TOF-MS, MDL of <20 pptv at 1 s for most species, <5 pptv at 1-min averages. GC-MS MDLs depend on pre-concentration volume (typically 1–2 L at STP) — typically 1–5 pptv for C₃–C₁₀ NMHCs.
Ancillary Meteorological Data
All VOC datasets co-registered with meteorological parameters (T, RH, wind speed/direction, radiation, boundary layer height from radiosonde or lidar). Meteorological flags used to exclude local contamination events and identify boundary layer dynamics in concentration variability. Backward trajectories (FLEXPART, HYSPLIT) computed for each campaign period.
Atmospheric Chemistry & Transport Models
Measurements alone cannot answer questions about the broader atmospheric context — why VOC concentrations are elevated at a given site, how they affect regional ozone and aerosol budgets, or what the climate feedbacks are. Models bridge observations to process understanding and predictive capability. My work employs three complementary modelling frameworks, each operating at a different scale and level of physical complexity.
EMAC is a global atmospheric chemistry–climate general circulation model built on the MESSy (Modular Earth Submodel System) framework, developed by a consortium of European institutes (MPI-C Mainz, Jülich, DLR). It couples the ECHAM5 GCM with online chemistry, aerosol microphysics, and tracer transport. EMAC runs as a 3D global model at resolutions typically between T42–T106 (∼2.8°–1.1°), with 31–90 vertical layers extending into the stratosphere and mesosphere.
In my research, EMAC simulations are used to contextualise field measurements — determining the large-scale atmospheric composition at the campaign sites, attributing observed VOC levels to emission source regions via tracer diagnostics (ATTILA Lagrangian trajectory module), and evaluating whether the model correctly reproduces measured isoprene and monoterpene mixing ratios over the Indian subcontinent and ocean.
CAABA-MECCA is an atmospheric chemistry box model developed at the Max Planck Institute for Chemistry, Mainz. MECCA is the chemical mechanism module (part of MESSy) that can be used standalone as a zero-dimensional box model via the CAABA framework. It implements state-of-the-science gas-phase chemistry (MCM-compatible, including isoprene oxidation schemes, halogen chemistry, and tropospheric–stratospheric chemistry) with an arbitrary number of reactions and species.
The box model is invaluable for interpreting field observations: given observed VOC mixing ratios, temperature, RH, and photolysis rates (J-values), CAABA-MECCA simulates the photochemical evolution of the air mass, allowing estimation of unmeasured radical species (OH, HO₂, RO₂), production and loss rates, and chemical lifetimes. Sensitivity studies with different VOC/NOₓ ratios help classify observed ozone chemistry as VOC-limited or NOₓ-limited.
WRF-Chem is a fully coupled, online regional meteorology–chemistry model developed by NCAR, NOAA-ESRL, PNNL and collaborators. Unlike offline transport models, WRF-Chem computes meteorology and chemistry simultaneously (online coupling), enabling two-way feedbacks between aerosols and radiation/cloud microphysics. Regional simulations at 1–50 km resolution are typical, with multiple chemical mechanism options (RADM2, CBMZ, SAPRC, MOZART-4 gas-phase; MOSAIC, VBS aerosol modules).
WRF-Chem is used to downscale from global EMAC fields to regional scales relevant to the Indian subcontinent, enabling interpretation of campaign measurements against the regional emission and transport context. Simulations initialised with MEGAN-v3 biogenic VOC emissions and fire emission inventories (FINN, GFED) are evaluated against PTR-TOF-MS field data. Nested domains allow capturing local topographic and land-surface effects on VOC transport over the complex terrain of the Western Ghats.
MEGAN (Guenther et al.) is the standard algorithm for calculating biogenic VOC emission fluxes from vegetation. It computes isoprene, monoterpene, sesquiterpene, and OVOC emissions as a function of plant functional type (PFT), leaf area index (LAI, from MODIS), temperature, solar radiation, leaf age, CO₂ concentration, and soil moisture. MEGAN v2.1 and v3 are widely implemented within WRF-Chem and EMAC.
Eddy covariance flux measurements from my tower campaigns (Western Ghats, tropical forest sites) are used to evaluate MEGAN's isoprene and monoterpene emission factors for Indian vegetation types — a region with limited in-situ flux validation data. Discrepancies between measured and modelled fluxes provide constraints to update emission factor databases for South/Southeast Asian plant functional types.
FLEXPART (FLEXible PARTicle dispersion model, NILU/Vienna) and HYSPLIT (Hybrid Single-Particle Lagrangian Integrated Trajectory, NOAA ARL) are Lagrangian trajectory and dispersion models driven by ECMWF ERA5 or GFS reanalysis meteorological fields. They compute backward or forward trajectories of air masses, emission sensitivity footprints (source receptor relationships), and dispersion of pollutant plumes.
In my campaigns, FLEXPART backward trajectories are computed for every campaign measurement period to identify the origin of sampled air masses. Source regions for elevated VOC or biomass burning signals are identified using the emission sensitivity maps. This provides the essential meteorological context for interpreting PTR-TOF-MS concentration variability and attributing signals to continental, marine, or biogenic source regions.
The Master Chemical Mechanism (MCM, University of Leeds) is the most explicit gas-phase tropospheric chemical mechanism available, containing ~17,000 reactions and ~6,700 species describing the oxidation of 142 primary VOC precursors in near-molecular detail. It is the reference mechanism for evaluating simpler parameterised mechanisms (RADM2, CBM-IV, SAPRC) used in 3D models.
CAABA-MECCA box model simulations can be run with MCM v3.3 chemistry to provide the most chemically detailed interpretation of field observations. For isoprene specifically, the IUPAC-recommended MCM isoprene oxidation scheme (Jenkin et al., 2015) — including the Leuven Isoprene Mechanism (LIM1) pathway for OH recycling under low-NOₓ conditions — is used to evaluate whether the box model can reproduce observed OH concentrations in tropical forest air masses.
Multi-Scale Modelling Strategy
The modelling framework operates as a nested hierarchy — from process-level box models constrained directly by field observations, through regional NWP-coupled chemistry, up to global climate-chemistry. Measurements at every scale provide evaluation benchmarks.
Model–measurement consistency challenge: A persistent issue is that global models (EMAC, WRF-Chem) underestimate isoprene mixing ratios in tropical regions by factors of 2–5×. This can stem from errors in MEGAN emission factors, model-resolved boundary layer heights, or chemical loss rate parameterisations. PTR-TOF-MS field data from undersampled regions (Arabian Sea, Indian tropical forests) provide critical constraints to diagnose these discrepancies — a central objective of my research programme.
The atmosphere does not distinguish between what we can and cannot measure. Every pptv of unreported isoprene, every undetected monoterpene oxidation product, shapes ozone budgets and aerosol formation in ways our models must account for. The path forward is not choosing between instruments and models — it is building the feedback loop between them.
— On integrating PTR-TOF-MS field campaigns with atmospheric chemistry models