Measuring particles in a mixed-phase cloud requires more than determining how much liquid water or ice is present. Researchers also need to distinguish the aerosols that remain between cloud particles from the particles incorporated into droplets or ice crystals. The SPIDER phase separation inlet was designed to separate these cloud elements and make them available for simultaneous analysis through distinct sampling channels.
Although developed for atmospheric research rather than industrial process lines, SPIDER demonstrates a useful instrumentation principle: combine staged particle selection, controlled environmental conditions and monitored flows to turn a complex sample into separately measurable streams. Understanding how the system works also highlights the importance of validating each component and the assembled instrument.
Why separate cloud particles and aerosols?
Mixed-phase clouds contain liquid droplets and ice crystals at the same time. Their proportions affect cloud development, while the aerosols that form droplets or promote ice formation provide information about the processes taking place. In cloud terminology, particles that have not activated into droplets or ice are called interstitial aerosols. Droplet and ice crystal residuals are the aerosol particles left behind when the water associated with a cloud particle is removed.
Conventional measurements can estimate cloud water or ice content, but do not necessarily identify the aerosols associated with each phase. A counterflow virtual impactor (CVI), or its pumped laboratory counterpart, the PCVI, separates particles according to their inertia and size. Larger activated cloud particles can pass through a flow-defined barrier, while smaller aerosol particles are rejected. The particle size at which transmission reaches 50% is known as the D50.
How the SPIDER phase separation inlet works
SPIDER combines four stages: an omnidirectional inlet, a large pumped counterflow virtual impactor (L-PCVI), a cooled evaporation chamber and a PCVI. Together, these components direct interstitial aerosol, droplet residuals and ice crystal residuals into separate outlets.
1. Inlet and initial particle selection
The omnidirectional inlet accepts air from different directions while limiting the entry of large particles, including precipitation and blowing snow. Its narrowing opening and internal geometry cause particles that cannot follow the airflow to impact and be removed. This helps protect downstream sampling stages from unwanted large material.
The incoming sample then enters the L-PCVI. Its counterflow creates a stagnation plane that blocks smaller particles while allowing sufficiently large cloud droplets and ice crystals to pass. In the SPIDER configuration, the L-PCVI’s approximate D50 is 20 micrometres. Its rejected flow provides the interstitial aerosol channel; particles transmitted onward are treated as activated cloud elements.
2. Separating droplets from ice
Particles that pass the L-PCVI enter a chamber maintained at -16 °C. The chamber uses the Wegener-Bergeron-Findeisen process: under suitable temperature and humidity conditions, ice crystals grow at the expense of liquid droplets because ice has a lower saturation vapour pressure. In the instrument, the resulting difference in particle size allows droplets to evaporate while ice crystals remain.
A second PCVI separates the remaining larger ice particles from the smaller particles left by evaporated droplets. Its approximate D50 is 2 micrometres. Ice crystals pass into the PCVI sample flow, while the smaller droplet residuals are removed through its pump flow. The system therefore provides three outlets: the L-PCVI pump flow for interstitial aerosol, the PCVI pump flow for droplet residuals and the PCVI sample flow for ice crystal residuals.
Flow control and component verification
Because particle selection depends on the balance between input, add, pump and sample flows, operating conditions are central to performance. The effective counterflow in a PCVI is determined by the relationship between add flow and sample flow. Changing the ratio of add flow to input flow changes the inertial barrier and therefore the particle cut size. Operators must establish and verify flows carefully rather than treat the inlet as a passive sampling tube.
The SPIDER startup procedure reflects that requirement. Communication with temperature sensors, mass flow controllers and measurement instruments is established first. Vacuum and pressurized air are then supplied, and the PCVI and L-PCVI flows are brought online in sequence, with add flow set before pump flow. Pressure measurements and achieved flow rates are checked for signs of blockage or other flow problems. The cooled chamber is then brought to operating temperature and flushed with dry, filtered air.
Laboratory tests examined the L-PCVI, evaporation chamber and PCVI individually. Tests with generated particles supported the L-PCVI’s ability to reject particles below its cut size and transmit larger droplets. Modelling and experiments assessed droplet evaporation and the passage of ice crystals through the chamber and PCVI. The tests also considered whether flow-induced forces could disrupt droplets; the calculated Weber numbers for the L-PCVI remained well below the stated breakup threshold.
Field deployment: performance and limitations
SPIDER was deployed for three weeks at Storm Peak Laboratory in Colorado, with sampling conducted during both clear and mixed-phase cloudy periods. Measurements showed elevated ice-residual counts during cloudy conditions and changes in the ice residual size distribution that were consistent with cloud particle observations. The campaign demonstrated that the assembled system could operate at a remote field site and separate distinct sample streams.
Field interpretation still requires care. Ice-residual counts were also observed during some clear periods. The researchers considered possible contributions from windblown snow and the PCVI’s particle concentration effect; during the study, blowing snow was not associated with higher ice-residual concentrations. The findings do not eliminate the need to investigate performance at higher wind speeds or under other environmental conditions.
The work also identified areas for further development. More complete characterization of the L-PCVI would help define its cut size across operating conditions, while additional evaporation tests could refine the limits on droplet sizes that can be fully evaporated. Higher time and size resolution for the droplet channel, and an extended sizing range for the ice channel, were identified as priorities for future measurements.
Practical implications for measurement system design
SPIDER’s value lies in treating sample separation as a coordinated system rather than a single instrument function. Its performance depends on flow settings, particle size, chamber temperature and humidity, and the behaviour of each stage. Component-level verification is therefore an important complement to testing the complete assembly. Separate outlets also make it possible to connect different measurement instruments to different particle populations, including instruments for particle sizing or chemical analysis.
For technical teams designing complex sampling systems, the broader lesson is to define what each stage must reject or transmit, verify operating conditions at each transition, and account for how environmental conditions affect the sample. SPIDER demonstrates a proof of principle for simultaneously sampling interstitial aerosols, droplet residuals and ice crystal residuals. Further characterization and higher-resolution measurements remain necessary to strengthen its use in studies of cloud particle composition and formation.