Influence of Ventilation on Formation and Growth of 1–20 nm Particles via Ozone–Human Chemistry

Ozone reaction with human surfaces is an important source of ultrafine particles indoors. However, 1–20 nm particles generated from ozone–human chemistry, which mark the first step of particle formation and growth, remain understudied. Ventilation and indoor air movement could have important implications for these processes. Therefore, in a controlled-climate chamber, we measured ultrafine particles initiated from ozone–human chemistry and their dependence on the air change rate (ACR, 0.5, 1.5, and 3 h–1) and operation of mixing fans (on and off). Concurrently, we measured volatile organic compounds (VOCs) and explored the correlation between particles and gas-phase products. At 25–30 ppb ozone levels, humans generated 0.2–7.7 × 1012 of 1–3 nm, 0–7.2 × 1012 of 3–10 nm, and 0–1.3 × 1012 of 10–20 nm particles per person per hour depending on the ACR and mixing fan operation. Size-dependent particle growth and formation rates increased with higher ACR. The operation of mixing fans suppressed the particle formation and growth, owing to enhanced surface deposition of the newly formed particles and their precursors. Correlation analyses revealed complex interactions between the particles and VOCs initiated by ozone–human chemistry. The results imply that ventilation and indoor air movement may have a more significant influence on particle dynamics and fate relative to indoor chemistry.


Total number of pages: 14
Total number of figures: 12 Total number of tables: 1

This Supporting Information file includes:
Section S1.Additional experimental details related to measurement of VOCs Section S2.Methodology for estimating particle growth rate Figure S1.Schematic layout of the climate chamber with sampling locations Figure S2.Illustration of experiment procedures Figure S3.Time series of ultrafine particle concentrations and size distributions at 3.0 h -1 air change rate with mixing fans off Figure S4.Time series of ultrafine particle concentrations and size distributions at 3.0 h -1 air change rate with mixing fans off (replicate) Figure S5.Time series of ultrafine particle concentrations and size distributions at 3.0 h -1 air change rate with mixing fans on Figure S6.Time series of ultrafine particle concentrations and size distributions at 3.0 h -1 air change rate with mixing fans on (replicate) Figure S7.Time series of ultrafine particle concentrations and size distributions at 1.5 h -1 air change rate with mixing fans on Figure S8.Time series of ultrafine particle concentrations and size distributions at 1.5 h -1 air change rate with mixing fans on (replicate) Figure S9.Time series of ultrafine particle concentrations and size distributions at 0.5 h -1 air change rate with mixing fans on (replicate) Figure S10.Correlation between quasi-steady-state concentrations of particles and ozone loss and between particle emission rates and ozone removal rates Table S1.Physiological data of participants in the experiments

Section S1. Additional experimental details related to measurement of VOCs
We monitored mixing ratios of indoor VOCs using a Vocus proton transfer reaction time-of-flight mass spectrometer (Vocus PTR-ToF-MS, Tofwerk AG and Aerodyne Research, Inc.) to capture gas-phase products from ozone-human chemistry.Any VOC having proton affinity higher than water can undergo proton transfer reactions with water and be detected at its protonated mass by a mass spectrometer. 1Compared to previous PTR instrumentation, Vocus PTR uses a new reagent ion source and a focusing ion-molecule reactor (FIMR), which significantly improves the sensitivity and detection efficiency. 2 During the experiment period, the ionization source pressure of the Vocus PTR was regulated to 2.0 mbar.Every 4s, a mass spectrum was collected in the range of 11-500 Th.The mass resolution was ~ 10000 at m/Q 500.The Vocus PTR sampled side stream air using 0.65 m ¼' perfluoroalkoxy (PFA) tubing at the flow rate of ~ 100 sccm from the main inlet (1/2" PFA) stream sucking the air either from the chamber exhaust or from the supply air at the flow rate of 12.5 L min -1 by an external pump.Additionally, a filter (material: PTFE) was used to discard particles that could potentially clog the capillary at the entrance of the Vocus.A calibration was performed before, during and after the campaign (in total 4 times) using a gas mixture standard (Apel-Riemer Environmental Inc), including acetaldehyde, methanol, ethanol, acetonitrile, acetone, acrylonitrile, isoprene, DMS, methyl vinyl ketone, methyl ethyl ketone, benzene, m-xylene, alpha-pinene, 1,2,4-trimethylbenzene, octamethylcyclotetrasiloxane (D4), decamethylcyclopentasiloxane (D5), 1,2,4-trichlorobenzene, and beta-caryophyllene.6-MHO and 4-OPA were calibrated with individual standard gas bottles.For data processing, we used Tofware (version 3.2.5;Tofwerk AG and Aerodyne Research, Inc.) in the Igor Pro 7.08 environment (WaveMetrics, OR, USA).In order to identify if a VOC is related to human emissions, only masses having an increase higher than 3 times the standard deviation of that mass during the empty chamber period were further considered for chemical formula assignment.In terms of quantification of VOCs not presented in the gas standard, the mixing ratios were calculated based on the experimentally derived transmission curve using either known rate coefficient of that compound reacting with protonated water in the literature or a fixed rate coefficient at 2.5 × 10 -9 cm 3 molecule -1 s -1 .

Section S2. Methodology for estimating particle growth rate
owever, when the ultrafine particle concentrations are not high and have considerable fluctuations during a formation event, such as in our study, the t 50 method may bring uncertainties in determining the maximum concentration for normalization and the 50% appearance time.It is expected that in a particle formation event, the time-series plot of larger size particles would have a similar shape to the smaller size particles, but with a time lag (indicating particle growth).We compared the results obtained from the t 50 method and the cross-fitting method.In the most obvious particle formation event in mixing-fan-off experiment with 3.0 h -1 air change rate, the results of particle GR within 3-10 nm were similar: 42.8 nm/h from the t 50 method and 41.5 nm/h  All experiments were performed with the same group of six participants.In the experiment with mixing fans off, one participant (No. 6) moved between the table and the sampling station, in order to investigate the potential difference in ultrafine particle levels between the bulk air and the peri-human microenvironment.In sessions with ozone, ozone was injected 10 min after the participants entered the chamber, targeting a steady-state level of 24-30 ppb inside the occupied chamber.The chamber conditions were set the night before each experiment to ensure that they had reached steady state at the beginning of the experiment.The chamber door was closed during the entire experiment.After the volunteers exited the chamber, the door was kept closed for 30 min and the decay of particle concentrations was measured.air change rate with mixing fans on (replicate).A11 nCNC system was unable to measure 1-3 nm particles in this experiment, whereas >3 nm particles were measured by SMPS and the diameter was mobility size.Shaded area in the top chart indicates the time when the chamber was occupied; and the upside-down triangle represents the moment when ozone was injected into the chamber.Fig. S7.Time series of ultrafine particle concentrations and size distributions at 1.5 h -1 air change rate with mixing fans on.1-3 nm particles were measured by A11 nCNC system and the diameter was activation size, whereas >3 nm particles were measured by SMPS and the diameter was mobility size.Missing data of 1-3 nm particles in the morning session was due to a data storage issue of the instrument.Shaded area in the top chart indicates the time when the chamber was occupied; and the upside-down triangle represents the moment when ozone was injected into the chamber.Note that in the top chart, the presented 1-3 nm particle concentration was divided by 20.

Fig. S8
. Time series of ultrafine particle concentrations and size distributions at 1.5 h -1 air change rate with mixing fans on (replicate).A11 nCNC system was unable to measure 1-3 nm particles in this experiment, whereas >3 nm particles were measured by SMPS and the diameter was mobility size.Shaded area in the top chart indicates the time when the chamber was occupied; and the upside-down triangle represents the moment when ozone was injected into the chamber.Fig. S9.Time series of ultrafine particle concentrations and size distributions at 0.5 h -1 air change rate with mixing fans on (replicate).1-3 nm particles were measured by A11 nCNC system and the diameter was activation size, whereas >3 nm particles were measured by SMPS and the diameter was mobility size.Shaded area in the top chart indicates the time when the chamber was occupied; and the upside-down triangle represents the moment when ozone was injected into the chamber.There was a 10-min bathroom break at 12:20.Note that in the top chart, the presented 1-3 nm particle concentration was divided by 20.

Fig
Fig S-I.A nucleation event day (31 Mar.2011) at SMEAR II in Hyytiälä.(c) Normalized concentration in each size bin (maximum concentration in each bin set to 1); the 50% appearance times of the particles are marked with red circles.(d) The GRs determined with the appearance time method GR sr,50 for two separate size classes 1-1.3 nm and 1.3-2 nm.Extracted from Fig. 7 in Lehtipalo et al.3 Fig S-II.An example of estimating particle growth rate using the cross-fitting method in one mixing-fan-off experiment with 3.0 h -1 air change rate.(A) Time-series correlation coefficient between 10 nm and 3 nm particle concentrations with adjustable time lags.The highest coefficient is marked with a dashed line, and the corresponding time lag indicates the delayed formation of 10 nm particles relative to 3 nm; (B) linear fitting between particle sizes and their corresponding time lags, in which the slope represents particle growth rate in the size range of 3-10 nm.

Fig. S1 .
Fig. S1.Schematic layout of the climate chamber with sampling locations for the experiments.The air was supplied via the supply diffuser and exhausted through a single outlet in the ceiling.Six participants were seated at three tables during the experiments.A11 nCNC, SMPS, and VOCUS were the instruments to measure real-time 1-3 nm nanocluster aerosols (NCAs), 3-55 nm ultrafine particles, and volatile organic compounds (VOCs), respectively.

Fig. S2 .
Fig. S2.Illustration of experiment procedures in the scenarios of different air change rate (ACR) and mixing fan status.All experiments were performed with the same group of six participants.In the experiment with mixing fans off, one participant (No. 6) moved between the table and the sampling station, in order to investigate the potential difference in ultrafine particle levels between the bulk air and the peri-human microenvironment.In sessions with ozone, ozone was injected 10 min after the participants entered the chamber, targeting a steady-state level of 24-30 ppb inside the occupied chamber.The chamber conditions were set the night before each experiment to ensure that they had reached steady state at the beginning of the experiment.The chamber door was closed during the entire experiment.After the volunteers exited the chamber, the door was kept closed for 30 min and the decay of particle concentrations was measured.

Fig. S3 .
Fig. S3.Time series of ultrafine particle concentrations and size distributions at 3.0 h -1 air change rate with mixing fans off.1-3 nm particles were measured by A11 nCNC system and the diameter was activation size, whereas >3 nm particles were measured by SMPS and the diameter was mobility size.Shaded area in the top chart indicates the time when the chamber was occupied; and the upside-down triangle represents the moment when ozone was injected into the chamber.

Fig. S4 .
Fig. S4.Time series of ultrafine particle concentrations and size distributions at 3.0 h -1 air change rate with mixing fans off (replicate).1-3 nm particles were measured by A11 nCNC system and

Fig. S6 .
Fig. S6.Time series of ultrafine particle concentrations and size distributions at 3.0 h -1 air change rate with mixing fans on (replicate).A11 nCNC system was unable to measure 1-3 nm particles in this experiment, whereas >3 nm particles were measured by SMPS and the diameter was mobility size.Shaded area in the top chart indicates the time when the chamber was occupied; and the upside-down triangle represents the moment when ozone was injected into the chamber.

Fig. S10 .
Fig. S10.Correlation between quasi-steady-state concentrations of particles and ozone loss (upper) and correlation between particle emission rates and ozone removal rates (ozone loss times air change rate, lower) across all the experiments with mixing fans on.