Relation Between Humidity and Size of Exhaled Particles

Abstract

Background:

Aerosol particles are generated in human airways, and leave the body with exhaled air. These particles may carry indicators of various lung conditions. To fully utilize the information provided by endogenously produced exhaled particles, it is important to understand their formation mechanism and physical properties. The scope of this work was to measure number size distributions of exhaled aerosol particles at various surrounding relative humidities (RH) in order to gain some knowledge of the size distribution at the point of particle generation.

Methods:

Number size distributions of exhaled particles were measured at various RHs, using an optical particle counter. Breathing with airway closure was employed.

Results:

A relation between particle volume and RH was fitted to experimental data and used to predict how exhaled droplets behave at RHs not easily accessible by experiments. The diameter of an exhaled particle is reduced by a factor of 0.42 when the RH is changed from 99.5 to 75% at 309 K. Calculations also show that the droplets are concentrated solutions near saturation at 75% RH.

Conclusions:

It is concluded that the particles are supersaturated liquid particles, rather than crystalline solids, in ambient air with RH below 75%. A size distribution related to the aerosol at the moment of formation is given. A successful detailed formation mechanism should be able to accommodate the size distribution predicted at 99.5% RH.

Introduction

It is well established that aerosol particles are generated in human airways as we talk, cough, sneeze, or perform breathing manoeuvres in general. The particles leave the body with exhaled air.^(1–6) Exhaled aerosol may carry pathogens that transmit airborne infectious diseases.^(7,8) In addition, their potential as carrier of indicators of various lung conditions is discussed in the literature.^(3,9,10) However, any information provided by exhaled particles cannot be fully utilized unless their formation mechanisms and physical properties are well understood.

High air velocities in the airways associated with, for example, sneezing and coughing, produces a high number of particles.^(11–13) During such activities, dynamic compression of the airways cause shaking and vibrations of the airway walls, probably producing droplets from the respiratory tract lining fluid (RTLF) of primarily central airways.⁽¹⁴⁾ Another plausible mechanism, associated with slow breathing, is based on the formation of “film droplets” generated by collapsing liquid films. Recent experimental and theoretical work gives strong indirect evidence for such a mechanism in connection with airway closure and subsequent reopening.^(4,5,15,16) Airway closure takes place during deep exhalation when the walls of terminal bronchioles in dependent parts of the lungs meet and the passage becomes blocked.⁽¹⁷⁾ When the bronchioles reopen during the next inhalation, it is proposed that a film of RTLF is spanned across the passage. When this film ruptures, droplets may form.

RTLF constitutes a thin liquid layer that lines the air-facing surface of the lungs. It is distributed continuously throughout the respiratory tract but its chemical composition and physiology changes between the conducting airways and the alveolar regions.⁽¹⁸⁾ In the conducting airways, RTLF is believed to consist of two phases:⁽¹⁹⁾ a periciliary watery sol, which surrounds the epithelial cilia, and an overlying mucus blanket or gel. The gel phase contains bronchial glycoproteins, some serum proteins and protein bound to the bronchial glycoproteins, whereas the sol phase contains any soluble components of bronchial secretion, together with serum proteins.⁽²⁰⁾ In the alveolar regions, RTLF is a watery phase covered by an overlying film of surfactant. The pulmonary surfactant is composed of phospholipids (∼80%), cholesterol (∼10%), proteins (∼10%), and small amounts of carbohydrates. The lipid composition is dominated by phosphatidylcholine (PC), but there are also significant amounts of phosphatidylglycerol.⁽¹⁸⁾ The alveolar RTLF is isotonic, that is, has an osmolality close to 0.287 Os kg⁻¹. It appears that most of this osmolality is provided by NaCl, and there is little evidence that molecular, uncharged osmolytes give a substantial contribution to the total osmolality.⁽²¹⁾ Based on the given information of RTLF, it is suggested that the chemical composition of exhaled particles, which presumably consist of RTLF, depends on exactly where in the airways they are formed.

The equilibrium diameter of an aqueous droplet containing nonvolatile solutes varies with relative humidity (RH).⁽²²⁾ When droplets are exhaled, they usually enter an environment where the RH is lower than in the respiratory tract. Evaporation of water, and therefore a decrease in particle diameter, is to be expected until the increasing concentration of nonvolatile substances has caused the water activity in the droplet to match that of the surrounding air. Based on the properties of pure water droplets, evaporation is fast.⁽²²⁾ For example, in dry air at 313 K, a 1-μm pure water droplet evaporates completely in less than 0.5 msec. Nicas et al.⁽⁸⁾ estimates that expiratory particles quickly decrease to half their diameter, corresponding to an eightfold increase in solute concentration, due to water loss.

The scope of this work was to measure number size distributions of exhaled aerosol particles and to gain some knowledge of the particle size distribution at the point of generation. This will aid the understanding of the underlying formation mechanism. Furthermore, particle size is a crucial factor influencing, for example, life time in air as well as choice and efficiency of collection methods, for example for ensuing analysis. Rather than trying to master the difficulties involved in working very close to saturation, experiments were made at easily controlled humidities and the results extrapolated to the range of interest.

Materials and Methods

System layout

The experimental system was located in a 1.75-m² walk-in climate chamber that may be thermostatted at any temperature between 243 and 323 K.⁽²³⁾ It was desired to observe size distributions of particles close to exhalation temperature and humidity as well as at less humid conditions. The dew point of exhaled air was measured to 307 ± 0.5 K, averaged over a respiratory cycle; thus, the standard chamber temperature was set to 308 K to maintain a margin of safety to avoid condensation in the equipment.

As shown in Figure 1, the subject located at (A), inhaled through a high efficiency particulate air (HEPA) filter (B), and exhaled through a two-way, nonrebreathing valve (C) (Hans Rudolph inc. series 2630). The exhaled air passed through a two-way ball valve (D), set either to send the air back into to the room or to direct it into the chamber. Items (C) and (D) were insulated and kept at 308 K. A 1 liter flexible, diffusion tight bag (E) served as a buffer for the downstream instrumentation. An arrangement of a two-way HEPA filter (F) and water locks (G) allowed excess air to escape when the bag was filled to capacity and filtered air to enter the system and avoid choking the instrument pumps when the bag was empty. Allowing the buffer bag to become empty and at the same time blocking the mouthpiece at (A) served as a leak test since the pressure drop over (G) then influenced the whole setup. Any unfiltered room air leaking into the system would have been detected by the instrument. Located after the buffer bag was a diffusion drier (H) in parallel to a copper tube (I), followed by an oven (J) consisting of a copper tube with a heating element and separate temperature control.

FIG. 1.

Schematic drawing of the experimental setup (not to scale). (A) Location of the subject. (B) High efficiency particulate air (HEPA) filter. (C) Automatic two-way nonrebreathing valve. (D) Manual two-way ball valve. (E) Buffer bag. (F) Two-way HEPA filter. (G) Water locks. (H) Diffusion drier. (I) Copper tube. (J) Oven. (K) Dew point meter. (L) Optical particle counter and sizer. (M) Pump for regulating total flow in setup.

The dew point of the air entering the aerosol instrumentation was measured with a cooled mirror hygrometer (Optidew High Performance Optical Dew-Point Transmitter, Michell Instruments, Ely, UK) (K) placed after the oven.

A Grimm dust monitor model 1.108 (Grimm Aerosol Technik GmbH, Ainring, Germany) (L) provided size and concentration information by means of single-particle light scattering. It was operated in a 6-sec averaging mode and delivered concentration data in 15 size intervals, measuring droplet diameters from 0.3 to greater than 20 μm (Table 1). The device drew a flow of 1.2 L min⁻¹. A pump (M) regulated the total airflow in the setup.

Table 1.

Size Bins of the Grimm Dust Monitor Model 1.108

Limits (μm)

0.3–0.4

0.4–0.5

0.5–0.65

0.65–0.8

0.8–1

1–1.6

1.6–2

2–3

3–4

4–5

5–7.5

7.5–10

10–15

15–20

>20

Measurement modes

The experimental setup could be configured in three different modes (cf. Fig. 1). In the first mode, exhaled air was directed through copper tube (I) and through the oven (J) at chamber temperature 308 K, and particles were measured without interfering with humidity. In the second mode, air and exhaled particles were dried in diffusion drier (H), loaded with either silica gel or P₄O₁₀ as drying agent. Here, particles were expected to shrink as they lose water due to evaporation. In the third mode, the setup remained the same as in the second mode, but with the oven temperature set to 348 or 423 K. The intention was that the higher temperature would remove as much as possible of the remaining water, causing a phase transition and further size reduction of the particles.

With a sample flow of 1.95 L min⁻¹, the residence times in the unheated copper tube, diffusion drier, and oven were approximately 2.7, 5.8, and 3.9 sec, respectively. The residence time in the diffusion drier was sufficient to decrease RH to <10% when using P₄O₁₀, whereas silica gel was not equally efficient. The efficiency of uptake by the drier was related to the process behind the water removal. In the case of P₂O₅, water was removed through an irreversible chemical reaction forming phosphoric acid, while silica gel adsorbed water in a reversible manner. The residence time in the oven was adequate to heat the exhaled air to the set oven temperature.

Losses of particles in the experimental setup through gravitational settling were estimated to <5% for 2-μm particles. Diffusion losses were estimated to <10% for 0.01-μm particles. Comparative measurements showed that the difference in residence time between the diffusion drier and the copper tube did not cause any measurable differences in particle concentration.

Subjects and breathing

The present investigation was not intended to characterize individuals or groups of individuals but rather to study the physical properties of exhaled aerosol. The study was approved by the local ethics committee at the Sahlgrenska Academy at the University of Gothenburg.

Three nonsmoking subjects with no symptoms of respiratory disease participated in the experiments. Basic data on the subjects are presented in Table 2. Spirometry (Spirare 3, ultrasonic flow sensor, Diagnostica, Oslo, Norway) was used to determine forced vital capacity (FVC) and forced expired volume in 1 sec (FEV₁), which are indicators of the ventilatory capacity of the respiratory system. The values of FVC and FEV₁ of the subjects in this study were all within the predicted normal limits.⁽²⁴⁾

Table 2.

Characteristics of Study Subjects

Sub.no.	Sex	Age (years)	Height (cm)	FVC %pred	FEV₁ %pred
1	M	61	183	107	109
2	F	29	169	102	107
3	M	38	203	120	123

The values of FVC (forced vital capacity) and FEV₁ (forced expired volume in 1 sec) are given as percentages of the predicted normal value, which is a function of sex, age, and height.⁽²⁴⁾

Before particle measurements, airways were cleansed from possible residual ambient particles. This involved 3 min of tidal breathing, inhaling filtered room air and exhaling back into the room through the valve (D) (cf. Fig. 1).⁽²⁵⁾ Valve (D) was then turned and the subject continued to exhale into the instrument setup.

Subjects were asked to exhale slowly into the experimental setup until no more air could be expired. This technique results in extensive airway closure, and has been shown to produce a high number of particles compared to tidal breathing.^(5,15) Subjects were wearing a nose clip to prevent unfiltered room air from entering the airways.

A series of measurements (A–D) were conducted (Table 3). Each series consisted of at least four individual experiments: one of each mode as well as additional measurements of one or two modes, as a measure of repeatability. An experiment lasted for 20 min.

Table 3.

Experimental Data

Series	Sub.no.	Exp. no.	Exp. mode	Oven temp.(K)	RH (%)	C_tot (#/liter)	C_0.3–0.4 (#/liter)	C_0.4–0.5 (#/liter)
A	1	1	1	308	72.7	8948	4142	2258
		2	2	308	17.1	7230	3387	1815
		3	3	348	26.3	9446	4377	2430
		4	3	423	28.7	9177	4245	2339
		5	1	308	79.3	11189	5094	2807
B	2	6	1	308	80.5	5649	2687	1556
		7	2	308	19.7	3348	1572	877
		8	3	348	26.1	4120	1939	1003
		9	3	423	29.4	3370	1614	742
		10	1	308	84.8	8967	4123	2231
C	1	11	1	308	80.6	8978	4120	2262
		12	2	308	27.9	5678	2655	1368
		13	1	308	76.5	10265	4760	2594
		14	3	423	34.6	9054	4166	2277
D	3	15	1	308	71.6	2165	1032	545
		16	2	308	5.2	1753	845	431
		17	3	423	6.4	2166	978	530
		18	2	308	6.2	1633	767	411
		19	1	308	75.2	2239	1056	532

The study is divided into four series of measurements, A–D, and includes a total of 19 experiments at three different modes. The table provides information on oven temperature, relative humidity (RH), total particle concentration (C_tot) and particle concentration in the first and the second size interval of the Grimm dust monitor (C_0.3–0.4 and C_0.4–0.5, respectively).

Calculations

Very low particle concentrations and the location of size distributions generated by the breathing pattern employed called for unusual procedures for data evaluation. The size distributions of exhaled aerosol peak close to 0.3 μm,⁽⁵⁾ but light scattering instruments in general lose sensitivity to particles below this size due to the low scattering power of small particles.⁽²⁶⁾ Thus, a larger or smaller fraction of the size distribution may not be measured, depending on particle size, that is, as a function of RH. Average diameters calculated with part of the distribution missing are useless as a quantitative measure of size.

It was assumed that the exhaled particle number concentration does not change between experiments in a series for an individual. Short time variations occur over the 20-min measurement period because the breathing occasionally has to be adjusted to the actual oxygen demand of the subject. This result in periods of higher or lower concentration, compared to the concentration generated by the desired breathing pattern. The size distributions, on the other hand, do not change significantly during these periods. To handle the “uncharacteristic” concentrations, data were filtered to remove the 15% highest and lowest concentrations before further evaluation. The data used for calculations were averages of the remaining concentrations. A representative data set is displayed in Figure 2. The typical difference between an unfiltered and a filtered average was 3%, with a maximum of 7%, indicating that excursions from the characteristic data were not serious, even in the unfiltered data.

FIG. 2.

A typical data set. The figure illustrates the variation in particle concentration throughout Experiment 1. The 15% lowest and highest data points are indicated by the dashed lines.

A Grimm dust monitor 1.108 delivers concentration data in size bins and data were evaluated under the assumption that the distribution function between 0.3 and 0.4 μm (the smallest size bin) was constant across the interval. This is a reasonable assumption because the distribution function is expected to peak between 0.2 and 0.5 μm.⁽⁵⁾ It was further assumed that all particles in the distribution change volume by a constant factor upon a change in RH in the surrounding air. This is expected if all particles have the same composition. The Kelvin effect that increases the vapor pressure over a curved surface has little effect because the particles considered here are all quite large in relation to the size where this effect is important, that is, <0.1 μm.⁽²²⁾ The volume change due to a decrease in RH gives rise to a diameter change at 0.3 μm by a factor A. The factor A may be expressed in terms of the loss of observed particles caused by the distribution shifting over the formal lower sensitivity limit at 0.3 μm and is described by Equation (1): \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} A = \frac {3C_{0.3 - 0.4 , high}} {C_{tot , high} - C_{tot , low} + 3C_{0.3 - 0.4 , high}} \tag {1} \end{align*} \end{document}

where C_tot,high is the total concentration above 0.3 μm at high RH, C_tot,low is the total concentration above 0.3 μm at low RH and C_{0.3–0.4,high} is the concentration in the 0.3–0.4-μm interval at high RH. A high RH always represents a first mode experiment, whereas a low RH may characterize a second or a third mode experiment. Equation (1) was derived by integrating the derived particle distribution function (C_{0.3–0.4,high})/[A(0.4–0.3)] from 0.3A to 0.3 and equating the result to the loss of total number concentration on drying. This equation is only valid for 0.75 < A < 1. When A < 0.75 resulted from Equation (1), the upper bound of the smallest size bin has shifted below the lower bound of that bin, and hence, below the detection limit of the Grimm dust monitor. In this case, C_{0.3–0.4,high} was subtracted from C_tot,high and the formalism in Equation (1) applied for bin 0.4–0.5 μm. The error introduced by the assumption of a constant distribution function between 0.3 and 0.4 μm disappears at A = 0.75 because then, the integral of the distribution function is equal to the measured concentration in the 0.3–0.4 μm bin.

A simple model was set up to describe the volume of a liquid particle as a function of the surrounding RH. The model is based on Raoult's law: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} P_{H2O} = P_{H2O}^0 * \gamma * X_{H_2O} = P_{H2O}^0 * \gamma * \frac {n_{H2O}} {n_{H2O} + n_{solute}}. \tag {2} \end{align*} \end{document}

Here, P⁰_H2O and P_H2O are the water vapor partial pressure above pure water and over a solution having liquid water molar ratio X_H2O, respectively. γ is the activity coefficient for water that is needed to account for the nonideality of the solution, whereas n_H2O and n_solute are the number of moles of water and dissolved species in the solution. The activity factor is difficult to estimate accurately because the solute composition is unknown. However, it appears that NaCl is a major component.⁽²¹⁾ Equation (3) is a fit to data for aqueous NaCl taken from Tang et al.⁽²⁷⁾ and used as an approximation of γ. It should be noted that the relation is valid also for strongly supersaturated solutions. \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \gamma = - 4.427*10^{- 5}* (RH)^2 + 1.443*10^{- 2}*RH \tag{3} \end{align*} \end{document}

To obtain volume V_H2O as a function of P_H2O, the relation V_H2O = (n_H2O * M_H2O)/ρ_H2O was used, where M_H2O is the molar mass and ρ_H2O is the density of liquid water. Further, a volume V_solute is assigned to n_solute and the particle volume V_part is approximated by the sum of the water and solute volumes. It is assumed that the solute occupies an equal volume regardless of being solid or dissolved, resulting in Equation (4). \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} V_{part} = V_{H_2O} + V_{solute} = \frac {P_{H2O} * M_{H2O} * n_{solute}} {\left( P_{H2O}^0 * \gamma - P_{H2O} \right) \rho_{H2O}} + V_{solute} \tag {4} \end{align*} \end{document}

Results

Particle concentrations

The measured particle concentrations of interest are displayed in Table 3. Given a fixed RH, the size distribution was expected to be approximately the same for all subjects,⁽⁵⁾ whereas the total number concentration could vary.⁽¹²⁾ The interindividual variation in particle concentration was evident also in this work, despite the small number of participants. On average, subject no. 1 consistently emitted almost twice as many particles as subject no. 2 and nearly five times as many particles as subject no. 3.

Particle volume change

Equation (1) was applied to calculate the observed resulting diameter ratio when reducing the RH between a mode 1 and a mode 2 or 3 experiment. The results are given in Table 4, together with the corresponding volume ratios. Then Equation (4) was applied to produce calculated volumes and ratios. The absolute value given to the amount of solute n_solute is of no concern for the volume ratios as long as the same value is used for both volumes in a ratio. It is, however, linked to the value of V_solute that in principle is a measure of the volume occupied by n_solute in the liquid. Here it is best seen as a fitting parameter adjusted to produce a “best fit” of Equation (4) to the experimental data. This was done by a least squares procedure, with n_solute kept fixed and finding a value of V_solute that gave the minimum ΣΔ², where Δ is the difference between observed and calculated volume ratio for all mode 1 to mode 2 experiments. The resulting R² value for the nonlinear fit was 0.67. The average deviation between observed and calculated ratios was 3.9% with a maximum of 11%. Equation (4) was used to produce Figure 3, with the best fit combination of n_solute = 1 mol and V_solute = 1100 cm³. The calculated volumes were normalized by the volume at 75% RH.

FIG. 3.

Relative particle volume as a function of relative humidity. The volumes were normalized by the volume at 75%.

Table 4.

Volume Change Calculations

Series	Sub.no.	High RH exp.	Low RH exp.	Low RH mode	RH high (%)	RH low (%)	Observed diameter ratio A	Observed volume ratio	Modeled volume ratio
A	1	1	2	2	72.7	17.1	0.88 ± 0.003	0.68	0.55
		1	3	3	72.7	26.3	1.04 ± 0.020	1.13	0.59
		1	4	3	72.7	28.7	1.02 ± 0.018	1.05	0.61
		5	2	2	79.3	17.1	0.79 ± 0.003	0.50	0.42
		5	3	3	79.3	26.3	0.90 ± 0.004	0.72	0.45
		5	4	3	79.3	28.7	0.88 ± 0.003	0.69	0.46
B	2	6	7	2	80.5	19.7	0.77 ± 0.007	0.46	0.41
		6	8	3	80.5	26.1	0.84 ± 0.001	0.58	0.43
		6	9	3	80.5	29.4	0.78 ± 0.007	0.47	0.44
		10	7	2	84.8	19.7	0.65 ± 0.018	0.27	0.32
		10	8	3	84.8	26.1	0.70 ± 0.015	0.34	0.33
		10	9	3	84.8	29.4	0.65 ± 0.018	0.27	0.34
C	1	11	12	2	80.6	27.9	0.79 ± 0.004	0.49	0.43
		11	14	3	80.6	34.6	1.00 ± 0.016	1.02	0.46
		13	12	2	76.5	27.9	0.76 ± 0.005	0.43	0.52
		13	14	3	76.5	34.6	0.92 ± 0.006	0.78	0.55
D	3	15	16	2	71.6	5.2	0.88 ± 0.016	0.69	0.53
		15	17	3	71.6	6.4	0.98 ± 0.061	0.95	0.54
		15	18	2	71.6	6.2	0.85 ± 0.005	0.62	0.54
		19	16	2	75.2	5.2	0.87 ± 0.010	0.65	0.47
		19	17	3	75.2	6.4	0.96 ± 0.051	0.89	0.47
		19	18	2	75.2	6.2	0.84 ± 0.001	0.59	0.47

A low RH experiment (mode 2 or 3) is compared to a high RH experiment (mode 1) in the same series. The observed diameter ratio, A, is calculated using Equation (1) and the modeled volume ratio is calculated using Equation (4). The error in A is calculated as the error propagation from an estimated 10% error in particle concentration.

Particle size distribution

Figure 4 illustrates the particle size distribution averaged over all first mode experiments. The solid curve in the figure shows the actual measured distribution. Because measurements took place at slightly different RH, all individual size distributions have been adjusted to represent the size distribution expected at 75% RH, using the model described above. The dashed curve is the distribution corrected for the shrinkage caused by the lower RH outside of the body. It represents the original distribution at the moment of formation in the saturated environment inside the human body, that is, at 99.5% RH, but is not corrected for any particle loss in the respiratory system.

FIG. 4.

Particle size distribution averaged over all mode 1 experiments. The solid line represents the particles at 75% RH. The dashed line show the original distribution of the particles, that is at 99.5% RH. The figure also illustrates the minumum and maximum values for individual experiments.

Discussion

The experiments showed that study of size distributions of exhaled particles offers a measurement challenge. This is so because the size distribution is positioned in a way that few, if any, single instruments can measure the entire distribution with adequate size resolution. In addition, low number concentrations in the exhaled air complicate matters. A Scanning Mobility Particle Sizer (SMPS) system, which is the standard instrument for measurements in the 0.01- to 0.45-μm range, observes only charged particles in a small interval of the size distribution at any given time. This makes statistically sound measurements quite time consuming and impractical from the breathing subject's point of view. Attempts were made to use an SMPS system in the present study, but with a typical count of 20–100 particles during a 20-min measurement period, it became evident that the total scanning time was too short to provide reliable size distributions of the exhaled aerosol.

Particle concentrations

In the first mode experiments, RH was measured to between 71 and 85% (Table 3), which is lower than the RH expected in exhaled air at body temperature.⁽²⁸⁾ This was probably caused by air, originating from the warm distal parts of the lung, being temporarily cooled by the colder inner walls of the upper airways and mouth, resulting in loss of water by condensation before expiration.^(29,30) Then, reheating in the experimental equipment reduced the RH and caused particles to lose water. The concentration of solutes had to be considerable to cause the reduction of water activity in the droplets that was apparent for the first mode experiments in Table 3. As an example to put the first mode RHs into context, a saturated NaCl solution has a concentration of 6 M and is in equilibrium with water vapor in the surrounding air at 74.9% RH and 308 K,⁽³¹⁾ that is, the water activity in the particles is 0.749. In addition to NaCl, the expired droplets contain surfactants, for example phosphatidylcholine and proteins.^(3,18,) Although the concentration of solutes is high, no crystallization is expected even though some components may be supersaturated. Aqueous NaCl droplets as large as 20 μm, and concentrations double that at saturation have been observed.⁽²⁷⁾ Small droplets containing mixtures of dissolved salts and molecular compounds but lacking crystallization nuclei often stay supersaturated, metastable liquids at RHs far below the deliquescence point.⁽³²⁾ Thus, all first mode experiments were considered to have the particles as concentrated liquid droplets.

During the second mode experiments, particles lost water due to evaporation when passing the diffusion drier charged with silica gel or P₄O₁₀. The particles shrunk, a larger part of the size distribution fell below the detection limit of the instrument, and consequently, a lower total concentration was registered. No sudden change in size or concentration that would indicate a phase transition could be observed on lowering the RH. Thus, the particles resulting from the second mode experiments were expected to be supersaturated.⁽³²⁾

The third mode experiment in each series involved heating to drive off as much as possible of remaining water from the particles. During passage of the oven the RH decreased to low values, typically 3–4%, before the air was again cooled. Here, the observed total concentrations increased, which was unexpected if only size reduction caused by additional water loss was taken into account. It is speculated that the apparent concentration increases reflect a transformation from supersaturated liquid to a dry amorphous or partly crystalline solid. The optical properties differ between solids and liquids and the refractive index that controls the light-scattering changes significantly. Solids generally scatter more efficiently than do liquid objects of comparable size. It is suggested that, although the physical size of the particles was slightly reduced due to further loss of water, this was more than compensated for by more efficient scattering by particles then containing solids. Thus, the third mode experiments could not easily be used for any quantitative calculations.

Particle volume change

To predict the original size distribution of the droplets by aid of Equation (4), it was necessary to estimate the water activity in the RTLF. The solute concentration in the droplets at the moment of formation should be identical to that of the RTLF. The osmolality of the RTLF reflects its water activity. Values found in the literature range from 0.200 Osmol kg⁻¹ to 0.330 Osmol kg⁻¹.^(33,34) This corresponds to water activities of 0.996 and 0.994, respectively.⁽³⁵⁾ Reference is made to isotonic NaCl solution that has a concentration of 0.155 M (0.9% w/w). The corresponding osmolality is 0.287, giving an approximate water activity of 0.995, that is, the RH in equilibrium with an isotonic solution of any kind, for example RTLF, at a given temperature is 99.5%.⁽³¹⁾

Temperature and absolute water vapor concentration encountered by the droplets, on their journey to the outside of the body, depend on time and location. During inhalation, the air is heated and moisturized by heat and mass transfer from the airway walls to reach body temperature and full saturation in the lower airways, whereas the opposite processes take place during part of the exhalation, at least in the central airways.⁽³⁰⁾ In the present case, this is of little concern because the droplets are generated from a liquid that supports 99.5% RH at the moment of formation. The transport to the mouth may take the droplets through sections with varying RH but the humidity at the start and at the point of measurement are known. Adjustment of droplet diameter to the prevailing RH is fast.⁽²²⁾ Thus, volume and diameter of any particle at the location of formation may be calculated from knowledge of diameter and RH at the point of measurement.

The RH values during the first mode experiments were measured to between 71 and 85%, and 75% was taken as the reference RH. The corresponding diameter at 99.5% RH was a factor 2.4 larger or, alternatively, the particles shrunk by a factor 0.42 traveling from the point of formation to the point of measurement.

Nicas et al.⁽⁸⁾ suggest, based on theoretical considerations and review of some earlier experimental data, that exhaled particles quickly decrease in diameter due to water loss. Their calculations show a size reduction of 0.47 and 0.61 at 30 and 70% RH, respectively. The calculations were based on estimations of the concentration of major components of mucus. The shrinkage is comparable to the present work but relies heavily on the assumed composition of the solute in the particles. The results by Nicas et al.⁽⁸⁾ were based on particles that exit the body at high velocities, and assumed there was a large amount of mucus present in the particles. This would indicate a point of origin of the particles in the cilia-covered airways of the transitional zone, whereas the particles treated here are assumed to be formed in the terminal bronchioles.

Particle size distributions

There are two sources of distortion of the size distribution shown in Figure 4. One comes from the instrument being unable to detect particles smaller than 0.3 μm. This leaves a significant portion of the size distribution not being measured. The other source is due to particle losses during transport from the point of generation to the mouth. Such loss was noted by Johnson and Morawska,⁽⁴⁾ who found a decrease in concentration and shift in size distribution on breath-holding. As expected, an increased residence time of particle-containing air in the lung resulted in general particle deposition and with a preference for loss of larger particles. It was shown that considerable deposition takes place on a time scale of seconds. A distribution corrected for alveolar deposition was presented by these authors. The correction was based on the assumption that gravitational settling occurs in a well-mixed 0.2-mm diameter spherical volume, representing an alveolus. The authors further assumed that the diameter of a droplet has decreased by a factor 0.5 when exposed to ambient RH.⁽⁸⁾ The average distribution presented here, likewise corrected for shrinkage from drying, was not corrected for particle loss through diffusion or sedimentation in the airways because the breathing method employed did not include any control of residence time in the lung. It is also recognized that a model taking into account, for example, location of the formation region and dependence of settling distance on airway direction is desirable to perform a correction for deposition.

Conclusions

The size reduction by a factor of about 0.5 for an exhaled particle, predicted by theoretical considerations in earlier work,⁽⁸⁾ is confirmed by the present experiments. Here, the particles shrink by a factor 0.42 when they leave the body for a surrounding with 75% RH. It is suggested that the particles quickly equilibrate to the surrounding environment, and that evaporation is a rapid process. Exhaled particles will be present as concentrated, possibly supersaturated droplets already at 75–80% RH. There appear to be severe requirements on the behavior of the film droplet mechanism suggested to produce the particles on breathing with airway closure. The mechanism should, with appropriate corrections, be able to reproduce the size distribution observed at 99.5% RH. Further, the average lifetime of the films spanned on airway opening needs to be of the order of a few seconds or less to rupture and generate particles during an inhalation. Another consideration is the size of the terminal bronchioles that restrict the physical dimensions of the films. Clearly, further work is needed to establish the details of the film droplet mechanism as a source of exhaled particles.

Footnotes

Acknowledgments

The authors gratefully acknowledge the participation of the volunteers in this study. The work was supported by the Centre for Environment and Sustainability (GMV) in Gothenburg, Sweden, and by the Swedish Research Council Formas (2009-789).

Author Disclosure Statement

A.-C. Olin and E. Ljungström declare that they are co-applicants for a patent on the use of exhaled particle collection and analysis for diagnosing and monitoring disease.

References

Fairchild

, Stampfer

. Particle concentration in exhaled breath—summary report. Am Ind Hyg Assoc J, 1987; 48:948–949.

Papineni

, Rosenthal

. The size distribution of droplets in the exhaled breath of healthy human subjects. J Aerosol Med, 1996; 10:105–116.

Almstrand

, Ljungström

, Lausmaa

, Bake

, Sjövall

, Olin

. Airway monitoring by collection and mass spectrometric analysis of exhaled particles. Anal Chem, 2009; 81:662–668.

Johnson

, Morawska

. The mechanism of breath aerosol formation. J Aerosol Med Pulm Drug Deliv, 2009; 22:1–9.

Holmgren

, Ljungström

, Almstrand

A-C

, Bake

, Olin

A-C

. Size distribution of exhaled particles in the range from 0.01 to 2.0 μm. J Aerosol Sci, 2010; 41:439–446.

Schwarz

, Biller

, Windt

, Koch

, Hohlfeld

. Characterization of exhaled particles from the healthy human lung—a systematic analysis in relation to pulmonary function variables. J Aerosol Med Pulm Drug Deliv, 2010; 23:371–379.

Riley

. Airborne infection. Am J Med, 1974; 57:466–475.

Nicas

, Nazaroff

, Hubbard

. Toward understanding the risk of secondary airborne infection: Emission of respirable pathogens. J Occup Environ Hyg, 2005; 2:143–154.

Scheideler

, Manke

, Schwulera

, Inacker

, Hammerle

. Detection of nonvolatile macromolecules in breath—a possible diagnostic-tool. Am Rev Respir Dis, 1993; 148:778–784.

10.

Hunt

. Exhaled breath condensate: An evolving tool for noninvasine evaluation of lung disease. J Allergy Clin Immunol, 2002; 110:28–34.

11.

Clarke

, Jones

, Olive

. Resistance to two-phase gas-liquid flow in airways. J Appl Physiol, 1970; 29:464–471.

12.

Edwards

, Man

, Brand

, Katstra

, Sommerer

, Stone

, Nardell

, Scheuch

. Inhaling to mitigate exhaled bioaerosols. Proc Natl Acad Sci USA, 2004; 101:17383–17388.

13.

Chao

CYH

, Wan

, Morawska

, Johnson

, Ristovski

, Hargreaves

, Mengersen

, Corbett

, Li

, Xie

, Katoshevski

. Characterization of expiration air jets and droplet size distributions immediately at the mouth opening. J Aerosol Sci, 2009; 40:122–133.

14.

Moriarty

, Grotberg

. Flow-induced instabilities of a mucus-serous bilayer. J Fluid Mech, 1999; 397:1–22.

15.

Almstrand

, Bake

, Ljungström

, Larsson

, Bredberg

, Mirgorodskaya

, Olin

. Effect of airway opening on production of exhaled particles. J Appl Physiol, 2010; 108:584–588.

16.

Haslbeck

, Schwarz

, Hohlfeld

, Seume

, Koch

. Submicron droplet formation in the human lung. J Aerosol Sci, 2010; 41:429–438.

17.

Dollfuss

, Milic-Emili

, Bates

. Regional ventilation of lung studied with boluses of 133xenon. Respir Physiol, 1967; 2:234–246.

18.

, Bidani

, Herring

. Innate host defense of the lung: Effects of lung-lining fluid pH. Lung, 2004; 182:297–317.

19.

Lucas

, Douglas

. Principles underlying ciliary activity in the respiratory tract. II. A comparison of nasal clearance in man, monkey and other mammals. Arch Otolaryngol Head Neck Surg, 1934; 20:518–541.

20.

Kim

. Lung mucus: a clinician's view. Eur Respir J, 1997; 10:1914–1917.

21.

Tarran

, Grubb

, Gatzy

, Davis

, Boucher

. The relative roles of passive surface forces and active ion transport in the modulation of airway surface liquid volume and composition. J Gen Physiol, 2001; 118:223–236.

22.

Hinds

. Condensation, evaporation. Aerosol Technology— Properties, Behaviour and Measurement of Airborne Particles. John Wiley & Sons, Inc.: New York, 249–273. 1982.

23.

Jonsson

ÅM

, Hallquist

, Ljungström

. The effect of temperature and water on secondary organic aerosol formation from ozonolysis of limonene, delta(3)-carene and alpha-pinene. Atmos Chem Phys, 2008; 8:6541–6549.

24.

Quanjer

, Tammeling

, Cotes

, Pedersen

, Peslin

, Yernault

. Lung-volumes and forced ventilatory flows—report working party standardization of lung-function tests european-community for steel and coal - official statement of the european respiratory society. Eur Respir J, 1993; 6:5–40.

25.

Invernizzi

, Ruprecht

, De Marco

, Paredi

, Boffi

. Residual tobacco smoke: Measurement of its washout time in the lung and of its contribution to environmental tobacco smoke. Tob Control, 2007; 16:29–33.

26.

Finlayson-Pitts

, Pitts

. Chemistry of the Upper and Lower Atmosphere. Academic Press: San Diego, CA, 614. 2000.

27.

Tang

, Munkelwitz

, Wang

. Water activity measurements with single suspended droplets—the NaCl-H₂O and KC1-H₂O systems. J Colloid Interface Sci, 1986; 114:409–415.

28.

Morawska

, Johnson

, Ristovski

, Hargreaves

, Mengersen

, Corbett

, Chao

CYH

, Li

, Katoshevski

. Size distribution and sites of origin of droplets expelled from the human respiratory tract during expiratory activities. J Aerosol Sci, 2009; 40:256–269.

29.

Hanna

, Scherer

. Regional control of local airway heat and vapor losses. J Appl Physiol, 1986; 61:624–632.

30.

Ferron

, Haider

, Kreyling

. Inhalation of salt aerosol particles - I. Estimation of the temperature and relative humidity of the air in the human upper airways. J Aerosol Sci, 1988; 19:343–363.

31.

Weast

. Handbook of Chemistry and Physics. The Chemical Rubber Co.: Cleveland, OH, D213–D214. 1971.

32.

Pavia

, Lampman

, Kritz

, Engel

. Introduction to Organic Laboratory Techniques. Thomson Brooks/Cole: Belmont, CA, 664–665. 2007.

33.

Knowles

, Robinson

, Wood

, Pue

, Mentz

, Wager

, Gatzy

, Boucher

. Ion composition of airway surface liquid of patients with cystic fibrosis as compared with normal and disease-control subjects. J Clin Invest, 1997; 100:2588–2595.

34.

Jayaraman

, Song

, Verkman

. Airway surface liquid osmolality measured using fluorophore-encapsulated liposomes. J Gen Physiol, 2001; 117:423–430.

35.

Kiss

, Hansson

H-C

. Application of osmolality for the determination of water activity and the modelling of cloud formation. Atmos Chem Phys Discuss, 2004; 4:7667–7689.