Accurately quantifying infant ingestion and inhalation of indoor dust requires a mechanistic understanding of how dust particles are transported and redistributed within the near-floor microenvironment where infants crawl, play, and explore. To address this, we developed a size-resolved mass balance framework to evaluate dust migration between spatial zones critical to exposure, including floors, indoor air, mouthed objects (e.g., hands, toys), saliva, and the respiratory system. This approach builds on the principle of mass conservation, where net dust accumulation within a spatial zone (e.g., flooring surfaces, indoor air) is determined by the balance between source processes delivering dust and loss processes removing it. For example, indoor air receives dust via resuspension from crawling and walking, while removal occurs through deposition, ventilation, and air filtration [1, 28, 31, 77]. By integrating behavioral observations, environmental measurements, and dust physical characterization, this framework enables size-resolved predictions of dust ingestion and inhalation exposures under real-world conditions.

The model addresses two primary components of dust exposure: ingestion and inhalation. For ingestion, it describes the transfer of size-resolved dust particles from flooring surfaces to mouthed objects, such as infants’ hands and toys, and subsequently to their mouths, where particles may be swallowed. Contact transfer fractions—quantifying the proportion of particles transferred per contact event—depend on dust particle size and morphology, surface properties (e.g., smooth versus carpeted flooring), contact pressure and duration, relative humidity, and films like saliva or skin oils [27, 56]. Controlled experiments, including robotic simulations and dust or soil adherence tests, provide size-specific contact transfer fractions, which integrate with behavioral metrics such as floor contact and mouthing frequencies from video analysis to estimate realistic dust ingestion rates using empirical and mechanistic approaches.

For inhalation exposure, the model accounts for dust resuspension triggered by infant locomotion, which generates mechanical disturbances and near-floor airflow turbulence, detaching settled dust particles. Similar to contact transfer fractions, resuspension fractions—quantifying the proportion of particles released into the air per disturbance—depend on dust particle characteristics, surface properties, and environmental factors [1, 31, 46, 77]. Crawling can release particles smaller than 20 µm into the infant breathing zone, forming dense dust clouds that elevate exposure risks [1,2,3]. Particles <10 µm are concerning as they remain airborne longer and penetrate deep into the respiratory system, while coarser particles >20 µm resettle quickly but dominate ingestion pathways [1, 78]. Chamber-based experiments provide size-specific resuspension data, which, combined with deposition rates, ventilation rates, and infant breathing parameters, enable robust predictions of inhalation exposure to resuspended dust. For example, Wu et al. [1] found that crawling-induced resuspension fractions varied from 10−6 to 10−2 across the 0.5–10 µm size range and among 12 different carpets, underscoring the strong influence of particle size and surface type on near-floor resuspension dynamics.

Infants experience simultaneous ingestion and inhalation of dust during locomotion and play. Traditional exposure models evaluate these pathways independently, overlooking their concurrent nature. As infants crawl, settled floor dust adheres to their hands and transfers to their mouth during mouthing events, while the same movement resuspends dust into the near-floor breathing zone, increasing inhalation exposure. Our unified mass balance framework links contact transfer and resuspension processes, enabling predictions of total dust exposure during dynamic activity. By integrating size-resolved particle processes with behavioral and environmental data, this framework comprehensively evaluates ingestion and inhalation pathways, capturing the complexities of indoor dust migration and redistribution in the home.

Figure 5 illustrates the five spatial zones modeled within this framework: Zone 1 (Surface): Floor (F), Zone 2 (Volume): Indoor Air (A), Zone 3 (Surface): Mouthed Objects (O), Zone 4 (Volume): Mouth–Saliva (S), and Zone 5 (Volume): Respiratory System (R). The framework applies the principle of mass conservation, represented through first-order differential equations, to account for the net accumulation of dust on surfaces (e.g., floors) and within volumes (e.g., indoor air). For surface zones, dust mass is calculated as the product of surface-based dust mass concentrations (m, g/m2 or mg/m2) and the surface area (A, m2). In volume zones, dust mass is expressed as the product of volume-based dust mass concentrations (C, mg/m3) and the corresponding volume (V, m3). Source terms (S, mg/h) represent processes that deliver dust to a given zone, such as resuspension from flooring surfaces into the indoor air zone. Loss terms (L, 1/h) represent removal processes, including deposition, ventilation, and air filtration, which remove dust particles from the indoor air zone [1, 28, 31, 49].

Illustrated spatial zones include the floor, indoor air, mouthed objects, mouth–saliva, and respiratory system. Source (S) and loss (L) processes are represented as arrows delivering dust to a zone (source) or removing dust from a zone (loss). i denotes the particle size dependence of source and loss rates, while subscripts F, A, O, S, and R indicate each spatial zone. Dust mass concentrations (surface-based: m or volume-based: C) are defined for each zone’s area (A) or volume (V). Input parameters for source and loss terms are detailed in Table 1. Note: time dependency (t) of each term is not explicitly shown.

To capture the size-resolved nature of dust transport, accumulation, and removal, dust mass concentrations, source terms, and loss terms are treated as size-dependent variables, denoted with the subscript i. Coupled mass balance equations describe the dynamic movement of dust across spatial zones, with selected terms modeled as functions of time (t). For the floor, indoor air, and mouthed objects, these equations form the basis for predicting size-specific dust transport and exposure processes:

$$\begin}}}\,}}}\,\left(}}}\right)\!\!:}}}\,(}}})\!\!: & _\cdot \frac_\left(t\right)}}=_(t)-_(t)\cdot _\left(t\right)\cdot _\end$$

(1)

$$\begin}}}\,}}}\,\left(}}}\right)\!\!:}}}\;}}}\,(}}})\!\!: & _\cdot \frac_\left(t\right)}}=_(t)-_(t)\cdot _\left(t\right)\cdot _\end$$

(2)

$$\begin}}}\,}}}\,\left(}}}\right)\!\!:}}}\; }}}\,(}}})\!\!: & _\cdot \frac_\left(t\right)}}=_(t)-_(t)\cdot _\left(t\right)\cdot _\end$$

(3)

The subscript (F, A, O) denotes the corresponding spatial zone. Dust mass concentrations are assumed to be uniform within each zone. For example, Zone 2, representing indoor air, is treated as well-mixed within the infant near-floor microenvironment to simplify mass balance calculations and ensure practical application for exposure assessment [1, 2, 30, 31]. While this assumption streamlines the model, each zone can be partitioned into sub-zones to capture spatial variations in dust concentrations, as demonstrated in recent work on infant crawling-induced dust resuspension [1]. The sizes of Zones 1 (floor surfaces) and 2 (indoor air) can be estimated using typical dimensions of living rooms and infant bedrooms. For simplified analyses, certain source and loss rate terms may be treated as approximately constant over time. AF can be defined as the floor area available for resuspension or contact transfer.

Common size-resolved source and loss terms for Zones 1 to 3 are summarized in Table 1. For example, in Zone 1 (floor surfaces), sources include dust track-in, modeled as a net track-in rate (Ti), and deposition of airborne particles, represented using a first-order deposition loss rate coefficient (βi). Loss terms account for dust removal via vacuuming or cleaning, modeled as a vacuuming/cleaning removal rate coefficient (vi); resuspension induced by infant locomotion, quantified using resuspension fractions (rr,i), defined as the fraction of dust particles on the floor surface that is in contact with the infant (e.g., foot, hand) and are released into the air by one stroke of a repetitive contact motion; and contact transfer to mouthed objects, described by contact transfer fractions (ɸFO,i), defined as the fraction of dust particles on the floor surface that is in contact with an object and are transferred to the object by one contact of a repetitive contact motion. Inputs are drawn from chamber studies of dust redistribution (e.g., βi and rr,i) [1, 28, 31], video analysis of infant behavior (e.g., fOS: object-to-mouth contact frequency), and dust physical characterization e.g., mF,i from LDPS analysis of dust samples [24, 25]. Contact transfer and resuspension fractions vary by orders of magnitude with dust particle size, contact dynamics, and environmental conditions, underscoring the complexity of dust redistribution in infant near-floor microenvironments.

For a typical infant locomotion or play event in the home, the model predicts dust mass concentrations on mouthed objects (mO,i) that contact the floor (e.g., hands and toys) and in indoor air (CA,i) following resuspension events. Dust ingestion rates and inhalation rates can then be estimated by incorporating incoming source terms for Zone 4 (saliva, SS,i, mg/h) and Zone 5 (respiratory system, SR,i, mg/h), as outlined in Table 1. For these zones, dust accumulation and loss terms are not considered. Ingestion rates are calculated by summing contributions from various objects, each characterized by size-resolved dust concentrations (mO,i), contact transfer fractions (ɸOS,i, defined as the fraction of dust particles on the object surface that is in contact with the mouth and are transferred to the mouth by one contact of a repetitive contact motion), and contact frequencies (fOS). Inhalation rates incorporate size-specific airborne dust concentrations (CA,i) and age-specific breathing rates (QR) [1, 2]. While size-integrated ingestion rates (SS,i integrated over a specified particle size range) are well-documented [25, 79], data on size-resolved ingestion rates (SS,i) and inhalation rates for resuspended indoor dust (SR,i) remain limited. This mechanistic framework provides an integrated approach to quantify the likelihood of dust ingestion—via hand-to-mouth or object-to-mouth transfer—and dust inhalation through resuspension into the near-floor breathing zone. This integration underscores our transdisciplinary approach to comprehensively evaluating dust dynamics and exposure pathways within infant microenvironments.

To support development of size-resolved mass balance models for infant dust exposure, we are conducting controlled laboratory experiments to quantify dust contact transfer and resuspension fractions under infant-relevant conditions (Fig. 6). These experiments take place in a custom-built inert controlled environmental chamber (Fig. 6A) with a particle-free clean air supply to maintain near-zero background particle levels (Fig. 6B). A robotic infant contact simulator, mounted on a linear stage, replicates crawling motions with programmable contact time, frequency, impulse, and pressure to mimic infant locomotion (Fig. 6C). Flooring surfaces are seeded with indoor dust collected from homes. A suite of aerosol instruments, including an optical particle sizer (OPS), aerodynamic particle sizer (APS), and wideband integrated bioaerosol sensor (WIBS), continuously sample chamber air to measure real-time resuspended dust mass concentrations (CA,i), from which size-resolved resuspension fractions (rr,i) are estimated (Fig. 6D). In parallel, contact transfer fractions (ɸFO,i) are estimated from floors to surrogate skin and toy surfaces. The robotic simulator performs controlled contact events with settled dust, and adhered particles are recovered and characterized using laser diffraction analysis. These measurements provide critical mechanistic inputs for estimating dust ingestion and inhalation rates within the mass balance framework.

A Custom-built inert controlled environmental chamber with electropolished stainless steel surfaces. B Particle-free clean air supply delivered via a zero air generator and HEPA filtration. C Robotic infant contact simulator equipped with a linear stage for horizontal motion and a servomotor for rotational motion to replicate infant crawling dynamics. D Real-time analysis of resuspended dust mass concentrations using a wideband integrated bioaerosol sensor (WIBS), aerodynamic particle sizer (APS), and optical particle sizer (OPS), shown left to right.

To illustrate potential health risks associated with indoor dust exposure, we calculated the average daily ingested dose (ADD, mg/kg/day) of heavy metals using the measured elemental concentrations from Fig. 4. While the size-resolved dust ingestion rates (SS,i) can be predicted through our mass balance framework, for this illustrative example, we used a common size-integrated (SS) literature value for the daily dust ingestion rate (200 mg/day [80]):

$$=\frac_}}\times ^\,$$

(4)

where E is the measured mass fraction of heavy metals and other elements (µg/g), BW is the average infant body weight (here taken as 12 kg), and 10−6 is a unit conversion factor. For non-carcinogenic risk assessment, the hazard quotient (HQ) was calculated by dividing ADD values by element-specific reference dose (RfD) values [66]. Using the elemental concentrations from Fig. 4, the resulting HQ values were all below 1, indicating that acute adverse health effects are unlikely at these dust ingestion rates. Similarly, the lifetime cancer risk (LCR) was estimated as the product of ADD and element-specific cancer slope factors (SF). The calculated LCR values were below 10−4, which is within the range considered acceptable for carcinogenic risk.

This preliminary dose calculation demonstrates the potential utility of combining dust ingestion rates with measured elemental concentrations (Fig. 4) to estimate exposure risks. However, our project aims to advance beyond this simplified example by providing size-resolved (ADDi) and scenario-specific dose estimates. By integrating the mass balance framework with detailed environmental and behavioral data, we will enable more refined assessments that account for individual exposure scenarios and specific indoor environments, providing a comprehensive understanding of dust pollutant risks to infants.