Predicting Airborne Volatile Organic Compound Transport in Highly Sensitive In Vitro Processes and Biopharmaceutical Manufacturing Facilities with Kinetic Models

Abstract

Introduction:

Volatile organic compounds (VOCs) are ubiquitous in indoor air spaces, including the most sensitive medical environments. In assisted reproductive technology facilities, VOCs are known to greatly decrease embryo implantation rates and may impact embryo growth rate as well as miscarriage rates. There is reason to believe that the observed impacts of VOCs translate into parallel industries as well, such as the emerging cell and gene therapy industry. Previous modeling efforts on VOC transport into cell culture media components have provided estimates on the expected equilibrium concentrations from airborne VOC exposure. However, a fundamental knowledge gap remains regarding the rate of VOC partitioning into cell cultures.

Materials and Methods:

In this work, we present an enhanced modeling approach to quantify the partitioning kinetics of selected VOCs in cell cultures consisting of a multiphase system with an oil overlay and a water-based culture media, as well as a water-based media without overlay.

Results:

Preliminary results from eight prevalent and cytotoxic VOCs indicate rapid equilibration into the system with equilibrium achieved within a timescale of seconds to minutes.

Discussion:

These results suggest that practitioners must take steps to maintain VOC-free air for the protection of cell cultures to avoid adverse outcomes, as open-air processes may be compromised even during brief exposure to VOCs.

Conclusion:

VOCs have been modeled to rapidly equilibrate into cell cultrue systems, potentially disrupting normal cellular development.

Keywords

IVF kinetics life sciences modeling VOCs

Introduction: Volatile Organic Compounds in Sensitive Health Care Settings

Indoor air quality (IAQ) is an important factor in health care efficacy, and poor IAQ has been widely reported to adversely affect patient outcomes.^1–3 Health care settings are particularly vulnerable to deteriorating IAQ due to continuous 24-hour operations, which allow little to no downtime for air quality recovery.¹ Volatile organic compounds (VOCs) receive particular attention, as even in the most sensitively regulated medical environments, VOCs are emitted from internal sources, such as containerized gas storage, alcohol-based sanitizers, plasticware, and vinyl flooring. External infiltration of VOCs can also occur because of a lack of adequate air locking, gaps in ceiling lighting, and lack of a dedicated air handling unit.^1–3

A meta-analysis of available literature concerning IAQ in health care shows frequent noncompliance with World Health Organization guidelines for IAQ.¹ While much of the data pertains to humidity, CO₂, particulate matter, and microbiological load, “indoor intrinsic emissions,” or VOCs, are also identified as significant contributors to IAQ.¹ One study has profiled common mixtures of VOCs present in a hospital setting, finding 40 distinct VOCs across 6 sampling sites.³ While concentrations of each individual VOC were lower than those known to be harmful to humans, the complex mixture effects were not known. Modeling is needed to better understand the complex mixtures of VOCs present in sensitive health care environments and their impacts.³ Specifically, an investigation of the fate and transport of common VOCs in the context of health care is warranted.

An emerging area for health care air quality research is in assisted reproductive technology (ART), specifically in vitro fertilization (IVF). IVF processes are particularly sensitive to air exposure within the incubator, which may harbor VOCs introduced via containerized gas injection. Furthermore, as the incubator is opened, gas exchange with the ambient room air will occur. While modern ART facilities use a layer of oil to overlay the water-based culture media to insulate and buffer against temperature and pH changes, the efficacy of this strategy for protecting the culture against airborne contaminants is not well established.^2,4 Studies have identified trends between IVF fertilization rates and ambient air quality, noting that both brief and sustained VOC release events significantly impact successful implantation rates.^5–6 As a result, the installation of a “cleanroom” environment is touted as instrumental in maintaining high clinical pregnancy rates.⁶ The Cairo Consensus, an expert report compiled at the 2017 Upper Egypt Assisted Reproduction Symposium, outlines the best practices for upkeeping air quality in an IVF laboratory. To date, it is one of the most comprehensive pieces supporting maintenance of an optimal culture environment for embryogenesis. Overall, the Cairo Consensus recommends maintaining total VOC levels <500 μg/m³ (400–800 ppb) and aldehydes <5 μg/m³.² It is crucial to note that while these recommendations have been made, there is currently little data available to qualify whether these limits are stringent enough to maintain successful culture across the board.

Applications in the life sciences industry

During the past decade, the prevalence of gene therapy clinics worldwide has increased by more than 40%, alluding to the rapid acceleration of research and development in this space in recent years.⁷ Gene therapy medications have proliferated significantly since, taking various forms dependent upon application. For instance, following formal recognition in 2013, chimeric antigen receptor T cell therapy has shown great promise in treating patients with leukemia and lymphoma.⁷ Additionally, Clustered Regularly Interspaced Short Palindromic Repeats technology has also recently been applied in cancer treatment, with the first successful treatment of lung cancer using genome-edited cells occurring in October of 2016.⁷ To date, there are over 3900 gene therapy trials, up from 2600 in an 2017 review conducted by the same authors.⁷

These developments provide great promise in advancing personalized medicine and revolutionizing the treatment of diseases related to gene expression, including cancers and immunodeficiencies.^7,8 However, with such rapid advances being made, there is a need to highlight the effects of VOC exposure within good manufacturing practices (GMPs) to maintain product safety and ensure high-quality treatments across the board. Several studies have outlined the possible epigenetic effects of VOCs on cells in batch experiments, suggesting that the quality and viability of gene medicine products could be impacted by poor air quality.^9–11 Immense, multivariate challenges arise with manufacturing high-quality cellular therapy medicines, including difficulties in maintaining stability, purity, and safety of gene therapy products.⁸ Better understanding of the environmental partitioning of VOCs into cell cultures is needed to apply methods for improving IVF outcomes to the life sciences industry.

Cellular and subcellular impacts of VOC exposure

Many studies attempted to describe the impacts of VOCs on cellular and embryonic health through animal and human studies.^12–14 Attention has also been directed toward VOC effects on individual components of the cell (subcellular), including proteins, DNA, and RNA.¹⁴ One mouse model study found that exposure to VOC concentrations 30–100 times the ambient air standards dysregulated 69 microRNA types in the lung, resulting in chronic inflammation, improper gene expression, and elevated cancer risk.¹⁵ As limited research exists regarding epigenetic effects of VOCs on cultured cells, modeling of VOC partitioning into cells can provide key insight in cases regarding genetic abnormalities that hinder optimal embryo or cell development.^9,15 Further research on describing the infiltration of VOCs from the air phase into cell cultures will greatly enhance our understanding of the risks to cells in vitro and help in developing approaches to mitigate them.

When considering IAQ, it is prudent to consider the biomagnification of VOCs that have partitioned into culture media, as well as mixture effects, as VOCs are rarely found in complete isolation but instead often as co-contaminants alongside other species. Despite earlier studies suggesting that cytotoxicity may be additive, there is a growing body of research suggesting that the concentration additive method does not adequately describe VOC toxicity, as synergistic (i.e., causing more harm to the affected cell) and antagonistic (i.e., causing less harm to the affected cell) effects may occur,¹⁶ and accounting for these effects is important in enhancing toxicology models.^14,17,18

Research objectives

A knowledge gap exists regarding both the mechanisms by which VOCs partition into cell cultures from the air phase and the resulting effects on cells that lead to diminished clinical outcomes. Previous modeling by the authors has sought to quantify equilibrium partitioning of VOCs into cellular in vitro processes.¹⁹ However, the timescale by which VOCs partition into culture media or oil overlays has not been previously defined. For example, when sudden spikes of VOCs or unexpected exposure to airborne VOCs occurs—over the course of seconds, minutes, or hours—practitioners currently do not have a way to define appropriate responses. Therefore, it is necessary to consider VOC partitioning kinetics to provide insight of the timescale by which airborne VOCs enter in vitro processes and therefore improve upon GMP standards in the life sciences industry. This research aims to provide a model to describe the kinetics of VOC partitioning into multiphase systems that occur with common in vitro processes.

Methods

Equilibrium partitioning modeling

Modeling VOC partitioning in cell cultures can provide insight into better quality control methods and better patient outcomes in the ART field. A mathematical model has previously been proposed using known principles of chemical equilibrium partitioning to describe the partitioning of VOCs into cell cultures.¹⁹ In this work, it was shown that thermodynamic equilibrium partition coefficients can predict the concentration of contaminant that is able to reach the cell by passing from the ambient air through a layer of water-based cell culture media (i.e., a biphasic system), which may also contain an oil overlay (i.e., a triphasic system), and into or onto the cell.¹⁹ The biphasic system is still used in older ART facilities as well as in the cell and gene therapy industry, while the triphasic system is implemented in most ART facilities.¹⁹

In biphasic systems, direct air/water interfacial partitioning is governed by Henry’s law, which states that the aqueous concentration of a compound in solution is proportional to the partial pressure of the volatile compound in the headspace above solution at equilibrium.²⁰ The VOC concentration in the air phase is assumed to be constant. Henry’s law is expressed in Equation 1:

K_{H i} = \frac{p_{i}}{C_{i w}}

(1)

where

K_{H i}

is the Henry’s law coefficient,

p_{i}

is the partial pressure of the compound “i” at equilibrium and

C_{i w}

is the concentration of “i” in water.²⁰ This may be converted into a dimensionless constant by dividing by the temperature, in Kelvin, and ideal gas law constant “R,”

0.0821 \frac{L * atm}{mol * K}

, given that both concentrations are in units of moles per liter and the pressure is given in atmospheres.

The first partitioning step in a triphasic system is between air and mineral oil, where oil can be approximated as a nonpolar organic solvent, typically n-hexadecane.²¹ Mathematically, the partition coefficient is defined in Equation 2:

K_{ial} = \frac{C_{i a}}{C_{i l}}

(2)

where

K_{ial}

is the air–oil partition coefficient,

C_{i a}

is the molar concentration of species “i” present in air, and

C_{i l}

is molar concentration present in liquid hexadecane (oil) at equilibrium.²¹ Note that

K_{ial}

is analogous with the air–oil partition coefficient,

K_{a o}

as seen in other literature,¹⁹ but the hexadecane phase is labeled here as “l” to avoid confusion with octanol. Similarly, partition coefficients between hexadecane and water are utilized to demonstrate partitioning from the oil into the water-based culture media, which is approximated as water for this purpose.²¹ Air–hexadecane and hexadecane–water partitioning trends have been published by Abraham et al. for hundreds of species.²² Once the VOC has entered the culture system through the first boundary crossing, it becomes available to interact with the cell.

Modeling VOC diffusion kinetics in multiphase systems

Accounting for temporal changes in concentrations requires modeling of VOC transport through a single phase and across phase boundaries. Transport through a single phase is considered via molecular diffusion, and this is represented by Fick’s law in Equation 3^20,23,24:

F_{x} = - D \frac{d C_{i}}{d x}

(3)

Here, F_x refers to the diffusive flux (mass/area/time) and D is the diffusivity coefficient, typically given in units of length²/time (cm²/s). Diffusivity, in this sense, is assumed to be within one single phase, and diffusivities can differ by several orders of magnitude depending on the medium in which a species is diffusing.²⁰ As diffusivities may not be available for every VOC, models exist to estimate diffusivities of different chemicals based on molar mass. Equations 4 and 5 give relationships between diffusivity and molar mass in air (g) and water (w), respectively.²⁰

D_{i g} = (0.83 \pm 0.09) {M_{i}}^{- (0.51 \pm 0.02)}

(4)

D_{i w} = (7.0 \pm 0.08) \times 1 0^{- 5} {M_{i}}^{- (0.45 \pm 0.02)}

(5)

where M_i is the molar mass of a compound in g/mol and D values for air and water are expressed as cm²/s.²⁰

Estimating the diffusion coefficient of a volatile compound in organic liquid or biological material may be accomplished with a model given by Castillo et al., which uses molar volumes of the VOC and the solvent as key parameters for diffusivity, as opposed to molecular weight. Here, this model is given as Equation 6²⁵:

D_{i l} = 7.21 * 1 0^{- 12} T^{1.63} {μ_{L}}^{- 0.702} {V_{m, VOC}}^{- 1.623} {V_{m, L}}^{0.510}

(6)

where T is temperature in Kelvin, μ_L is solvent viscosity in mPa·s, and V_m is molar volume in cm³/mol.²⁵ This adaptation of the Siddiqi and Lucas model, which has been calibrated to minimize error by Castillo et al., was chosen for this modeling effort due to the parameters being easily accessible, and an average relative error of just 4.2% was reported between these model results and experimental diffusivities.²⁵

Transfer across a boundary, for example, the air/water interface, is modeled using boundary mass transfer models. A popular model, developed by Lewis and Whitman in 1924, is the two-film theory of mass transfer in gas–liquid systems.²⁶ This model gives a mass transfer coefficient in terms of the diffusivity of a species within a thin film at the gas–liquid interface. The schematic shows the concentration gradients on either side of the interface within the film, indicative of the model’s assumption that molecular diffusion occurs within these boundary layers. The concentration jump at the interface is described via equilibrium partitioning.²⁰ This is depicted in Figure 1.

FIG. 1.

Graphical depiction of the two-film theory, representing the interface between the gas phase and the liquid phase.

Equation 7 gives the two-film mass transfer coefficient on the gas side, and Equation 8 gives the same relation for the liquid side:

k_{g} = \frac{D_{i g}}{δ_{g}}

(7)

k_{L} = \frac{D_{i L}}{δ_{L}}

(8)

where k_g is gas side two-film mass transfer coefficient in units of length/time, D_ig is diffusivity of a species in the gas film, and

δ_{g}

is gas side film thickness expressed in units of length, with the same convention used for the liquid side film equation.^20,26 To define the overall mass transfer coefficient, k_gl, the relationship given in Equation 9 is useful²⁰:

\frac{1}{k_{g l}} = \frac{1}{k_{g}} + \frac{K_{igl}}{k_{l}}

(9)

where K_igl is a generalized partition coefficient describing equilibrium of compound “i” between any gas phase “g” and any liquid phase “l.”

The two-film theory has been adapted to describe mass transfer within gas–liquid–liquid systems, that is, gas interfacing with water and oil, suggesting that it may be useful in describing mass transfer in multiphase systems.²⁷ It is important to consider that the addition of an oil phase may increase the magnitude of mass transfer into the system if the VOC has a high affinity for oil.²⁷

Film thickness is a critical parameter in these calculations and historically has been difficult to measure accurately.²⁸ Computational fluid dynamics models have been used to describe boundary layer thicknesses for a variety of systems.²⁸ For example, Mahmoodlu et al. obtained air–water mass transfer coefficients for TCE and toluene in continuously stirred and unstirred, or batch reactor, air–water systems.²³ As the culture system can be modeled as a batch reactor, the unstirred system was of greater interest for this study. Mahmoodlu’s system used was an unstirred batch reactor consisting of a pool of water in a column, with air headspace above the pool, and a pure VOC in an open glass vial suspended above the pool. Air phase and liquid phase mass transfer coefficients from Mahmoodlu et al. fall in the same order of magnitude as results from Liu et al. that reported mass transfer coefficient computation for small liquid pools subjected to formaldehyde in indoor air. Experiments performed by Mahmoodlu et al. and Liu et al. are of a similar scale to the small, static, cell culture system, as opposed to other environmental estimates looking at mass transfer over large natural bodies of water, where wind and wave motion influence the boundary layer.^20,29

For this study, we estimated order-of-magnitude mass transfer coefficients using data from similar experiments. Equations 7 and 8 were applied to Mahmoodlu et al.’s experimental data to obtain average gas side film thickness of 8.395 × 10⁻³ m and liquid side film thickness of 6.379 × 10⁻⁵ m for the batch system, which could then be used to estimate mass transfer coefficients for other VOCs based on diffusivities.^23,29 We assume that mass transfer coefficients are in a similar order of magnitude due to the similarity of the oil and water diffusivities.

Mahmoodlu’s study also focused on determining the time of equilibration through experiments and mass balance models. The findings of this study indicate that headspace VOC concentration is equilibrated quickly, on the order of about 50 minutes; however, the dissolution kinetics of the VOCs across the air–water interface in undisturbed water were found to be slower, on the order of about 1000 minutes. This discrepancy indicates that molecular diffusion is the governing process in mass transfer.²³

Studies have attempted to better quantify the time of equilibration by developing mass balances that include diffusion models. Mahmoodlu et al. present mass balance equations for VOC dissolution in an air–water system. The mass balances for the VOC concentrations in air and water are given as Equations 10 and 11²³:

V^{a} \frac{d C^{a}}{d t} = A^{a p} k^{e} (C_{\max}^{a} - C^{a}) - A^{a w} k^{d} (\frac{C^{a}}{H_{c}} - C^{w})

(10)

V^{w} \frac{d C^{w}}{d t} = A^{a w} k^{d} (\frac{C^{a}}{H_{c}} - C^{w})

(11)

where V^a and V^w are the volumes of air and water, respectively, C^a and C^w are VOC concentrations in air and water, t is time, A^ap is the interfacial area between the VOC reservoir and air headspace, A^aw is air–water interfacial area, k^e is rate coefficient for evaporation of VOC into the headspace, k^d is the dissolution coefficient for VOC in water, and H_c is the Henry’s law constant for the VOC. These differential equations can be solved mathematically to obtain the concentrations of VOCs in each media at a specific time.

Kinetic model development

For the purposes of this model, it is assumed that the air phase concentration remains constant as the sinking capabilities of small petri dishes are not expected to be significant, and a relatively low level of compound is expected to partition out of the air. Furthermore, air is often continuously supplied and replenished in an incubator or clean room at a rate sufficient to provide consistent air phase compounds. In essence, the air is assumed to be infinitely available to the culture environment. Therefore, this model describes an open system. The small scale of the culture system with respect to the air phase reinforces this assumption, as the modeling describes the entry of volatile compounds from a nearly infinite air supply into microliters of uncovered media or milliliters of oil overlay. In the case that it is assumed that initial concentrations of “i” in each medium are 0, the first term drops from each equation entirely. At t = 0, there is assumed to be no real exchange velocity, in that the expression for exchange velocity is time dependent. Therefore, it may be assumed that no partitioning has occurred at t = 0, and at this time, every concentration exists at its initial state.

In this system, mass transfer of VOCs between phases is described by the two-film theory. The boundary layer that defines phase separation in two-film theory is based on the premise that each side is at steady state, that is, well mixed. The system described is a batch reactor in that the culture components are not stirred. Because the air layer is assumed to be well mixed and the other culture system components are static, with the assumption that VOC concentration is constant throughout the liquid phase at equilibrium, the only boundary for a VOC to cross that is described by two-film theory is the air–oil boundary or air–water boundary in cases where oil is not present.

For this study, two key mass balances were developed and solved for VOC concentration as a function of time. They describe the mass transfer of a given VOC species “i” between air and water, in the case of the biphasic system, or air and oil overlay, in the case of the triphasic system, in relation to exposure time. In this work, only the mass transport of VOC into the oil phase for a triphasic system, or the water-based culture media phase for a biphasic system, was considered, that is, only the top-level partitioning step in what would be a series of sequential partitioning steps ending with the VOC reaching the biological matter in culture. This initial partitioning step is of greatest interest because the equilibration of VOC into the dish, whether overlaid with oil or not, introduces the contaminant to the cell culture system, allowing for inevitable VOC diffusion through the culture media and interactions with the cells themselves.¹⁹ At a given temperature and pressure, which we assume would be maintained constantly in an ART or biopharmaceutical manufacturing facility, chemical equilibrium would be maintained, thermodynamically fixing the VOC into the system. Equation 12 describes mass transfer between air and water in a biphasic system.

\frac{d C_{i w}}{d t} = \frac{k^{a w} A^{a w} (\frac{p_{i} m a x}{R T K_{H i}} - C_{i w})}{V_{w}}

(12)

where

\frac{{d C}_{i w}}{d t}

is the rate of concentration change of VOC “i” in water (mol/m³),

C_{i w}

is water phase concentration of “i,” which is assumed to be constant throughout the water phase,

k^{a w}

is the two-film theory mass transfer coefficient between air and water for “i” (m/s),

A_{a w}

is the interfacial surface area between air and water (m²),

t

is time (s),

t_{0}

is the initial time (s),

V_{w}

is the volume of water (m³),

p_{i} m a x

is the constant partial pressure of “i” in air phase (atm),

R

is the ideal gas law constant,

8.21 \times 1 0^{- 5} \frac{m^{3} atm}{mol K}

T

is the temperature (K), and

K_{H i}

is the Henry’s law constant for species “i.”

The solution of Equation 12 with initial condition $C_{i w} (t = 0) = C_{i w 0}$ for all t > t₀ is presented in Equation 13:

C_{i w} (t) = C_{i w 0} e^{\frac{- k^{a w} A_{a w}}{V_{w}} (t - t_{0})} + \frac{p_{i} m a x}{R T K_{H i}} (1 - e^{\frac{- k^{a w} A_{a w}}{V_{w}} (t - t_{0})})

(13)

where

C_{i w} (t)

is the time-dependent water phase concentration of VOC “i” (mol/m³) and

C_{i w 0}

is the initial concentration of “i” in water phase (mol/m³).

The mass balance for a triphasic system is modified to describe mass transfer across the air–oil interface, rather than the air–water interface. The same assumptions made when developing the biphasic mass balance hold true here. This is defined in Equation 14:

\frac{d C_{i l}}{d t} = \frac{k^{a l} A_{a l} (\frac{p_{i} m a x}{R T K_{ial}} - C_{i l})}{V_{l}}

(14)

where

\frac{{d C}_{i l}}{d t}

is the rate of concentration change of VOC “i” in oil (mol/m³),

C_{i l}

is oil phase concentration of “i,”

k^{a l}

is the two-film theory mass transfer coefficient between air and oil for “i” (m/s),

A_{a l}

is the interfacial surface area between air and oil (m²),

V_{l}

is the volume of oil (m³), and

K_{ial}

is the equilibrium partition coefficient between air and oil for VOC “i.”

Solution of Equation 14 with initial condition $C_{i l} (t = 0) = C_{i l 0}$ for all t > t₀ gives the time-dependent concentration in the system. Equation 15 gives thetime-dependent concentration in oil.

C_{i l} (t) = C_{i l 0} e^{\frac{- k^{a l} A_{a l}}{V_{l}} (t - t_{0})} + \frac{p_{i} m a x}{R T K_{ial}} (1 - e^{\frac{- k^{a l} A_{a l}}{V_{l}} (t - t_{0})})

(15)

where

C_{i l} (t)

is the time-dependent oil phase concentration of VOC “i” (mol/m³) and

C_{i l 0}

is the initial concentration of “i” in oil phase (mol/m³).

Figure 2 depicts the described model graphically, showing all transport processes.

FIG. 2.

VOC diffusion across boundaries in IVF culture systems. IVF, in vitro fertilization; VOC, volatile organic compound.

Model parameters

The model described has many parameters, including some that are system specific and some that are relevant to the VOC of interest, and it is warranted to discuss the use of specific constant values. In mass transfer, surface area is an important factor in describing the flux of the VOC between phases. Assuming 10 mL of oil overlay is used (0.00001 m³), and assuming this oil forms a flat layer over the culture media and embryo, the dimensions of a typical culture dish can provide the surface area. If a cylindrical dish has a diameter of 60 mm, top-level surface area ( $A = π r^{2}$ ) of the dish is equal to 2.83 × 10⁻³ m². Using a volume of 50 μL (5 × 10⁻⁸ m³) culture media, a similar calculation was used to derive the approximate media surface area in the case that oil is not used. If a cylindrical dish has several wells of approximately 5 mm in diameter, top-level surface area ( $A = π r^{2}$ ) of each well is equal to 1.96 × 10⁻⁵ m².

The model was set to 37°C, or 310.15 K, which is a typical incubation temperature used in ART facilities.¹⁹ Viscosity of the hexadecane oil is reported as 3.474 mPa s.²⁰ Molar volume used to calculate oil diffusivity is reported as molar mass in g/mol divided by density in g/cm³, $\frac{M}{ρ}$ , and is approximated here as 294.04 cm³/mol for hexadecane.²⁰

Table 1 summarizes all fixed parameters that are used in this model.

Table 1.

Fixed Model Parameters

Parameter	System	Value
Surface area, A_aw	Biphasic	2 $\times$ 10⁻⁵ m²
Surface area, A_al	Triphasic	2.83 $\times$ 10⁻³ m²
Volume, V_w	Biphasic	5 $\times$ 10⁻⁸ m³
Volume, V_l	Triphasic	1 $\times$ 10⁻⁵ m³
Initial VOC in media, C_iw0	Biphasic	0 mol/m³
Initial VOC in oil, C_il0	Triphasic	0 mol/m³
Airborne VOC, p_i^max	Both	500 ppb
Ideal gas constant, R	Both	8.21 $\times$ 10⁻⁵ m³ atm/mol K
Temperature, T	Both	310 K

VOC, volatile organic compound.

Selection of VOCs for study

Eight VOCs were studied in this system: acetaldehyde, acrolein, ethanol, formaldehyde, isopropanol (isopropyl alcohol), phenol, styrene, and toluene. Acetaldehyde was chosen as important for research, as comprehensive embryotoxicity studies have indicated that it is a contaminant of interest, suggesting a great need to monitor its presence in ART facilities.^12,30 The Cairo Consensus has indicated this need by showing that in ART facilities, acetaldehyde is found with the second-highest average airborne concentration, just below formaldehyde.² It is also a metabolite of ethanol, the most prevalent VOC in ART facilities.² Acrolein has been well studied as an embryotoxic agent and has relatively strong affinity for biological matter, similar to phenol and toluene.¹⁹ Isopropanol is widely used as a disinfectant in laboratory spaces.² Styrene is ubiquitous in laboratory spaces being that it volatilizes from new plasticware due to incomplete polymerization of polystyrene in production, shown to be highly embryotoxic and highly prevalent in ART facilities with great affinity to partition into the cell.^2,19

The thermodynamic properties of the selected VOCs differ, specifically in polarity: acetaldehyde, acrolein, ethanol, formaldehyde, isopropanol, and phenol are polar molecules, while toluene and styrene are nonpolar. Polarity of a VOC is associated with its hydrophilicity, or affinity to partition into water: more polar VOCs will typically be more hydrophilic, while nonpolar VOCs are more hydrophobic and have a greater relative affinity for partitioning into organic phases such as oil.

Table 2 provides chemical and thermodynamic partitioning properties of the eight chosen VOCs at 37°C, which is a typical incubation temperature utilized in IVF culture.^31–38 It is worth noting that the molar volumes are computed from available density data at typical room temperature (20°C–25°C), while partition coefficients are listed at the modeled temperature of 37°C, as they may be more easily converted given a set temperature. The order of magnitude changes of the molar volumes between 20°C and 37°C is expected to be insignificant in most cases (Table 2).

Table 2.

Chemical Properties of Chosen Volatile Organic Compounds^19,31–38

Compound	Molecular weight (g/mol)	Molar volume (cm³/mol)	K_Hi	K_ial	k^aw (m/s)	k^al (m/s)
Acetaldehyde	44.05	55.90	0.00544	0.0442	0.00103	1.11 $\times$ 10⁻⁴
Acrolein	56.06	55.82	0.103	0.0177	1.53 $\times$ 10⁻⁴	4.96 $\times$ 10⁻⁴
Ethanol	46.07	58.32	4.68 $\times$ 10⁻⁴	0.0347	0.00136	3.03 $\times$ 10⁻⁴
Formaldehyde	30.02	28.86	1.95 $\times$ 10⁻⁵	0.431	0.00174	9.24 $\times$ 10⁻⁵
Isopropanol	60.10	76.56	7.70 $\times$ 10⁻⁴	0.0174	0.00116	3.53 $\times$ 10⁻⁴
Phenol	94.11	87.77	3.31 $\times$ 10⁻⁵	0.000807	9.71 $\times$ 10⁻⁴	8.74 $\times$ 10⁻⁴
Styrene	104.2	158.5	0.143	0.000200	8.60 $\times$ 10⁻⁵	8.76 $\times$ 10⁻⁴
Toluene	92.14	105.9	0.446	0.000962	3.66 $\times$ 10⁻⁵	8.30 $\times$ 10⁻⁴

Results

A key outcome of kinetic modeling is estimating the time required for VOC concentrations in biphasic and triphasic systems to reach equilibrium. Here, equilibrium was considered reached when 99% of the maximum concentration was achieved in the top layer: oil for the triphasic system and water-based culture media for the biphasic system. Maximum concentrations (C_max) were defined as the equilibrium concentrations. Table 3 and Figure 3 show the equilibration times for the eight VOCs in both the biphasic and triphasic systems.

Table 3.

Equilibration Times in Biphasic and Triphasic Systems

VOC	System	Approximate time of equilibration (s)
Acetaldehyde	Biphasic	12
	Triphasic	147
Acrolein	Biphasic	76
	Triphasic	105
Ethanol	Biphasic	9
	Triphasic	171
Formaldehyde	Biphasic	7
	Triphasic	559
Isopropanol	Biphasic	10
	Triphasic	47
Phenol	Biphasic	12
	Triphasic	60
Styrene	Biphasic	135
	Triphasic	19
Toluene	Biphasic	121
	Triphasic	63

VOC, volatile organic compound.

FIG. 3.

Equilibration times in biphasic and triphasic systems.

To visualize the rate of concentration increase approaching the maximum equilibrium concentration, it is useful to plot the fraction of maximum concentration (C/C_max) computed by the kinetic model at specific time intervals. Each curve is unique for each VOC and system. As the curve slope levels and approaches a horizontal asymptote at C/C_max = 1, concentration increase is slowing, and the “time of equilibration” is near. Time intervals for the curves were chosen to best frame the concentration increase and subsequent leveling. VOCs are often classified by polarity, as polarity generally correlates with hydrophilicity. Figures 4 and 5 show the equilibration curves for polar (hydrophilic) and nonpolar (hydrophobic) VOCs in a biphasic system as visual examples of the approach toward equilibrium. Figures 6 and 7 show the same phenomena of VOC partitioning across the air–oil interface in a triphasic system.

FIG. 4.

Equilibration of polar VOCs in media: Biphasic system. VOC, volatile organic compound.

FIG. 5.

Equilibration of nonpolar VOCs in media: Biphasic system. VOC, volatile organic compound.

FIG. 6.

Equilibration of polar VOCs in oil: Triphasic system. VOC, volatile organic compound.

FIG. 7.

Equilibration of nonpolar VOCs in oil: Triphasic system. VOC, volatile organic compound.

The fastest recorded equilibration time was that of formaldehyde. Air–media (biphasic system) partitioning of formaldehyde occurred at 7 seconds, and air–oil partitioning of formaldehyde (triphasic system) took the longest of any VOC at 559 seconds. More hydrophilic VOCs were predicted to partition into media faster than oil. Acrolein partitioned into uncovered media in 76 seconds and oil in 105 seconds, ethanol entered media in 9 seconds and oil in 316 seconds, isopropanol entered media in 10 seconds and oil in 47 seconds, and phenol entered media in 12 seconds and oil in 60 seconds. Conversely, more hydrophobic VOCs were predicted to enter the oil phase more rapidly. Styrene partitioned into media in 135 seconds according to the model, but oil in just 19 seconds. Similarly, toluene entered media in 121 seconds and oil in 63 seconds.

Discussion

The primary advantage of kinetic modeling is mapping the concentration changes over time in a system to estimate how long it takes for equilibration to be reached in a given phase. It was found that different VOCs partition to different phases in different time frames, yet all within several minutes when considering the scale of the cell culture. Greater affinity for a phase, expressed via partition coefficients, is tracked with a quicker equilibration time in said phase. For instance, styrene is more hydrophobic than acetaldehyde, suggesting that it may partition into an oil-overlaid system more rapidly. While the oil in a triphasic system may act as a sink for much of the hydrophobic, lipophilic compound, it does not completely protect the culture media from styrene infiltration, as shown in equilibrium partitioning modeling. As molecular diffusion dominates within the culture system, an oil overlay may draw certain compounds to the system more rapidly. Because the components of the culture are not continuously mixed or disturbed, equilibration across the boundary between oil and media in a triphasic system is solely a function of diffusivity, and there is no two-film boundary that must be overcome once the VOC is out of the air phase. However, it must be emphasized that the scope of this work only encompasses the top-level phase partitioning: that is, how quickly the VOC will equilibrate into the media in a biphasic system or how quickly the VOC will equilibrate into the oil alone in a triphasic system, when being continuously supplied from the air phase. The kinetics of passing from the oil to the media are not entirely clear. Equilibrium modeling, alone, does not account for diffusive flux, negating interfacial surface areas as well as diffusivities, limiting its potential to predict real-time dynamic concentrations.

Several components of the kinetic model are highly scale dependent. A key consideration in the model is the volume-to-surface area ratio or the depth of penetration in the dish. Considering surface area in the model goes beyond the use of equilibrium partition coefficients to quantify flux.³⁹ When this ratio is maximized, that is, the surface area of the fluid boundary is relatively small compared to the total fluid volume, the time of equilibration is increased. This indicates that minimizing the amount of liquid area exposed to open air may reduce cell culture exposure to airborne VOCs. Essentially, the shape of the container has a notable impact on the flux of VOCs into the system. Even when a prescribed volume of liquid is used, the surface area may be minimized, possibly protecting the culture from VOC load in the case of acute spikes. There may be benefits to altering the exposed surface area between air and liquid to prevent VOC entry, should a small amount of VOC enter a space and be rapidly diffused.

This study makes a first attempt at modeling kinetics of VOCs in the context of IVF based upon similarly scaled experiments. However, this behavior is greatly dependent on mass transfer coefficients, which are best determined experimentally for unique VOCs in unique contexts. The only previous work in modeling VOC kinetics for IVF estimated mass transfer coefficients based upon a model used for large natural bodies of water and only considered air–water partitioning.⁴⁰ This warrants further experimentation to develop system-specific parameters to establish more confidence in results.

Modeling is deemed necessary at this scale due to the experimental difficulty of measuring the film thickness associated with VOC flux in such a small dish. To quantify the order of magnitude differences expected with changes in overall surface area and volume, as a bridge to other literature that has performed similar experimentation at a larger scale, Figure 8 expands the model for acetaldehyde to illustrate differences in kinetics expected across different fluid depths. To develop this analysis, different fluid depths were utilized based on vessel dimensions, with all other parameters remaining constant.

FIG. 8.

Acetaldehyde partitioning kinetics in water in differently sized vessels.

This demonstrates differences across several orders of magnitude for even a very hydrophilic VOC partitioning into water when the vessel dimensions are altered, reinforcing the system-specific nature of this phenomenon. This is reflected through Einstein’s equation, given as Equation 16:

t = \frac{L^{2}}{2 D}

(16)

where t is the characteristic time, L is the length of diffusion path (i.e., the depth of the fluid), and D is the diffusivity of the compound through a given medium. This equation demonstrates that as fluid depth increases by a factor, the time to diffuse through the fluid completely increases by the square of that factor.

Overall, the modeling showed that VOCs may enter IVF culture media within seconds to minutes. This is sensible in that the cultures have extremely small volumes when compared to other benchtop systems that have been modeled.^23,24,29 Table 4 shows the impact on time of equilibration achieved by maximizing all the model parameters.

Table 4.

Trends in Model Parameters

By increasing parameter…	System-specific?	Impact
p_i^max	No	No change
T	No	−
K_Hi or K_ial	No	+
k_aw or k_al	Yes	−
A_aw or A_al	Yes	−
V_w or V_l	Yes	+

+, slower to equilibrate with an increase in a parameter; −, faster to equilibrate with an increase in a parameter.

Conclusions

While air quality plays a clear role in the success of cell development, there is limited understanding of VOC partitioning into embryo or cell cultures and the subsequent effects that the VOCs may have after contacting the embryos or individual cells. In previous work, a model was proposed to quantify the static equilibrium concentrations of VOCs present in the oil overlay, culture media, and biological cell matter. This work serves to enhance this model by describing the kinetic behavior, that is, time dependency, of this partitioning. Preliminary results show a strong indication for the modeled VOCs to enter the cell culture from the ambient air at extremely fast rates, where they may compromise cell viability and health. This scale implies that VOCs cannot be “outrun” in the IVF laboratory, and the best practice is eliminating their presence in the first place to adequately protect the culture environment.

Kinetic modeling may be enhanced in future work with experiments to obtain the mass transfer coefficients of VOCs in this unique system, where such modeling has not been applied in literature. This will allow for confirmatory experimentation to be conducted, in which concentrations over time may be mapped to evaluate the fit of the modeling. The model may also be extrapolated outward to accommodate a wider range of common VOCs in health care. Equilibrium modeling outlined by Fox et al. describes 23 VOCs of interest.¹⁹ More of these VOCs, among others, may be incorporated into kinetic modeling and validated with gas chromatography–mass spectrometry testing to examine their various unique rates of infiltration into cell cultures. Such information would be vital in predicting the net effects resulting from spikes of complex mixtures of VOCs into sensitive environments. For instance, kinetic data would allow practitioners to determine how quickly their cultures may be compromised in the event of an irregular or unexpected influx of VOC due to a nearby point source. Further experimentation must be designed to assess triphasic systems that have been exposed to VOCs for varying amounts of time at set air-phase concentrations. These experiments could result in the determination of calibration factors that would improve the fit of the model for extrapolation to various contaminant exposures across varying time scales.

By considering the impact that oil offers in shielding cells from VOCs, practitioners may take steps beyond controlling ambient air quality to further protect their cultures. However, this model also shows that oil, regardless of its composition or weight, is not an impenetrable shield against the infiltration of VOCs. Hydrophobic species may enter the culture system and equilibrate more rapidly when the continuously mixed air phase is exposed to oil. Light or heavy oil still provides a sink to store hydrophobic VOCs at equilibrium, even if system kinetics are enhanced. Other standard operating procedures may be updated as well, such as the regulation of isopropyl alcohol sanitation to minimize airborne alcohol concentrations. With a kinetic modeling approach to accurately predict the concentrations of indoor air pollutants that infiltrate cell cultures, clinical outcomes in highly sensitive medical procedures may be greatly improved.

Authors’ Contributions

J.S.R.: Writing, review, editing, visualization, and analysis. J.T.F.: Writing, review, editing, conceptualization, supervision, and analysis. H.T.H.: Writing, review, editing, visualization. K.C.W.: Conceptualization, visualization, review, and editing. D.G.B.: Conceptualization and editing.

Footnotes

Acknowledgments

The authors would like to thank the practitioners and colleagues who asked the critical questions specific to the mechanics of VOC equilibrium and cytotoxicity.

Author Disclosure Statement

J.T.F.’s research has been funded in part by LifeAire Systems LLC and the State of Pennsylvania. K.C.W. and H.T.H. are employed by LifeAire Systems LLC and own stock in the company. J.S.R. is employed by LifeAire Systems LLC. D.G.B. has no potential competing interests to declare.

Funding Information

This work was supported by a grant made possible through the Pennsylvania Infrastructure Technology Alliance, a program of the Pennsylvania Department of Community and Economic Development that sponsors industry–academic collaborations. As part of this funding, LifeAire Systems LLC in part sponsored this effort at Lehigh University.

References

Fonseca

, Abreu

, Guerreiro

, et al. Indoor air quality in healthcare units—A systematic literature review focusing recent research. Sustainability, 2022; 14(2):967; doi: 10.3390/su14020967

Mortimer

, Cohen

, Mortimer

, et al. Cairo consensus on the IVF laboratory environment and air quality: Report of an expert meeting. Reprod Biomed Online, 2018; 36(6):658–674; doi: 10.1016/j.rbmo.2018.02.005

Bessonneau

, Mosqueron

, Berrubé

, et al. VOC contamination in hospital, from stationary sampling of a large panel of compounds, in view of healthcare workers and patients exposure assessment. PLoS One, 2013; 8(2):e55535; doi: 10.1371/journal.pone.0055535

Shen

, Yuan

, Zeng

. An effort to test the embryotoxicity of benzene, toluene, xylene, and formaldehyde to murine embryonic stem cells using airborne exposure technique. Inhal Toxicol, 2009; 21(12):973–978; doi: 10.1080/08958370802687493

Cohen

, Gilligan

, Esposito

, et al. Ambient air and its potential effects on conception in vitro. Hum Reprod, 1997; 12(8):1742–1749; doi: 10.1093/humrep/12.8.1742

Boone

, Johnson

, Locke

A-J

, et al. Control of air quality in an assisted Reproductive Technology Laboratory. Fertil Steril, 1999; 71(1):150–154; doi: 10.1016/s0015-0282(98)00395-1

Ginn

, Mandwie

, Alexander

, et al. Gene therapy clinical trials worldwide to 2023: An update. J Gene Med, 2024; 26(8):e3721; doi: 10.1002/jgm.3015

Emerson

, Glassey

. Bioprocess monitoring and control: Challenges in cell and gene therapy. Curr Opin Chem Eng, 2021; 34:100722; doi: 10.1016/j.coche.2021.100722

Gostner

, Zeisler

, Alam

, et al. Cellular reactions to long-term volatile organic compound (VOC) exposures. Sci Rep, 2016; 6(1):37842; doi: 10.1038/srep37842

10.

Hyland

. A two-phase network theory of atopy and asthma causation: A possible solution to the impact of genes, hygiene and air quality. Clin Exp Allergy, 2001; 31(10):1485–1492; doi: 10.1046/j.1365-2222.2001.01234.x

11.

Kuang

, Li

, Lv

, et al. Exposure to volatile organic compounds may be associated with oxidative DNA damage-mediated childhood asthma. Ecotoxicol Environ Saf, 2021; 210:111864; doi: 10.1016/j.ecoenv.2020.111864

12.

Lau

C-F

, Vogel

, Obe

, et al. Embryologic and cytogenetic effects of ethanol on preimplantation mouse embryos in vitro. Reprod Toxicol, 1991; 5(5):405–410; doi: 10.1016/0890-6238(91)90003-x

13.

Reimers

, Flockton

, Tanguay

. Ethanol- and acetaldehyde-mediated developmental toxicity in Zebrafish. Neurotoxicol Teratol, 2004; 26(6):769–781; doi: 10.1016/j.ntt.2004.06.012

14.

Croute

, Poinsot

, Gaubin

, et al. Volatile organic compounds cytotoxicity and expression of HSP72, hsp90 and grp78 stress proteins in cultured human cells. Biochim Biophys Acta, 2002; 1591(1–3):147–155; doi: 10.1016/s0167-4889(02)00271-9

15.

Wang

, Li

, Liu

, et al. Modulation of microRNA expression by volatile organic compounds in mouse lung. Environ Toxicol, 2014; 29(6):679–689; doi: 10.1002/tox.21795

16.

Brown-Woodman

, Webster

, Pickler

, et al. In vitro assessment of individual and interactive effects of aromatic hydrocarbons on embryonic development of the rat. Reprod Toxicol, 1994; 8(2):121–135; doi: 10.1016/0890-6238(94)90019-1

17.

Pariselli

, Sacco

, Ponti

, et al. Effects of toluene and benzene air mixtures on human lung cells (A549). Exp Toxicol Pathol, 2009; 61(4):381–386; doi: 10.1016/j.etp.2008.10.004

18.

Khalil

, Nasir

. (2010). Toxicity of volatile organic compounds (VOCs) mixtures using human derived cells. WIT Transactions on Ecology Environment, 132, 3–12; doi: 10.2495/etox100011

19.

Fox

, Ni

, Urrutia

, et al. Modeling the equilibrium partitioning of low concentrations of airborne volatile organic compounds in human IVF laboratories. Reprod Biomed Online, 2023; 46(1):54–68; doi: 10.1016/j.rbmo.2022.05.006

20.

Schwarzenbach

, Gschwend

, Imboden

. (2017). Environmental Organic Chemistry (3rd ed.). Wiley.

21.

Scarica

, Monaco

, Borini

, et al.; SIERR, Società Italiana di Embriologia Riproduzione e Ricerca. Use of mineral oil in IVF culture systems: Physico-chemical aspects, management, and safety. J Assist Reprod Genet, 2022; 39(4):883–892; doi: 10.1007/s10815-022-02479-z

22.

Abraham

, Whiting

, Fuchs

, et al. Thermodynamics of solute transfer from water to hexadecane. J Chem Soc, Perkin Trans 2, 1990(2):291; doi: 10.1039/p29900000291

23.

Mahmoodlu

, Pontedeiro

, Pérez Guerrero

, et al. Dissolution kinetics of volatile organic compound vapors in water: An integrated experimental and computational study. J Contam Hydrol, 2017; 196:43–51; doi: 10.1016/j.jconhyd.2016.12.004

24.

, Kropscott

. Evaluation of the three-phase equilibrium method for measuring temperature dependence of internally consistent partition coefficients (k_ow, k_oa, and k_aw) for volatile methylsiloxanes and trimethylsilanol. Environ Toxicol Chem, 2014; 33(12):2702–2710; doi: 10.1002/etc.2754

25.

Rodriguez Castillo

A-S

, Biard

P-F

, Guihéneuf

, et al. Assessment of VOC absorption in hydrophobic ionic liquids: Measurement of partition and diffusion coefficients and simulation of a packed column. Chem Eng J, 2019; 360:1416–1426; doi: 10.1016/j.cej.2018.10.146

26.

Mahmud

, Erguvan

, MacPhee

. Performance of closed loop venturi aspirated aeration system: Experimental study and numerical analysis with discrete bubble model. Water (Basel), 2020; 12(6):1637; doi: 10.3390/w12061637

27.

Dumont

, Delmas

. Mass transfer enhancement of gas absorption in oil-in-water systems: A Review. Chem Eng Proces: Process Intensification, 2003; 42(6):419–438; doi: 10.1016/s0255-2701(02)00067-3

28.

Wang

, Zheng

, Xu

, et al. (2019). Liquid film thickness measurement for gas-liquid two phase flow using ultrasound. 2019 IEEE International Instrumentation and Measurement Technology Conference (I2MTC); doi: 10.1109/i2mtc.2019.8827025

29.

Liu

, Guo

, Roache

, et al. Henry’s law constant and overall mass transfer coefficient for formaldehyde emission from small water pools under simulated indoor environmental conditions. Environ Sci Technol, 2015; 49(3):1603–1610; doi: 10.1021/es504540c

30.

Chhibber

, Gilani

. Acrolein and embryogenesis: An experimental study. Environ Res, 1986; 39(1):44–49; doi: 10.1016/s0013-9351(86)80006-8

31.

U.S. National Library of Medicine. (n.d.). Acetaldehyde. National Center for Biotechnology Information. PubChem Compound Database. Available from: https://pubchem.ncbi.nlm.nih.gov/compound/Acetaldehyde

32.

U.S. National Library of Medicine. (n.d.). Acrolein. National Center for Biotechnology Information. PubChem Compound Database. Available from: https://pubchem.ncbi.nlm.nih.gov/compound/7847

33.

U.S. National Library of Medicine. (n.d.). Ethanol. National Center for Biotechnology Information. PubChem Compound Database. Available from: https://pubchem.ncbi.nlm.nih.gov/compound/Ethanol#section=Density

34.

U.S. National Library of Medicine. (n.d.). Formaldehyde. National Center for Biotechnology Information. PubChem Compound Database. Available from: https://pubchem.ncbi.nlm.nih.gov/compound/Formaldehyde

35.

U.S. National Library of Medicine. (n.d.). Isopropyl Alcohol. National Center for Biotechnology Information. PubChem Compound Database. Available from: https://pubchem.ncbi.nlm.nih.gov/compound/Isopropyl-Alcohol#section=Density

36.

U.S. National Library of Medicine. (n.d.). Phenol. National Center for Biotechnology Information. PubChem Compound Database. Available from: https://pubchem.ncbi.nlm.nih.gov/compound/Phenol

37.

U.S. National Library of Medicine. (n.d.). Styrene. National Center for Biotechnology Information. PubChem Compound Database. Available from: https://pubchem.ncbi.nlm.nih.gov/compound/Styrene#section=LogP

38.

U.S. National Library of Medicine. (n.d.). Toluene. National Center for Biotechnology Information. PubChem Compound Database. Available from: https://pubchem.ncbi.nlm.nih.gov/compound/Toluene

39.

Starokozhev

, Sieg

, Fries

, et al. Investigation of partitioning mechanism for volatile organic compounds in a multiphase system. Chemosphere, 2011; 82(10):1482–1488; doi: 10.1016/j.chemosphere.2010.11.033

40.

Worrilow

, Urrutia

, Huynh

, et al. Modeling of airborne embryotoxic volatile organic compounds (VOCs) in the IVF culture environment—Their concomitant cytotoxic concentration within the growth media and embryo [Abstract]. Fertil Steril, 2019; 112(3):E122–E123; doi: 10.1016/j.fertnstert.2019.07.439