Fast sorption measurements of VOCs on building materials: Part 2 – Comparison between FLEC and CLIMPAQ methods

Highlights

• First comparison of two methodologies used to measure VOC sorption parameters on indoor surfaces.

• Highlighting on the robustness and the advantages of the FLEC-PTRMS method for field measurements.

• Usefulness of FLEC derived parameters as data inputs for indoor air quality models.

Abstract

A new method was developed to measure on the field VOC sorption coefficients (k_a; k_d) on the surface of a material by coupling a Field and Laboratory Emission Cell (FLEC) to a Proton Transfer Reaction-Mass Spectrometer (PTR-MS) as presented in the first part of this study. In this second part, the method is compared to the classical method based on a CLIMPAQ chamber coupled to an on-line GC analyzer. Different models were used to determine the sorption parameters from experimental data taking into account the sink effect on empty chamber walls and the presence of a boundary-layer. Determined sorption equilibrium coefficients K_e (k_a/k_d) for a mixture of BTEX on a gypsum board was found to be in good agreement between both methods. However, the CLIMPAQ method seems to be less robust than the FLEC method in the determination of sorption coefficients since more than one couple of (k_a; k_d), showing the same ratio K_e can retrieve the same CLIMPAQ experimental data. Giving this result, the question arises about the reliability of the literature data determined using emission test chamber which could be one of the reasons behind the discrepancies found between experimental indoor concentrations and predicted ones using chamber derived parameters.

Introduction

Nowadays, people spend between 60 and 80% of their time in indoor areas that can be contaminated by a large range of pollutants having hazardous effects on human health [1], [2], [3], [4]. Pollutants, especially Volatile Organic Compounds (VOCs), are ubiquitous contaminants for indoor areas [5]. Other than their emissions by surfaces such as building materials, VOCs can interact with these indoor surfaces through different processes including adsorption and desorption. Therefore sorption processes can be key drivers of indoor air concentrations since building materials can act as both a source or a sink for VOCs [6].

In the literature, studies were carried out using either static or dynamic (flow-through) chamber experiments to determine sorption rate coefficients (k_a and k_d) of VOCs on indoor materials or equilibrium partition coefficients (K_e) [7], [8], [9], [10]. In these studies, experimental data were analysed using different mathematical sinks models to extract sorption parameters [11], [12]. However, indoor air quality models using chamber derived parameters failed to predict real indoor concentrations [13] and differences as high as a factor of 9 were observed [14]. First, this disagreement can be related to the non-representativeness of laboratory experiments compared to field conditions (type of material, implementation conditions, ageing due to environmental conditions, etc.) and highlights a need to reliably measure sorption parameters in the field under real conditions [14]. Secondly, sorption on experimental chamber internal walls was considered in some studies as insignificant [15] while others [7], [16], [17], [18] reported a sink effect on chamber walls for different VOCs such as ethylbenzene, n-dodecane, α-pinene, 1,2,4-trichlorobenzene…etc. Thus, any underestimation of the chamber sink can introduce biases in measuring sorption coefficients of materials. Thirdly, the use of inappropriate mathematical models to extract sorption coefficients from experimental data may also introduce biases that could be responsible for the discrepancies mentioned above. Blondeau [11] demonstrated that models relating macroscopically the bulk air and the surface concentrations through adsorption and desorption constants are not scientifically sound. Zhang YP [19] highlighted that the mass transport process taking place between the material surface and the bulk air has to be accounted for, which is not usually described in the mathematical models used to analyse chamber experiments. In fact, several models has been developed without taking into account for the mass transfer coefficient but the adsorption (k_a) and desorption (k_d) rates [9], [10], [20] or the diffusion coefficient (D_m) in the building material [21]. Later on, Deng [22] proposed an improved model that considers for the convective mass transfer coefficient (h_m) through the boundary layer present on the surface of a material as well as the diffusion and the partitioning coefficients. Therefore, models combining local sorption equilibriums should be used in combination with the mass transport equations to fit the experimental measurements.

Trying to shed some light on the discrepancies found in literature, a new methodology based on a coupling between a FLEC (Field and Laboratory Emission Cell) and a PTR-MS (Proton Transfer Reaction – Mass Spectrometer), was developed. The experimental setup is detailed in the companion paper (Part 1) [23], and only few details are given here. The FLEC inlet is connected to two gas generation systems, to be supplied either with humid clean air at constant flow rate and stable relative humidity or with a diluted VOC mixture; the outlet is connected to the PTR-MS to quantify VOC concentrations exiting the cell. A sorption experiment performed using the FLEC is similar to the classical experiments described in the literature and using a test emission chamber of several liters of volume. The method was tested by performing sorption experiments on a gypsum board and vinyl flooring using a mixture of BTEX at ppb levels. Adsorption and desorption coefficients were derived from experimental output concentration profiles using the Tichenor model [10] that can be applied to the FLEC cavity to determine elementary sorption coefficients. Sorption coefficients were successfully determined and independently on flow conditions and on VOCs concentration, with an experimental error lower than 15%. The limits of applicability were also assessed for this method and showed that sorption parameters (k_a, k_d) in the range of (0.01 m h⁻¹; 0.01 h⁻¹) and (0.09 m h⁻¹; 0.09 h⁻¹) can be measured using 2 and 20 s of time resolution respectively, with an accuracy better than 10%. The FLEC-PTRMS method is suitable for field applications and allows a reduction of the measurement time to 0.5–12 h compared to several days for the emission test chamber method [7], [24], [25].

The objective of the present study is to compare sorption coefficients measured using the FLEC-PTRMS and the emission test chamber methodologies for a common VOC mixture and the same unpainted gypsum board. In fact, the FLEC was developed and used in previous studies to measure materials emissions [26] as an alternative method to emission test chambers [27]. Under this framework, the two methods were only compared for emission measurements in several studies [28], [29], [30]. Since the FLEC had never been used as a tool to measure in-situ VOC sorption parameters, a comparison for sorption measurements has yet to be performed. To achieve this new objective, a 50-L CLIMPAQ chamber (Chamber for Laboratory Investigations of Materials, Pollution and Air Quality) coupled to a compact gas chromatography analyser (AirmoVOC, Chromatotech) was used to perform the sorption experiments. The comparison of the two methods involved two approaches:

• The comparison between CLIMPAQ experimental concentration profiles and the concentration profiles that should be obtained considering the sorption parameters derived using the FLEC-PTRMS method and,

• The comparison of the sorption coefficients derived from the CLIMPAQ data using two models (that account for the sink effect on the chamber walls and for the boundary-layer mass transport) with the FLEC derived parameters.

A highlight on the usefulness of the FLEC derived parameters for modelling sorption processes in real buildings is also given.