Application of a data-processing model to determine the optimal sampling conditions for liquid phase trapping of atmospheric carbonyl compounds

Abstract

The reactivity of two fluorescent derivatization reagents, 2-diphenyl-1,3-indandione-1-hydrazone (DIH) and 2-aminooxy-N-[3-(5-dimethylamino-naphtalene-1-sulfonamino)-propyl]-acetamide (dansylacetamidooxyamine, DNSAOA), was studied towards selected atmospheric carbonyl compounds. The results were compared to those obtained using the 2,4-dinitrophenylhydrazine (2,4-DNPH) UV–vis reagent, a standard well-established technique used to detect atmospheric carbonyl compounds. The experimental rate constant were integrated into a data-processing model developed in the laboratory to simulate the trapping efficiencies of a mist chamber device as a function of temperature, reagent and solvent type among others. The results showed that in an aqueous solution, DNSAOA exhibits a higher reactivity towards carbonyl compounds without the addition of an acidic catalyst than 2,4-DNPH. It was observed that DNSAOA can trap efficiently water-soluble gaseous compounds (for example formaldehyde). However, because of a high initial contamination of the reagent caused by the synthesis procedure used in this work, DNSAOA cannot be used in high concentrations. As a result, very low trapping efficiencies of less reactive water-insoluble gaseous compounds (acetone) using DNSAOA are observed. However, the use of an organic solvent such as acetonitrile improved the trapping efficiencies of the carbonyl compounds. In this case, using DIH as the derivatization reagent (DNSAOA is not soluble in acetonitrile), trapping efficiencies greater than 95% were obtained, similar to 2,4-DNPH. Moreover, fluorescence associated with DIH derivatives (detection limits 3.33 × 10⁻⁸ M and 1.72 × 10⁻⁸ M for formaldehyde and acetone, respectively) is further advantage of this method for the determination of carbonyl compounds in complex matrix compared to the classical UV–vis detection method (detection limits 3.20 × 10⁻⁸ M and 2.9 × 10⁻⁸ M for formaldehyde and acetone, respectively).

Introduction

The atmospheric chemistry and photochemistry of low molecular mass carbonyl compounds including aldehydes and ketones is well documented. Interest in studying atmospheric chemistry of carbonyl compounds stems from: (1) their importance in atmospheric chemistry, (2) high reactivity, (3) complexity of their mechanism and dynamics, i.e. many photochemical reactions occur on multiple potential energy surfaces and generate multiple sets of products, and (4) their potential impact on human health.

During the past two decades, different strategies have been adopted in the analysis of carbonyl compounds in ambient air [1], [2]. The most common method uses the 2,4-dinitrophenylhydrazine (2,4-DNPH) derivatization agent coated on a solid sorbent. After trapping, the derivatization products are eluted with an organic solvent and analyzed by liquid chromatography (LC) coupled to an UV–vis detector [3], [4], [5], [6], [7]. Unfortunately, interferences mainly induced by oxidant such as ozone [8], [9], [10] and NOx (NOx = NO + NO₂) [11] were reported. They can be drastically reduced using KI annular denuders as oxidant scavengers but it was reported that this procedure could trap a fraction of the gaseous carbonyl compounds [8], [10], [12]. Consequently, the sampling procedure based on a gas–liquid scrubber sampling technique seems more reliable because it is free of important oxidizing artefacts [12], [13], [14], [15]. With this technique, the role of the derivatization process is not only to stabilize the carbonyl compounds but also to improve the trapping efficiency by avoiding the saturation of the trapping solution by the carbonyl compounds. A calibrated mathematical algorithm simulating the trapping efficiencies of carbonyl compounds into the liquid phase was developed in a previous work [16]. It was used to determine and compare the trapping efficiencies of two kinds of gas–liquid scrubbers: mist chamber and glass tube. This work highlighted the ability of the mist chamber to concentrate samples and thus to improve the sensibility of the analytical method. Using the data-processing model previously validated, the present work is focused on the determination and the comparison of the performances of various derivatization reagents to trap volatile carbonyl compounds in a mist chamber.

Numerous derivatization reagents were developed for the analysis of carbonyl compounds such as: 2,4-dinitrophenylhydrazine [14], [17], dansylhydrazine (DNSH) [1], [2], [18], [19], [20], [21], 2-diphenyl-1,3-indandione-1-hydrazone (DIH) [1], [2], [21], [22], [23], N-(5-dimethylamino-1-naphtalenesulphonamido)-3-oxapentane-1,5-dioxyamine (dansyloxyamine, DNSOA) [24], 2-aminooxy-N-[3-(5-dimethylamino-naphtalene-1-sulfonamino)-propyl]-acetamide (dansylacetamidooxyamine, DNSAOA) [25], 4-N,N-dimethylamino-6-(4′-methoxy-1′-naphtyl)-1,3,5-triazine-2-hydrazine (DMNTH) [26], [27]. Because of their fluorescence properties, their solubility in water and/or organic solvents or their kinetic rate reaction towards carbonyl functional group, three of them (2,4-DNPH, DIH and DNSAOA, Fig. 1) seem particularly fit for the analysis of atmospheric carbonyl compounds using an integrative trap such as a mist chamber.

The objective of the present work is to compare the performances of 2,4-DNPH, the most commonly used derivatization reagent, with those of DIH and DNSAOA for the liquid phase sampling of gaseous carbonyls compounds. This work was carried out using formaldehyde and acetone as model molecules because of their different physico-chemical properties (formaldehyde, a water-soluble gaseous compound (Henry's constant >10³ mol L⁻¹ atm⁻¹) highly reactive and acetone, a low water-soluble gaseous compound (Henry's constant ≈30 mol L⁻¹ atm⁻¹) fairly unreactive). The comparisons were performed using the mathematical algorithm developed and validated by François et al. to simulate the trapping efficiencies of the carbonyl compounds under study in a mist chamber [16]. Laboratory experiments were carried out on DIH and DNSAOA to determine the rate constant of derivatization while those of DNPH were found in literature. The studied parameters were: (1) the solubility of the compounds under study in the trapping solutions (K_S), (2) the rate constant (k_D) for each derivatization reagent and (3) the influence of the temperature on trapping efficiencies.