Performance of Air Fresher System for the Removal of Various Odorants Released from Foodstuffs

2. 1 Sample Collection

For the purpose of this study, four types of foodstuffs (soy bean soup, steamed egg, kimchi, and fermented skate fish) were purchased from a local market on the day of the AF test. Sampling was conducted in a room (capacity of≈20 m³) at a mean temperature of 22± 1℃. All four foodstuffs were placed on a table in the closed room for 10 min to ensure the generation of sufficient odorant levels before air samples were collected using both sorbent tubes and bag samplers. Simultaneously, PM_2.5 was also quantified using a PM sensor. Each AF system was repeatedly operated for 30 min after removing the food samples from the room. Lastly, air samples were again collected and analyzed using each analytical instrument.

2. 2 Instrumental Set-up for Analysis

In this study, a total of 17 odorous compounds (including two reference compounds) were measured by three different analytical methods: (1) air server (AS)/thermal desorber (TD)/gas chromatography (GC)/pulsed flame photometric detector (PFPD) for five sulfur compounds (hydrogen sulfide, methanethiol, dimethyl sulfide, and dimethyl disulfide), (2) TD/GC/mass spectrometry (MS) for 13 odorants (aldehyde, ketone, ester, alcohol, carboxylic groups, and trimethylamine), and (3) ultraviolet (UV)/visible (Vis) spectrophotometry for ammonia analysis based on the indophenol blue method (Kim and Kim, 2014; Kim and Kim, 2013, 2012). In addition to GC and spectrophotometry-based analysis, air dilution sensory (ADS) tests were performed based on the olfactometry threshold method by diluting odor samples with odorless pure air to determine the dilutionto-threshold ratio (D/T ratio) values (Kim and Park, 2008). Detailed information of these analytical procedures regarding the following targets is available in our previous work (Jo et al., 2013) as well as in the supporting information of the present work (Table 2S).

2. 3 Data Analysis

Apart from the preliminary quantitation of the differences between the input and output odorant concentrations, three analytical parameters, including percent sorptive removal (%SR), odor intensity (OI; (Nagata, 2003)) and odor activity value (OAV; (Qian and Wang, 2005)) were assessed in order to estimate the performance of the different AF units toward all different target odorants. Generally, the SR of an odorant is an indication of adsorption efficiency of the air purifier system. The SR was estimated as the difference in the concentration data for each odorant as given in equation 1:

where X₁ and X₂ represent the concentrations of odorant before (input) and after (output) exposure to the AF system, respectively. Following this simple estimation, the use of Student t-tests was employed to statistically investigate the significant differences between the efficiencies of each pair of treatment conditions (i.e., paired combination among different AF units (A, B, C and D) as well as the natural degassing (E) condition). In the case of ultrafine dust removal (DR), %DR was also estimated.

The concept of OI and OAV pair can be used to numerically quantify or describe the strength of known odorants emitted from identifiable sources. The full details of these concepts have already been described in our recent work (Adelodun et al., 2015). Therein, we stated that by the use of odorant concentration, the odor intensity (OI, measured in OU m^-3) of all target compounds is calculated according to Stevens (1960), where psychophysical phenomena (stimuli) such as odor strength follows a power law (Stevens, 1960). The law is expressed as:

where X is the odorant concentration in mg m^-3, k is the stimulus-dependent constant (Dravnieks et al., 1986), and n is the characteristic constant of each odorant. The relationship between perceived psychological intensity and odorant concentration was later modeled by the Weber-Fechner law (modified from the Steven’s law) as:

where a is the Weber-Fechner coefficient and b is the intercept constant (Jiang et al., 2006). By the calculation of OI, the extent of reduction in odor intensity due to the purifiers could also be assessed in addition to the pattern for their concentration data. OAV (Qian and Wang, 2005), which is a dimensionless expression of odor strength, was obtained by the ratio of the odorant concentration after exposure (to a purifier, X₂) to the odorant threshold value (Nagata, 2003; Nagata and Takeuchi, 2003). Regardless of the extent by which the odor concentration is reduced after treatment, the odorant removal system is deemed “not environmentally efficient” if OAV>1, and vice versa. Hence, lower OAV values indicate a more efficient air purifier system in lowering odorant (and dust pollution) menace in the environment. Along with sorptive removal (%SR), these odor indices (OI and OAV) were used to judge the performance of the four commercially available air purifiers. Similarly, we also estimated percent reduction and we made comparisons with data presented in the open literature.

3. 1 R emoval of Odorant and Ultrafine Dust Particles by Air Purifiers

Using the analytical procedures described in Table 2S, the reduction in target odorant concentration (in logarithm form) by the four AFs (A to D) and natural degassing (E) was monitored and recorded (Fig. 1S). The concentrations (in ppb) before (- 1) and after (- 2) use of each AF are listed in Table 2. We observed that all AFs did not show any significant response to five odorants (PA, IA, VA, CA, and BZA). All the odorants had an unchanged concentration value between initial and final measurements under our experimental conditions (1.59, 0.90, 0.75, 0.94, and 0.97 ppb, respectively) between blanks and samples. Apart from these five targets, ammonia was the only odorant that did not undergo any observable changes from natural degassing conditions (i.e. E1=E2=1,000 ppb). As one of the most influential odorants most commonly encountered, a significant reduction in the concentration levels of ammonia would be an appreciable challenge for AF. However, as there was no detectable change in the concentration of ammonia in natural degassing, the performance of each AF cannot be directly compared relative to such conditions, i.e. below detection limit (BDL) as written in the footnote of Table 2. As the odorants that were BDL were not evaluated further, the total number of targets was reduced to 18 odorants for further statistical analyses.

Table 2.

Comparison of concentration levels of odorants and VOC measured from air samples.

The %SR values were calculated (equation 1) to estimate the extent of reduction in the target odorant levels due to the AF operation (or degassing). The obtained results are provided in Fig. 1 (Table 3S). Because adsorption is known to be influenced by both surface chemical groups and atomic or molecular weights, comparisons among the odorants were made and depicted according to (a) chemical functionalities and (b) decreasing order of molecular weights. It was obvious from the haphazard distribution of %SR that the extent of sorptive removal was neither directly nor solely dependent on such criteria. In certain cases, the AFs lowered the level of odorants relative to those reduced by natural degassing. As such, the results suggest that the effectiveness of AF with respect to odorants is not satisfactory relative to PM. Exceptions to this affirmation were observed where one or two AFs tended to either increase the levels of specific odorants (by emission) or were inefficient (slower scavenging rate) when compared to natural degassing. According to Fig. 1(b), such cases included those of HPA (all AFs), HXA (C & D), DMDS (A), BA (B & C), DMS (B), acetone (B & D) and AA (C & D). Here again, since no AF showed superior preference toward a specific chemical group of odorants, a conclusive remark regarding discrepancies in efficiency and specificity among the AFs toward the various odorant targets could not be drawn. However, when considering differences in their chemical properties (Fig. 1(a)), it was found that all AFs performed very well (between 60% and 80%) against VFAs, TMA, H₂S, and NH₃. However, in the case of AA, BA, acetone, DMS, and DMDS, a negative %SR was commonly seen. A negative %SR value in this work signifies an occurrence of self-emission of the specific odorant(s) by the AF or a situation whereby the rate of desorption was faster than the adsorption rate. The AF unit B demonstrated the worst performance in this regard. On average (~44% SR), the A product exhibited the highest removal efficiency amongst the different products. Apart from this rough estimation, the other AFs performed similarly well. By general comparison of their %SR, the removal of IBA was moderate (~78% SR) while that of DMS was worst (~23% SR).

Fig. 1.

Comparison of odor removal by natural degassing and air freshers with respect to (a) chemical group and (b) molecular weight.

Based on random removal efficiencies of the air purifying systems against various target odorants, Student’s paired t-tests were used (on the %SR) to determine if significant differences were present between the treatment pairs, i.e. between five sets of data (resulting in 10 pairs: A-B, A-C, A-D, B-C, B-D, C-D, E-A, E-B, E-C, and E-D) by examining the removal efficiency of the odorants after treatment. The results derived from this test on a population or a number of the 18 target compounds (df=17 and confidence limit of 95%) are listed in Table 4S. Accordingly, significant differences in the performance between each pair (t_cal) were greater than the tabulated values (t_tab) (i.e., t_cal>t_tab) in all cases. Performance-wise, among the AFs (based on the difference between t_cal and t_tab as well as between each pair of sorption efficiency), those for B and C products were the most closely related, while the highest discrepancy existed between the A and C products. It was also confirmed from this analysis that their correlation with natural degassing (E) was almost non-existent, as in most cases (especially for the E-A pair), t_cal>>t_tab. Judging from the large differences in the two terms between the pairs of natural degassing and the AFs (relative to the differences between each AF pair), the following conclusions can be drawn. Most importantly, although the odor removal ability of the AFs was higher than that of the natural scavenging process, they were yet ineffective in completely removing or significantly ameliorating the odor that originated from the selected food sources.

The response of the air purifier systems toward ultrafine particulates was also evaluated in reference to odorant data. From the actual concentration values provided in Table 2, we estimated percent dust removal (%DR) as shown in Fig. 2, and the values were compared along with the concentration values (before and after removal) to assess the relationship between the two conditions. When the extent of reduction was compared in terms of concentration levels amongst the AFs, purifier A was found to show the best removal ability toward ultrafine dust particles compared to others as it lowered the concentration by ~3.0 mg/m³. Note that product D exhibited the worst performance in this respect at an ~0.5 mg/m³ reduction of concentration. Despite such differences, product D exhibited approximately the same %DR (~95%; range of ~90-96 %) as products A, B and C. As such, most of the AFs effectively treated ultrafine dust relative to the odorants tested in this work. In the absence of AFs, the removal of PM_2.5 by natural means was found to be insignificant. We suggest that these particles are extremely fine and lightweight, and therefore tend to exhibit long residence time in the air by remaining suspended unless treated by the AF system.

Fig. 2.

Comparison of PM2.5 concentrations (before and after treatment) and removal efficiency (%DR) by each air fresher.

3. 2 E valuation of Odor Intensity and Odor Activity Value

The paired indices of odor strength (OI and OAV) have often been studied in the field of odor chemistry to control and assess the actual effect of odor pollution instead of evaluation based on concentration values (Jiang et al., 2006; Dravnieks et al., 1986; Stevens, 1960b). These calculations enable the quantification of olfactory stimuli strength for easy evaluation of intensity as well as comparison with other odorants. In this study, OI values were estimated both before

and after treatment using equation 3 and the detailed results are described in Table 3. Again, some odorants measured at or below the detection limit were not assessed here. In addition, CH₃SH and NH₃ were the only undetected odorants when natural degassing and the AF treatments were applied. Since the odorants were initially present at varying concentrations, calculating the percent reduction in terms of OI also quantitated the effect of the AF systems on each odorant. The results are depicted in Fig. 3(a). The odorants are arranged in descending order of %OI reduction to simulate the decreasing order of human tolerance. This comparison was also made along with those derived from natural degassing conditions. By virtue of OI values, the effect of the AF on various odorants was found to be random, as no AF showed either a particular preference or determent toward/against a particular or group of odorants. The after-treatment intensities (OI_initial-OI_final) of IBA, along with the two commonly encountered inorganic odorants (H₂S and NH₃), were found to be the highest while those of sulfur-containing compounds (DMDS, CH₃SH, and DMS) were the lowest (Table 3 and Fig. 3(a)). However, the order of OI reduction was IBA>H₂S>TMA>NH₃>PPA, and was found as the all-positive reduction results from the exercise. The increase in OI after treatment was random; consequently, we could infer that the effect of odor reduction in terms of OI-based assessment is not effective to rank the performance between the different AFs. The performance of all the AF units in lowering OIs of target odorants were nonetheless better than those of natural degassing. Such affirmation specifically holds true for IBA, H₂S, TM, NH₃ and PPA. Also, by comparing the ability of AFs to lower OI values, it was found that on average, purifier A was regarded as the product with the highest performance as it recorded superior values in OI reduction over the other products (Fig. 3(a)).

Table 3.

Odor intensity (OI) values of main odorants emitted from gaseous samples.

Fig. 3.

Comparison of % reduction of odor intensity values of target odorants from (a) the current work with (b) similar results in the literature (McGinley and McGinley, 2006) (Note: Situations where the calculated % reduction in OI value is negative are not indicated in this figure).

In a similar work, three air purifiers (A, B, and C) and one fragrance (D) were tested against odors emitted from food samples (McGinley and McGinley, 2006). The effect of the treatment was monitored with respect to the extent of the odorant reduction over a period of 2 h and the reduction in %OI was further computed and plotted (Fig. 3(b)). In the current work, we estimated the percent reduction in OI odors between the four AFs used as well as that of natural degassing. In order to ensure quantitative comparison of our work with the reference (see dotted arrow in Fig. 3(b)), the percent reduction in odor after 20 min (as the period of experimental exposure used in current work) was compared. It was reported that the three (A, B, C) AFs showed an average reduction in OI of 4.6% (range ~3.1%-6.7%) while D exhibited a distinct value of 37.3%. Despite the fact that we selected multiple target compounds for comparison, the AFs used in our study showed on average a comparatively similar percent reduction of OI.

The OAV values were also computed to assess how effectively odor was reduced after treatment (Table 4). More importantly, investigation into how significantly the AFs faired in elevating the sanity of the test atmosphere was conducted. Similar to the assessment based on OI, the results were also compared based on percent reduction in OAV as provided in Fig. 2S. Here, PA, IA, VA, and CH₃SH were the only inestimable odorants due to aforementioned reasons (under sorptive removal and OI analyses). Unlike the findings based on OI analysis, the OAV results showed a more appreciable response of the AFs towards the odorants. A percent reduction as high as 87% was exhibited by most AFs in terms of OAV, and if the results are compared in terms of relative ordering, the most significant reduction in OAV after the treatment was seen from IBA, H₂S, IVA, BTA, TMA, and NH₃. At the other extreme, odorants such as DMS, AA, acetone, and BA (in no specific order) demonstrated the least reduction in OAV values, and remain as difficult targets to treat based on this criterion.

Table 4.

Odor activity values (OAV) for the main odorants emitted from gaseous samples.