Environmental Science and Pollution Research

, Volume 23, Issue 7, pp 6300–6311

Measurements of VOC/SVOC emission factors from burning incenses in an environmental test chamber: influence of temperature, relative humidity, and air exchange rate

  • A. Manoukian
  • D. Buiron
  • B. Temime-Roussel
  • H. Wortham
  • E. Quivet
Research Article

DOI: 10.1007/s11356-015-5819-2

Cite this article as:
Manoukian, A., Buiron, D., Temime-Roussel, B. et al. Environ Sci Pollut Res (2016) 23: 6300. doi:10.1007/s11356-015-5819-2

Abstract

This study investigates the influence of three environmental indoor parameters (i.e., temperature, relative humidity, and air exchange rate) on the emission of 13 volatile organic compounds (VOCs) and semi-volatile organic compounds (SVOCs) during incense burning. Experiments have been carried out using an environmental test chamber. Statistical results from a classical two-level full factorial design highlight the predominant effect of ventilation on emission factors. The higher the ventilation, the higher the emission factor. Moreover, thanks to these results, an estimation of the concentration range for the compounds under study can be calculated and allows a quick look of indoor pollution induced by incense combustion. Carcinogenic substances (i.e., benzene, benzo(a)pyrene, and formaldehyde) produced from the incense combustion would be predicted in typical living indoors conditions to reach instantaneous concentration levels close to or higher than air quality exposure threshold values.

Keywords

Incense Combustion Indoor air quality 

Introduction

Indoor air quality is influenced by many combustion sources such as tobacco smoking, cooking, heating, or candle/incense burning (Willers et al. 2006; Petrick et al. 2011; Petry et al. 2014). Burning incense indoors is a common practice in many countries, more particularly in the Middle East, Asia, and parts of Africa. It is used for various reasons covering religious, spiritual, aesthetic, and therapeutic practices, or simply for home fragrance. This source, though being small, emits organic compounds in both gas and particle phases that cause adverse health effects such as airway disease and cancer (Friborg et al. 2008; Lin et al. 2008; Cohen et al. 2013; Pan et al. 2014). Many toxic compounds have been identified in the incense smoke (Chuang et al. 2011; Hsueh et al. 2012; Manoukian et al. 2013 and references therein; Cohen et al. 2013) and arise from incomplete combustion, evaporation, and/or thermal decomposition processes (Derudi et al. 2013). It follows that smoke emitted from incense combustion has been considered by the World Health Organization as a potentially serious indoors problem for public health (WHO–World Health Organization 2010). Previous studies have indicated that the chemical composition and the range of levels of pollutants in incense smoke not only depend on incense composition and manufacturing process (e.g., elemental composition of herbs/wood powder, nature of fragrance materials, nature of adhesive powder, nature of stick…) (Hsueh et al. 2012; Yang et al. 2012) but may also vary according to the environmental conditions (i.e., temperature, relative humidity, air-flow rate, and oxygen content) (Yang et al. 2005, 2007, 2013).

The aim of this study is to investigate the effects of temperature, relative humidity, and air exchange rate on the VOC/SVOC emission factors of incense combustion. The burning emission experiments were performed in an environmental test chamber. The influence and interactions between these three environmental parameters were estimated using a classical two-level full factorial design.

Materials and methods

Materials

The incense investigated was in stick form. Its core, made of bamboo stem, was covered by layers of wood powder (m ≈ 500 mg). They are homogeneously flavored by a dispersing technique of the perfume in the wood powder without solvent addition. All tested incense sticks come from the same batch and have the same characteristics (shape, length, diameter, color, and fragrance). The average burning duration of incense stick is 27.6(±2.8) min. Emission characteristics from these incense sticks are detailed in Manoukian et al. (2013).

Experimental chamber

Description and general characteristics

All experiments were performed in a VCE1000 emission test chamber (Vötsch Industrietechnik, Fig. 1). This experimental chamber is designed to study material emission in well-controlled conditions. In this chamber, temperature can be controlled between +20 and +130 °C (with a deviation ranging from ±0.1 to ±0.3 °C), relative humidity (RH) can be set between 5 and 95 % (with a deviation from 1 to 3 %), and air exchange rate (AER) is adjustable between 0.1 and 2.0 h−1.
Fig. 1

Scheme of the experimental test chamber VCE1000

Briefly, the VCE1000 experimental chamber consists in a 916-L test space chamber housed in an air jacket temperature control system. The inner walls of the test chamber are made of electro polished stainless steel to minimize the surface area and thereby to lessen the potential effect of organics wall losses, as already demonstrated in sorption studies (De Bortoli et al. 1996; Popa and Haghighat 2002). An oil-free compressor (Dürr technik, TA-100 K) and a set of cartridges and filters supply clean air into the chamber. Clean air flow enters the chamber through a ¼″ Swagelok® bulkhead union port located at the rear of the chamber. The flow is set by a mass flow controller up to 30 L min−1. The outlet port located on the side of the chamber was used for sampling incense smoke. A magnetic fan ensures a continuous air mixing in the test chamber, providing a homogeneous mixture for sampling. Connection, tubing, and internal surfaces in contact with the air flux are designed without plastic to prevent the release of organic compounds such as phthalates, which could induce contamination during the emission tests.

Incense was placed on the shelf at the center of the chamber and was lighted by a lighter.

Cleaning and maintenance

After each experiment, the stainless steel walls of the test chamber are cleaned three times with ethanol. After this cleaning, AER is maintained at 1.8 h−1 for at least 12 h to remove traces of solvent.

Selected pollutants

In a recent study, incense smoke composition has been described in details (Manoukian et al. 2013). Based on this study and on the literature (see Manoukian et al. 2013 and references therein), 20 VOCs and SVOCs have been selected in the present work according to their frequencies of occurrence, their concentration levels, or their potential toxicity for humans.

They were used to represent both gaseous and particulate emissions produced by incense combustion and to test the influence of temperature, relative humidity, and AER on the emission factors.

These 20 compounds under study are split into four chemical groups:
  • for the gas phase:
    • ○ carbonyls: formaldehyde, acetaldehyde, acrolein, acetone, propionaldehyde, benzaldehyde, crotonaldehyde, and hexaldehyde;

    • ○ monocyclic aromatic hydrocarbons (MAHs): benzene, toluene, ethylbenzene, and xylenes;

  • for the particle phase,
    • ○ polycyclic aromatic hydrocarbons (PAHs): phenanthrene, fluoranthene, pyrene, benzo(a)anthracene, chrysene, benzo(b,k)fluoranthene, benzo(e)pyrene, and benzo(a)pyrene;

    • ○ phthalates: diethyl phthalate, diisobutyl phthalate, dibutyl phthalate, butylbenzyl phthalate, and diethylhexyl phthalate.

Some of these compounds are classified by the International Agency for Research on Cancer (WHO IARC Monograph Working Group 1995, 1999a, b, 2000, 2010, 2012 as harmful to human health. This is the case for benzene, benzo(a)pyrene, formaldehyde (category 1, human carcinogen), acetaldehyde, benzo(a)anthracene, chrysene, diethylhexyl phthalate, ethylbenzene (category 2B, possible human carcinogen), acrolein, benzo(e)pyrene, butylbenzyl phthalate, crotonaldehyde, fluoranthene, phenanthrene, pyrene, toluene, and xylenes (category 3, not classifiable).

Sampling and analysis methods

Carbonyls

Gaseous carbonyl compounds were collected on DiNitroPhenylHydrazine (DNPH) coated short-body cartridges (Sep-Pak, Waters), at a flow rate of 1 L min−1. After liquid extraction with 5 mL of acetonitrile (ACN) (LC-MS Chromasolv, 99.9 %, Sigma Aldrich) and filtration on Nylon filter (NY 4-mm syringe filters, 0.45 μm, Phenomenex), samples were analyzed using high-performance liquid chromatography coupled with diode array detector (HPLC-DAD; Finnigan Surveyor PDA, ThermoScientific). The column was a Kinetex C18 (75 × 3.0 mm i.d., 2.6 μm, Phenomenex) thermostated at 32 °C. The flow rate was 0.6 mL min−1, and the injection volume was 10 μL. The eluent was a mixture of water and ACN using the following gradient profile: 0–5 min, 30 % ACN, 5–12 min, 30–35 % ACN, 12–25 min, 35–40 % ACN, 25–40 min, 40–60 % ACN, 40–50 min, 60–100 % ACN. The UV detection was centered at 360 nm. A standard mixture of target carbonyls (TO11/IP-6A Aldehyde/Ketone–DNPH Mix, Supelco) was used for the calibration.

Monocyclic aromatic hydrocarbons

Sampling was performed on-line by pumping air from the test chamber at a flow rate of 20 mL min−1 through the cold (−30 °C) trap (TurboMatrix Air Monitoring Trap, Perkin Elmer) of an automated thermal desorptionautomated thermal desorption (ATD 400, Perkin Elmer). After the sample collection, the cold trap was rapidly heated (40 °C s−1) to 350 °C in order to inject the sample into a chromatographic column (Elite-1, 60 m × 0.25 mm i.d., Perkin Elmer) of a gas chromatograph coupled with flame ionisation detector (GC-FID; Autosystem XL, Perkin Elmer), via a splitless injector held at 250 °C. The initial oven temperature was set at 50 °C for 10 min, ramped to 150 °C at 6.6 °C min−1, increased at 25 °C min−1, and held for 12 min at the final temperature of 200 °C. The FID detector was operated at 250 °C with air and hydrogen flow rates at 400 and 40 mL min−1, respectively. A gas standard mixture (TO-14A Aromatics Mix (14 components), 100 ppb in nitrogen, Restek) was used for calibration.

Polycyclic aromatic hydrocarbons and phthalates

Particles were collected on quartz glass fiber filters (47 mm, 2500QAT-UP, Pallflex), at a flow rate of 2 L min−1. A charcoal denuder placed upstream trapped gaseous compounds and thus reduced sampling artifacts. The samples were spiked before extraction with 100 μL of an isotope-labeled standard, tetracosane-d50 (2.5 μM). Then they were extracted using accelerated solvent extraction (ASE 300, Dionex; solvent: 1/1 (v/v) mixture dichloromethane (DCM)/acetone, T = 100 °C, PN2 = 100 bars, 5 min) and concentrated under nitrogen flow at 40 °C to a volume of 500 μL using a TurboVap II Concentration Workstation (Caliper LifeSciences). During the concentration step, DCM was exchanged into ACN. Samples were analyzed with a GC (Thermo Trace GC 2000, Thermo Scientific) coupled to a mass spectrometer (Polaris Q, Thermo Finnigan) equipped with a TR-5MS capillary column (30 m × 0.25 mm i.d., Thermo Scientific) with helium as carrier gas. The Polaris Q ion trap mass spectrometer was operating in the electron impact mode (70 eV). Aliquots of 1 μL were injected in split mode (split ratio 50) at 280 °C. The following temperature program was used initial temperature, 65 °C, held for 2 min; ramped to 300 °C at 6 °C min−1; and held at this temperature for 20 min. Standard mixtures were used to calibration (EPA 525 Semivolatiles Calibration Mix without Pesticides (25 components), 1000 μg mL−1 in acetone, Supelco; MA EPH aliphatic hydrocarbon standard (14 components), 1000 μg mL−1 in hexane, Restek).

Emission factor calculation

The individual emission rates were calculated according to the law of conservation of mass (Eq. 1).
$$ \frac{d{C}_{\mathrm{in}}(t)}{dt}=\frac{E_{\mathrm{R}}}{V}-\left({k}_{\mathrm{AER}}+k\hbox{'}\right)\times {C}_{\mathrm{in}}(t) $$
(1)
where Cin (μg m−3) is the indoor VOC/SVOC concentration, t (h) is the time, ER (μg h−1) is the mean emission rate, V (m3) is the volume of the test chamber (V = 0.916 m3), kAER (h−1) is the air exchange rate, and k’ (h−1) is the kinetic rate constant including all other additional loss mechanisms potentially involved, like chemical degradation or photolysis. Assuming that all experiments are carried out in the dark that air flows are consistently controlled (i.e., [O3] = 0) and that the material composition of the test chamber was chosen in a way to minimize the sink effects on the inner walls, k’ is considered to be negligible, and then set to 0.
Assuming that the emission rate ER is obtained according to the Eq. (2):
$$ {E}_{\mathrm{R}}={E}_{\mathrm{F}}\times {v}_{\mathrm{combustion}} $$
(2)
where EF (μg g−1) is the emission factor of the compound under study and vcombustion (g h−1) is the mean incense burning rate during sampling.
In a dynamic equilibrium and considering Eq. (1) and (2), the emission factor EF can be calculated as a specific emission rate normalized to the incense burning rate and can be written as follows (Eq. 3):
$$ {E}_{\mathrm{F}}=\frac{k_{\mathrm{AER}}\times V\times {C}_{\mathrm{in}}}{v_{\mathrm{combustion}}} $$
(3)

Chemometric study

Temperature (T), RH and AER could influence the VOC/SVOC emissions during the incense combustion. These three parameters cannot be considered separately. A chemometric approach using experimental design represents one possible strategy to provide relevant information with a limited number of experiments.

Factors considered for the study and experimental domain

With a classical two-level full factorial design of eight experiments (23), it is possible to study three experimental parameters (Ui), i.e., T, RH and AER, and their interactions. The levels (Xi) of the three parameters, which represent realistic extreme indoor conditions, are presented in Table 1. The observed response (Y) is the emission factor of each emitted compounds expressed in microgram per gram of burned incense for carbonyls and MAHs and nanogram per gram of burned incense for PAHs and phthalates.
Table 1

Parameters and experimental domain

Parameter

Level (Xi)

Low (−1)

High (+1)

U1

Temperature (°C)

15

35

U2

RH (%)

40

70

U3

AER (h−1)

0.25

1.5

The low temperature level was chosen based on indoor thermal comfort and human health criteria. Several studies have shown that between 15 and 25 °C, the mortality rate was minimal (Ballester et al. 2003; Rudge and Gilchrist 2007). The high level was chosen to include summer conditions.

The relative humidity is also an important parameter for indoor comfort, which can provoke the proliferation of dust mites and mold with adverse human effects (Baughman and Arens 1996; Le Bail 2010). According to the French Indoor Air Quality Observatory, the optimal indoor RH ranges between 40 and 70 %. Accordingly, these two extreme values were selected for the study.

In order to include a majority of dwelling conditions, the lowest and the highest AER were set at 0.25 and 1.5 h−1, respectively. Indeed, a large range of ventilation conditions is generally observed because AER depends on many parameters such as the intrinsic characteristics of dwellings, the nature of air conditioning, or the resident activities (Lucas et al. 2009; Sundell et al. 2011; Dimitroulopoulou 2012).

Model and strategy used

The synergic model used was:
$$ \mathrm{Y} = {\mathrm{b}}_0 + {\mathrm{b}}_1{\mathrm{X}}_1 + {\mathrm{b}}_2{\mathrm{X}}_2 + {\mathrm{b}}_3{\mathrm{X}}_3 + {\mathrm{b}}_{12}{\mathrm{X}}_1{\mathrm{X}}_2 + {\mathrm{b}}_{13}{\mathrm{X}}_1{\mathrm{X}}_3 + {\mathrm{b}}_{23}{\mathrm{X}}_2{\mathrm{X}}_3 + {\mathrm{b}}_{123}{\mathrm{X}}_1{\mathrm{X}}_2{\mathrm{X}}_3 $$
(4)
Where bi coefficients are the effects of parameters Xi on the observed response Y (here, the emission factor of each emitted compounds), and bij and bijk coefficients are the interactions between these parameters.

To estimate the reproducibility of experiments, each extreme condition was tested three times. According to this margin of error and before proceeding with the prioritization of the coefficients, a statistical analysis was carried out to estimate the probability that the coefficient value is significant (α = 95 %).

Results and discussion

Quantified compounds and emission factors

A total of 13 compounds were quantified in the incense smoke samples with formaldehyde, acetaldehyde, propionaldehyde, acetone (carbonyls) and benzene, toluene, ethylbenzene (MAHs) in the gas phase, and benzo(a)pyrene, benzo(a)anthracene, chrysene (PAHs) and diethyl phthalate (DEP), diisobutyl phthalate (DIBP), diethylhexyl phthalate (DEHP) (phthalates) in the particle phase.

Minimum and maximum individual emission factors (μg g−1 or ng g−1 of burned incense) are summarized in Table 2. Minimum and maximum values show large gaps (greater than 100 for some phthalates) pointing out a large influence of the chosen parameters. Formaldehyde and benzene, which are carcinogenic (category 1, human carcinogen, WHO IARC Monograph Working Group 2012), are among the compounds having the highest emission factors.
Table 2

Emission factors of compounds emitted during burning experiments (n.d., not detected)

Compounds in gas phase

Emission factor (μg g−1)

Minimum

Maximum

Carbonyls

Formaldehyde

772

2550

Acetaldehyde

390

1150

Acetone

184

668

Propionaldehyde

48

158

MAHs

Benzene

654

1826

Toluene

366

1194

Ethylbenzene

112

1156

Compounds in particle phase

Emission factor (ng g−1)

Minimum

Maximum

Phthalates

Diethyl phthalate (DEP)

88

11,240

Diisobutyl phthalate (DIBP)

354

4274

Diethylhexyl phthalate (DEHP)

1206

31,080

PAHs

Benzo(a)pyrene

n.d.

766

Benzo(a)anthracene

n.d.

938

Chrysene

n.d.

718

Among the compounds quantified in the gas phase, the emission factors measured (Table 2) are consistent with the values presented in literature and corresponding to different conditions of temperature, relative humidity, and air exchange rate (Lee and Lin 1996; Eggert and Hansen 2004; Lee and Wang 2004; Yang et al. 2007). For the particulate compounds, a comparison with literature is more difficult because only very few data are available. To our knowledge, only Yang et al. (2013) have reported emission factors for some PAHs. For the compounds investigated in the present work, the authors reported emission factors at 28.9 °C ranging between 162–186 ng g−1, 391–427 ng g−1, and 406–438 ng g−1 for benzo(a)pyrene, benzo(a)anthracene and chrysene, respectively. These values are of the same order of magnitude as those measured in our study (Table 2).

To determine the relative importance of the main parameters influencing the emission factors, the experimental results were statistically processed according to a classical two-level full factorial design.

Full factorial design modeling

Each experimental design provides eight (23) results (Y) based on three parameters (T, RH, and AER). According to the experimental conditions, the results obtained for each compound can be represented, in a first approach, by a cubic plot. However, it is not possible to present the individual results for each molecule. Because of the similarity of the behavior of the compounds within the four classes of compounds (carbonyls, MAHs, phthalates, and PAHs), only the individual results of formaldehyde, benzene, DIBP, and chrysene will be presented (Fig. 2). Each corner represents the average emission factor (average of three replicates) for experimental conditions (T/RH/AER) under study. In the cube, the opposite faces make it possible to compare the individual influence of each parameter and to identify the experimental conditions the most (or the least) favorable in terms of emission factor and consequently in terms of indoor air quality.
Fig. 2

Cubical factor space (in gray, the highest emission factors obtained for the most influent parameter)

AER appears to be the most influent parameter for formaldehyde, benzene, and DIBP emissions (minimum for the lowest AER and maximum for the highest AER), whereas temperature is the most influent parameter for chrysene only (maximum for the lowest temperature and minimum for the highest temperature).

The determination of bi coefficients associated to their probability of significance (Tables 3) allows considering the main effects (b1, b2, and b3) which are displayable by comparing opposite faces in the cubic plot (Fig. 1). However, there are also bij and bijk coefficients associated to two-way interactions (b12, b23, and b13) and three-way interaction (b123) between parameters.
Table 3

Coefficient bi value and their significance probability

 

Formaldehyde

Acetaldehyde

Acetone

Propionaldehyde

Benzene

Toluene

Ethylbenzene

DEP

DIBP

DEHP

Benzo(a)pyrene

Benzo(a)anthracene

Chrysene

 

Coefficient value

Significance probability (%)

Coefficient value

Significance probability (%)

Coefficient value

Significance probability (%)

Coefficient value

Significance probability (%)

Coefficient value

Coefficient probability

Coefficient value

Significance probability (%)

Coefficient value

Significance probability (%)

Coefficient value

Significance probability (%)

Coefficient value

Significance probability (%)

Coefficient value

Significance probability (%)

Coefficient value

Significance probability (%)

Coefficient value

Significance probability (%)

Coefficient value

Significance probability (%)

b0

1542.4

732.4

415.2

88.2

1158.4

772.4

347.2

4845

1584.8

8376.4

159.8

201.2

170.2

b1

−21.2

51.8

−3.6

15.5

38.4

99.8

10.4

100

−47.2

93.2

−78.2

100

−66.6

96.6

−1002.2

98.1

180

88.2

4221.1

100

−85.0

99.4

−98.7

97.9

−110.6

99.9

b2

−16.4

41.8

23.6

78.1

56.4

100

1.8

52.9

123.4

100

85

100

41.2

82.9

−761.8

93.5

166.8

85.4

2203.1

98.5

21

65.9

43.4

72.2

7.8

21.1

b3

569.4

100

237.4

100

102.8

100

36

100

289.8

100

224.8

100

183.2

100

2238.8

100

956.2

100

2613.9

99.5

35.9

80.5

42.6

71.3

27

64

b12

−127.0

99.9

−20.2

71

10

66.1

−0.8

26.8

16.2

48.9

50.4

99.9

3.6

9.6

−843.1

95.6

20.8

14.9

1401.8

89.6

−28.4

70

−23.3

44.5

−14.6

38.5

b13

−3.8

10.4

−28.6

85.9

−15.2

84.8

6.6

98.8

−62.0

98

−63.4

100

−64.4

96

−753.9

93.2

93

59.3

1072

79.4

−44.3

88.6

−61.2

86.7

77.2

98.4

b23

22.8

55.3

0.4

1.8

22.2

95.6

−7.4

99.4

78.6

99.5

31.2

97.2

36

77.3

−1048.6

98.5

129.8

74.8

1331.1

87.9

34.8

79.2

32.2

58.3

7.2

19.3

b123

−33.4

72.8

−0.6

2.5

13.4

79.5

−3.0

80

25.4

69.1

80.6

99.4

−0.8

2

−354.8

63

102.4

63.8

1238.3

85.3

−35.1

79.5

−1.5

3.1

−6.6

18

Significant factors (>95 %) are in italics

For example, for formaldehyde, the emission factors are significantly influenced both by the main effect AER (b3) and the two-way interaction T/RH (b12). The effect of a factor can be defined as an average response (the coefficient value) when the factor changes from the lowest to the highest level. As a result, the model predicts that when the AER changes from 0.25 to 1.5 h−1 (from low to high level), the emitted mass of formaldehyde per gram of burned incense increases, on average, of 1138.8 μg (corresponding to two times the effect of the main factors b3, Tables 3).

Interpretation of the main effects

Statistical analysis of the results shows that the emission factors of the 13 compounds quantified are influenced by at least one main effect. In general, AER is the most important main effect affecting emissions of carbonyls, MAHs, and phthalates. The positive influence indicates that the emitted mass increases significantly with the increase of AER (from 0.25 to 1.5 h−1). For PAHs, temperature is the most important parameter that governs their emissions. The negative influence means that the emitted mass increases significantly when the temperature decreases (from 35 to 15 °C). These statistical results confirm the initial findings from the cubic plots (Fig. 1).

Influence of air exchange rate

For all of the 13 compounds quantified, the increase of AER from 0.25 to 1.5 h−1 (coefficients b3) induces systematically an increase of the mass emitted, even though PAHs coefficients are not significant. Formaldehyde and benzene are the compounds the most affected by the increase of AER, with a mean increase of 1138.8 and 597.6 μg g−1 of burned incense, respectively. Yang et al. (2005 and 2007) explain the influence of AER on the emission rate of the studied VOC/SVOC by a change in the combustion process, higher AER leading to a decrease of the combustion efficiency. Because emissions are reported as a function of the mass of incense burned (emission factor), the increase of the emission rate cannot be induced by a faster burning of incense due to a higher air flow rate (so AER in our test chamber). Yang et al. (2005) suggest that the additional supply of oxygen molecules induced by the increase of air flow rate to the incense burning tip does not offset the decrease of flame temperature, thus resulting in a poor combustion. These findings were confirmed by a study of Derudi et al. (2013), dedicated to a review of the several origins identified for the emission of compounds (mainly PAH), according to burning characteristics: on the one hand, vaporization and pyrolysis phenomena that occur at relatively low temperature, and on the another hand, production by an incomplete combustion that occurs at high temperature. Review of the available data suggests that products from combustion of incense are mainly emitted by high temperature processes, i.e., produced by an incomplete combustion. Such observations combined with the present results may illustrate that the higher the AER, the less efficient the combustion, and the higher the mass of compounds emitted.

Influence of temperature

It was expected that the temperature of the test chamber is negligible compared to the one of the combustion zone (incense burning tip). For some compounds (formaldehyde, acetaldehyde, benzene, and DIBP), the temperature coefficient (b1) is indeed not significant. For the other compounds, temperature dependence is sometimes observed but the effects depend on the compounds under study. The lower the temperature (here low level at 15 °C), the higher the emission factor of all PAHs, DEP, toluene, and ethylbenzene. Conversely, the emission factors are at their highest level for acetone, propionaldehyde, and DEHP when the temperature is warmer (35 °C).

It should be noted here that PAHs and phthalates are quantified only in the particle-phase while these compounds are semi-volatile. This could explain the negative effect of temperature on the emission rates of PAHs and phthalates (except for DEHP). Indeed, the vapor pressure of these compounds increases exponentially with the temperature. As a result, a higher temperature would shift the gas-particle equilibrium of these SVOCs towards the gas phase. Based on this potential artifact, the experimental system does not seem appropriate for capturing the influence of temperature on SVOC (here, PAHs and phthalates) emission factors.

Influence of relative humidity

RH has not been found to significantly affect the emission factors for the compounds under study. A positive influence (from 30 to 70 %) is only noted for acetone, benzene, toluene, and DEHP. This influence is still small compared to the influence of AER because the coefficients b2 (RH) are lower than the coefficients b3 (AER).

Lin and Wang (1994) concluded to a decrease in the emission factors of aliphatic aldehydes with increasing humidity. However, the lack of the temperature control during their combustion experiments makes it difficult to draw definitive conclusions from this study. Yang et al. (2013) estimated the effect of RH on PAH emissions both in gas and particle phase. They concluded that the burning incense at high RH may minimize PAH emissions. However, according to their results, the emission factors of benzo(a)pyrene, benzo(a)anthracene, and chrysene (PAHs) in particle phase were not significantly different when RH rises from 21.3 to 90.5 %, at a constant temperature of 28.9 °C (Yang et al. 2013). The same behavior is observed in the present work for RH ranging between 40 and 70 %.

Interpretation of interaction effects

None of the compounds under study had significant three-way interaction (T/RH/AER). Considering the 13 quantified compounds, the two-way interactions (T/RH, T/AER, and RH/AER) are observed at least once. As a reminder, a two-way interaction occurs whenever the effect of one independent variable influences the level of the other.

For most of the compounds under study, the value of the bij coefficients is close to the limit of significance, highlighting their small influence. However, among these interactions, the two-way interaction between T and AER (b13) is the most frequently significant. All MAHs and chrysene followed the same trend. Statistically, the highest emission factors are obtained under the experimental condition that combines the lowest temperature (15 °C) with the highest AER (1.5 h−1) (Fig. 3). For the four considered compounds, this interaction is a positive interaction, which reflects the addition of both T and AER individual positive effects on the emission factors, as preliminary calculated. More among the other two-way interactions, no crossover interaction occurs.
Fig. 3

Two-way interaction T/AER (in gray, the highest significant emission rate)

Indoor implications

Based on the determined individual emission factors (EF, μg g−1 or ng g−1), eight different application scenarios are possible to evaluate the influence of experimental parameters (i.e., T, RH, and AER) on the concentration of compounds under study. Assuming that ventilation is the only removal mechanism occurring in the air compartment of the test chamber, the concentration profile during the incense burning is governed by the following Eq. (4) from Eq. (1):
$$ {C}_{\mathrm{in}}(t)=\frac{E_{\mathrm{R}}}{k_{\mathrm{AER}}\times V}\times \left(1-{e}^{-{k}_{\mathrm{AER}}.t}\right) $$
(5)
where t (h) is the time, Cin(t) (μg m−3 for carbonyls and MAHs, or ng m−3 for phthalates and PAHs) is the concentration of compounds under study in the room, kAER (h−1) is the air exchange rate constant, ER (μg h−1 or ng h−1) is the mass emission rate of the compound under study, and V (m3) is the volume of the room.
Considering that the typical incense stick mass is 500 mg, the combustion time is 30 min corresponding to an incense burning rate (vcombustion) at 1 g h−1. According to the technical specification of European Committee for Standardization (CEN/TS 16516, 2013) for the determination of emissions into indoor air, the room volume (V) is fixed to 30 m3. Focusing for example on benzene results, the lowest and highest emission rates obtained are equal to 800(±128) μg h−1 and 1718(±94) μg h−1, respectively. These data appear very consistent with the previous value of 937(±5) μg h−1 obtained in a full-scale experimental room (Manoukian et al. 2013), with the same batch of incense sticks, and in typical residential indoor conditions (i.e., 16.2 °C, 50 % RH, and 0.8 h−1). Equation (4) allows modeling the concentration profiles of benzene corresponding to the two bound values of the emission factors. The modeled peak concentrations of benzene after 30 min of incense combustion are 12.5 and 20.1 μg m−3, for the lowest and the highest emission factor, respectively (Fig. 4). Once again, these data are in good agreement with the experimental result (12.9 μg m−3) reported by Manoukian et al. (2013). The same trend is obtained for toluene (10.6 μg m−3; another compound selected as a model by Manoukian et al. (2013)) which successfully validates measurements carried out within the test chamber.
Fig. 4

Concentration profiles of benzene for one incense burning

As a result, thanks to the measurement of emission factor in different indoor conditions related to experimental parameter ranges, the peak concentration ranges for the compounds under study can be calculated (Table 4). Based on those results, and considering typical conditions (i.e., temperature, relative humidity, air exchange rate, volume of the room…), incense emissions of carcinogenic substances (i.e., benzene, benzo(a)pyrene, and formaldehyde) could be compared to guideline exposure threshold values. For benzene and benzo(a)pyrene, no safe level of exposure can be recommended for indoor air quality (WHO–World Health Organization 2010). In ambient air, limit concentrations of airborne benzene and benzo(a)pyrene are 5 μg m−3 and 1 ng m−3 for annual average, respectively (European Union 2000, 2004). For formaldehyde, the threshold value is 100 μg m−3 for any 30-min period (WHO–World Health Organization 2010). The maximum concentration of formaldehyde emitted by the total combustion of one incense stick calculated in this work remains well below the exposure threshold value, while benzene and benzo(a)pyrene values are close to or higher than the corresponding air quality standards. Nevertheless, estimated concentrations are instantaneous values and it must be underlined that limit standard values are related with time of exposure. Of course additional risk assessment studies are needed to clearly determine the toxicological significance of incense emissions of carcinogenic substances.
Table 4

Modeled concentrations for the lower and the higher emission factor measured (mincense = 500 mg; combustion time = 30 min; room volume = 30 m3)

  

T (°C)

RH (%)

AER (h−1)

 

Compounds

Emission factor (μg g−1)

Experimental parameters

Peak concentration (after 30 min burning) (μg m−3)

T (°C)

RH (%)

AER (h−1)

Formaldehyde

822 ± 66

Lower

35

70

0.25

12.9 ± 1.0

2304 ± 226

Higher

15

70

1.5

27.0 ± 2.6

Acetaldehyde

418 ± 34

Lower

15

40

0.25

6.7 ± 0.5

1046 ± 104

Higher

15

70

1.5

12.3 ± 1.2

Acetone

222 ± 34

Lower

15

40

0.25

3.5 ± 0.5

644 ± 28

Higher

35

70

1.5

7.6 ± 0.3

Propionaldehyde

42 ± 4

Lower

15

40

0.25

0.7 ± 0.1

150 ± 6

Higher

35

40

1.5

1.8 ± 0.1

Benzene

800 ± 128

Lower

15

40

0.25

12.5 ± 2.0

1718 ± 94

Higher

15

70

1.5

20.1 ± 1.1

Toluene

470 ± 48

Lower

35

40

0.25

7.4 ± 0.8

1164 ± 28

Higher

75

70

1.5

13.6 ± 0.3

Ethylbenzene

152 ± 42

Lower

35

40

0.25

2.4 ± 0.7

736 ± 376

Higher

15

70

1.5

8.6 ± 4.4

Compounds

Emission factor (ng g−1)

Experimental parameters

Peak concentration (after 30 min burning) (ng m−3)

T (°C)

RH (%)

AER (h−1)

DEP

2080 ± 1352

Lower

15

40

0.25

32.6 ± 21.2

9452 ± 2986

Higher

15

40

1.5

110.8 ± 35.0

DIBP

424 ± 90

Lower

15

40

0.25

6.6 ± 1.4

3234 ± 1034

Higher

35

70

1.5

37.9 ± 12.1

DEHP

1904 ± 610

Lower

15

40

0.25

29.8 ± 9.5

22458 ± 8830

Higher

35

70

1.5

263.3 ± 103.5

Benzo(a)pyrene

58 ± 8

Lower

35

70

1.5

0.7 ± 0.1

444 ± 296

Higher

15

70

1.5

5.2 ± 3.5

Benzo(a)anthracene

34 ± 22

Lower

35

40

1.5

0.4 ± 0.3

504 ± 394

Higher

15

70

1.5

5.9 ± 4.6

Chrysene

2 ± 2

Lower

35

70

1.5

0.02 ± 0.02

422 ± 260

Higher

15

70

1.5

4.9 ± 3.0

Moreover, according to Table 4 and Fig. 4, for carbonyls, MAHs and phthalates, the higher the AER, the higher the emission factor, and the higher the peak concentration at the end of one incense stick burning (30 min). Several studies have investigated the relationship between ventilation rate, emission factor, and indoor concentration of VOCs (Sherman and Hodgson 2004; Gilbert et al. 2008; Lin et al. 2009). A positive correlation between AER and VOCs emission factor was also observed while, in the same time, a negative relationship between the VOCs mean concentrations and AER was obtained. This former result seems to contradict our observations (Fig. 4). Nevertheless, in the present work, if several incense sticks are used in series (i.e., an incense is lighted as soon as the previous is off), an opposite trend is observed between concentration profiles of the lowest and the highest emission factor measured. For benzene, after 1.55 h of continuous combustion (e.g., use about four incense sticks successively), the concentration profile calculated using the lowest emission factor (and the lowest AER) becomes higher than the one calculated using the highest emission factor (and the highest AER) (Fig. 5). This result highlights two opposite consequences of the increase of ventilation: (i) a change in the burning process of the incense that promotes VOC/SVOC emissions (i.e., the highest emission factor), (ii) a more rapid dilution of emitted compounds induced by the higher quantity of fresh air supplied to the room, which then improves the reduction of VOC/SVOC concentrations. As illustrated for benzene in Fig. 5, the first process prevails during the early phase of the combustion, until a critical equilibrium concentration is reached (occurring at 1:55 h in this work). Over this threshold, the trend is reversed and the elimination process of pollutants becomes dominant, compared to the emission efficiency.
Fig. 5

Concentration profiles of benzene for low and high EF conditions

Conclusion

This work highlights through a statistical study of the influence of three indoor parameters (i.e., temperature, relative humidity, and air exchange rate) on 13 compounds emitted during incense burning in an experimental test chamber, the predominant effect of ventilation on emission factors. The higher the ventilation, the higher the emission factor. Moreover, thanks to those experiments, an estimation of indoor pollution levels induced by incense combustion can be calculated. Results show that some of the carcinogenic substances studied, namely benzene and benzo(a)pyrene produced by incense emissions, reach instantaneous concentration values close to or higher than the corresponding air quality standards defined for specific time of exposure.

Acknowledgments

The AMBISAFE project is labialized by competitiveness PASS cluster and was funded by a research consortium (Albhades Provence, APF arômes et parfums, Bougie & Senteur, L’Occitane en Provence, TERA Environnement, Terre d’Oc). Audrey Manoukian is grateful to the Région Provence-Alpes-Côte-d’Azur and the Albhades Provence laboratory for the doctoral grant obtained to carry out this study.

Copyright information

© Springer-Verlag Berlin Heidelberg 2015

Authors and Affiliations

  • A. Manoukian
    • 1
  • D. Buiron
    • 1
  • B. Temime-Roussel
    • 1
  • H. Wortham
    • 1
  • E. Quivet
    • 1
  1. 1.Aix Marseille UniversitéCNRS, LCE UMR 7376MarseilleFrance

Personalised recommendations