1. Introduction

Most people in modern societies spend approximately 90% of their time indoors on a daily basis ^[1]. Thus, indoor air quality has become a major concern. In this regard, the car indoor environment has attracted considerable attention because cars have become the main means of transportation in our society ^[2,3,4]. Overall, people spend 5% of their time in cars on a daily basis, which proportionally follows the times spent in homes and offices ^[5].

The indoor air in a car cabin is polluted by organic compounds, especially volatile organic compounds (VOCs) and aldehydes ^[3,6]. Several studies have revealed that vehicle exhaust gases could be significant sources of VOCs such as benzene, toluene, and xylenes ^[7,8,9,10]. In addition to these gases, Rahman and Kim ^[11] found that exhaust gases diffusing back into a car cabin, or back diffusion gases, polluted the air in a car cabin and contained high concentrations of toluene, formaldehyde, acetaldehyde, propionaldehyde, and butylaldehyde. Moreover, organic compounds emission from car materials such as carpets, paints, leather, plastics, and foam in the seats are air pollutants in car cabins ^[12,13,14]. In addition to these various pollutants, the nature of car construction, such as high surface area and small volume, easily leads to elevated concentrations of organic compounds emitted from car interiors. Because most VOCs are toxic to humans when inhaled at sufficient concentrations, high concentrations of these pollutants may cause adverse health effects to passengers ^[2,15].

Compared with the indoor environment of dwellings, only limited studies have addressed the indoor environment of car cabins. Zhang et al. ^[4] reported that under static conditions, emissions from the interior materials of car cabins contribute significantly to air pollution in many compact car models sold in China. In their study, however, the number of air pollutants measured was limited to formaldehyde, benzene, toluene, and xylenes. Because many kinds of VOCs exist in car cabins, a more comprehensive quantification of the organic compounds is needed. Conversely, Yoshida and Matsunaga ^[16] measured 162 organic compounds including many aliphatic and aromatic hydrocarbons in a new privately used car. Interestingly, high concentrations of n-nonane (458 μg/m³), n-decane (1301 μg/m³), n-undecane (1616 μg/m³), n-dodecane (716 μg/m³), n-tridecane (320 μg/m³), 1-hexadecene (768 μg/m³), ethylbenzene (361 μg/m³), xylene (4003 μg/m³), and 2, 2’-azobis (isobutyronitrile) (429 μg/m³) were also detected. The total volatile organic compound (TVOC) concentration approximated 14, 000 μg/m³, which significantly exceeded the advisable value established by the Ministry of Health, Labour and Welfare of Japan (400 μg/m³). Chen et al. ^[2] determined the concentrations of seven VOC species including benzene, toluene, ethylbenzene, xylenes, styrene, butyl acetate, and undecane and TVOC in 38 taxis in China. Benzene, toluene, ethylbenzene, xylene, styrene, butyl acetate, undecane, and TVOC concentrations in taxi cabins averaged 82.7, 212.3, 74.7, 182.3, 24.7, 33.5, 61.3, and 1441.7 μg/m³, respectively. These findings indicated that in-car airborne VOC and TVOC pollution may induce healthproblems in passengers and drivers.

In this study, a comprehensive measurement of VOCs in car interiors was carefully conducted for 24 unoccupied cars. The number of target organic compounds amounted to 55. The effects of vehicle age on organic compound concentrations were also discussed. Finally, a screening-level risk assessment was also performed to evaluate organic compound concentrations in car cabins.

---

2. Materials and Method

---

2.1. Air sampling and analytical methods

Air samples were collected from 24 unoccupied cars (n = 24, Table 1). Sampling was conducted at several outdoor parking lots in Yokohama and Kawagoe, Japan from November 17 to 27, 2013. Outdoor temperatures during the sampling ranged from 14.3 to 23.0 °C while indoor temperatures spanned from 13.7 to 25.8 °C. During the sampling period, weather conditions were mostly cloudy. Only one driver of the investigated cars smoked regularly (Car 9). During sampling, an air freshener was found in Cars 2, 9, and 11.

Most drivers spend 5% of their day in their cars ^[5]. Taking into account round trips, this corresponds to about 30 min in a single private-use car on a usual day for routine activities, such as traveling from home to workplaces, shopping, and returning home. To reflect these driving habits, VOC and aldehyde emissions were measured for 30 min. Before air sampling, the engine was turned off and the car cabin was ventilated for 10 min by opening the doors. Next, all the doors and windows were closed. Sampling devices for VOCs, TVOC, and aldehydes were positioned near the driver’s headrest to sample the air around the driver’s breathing zone.

---

2.1.1. Aldehyde sampling and analysis

Aldehydes were sampled by collecting the interior air of the vehicles using a 2, 4-dinitrophenylhydrazine (DNPH) cartridge (InertSep mini AERO DNPH, GL Sciences Inc., Tokyo, Japan) incorporating an ozone scrubber (InertSep mini Ozone Scrubber, GL Sciences Inc., Tokyo, Japan). A portable pump (MP-Σ300NII, Sibata Scientific Technology Ltd., Saitama, Japan) was turned on for 30 min at an air flow rate of 1 L/min. Subsequently, the samples were extracted from the cartridges using acetonitrile (5 mL) and analyzed by high-performance liquid chromatography (HPLC) using an Agilent 1100 apparatus (Agilent Technologies Inc., Santa Clara, CA, USA) equipped with a column (Atlantis T3, 100 mm × 4.5 mm i.d., film thickness: 3 μm, Waters Corporation, Milford, MA, USA) and a UV detector (360 nm). The HPLC analysis was performed using 1:1 acetonitrile/water (v/v) as an isocratic mobile phase at a flow rate of 0.8 mL/min, an injection volume of 20 μL, and a column temperature of 40 °C.

---

2.1.2. VOC sampling and analysis

The interior air of the vehicles was collected for 2 min in 10 L Flek polyester bags (Flek-Sampler, Omi Odor-Air Service Corporation, Shiga, Japan) using a hand-operated air pump (DC1-NA, Omi Odor-Air Service Corporation, Shiga, Japan) at an air flow rate of 1 L/min from 28min after starting aldehyde sampling. Air samples were analyzed using two gas chromatographs equipped with a flame ionization detector (GC-FID) (HP5890A, Hewlett-Packard, San Fernando, CA, USA). Specifically, highly polar VOCs were characterized using a GC-FID equipped with an HP-1 column (15 m × 0.32 mm i.d., film thickness: 5 μm, Hewlett-Packard, San Fernando, CA, USA). Their low-polarity counterparts were analyzed using a GC-FID fitted with an HP-5 column (15m×0.32 mm i.d., film thickness: 1.05 μm, Hewlett-Packard, San Fernando, CA, USA). For all compounds, the GC-FID oven temperature was initially held at 35 °C for 5 min and increased to 180 °C at a rate of 20 °C/min, 200 °C at a rate of 2 °C/min, and finally 220 °C at a rate of 10 °C/min. The injector and detector temperatures were 250 °C. Helium (He) acted as a carrier gas. The air samples were injected to GC-FID by an auto injection system.

---

2.1.3. TVOC, temperature, and relative humidity measurements

The TVOC concentrations in the interior air of the vehicles were measured using a TVOC meter (FTVR-01, Figaro Engineering Inc., Osaka, Japan) for concentrations ranging from 0 to 9999 μg/m³. The limit of detection (LOD) and resolution amounted to 1 μg/m³. Temperatures were also determined using the TVOC meter.

Hori et al. ^[17] related values obtained using the same TVOC meter and data acquired by active sampling by measuring VOC concentrations in different rooms through these approaches. They found a good correlation between both sets of measurements regardless of differences in VOC compositions between individual samples. They also found that the detection of VOCs with ultralow or ultrahigh sensitivity using the TVOC meter provided values that may differ from actual TVOC concentrations. However, they concluded that the TVOC meter could be used to monitor TVOC concentrations without considering individual VOCs. Therefore, this instrument was chosen to measure TVOC concentrations in car cabins.

---

2.2. Quality assurance/quality control

The external calibration method and the linear regression model were used for the calibration for VOCs and acetaldehydes. The calibration ranges of acetaldehyde, benzaldehyde, and formaldehyde were 2-330, 33-330, and 2-330 μg/m³, respectively. The values of R² for acetaldehyde, benzaldehyde, and formaldehyde were more than 0.999. The LOD for VOCs was calculated based on a signal-to-noise ratio (S/N) of 3:1, while the limit of quantification (LOQ) was determined to be 10 times the LOD. In the case of aldehydes, LOD and LOQ were defined as 3σ and 10σ of blank level samples (n = 5), respectively. The method detection limit (MDL) and method quantification limit (MQL) were calculated using LOD, LOQ, extract volumes, and air sampling volume. Obtained MDL and MQL for aldehydes and VOCs are listed in Table 2.

Before each VOC sampling round, sampling bags were flushed twice with ultrapure nitrogen. In the preliminary accuracy tests, the coefficients of variation for VOCs, the ratio of the standard deviation to the mean, were observed to be well below 5%. Slightly higher coefficients of variation (5.5-6.0%) were observed for 1, 2, 4-trimethyl benzene and 1, 3, 5-trimethyl benzene. These coefficients of variation amounted to 1.3% and 2.6% for formaldehyde and acetaldehyde, respectively. Recovery efficiencies reached 101% and 93% for spiked formaldehyde (3.8 μg) and acetaldehyde (7.6 μg), respectively. In blank DNPH cartridges, all aldehyde concentrations measured in this study were below MDL. A breakthrough test ensured that no breakthrough would occur from DNPH cartridges during the experimental procedure. The air flow rates of pumps used for air sampling were calibrated before sampling and measured after sampling to ensure that they were unchanged during the sampling. The TVOC meter was calibrated daily using the air passing through the activated carbon filter as a zero gas.

---

2.3. Statistical analysis

All statistical analyses were performed using Excel (Microsoft Office 2013). To include all data, values that were non-detected (N.D.) or below MQL were replaced by half MDL and half MQL in statistical calculations, respectively ^[5,18,19,20]. Non-parametric tests were used because of limited sample numbers. The Spearman rank order correlation coefficient was calculated to correlate VOC, aldehyde, and TVOC concentrations; car age; and interior temperature. The criterion for significance was p < 0.01 or p < 0.05.

---

3. Results and discussion

---

3.1. Detection frequency of organic compounds in car cabins

Figure 1 shows the detection frequencies of organic compounds in the 24 investigated car cabins. Acetaldehyde (3.81-36.0 μg/m³), formaldehyde (3.26-26.7 μg/m³), n-tetradecane (< MQL to 47.7 μg/m³), toluene (4.23-78.3 μg/m³), and n-undecane (< MQL to 6.24 μg/m³) were detected in all car cabins. In addition, dichrolomethane (N.D. to 69.4 μg/m³), n-dodecane (N.D. to 8.45 μg/m³), ethylbenzene (N.D. to 8.10 μg/m³), 2-ethylhexanal (N.D. to 6.85 μg/m³), n-hexane (N.D. to 22.7μg/m³), methylcyclohexane (N.D. to 7.87 μg/m³), methylcyclopentane (N.D. to 13.5 μg/m³), 1-methyl-2-ethylbenzene (N.D. to 31.7 μg/m³), 1-methyl-3-ethylbenzene (N.D. to 20.6 μg/m³), 1-methyl-4-ethylbenzene (N.D. to 83.9 μg/m³), methylethylketone (N.D. to 13.2 mg/m³), n-octanal (N.D. to 29.9 μg/m³), n-octane (N.D. to 21.4 mg/m³), 1-propanol (N.D. to 15.1 μg/m³), m, p-xylene (N.D. to 17.4 mg/m³), and o-xylene (N.D. to 5.27 μg/m³) were observed in more than 80% of the cars. These high detection frequencies implied that the measured organic compounds mainly originated from car structural materials. In contrast, benzaldehyde (MDL = 8.56 μg/m³) was not detected in any car cabin.

---

3.2. Organic compound concentrations in car cabins

Organic compound and TVOC concentrations in the 24 car cabins are listed in Table 2 (see individual data in Table S1, Supporting Information). An unidentifiable broad peak was detected near 22 min corresponding to the retention time of benzene for Car 18, preventing the quantification of benzene in this vehicle.

Organic compound concentrations ranged from N.D. to 342 μg/m³. 3-methylpentane observed in Car 8 exhibited the highest concentration (342 μg/m³). However, its median concentration amounted to 1.97 μg/m³, making it the 18th most abundant organic compound. According to Zuraimi et al. ^[21], this compound may mainly originate from adhesives and solvents. Also, its emission rates highly depend on the country producing these materials. For example, materials fabricated in Europe (EU) and Singapore presented emission rates of 166.4 and 137 μg/m²/h, respectively. Tsai et al. ^[22] reported that motorcycle exhaust contained high concentrations of 3-methylpentane (1900-274, 100μg/m³). In addition to 3-methylpentane, 2-methylpropane (310 μg/m³), p-dichlorobenzene (209μg/m³), 1, 2, 3-trimethylbenzene (163 μg/m³), and n-tridecane (113 μg/m³) were also detected in high concentrations. 2-methylpropane and 1, 2, 3-trimethylbenzene are also present in exhaust gas ^[23]. p-dichlorobenzene is typically used as a moth repellent and a deodorizer ^[24,25]. n-tridecane may result from vinyl-based materials ^[26]. Toluene displayed the highest median concentration (23.5 μg/m³). This compound may originate from many kinds of emission sources, such as tobacco smoke, solvent-based paints, and consumer products ^[24]. As mentioned above, it is also the main component of car exhaust gas. Acetaldehyde (13.4 μg/m³), n-hexane (11.5 μg/m³), and formaldehyde (9.41 μg/m³) exhibited the next-highest median concentrations. In addition to their wide use as adhesives for wood products, formaldehyde and acetaldehyde are also utilized as surface coatings for furniture and floors. They are present in composite wood, products containing urea-formaldehyde resin, and tobacco smoke ^[24,27]. n-hexane, which exists in car exhaust gas, is also used as an adhesive and solvent ^[22,28].

Most median concentrations obtained in this study were lower than data reported for new cars by Yoshida and Matsunaga ^[16] and You et al. ^[29], with the exception of 1, 1, 1-trichloroethane. In general, most chemicals detected in car cabins were considered to originate from the original structural materials. New cars tend to display higher organic compound concentrations than their old counterparts. Here, the car age averaged 5.56 years, except used Cars 1 and 5, which did not have any clear information on their ages. This may largely explain the low organic compound concentrations obtained compared with literature values for new cars. On the other hand, except for ethylbenzene, most organic compounds measured in this study presented comparable median concentrations to values for two old cars reported by You et al. ^[29]. The influence of car age on these concentrations will be discussed below. Moreover, temperature differences during air sampling may impact results. Temperatures during sampling in this study ranged from 14.3 to 23.0 °C instead of remaining at 25 °C ^[29]. High temperatures enhance organic compound emission rates from materials, which is expected to increase high organic compound concentrations. This may explain why most median concentrations in this study were lower than those reported by You et al. ^[29]. The relationship between temperature and organic compound concentration will be discussed later.

---

3.3. Ratios betweenindoorandoutdoor organic compoundconcentrations

Table 3 shows the ratios between indoor and outdoor concentrations (I/O ratios) for the 55 organic compounds. Outdoor concentrations are shown in Table S2. An I/O ratio higher than 1.0 shows that the pollutant is likely inside the car cabin whereas an I/O ratio less than or equal to 1.0 indicates that the pollution source is outside the car cabin. Maximum I/O values were extremely high for most organic compounds, except for benzaldehyde. Acetaldehyde (median = 3.0), benzene (median = 3.3), 2-ethylhexanal (median = 5.1), 1, 2, 3-trimethylbenzene (median = 5.0), 1, 3, 5-trimethylbenzene (median = 10.0), and n-undecanal (median = 10.0) exhibit median I/O ratios exceeding 1.0. This implies that these compounds originated from interior sources such as back diffusion gases and adsorbed vehicle exhaust gases.

---

3.4. Correlation between organic compound concentrations

Correlation coefficients (r) between organic compounds are shown in Table S3. Two groups of compounds presented high correlation coefficients. The first group consisted of n-decane, butylacetate, and naphthalene (n-decane and butylacetate: r = 0.81, p < 0.01; naphthalene and butylacetate: r = 0.72, p < 0.01; naphthalene and n-decane: r = 0.87, p < 0.01). The other group included dichloromethane, n-nonane, and toluene (dichloromethane and n-nonane: r = 0.70, p < 0.01; dichloromethane and toluene: r = 0.84, p < 0.01; n-nonane and toluene: r = 0.78, p < 0.01). The high correlation coefficients within each group imply that their components may have common emission sources. Additionally, good correlation pairs, such as diethylether-1, 3, 5-trimethylbenzene (r = 0.72, p < 0.01), n-pentane-ethylacetate (r = 0.77, p < 0.01), n-nonane-isopropylbenzene (r = 0.71, p<0.01), n-propylbenzene-isopropylbenzene (r = 0.73, p < 0.01), n-tridecane-1-methyl-3-ethylbenzene (r = 0.71, p < 0.01), n-nonane-n-propylbenzene (r = 0.71, p<0.01), m, p-xylene-n-nonane (r = 0.70, p < 0.01), and m, p-xylene-n-propylbenzene: r = 0.71, p<0.01, were also found.

---

3.4.1. Correlation between temperature and concentrations of organic compounds

In general, a higher interior temperature leads to higher concentrations of VOCs and aldehydes. The higher interior temperature can enhance the emission rate of organic compounds from the interior materials in car cabin. However, in this study, not only significant positive correlations (formaldehyde: r = 0.58, p < 0.01; methylcyclohexane: r = 0.52, p < 0.05) but also significant negative correlations (1, 2, 3-trimethylbenzene: r = −0.57, p < 0.01; tetrahydrofuran: r = −0.43, p<0.05) were found. Parra et al. ^[30] also found the negative correlation between interior temperature and organic compounds (benzene, toluene, ethyl benzene, xylenes) in buses. The higher interior temperature may increase not only the emission rate of organic compounds but also the photochemical degradation rate, which plays an important role in reducing the concentrations of organic compounds ^[30].

---

3.4.2. Correlation between car age and organic compound concentrations

Organic compound concentrations tend to decrease with increasing car age because of long-term ventilation or emission ^[2,5]. However, no significant negative correlations were observed between car age and organic compound concentrations in this study. Li et al. ^[31] found that toluene, ethylbenzene, xylene, and trimethylbenzene concentrations decreased quickly with time and dropped by 23%, 32%, 10%, and 50% in one year, respectively. Chen et al. ^[2] also demonstrated that, in the case of benzene, toluene, ethylbenzene, xylenes, styrene, butyl acetate, and undecane, VOC and TVOC concentrations decreased rapidly when the car age increased from 1.5 to 23.9 months. However, these concentrations did not decrease significantly when the car age rose from 23.9 to 39.4 months. As a result, a significant decrease in organic compound concentrations is likely to occur for one to two-year-old cars. In this study, the average car age approximated 5.6 years. Therefore, the large number of older cars masked any important negative correlation between car age and organic compound concentrations. In fact, TVOC concentrations in Cars 13, 16, and 21, which were 0.08, 0.25, and 0.42 years old, respectively, surpassed those of the other cars.

---

3.5. TVOC concentrations in car cabins

Figure 2 shows the TVOC concentrations in the 24 car cabins along with the advisable value established by the Ministry of Health, Labour and Welfare of Japan (400 μg/m³, red line). Fourteen cars (58%) exceeded the provisional guideline value. Car 13, which had been purchased one month before testing, displayed the highest TVOC concentration (1136 μg/m³). Purchased over 10 years before sampling, Cars 11 and 15 presented TVOC concentrations that exceeded the provisional guideline value. Cars 2, 9, and 11, in which an air freshener existed during sampling, exhibited relatively high TVOC concentrations. The correlation between car age and TVOC concentration was not significant (r = −0.12, p = 0.61). However, new cars (cars younger than one year old, Car 13) showed significantly higher TVOC concentrations, consistent with the extremely high values obtained by Yoshida and Matsunaga (14,081 μg/m³) ^[16] and You et al. (4940 μg/m³) ^[29]. For new cars, these concentrations may stem mainly from the emission of organic compounds from the original car materials. On the other hand, they likely depend on the way drivers use their cars for vehicles older than one year. In other words, the driver’s habits strongly impact the car indoor environment in these older cars.

An analysis of the correlation between TVOC and individual VOC concentrations revealed moderate correlations (p values lower than 0.05) for 2-ethylhexanal (r = 0.46, p < 0.05), n-nonanal (r= 0.57, p < 0.01), n-tetradecane (r = 0.54, p < 0.01), toluene (r = 0.51, p < 0.05), n-tridecane (r=0.46, p < 0.05), and n-undecane (r = 0.44, p < 0.05) (Table S3, Supplementary materials).

---

3.6. Risk assessment for organic compounds in car cabins

A screening-level risk assessment was conducted to evaluate the health risk of drivers and passengers associated with exposure to organic compounds in cars. Exposure to organic compounds via inhalation of car indoor air was determined using:

${I_{inh}} = {{{R_{inh}} \cdot {C_{air}} \cdot {t_{car}}} \over W}$

(1)

where I_inh is the daily intake rate of organic compounds via inhalation in a car cabin [μg/kg-body weight/day], W is the body weight [kg], C_air is the concentrations of organic compounds in cabin air [μg/m³], R_inh is the daily inhalation rate [m³/day], and t_car is the exposure time in a car cabin [day].

Human exposure to organic compounds via inhalation in car cabins was assessed using two scenarios. (1) Typical exposure considered median concentrations between measured cars. (2) The worst case scenario accounted for the highest concentration. Both scenarios used the average body weight (50 kg) and inhalation rate for a Japanese person (15 m³/day) ^[32,33]. The average residence time in a car cabin was estimated as (1) typical (0.05 day; 5% of 24 h) or (2) occupational (0.33day;8 h).

Azuma et al. ^[32] determined the estimated human no observed adverse effect level (NOAEL) via inhalation exposure, NOAEL_inh [mg/m³] using a reference human body weight of 50 kg and respiration rate of 15 m³/day for Japanese people. They accounted for three uncertainty factors. (1)Uncertainty factor 1 (UF1) was applied to lowest observed adverse effect level (LOAEL) when a NOAEL was unavailable. (2) Uncertainty factor 2 (UF2) was applied to extrapolation across durations. (3) Uncertainty factor 3 (UF3) was applied to extrapolation from animal studies to a human situation. Here, reference doses (RfDs) [mg/kg-body weight/day] were calculated using NOAEL_inh as

$RfD=\frac{NOAE{{L}_{\text{ inh}}}\times 1000\times 15\text{ }{{{\text{m}}^{\text{3}}}}/{\text{day}}\;}{50\text{ kg}\times UF}$

(2)

where UF is an uncertainty factor for individual uncertainty components [-]. In this study, a default factor of 10 was used for UF.

The hazard quotient (HQ) [-] was evaluated as

$HQ = {{{I_{{\rm{inh}}}}} \over {RfD}}$

(3)

Table 4 summarizes results of the screening-level risk assessment for driver and passenger exposure to organic compounds in cars. Figure 3 shows HQ values for the organic compounds. The HQ values for formaldehyde in both exposure scenarios exceeded one for occupational use (1.3 and 3.4 for typical and worst case scenarios, respectively). For exposure scenarios involving typical use, no organic compounds showed HQ values above one. However, passengers may be exposed to these organic compounds outdoors as well as indoors. Therefore, a more accurate risk assessment requires a more comprehensive exposure assessment, such as a personal exposure assessment.

---

4. Conclusion

Fifty-five organic compounds, such as VOCs and aldehydes, were measured in 24 car cabins in November 2013 in Yokohama and Kawagoe, Japan to assess car indoor environment. Acetaldehyde, formaldehyde, n-tetradecane, toluene, and n-undecane were detected in all car cabins. TVOC concentrations observed in 14 car cabins exceeded the advisable value established by the Ministry of Health, Labour and Welfare of Japan (400 μg/m³). The highest TVOC concentration (1136 μg/m³) was obtained in a new car (1 month after purchase). However, TVOC concentrations surpassed the advisable value in cars purchased over 10 years ago. For formaldehyde, HQ values in both exposure scenarios for occupational use were higher than one, indicating the need for a more comprehensive exposure assessment for this compound. The current study is limited in terms of number of vehicles and kinds of organic compounds possibly existing in car cabins. Nonetheless, its results provide useful information on actual car indoor environments.

---

Acknowledgement

This work was supported by the Japan Society for the Promotion of Science KAKENHI (Grant Number 25241023). We thank Ms. Kanako Naito for technical support.

---

Conflict of interest

All authors declare no conflicts of interest.

---