FormalPara Article Highlights

  • Charcoal fuel generates particulate matter, carbon monoxide and heavy metals which pollutes indoor air pollution especially in kitchens

  • While the concentration of heavy metals was lower than recommended EPA concentrations, it is important to note that chronic exposure is a health risk

  • In order to reduce this pollution, kitchens should be well ventilated; we recommend that people in rural areas should endeavor to increase the kitchen size and improve ventilation

1 Introduction

Air pollution although more pronounced in low- or middle-income countries in Asia and Africa is a global challenge. Poor air quality indoors is estimated to be responsible for 2.7% and 3.7% of the global and developing countries disease burden respectively [1]; it is associated with respiratory diseases such as lung cancer and pneumonia [2] which are detrimental to health. Indoor air pollution due to inefficient and poorly ventilated stoves burning biomass fuels such as wood, crop waste and dung, or coal is responsible for the deaths of an estimated 1.6 million people annually; more than half of which occur among children under 5 years of age [3].

Biomass in it various forms provides over 90% of primary energy in most developing nations [4] where the majority of people depend on charcoal, fire wood or agricultural wastes for their energy needs [5]; for example, 93% of the energy consumption in Uganda is wood fuel (firewood and charcoal) and agricultural wastes [6, 7] and in urban areas, 65.7% of the households use charcoal while 33.4% use firewood for cooking [8] yet biomass is a source of air pollutants [3, 8,9,10]. Knowledge about specific biomass contribution to air pollution in Uganda is limited, thus emissions from kitchens using charcoal as the main source of energy in Mbarara Municipality Western Uganda were investigated to determine the contribution of charcoal usage to indoor air pollution.

Biomass contributes to indoor air pollutants such as particulate matter and carbon monoxide resulting from its combustion [11]. Particulate matter and carbon monoxide (CO) are deleterious to health for example their exposure has been associated with heart attacks, wheezing, coughing, asthma, still births and lung cancer [2, 12, 13]. Particulate matter (PM) consists of a mixture of solid and liquid particles suspended in the air for extended periods of time [7]. Solid particulates are composed of those with a diameter of less than 2.5 μm (PM2.5), often called fine PM, and those with a diameter of less than 0.1 μm called ultra-fine particles [14]. They can penetrate deeper into respiratory tract leading to respiratory failure [15], hence the concern on possible hazardous effect on health if inhaled.

In addition to the size of particulate matter, its components such as heavy metals are among the risk factors to human health and studies have shown a higher concentration of these metals in indoor air compared to outdoor air [16]. The toxic metal content of PM2.5 has been associated with adverse respiratory health effects according to the International Agency for Research on Cancer (IARC), which classified several metals, including chromium, cadmium, lead and nickel, as potential cancer causing agents [7, 17] showing the need to determine heavy-metal composition in particulate matter.

Ambient air quality affects indoor air and indoor air has been found to contain higher concentrations of pollutants compared to outdoor air [18, 19]; High ambient air pollution levels have been reported in North Africa and Southeast Asia [20] and there is evidence that air quality in some African cities like Lagos in Nigeria, is deteriorating where PM2.5 concentrations have been estimated at 100 mg/m3 compared to <20 mg/m3 in most European and North American cities [21]; there is need for air quality monitoring in African cities.

Apart from ambient air quality activities occurring indoors also influence the air quality; for example BTEX was found in indoor air of beauty salons in Iran [22]. In a study investigating PM2.5 in urban homes in Egypt a correlation was observed between kitchen activities (smoking and cooking) and PM2.5 levels [23] and variations in cooking fuel and kitchen size produced large differences in indoor air pollution in rural Madagascar [24]. In this study kitchen characteristics like the wall, floor and roof types in addition to the volume and fuel used were investigated to better understand their effect on indoor air quality.

Although there is limited information about air quality in Uganda, the 2018 world air quality report ranked Uganda, second after Nigeria among the African countries with worst particulate matter pollution [25]. An assessment of outdoor air in two Ugandan cities (Kampala and Jinja) indicated the mean PM2.5 concentration (132.1 μg/m3) as 5.5 times higher than the WHO cut-off limits of 24 μg/m3 [26]. Higher concentrations of PM2.5 were observed in Kampala compared to Jinja and this was attributed to dust emission from unpaved roads, high traffic and industrial emissions, and burning of household garbage. The limited air quality information in Uganda and the scarcity of data on indoor air quality motivated this investigation in Mbarara Municipality.

In Mbarara town 77.3% of population depends on charcoal fuel [27] and the town is the second to Kampala in industry. The increasing small-scale industries and biomass use in Mbarara can generate hazardous air pollutants thus a need for air quality monitoring in Mbarara. Furthermore, since outdoor air can influence indoor air, there was need to assess indoor air quality because the World Health Organization (WHO) ranks indoor air pollution (from biomass fuel combustion) and urban outdoor air pollution 10th and 14th, respectively, among 19 leading risk factors for global mortality [28]. This study therefore determined particulate matter, its metal components, and carbon monoxide concentrations from households using biomass energy in Mbarara Municipality to evaluate the contribution of charcoal biomass to indoor air pollution.

2 Materials and methods

2.1 Study site

The study was conducted in Mbarara Municipality western Uganda (Fig. 1), about 270 km, by road, southwest of Kampala, Uganda’s capital city. Mbarara Municipality is divided into six divisions which include Biharwe, Nyakayojo, Kakiika, Kakoba, Kamukuzi and Nyamitanga. In August 2014, the national population census put the population at 195,013 [29].

Fig. 1
figure 1

Map showing Mbarara Municipality—the study area

Full size image

2.2 Study design

This was a descriptive and experimental study which described kitchen characteristics and determined indoor particulate matter PM2.5, its heavy metal composition and CO in kitchens that use charcoal as their main cooking fuel Mbarara Municipality. The study was conducted between December 2017 and April 2018 in six villages randomly selected from the six divisions of Mbarara Municipality. PM2.5 and its heavy metal content were measured in December 2017 (dry season, with average rainfall 76 mm) and in February, March and April 2018 (wet season, with average rainfall 132 mm). CO in household kitchens was measured from February to April 2018. Sixty households (10 from each village) which use charcoal as their main source of energy were selected for the study, and permission to install devises for measuring the pollutants in households was obtained from family heads.

2.3 Sample size determination

Sample size was determined using a formula by Smith (2013) [30] as shown in Eq. (1)

$$n=\frac{Z^2{\sigma}^2}{B^2}$$
(1)

where n = sample size

Z = Z-score (1.96)

σ = standard deviation (0.198)

B = margin of error (0.05).

The study was described to all participating households and informed consents for participation were secured. Ethical approval was sought and obtained from Mbarara University of Science and Technology Research and Ethics Committee (MUREC1/7). The study used interviews, focus group discussions, self-administered questionnaires, indoor air monitors, and ICP-OES to obtain qualitative and quantitative data.

2.4 Determination of household demographics and kitchen characteristics

A questionnaire was administered to each of the family heads to obtain data about house demographics and kitchen characteristics. The questionnaire was pretested before the main study among ten randomly selected houses to correct method errors and was administered in both English and Runyankore (the local language spoken in Mbarara). In addition to questionnaires, six focus group discussions (FGDs) consisting of family heads were conducted in each of the selected cells to interrogate the data obtained by questionnaires about house characteristics. All FGDs consisted between six to seven participants and discussions were conducted in the local language (Runyankole).

2.5 Determination of PM2.5 in indoor air of kitchens using charcoal as the main source of energy

Indoor concentrations of fine particulate matter (PM2.5) and temperature were measured following the method described by Balakrishnan et al. [31], using University of California Berkeley Particle and Temperature Sensor (UCB-PATS, Berkeley Air Monitoring Group, Berkeley, CA, USA) with size-selective in-let conditioner. Monitors were placed at a height of 1.5 m above the ground and 1 m from the point of cooking which is within the breathing range of a person seated while cooking. They were calibrated with combustion aerosols of charcoal against gravimetric measurements of PM2.5 in the laboratory before being used in the field. Particle coefficients were derived for each instrument in the field through colocation of UCB-PATs monitors and gravimetric samplers with a flow rate of 1.5 L/min. Each participating household was monitored for 24 h.

Using a laboratory calibrated rota meter, flow rates were measured before and after sampling. The in-lets drew PM2.5 onto cellulose/PTFE collection filters fitted in the cyclones of size-selective in-lets of diameter 2.5 μm connected to the sampler. The filters were conditioned in a temperature and humidity controlled room for 24 h before weighing. The filters were then weighed three times each before and after sample collection using a micro balance (Metler Toledo XS3DU) in a temperature and humidity controlled room and average values of PM2.5 were calculated. PM samples were then stored in plastic petri-dishes at 20 °C in a refrigerator before analyzing them for heavy metals.

The UCB-PATS were first kept in ziplock bags for 1 h to allow the monitor adjust itself to measure zero particulate matter before and after deployment. After monitoring, all data files were batch-processed using a customized software package developed for this device [31].

2.6 Determination of heavy metals in PM2.5 collected from kitchens using charcoal as the main source of energy

2.6.1 Sample collection and preparation

Particulate matter was collected on filters as described in section 2.5 (Determination of PM2.5 in Indoor Air of Kitchens Using Charcoal). A procedure recommended by U.S. Environmental Protection Agency (EPA/625/R-96/110a) was used as the conventional acid extraction method to extract metals from the filters. Nitric acid (1 M) was added to a filter in a digestion glass vial. The sample was then heated at 90 °C for 15 min. Concentrated nitric acid (2.5 mL) was then added, and the sample refluxed for another 30 min. The sample was then cooled and concentrated nitric acid (2.5 mL) added again. The sample was refluxed for an additional 30 min. This was followed by addition of 30% hydrogen peroxide (0.5 mL) to the cool sample and the sample re-heated until no effervescence. Hydrogen peroxide (1.0 mL) was then added to the sample three more times and sample heated between each aliquot addition of 30% hydrogen peroxide. After cooling, the sample was diluted up to 50 mL with deionized water. The procedure was repeated for each of the filters from the households studied.

2.6.2 ICP-OES analytical method validation

Calibration solutions were prepared by diluting standard metal solutions to concentrations of 0.1 mg/L, 1.0 mg/L, 2.0 mg/L, 20.0 mg/L and the machine was checked after the initial calibration for every ten samples. In the case of more than ±10% deviation, the ICP-OES (8300 Perkin Elmer, USA) would be re-calibrated. The limit of detection (LOD) for each element was obtained from, LOD = 3S/m, where S is the standard deviation of the blank readings and m represents the gradient of the calibration curve for each element. The limit of quantification (LOQ) was calculated using LOQ = 10S/m.

2.6.3 Analysis for metals in PM 2.5

The samples were analyzed using an Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES). A tube connected to the ICP-OES was dipped into each sample to aspirate the sample into the machine which would analyze and produce the analyte concentration. The average concentration for triplicate measurements for a particular sample was then computed.

2.7 CO monitoring in indoor air of kitchens using charcoal as the main source of energy

Carbon monoxide concentrations were measured using a portable, battery-operated, data-logging Drager Pac 7000 (SKC, Inc.; Eighty-Four, PA, USA) placed at a height of 1.5 m above the ground and 1 m from the point of cooking. The Pac 7000 recorded and logged the peak concentration that occurred within each minute during the monitoring period following the method described by Balakrishnan et al. (2015) [31]. Each participating household was monitored for 24 h.

2.8 Data analysis

Data was collected, tabulated and entered into Microsoft Excel. Means, variance and standard error were used to assess the spread of the data. The mean of parameters (±SE) and one-way analysis of variance (ANOVA) were used to compare the mean values of observations. Differences in mean values were considered significant if calculated p-values were < 0.05.

3 Results and discussion

3.1 Household demographics and kitchen characteristics

Information on household demographics and kitchen characteristics is summarized in Table 1. The average number of people living in each house was 3.6 ± 1.5 which is in agreement with the national average of 4.0 [32] and there were no smoking occupants in all households investigated in this study. Some of the households (37%) were renting and some houses had no kitchens which influenced cooking from the main house. 43.3% of household owners attained primary education 38.3% secondary education and 15% degree and 3.3% attained diplomas.

Table 1 Household demographics and kitchen characteristics
Full size table

Kitchen volumes ranged from 12 to 81 m3 and their roofs (96.6%) were metallic while only 3.4% were concrete. Kitchen walls were made of mud (5%), wood (20%) or concrete (74%) and floors were wood (3.4%), mud (41%), and cement (56%). 23% of the respondents did not open doors while cooking but only 3.4% did not open doors. 22% of the houses studied were located near the road and only 6.8% showed cross ventilation. 21% of the respondents cooked from the main house and 79% used kitchens.

Charcoal was the most commonly used cooking fuel and this was corroborated during focus groups discussions. There was high demand for charcoal compared to wood; other fuels like electricity and paraffin were considered to be expensive. Participants in the focus group discussion also stated that financial status and lack of kitchens determined the type of fuels used. When asked “why do people in Mbarara Municipality use different fuels?” for example one respondent said:

for me I don’t have a kitchen…I cook from where I sleep, how can I use wood? Wood even produces a lot of smoke which can darken the walls and my clothes.

Two participants revealed that they use paraffin stoves only when they do not have enough time to light charcoal stoves.

Responding to the question “How do you light charcoal?”, 58% said they use wood husks or papers, 30% use polythene papers, 5% use paraffin and others were undecided. The method used to light charcoal can contribute to kitchen indoor pollutants for example; burning of polythene papers has been associated with heavy metal emission [33].

Most of the focus group participants (73.7%) responded positively when asked “Do you believe in opening windows and doors while cooking?” These participants associated window and door opening with circulation of fresh and cool air into kitchens but had no knowledge about the other effects of outdoor air on indoor air like transfer of pollutants from ambient air into indoor air. Some participants however were opposed to the opening of windows and doors during cooking because they did not want people to know that they cook and they believed that charcoal does not produce a lot of smoke which would require opening kitchen windows. For example, one participant said:

for me I don’t see the importance of opening the widows…what if it is raining or at night…you cannot. May be if there is no electric power I can open to get light

This shows limited knowledge on the importance of kitchen ventilation on air circulation.

All participants recommended the kitchen as the best place to cook from when responding to the question “What is the best place to cook from?” but they gave varying views when responding to the question “Why do people cook from the main house and not in kitchens?”. Some attributed the cooking from main house to financial status, for example one participant said in a sad tone:

I have failed to pay rent for the main house. Can I manage to pay for another room to be used as a kitchen? It is not possible.

Other people attributed cooking from the main house to limited space, for example one participant said: “usually there is no enough space and even the little space available you are planning to build another room for renting”. Financial status is a limiting factor to possession of kitchens although it was also revealed during the discussions that some people were renting and even if they had money to construct kitchens, they were unable to.

All focus group participants were in agreement that kitchen air can be contaminated when responding to the question “Do you agree that kitchen indoor air can be contaminated?” They revealed that cooking and sweeping are the common sources of kitchen air pollutants when asked “What could be the sources of kitchen air contaminants?”. For example, one participant said:

When you are lighting charcoal stove, the wood husks you are using produce a lot of smoke. In fact sometimes I am forced to light it from outside to reduce on the smoke in the kitchen.

From the discussions, the participants demonstrated awareness about some indoor air pollutants although knowledge about the outdoor sources of pollutants like nearby traffic and industrial emissions was limited. Participants were also aware of some effects of kitchen smoke; responding to the question “What could be the effects of kitchen smoke?” Most responses were about eye irritation, staining/darkening of clothes and kitchen walls by the smoke. Participants had limited knowledge on other health effects of kitchen smoke inhalation like lung cancer and pneumonia.

Focus group discussions generated richer data that supplemented the data collected by interviews and self-administered questionnaires; for example, the large number of people cooking from main house which was generated by questionnaires was attributed to financial status by the focus groups discussions. It was also revealed that participants had little information about indoor air pollution while some seemed unbothered.

3.2 Mean 24 h indoor PM2.5 concentration in different kitchens in Mbarara Municipality

During the 2017 survey, 60 households were monitored however only 58 households were present in the follow-up 2018 survey because two households shifted. Kitchen concentrations of PM2.5 are reported in Table 2.

Table 2 Mean 24 h indoor PM2.5 concentrations in indoor air of kitchen in Mbarara Municipality
Full size table

Nyakayojo had the highest average indoor concentration of PM2.5 (0.84 mg/m3) while Kamukuzi had the least average concentration of PM2.5 (0.29 mg/m3) in both dry season and wet season as shown in Table 3. This can be attributed to the kitchen characteristics like mud floors and supplementing charcoal with wood fuel which were predominant in Nyakayojo. On average, there was a significant difference in PM2.5 concentration across all divisions (p = 0.004) due to differences in fuels used to supplement charcoal and other kitchen characteristics like size and floor type.

Table 3 Mean 24 h indoor PM2.5 concentrations per division in Mbarara Municipality
Full size table

The mean 24-h particulate matter concentration was higher than the World Health Organization (WHO) 24-h Air Quality Guideline of 0.024 mg/m3 (Tables 2 and 3 and Fig. 2). This shows that charcoal fuel contributes to pollution of indoor air in kitchens in Mbarara Municipality. Higher PM2.5 concentrations have also been observed in other studies in rural Sierra Leone where an average of 1.686 mg/m3 was observed in kitchens using charcoal fuel [34]. However, some studies have observed lower concentrations than observed e.g. a study assessing indoor air in urban slums of Nairobi found PM2.5 concentration of 0.1265 mg/m3 [35]. These differences can be attributed to locations of kitchens in these studies like being near emission sources [19, 24] in addition to kitchen characteristics such as ventilation and use of other fuels in addition to charcoal may affect the quantity of PM.

Fig. 2
figure 2

Mean 24 h PM2.5 concentration per division for dry season and wet season

Full size image

3.3 Influence of household demographics on PM2.5 concentrations

Most household owners attained primary education (43.3%) followed by secondary education (38.3%), diploma (03%), and then degree (15%) indicating low literacy levels in the area. Households where family heads attained degree had higher mean PM2.5 concentration (0.62 mg/m3) followed by primary education (0.59 mg/m3) then secondary (0.336 mg/m3) and lastly diploma (0.15 mg/m3). This is because most households with graduate heads cooked for a long period (6–7 h) leading to which generation of more of the pollutant compared to other households like those with diploma that cooked for a short period (3–4 h). During focus group discussions, long hours of cooking was attributed to cooking foods like meat and beans which take long to be ready and leaving fire on to keep food warm.

The mean PM2.5 concentration was higher in households with seven occupants (1.01 mg/m3) followed by those with four (0.66 mg/m3), then households with six (0.52 mg/m3) followed by those with two (0.39 mg/m3), three (0.38 mg/m3), five (0.19 mg/m3) and lowest in households with one occupant (0.14 mg/m3). On average, there was a significant difference in PM2.5 concentration across all households with different number of occupants (p = 0.01). This was expected because of the observed differences in cooking times in different households. Households with one occupant cooked once or twice a day implying that less amounts of the pollutant were produced compared to other households that cooked many times a day. Also, households with many occupants cooked for longer hours which can be attributed to food prepared that require longer hours to be ready. The mean PM2.5 concentration in households with five occupants was lower than that for those with four occupants. This is because 25% of households with five occupants had mud floors that generated less PM2.5 concentration compared to households with four occupants where 44.4% had mud floor that generated much PM2.5 concentration.

3.4 Influence of kitchen characteristics on PM2.5 concentrations

It was observed that more kitchens supplemented charcoal fuel with wood (26.3%) than kerosene (5.3%) as shown in Table 1. The results are similar to those observed in a similar study conducted in Korogocho a Nairobi slum where a larger percentage of households (62.5%) used charcoal and wood compared to that using kerosene (12.5%) [35] and this is attributed to high prices of kerosene compared to other fuels. This supports results from a study which classified wood and charcoal as most important sources of energy in most African cities [4, 6, 36]. A large number of households (39.5%) cooked from rooms used for sleeping which increases their exposure to indoor air pollutants. Most households cooked with either windows or doors closed which limits ventilation yet poor ventilation affects air quality [37] and this could increase exposure levels to indoor air pollutants.

3.4.1 Influence of kitchen size on PM2.5 concentration

The average kitchen volume was 46 ± 17 m3. Increase in kitchen volume decreases PM2.5 concentrations as shown by a negative significant correlation between PM2.5 concentration and kitchen volume (r = −0.7; p = 0.000). This is because increased kitchen sizes can allow quick dilution, circulation and exit of polluted air from kitchens and this reduces exposure level to PM2.5.

3.4.2 Effect of cooking fuel on indoor PM2.5 concentrations

Kitchens using charcoal and wood had higher PM2.5 concentrations (1.044 mg/m3 ± 0.46) compared to those using charcoal only (0 .284 mg/m3 ± 0.28) showing that wood contributed to PM2.5 concentrations in indoor air. There was no significant difference between PM2.5 concentrations in kitchens using charcoal only and those using both charcoal and kerosene (p = 0.972) and this can be attributed to irregular use of kerosene indicated in focus group discussions. On average, there was a significant difference in PM2.5 concentration across all the cooking fuel types (p = 0.00). This was expected since different fuel types have different components. All the cooking fuel types produced PM2.5 concentrations that were higher than the WHO recommended levels which exposes households to harmful effects of PM2.5.

3.4.3 Effect of roofing material on indoor PM2.5 concentrations

Many kitchens (56) had metal roofs and only two had concrete roofs and the observed PM2.5 concentration was higher (1.8 mg/m3) in the few kitchens with concrete roof than in those with metal roof (0.52 mg/m3). Metal roofs had some vents that could have allowed quick circulation and exchange of air pollutants reducing their concentrations in indoor air compared to concrete roofs which did not have vents resulting into high PM2.5 concentrations. Also, most of concrete roofs were not smooth and in wearing off condition, hence could have generated indoor air pollutants raising their concentration in indoor air. PM2.5 concentration was higher (1.61 mg/m3) in kitchens with mud floor and concrete roof than in those with mud floor and metal roof (0.77 mg/m3) indicating increased pollution in kitchens with concrete roofs. Therefore, there is need to allow more ventilation in kitchens with concrete roofs to enable quick air exchange of indoor air pollutants and smoothening of concrete to reduce on wearing off.

3.4.4 Effect of kitchen wall on indoor PM2.5 concentrations

Kitchens with mud walls had higher PM2.5 concentration (1.6 mg/m3), followed by concrete walls (0.48 mg/m3), then wood (0.47 mg/m3) as shown in Table 1. Also, kitchens with mud floor and mud wall had higher PM2.5 concentration (1.8 mg/m3), followed by mud floor with concrete walls (0.8 mg/m3), then mud floor with wood wall (0.44 mg/m3). This shows that wood wall was better than mud and concrete walls (which were in poor condition). The high PM2.5 concentration in kitchens with mud and concrete walls can be attributed to high rate of wearing off of the walls observed in most kitchens compared to wood. Also, during cleaning, air pollutants can be generated much more from concrete and mud which were observed to have a high rate of wearing off compared to wood. Concrete kitchens should be cleaned gently to reduce on the rate of wearing off.

3.4.5 Effect of kitchen floor on indoor PM2.5 concentrations

Kitchens with mud floors had higher PM2.5 concentration (0.83 mg/m3), followed by wood floor (0.4 mg/m3), followed by cement (0.30 mg/m3). High PM2.5 concentration in kitchens with mud floors is as a result of sweeping where mud floors generate a lot of dust compared to other floors. This is supported by information generated from focus group discussions where most house owners reported that they sweep daily. Therefore controlled sweeping by first applying some water to reduce on dust is recommended.

3.4.6 Effect of ventilation on indoor PM2.5 concentrations

Kitchens that cooked with windows and doors open had lower PM2.5 concentration (0.37 mg/m3) than those that cooked with windows and doors closed (0.78 mg/m3). On average, there was a significant difference in PM2.5 concentration across kitchens with closed and open windows (p = 0.002) as expected since open doors and windows allow quicker air exchange of air from indoor air reducing their concentration in indoor air. Low PM2.5 concentration in kitchens with open windows and doors also implies that no significant amount of PM2.5 infiltrated indoor air from outdoor.

Also, 95% of the kitchen windows/doors did not allow cross ventilation and had higher PM2.5 concentration (0.55 mg/m3) than those that allowed cross ventilation (0.14 mg/m3). On average, there was a significant difference between PM2.5 in kitchens with closed windows/doors and those with open windows/doors (p = 0.016). Kitchens should therefore be constructed with doors/windows facing one another to allow cross ventilation which also allows quicker exchange of indoor pollutants.

3.5 Heavy metal concentrations in PM2.5

PM2.5 from each kitchen was analyzed for metals Zn, Cu, Pb, Fe, Cd and Cr using an ICP-OES. The calibration curves for all the six metals demonstrated good linearity over the concentration range (0.1–20.0 mg/L) with correlation coefficients (R2) in the range of 0.996–0.999. The limits of detection (LODs) were determined using the calibration data and were 0.000002 mg/L for Cu, 0.000002 mg/L Cd, 0.000006 mg/L Pb, 0.000002 mg/L Zn, 0.000003 mg/L Fe and 0.000005 mg/L Cr.

Concentrations of heavy metals in PM2.5 in different kitchens are reported in Table 4. Among the kitchens considered, a large number of kitchens (62.1%) had iron present in its PM and 6.9% had cadmium present in its PM. Although all the elements can originate from biomass combustion as shown by previous studies [37], iron is also classified among mineral components, copper among traffic-related components, lead among tracer of heavy oil combustion, and cadmium, zinc and chromium among industrial elements [37]. Cadmium, zinc and chromium were present in kitchens located near areas with motor vehicle emissions which are sources of these metals. The outdoor air polluted with them could have entered into indoor air and this is supported by results of a related study investigating indoor air quality in commercial kitchen which indicate higher amount of indoor PM2.5 in food centers located towards the roadside pavement and this was attributed to infiltration of outdoor air into indoor [38].

Table 4 Mean concentrations of metals in PM2.5 in different kitchens
Full size table

Among the kitchens with mud floors, 88% had iron present in their PM2.5 indicating that floor dust could be one of the potential sources of iron metal. Previous studies have associated dust to heavy metal presence in particulate matter and results of a study measuring heavy metal concentrations in indoor dust in Malaysia indicate higher concentration of iron compared to other metals [39]. Therefore the observed large number of kitchens with mud floors could have generated a lot of dust leading to contamination of indoor air with iron. This is likely to be substantial where there is daily sweeping of kitchens that results into a lot of dust. In addition to sweeping, burning of plastics is another major source of heavy metals like Cd, Cr, Pb, and since some of the households were burning plastics like polyethene bags when making charcoal fire, they are at a risk of heavy metal pollution. No metal was detected in four kitchens which had cemented floors while only one kitchen with mud floor in Kamukuzi Division had all the metals detected in its indoor PM2.5 indicating that floor type influenced the presence of metals in kitchen indoor air.

Generally, mean metal concentrations for all metals were lower than recommended exposure levels by EPA (Table 4) but these low concentrations if inhaled continuously can bio-accumulate to hazardous levels hence the need for their emission reductions. A similar study in China observed higher concentrations of lead (0.09 μg/m3) in indoor PM2.5 but lower zinc (0.24 μg/m3), copper (0.02 μg/m3) and chromium (0.03 μg/m3) concentrations [40] than those observed in this study. However, the PM2.5 analyzed came from different environments using gas fuel as opposed to charcoal fuel used in this study.

3.6 Mean 24 h carbon monoxide concentration in kitchens

Between February and April 2018 we monitored 56 households using the Pac 7000 that recorded and logged the peak concentration that occurred within each minute during the monitoring period following the method by Balakrishnan, K., et al. [31]. Each participating household was monitored for 24 h and the mean 24-h CO concentrations for each kitchen are reported in Table 5.

Table 5 Mean 24 h indoor CO concentrations in kitchens in Mbarara Municipality
Full size table

The concentration of CO varied within households in different divisions because of different kitchen characteristics and fuels although the differences were not statistically significant (p = 0.102) as shown in Table 5. Nyamitanga had the highest CO concentration (62.6 ppm) while Kakiika had the lowest (15.8 ppm). This can be attributed to the use of kerosene in some households in Nyamitanga, which is a significant contributor of CO to indoor air in addition to charcoal and wood compared to Kakiika where all households use only charcoal and wood.

It is important to note that the mean 24-h CO concentrations (41.5 ppm) were higher than the World Health Organization (WHO) 24-h Air Quality Guideline of 6.34 ppm. This is also similar to results from a related study in Sierra Leon [34] Furthermore, high CO levels (44 ppm) were observed in another study assessing indoor air in peri-urban areas of Kampala [41]. All the studies examined households using charcoal fuel and indicate a high risk of CO poisoning. Therefore, more efforts should be taken to control CO levels in kitchens.

3.7 Influence of household demographics on CO concentrations

Households where family heads attained diploma had higher mean CO concentration (65.7 ppm) followed by primary education (42.41 ppm) then secondary (40.87 ppm) and lastly degree (23.8 ppm). All the households with diploma cooked from the main houses and used charcoal fuel which generates more CO compared to wood. Taylor and Nakai [34] in their investigation of levels of toxic air pollutants in kitchens using traditional stoves in rural Sierra Leone, also demonstrated that wood produces more PM while charcoal generates more CO.

The mean CO concentration was higher in households with six occupants (86.68 ppm) followed by those with four (43.54 ppm), then households with three (40.05 ppm) followed by those with seven (38.00 ppm), two (33.15 ppm), five (4.22 ppm) and lowest in households with one ocupant (1.99 ppm). On average, there was a significant difference in CO concentration across households with different number of occupants (p = 000). Lower CO concentrations observed in households with one participant was caused by fewer cooking times (one to two times) compared many coking times in households with many occupants.

3.8 Effect of kitchen characteristics on CO concentrations

Mean 24-h CO concentrations per kitchen and household characteristic are summarized in Table 1.

3.8.1 Effect of cooking fuel on CO concentration

Type of cooking fuel is an important factor that influences CO levels in kitchens. High CO levels were observed in kitchens using charcoal and kerosene (146.4 ppm) compared to those using charcoal and wood (55.4) or charcoal alone (25.2) and on average, there was a significant difference in CO concentration across all the cooking fuel types (p = 0.00) as expected.

3.8.2 Influence of kitchen volume on CO concentration

Increase in kitchen volume had no effect on CO concentration. This was shown by a low insignificant correlation observed when kitchen volume was compared with CO concentration, (p = 0.074). This was not expected but it could have been influenced by other kitchen characteristics.

3.8.3 Effect of roofing material on indoor CO concentrations

Carbon monoxide concentration was 6.6 ppm in 2 kitchens with concrete roof and 39.1 ppm in those with metal roof which were 56 kitchens. Concrete roofs do not allow quick exchange of indoor air pollutants compared to metal roofs which have some vents. Kitchens with concrete roofs should therefore be constructed with enough ventilators to allow quick exchange of CO from indoor air.

3.8.4 Effect of ventilation on Indoor CO Concentrations

In Kitchens where cooking was done while windows and doors were open had lower CO concentration (37.4 ppm) than those that cooked with windows and doors closed (47.8 ppm) The difference in CO concentration across kitchens with closed and open windows was not significant (p = 0.363). This indicates that ventilation did not affect CO concentration significantly, 95% of the kitchen windows/doors did not allow cross ventilation and had higher CO concentration (43.3 ppm) than those that allowed cross ventilation (9.9 ppm) and this highlights the importance of cross ventilation in reducing CO concentrations by allowing quick circulation and exit of polluted air.

3.8.5 Effect of cooking time on indoor CO concentration

Increase in cooking time increases indoor CO concentration. This was evidenced by high CO concentration (46.56 ppm) in kitchens that cooked four times per day than in those that cooked fewer times. This increases the risk of exposure to CO.

4 Conclusion and recommendations

This study determined the effect of charcoal fuel on concentrations of PM2.5, CO and heavy metals in kitchen indoor air in Mbarara Municipality Uganda. The results showed that charcoal fuel contributed to fine particulate matter (PM 2.5), heavy metals and carbon monoxide in kitchen indoor air thus it is a risk factor to human health. The study results also confirmed that kitchen characteristics like fuel, ventilation and volume affect the concentration and type of indoor air pollutants. There is a need for monitoring indoor air quality to ensure it meets standards.

While the observed results provide important information on indoor air quality in Mbarara town, there is need for further studies on personal exposure to air pollutants to clearly assess the health implications.

The study limitation was that PM2.5 and CO could not be measured throughout the year; instead measurements were made only in December 2017 for dry season and from February to April 2018 for wet season. Further research is needed to investigate long term variations on PM2.5 and CO concentrations by studying all the months throughout the year.