Determination of CO 2 , CH 4 and N 2 O : a Case Study for the City of Rio de Janeiro Using a New Sampling Method

A new sampling method using polypropylene syringes and chemical analysis by gas chromatography was proposed and applied to determine CO2, CH4 and N2O. The samples’ stability was evaluated and compared with the stability of the samples in stainless steel canisters and Tedlar® bags. This methodology was applied to a case study in the urban area of the city of Rio de Janeiro, Brazil. In 2012, annual averages of 507.6 ± 13.2 μmol mol for CO2, 1.95 ± 0.06 μmol mol for CH4 and 325.3 ± 3.4 nmol mol for N2O were obtained. The results for CO2 and CH4 showed peak values in the warmer months and lower values in winter. This CO2 behavior is the opposite of that observed in studies performed in many other cities around the world and may be climate-related. CH4 and N2O inputs are probably due to the evaporation of liquid and solid waste both from landfill sites located within the city boundaries and from polluted rivers and canals. Additionally, other anthropogenic sources may be considered, such as the release of CH4 by the gas fuel network of the urban area and the emissions due to the use of compressed natural gas by light vehicles.


Introduction
According to the National Oceanic and Atmospheric Administration (NOAA), the highest globally averaged temperature for land and ocean surfaces since records began in 1880 was measured in 2014. 1 Additionally, in 2014, for the first time since carbon dioxide has been measured at Mauna Loa Observatory, Hawaii, its peak levels surpassed 400 µmol mol -1 during three consecutive months in the atmosphere. 2 The highest carbon dioxide global average concentration was 398.83 µmol mol -1 in May 2014, according to NOAA. 2 Although this value has not yet reached 400 µmol mol -1 , these global concentrations of the main greenhouse gases (GHG) represent the highest historical values.
Carbon dioxide (CO 2 ) is the most important anthropogenic greenhouse gas in the atmosphere; 3 however, two other gases, i.e., methane (CH 4 ) and nitrous oxide (N 2 O), are important, and they are all closely related to anthropogenic activities, such as fossil fuel combustion, agriculture, biomass burning and waste management.
Brazil is the largest country in Latin America.It is the world's fifth largest country by both geographical area and population. 4It is considered part of the BRICS (Brazil, Russia, India, China and South Africa) group of countries, which are similar in terms of area, large populations and growing economies and, together, are responsible for 18% of the world's gross domestic product (GDP). 5he BRICS group emits an average of 4.5 metric tons per capita of CO 2 , whereas the entire world, including the developed countries, emits an average level of 4.9 metrics tons per capita, according to the World Bank. 6This points the need for the effective monitoring of GHG emissions in these countries.
Many studies have been published involving the measurement of GHG that originate from different sources, including stationary sources, mobile sources and ambient air.For example, GHG measurements have been included in studies involving emissions from human activities in the soil, [7][8][9][10][11] ships, 12,13 hydroelectric reservoirs, 14 vehicles [15][16][17] and thermoelectric power plants. 18However, to our knowledge, the available literature contains no information on the variation in GHG concentrations in Rio de Janeiro and other Brazilian cities, despite their high level of urbanization and economical activities.The annual emissions of CO 2 , CH 4 and N 2 O in Rio de Janeiro have been estimated as 8,734 Gg, 112,616 t and 800 t, respectively. 19The total emissions was calculated Vol. 27, No. 4, 2016   as 11,352 Gg CO 2 , which means that the CH 4 and N 2 O contribution to GHG is approximately 25%. 19 variety of methods for determining trace gases in the atmosphere has been developed.Discontinuous methods can be performed, without pre-concentration, by using stainless steel canisters or Tedlar ® bags to collect air samples and for further chromatographic analysis. 20he collection of ambient air samples in canisters is common practice, and it is mainly done for the determination of volatile organic compounds (VOCs) in the atmosphere. 21,22These canisters have advantages such as inertness, ruggedness, reuse and durability.Despite the high collection efficiency, there are some disadvantages when it is necessary to collect multiple samples at different sampling points, such as the high costs and greater difficulties in transporting the samples.
][25] Commercial analyzers, which have the advantage of providing continuous measurements, are available for automatic GHG monitoring.The high cost of these instruments and the installation requirements are their main disadvantages.
GHG are not currently monitored in Brazilian cities.To our knowledge, there are no data for Rio de Janeiro, which is the second largest city in Brazil.The following investigation was conducted to evaluate the efficacy of low-cost polypropylene (PP) syringes for the sampling and storage of CO 2 , CH 4 and N 2 O.The method was applied in a study case in Rio de Janeiro to evaluate the concentration of these three compounds at representative locations to determine the hourly concentrations over a typical day and seasonal variations throughout one year.

Sampling media
In this experiment, 6.0 L internally electropolished (Restek SUMMA ® coating) stainless steel canisters, 3.0 L Tedlar ® bags (SKC), and 100 mL medical graduate PP syringes with Luer caps were used.The syringes' features include easy handling and transportation and a very low cost per unit.These types of syringes are also used in medical procedures and contain a PP body and rubber plunger.PP syringes might be a suitable substitute when continuous measurements are not available or when sampling with other well-developed sampling media, such as canisters and Tedlar ® bags, cannot be implemented due to the related costs or difficult-to-reach sampling locations.For a blank test, each sampling material was filled with 99.999% synthetic air provided by Linde Gas.

Chemical analyses
The samples were analyzed by gas chromatography with multiple detectors.Chemical analyses were performed using an Agilent 7890A Gas Chromatograph with three detectors: a thermal conductivity detector (TCD) to measure CO 2 at high levels as well as molecular oxygen and nitrogen, an electron capture detector (ECD) to measure N 2 O and a flame ionization detector (FID) to measure CH 4 and CO 2 at low levels.
The chromatography system used two separated channels with 1/8" packed columns (HayeSep Q80/100).The first channel had two valves for the TCD and FID, which were organized in series to measure CO 2 by using a methanizer to convert CO 2 to CH 4 .The other channel, which also had two valves, was used to measure N 2 O on the micro ECD.Two pre-columns were used to retain heavier compounds and to purge oxygen and water through the vent.
The system used a 1.0 mL sampling loop, and the columns, valves, FID, ECD and TCD were operated at 60 °C, 100 °C, 250 °C, 350 °C and 200 °C, respectively.Helium 5.0 was used as the mobile phase at a flow rate of 21 mL min -1 .Quantification was performed using triplicate external calibration curves and three standard reference GHG mixtures provided by Linde Gas.The certified concentrations for CO 2 , CH 4 and N 2 O were 351 ± 7; 1.510 ± 0.075 and 0.250 ± 0.012 µmol mol -1 , respectively, according to the standard number 2040/11; 401 ± 7, 1.760 ± 0.088 and 0.300 ± 0.015 µmol mol -1 , respectively, according to the standard number 041/11; and 451 ± 9; 2.010 ± 0.100 and 0.350 ± 0.017 µmol mol -1 , respectively, according to the standard number 04/11.The determination coefficients (R 2 ) for the triplicate calibration curves were higher than 0.99 for CO 2 and CH 4 and 0.98 for N 2 O.

Stability tests
First, a stability test was performed comparing the most common sampling methods (stainless steel canisters and Tedlar ® bags) with the new proposed method (PP syringes) by checking the sample stability for 10 days of a certified standard mixture containing CO 2 (401 ± 7 µmol mol -1 ), CH 4  (1.760 ± 0.088 µmol mol -1 ) and N 2 O (300 ± 15 nmol mol -1 ) in synthetic air, supplied by Linde Gas (Linde, certificate number 2041/11).
Then, to verify the interaction of the GHGs with other gases and their stabilities, control atmospheres were prepared with different concentrations, stored in the three sampling medias for 10 days, and compared with the initial concentrations.
The control atmospheres were prepared inside 10 L black Tedlar ® bags using 99.999% nitrogen as the dilution gas.The others gases (CO 2 , CH 4 , N 2 O, CO, NO and NO 2 ) were supplied by Linde Gas.
A Sierra Mass Flow controller was used in the dilution process.The GHG were determined by gas chromatography, as previously described, and CO, NO, and NO 2 were determined using Thermo Scientific analyzers (models 48i for CO and 42i for NO and NO 2 ).Water content was controlled by injection, using a syringe at 20 °C and 101 kPa.The composition of the control atmospheres is shown in Table 1.The selected values for GHG concentrations were in the same range as those expected for the atmospheric samples.

Ambient air collection
PP syringes were used to collect ambient air samples in Rio de Janeiro.The city has a population of 6,453,682 26 and contains 2,451,155 vehicles 27 in an area of 1,224 km 2 .Air circulation is greatly affected by topographical conditions; the Tijuca Forest, a mountainous rainforest area, forms a natural barrier to air circulation and divides the city into northern and southern sections. 28As described below, all of the samples were collected in the northern area of the city in locations that are strongly affected by vehicle emissions.
No samples were collected during rainy days or days with atypical anthropogenic emissions.The distance from the sampling positions to obstacles was at least twice the height of the obstacle, and the surrounding environment was free of interference from obstacles.Sampling was performed at a height of approximately 1.5 m above the ground, and the syringe was rinsed five times with the same ambient air before sampling.A 0.2 µm fiberglass filter was linked to the syringe to prevent the entry of particulate matter, which could damage the chromatographic system.The duration of the sampling for one syringe was approximately 10 seconds.After sampling, the syringes were sealed with Luer caps, protected from light with aluminium foil and stored at room temperature.
Initially, 3 samples were collected in 2011 at different times in each of seven districts in the city, as described in Table 2.The sampling locations were mainly residential areas with intense vehicular traffic and commercial activity that are representative of the city.A map of the city indicating the sampling locations is shown in the Supplementary Information section, Figure S1.The exact localization of each sampling point is shown in the Supplementary Information section, Table S1.A brief description of each location is presented in Table 2. Data collected during the sampling period at three monitoring stations of the Environmental Municipal Secretary (SMAC) showed a predominance of weak winds (0-4 km h -1 ) in the southeast and northwest directions and ambient temperatures between 20 and 35 °C.The mixing height profiles were measured using radiosondes at Rio de Janeiro International Airport, which is located approximately 10-15 km from the sampling sites. 29At 10:00 a.m. on each day, the mixing heights were 388-858 m, as detailed in Table S1.Other meteorological data are presented in Table S2.
Later, to check the evolution of the hourly concentrations, 12 samples were collected during the period from 6:00 a.m. to 6:00 p.m. on January 3, 2012, at the Maracanã Campus of the Rio de Janeiro State University, where an automatic air quality monitoring station was located.The mixing height at 10:00 a.m. was 290 m.Meteorological and criteria pollutants concentration data were collected at the station.The minimum temperature was 20 °C at 6:00 a.m., and the maximum temperature was 30 °C at midday.Wind was from the southeast direction and was weak (approximately 1.5 km h -1 ).Finally, 38 samples were collected at Maracanã Campus of the Rio de Janeiro State University during the year of 2012.Samples were always collected in triplicate on Mondays at 1:00 p.m.Samples were collected at a height of 70 m to avoid the possible contribution of local sources of GHG.Additionally, no samples were taken on rainy days or holidays (14 days).

Statistical analysis
The experimental data were analyzed using Statistica 7.0 (Statsoft) to define descriptive statistics before extended analyses.A multivariate statistical analysis was performed that included linear correlation analysis (Pearson coefficients) and principal component analysis (PCA).Both calculations were achieved using the individual experimental values for each sample and sampling site.

Results and Discussion
A typical chromatogram for the GHG standard, in which the three detectors operated simultaneously, is shown in the Supplementary Information section, Figure S2.
The results for the stability test performed with the GHG standard are shown in Table 3.The values for the three sampling media are the arithmetic means and standard deviations for six results obtained over 10 days of storage.
These results show that the stability of GHG samples is similar in the syringes to that found in the more usual sampling devices (stainless steel canisters and Tedlar ® bags).The difference between the initial value and the mean value is always less than 3%, which suggests that syringes are an effective sample storage approach. 30For volatile organic compounds, a difference less than 25% is considered satisfactory. 22he results for the stability test of GHG in the control atmospheres (Table 1) are shown in Table 4.The values are the mean concentrations for three replicate determinations of samples after 10 days of storage and the deviations from the initial value.
Considering the mean values and concentrations of the standard mixture and the uncertainties from the manufacturer (1.75% for CO 2 , 5.11% for CH 4 and 5.00% for N 2 O), it can be concluded that all of the sample media evaluated were effective in storing GHG for a period of 10 days under different concentration conditions.
No effect was observed in the presence of different levels of CO 2 , CH 4 , N 2 O, CO, NO and NO 2 .However, in the mixtures with higher water content, which was equivalent to a relative humidity of 80%, a reduction in N 2 O value was observed.In the ATM 6 mixture, a reduction of 18.1, 15.2 and 16.2% was observed for the canister, Tedlar ® bag (14.33%) and PP syringe, respectively.This may be explained by possible water condensation on the inner wall and the partial solubilization of N 2 O.After the calibration procedure, samples were collected at seven locations in Rio de Janeiro at three different hours of the day.The mean concentrations are shown in Table 5.
The CO 2 concentrations were higher than 400 µmol mol -1 for all samples.The reported values, which were between 475 and 609 µmol mol -1 , are typical of urban areas, as previously determined in many other places throughout the world.The persistence of CO 2 over urban areas as a result of anthropogenic contribution to the CO 2 budget has been reported in many cities and is known as the urban dome.The typical concentrations of CO 2 in urban areas are between 350 and 600 µmol mol -1 and depend on meteorological parameters and urban agglomeration. 31,32In this work, samples were collected in the spring (October and November) in residential areas that are highly affected by vehicle emissions, which were expected to be the main source of CO 2 . 33s shown in Table 5 and in the Supplementary Information section, Figure S3, when the sample locations are compared, higher levels of CO 2 can be observed in two districts, Bonsucesso and Méier, which is probably due to the preferential dispersion of pollutants affected by the topography of these districts and a higher local anthropogenic contribution related to intense commercial activity, high vehicular flux and traffic jams.
According to literature data, the typical concentrations of CH 4 in urban areas are between 1.70 and 2.50 µmol mol -1 and are influenced by meteorological parameters and urban agglomeration.The most significant sources of methane come from the anaerobic decomposition of organic matter. 3he values obtained in this work were in the range of 1.83-2.44µmol mol -1 , which is in good agreement with the reported values for other urban centers.As shown in Table 5 and in the Supplementary Information section, Figure S4, the higher values were obtained in the Penha, Maracanã and Cidade Nova Districts, and they may be attributed to the proximity of polluted rivers with a high organic matter content.The available emission inventories for the city attribute approximately 90-95% of the total CH 4 emissions to the disposal of residential, commercial and industrial solid waste in sanitary landfills and to untreated liquid waste. 33s shown in Table 5 and in the Supplementary Information section, Figure S5, N 2 O concentrations were in the range of 292-360 nmol mol -1 , except for one atypical value (468 nmol mol -1 ).The anthropogenic contribution of N 2 O is mainly related to the disposal of solid and liquid waste. 33iurnal variations in the concentrations were also observed, as shown in Figure 1.For CO 2 , the concentrations varied between 490 and 616 µmol mol -1 .The results indicate that for this compound, higher values were observed in the morning and evening hours, which indicates the effect of the rush hours, and the lowest values were obtained at midday, as previously observed in other cities. 34 The methane concentrations varied between 1.50 and 1.77 µmol mol -1 , which is in the same range as the natural levels (1.65 µmol mol -1 ).A slight diurnal variation due to the boundary layer height and potential photochemical decay was observed.These effects are probably compensated by methane release through evaporation, which, in turn, depends on temperature.N 2 O concentrations varied between 335 and 349 nmol mol -1 , with a slight decrease near midday that was probably due to the superposition of two effects: the dilution effect of the boundary layer height and the increase of emissions through the evaporative process.Correlation coefficients were calculated between CO 2 , CH 4 and N 2 O and the criteria pollutants' concentrations, which were determined at the monitoring station.The CO 2 concentrations showed correlation coefficients over the 0.6 threshold with all of the primary pollutants, CO (0.84), NO x (0.91) and inhalable particulate matter (0.66).Because the main emission source in the studied area of the criteria pollutants are vehicles, it could be assumed that this is also the main source of CO 2 and that anthropogenic emissions are the primary source of the urban CO 2 dome.The correlation of CH 4 with the criteria pollutants was poor except for with NO x (0.60).N 2 O showed correlations with CO (0.74) and NO x (0.78).
Figure 2 shows the results of samples collected from January to December 2012 at the Maracanã campus of Rio de Janeiro State University (UERJ).6][37][38] These annual cycles vary according to geographical location, altitude and meteorological conditions.Correlation coefficients were calculated between our data and the meteorological data (temperature, relative humidity and wind speed).These correlations were poor, and the analysis of regression coefficients showed that all of the values were lower than the 0.6 threshold.The highest value was observed for the CO 2 -temperature correlation (0.46).A high correlation was observed for CO 2 and CH 4 (0.85), which may be due to a common source or to an indirect correlation with another parameter, such as the mixed layer height.This correlation suggests that anthropogenic activities, particularly vehicle traffic, are an important source of methane in the urban atmosphere. 34 This CO 2 behavior was the opposite of that observed in studies performed in many other cities around the world.In general, for CO 2 , the largest values were recorded in autumn, at the end of the biological cycle of plants, and during the winter, and the CH 4 concentrations showed a seasonal variation, with maxima in the summer and minima in the autumn and spring. 34The observed behavior in Rio de Janeiro may be related to the climatic and geographical characteristics.The city has a tropical wet and dry/savanna climate, with a pronounced dry season in the low-sun months, no cold season and a wet season (October-March) during the high-sun months.The city surrounds Tijuca National Park (TNP), a secondary rainforest, which occupies 32 km 2 (3.5% of the area of Rio de Janeiro), and is home to hundreds of species of plants and wildlife, of which many are threatened by extinction and are found only in the Atlantic Rainforest. 39The concentrations of CO 2 and CH 4 were higher during the high-sun months (November-March).The annual CO 2 cycle with the highest values in summer and the lowest values in winter may reflect the high increase in energy consumption, especially that generated by thermo-electrical units, related to the use of refrigeration systems and the increase in tourism in the city.It may also be related to the growth and decay of vegetation in TNP, which is surrounded by the city.The TNP vegetation is characterized by large trees, palms, ferns, epiphytes and lianas.The more frequent plant families are Leguminosae, Sapotaceae, Bombacaceae, Lecythidaceae, Mraceae and Melastomataceae. 40To our knowledge, no studies about CO 2 sequestration have been conducted for this forest.As discussed in the literature, the global estimates of atmospheric CO 2 exchange indicate that the tropics are near equilibrium or are a source with respect to carbon, though there remain considerable uncertainties as to the contribution of different processes. 41egarding CH 4 , the variations may be related to the input from evaporation of liquid and solid waste from landfill sites located within the city boundaries and from polluted rivers and canals.Additionally, other anthropogenic activities that may be considered are the release of methane by the gas fuel network of the urban area and emissions due to the use of compressed natural gas in taxis and light vehicles. 42,43he mean concentration values for each compound, calculated as the arithmetic mean of data obtained from January to December 2012, and the concentrations from the pre-industrial period and global values are shown in Table 6. 44he pre-industrial global-scale trace-gas concentrations from prior to 1750 are assumed to be practically uninfluenced by human activities such as increasingly specialized agriculture, land clearing, and combustion of fossil fuels.Recent tropospheric concentrations of most gases tend to vary systematically over the course of a year, and the global values given in Table 6 represent averages over a specific 12-month period for all gases (the year of 2013 for CO 2 and 2012 for CH 4 and N 2 O).
Comparing the data in Table 6, it can be observed that in Rio de Janeiro, the concentration of CO 2 , the main compound responsible for the greenhouse effect, is 30% higher than the mean global value, thus confirming the presence of a CO 2 dome over the studied urban area.For CH 4 and N 2 O, the values obtained in this work are, respectively, 11.5 and 1% higher than the global concentrations.Values reported in this work should be considered with caution because samplings were always made at noontime on non-rainy workdays and were thus not obtained in the conditions of global values and are not representative of diurnal and weekly variations.However, they do indicate a contribution of urban emissions to CO 2 and CH 4 levels.

Conclusions
The proposed method of sampling provided good results, lower costs and easier implementation in comparison to other sampling devices.It seems to be a good alternative for investigating the CO 2 , CH 4 and N 2 O concentrations in cities where global monitoring is not performed, such as Brazilian and other Latin American cities.
In the studied urban area in Rio de Janeiro, the concentration of CO 2 , which is the main compound responsible for the greenhouse effect, is 30% higher than the mean global value, confirming the presence of a CO 2 dome over the studied urban area.The results for CO 2 and CH 4 show peak values in the warmer months and lower values in winter.This CO 2 behavior is the opposite to that observed in studies performed in many other cities around the world and may reflect the summer increase in energy use due to the higher temperatures and the vegetation cycle in the rainforest, which is part of the urban area.These particular characteristics suggest that further study of the role of the urban rainforest, wetlands and rivers in the air quality of the city may be significant and that the carbon cycle of the vegetation in the rain forest should be studied.

Table 1 .
Composition of the control atmospheres used for the stability tests in µmol mol -1 ; CO 2 , CH 4 and N 2 O concentrations were determined by gas chromatography at the conditions described in the text

Table 2 .
Description of sampling districts in Rio de Janeiro

Table 3 .
Mean concentration and standard deviation values of CO 2 , CH 4 and N 2 O in three sampling media along 10 days of storage at ambient temperature protected from light

Table 4 .
Mean concentrations of CO 2 , CH 4 and N 2 O in the control atmospheres and deviations from the initial value for three replicate determinations of samples after 10 days of storage a PP: polypropylene.

Table 5 .
Concentrations of CO 2 , CH 4 and N 2 O in seven districts of Rio de Janeiro.Samples were collected using syringes, as described in the text a ND = not determined.
Concentrations of CO 2 , CH 4 , and N 2 O on January 3, 2012, at the Maracanã Campus of the Rio de Janeiro State University (UERJ).Concentrations of CO 2 , CH 4 , and N 2 O in 2012 at the Maracanã of the Rio de Janeiro State University.Samples were collected at a height of 70 m.

Table 6 .
CO 2 , CH 4 and N 2 O tropospheric concentrations in the preindustrial period (pre-1750), average global data and values determined in Rio de Janeiro (annual averages)Averages over a specific 12-month period for all gases (the year of 2013 for CO 2 and 2012 for CH 4 and N 2 O).44 a