1. Review of Previous Studies
    There are, to date, few published measurements of benzene in ambient air, and even fewer data which relate to concentrations in the vicinity of petrol filling stations.

    One of the earliest studies measured concentrations of benzene continuously for a period of one year, starting in 1991, at a roadside site on Exhibition Road in London, and reported an annual average concentration of 4.1 ppb. Since this time, the Department of the Environment has established a national network of continuous hydrocarbon monitoring stations. These monitoring stations are located at urban background sites i.e. located away from the immediate influence of roads or other local sources. A summary of the annual average, and peak hour concentrations measured at each site during 1995 are given in Table 1.

    Table 1: Summary of Benzene Concentrations at Hydrocarbon Network Sites (January to December 1995)
    Annual mean (ppb) Max. hour (ppb)
    Birmingham East 1.0 34
    Middlesborough 1.1 44
    London UCL 1.7 19
    London Eltham 1.0 15
    Edinburgh 0.7 11
    Belfast South 0.9 34
    Cardiff East 1.2 24
    Harwell Rural Site 0.4 5

    Stanger Science and Environment have been monitoring benzene levels in the Greater London area since 1991 as part of the London Wide Environment Programme operated on behalf of the London Boroughs. Benzene levels are measured using diffusion tubes packed with Chromosorb 106 which has an excellent affinity for benzene. The diffusion tubes are exposed for two week periods although previous work has shown that the uptake rate for benzene on to Chromosorb 106 differs by less than 1% for exposure periods of one, two and four weeks2. For most adsorbents uptake rates decline rapidly over the first 16 to 24 hours of sampling, after which rates tend to stabilise.

    During 1995 a total of 80 sites were maintained, which included 44 roadside sites, 32 background sites and four petrol station sites (forecourt perimeter). The petrol stations were fitted with neither Stage I or II controls. Annual mean benzene concentrations ranged from 2 to 6 ppb at the roadside sites, 1 to 3 ppb at the background sites and 2 to 5 ppb at the petrol station sites3. During 1995 duplicate tubes were exposed on 26 occasions and showed a coefficient of variation of less than 4%.

    Concentrations of benzene near to petrol filling stations in Middlesborough have been reported by Little and Cram (1995)4. Monitoring was carried out at two petrol stations, neither of which were equipped with Stage I or II controls, during November and December 1994 using passive diffusion tubes with Tenax as the adsorbent . Although a popular adsorbent for other VOCs, Tenax is not particularly suitable for benzene as it has a much lower specific surface area than Chromosorb 106; 35 m2/g compared to 750 m2/g respectively.

    In the Middlesborough study diffusion tubes were exposed over six 48-hour periods at background, roadside, forecourt perimeter and pump locations. The highest levels were recorded at the pumps (12.8 ppb), with measured concentrations at the forecourt perimeter and roadside of 7.8 and 6.8 ppb respectively. There was considerable variation in the measured data from one sampling period to the next. Concentrations at the background sites were much lower (1.6 ppb) corresponding with data measured in a previous survey.

    A study to measure benzene levels around petrol stations is currently being co-ordinated by Bristol City Council, and has been ongoing since August 19955. Monitoring is being carried out using Chromosorb-filled passive diffusion tubes, exposed over two week periods, at two major petrol stations in the Bristol area. Station A is described as 'medium-sized' with six service bays (four nozzles at each), whilst Station B is larger, with six service bays (six nozzles at each) and 24 hour opening. There are no vapour recovery systems fitted at either station. At each station sampling is being carried out at the station forecourt (close to the fuel pumps), and at two off-station sites at a distance of between 50 and 80 metres. Limited sampling has also been carried out at the pump nozzle. These data are shown summarised in Table 2 for the 24 week period between 22 August 1995 and 21 May 1996.

    The average concentration measured at the pumps of the larger station (Station B) is almost 3 times higher than at Station A, and is perhaps related to the greater number of nozzles and 24 hour operation at Station B, which will lead to much higher fuel sales. Limited sampling at the pump nozzle has indicated much higher benzene concentrations than measured at the pump stand, but such levels are probably not indicative of public exposure. Concentrations measured off-site were much lower than at the forecourt, and are likely to have been influenced by emissions from local roads rather than the petrol stations.

    Table 2: Summary of Benzene Concentrations at Two Service Stations in Bristol
    Sampling Location Concentration range (ppb) Average concentration (ppb)
    Station A    
    forecourt 2.6 - 20.1 10.4
    Off-site (1) 1.1 - 3.4 1.9
    Off-site (2) 1.5 - 5.3 2.5
    Nozzlea 10.4 - 23.4 18.1
    Station B    
    forecourt 9.8 - 43.4 28.4
    Off-site (1) 2.1 - 8.1 3.8
    Off-site (2) 3.1 - 8.8 4.8
    Nozzleb - 66.8
    Notes:
    (a) Sampling carried out for 3 periods only
    (b) Sampling carried out for 1 period only
    CONCAWE have produced several reports relating to measurements of benzene levels in and around petrol service stations. A preliminary study was published in 19946 which reported the results of a pilot study which measured concentrations at a petrol station in the UK, and a main study which measured concentrations at twelve service stations in nine countries. The methodology used in the pilot study involved active sampling onto Chromosorb 106, at nine fixed positions for consecutive three hour periods over 24 hours. The precise location of the service station was not stated, although the forecourt was reported to be two miles from the nearest conurbation and in the vicinity of only one major road. During the monitoring period, the wind direction was generally over the busy adjacent road, and the sampling positions included locations at the centre of the forecourt, upwind of the forecourt at the roadside and beyond, and downwind of the forecourt. Mean benzene levels are summarised in Table 3.

    Table 3: Mean Benzene Concentrations Measured Around UK Service Station (Pilot Study)
    Sampling Location Mean Concentration (ppb)
    Centre of forecourt 15.1
    Downwind (50m) 2.2
    Downwind (100 m) 1.8
    Upwind (roadside) 2.9
    Upwind 0.4

    Within the main survey, the sampling strategy was adapted to include two consecutive 12 hour samples, collected in both the summer and winter periods. Of the 12 stations included, eleven were in residential areas, and one was a motorway service station. Two stations were reported to be equipped with Stage II vapour recovery controls. In this study, six sampling points were selected at the boundary of the service station and the adjacent road, but no samples were collected at the centre of the forecourt. The results are summarised in Table 4 and show a considerable spread in the data, presumably related to the precise sampling locations, fuel sales, traffic flows on the adjacent roads etc.

    Table 4: Summary of 24 hour Benzene Concentrations at 12 Service Stations (Summer and Winter Periods)
    Sampling Location Concentration Range (ppb) Mean Concentration (ppb)
    Background 1 0.5 - 8.3 2.4
    Background 2 0.7 - 11.7 2.6
    Roadside (traffic only) 0.8 - 16.2 4.4
    Downwind boundary (traffic plus forecourt) 2.2 - 28.0 6.3
    Boundary 0.5 - 36.6 5.0
    Boundary (roadside) 0.9 - 17.8 4.5


    Following completion of the preliminary study, CONCAWE commissioned a more detailed one-year survey of benzene levels around a single service station in the UK7. The aim of the study was to determine long-term benzene-in-air concentrations, and to assess the contribution of the service station emissions to ambient levels.

    The selected service station was within a semi-rural environment, on a straight section of road, with no other benzene sources in the immediate vicinity. The station was equipped with neither Stage I nor Stage II controls. Sampling was carried out using passive diffusion tubes (filled with Chromosorb 106) at 4 locations around the service station. These locations included 3 sites at the sides and rear of the forecourt area, and one site on the opposite side of the adjacent road. A further 4 samples were collected at a 'green field site' 300 metres away from the station in order to determine the influence of emissions from road traffic. A ninth sample was collected at a site away from both the road and the service station, in order to reflect background benzene-in-air concentrations. Sampling was carried out over 26 consecutive two-week periods during 1993 . The data are summarised in Table 5.

    Table 5: Summary of Benzene Concentrations at a UK Filling Station (Main Study)
    Sampling Location Concentration Range (ppb) Annual Mean (ppb)
    Service station sites    
    Site 1 <0.3 - 3.2 2.1
    Site 2 <0.3 - 2.6 0.7
    Site 3 <0.3 - 2.9 1.1
    Site 4 <0.3 - 1.7 0.5
    Roadside control sites    
    Site 1 <0.3 - 0.8 0.4
    Site 2 <0.3 - 0.8 0.3
    Site 3 <0.3 - 3.2 0.5
    Site 4 <0.3 - 1.1 0.4
    Background sites    
    Site 1 <0.3 - 0.8 <0.3

    The results of this study indicated that the major influence on data measured at each site appeared to originate from local activity i.e. within a few metres. Consequently there was a correlation between the benzene concentrations measured within the forecourt and location of the connection points for deliveries and tank vents, with the highest values recorded at Site 1 located closest to the loading connections and tank vents.
    Analysis of the complete data set indicated that the service station activities contributed approximately 65% of the measured benzene concentrations at the service station, with 15% being derived from adjacent road traffic, and the remaining 20% from the local background.

    In addition to passive sampling with diffusion tubes the study also involved active sampling for twelve hour periods every thirtieth day. Comparison of the two data sets showed that the diffusive monitoring results were lower than the results obtained through active sampling. This was partly explained by the relative periods of time during which the service station was open; during the diffusion tube sampling the service station was open for 63% of the time whilst during the active sampling the service station was open for 83% of the time. In addition the active data were more variable as the fuel throughput during sampling was more variable and the data were more influenced by short term emissions.


Report prepared by Stanger Science and Environment
Site prepared by the National Environmental Technology Centre, part of AEA Technology, on behalf of the UK Department of the Environment, Transport and the Regions