LIDAR MEASUREMENTS OF INDUSTRIAL BENZENE EMISSIONS

The ability to measure benzene concentrations was added to the RIVM mobile DIAL system. In a ten-days campaign, it was used to measure benzene emissions in the Rijnmond, a heavily industrialised area in the South-west of the Netherlands with petrochemical industry, petrochemical products storage and the port of Rotterdam. On two of the ten days, benzene emissions were found. Combined with measurements of wind speed and wind direction, the Lidar measurements indicated the possible origins of these emissions. This makes the Lidar a valuable tool, augmenting the data collected at fixed monitoring stations.


INTRODUCTION
Benzene is a widely used but hazardous and carcinogenic substance.During production, processing or transhipment, some benzene often escapes into the air.In 2012, the total emission of benzene to the air in the Netherlands amounted to slightly more than two million kg.Exposure to benzene should be avoided as much as possible.This makes it important to know where sources of benzene are located.For this reason, the DCMR, the environmental protection agency of the local authorities in the Rijnmond area, ordered RIVM to conduct a Lidar measurement campaign, to map benzene sources.This campaign was in addition to the air quality monitoring network that is present in the area.

METHODOLOGY
The measurement campaign was conducted with the RIVM mobile Lidar system.This uses the DIAL technique to make two-and threedimensional scans of the benzene concentration in the air.This can be used to map benzene plumes (see Figure 1).When combined with a measurement of wind speed and wind direction, an emission rate in kg/h can be determined.Figure 2 shows a schematic representation of the mobile DIAL system.The light source is a pump laserdye laser combination of a Spectra-Physics Quanta-Ray Pro 250 pulsed Nd:YAG laser pumping a Sirah PrecisionScan PRSC-D-30 dye laser.For benzene measurements, the dye Coumarin 307 is being used.This generates green light with a wavelength of 506 nm, which is frequency doubled to 253 nm.This UV light is separated off the residual green light and is used for the measurements.It has an energy of about 5 mJ/pulse, a linewidth of 0.0012 nm and a pulse duration of about 10 ns.The laser repetition rate is 30 Hz, pulses are tuned between the on and off wavelengths on each consecutive pulse.The UV laser beam has a divergence of about 0.5 mrad.The wavelength of each pulse is checked with a Cluster LM007 wavemeter.If the wavelength has drifted too much, the dye laser is tuned to the proper wavelength.The spectroscopy branch shown in Figure 2 is used to determine the effective molecular cross section of the gas being measured.The wavelengths selected for the measurements were 252.95052 nm as on wavelength and 252.83175 nm as off wavelength (both wavelengths in vacuum).A short literature study was performed to determine possible interference by a number of gases that may be present in the industrial environment where the measurements took place.Their spectra are shown in Figure 3, their relative sensitivities in Table 1.As can be seen, the interference by these compounds is negligible.The instrument is housed in a fully self-supporting measurement vehicle (Figure 4).It is able to measure sulphur dioxide [5], nitrogen dioxide [6] and ammonia [7].For the campaign described here, the ability to benzene was added to the instrument.Measuring a single two-dimensional concentration slice took 52 to 90 s, depending on the number of directions (the blue lines in Figure 1).During this campaign, vertical wind profiles were measured using a SODAR, Sound Detection And Ranging.
The SODAR used employed the Doppler shift of reflected sound pulses to construct vertical wind speed and wind direction profiles.See [8] for a discussion of the SODAR used in this campaign.
The current campaign encompassed ten measurement days, all in the Rijnmond area (Figure 5).This is a heavily industrialised area, with the extensive facilities of the port of Rotterdam, petrochemical and other industry, storage of petrochemical products and the associated transport movements.Residential areas are often in close proximity to industrial activities, especially in the east of the Rijnmond area.Measurements were in most cases conducted from public roads, but on a number of occasions the Lidar was positioned within the perimeter fence of an industrial site.Figure 6 shows an example of a measurement location.In this case, two measurement directions are used.Any benzene found in direction 1 but not in direction 2 is assumed to originate in the area between the two directions, as the wind will carry any benzene from the source through the measurement plane.

RESULTS
In Figure 7, an example of a benzene plume is shown.The Lidar measures concentrations between 300 and 900 m distance.The detection limit of the instrument was estimated to be 12 µg/m 3 .The highest concentration measured during the campaign was 3400 µg/m 3 .In Figure 8, an example is shown of an emission event.This event lasted only 25 minutes, other events lasted longer.Emission events were found on two of the ten measurement locations.The detection limit of the instrument for benzene emissions was estimated to be 1.6 kg/h.The highest emission measured was 25 kg/h.The total uncertainty of the emission rate was estimated to be 40% (1 sigma).Table 2 lists a number of contributions to this uncertainty.The largest source is the uncertainty in the representativity of the wind measurement, i.e. the observation that the wind field is disturbed by structures in the vicinity of the measurement slice.This may lead to a difference between the wind profile being measured and the wind profile at the position where the benzene concentration are measured.This contibution is difficult to quantify, it is recommended to conduct further research to reduce this source of uncertainty.

CONCLUSIONS
The Lidar was successfully deployed to measure benzene emissions.Combined with measurements of wind speed and wind direction, the Lidar measurements indicated the possible origins of these emissions.This makes the Lidar a valuable tool, augmenting the data collected at fixed monitoring stations.
The measured emissions were high, the highest instantaneous emission rate observed amounted to 10% of the average total emission rate in the Netherlands.However, because of their high variability in time, these emission rates cannot be extrapolated to a daily, monthly or yearly emission.

ACKNOWLEDGEMENT
The measurements described in this paper were commissioned by DCMR, the joint environmental protection agency of the province of South Holland and 16 municipalities in the Rijnmond area.

Figure 1 .
Figure 1.Measuring the emission from an industrial source.A mapping of the emitted plume is combined with measurements of wind speed and wind direction to determine an emission rate.

Figure 2 .
Figure 2. Schematic representation of the mobile Lidar DIAL system.

Figure 4 .
Figure 4.The Lidar in an industrial area.

Figure 5 .
Figure 5. Map of the Rijnmond area and its position in the Netherlands.The most heavily industrialised area is outlined in black.

Figure 6 .
Figure 6.Example location.The red lines indicate the measurement directions.The arrow shows the wind direction during the measurements.

Figure 7 .
Figure 7. Example of a benzene plume.

Figure 8 .
Figure 8. Example of the development of an emission event.Time is in minutes after the start of the event.

Table 2 .
Various sources of uncertainty and their contributions to the total uncertainty in the measured emission rate.