3. Results and discussion
3.1 Building structure and sampling regime
Fig. 1 depicts the spatial situation on the third floor below the water reservoir (drip floor) from which the boiler feed water was dispensed to railway steam engines. The circular room exhibits staircases from the second floor and to the fourth floor containing the reservoir tank for the water supply.
Figure 1: Layout of the third floor, PAH sources and sampling spots.
This building displays a complex air flow situation and was designed for a constant air exchange with the outside via an open lantern. The air exchange rate is unknown to its extend and within the building there was never any sensing of air movement by the authors. The maximum indoor temperature on the drip floor follows largely the outside temperature and ranged in 2019 between 5 °C in January and 30 °C in early July. It was decided to measure the airborne PAH concentration near the wall (Fig. 1, position P) as this would be the area basically used by museum visitors.
The PAH source is the original jute fabric soaked with a bituminous material wrapped around the steel tubes that promoted groundwater to the reservoir. This material and also the cardboard display a noticeable tar-like scent and releases PAHs contained in the soaked fabric by diffusion into the room and eventually to the outside. Table 1 comprises the content of the 16 most often quantified PAH congeners in the jute fabric. The PAH congener pattern is typical for hard coal tar22,23 with phenanthrene, fluoranthene and pyrene as most abundant congeners and with phenanthrene dominating over anthracene, benzo[b ]fluoranthene over benzo[k ]fluoranthene and indeno[1,2,3-cd] pyrene over dibenz[ah ]anthracene. The levels of naphthalene, acenaphthylene, and acenaphthene are probably largely reduced since production due to their higher volatility. The levels of airborne PAHs are influenced by complex processes including diffusion and evaporation from the partly more than 115 years old source material, and the interaction with the inorganic and organic construction materials such as concrete, brick, iron and wood, as well as the omnipresent dust particles.
Table 1 : Content of PAH congeners in the jute fabric (means and standard deviations, n=3) and respective vapor pressures at 25 °C
Therefore, in order to estimate the actual airborne PAH as basis for a risk assessment for visitors cannot be done using documented vapor pressure data for PAHs (Table 1) and diffusion models but requires measurement data. Since the building was not easily accessible during the restoration period and never meant to be accessed by visitors the technical execution of pump sampling to arrive at a PAH related risk assessment on the drip floor with the equipment described in section 2.3 was rather tedious. Thus, the situation was used to investigate if SBSE could offer an easily accessible complement to indoor risk assessment regarding harmful (semi)volatile substances. Therefore, it was firstly tested if there was any significant PAH extraction via SBSE devices and what time frame would be needed. A preliminary exposure of SBSE devices on the drip floor over up seven days revealed detectable amounts of all 16 PAH congeners already after one day of exposure. This encouraged two SBSE sampling campaigns, each displaying exposure periods of one day intervals up to five or seven days, respectively. The campaigns were run temporally apart (July 2019 and September 2019) such that the outside temperature influence could be assessed. On each measurement day three SBSE devices were sampled to enable estimation of analytical repeatability. In addition to the sampling on the wall (Fig. 1, position S1) SBSE devices were placed along a cross section of the room (Fig. 1, position S3) in order to assess the dependence of PAH content from the distance of the source (Fig. 1, positions A2). Likewise, pump sampling followed by quantification of airborne PAH levels was performed next to the site of SBSE sampling (Fig. 1, position P) but temporally apart. Again, measurements were performed with three replicates each on different days in order to evaluate repeatability of and temperature influence on determined airborne PAH concentrations.