Managing Exposure to Benzene and Total Petroleum Hydrocarbons at Two Oil Refineries 1977–2014

Air concentrations of and inhalation exposure to total petroleum hydrocarbons (TPH) and benzene was monitored separately at two oil refineries from 1977 to 2014. Prevention policies and control measures that may explain changes were surveyed. The aim was to evaluate how the application of of Occupational Health and Safety Assessment Series OHSAS 18001.04 principles as well as Environmental protection Agency EPA and European Oil Company Organisation for Environment, Health and Safety CONCAWE practices have influenced air concentrations. Benzene air concentrations declined in 11 of 17 units, six of which were associated with declining exposures. Benzene air concentrations declined across all units on average by 46%. This amounts to an average yearly decline of 1.7%. TPH air concentrations declined in 10 of 17 units, seven of which were associated with declining exposures. The average decline in TPH air concentrations was 49%, corresponding to 1.3% per year. As a result, average working day exposure in 10 of 17 units have declined significantly and today, benzene and TPH exposure in most units are well below 10% of the current Occupational Exposure Limit (OEL8h:s). A decline in air concentrations have coincided with consistent implementation of control measures. Such measures include on-line monitoring of leaks; benzene recovery; floating container roofs; improved valves and seals; hermetic pumps; recovery of loading gases and instalment of torches in terminals; cutback in coke combustion; a new production line spanning directly from the dock to aromatics production; and recovery of loading gases in the doc. Other tools in exposure management include personal leak monitors, on-line measurements, monitoring campaigns, risk assessment, and availability and user training of protective equipment. However, improvements are still needed. Hydrocarbon or benzene air concentrations have not declined in 8 of 17 units, in some of which concentrations exceed 10% of the relevant OEL8h value. In addition, for benzene even 10% of the current OEL, 0.1 ppm, might still possess a risk. With this in mind, methods to estimate exposure at the refineries need to be improved to enable measuring benzene concentrations <0.1 ppm. Shut downs of the refinery have been associated with peaks in exposure concentrations. Consequently, effort should be placed on safe working methods pertaining to shutdowns. Also, the connection and detachment of hoses continues to be problematic from the point of view of controlling exposure.

During regular operation conditions, exposure in modern refineries to benzene and other hydrocarbons is mostly below Finnish OEL 8h values or corresponding values stipulated by, for instance, EU, Occupational Safety and Health Administration (OSHA), National Institut Occupational and Security Health (NIOSH) or American Conference of Governmental Industrial Hygienists (ACGIH) [1,8]. Exposures of in-house personnel to benzene has been reported to be lower during turnarounds and shutdowns than routine production [4,9,10]. During normal operations, high benzene or TPH exposures have been recorded in the loading and unloading of raw oil and products, as well as sample collection, installing and removing of blinds, tank cleaning, pipefitting and welding, waste treatment, and the cleaning of equipment [4,8]. During shut downs and turnarounds, plumbing, scaffold work, welding, drain fitting, and general field work can be challenging with respect to managing benzene and TPH exposure [3,8].

Exposure to Benzene and Relevant Health Implications
Benzene concentration of crude oil can be up to 4 g/L [11]. Within the EU, the maximum benzene concentration in petrol is currently 1% (EU DIR 98/70/EY). Benzene is formed during heating and incomplete combustion of organic materials and as such, occurs ubiquitously in non-rural environments in concentrations below ca. 1.5 ppb [12].
Absorption of benzene through the lungs is rapid. The proportions absorbed are inversely related to the exposure concentration [6,7]. At exposure levels of~1-10 ppm approximately 50% of benzene is absorbed [13]. Short term (60 min) exposure to 50-150 ppm may cause headache, lassitude and weariness [14]. From epidemiological studies there is clear evidence on the association between benzene exposure and acute myeloid leukemia, acute non-lymphocytic leukemia, as well as myelodysplastic syndrome [15]. It has also shown to cause non-malignant haematotoxic effects, including leukopenia [16].

Background and Aims of the Study
The present study was initiated by the finding that workers employed at the refineries during 1967-1982, had a 2-3 fold risk of attaining kidney cancer, when compared to the general (reference) population [17,18], and that this increase may be associated with exposure to hydrocarbons (30-200 • C) and/or benzene [18]. These previous results reflect exposures up to the 1980s. As current exposures to hydrocarbons, including benzene, may be different, we wanted to examine air concentrations as well as inhalation exposure to benzene and TPH during recent decades . Particular interest was directed to the time-dependent development of concentrations and to control measures that may explain changes, when found. As all possible improvements are based on the safety management system of the company, we attained as much information as possible about measures taken that were initiated by corporate occupational health and safety strategies.
In short, the aim was to evaluate how the application of Occupational Health and Safety Assessment Series (OHSAS) 18001 and 18004 principles as well as Environmental protection Agency (EPA) and European Oil Company Organisation for Environment, Health and Safety (CONCAWE) exposure management practices have influenced air concentrations of and inhalation exposure to the examined agents.

The Company
The Neste Ltd. (Espoo, Finland) refineries in Porvoo and Naantali produce fuels, solvents, bitumen and lubricants, including the most important oil products. The refinery in Naantali was put into use in 1957, while the Porvoo refinery has operated since 1966. The Naantali refinery was originally designed to refine fuels, lubricants and bitumen products from crude oil, with an annual capacity of 800,000 tons. In 2014, the total capacity of oil products and renewable products of the two refineries were 15 million and 2 million tons per year, respectively. In 2014, the main product groups were motor gasoline and gasoline components (26%), diesel fuel (37%), renewable fuels (12%) and heavy fuel oil (8%).
The company has a certified Occupational Health and Safety management system based on OHSAS 18001 (2007) [19]. In addition to risk assessment and management, key components of safety management at the refineries include a work permit practice, safety discussions, safety observation tours, and inspections. Workers, including subcontractors, are yielded work permits only when they have been sufficiently trained in safe working methods. At the onset of a new production unit, relevant exposures are measured to either confirm the sufficiency of exposure management measures or assist in the planning of further measures.

Basic Exposure Monitoring Programmes, Online Monitoring of Emissions
Exposure monitoring in the refineries is based on annual monitoring programmes planned by HSE Services in cooperation with the manufacturing units and the occupational health care unit. HSE Services gathers all occupational hygiene samples in the refineries. The timetable of exposure monitoring following initial measurements at a new unit, are dependent on the concentrations in the initial measurements as well as the implementation schedule of exposure management measures. In each case, measurements are followed up within five years.
In addition to the units and working tasks included in the basic monitoring programs, exposure monitoring is performed in conjunction with accidents or production disturbances or changes with a potential to increase exposure.
To prevent excessive exposure and detect leaks, hydrocarbons and various flammable gases, carbon monoxide, hydrogen sulfide and oxygen are continually measured on-line using personal Leak Detection And Repair monitors (LDAR monitors) carried by all workers in the refinery area [20,21]. In addition, ca. 300 monitors are installed in stationary points in the refinery area, critical with respect to exposure and/or leaks.
Benzene concentrations in the production area are currently monitored by an on-line gas chromatograph connected to 11 stationary sampling points. The measuring range of the chromatograph is 0.5 ppm. The data is visible to the unit operators. The alarm limit pertaining to these measurements is 0.3 ppm.

Sample Collection and Analysis in the Present Study
Altogether 4588 samples were taken and analysed by Neste Ltd. HSE Services during the relevant time span   (Tables 2 and 3). Out of these 4588 samples, 1534 (33%) describe average working day exposure as measured from the breathing zone of workers not wearing respiratory protective equipment ( Table 3). The samples were collected indoors or outdoors, either from stationary points or from the breathing zone of workers at 18 different working units. The measurements included tasks from all units with the aim to measure air concentrations or to distinguish average 8 h (or 12 h) inhalation exposure during normal work shifts at the refineries. The samples were taken mainly in association with annual monitoring programs. Results describing average working day exposure were treated separately ( Table 3). All samples were withdrawn by occupational hygienists or technicians from Neste HSE Services.
The results that include breathing zone samples outside of respirators as well as stationary point samples not necessarily descriptive of average 8 or 12 h working day exposure are referred to as "air concentrations" throughout. Samples collected from stationary points were collected at a height of 1.5 m above the floor or ground.
Samples were withdrawn to SKC-226-01 charcoal adsorbent tubes at a sampling rate of ca. 0.1 L/min using SKC 222 (SKC Inc., Eighty Four, PA, USA) or Gilian LFS-113 DC (Sensidyne LP, St. Petersburg, FL, USA) sample pumps, or other SKC or Gillian (Sensidyne LP) pumps with a similar sampling range (0.05-0.2 L/min). Aliphatic and aromatic hydrocarbons with a boiling point of ca. 35-220 • C, were extracted from the adsorbent in 4 mL glass tubes with 1.5 mL of carbon disulfide. Main components in the samples were dependent on the production unit and work-task examined and included for example benzene, toluene, ethylbenzene, xylenes, methyl tert-butyl ether (MTBE), tert-amyl methyl ether (TAME), in addition to aliphatic hydrocarbons. Samples were analysed using a gas chromatograh equipped with Flame Ionization Detectors (FID) and one or two columns. If two columns were used, they were of dissimilar polarity characteristics: for instance, HP-InnoWax (30 m × 0.32 mm i.d., 0.5 µm; Agilent Technologies, Santa Clara, CA, USA) and HP 5 (30 m × 0.32 mm i.d., 1 µm; Agilent Technologies).  Compounds were identified and quantitated using pure calibrators at a minimum of three different concentrations, spanning the concentration range of 0.01-10 µL/sample. The Limit Of Quantitation (LOQ) of the analysis corresponded to the smallest calibrator and was generally ca. 8 µg/sample. For benzene, the detection limit was 13 µg/sample, corresponding to ca. 0.1 ppm for a sample withdrawn during a workday (7 h) with a sampling rate of 100 mL/min. Blanks and control standards (one in each run) were spiked onto sorbent tubes and prepared analogically to the samples.
Samples were kept refrigerated (−20 • C) until analysis. The samples were analysed with an HP or Agilent 6890 gas chromatograph operated using MS Chemstation software (Hewlett-Packard, Palo Alto, CA, USA), and equipped with two FID detectors, one for each column. The operating conditions were set and the results calculated as described in the reference methods NIOSH 1500-1501.

Testing the Time-Dependency of the Concentration Measurement and Calculating the Change during the Measured Time Interval
A linear regression analysis was performed on the sampling date (x) and concentration (y) data.
The Pearson correlation coefficient was tested against its sample-size dependent p-value (critical value, 95% level of significance, two tailed test) as described by Warner (2013) [22]: ρ x,y = √ (t 2 /(t 2 + n − 2)) at the t-test critical value (t) corresponding to the relevant degree of freedom. The test was performed to distinguish units where the concentration of benzene and/or total volatile hydrocarbons significantly increased or decreased as a function of the date of sampling during the time-interval of the measurements, from those where a statistically significant change could not be verified.
If a linear relationship (sampling date vs. concentration) was established, the proportional (%) change in concentrations was estimated by comparing the value at the starting point of the regression interval (x 1 ,y 1 ) to the concentration at the end point (x 2 ,y 2 ):%Ch = (y 1 − y 2 )/y 1 × 100.

Production Lines 1-5
The production lines (1-5) differ in their output and also in the raw materials used and, consequently, in their benzene emissions as well as the amount and profile of volatile hydrocarbons emitted to their immediate environment. Production lines 1-3 produce mainly gasoline and production line 4 mainly diesel fuel. Production line 5 yields bitumen, small engine gasoline and solvents, in addition to diesel fuel. Production lines 1-4 are situated in Porvoo, while line 5 is in Naantali.

Production Lines 1-5
Characteristic for the production lines in Porvoo (lines 1-4) is that while accidents, shut downs and operating disturbances have on occasion led to high benzene and TPH concentrations, the air concentrations during normal working shifts have been fairly low during all of the years studied (Figures 1 and 2 and Table 2), with the inhalation exposures for the most part being close to or similar to ambient concentrations (Table 3). Average and median benzene air concentrations during 37 years in production lines 1-4 were 0.36 ppm and 0.05 ppm, respectively, while the average and median TPH concentrations were 17 mg/m 3 and 1 mg/m 3 , respectively. The medians of the air concentrations amount to 5% (benzene) and 0.5% (TPH) of the current OEL 8h values, 1 ppm for benzene and 200 mg/m 3 for TPH.
In the Naantali production line (line 5), the average and median benzene air concentrations have been of the same magnitude as in the Porvoo production lines, namely 0.35 and 0.05 ppm, respectively. The average and median TPH concentrations have, however, been notably higher in production line 5 than in other production lines: average 56 mg/m 3 , median 15 mg/m 3 (7.5% of the current TPH OEL 8h value). In production lines 1-5, benzene concentrations have very rarely reached concentrations that necessitate the use of respiratory protective equipment. In fact, in production lines 2 and 4, no such incidents have been monitored (Table 2). partially explaining the low concentrations encountered. In addition, line 4 is the newest production line, most likely benefitting from the use of more recent technical solutions and components. During recent years (2008-2014), the average of benzene air concentrations in all production lines other than line 3 have been consistently less than 10% of the current OEL8h value (Table 4). In line 3, the 95% point is 32% of the OEL8h value. As benzene exposure, for the most part, have been low in all production lines, declining (5%) benzene exposures could be established only in line 5, where the initial exposures were higher than in the more recently built production lines in Porvoo (Table 3).    In production lines 1, 3 and 5, the decline in measured benzene air concentrations have been steep starting with 1977. In line 2, we see a clear but nonetheless not as steep decline in air concentrations with time. In line 4, no apparent change in air concentrations can be observed, but then again, the concentrations were low from the start ( Table 2). Line 4 produces mainly diesel fuels low in benzene, partially explaining the low concentrations encountered. In addition, line 4 is the newest production line, most likely benefitting from the use of more recent technical solutions and components. During recent years (2008-2014), the average of benzene air concentrations in all production lines other than line 3 have been consistently less than 10% of the current OEL 8h value (Table 4). In line 3, the 95% point is 32% of the OEL 8h value. As benzene exposure, for the most part, have been low in all production lines, declining (5%) benzene exposures could be established only in line 5, where the initial exposures were higher than in the more recently built production lines in Porvoo (Table 3). TPH air concentrations have declined in all production lines except line 4. However, particularly in production lines 1, 3 and 5, much remains to be done in order to reach TPH air concentrations similar to production line 4: systematically less than 10% of OEL 8h (Table 5). Line 4 produces a higher-boiling hydrocarbon mixture (diesel) than production lines 1-4, while line 5 produces solvents and other special products, including fuel oil and bitumen, in addition to diesel. This may in part explain why high concentrations of hydrocarbons with low boiling points may occur there (Tables 2 and 5). In addition, some of the pumps distributing input, output and side currents, are located in hangars in line 5 as opposed to outdoors in lines 1-4. Similar to benzene, a significant decline in TPH exposure (86%) could be established only in line 5 (Table 3).

Tank Yard, Wastewater Treatment, Maintenance and Technology Center (Porvoo)
These units represent diverse functions with few if any common denominators other than the fact that they all serve production units 1-4. Of these units, only wastewater treatment show a clearly declining trend in air concentrations ( Table 2).
Tank yard is responsible for blending the feeds from the different production lines to finished products and may be susceptible to stray leaks and emissions particularly in association with modifications in finished products, as well as during shut downs and maintenance operations. During the measured time span , benzene as well as TPH air concentrations have occasionally reached values closed to prevailing OEL 8h values in the tank yard. Even though only one percent of measurements in the tank yard have exceeded OEL 8h , this has resulted in average concentrations close to the current OEL 8h , while the medians are low in comparison (Tables 4 and 5). During recent years, starting with 2008, the 95% percentile of the air concentrations of both benzene and TPH in the tank yard have remained clearly above 10% of the current OEL 8h .  Tables 2 and 3). The wastewater treatment unit is continuously burdened with streams of high hydrocarbon content. Particularly benzene seems to represent a challenge, partially due to polymer production units located in the same factory area as the oil refinery. Benzene concentrations higher than the present OEL 8h were recorded relatively frequently up until 2008 but during recent years, benzene air concentrations have usually remained below 10% of the OEL 8h ( Table 4).
The Technology Center contains R&D units, as well as analytical and measuring services. Different solvents are used or produced in laboratory and/or pilot-scale operations. Benzene exposure has not been an issue for the past ten years in the Technology Center, but the TPH air concentrations have increased from 1977 to 2014. However, since 2004 high concentrations of TPH (more than 50% of current OEL 8h ) have been measured only in conjunction with two isolated incidents not descriptive of average working day exposure: a temporary operational break in an exhaust ventilator in a cool storage room and the uncontrolled purification of equipment using hexane (Tables 2, 4 and 5).
In the maintenance unit, benzene air concentrations-as well as exposure-has been low throughout the years, while TPH air concentrations and exposure show an increase based on the calculations (Tables 2 and 3). However, TPH air concentrations, and hence exposure, have never been found to exceed 10% of the current TPH OEL 8h . Highest average working day exposures (10.18 mg/m 3 ) have been measured in association with the application of corrosion inhibitors. In other words, TPH exposures in the maintenance functions were fairly low to begin with and have remained so during the measured time interval.

Harbour, and Distribution and Unloading Terminals (Porvoo)
Harbour functions as well as the distribution and unloading of petrochemicals include tasks that yield high air concentrations and are repeated several times during a normal working day. As high immediate benzene concentrations may be relevant to the manifestation of health implications, it is important that all units where benzene emissions is of concern, show a declining trend with time. This has been the case with the benzene air concentrations in the harbour functions (relative decline of measured concentrations 49%), as well as the distribution terminal (relative decline 99%, Table 2). However, in the unloading terminal benzene air concentrations or exposure have not declined, but then again, initial concentrations were lower there (Tables 2 and 3). Further, since 2008 concentrations exceeding 10% of the current benzene OEL 8h has been measured only once in the unloading terminal, from the breathing zone of an operator wearing a half mask.
While hydrocarbon air concentrations show a declining trend with time only in the distribution terminal, the concentrations during recent years have for the most part been low to moderate. The highest TPH concentrations have been recorded in association with the loading of trains, 1.220 mg/m 3 , corresponding to 0.5.110% of the current Finnish OEL 8h .
In the distribution terminal, air concentrations of both benzene and TPH were high up until 1993. After the maximum concentration of benzene in petrol was lowered from 2 to 1% in 1994, air concentrations have been lower. In the loading of ships, recovery of gaseous emissions was taken into use in 2014, the effect of which will be seen during upcoming years.

Naantali: Tank Yard, Harbour, Distribution Terminal, Wastewater Treatment and Maintenance
In the Naantali production line as well as in the tank yard and wastewater treatment functions, we see a clear decline in the air concentrations with time, resulting in a reduction of exposure as well (Tables 2 and 3). Most other measured concentrations in these units have reduced throughout the years as well. Exceptions to this trend are benzene air concentrations in the distribution terminal, benzene exposure and air concentrations in the maintenance functions, benzene exposure in the harbour, and TPH exposure and air concentrations in the harbour. In previous years, benzene and particularly TPH OEL 8h values have frequently been exceeded in the tank yard as well as in the distribution terminal. During recent years, 2005 and onwards, concentrations higher than OEL 8h have been recorded only in the distribution terminal. In units other than the distribution terminal, including the wastewater treatment unit, measured concentrations of benzene as well as other hydrocarbons have for the most part been fairly low (Tables 4 and 5).
In the distribution terminal, high concentrations of TPH, have been recorded in conjunction with vehicle dispatching. Measurements describing average working day exposure have, however, yielded low results in all units.

Discussion
In general, benzene and hydrocarbon exposure in Finnish industries have declined during the past decades. For instance, according to the FINJEM-database, the number of workers exposed to benzene have decreased more than 10-fold concurrently with the present study: from 10,013 (1960-1984) to 593 (2007-2009) [23]. On average, benzene exposure in industrial workplaces in Finland declined 52% and total volatile hydrocarbon exposure (category two solvent petrols) 99% during 1994-2011, as estimated from FIOH service measurements. Published reports by the CONCAWE indicate a similar development in oil refining in general [24]. In investigations on the benzene exposure at ExxonMobil refineries in Louisiana and Texas, up until 2006, similar mean air concentrations were reported, i.e., 0.01-0.4 ppm [4,9,10]. In those studies, from 1976 to 2007 (Texas) and 1977-2005 (Louisiana), declining exposures were established for job categories "laboratory technician" (Texas) and "process technician", excluding waste treatment area (Louisiana). However, no precise quantitative estimates are available enabling a comparison to the exposure measurements performed by Neste Ltd. during the whole time frame 1977-2014.
In the present study, benzene air concentrations declined in 11 out of 17 units, six of which were associated with a significant decline in benzene exposure as well. Benzene air concentrations declined across all units on average by 46%. This amounts to an average decline of 1.7% per year. TPH air concentrations declined in 12 out of 17 units, seven of which were associated with a decline in exposure. Across all units, the average decline in TPH air concentrations was 49%, corresponding to 1.3% per year.
In all sites apart from the Porvoo harbor and ships, benzene median values or the level at the end points of the regression curves are at or below 0.1 ppm. Using an old SCOEL (1991) estimate on cancer risk this represents a cancer risk in the range of 0.05-0.66 per 1000 in 40 years occupational exposure [6]. This risk estimate is based on the Crump and Allen analysis of so-called Plioform cohort and assumes linear dose response [25]. It should, however, be noticed that there are large uncertainties related to the magnitude of benzene caused cancer risk at low exposure levels. Different analyses on the dose-response of the carcinogenicity of benzene at these low levels has been published since then. For example, the German AGS derived so-called "tolerable cancer risk level" of 0.6 ppm resulting in four extra cancers per 1000 exposed workers, and an "acceptable cancer risk level" of 0.06 ppm resulting in four extra cancers per 10,000 exposed workers based on linear extrapolation from the dose-response data derived from several epidemiological studies. On the other hand, Dutch DECOS Committee recently proposed an OEL of 0.2 ppm for benzene based on the haematotoxicity of benzene [7]. Benzene was considered as a non-stochastic genotoxic carcinogen for which a threshold for carcinogenicity could be established. Since haematotoxicity often precedes benzene caused leukemia, preventing haematotoxicity was considered to minimise the risk of leukaemia [7]. In US, ACGIH has concluded that at a TLV level of 0.5 ppm odds of death from leukemia due to occupational benzene exposure would be indistinguishable from the odds of death from leukemia of non-exposed workers [26]. This conclusion was based on Plioform data. Overall, although no definitive conclusions can be made on the cancer risks at the levels below 0.1 ppm, it is advisable to minimize the exposure to benzene as low as reasonably achievable. In the wastewater treatment units a 7.28-fold decrease in exposure levels during the follow-up period was recorded. This is likely to have a significant impact on cancer risk as well. In most sites (10 out of 17) however, benzene exposure was below 0.1 ppm to begin with (<10% of OEL 8h ). In these units, no statistically significant decline in exposure could be recorded, regardless of whether the air concentrations declined or not (Tables 2 and 3).
As the present investigation spans nearly four decades, numerous outside factors have affected work related exposure to hydrocarbons in the refineries. Such factors include (effect, either positive or negative, in captions): lowering the maximum benzene concentration allowed in petrol gradually: 3.5% in 1980, 2.5% in 1985, 2% 1991, 1% in 1994 (−); increase of benzene-rich effluents from polymer production units to wastewater treatment plant in Porvoo (+); changes in production capacity and product portfolio (±); technological progress and better process engineering techniques (−); new legislation concerning safety and risk management as well as cancer registry (−); stricter OEL 8h values (−). In addition, measuring techniques and analysis methods of organic emissions have improved, and new techniques and equipment have become available enabling online measurement of air concentrations and in some cases even exposure. Consequently, it is likely that in 2014 we get a more reliable picture of air concentrations and work related exposure at the refineries, than we did in 1977.
In managing exposure, particular effort should be placed on control of emissions at the source [27]. In the refineries, such measures include careful planning of new production lines and process technological solutions. However, much can and has been done to improve conditions in older units as well, as can be seen from the development of air concentrations in production line 5 (Naantali). Providing other control measures are inadequate to prevent excessive exposure, the use of protective equipment is necessary. Neste Ltd. has a corporate policy for the use of protective equipment. Enforcing and supporting the use of protective equipment is often problematic. In fact, it has been suggested that "personal protective equipment should be used only sparingly and under appropriate circumstances, and never as a substitute for more reliable and effective engineering or administrative controls" [28]. In other words-and with reference to the Neste Ltd. policy for the use of protective equipment-the goals should be to lower benzene air concentrations below 10% of the current benzene OEL 8h and TPH air concentrations below 50% of the current TPH OEL 8h . This would help to focus the need for protective equipment mainly to infrequent tasks short in duration.

Conclusions
In conclusion, air concentrations of TPH and benzene have generally declined with time . In more than half of the units where a decrease in air concentrations were established, this decrease was accompanied by a decrease in exposure. When a decrease in exposure could not be established, the exposure was without exception less than 10% of the current OEL 8h to begin with.
However, there is still a need for improvements in many of the units as benzene concentrations below 0.1 ppm may still possess a risk to exposed workers. It should also be noted that according to some studies (repeated) short term (≤15 min) peak exposures may play a role in benzene carcinogenesis [29]. Therefore, consideration should be given also to minimizing leaks resulting in short term peak exposures. Further, benzene and/or TPH air concentrations did not improve in 8 out of 17 of the units, some of which are associated with average air concentrations exceeding 10% of the relevant OEL 8h value (Tables 4 and 5). Considering the harmfulness of the exposure agents in question, the average 8 h benzene exposure and air concentrations of benzene should as a minimum not exceed 10% of the OEL 8h , while the average air concentrations to TPH as well as TPH exposure should be kept below 50% of the OEL 8h . In addition, in no circumstances should the workers be continually exposed to concentrations exceeding OEL 8h when not protected. With this in mind, and considering that the OEL 8h of benzene will possibly be lowered in the near future, we suggest that when possible, the highest benzene air concentrations should be lowered in the following operations and units (in order of priority, see Table 4 Regardless of the unit or exposure agent, the maintenance shut downs of the refineries (in Porvoo 1993Porvoo , 1997Porvoo , 2001Porvoo , 2005Porvoo , 2010in Naantali 2000in Naantali , 2006in Naantali , 2012 have generally been associated with clear peaks in the air concentrations. Consequently, particular effort should be placed on safe working methods pertaining to shutdowns as well as the use and condition of respiratory protective equipment of subcontractors and in-house workers alike. Also, while recovery of loading gases in the distribution terminals and docks will most likely lower air concentrations in these units, the connection and detachment of hoses continue to be challenging from the point of view of controlling excessive exposure.
Benzene and hydrocarbon air concentrations in some of the units investigated will most likely continue to have room for improvement for years to come. In the Neste refineries, personal LDAR-monitors and on-line measurements, monitoring campaigns and risk assessment procedures, availability of protective equipment as well as training programs are effective tools in the monitoring and management of exposure. However, ensuring safe working conditions and working methods will continue to require persistent work. Many problematic work tasks and situations will still have to be addressed in order to improve high air concentrations and further lessen the need for protective equipment. Also, as an instrument in lowering exposure with recent risk estimates in mind, methods to estimate exposure at the refineries need to be improved to enable measuring benzene concentrations <0.1 ppm.