Joint treatment of landfarming and bioventing in karst soils contaminated by petroleum hydrocarbons

Transcription

1 Joint treatment of landfarming and bioventing in karst soils contaminated by petroleum hydrocarbons F. Cangialosi, G. Intini, A. Lattarulo, L. Liberti & M. Notarnicola Department of Environmental Engineering and Sustainable Development, Technical University of Bari, Taranto, Italy Abstract A case study of bioremediation of petroleum hydrocarbon contaminated soils of a dismissed oil refinery in the industrial area of Bari (S. Italy) is presented. The soil and soil gas quality data collected in the characterization activities indicated that the shallow sediments were mostly contaminated by sorbed hydrocarbons whereas volatile hydrocarbons were found in the fractures of the rocks, affected by karst phenomena, underlying the superficial soil. According to laboratory and field tests, in-situ landfarming was chosen to clean up the less contaminated and shallow soil volumes; the zones with highly contaminated soil were removed and cleaned up by on-site landfarming operations in a large concrete basin. Results from field tests and full-scale one-year-monitored bioventing system for the remediation of the soil gas from the deep vadose zone are also shown. Complete remediation of superficial soils decontaminated with the in-situ landfarming actions had occurred in 9 months, thanks to the engineered optimization of the suitable environmental conditions for enhancing the biomass growth. A faster activation of the microbial flora in the basin used for the on-site landfarming operations allowed a substantial reduction of TPH concentrations in soils from an average of 3500 mg/kg to less than 1000 mg/kg following 4 months of landfarming. Effectiveness of bioventing in karst soils is evaluated in terms of its impact on mass removal by biodegradation and volatilization. Keywords: hydrocarbons, bioventing, landfarming, karst soil, site remediation.

2 230 Geo-Environment 1 Introduction The environmental reclamation requires the assessment of suitable techniques, in relation to the nature of the compromised environmental matrix to clean up and the kind of contamination to remove. Generally speaking, bioremediation of a contaminated site is based on a continuous pollutants transformation sequence in simpler by-products, obtained in consequence of appropriately activated bacterium consortia [1]. A leading role in the activation and control of the microbiological process is played by physico-chemical parameters [2]: - soil temperature, which has to be included in a range from 10 C to 45 C - relative humidity, which has to be close to 80% of the field capacity, between 15% and 20% of the treated soil volume, in order to facilitate oxygen supply within the subsoil and enhance macronutrients (N, P, K) solubilization, when they are supplied in solid form - ph has to be between 6 and 9 - Redox potential in the soil, which shows the acclimatization kinetics of the microbial flora in the treated soil In the present paper the site remediation process of a dismissed oil refinery by means of biological treatments on surficial soils contaminated by sorbed petroleum compounds (landfarming) and deeper vadose zone soils impacted by vapor phase hydrocarbons (bioventing) is described. 2 Site characterization The investigated site, placed in the industrial area of Bari (S.Italy), was the seat of an oil refinery and afterwards an oil store tank. Stratigraphical and tectonic studies, related to the local lithology of the industrial area, show the presence of bioclastic Pleistocene deposits in calcareous facies in continuity on a level of brackish deposits in transgression on Mesozoic limestone deposits known as Calcari di Bari : they constitute the rigid basement of the area and are affected by karst phenomena revealed by interbedded cavities and fractures, often filled with terra rossa. During the investigation period, the groundwater level fluctuated seasonally over a 3 meters vertical interval (from 12 m to 15 m) under aquifer recharge and depletion phenomena. On the basis of historical information related to the industrial activity carried out in the industrial area, as well as the geological and hydrogeological data collected, a contamination by petroleum hydrocarbons spill from tanks and oil pipelines was assessed. The finer fractions in the shallow sediments were mostly affected by contaminants because of their higher surface area and clay content, thus enhancing the hydrophobicity towards organic compounds. In order to define the extent of soil contamination, an analytical survey has been arranged to detect TPH (Total Petroleum Hydrocarbons) in the surficial soils; chemical analyses have been carried out on 73 out of 110 investigated

3 Geo-Environment 231 samples in order to measure hydrocarbons concentrations (C2-12 and C12-20), BTEX (Benzene, Toluene, Ethyl Benzene, Xylene) and PAHs. The outcomes of the chemical analyses pointed out the presence of a widespread contamination in the refinery s central area and two contaminated areas located in the North and Northeast part (Fig.1). Since the draft of the remediation project preceded the coming into force of the national law on contaminated sites (Environment Ministry s Decree no. 471/99) [3], clean-up targets of Lombardia Region (Decree no. 6/96) [4] were adopted, which provide 1000 ppm as maximum acceptable value of TPH for site located in an industrial area with a freatic aquifer. The poor sorption capacity of calcareous rocks in the vadose zone suggested a partitioning in vapour phase (Volatile Organic Compounds, VOC) of NAPLs (Non Aqueous Phase Liquids) trapped as residual hydrocarbon in the fractures of deep soil. A monitoring wells system has been used to carry out a soil gas survey, measuring VOC, O 2 and CO 2 concentrations. A significant contamination was found in the central part of the site: very high concentrations of VOC (more than ppm) were measured in the deeper soils, while low O 2 concentrations and high CO 2 values pointed out a previous biological reclamation (natural attenuation) halted owing to the lack of oxygen. Figure 1: Site contamination by C 2 C 12 hydrocarbons (mg/kg). 3 Experimental plan In Figure 2 the areas of the site to be remediated by in situ and on site landfarming and bioventing treatments are shown. 3.1 Landfarming The less impacted surficial soils were remediated by means of in situ landfarming which took place trough the execution of operations similar to those

4 232 Geo-Environment regularly in use in agronomic practices for increasing the rate of aeration: soil tilling, periodic ploughing (every 45 days) and daily watering. Periodic ploughings, in particular, have allowed the disrupting of clods created during watering and evaporation processes. In landfarming, the health of the microorganisms is critical to the success of the process. Controlling parameters are those constituent concentrations or environmental conditions (ph, moisture, nutrients etc..) which act to limit the process and which can be manipulated to optimize the degradation of hydrocarbons. In order to ensure the bioavailability of macronutrients for the microbial degradation of organic compounds, nitrogen and phosphorus were supplied in solid form as conventional fertilizer and solubilized with the aid of rain watering installations (510 m 3 /d), useful as well for allowing O 2 enrichment in the soil. In situ treatment has involved an overall surface of 10 ha divided in six lots ranging from 0.05 km 2 (area 5) to 0.3 km 2 (area 4) (see Figure 2). Figure 2: Site areas interested by bioremediation. In the 6 treated areas, a trimestral monitoring program of 24 soil samples collection (0.25 kg per sample) has been carried out in order to assess in situ landfarming performances by measuring TPH, total aromatic hydrocarbons, nitrogen and phosphorus concentrations, as well as relative humidity, ph and by counting the microbial population (two samples for each area) [5]. On six of the twelve collected samples, counts of specific hydrocarbon degrader bacteria were performed. Nutrients (N and P) were monitored in the soil using field test kits commonly used in agriculture. On site treatment took place in a concrete basin 0.1 km 2 large, covered with an impermeable liner according to the Italian technical guidelines for solid waste landfills. The heavily contaminated surficial soils were excavated to a depth of 3 m, stockpiled for a volume of 2000 m 3 for each cycle of treatment and stored in the basin to an average depth of 30 cm. These preliminary operations have favoured the creation of an aerobic environment to support the degradation process. For soil treatment in the basin, monitoring plan has been carried out by analysing 10 samples took from the nodes of a grid 25 m wide.

5 Geo-Environment Bioventing Due to the large surface extension (8 ha) of the subsoil contaminated by VOC and the depth of the strata (over 10 m) to be reclaimed, the remediation plan was implemented by bioventilation treatment. Bioventing (BV) allows to remediate a great volume of soil by stimulating natural biodegradation of petroleum hydrocarbons by indigenous aerobic biomass, through oxygen supplying by means of air injection directly in the contaminated subsoil. Factors affecting the success of a BV approach as a remediation option include the characteristics of the contaminants, the properties of the soil, and site-specific hydrogeologic conditions [6]. Pilot scale tests are useful before full-scale application of this treatment for evaluation of its suitability for a specific contaminated site [7]. In order to evaluate the feasibility of bioventing for the remediation of deep vadose zone soil, extending from 2 m below ground surface (bgs) to depths ranging from 8 to 15 m, air permeability and in situ respiration tests were performed. The pilot test equipment consisted of an injection unit, venting and monitoring wells, pressure transducers, and gas composition monitoring systems. Two blowers (power of 3 and 5 hp), equipped with digital flow controllers, were connected to two venting wells (VW) via 2-inches underground PVC pipes. The venting wells were equipped with a screened interval from 5 to 15 m bgs, whereas the five monitoring wells (MW) were equipped with multiple-depths monitoring points (MP). Three screens were installed at discrete depth intervals between 6 and 14 m bgs: pressures at the monitoring wells were measured using an automatic logging system and soil gases such as O 2, CO 2, and VOC were measured using a landfill gas analyzer. The main concerns about the feasibility of bioventing in the deeper soils were related to the distribution of flow paths and diffusion of air within the fractured limestone. For these reasons the radius of pressure influence (ROI p ), defined as distance from the injection well where it is possible inducing a pressure variation in the soil, was considered the limiting factor for the success of the treatment. Two air permeability tests were conducted by injecting air into VW1 and measuring the pressure response at the 15 monitoring points. From the regression analysis of the data collected 10 hours after the test start-up, average values of 45 m and 20 m were assumed for the ROI p of the two blowers. Due to the rapid pressure response, a steady-state onedimensional flow equation for determining air permeability with both injection systems was selected [8]: a permeability value of m 2 was calculated for the test area. With the same wells configuration, in situ respiration tests were performed in order to provide field measurement of biodegradation rate. After air had been injecting in VW1 for 6 weeks, the blower was turned off and the oxygen loss in soil gas composition was measured at the MP screened intervals. When oxygen concentrations drop to below 5%, the oxygen itself is limiting for the hydrocarbons degradation; oxygen utilization rates were then computed on the basis of a zero-order kinetic for concentrations above 5% [9]. The oxygen utilization rate ranged from 2.9 to 6.7 %/d with an average value of 3.9 %/d, indicating that bioventing may be feasible at the site [10, 11].

6 234 Geo-Environment N-hexane was selected as a representative compound of the dominant contamination and the mass ratio of hydrocarbons to oxygen required for mineralization is 1/3.5. This ratio probably underestimates hydrocarbons biodegradation because almost a third of hydrocarbons may be used in cell production rather than CO 2 production [1]. Based on the above oxygen utilization rate and an air-filled porosity of 0.3, 2.7 milligrams of n-hexane per kilogram of soil could be degraded each day at this site. After these preliminary tests, the experimentation has been carried out for: - verifying the regular ventilation of all the subsoil interested by volatile compounds contamination; - monitoring the biological process with blowers turned on, to verify the occurrence of the conditions to activate the biomass and its growth. 4 Results and discussion 4.1 Landfarming With regard to in-situ treatment, the first analytical survey after three months from the process start-up has pointed out a sufficient development of the microbiological process (Tab.1). Table 1: In-situ landfarming: main performance data. TOTAL MICROBIAL HYDROCARBON TPH AROMATIC POPULATION DEGRADER AREA POINT HYDROCARBONS BACTERIA baseline 3 months baseline 3 months baseline 3 months baseline 3 months ppm ppm ppm ppm CFU/g CFU/g CFU/g CFU/g POINT POINT < 0, ND ND ND ND 1 POINT < 0, ND ND POINT < 0,4 < 0,4 ND ND ND ND POINT ND ND ND ND POINT POINT ND ND ND ND POINT < 0,4 < 0, ND ND POINT < 0, POINT ND ND ND ND 3 POINT ND ND POINT ND ND ND ND POINT < 0, POINT ND ND ND ND 4 POINT ND ND POINT ND ND ND ND POINT POINT ND ND ND ND 5 POINT ND ND POINT ND ND ND ND POINT < 0, POINT ND ND ND ND 6 POINT ND ND ND ND POINT ND ND ND = Not Detected Particularly in the area 3 it is evident an order of magnitude increase of hydrocarbons oxidant biomass which is almost doubled in areas 1, 2, and 5. On the other hand, in areas 4 and 6 a decrease of hydrocarbon degrader bacteria was

7 Geo-Environment 235 observed, probably related to depletion rate of contaminants indispensable to bacterial proliferation thus limiting the microbial growth. After only three months of landfarming remediation progress was satisfactory, since the number of hot spots passed from 6 to 4 and the maximum TPH soil concentration decreased from ppm to 7500 ppm. The moisture content, monitored at middle depth of the tillage zone, was more than 10% in every investigated point, thanks to watering operations. Both ammonia-n and nitrogen oxide-n, parameters controlling biomass development, showed values included in the optimum variation range. Decontamination levels achieved after 9 months indicated that in situ treatment was concluded earlier than the period estimated on laboratory tests basis (12-14 months). In on-site landfarming treatment carried out in the concrete basin a rapid development of biodegradation process was observed. Measurements of the microbial population on 4 soil samples showed that the microbial growth was in its conclusive stage after only 2 months of activity. The total biomass increase was to be related with the high decontamination rate achieved. The hydrocarbons degradation rates in the landfarming basin pointed out a positive trend: after 70 days all residual TPH concentrations were under 1000 ppm (Fig. 3) point 4 point 6 point 2 point 8 TPH (mg/kg) time (days) Figure 3: On-site landfarming: TPH degradation in surficial soils. The renewal of the global biological activity, before limited by the presence of contaminants, became possible thanks to reacquisition from the soil of appropriate agronomic characteristics [12]. It s noteworthy that after 4 months of treatment the soil was suitable for residential use of the site, since all TPH residual concentrations on 20 samples uniformly took from the on site treatment area were lower than MAC provided by Italian Decree no. 471/ Bioventing Based on the results of pilot tests, a bioventing system was designed to address the entire 8 ha area to a depth of 16 m; 32 air injection wells at a rate of

8 236 Geo-Environment approximately 300 m 3 /h have been drilled to obtain a ROI p of about 50 m. The air has been injected at the temperature of 55 C to allow subsoil mean temperature raising, improving environmental conditions for bioremediation start-up. Operation has been optimized via data generated through sampling of 20 multi-depth (3, 8 and 13 m) monitoring points in order to determine bioventing biodegradation rate. Unlike what happens in the granular soils, in the case of limestones, activated bacteria are present within fractures of the rocks interested by the presence of hydrocarbons, and bioremediation rate has to be deduced by soil gas data. Therefore a correct monitoring of the reclamation trend cannot be carried out through single-point analytical survey on soil samples due to the poor sorption capacity of calcareous rocks, but considering the presence of O 2, CO 2 and VOC in soil gas. Given the particular nature of the subsoil, one of the major doubt about the effectiveness on field scale was linked to the real possibility of venting all the contaminated soil volume: limestone permeability, in fact, could produce some preferential paths which could not allow the ventilation of the whole rocky block. Prior to system operation starting, the entire impacted area was nearly lacking in oxygen (interstitial pressure less than atm) and with an high CO 2 /O 2 ratio, showing former natural attenuation processes in act. 45 days after the system start-up, the oxygen concentration over the whole area increased to more than 15% and the bioremediation rates began to increase slowly (Fig. 4) O2 and CO2 in soil gas (%) CO2-3 m CO2-8 m O2-13 m O2-3 m O2-8 m CO2-13 m elapsed time (days) Figure 4: Bioventing: O 2 and CO 2 concentration in soil gas (multiple-depth monitoring survey with active blowers). At the same time a significant decrease in CO 2 quantity in all the 3 multipledepths monitoring points from 10% (in some cases even over 15%) down to values higher than CO 2 concentration in the atmospheric air (0.03%), used for subsoil ventilation, was detected. The above-described situation remained unchanged until the last monitoring survey, after almost one year from the beginning of the air injection. VOC concentrations have been measured monthly with active blowers for 12 months: a first order kinetic degradation [12] was observed, with residual values next to 0 at 8m deep and a decreasing trend not stabilized at 3m and 13m deep (Fig. 5).

9 Geo-Environment VOC (ppm V) elapsed time (days) Figure 5: Bioventing: VOC concentration in soil gas (multiple-depth monitoring survey with active blowers). Biodegradation rate has been estimated through oxygen requirement by the biomass. It must be pointed out that during the third of four respiration tests we assisted at a small increase of the oxygen demand within the surficial soil, with a more significant increase within the deeper clusters, caused by raising of water table level during spring months. During bioventing activity a seasonal variation of the oxygen consumption in biomass degradation was showed, more pronounced during summer months in the upper clusters (3m and 8m) owing to the intense microbial activity. By the end of first year of operation an estimated 2,700 ton of hydrocarbon have been biodegraded in CO 2 and H 2 O, with an oxygen consumption by biomass of about 7700 ton of oxygen. 5 Conclusions Bioventing and landfarming were successfully performed to provide an effective, integrated approach to bioremediation of petroleum hydrocarbon contaminated soils of a dismissed oil refinery in the industrial area of Bari (S. Italy). The results can be summarized as follows: A complete remediation of superficial soils decontaminated with in situ landfarming action has occurred in 9 months thanks to the execution of operations similar to those regularly in use in agronomic practices for increasing the rate of aeration: soil tilling, periodic ploughing and daily watering. A faster activation of the microbial flora in the on-site landfarming basin allowed a substantial reduction of TPH concentration in heavily contaminated surficial soils from an average of 3500 mg/kg to less than 1000 mg/kg after 4 months of treatment. The remediation of the subsoil contaminated by VOC was implemented by bioventilation treatment. After almost one year from the beginning of the air injection, a decreasing in CO 2 from over 10% down to values higher than CO 2 concentration in the atmospheric air was observed. Measurements

10 238 Geo-Environment carried out with active blowers, moreover, pointed out a decreasing trend for VOC, indicative of microbial process activation. By the end of first year of operation an estimated 2,700 ton of hydrocarbon have been biodegraded in CO 2 and H 2 O, with an oxygen consumption by biomass of about 7,700 ton of oxygen. References [1] Atlas R.M., Microbial degradation of petroleum hydrocarbons: An environmental Perspective, Microbiological reviews, 45, pp , [2] Robertiello A., Bioremediation of hydrocarbons contaminated sites (in Italian), Hoepli Technical Library, pp , [3] Environment Ministry s Decree no. 471/99, Regulation outlining criteria, procedures and modes for security measures, remediation and environmental recovery in contaminated sites, [4] Lombardia Regional Decree no. 6/96, Soil quality standards for land remediation in Lombardia region, [5] USACE, Design consideration for landfarms, ETL ,1996. [6] Dupont, R.R.. Fundamentals of bioventing applied to fuel contaminated sites. Environ. Progress, 12, pp [7] Rathfelder K.M., Lang J.R., and Abriola L.M.. A numerical model for the simulation of coupled physical, chemical and biological processes in soil vapor extraction and bioventing systems. J. Cont. Hydrol., 43, pp [8] Johnson P.C. A practical approach to the design, operation and monitoring of in situ soil venting systems. Groundw. Monit. Rev., 10 (2), pp , [9] Hinchee R.E., and Ong S.K.. A rapid in situ respiration test for measuring aerobic biodegradation rate of hydrocarbon in soil. J. Air Waste Mgmt. Assoc, 42, pp , [10] Miller R.N., Downey D.C., Hinchee R.E. and Leeson A. A summary of bioventing performance at multiple Air Force sites. Proceedings of Petroleum Hydrocarbons and Organic chemicals in ground water, Houston, TX, pp , [11] United States Environmental Protection Agency (USEPA), Manual - Bioventing Principles and practise, Vol. II Bioventing Principles EPA/625/r95/534a, [12] United States Environmental Protection Agency (USEPA), Manual - Bioventing Principles and practise, Vol. I Bioventing Design EPA/625/xxx/1, 1995.