Evaluation of the suitabil[i]ty of Fischer-Tropsch gas-to-liquid (GTL) Primary Column Bottoms as process cooling water : analysis of microbial community dynamics, fouling, scaling and corrosion

Thesis (M. Environmental Science (Water Science))--North-West University, Potchefstroom Campus, 2007

Water in South Africa is becoming limiting due to economic growth, social development and the country's water demand that exceed its water availability. Water conservation in the industry can be accomplished by the reuse of process water instead of direct treatment and discharge. By reusing a process effluent as cooling water in cooling towers, the water requirements of an industry, such as Sasol, will be lower and a zero effluent discharge scenario could be achieved. At Sasol, during the gas-to-liquid (GTL) conversion process, natural gas is converted to diesel and other products. During this process an aqueous effluent stream is produced in the Fischer-Tropsch (F-T) reactors known as Primary Column Bottoms. Primary Column Bottoms can be re-used as cooling water within cooling towers. Although this approach is technically feasible, the re-use of process water in cooling systems is characterised by major problems (fouling, scaling and corrosion) due to the complicated chemistry of the process water and the increased nutrient loads within the system. The aim of this study was to evaluate the suitability of Fischer-Tropsch gas-to-liquid Primary Column Bottoms as process cooling water by analysing the microbial community dynamics, fouling, scaling and corrosion. Due to the corrosive nature of this process effluent, stabilisation of the water was essential. To determine whether efficient stabilisation was attained, an accelerated corrosion test was performed. Influence of the external operating parameters within the cooling tower on the rate of fouling, scaling and corrosion were also determined. Structural and functional diversity of planktonic and sessile communities were studied by making use of conventional microbiological techniques (plate counts, MPN technique) and molecular methods (PLFA, DGGE). The accelerated corrosion test of 28 days conducted on mild steel and stainless steel (316L) corrosion coupons accelerated corrosion by immediately establishing the mature natural environment that causes corrosion. The test solution was stabilised as well as non-stabilised synthetic Primary Column Bottoms, in order to compare the effect of stabilisation. Scaling and corrosion indices were also calculated on stabilised and non-stabilised water to determine the scaling and corrosive tendencies of the water and how this correlates with the actual corrosion results obtained. According to the Langelier Saturation Index (LSI), Rymar Stability Index (RSI) and the Puckorius Scaling Index (PSI) the stabilised water was slightly scale forming with little corrosion and the non-stabilised water being more corrosive than scale forming. Average corrosion rate of the stabilised water was 0.032 m d y and 0.049 m d y for non-stabilised water. Average scaling rate was calculated as 7.269 mg/dm2/d for stabilised water and 5.853 mg/dm2/d for non-stabilised water. It can therefore be concluded that effective stabilisation was achieved since stabilised water was less corrosive than non-stabilised water which was also confirmed through experimental data (corrosion rates from accelerated corrosion test) and corresponded with the corrosive tendencies obtained from the scaling and corrosion indices. A lab-scale cooling tower was operated with stabilised synthetic Primary Column Bottoms as cooling water. Five experiments were conducted under varying flow rates and cycles of concentration. Influence of the external operating parameters (linear flow velocity and cycles of concentration) on fouling, scaling and corrosion rates of mild steel and stainless steel (316L) corrosion coupons and heat exchanger tubes were determined through weight loss measurements. Routine physico-chemical analyses, EDS (energy dispersive spectrometry) microanalysis as well as scaling and corrosion indices of each experiment were also compared, in order to evaluate the influence of cycles of concentration and linear flow velocity. Based on the results obtained, it was evident that the variation in cycles of concentration and linear flow velocity had a significant effect (p0.05) on the fouling, scaling and corrosion rates on the mild steel corrosion coupons and heat exchanger tubes. Experimental runs operated at low flow rates of 0.6mIs and 0.9 d s resulted in relative high fouling, scaling and corrosion rates. Operation at 3 and 4 cycles of concentration had the highest scaling and corrosion rates. The COD within the cooling tower was not removed by the microorganisms within the planktonic and sessile communities and resulted in a build-up of COD in the sump. Thus, the cooling tower cannot be used as a bioreactor to biologically degrade volatile organic acids and hydrocarbons. To evaluate the structural and functional diversity of the bacterial and fungal communities. plate counts, most probable number technique, phospholipid fatty acid (PLFA) analysis as well as denaturing gradient gel electrophoresis (DGGE) was used. According to PLFA profiles the community structure within the planktonic and biofilm samples of the experiments operated at low linear flow velocities were similar. The same percentages of Gram-positive, Gram-negative bacteria and fungi occurred, The community structure composition of the planktonic and sessile phases in the experiments operated at higher linear flow velocities was also similar. PLFA analysis concluded that the highest estimated viable biomass was in experiment 1 which had a low linear flow velocity of 0.6 d s . Shannon-Weaver index analysis of DGGE profiles (general structural diversity) indicated that the planktonic bacterial diversity of experiment I and 2 were the highest. Experiment I and 2 were operated at a linear flow velocity of 0.6 and 0.9 m/s respectively. The biofilm samples that had the highest Shannon-Weaver diversity index were experiment 1 and 5. Both experiment 1 and 5 were operated at a linear flow velocity of 0.6 m/s. Morphological changes between planktonic and sessile communities were monitored through scanning electron microscopy (SEM). SEM results illustrated that the planktonic and sessile microbial populations throughout the five experiments were similar, based on morphology. According to the results obtained from the MPN technique, the experiment operated at the lowest linear flow velocity had the highest numbers of sulphate reducing bacteria and also resulted in the highest corrosion rate. Both experiments that were operated at a low linear flow velocity of 0.6 d s had the highest bacterial numbers and also resulted in high fouling rates. However, no relationship exists between the percentage increase in the numbers of aerobic bacteria and the cycles of concentration at which the cooling tower was operated. These observations are supported by results from PLFA profiles that showed that the community structure within the planktonic and sessile samples of the experiments operated at low linear flow velocities were similar. The planktonic and sessile phases of these two experiments had similar levels of Gram-positive-, Gram-negative- bacteria and fungi. The community structure composition of the planktonic and sessile phases in the experiments operated at high linear flow velocities was also similar. PLFA analysis further demonstrated that the highest estimated viable biomass was in the experiment operated at a low linear flow velocity of 0.6 m/s. Shannon- Weaver index analysis of DGGE profiles (general structural diversity) also indicated that the planktonic bacterial diversity during operation at low linear flow velocities were the highest. Although scanning electron microscopy results illustrated that the planktonic and sessile microbial populations throughout the five experiments were generally similar. these results supported the observations of the other techniques. These techniques all supported the notion that corrosion rates may not be directly related to the total microbial biomass or the number of species on mild steel or stainless steel. Corrosion rates seem to be more profoundly affected by biofilm composition within the sessile phase. Based on the results obtained when using Primary Column Bottoms as cooling water, it were evident that variation in cycles of concentration and linear flow velocity had a significant effect (p>0.05) on the fouling, scaling and corrosion rates on mild steel corrosion coupons and heat exchanger tubes. Low linear flow velocities resulted in high fouling rates, increased bacterial numbers as well as high bacterial and fungal diversities. High cycles of concentration resulted in high scaling and corrosion rates and also had the result of similar community structure profiles. This research study could facilitate the selection of optimised operational parameters for the re-use of industrial process water (such as Primary Column Bottoms) as cooling water to minimise fouling, scaling and corrosion.

Masters