SITE INFORMATION AND CORRESPONDENCE_1985-1997

Home Browse Search

within a row; rows were alsostalled 6 inches apart. There was a 3-inch stagger in the pencil locations between the rows. Figure 3 shows a cross section of the PL-1 PL-2 PL-3 PL-A site along the Pencil Line. 0.0- 0.5- Sample .00.5Sample Collection and Analysis �! Samples were collected using a sampling manifold specifically designed for this study. Sample bottles were 1.0 placed in line,between the sampling point and a peristal- tic pump, so that air did not contact the sample at any 1.5 NJ _ time. Aflow-through cell was placed immediately PENCILS before the pump. This cell contained pH, temperature, and conductivity meters.The manifold was built mostly 2.0 of /s-inch O.D. stainless steel, but a few short lengths 0.0 0.5 1.0 1.5 2.01m1 of Viton and Tygon tubing were used to facilitate attach- ing it to the sampling points. A volume equal to three Figure 3. Cross section of the site along the Pencil Line. Blank times the volume of the sampling point was purged from pencils represent offset described in the text each point before any samples were collected. Geo- chemical parameters such as D.O., pH, temperature, and conductivity were analyzed in the field. D.O. was ease data interpretation (i.e., variations in solute con- measured using an iodometric method (APHA et al. centrations resulting from injection pulsing can be dif- 1985) modified by diluting the reagent to accommodate ferentiated from the concentration decreases resulting a 100-mL sample. The estimated detection limit was from transformation processes). In addition to increas- 0.1 mg/L. ing the injection frequency by a factor of two, the B/T, Benzene and toluene in the samples were concen- bromide, and nitrate concentrations were reduced by a trated by liquid-liquid extraction using pentane as the factor of two. Thus the average rate of mass injection organic phase. The pentane contained 25 mg/L of remained the same. The daily solute injections contin- m-fluorotoluene as an internal standard. Sample ued for approximately 12 days, ending on Day 37. extracts were analyzed for benzene and toluene using Although the injection frequency was changed, as a Shimadzu GC-9A equipped with a flame-ionization described previously, the method of B/T injection was detector. The method detection limit was estimated to consistent throughout the experiments. Benzene con- be 1.0 µg/L. None of the laboratory blanks analyzed centrations were 3947 . 284 µg/L and toluene 3810 contained benzene or toluene above the detection limit. ± 264 µg/L (standard deviations of about 7 percent) in The average error in the analyses, determined from the source wells on Day 15.As discussed above,injection laboratory spike samples, was less than 10 percent for concentrations for all solutes were cut in half on Day benzene and toluene. 25.Although source well concentrations were not subse- Bromide samples were analyzed in the laboratory quently monitored,it was assumed that the injected BIT using a Corning bromide ion specific electrode and a concentrations for each individual injection round were Markson pH/mV meter. At least five standards were relatively constant throughout the experiment, i.e., the used to generate the calibration curves. Although no same mass of B/T/NO3/Br was introduced into each of formal method was used to estimate the detection limit the 16 source wells. of the method as applied in this work, the analyses suggested that it was below 0.5 mg/L. Ground Water Velocities Ground water velocities, estimated from simple 2-D modeling using FLOWPATH(Waterloo Hydrogeo- Results and Discussion logic Software, Waterloo, Ontario), range from 20 to 30 cm/day in the experimental zone. Due to the semi- Solute Injection continuous injection technique, bromide breakthrough Solute injection started two days after the ORC curves (plots of bromide concentration vs. time) were sources were installed.For convenience,this is referred difficult to interpret by anything other than the most to hereafter as Day 0. Initially, injections were con- simple method;i.e., the ground water velocity was esti- ducted every other day to dose the aquifer with benzene, mated as the distance from the injection well to the toluene, and nitrate. This last compound was included monitoring point divided by the time of arrival of the to promote the growth of microorganisms. Bromide bromide pulse. Figure 4 shows a typical breakthrough injection started on Day 13. Hereafter, benzene and curve for bromide.The bromide data obtained from the toluene are referred to as B/I: various sampling locations indicated that average linear After Day 25,the injection frequency was increased ground water velocities varied from 25 cm/day to to once per day.More frequent injections result in more 38 cm/day, in reasonable agreement with the velocities constant solute concentrations across the site, which calculated by the model. WINTER 1994 GWMR ■ 123