By Weinberg, Howard S; Cook, Steven J; Singer, Philip C
ABSTRACT:
Many publicly owned treatment works in North America are exceeding permitted limits for total cyanide in their wastewater treatment effluents. A recently introduced rapid, segmented, flow- injection analysis procedure using UV digestion and amperometric detection of the membrane-separated cyanide was used to investigate the various scenarios by which elevated cyanide levels might be present in wastewater treatment plant effluent. A number of significant interferences can produce false positive bias during sample analysis with the traditional acid distillation technique, but are minimized or absent with the new analytical method. However, increased levels of cyanide were found in some chlorinated wastewaters compared to the levels before chlorination, suggesting a fast reaction mechanism associated with the disinfectant and some precursor in the wastewater. In particular, the contact of chlorine with nitrite in the presence of a carbon precursor appears to contribute to cyanide formation during wastewater treatment and sample handling. This paper explores the scenarios under which cyanide can form during wastewater treatment as well as those in which a false bias for total cyanide can be obtained during sample processing and provides guidance for appropriate sample handling, screening, and processing to ensure valid analytical results. Water Environ. Res., 77, 491 (2005).
KEYWORDS: total cyanide, free cyanide, nitrite, sulfide, thiocyanate, flow injection, chlorination, wastewater.
doi: 10.2175/106143005X67403
Background
The objectives of the research described in this paper are to understand the reasons for the occurrence of false positive levels of total cyanide and to prevent cyanide formation during chlorination of wastewater.
The major point sources of cyanide discharges to water are from wastewater treatment plants (WWTPs), iron and steel production, and organic chemical industries (Fiskel et al., 1981). Cyanide salts have such varied and diverse applications today that they are appearing in the influent waters of more WWTPs than in the past. Those industries that produce cyanide wastes use a variety of methods in their pretreatment processes to remove them before they are discharged from the plant. Among the more traditional processes are alkaline chlorination, which converts cyanide to the less toxic cyanate; electrolysis, which converts it to carbon dioxide; or ozonation. Thiocyanate may also reach WWTPs from coal processing, gold and silver extraction, and mining wastes. Complex cyanides may form when free cyanide comprising hydrogen cyanide (HCN) and cyanide ion (CN^sup -^) comes into contact with heavy metals from electroplating wastes. However, in spite of vigorous efforts by industrial pretreatment to limit the levels of cyanide reaching WWTPs, there are records dating back to the early 1970s that indicate detectable levels of cyanide in both primary and secondary effluents (Young, 1978). Because discharges from WWTPs may have adverse ecological effects on receiving waters, water quality criteria have been developed for cyanide with the aim of determining an upper limit on the levels of permissible cyanide discharge to these streams (U.S. EPA, 1984). The toxicity of cyanides is highly dependent on the cyanide species. Free cyanides, particularly HCN, are considered the most bioavailable and most toxic forms (Wild et al., 1994), while complexed cyanides known to be stable in natural waters, are significantly less toxic (Blaha, 1976).
The National Ambient Water Quality Criteria for cyanides are based on the results from bioassay tests using free cyanides. However, the U.S. Environmental Protection Agency (U.S. EPA) currently recommends applying the criteria based on the total recoverable cyanide measurement (U.S. EPA, 1984). This approach is overly conservative due to the lower toxicity of complexed cyanide forms and has resulted in a number of dischargers having stringent cyanide limits, even though the cyanide in many of these discharges may be in a nontoxic form. An increasing number of WWTPs are reporting difficulties in complying with these levels and some are facing legal action by public challengers because they are unable to control these permit violations. Part of this problem is the unrealistic burden placed on utilities and their contract analytical laboratories to determine total cyanide (CN^sub tot^) levels at or below the practical quantitation limit set by U.S. EPA (10 g/L) for the currently approved analytical methodology (U.S. EPA, 1983). Oxidizing agents can transform cyanide during storage and handling, necessitating the addition of a quenching agent for any residual disinfectant. However, there are numerous reports of interferences from the various quenching species used (Carr et al., 1997; Delaney et al., 1997; Water Environment Laboratory, 1994). Other species such as sulfide, certain oxidizing agents, nitrate or nitrite, thiocyanate, aldehydes, and ketones may interfere under the acid distillation conditions that are required for the accepted testing procedure, thus producing erroneous results. In addition to analytical problems, it is quite possible that thiocyanate, which is not included in the CN^sub tot^ measurement, undergoes incomplete oxidation in chlorinated wastewaters generating some free cyanide.
Therefore, the following issues related to cyanide compliance face WWTP operators:
* The existing analytical methodology “may not accurately reflect actual cyanide concentrations found in wastewaters” (Guidelines, 1995).
* The water quality criteria developed for cyanide may be suspect if they were developed using inadequate analytical methodologies.
Figure 1-Locations of disinfectant addition relative to the sample points in the WWTPs studied.
* Chlorination and/or dechlorination may generate or appear to generate cyanide during wastewater treatment.
This paper demonstrates inadequacies with analytical methodologies for cyanide through an investigation of practices across the industry and provides insight to sources of major bias during sample processing and analysis. This is emphasized through use of an alternative analytical method that is more reliable and amenable to rapid sample throughput and generation of quality- assured and controlled data. Furthermore, we have investigated scenarios where cyanide may be formed during wastewater treatment and subsequent sample handling.
Materials and Methods
Sample Collection. Glass vials measuring 40 mL that were used for cyanide and nitrite analysis were rinsed with tap water, soaked in a 10% nitric acid (certified American Chemical Society plus grade; Fisher Scientific, Pittsburgh, Pennsylvania) bath overnight, rinsed three times with deionized water (Dracor, Inc., Durham, North Carolina), and dried in an oven at 100 C. All other laboratory glassware used for analysis and sample collection, with the exception of the volumetric flasks that were air-dried at room temperature, underwent the same cleaning procedure.
Wastewater samples were collected by a grab-sampling technique in 1-L, high-density polypropylene or linear polyethylene containers (Fisher Scientific) containing sodium hydroxide and other pretreatment reagents as described in this paper. The samples were preserved in the pH range of 12 to 12.5 and held at 4 C in the dark until analysis within 48 hours of collection. The bottles were cleaned by first being soaked in a detergent bath (Alconox, Inc., New York, New York). They were then rinsed with tap water and soaked in a 10% nitric acid bath. Finally, the bottles were triple-rinsed with deionized water and dried by inversion over clean utility wipes.
Sample Collection Procedure. To determine if the inadequacies of the analytical method were a direct function of the application to which the method was put, a series of treatment plants employing either UV irradiation or chlorination for disinfection were surveyed for cyanide levels. Samples were collected at various points in the WWTP. The first sample collection point was immediately before the point of disinfection. For chlorination plants, the second sample was collected at the midpoint of the disinfection chamber and the third sample was collected at the location from which samples are collected for monitoring compliance with the effluent cyanide permit. For UV plants, the midpoint was omitted. For some plants, the sample used for compliance monitoring was collected at the end of the disinfectant contact chamber; at other plants, the sample was collected from the discharge pipe just before the receiving stream. Some plants using chlorination are required to dechlorinate before discharge, in which case the sample collected for compliance monitoring was taken just before the dechlorinating agent was added. A schematic illustration of the sample collection points at the WWTPs is given in Figure 1. Along with samples collected for cyanide analysis at the WWTPs, controlled chlorination experiments were performed in the laboratory on selected samples of secondary effluent that were collected before chlorination. A 1-L aliquot of this sample was chlorinated at the same level used on the day of sample collection, and the sample was held at 20 Cin a water bath for the same amount of time as in the plant’s chlorine contact chamber. To investigate the effect of nitrite and thiocyanate on cyanide formation during chlorination, the laboratory procedure was duplicated on separate samples of secondary effluent that were spiked with 700 g/L of sodium nitrite and 200 g/L of sodium thiocyanate.
The stabilization procedures, carried out at the time of sample collection unless otherwise stated, were as follows: (1) residual chlorine quenching with sodium meta-arsenite; (2) residual chlorine quenching and sulfide removal using lead carbonate and subsequent filtration; (3) residual chlorine quenching and nitrite removal using sulfamic acid; (4) residual chlorine quenching onsite and nitrite removal just before analysis; and (5) residual chlorine quenching, sulfide removal, and nitrite removal at the time of sample collection. The precise sample handling processes are displayed in Figure 2. Among these procedures, only the third is routinely carried out using standard analytical procedures even though the procedures “suggest” screening for the other interferents before analysis. Moreover, there is little consistency among treatment plants in terms of the type and amount of quenching agent used. The approved methods provide options and the choice of agent is typically a function of the contract laboratory procedural protocols rather than a measured response to a utility’s concerns. As a result, instances of use of quenching agent on already dechlorinated wastewaters were documented as was the use of sulfur- containing reagents for dechlorination on wastewaters that were disinfected with UV radiation. These inappropriate practices appeared to be the default unless plant or laboratory managers made special requests to their contract laboratories. Because the analytical method is highly susceptible to various oxidized states of sulfur, the presence of these reagents can be detrimental to the accurate measurement of CN^sub tot^ in the matrix. Details of each procedure used in this study are provided here.
Figure 2-Stabilization and preservation procedures used in this study (carried out onsite unless specified otherwise).
Residual Chlorine Removal. Sodium meta-arsenite (Aldrich Chemical Company, Milwaukee, Wisconsin) was used for quenching through the addition of 80L of 11.3 g/L as AsO^sub 2^^sup -^ into an empty 40- mL vial, which was subsequently filled with the sample before the addition of sodium hydroxide (NaOH).
Chlorine and Sulfide Removal. Lead carbonate (150 mg) (Aldrich Chemical Company) was placed in the bottom of an empty 40-mL vial along with the meta-arsenite described previously. This weight was initially measured accurately into a small spatula scoop so that subsequent additions would reproduce this weight by sight. Immediately after the addition of the sample to the vial, a stopwatch was started, the vial capped, and then the vial inverted three times to mix the contents. Within 20 seconds of adding the sample, the contents were filtered with a 0.45-m nylon syringe filter attached to a 10-mL polypropylene syringe and the filtrate collected in a 20-mL vial containing NaOH. At 1 minute, the filtration process must cease otherwise the precipitated lead sulfide will catalyze the formation of thiocyanate as described previously. Whatever filtrate volume has been collected by this time is used for subsequent analysis.
Chlorine and Nitrite Removal Onsite. Sulfamic acid (80 L of 6.35g/ L) (Aldrich Chemical Company) was placed at the bottom of a clean 40- mL vial along with the meta-arsenite, and the vial was filled with the sample prior to the addition of NaOH.
Chlorine and Sulfide Removal Onsite and Nitrite Removal at the Time of Analysis. This procedure was identical to that described for chlorine and sulfide removal except that, just before analysis, 80 L of the 6.35g/L sulfamic acid solution was added with further mixing.
Chlorine, Nitrite, and Sulfide Removal Onsite. Lead carbonate, sulfamic acid, and meta-arsenite were placed at the bottom of a clean 40-mL vial. After sample collection, the contents were treated as described for the chlorine and sulfide removal procedure.
Analytical Methods. Total Cyanide by Acid Distillation, Complexation, and Colonmetry. U.S. EPA Method 335.3 (U.S. EPA, 1983) was used by the participating utilities in this study for CN^sub tot^ determination. This method generates HCN from both free and complexed cyanides in the samples by acid distillation. At basic pH, the recovered cyanide undergoes conversion to cyanogen chloride by reaction with chloramine-T. A chromogenic reagent mixture (usually pyridine-barbituric acid) is then added, which generates a colored complex with intensity proportional to the concentration of the original cyanide.
Figure 3-Variability in recovery of 5 g/L cyanide (from sodium cyanide) spiked into chlorinated effluent.
In most laboratories, much of the procedure is still undertaken in a manual mode, which exposes the method to a potentially high source of experimental error and also exposes the analyst to the various hazardous chemicals involved. Furthermore, because the method suffers from a number of interferences, a variety of screening procedures is required before sample analysis can proceed. Ideally, this screening should take place before the actual sampling, and the prescribed procedures for removing the interferences from the matrix would then be implemented. In practice, some of the procedures for removing interferences are included in all analyses without any prior screening and are based on assumptions that are made about the presence and absence of such interferences. Some of these reagents can generate a false response for cyanide recovery and quantitation if present in excess. Additionally, some of the qualitative screening tests use test strips that are insufficiently sensitive to determine the presence of sub-milligram/liter levels of interference. Our research indicated that these levels can severely affect the accuracy of subsequent cyanide quantitation.
In spite of all the precautions that could be taken to minimize sample handling error, there are indications that U.S. EPA Method 335.3 (U.S. EPA, 1983) would still be unable to accurately determine the levels of cyanide in a complex matrix such as wastewater.
Total Cyanide by Flow Injection, UV Digestion, Membrane Separation, and Amperometric Detection. Given the cumbersome and lengthy nature of the acid distillation procedure, an alternative method developed by Solujic et al (1999) was evaluated for sensitivity, precision, and interference-free analysis in chlorinated effluents (Weinberg and Cook, 2002). This procedure takes a 250L-aliquot from a 5-mL sample and directly measures the total cyanide converted into free cyanide via flow-injection, UV digestion, and membrane separation of HCN into an elevated pH stream passing through an amperometric detector. Reliable and reproducible detection was shown down to 2 g/L in a variety of wastewater matrices, with analysis achieved in less than 5 minutes. At levels often found in municipal wastewater, sulfide, the only demonstrated interference in this method, is removed prior to analysis through precipitation and filtration as described previously. Any remaining soluble sulfide (up to 50 mg/L) is handled during analysis by the presence of bismuth nitrate in the carrier stream, which precipitates the sulfide and prevents its transfer across the membrane. Thiocyanate, at levels above 1 mg/L, can generate CN^sup – ^ in the UV digestor, although such levels are rarely found in municipal WWTP influents. Safety advantages with this procedure include the generation of HCN in a totally enclosed system as well as an effluent at high pH that is no longer hazardous and can be collected in a sealed bottle for disposal through Occupational Safety and Health Administration procedures without operator contact. Standard solutions of free cyanide must always be prepared in pH 12 solutions and handled by trained personnel using double gloving in a hood with appropriate ventilation.
Nitrite and Sulfide. A syringe-filtered sample (0.45 m) was loaded into a 100-L sample loading loop and subsequently injected into the eluent stream (sodium bicarbonate for nitrite and a mixture of sodium acetate, sodium hydroxide, and ethylene diamine for sulfide) of a model DX300 ion chromatograph (Dionex Corporation, Sunnyvale, California). Anion resolution was achieved on a 4-mm diameter analytical column (AS 12 for nitrite and AS7 for sulfide). Detection was by suppressed conductivity for nitrite and amperometry for sulfide, with practical quantitation limits of 20 and 100 g/L, respectively.
Thiocyanate. Suppressed conductivity detection using an ASlO analytical column (Dionex Corporation) with a carbonate/bicarbonate mobile phase modified with p-cyanophenol permitted quantitation down to 20 g/L in wastewater samples. Sample preparation was identical to that described for nitrite and sulfide.
Results and Discussion
Matrix Effects. Figure 3 demonstrates some shortcomings in the distillation technique of U.S. EPA Method 335.3 (U.S. EPA, 1983), with the results of some quality assurance studies in the analysis of chlorinated effluent from a local WWTP containing 800g/L sulfide. The figure illustrates the results of recovery of triplicate spikes of sodium cyanide (5g/L CN) into water that had undergone different degrees of sample pretreatment. The expected results, if 100% recovery of the spike was achieved, are shown by the points on the unbroken line. In all cases, recovery falls significantly short (20 to 60%) of the expected values. In tests that were run using the same pretreatment techniques applied to deionized water, there was no detectable CN^sub tot^ in any of the samples. When the water was spiked with 10g/L CN^sub tot^, recoveries after each of the pretreatments \were between 95 and 105%, except in the case of sulfide removal if the filtration time was in excess of 2 minutes. These results indicate that there are significant matrix effects that are not accounted for by the published procedure.
Plant Study. Sixteen WWTPs were evaluated in this study for total cyanide levels in their plant effluents during a 12-month period. These plants are listed in Table 1, which distinguishes between those practicing UV and chlorine addition for disinfection. A more detailed characterization by chemical analysis is given in Table 2. A letter identifies a specific utility, while the numeral indicates a specific plant belonging to that utility. Those plants marked with an asterisk (*) were practicing UV disinfection. All data are presented both as seasonal averages and also as ranges of levels measured during that season to indicate extremes. The biological oxygen demand (BOD) and total suspended solids (TSS) levels are presented to assist in determining any correlations between noncompliant levels of targeted analytes and organic load and solids content at the plant. There are no clear patterns of correlation. Ammonia and total Kjeldahl nitrogen (TKN) levels are presented to illustrate potential interferences of side reactions involving residual chlorine and ammonia during cyanide analysis. Again, there does not appear to be a correlation between these parameters.
It is apparent from the range of total cyanide values that every plant practicing chlorination exceeded the permit level (either 5 or 10g/L) at least once each year. Plant G was not required to measure total cyanide in the plant effluent and only self-reported values on one occasion during the year; on that occasion, values were below detection. Neither of the two plants practicing UV disinfection reported permit violations and, furthermore, plant E, in years subsequent to that for which this data is presented, switched to UV disinfection; total cyanide levels in effluent from plant E dropped to those below detection.
Impact of Sulflde on CN^sub tot^ Measurements. Sulfide is identified as a positive bias in both analytical procedures (APHA et al., 1998; U.S. EPA, 1983) and the standard approach for its removal involves precipitation with lead carbonate and filtration. However, if the precipitated lead sulfide is not immediately removed from solution it will catalyze the formation of thiocyanate (Wilmot et al., 1996), which is not included as part of the CN^sub tot^ measurement. Tests in which chlorinated wastewater containing 5 and 10 g/L free cyanide were spiked with up to 200 g/L thiocyanate showed less than 3% impact on the CN^sub tot^ value, which was within the 5% coefficient of variation obtained for replicate analyses (n = 5) of the original sample. However, to prevent negative bias due to loss of cyanide in this way, a rapid mixing of lead carbonate and subsequent filtration of the precipitated sulfide must be carried out as described in both procedures. Figure 3 confirms that this treatment does not cause significant loss of cyanide. In some recently published findings (Seto, 2002), scenarios have been postulated for the formation of cyanide from thiocyanate through oxidative mechanisms. At least under the conditions of wastewater treatment practiced in this study, there is no evidence that chlorination triggers such a process (Cook, 1999). In laboratory chlorinations of wastewaters that were spiked with 200g/ L thiocyanate, there was no discernible difference in the CN^sub tot^ levels measured in each of the treatment scenarios compared with an unspiked sample. This does not preclude the possibility of cyanide formation from thiocyanate in other chlorinated wastewaters, and suggests the need for similar studies in cases of permit violations that are not solved by targeting the mechanisms identified in this study as sources of positive bias.
Study of the Potential Formation of Cyanide during Chlorination. Samples collected from plant C2, which uses chlorination and dechlorination according to the points identified in Figure 1, were analyzed for total cyanide by the flow-injection technique and by using the various sample-processing scenarios depicted in Figure 2. Laboratory-controlled chlorination experiments using a 5mg/L as Cl^sub 2^ dose and a contact time of 30 minutes are also shown. The nitrite and sulfide levels in the water before chlorination were 736 and 900 g/L, respectively. Figure 4 provides a graphical display of the effects of the different sample treatment scenarios on total cyanide measurement for each sample point. The first bar in each set represents the effect of adding sodium meta-arsenite to the sample to remove residual chlorine; the second bar represents the use of arsenite and lead carbonate (to remove sulfide); and the third bar represents the use of arsenite and sulfamic acid (for removal of nitrite), followed by lead carbonate treatment for the removal of sulfide.
The source wastewater of plant C2 contains between 5 and 8 g/L total cyanide after accounting for the slight positive bias observed if the sulfide is not removed. However, after chlorination, this level rises to more than 40 g/L, indicating production of cyanide if the only sample treatment involves chlorine quenching. Removal of chlorine, sulfide, and nitrite at the time of sample collection appears to eliminate this reaction, while the removal of only chlorine and sulfide does not. Therefore, it appears that cyanide was formed not by chlorination of the wastewater, rather by some reaction between quenched chlorinated wastewater and nitrite or by a base-catalyzed reaction between nitrite and some precursor material in the wastewater when the sample was held at elevated pH before analysis. This would confirm the hypothesis first suggested by Carr et al. (1997). A similar trend was observed in the laboratory- chlorinated sample. In terms of impact on reported levels of total cyanide, this utility collects and analyzes samples of dechlorinated plant effluent and reported a level of 15g/L for this sampling event employing only chlorine-quenching in their sample handling. As observed in Figure 4, this overestimates the true level of total cyanide in the plant effluent determined when sulfide and nitrite are removed as part of the sample processing.
Table 1-Disinfection practiced at surveyed plants and the annual range of total cyanide levels.
To evaluate the hypothesis regarding impact of nitrite on total cyanide measurement, samples collected from plant C1 were treated for nitrite removal both at the time of analysis (as prescribed by the U.S. EPA method) and at the time of sample collection. Additionally, a nitrite-spiked wastewater collected before chlorination in the plant was chlorinated under controlled conditions in the laboratory. The results of these tests are presented in Figure 5, with the trends through the plant grouped together in one block for each sample stabilization approach employed. Nitrite and sulfide values are presented for water before chlorination. For samples in which only residual chlorine was removed, all sample points revealed detectable (>5 g/L) levels of total cyanide. A value of 7.2 g/L in the secondary effluent before chlorination was increased to 14 g/L after chlorination at 5mg/L. The sample chlorinated in the laboratory at the same dose also produced elevated levels of total cyanide (19 g/L), while the sample spiked with nitrite produced 21 g/L. The removal of sulfide, shown in the second grouping of Figure 5, reduced the levels of measured total cyanide in the first two samples to below 5g/L, but did not substantially affect levels apparent in the plant effluent that had been dechlorinated with sulfur dioxide before sample collection. The laboratory-chlorinated samples provide some clues as to the source of “apparent” cyanide formation in the other samples. The removal of sulfide before analysis keeps the total cyanide level closer to the values in the secondary effluent before chlorination, while the addition of nitrite to the sample before chlorination generates cyanide either during chlorination or subsequent storage. This is due, perhaps, to reactions with nitrogenous organics such as amino acids and polypeptides.
Table 2-Characterization of WWTP effluents used in this study.
When nitrite is removed at the time of sample collection, all cyanide levels are below 5 g/L. However, if nitrite remains in the chlorine-quenched samples right up to the time of analysis, the levels of cyanide are comparable to those in the samples that were only quenched to remove chlorine, providing a strong correlation between the presence of nitrite and cyanide formation. This is a clear indication that the cyanide formation reaction is occurring in the sample vial while it is being held for analysis and not in the original wastewater. More proof of this is shown for the samples in which the nitrite was removed immediately before analysis and which have a total cyanide level consistent with the “chlorine removal only” samples. The difference between the total cyanide levels in these two latter samples is not statistically significant. The clearest conclusion from this particular study is that when both sulfide and nitrite are removed at the time of sample collection, the levels of cyanide in all samples are close to the instrument detection limit (2 g/L).
Figure 4-Effect of chlorine, sulfide, and nitrite removal on total cyanide levels in plant C2 wastewater.
Figure 5-Total cyanide for plant C1 using different sample stabilization procedures at different points in the wastewater treatment process (nitrite = 84 g/L; sulfide = 850 g/L).
Figure 6-Total cyanide as a function of treatment process and sample handling at plant H2 (nitrite = 2.3 mg/L and sulfide below detection levels in wastewater before chlorination).
The same laboratory chlorination as described p\reviously but spiked with 700 g/L nitrite produced mixed results. The cyanide concentration in the sample from which only chlorine was removed was only 2 g/L higher than in the laboratory-chlorinated sample without the nitrite spike. A much greater difference might be expected if the true level of nitrite in the unspiked sample was similar to that in the prechlorinated sample. If, however, the nitrite level in the chlorinated water was elevated due to denitrification processes, then the effect observed in the laboratory-chlorinated sample would not be enhanced with the nitrite spike if the precursor for total cyanide formation had already become depleted. The implications of these results are that samples collected from the plant effluent for permit compliance using the existing procedures of Method 4500- CN^sup -^ (APHA et al, 1998) and U.S. EPA Method 335.3 (U.S. EPA, 1983) will indicate elevated levels of total cyanide due to “apparent” formation either during chlorination or during subsequent holding before analysis. The results for the analysis of total cyanide in plant H2 wastewater are presented in Figure 6. All plant H2 samples that were only treated for removal of chlorine indicated a cyanide concentration of ~8 g/L. Total cyanide concentration in the laboratory-chlorinated sample was twice as high and was most likely due to a higher laboratory chlorination dose (5mg/L) than was actually used at the plant. Those treatment plant samples from which sulfide was removed in addition to chlorine showed a slightly lower cyanide concentration (~6 g/L total cyanide) compared to the samples in which only chlorine was removed.
Table 3-A summary of the impact of nitrite on cyanide levels for the WWTPs surveyed in the project.
Those plant samples quenched of chlorine and treated with sulfamic acid to remove nitrite before analysis indicate elevated levels of cyanide compared to those samples in which chlorine and sulfide were removed. All samples treated for the simultaneous removal of chlorine, sulfide, and nitrite at the point of sample collection indicate levels of cyanide below 5 g/L. Therefore, the likelihood is that nitrite present in the secondary effluent before chlorination (measured at 2.3 mg/L) acts to produce a positive cyanide bias during sample collection and processing.
Impact of Nitrite on Cyanide Measurements in Chlorinated Waters. It was observed at most of the plants evaluated that the removal of nitrite at the time of sample collection decreased the “apparent” levels of total cyanide compared to samples in which nitrite was removed immediately prior to analysis and samples which were only quenched of residual chlorine. The comparative effect of three different treatments is presented in Table 3, which summarizes the total cyanide concentrations for samples that were collected in the chlorine contact chamber. Values shown are the average of three replicate analyses. Nitrite has the potential to generate nitrous acid during sample digestion and this is the basis for the prescribed handling of this interferent in the existing Method 4500- CN^sup -^ (APHA et al., 1998) and U.S. EPA Method 335.3 (U.S. EPA, 1983). However, this approach does not take into account potential cyanide-forming reactions between nitrite and carboncontaining components of the wastewater during prolonged holding before analysis as well as the elevated pH values at which the samples are held. The latter effect is demonstrated in Table 3 by the difference in measured cyanide levels as a result of the two nitritehandling procedures. Generally, the results of Table 3 also illustrate the potential positive bias that can be created by not removing nitrite from the solution before analysis. This is demonstrated by a comparison of the cyanide levels in the samples from which only chlorine or nitrite was removed.
The general observation is that the higher nitrite concentrations contributed to significant cyanide formation during processing of the sample taken from the mid-chlorine contact chamber locations. Table 4 summarizes total cyanide data for each of the three sampling points at all plants, which are listed according to the sample handling technique employed. If experiments were performed on water collected from the same plant on more than one occasion this is designated by “a” and “b”. There are specific trends across each row as the sample handling intensifies. Column 4 shows the results of analysis using the chemical addition techniques prescribed in Standard Methods (APHA et al., 1998) and U.S. EPA methods (U.S. EPA, 1983). Column 6 shows the results obtained in this study when nitrite was suppressed at the time of sample collection. In each data set, the first value represents the analysis of total cyanide in prechlorinated water, the second value chlorinated water, and the third value the wastewater discharged into a receiving stream. For some of the plants, it is apparent that cyanide was indeed present in the source waters because no aspect of sample processing changed the values analyzed (that is, in plants A1b, A2b, C2a, and H1b). From the observations for the other plants, it appears that the cyanide formation reaction that takes place in the treatment plant and subsequent sample processing is a two-step process. The first step of the reaction occurs in the chlorine contact chamber when an unknown, reactive, carbon-containing compound reacts with chlorine to form a cyanide precursor compound. The second step occurs when this newly created molecule reacts with nitrite in the sample vial to form cyanide. Because the reported elevated levels of total cyanide in chlorinated wastewaters often correlate with continued contact with nitrite, it is recommended that sulfamic acid be present in the bottles used for sample collection. This requires a change in the current practice of sample preservation. Without implementing this change, cyanide may form during sample processing of the collected chlorinated wastewater effluents. This may explain the elevated levels of cyanide in many of the wastewaters analyzed in this study and would certainly explain why this was not a problem for those samples which had undergone UV rather than chlorine disinfection.
Table 4-Relative effects of sample handling on measured total cyanide in various wastewaters*.
References
American Public Health Association; American Water Works Association; Water Environment Federation (1998) Standard Methods for the Examination of Water and Wastewater, 20th ed.; Washington, D.C. Ascorbic Acid Causes Cyanide False Positives (1994) Water Environment Laboratory Solutions; Jan/Feb; Water Environment Federation: Alexandria, Virginia.
Blalia, J. (1976) Mathematical Analysis of the Chemical System ‘Cyanide-Heavy Metals’ in Water-Determination of Components and Toxicity of the System-I. The Theoretical Solution. Water Res., 10, 815.
Carr, S. A.; Baird, R. B.; Lin, B. T. (1997) Wastewater Derived Interferences in Cyanide Analysis. Water Res., 31 (7), 1543.
Cook, S. J. (1999) Application of Flow Injection Analysis to Assess the Impact of Chlorination on Total Cyanide Concentrations in Municipal Wastewaters. M.S.E.E. Report, Department of Environmental Sciences and Engineering, University of North Carolina, Chapel Hill, North Carolina.
Delaney, M. F.; Zilitinkevitch, L.; McSweeney, N.; Epelman, P. (1997) Cyanide Formation from Chlorinated POTW Effluent. Presented at the Water Environment Federation Environmental Laboratories Conference, Philadelphia, Pennsylvania; pp 6-17, 6-26.
Fiskel, J.; Cooper, C.; Eschenroeder, A. (1981) Exposure and Risk Assessment for Cyanide; EPA-440/4-85-008; U.S. Environmental Protection Agency: Washington, D.C.
Guidelines Establishing Test Procedures for the Analysis of Cyanide under the Clean Water Act (1995) Fed. Regist., 60 (228), 60650.
Seto, Y. (2002) False Cyanide Detection. Anal. Chem., 74, 135A.
Solujic, L.; Milosavljevic, E. B.; Straka, M. R. (1999) Total Cyanide Determination by a Segmented Flow Injection-on-Line UV DigestionAmperometric Method. Analyst (London) G.B., 124, 1255.
U.S. Environmental Protection Agency (1983) Methods for Chemical Analysis of Water and Wastes. Environmental Monitoring and Support Laboratory: Cincinnati, Ohio.
U.S. Environmental Protection Agency (1984) Ambient Water Quality Criteria for Cyanide; EPA-440/5-84-028; Duluth, Minnesota.
Weinberg, H. S.; Cook, S. J. (2002) Segmented Flow Injection, UV Digestion, and Amperometric Detection for the Determination of Total Cyanide in Wastewater Treatment Plant Effluents. Anal. Chem., 74 (23), 6055.
Wild, S. R., Rudd, T, Neller, A. (1994) Fate and Effects of Cyanide During Wastewater Treatment Processes. Sd. Total Environ., 156 (2), 93.
Wilmot, J.C.; Solujic, L.; Milosavljevic, E.B.; Hendrix, J.L.; Rader, W.S. (1996) FormationofThiocyanateDuringRemovalofSiilfideasLeadSulfidePriorto Cyanide Determination. Analyst (London) G. B. ,121, 799.
Young, D. R. (1978) Priority Pollutants in Municipal Wastewaters; Annual Report; Southern California Coastal Water Resource Project: Westminster, California; pp 103-112.
Acknowledgments
Credits. The financial support of the North Carolina Urban Water Consortium through the University of North Carolina Water Resources Research Institute is gratefully acknowledged (WRRI Project Number 50210), as is the participation of the various utilities in the collection of samples and provision of plant operation parameters.
Authors. Howard Weinberg is an assistant professor in the Department of Environmental Sciences and Engineering at the University of North Carolina, Chapel Hill. Steven Cook is an environmental engineer with Malcolm Pirnie, Inc., Newport News, Virginia. Philip Singer is the Daniel A. Okun Distinguished Professor of Environmental Engineering in the Department of Environmental Sciences and Engineering at the University of North Carolina, Chapel Hill. Correspondence should be addresse\d to Howard Weinberg, Department of Environmental Sciences and Engineering, University of North Carolina, Rosenau Hall, Chapel Hill, NC 27599- 7431; e-mail: [email protected].
Submitted for publication April 14, 2003; revised manuscript submitted May 10, 2004; accepted for publication July 12, 2004.
The deadline to submit Discussions of this paper is January 15, 2006.
Copyright Water Environment Federation Sep/Oct 2005
Comments