Environ. Sci. Technol. 2003, 37, 107-115
Evaluation and Testing of Analytical Methods for Cyanide Species in Municipal and Industrial Contaminated Waters ANPING ZHENG AND DAVID A. DZOMBAK* Department of Civil and Environmental Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213
which experienced significant interference problems and/ or low recoveries. There was recovery of significant diffusible cyanide in microdiffusion tests with nickel-cyanidespiked samples, reflecting dissociation of this weak metalcyanide complex during the test and demonstrating that the test can recover some fraction of WAD cyanide in addition to free cyanide. The automated total cyanide methods, which involve UV digestion, achieved low detection limits for most waters but exhibited low recoveries for some waters.
RICHARD G. LUTHY
Introduction
Department of Civil and Environmental Engineering, Stanford University, Stanford, California 94305
Numerous cyanide species can occur in water, but identification and quantification of these species is not commonly practiced. Rather, the bulk measurement “total cyanide by distillation”, which does not differentiate different forms of cyanide (and excludes thiocyanate, cyanate, organocyanides, and some other forms), is usually employed to assess cyanide content of water. Since different forms of cyanide have different toxicity characteristics and physical-chemical properties, total cyanide analytical data have limitations for use in risk assessment and in evaluation of cyanide fate and transport in the aquatic environment and treatment processes. Analytical methods that focus on cyanide species and groups of species have been developed but are used relatively little in part because of concern about ability to achieve low (ppb level) detection limits in the complex matrices relevant to water quality management. The aim of this study was to investigate this concern through evaluation of the comparative performance of five species-specific cyanide analysis methods. Two automated techniques for total cyanide analysis were also evaluated. Generally, cyanide compounds can be classified as simple and complex cyanides. Simple cyanides are represented by the formula A(CN)x where A is an alkaline earth element or a metal and x is the number of cyanide groups, such as NaCN, KCN, etc. In solutions of simple metal cyanides, the CN group may occur also in the form of metal-cyanide complexes, e.g., Cu(CN)3-, Ni(CN)42-, Zn(CN)42-. The metal cyanides may dissociate depending on several factors. Some metal-cyanide complexes such as zinc-, nickel-, copper- and cadmiumcyanide are easily dissociable under acidic conditions and are classified as weak acid dissociable (WAD) species. Other metal cyanides are strongly complexed and difficult to dissociate, such as cobalt- and iron-cyanide complexes. The different types of cyanide species, free cyanide, WAD metal-cyanide, and strongly complexed metal-cyanide, as well as other cyanide-related compounds such as thiocyanate, cyanogen chloride and cyanate, exhibit great differences in their toxicity, reactivity, and environmental fate and transport (1-3). From an environmental perspective, the most toxicologically significant or ecologically important forms of cyanide are free- and WAD cyanide. Operationally, the term “cyanide” refers to all CN (-CtN) groups that can be determined analytically as the cyanide ion, CN-, via colorimetric, amperometric, or electrochemical measurement, usually after some pretreatment to liberate the CN- ion (4). The common total cyanide by distillation measurement (4), which is performed manually, involves first pretreating a water sample by digestion in strong acid under heat, which breaks down most (but not all) cyanide-bearing compounds, followed by distillation to release and capture the free cyanide, and a final determination of CN- via colorimetric, titrimetric or cyanide-selective
BERNARD SAWYER, WILLIAM LAZOUSKAS, AND PRAKASAM TATA Metropolitan Water Reclamation District of Greater Chicago, Cicero, Illinois 60804 MICHAEL F. DELANEY AND LARISSA ZILITINKEVITCH Massachusetts Water Resources Authority, Boston, Massachusetts 02129 JOHN R. SEBROSKI AND REBECCA S. SWARTLING Bayer Corporation, New Martinsville, West Virginia 26155 SHARON M. DROP Alcoa Inc., Alcoa Center, Pennsylvania 15069 JOHN M. FLAHERTY Exygen Research, Inc., State College, Pennsylvania 16801
Total cyanide analysis by distillation is used most commonly to assess cyanide content of water samples. This manual method is robust but slow and provides no information about cyanide speciation, a significant limitation in that cyanide species have substantially different toxicity characteristics. Seven alternative methods for the analysis of cyanide species or groups of species were evaluated in reagent water and five different contaminated water matrices, including five species-specific methods - weak acid dissociable (WAD) cyanide, free cyanide by microdiffusion, available cyanide, automated WAD cyanide by thin film distillation, metal cyanides by ion chromatography - and two automated techniques for total cyanide total cyanide by thin film distillation and total cyanide by lowpower UV digestion. The species-specific cyanide analytical techniques achieved low, ppb-level detection limits and exhibited satisfactory accuracy and precision for most contaminated waters. Analysis of low concentrations of cyanide species in raw wastewater was problematical for the available cyanide and ion chromatography methods, * Corresponding author phone: (412)268-2946; fax: (412)268-7813; e-mail:
[email protected]. 10.1021/es0258273 CCC: $25.00 Published on Web 11/21/2002
2003 American Chemical Society
VOL. 37, NO. 1, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
107
TABLE 1. Cyanide Analytical Methods Evaluated in the Study methodology
method code CN species measured reference
available cyanide, EPA Method OIA-1677 free cyanide by micro diffusion, ASTM D 4282-95 metal-cyanides by ion chromatography total cyanide by low power UV digestion total cyanide by UV digestion and thin film distillation, ASTM D4374-93 weak acid dissociable cyanide, SM 4500-CN-I WAD cyanide by thin film distillation, SM 4500-CN-I, automated c
Avail-CN FCN-MD MCN-IC TCN-LPUD TCN-TFD WAD-CN WAD-TFD
WAD CN free CN metal-cyanides total CN total CN WAD CN WAD CN
(5) (10) (11) (13) (12) (4) (4)
laboratory Bayer Corp. Carnegie Mellon Alcoaa MWRAb MWRDGCc Carnegie Mellon MWRDGCc
a Centre Analytical Laboratories, Inc. (State College, PA) performed the analysis for Alcoa. b Massachusetts Water Resources Authority. Metropolitan Water Reclamation District of Greater Chicago.
electrode methods. Various methods have been developed to measure different cyanide species, although most of the methods assess groups of species rather than single species. The U.S. Environmental Protection Agency (USEPA) has acknowledged that cyanide is a method-defined analyte, which means that cyanide measured is defined by the method used (5). The goal of this study was to assess the method detection limits, accuracy, and precision of five cyanide-species-specific and two automated-total cyanide analytical methods with a broad spectrum of contaminated water matrices such as publicly owned treatment works (POTW) influents, POTW effluents, and contaminated groundwaters from a former manufactured gas plant (MGP) site and from a waste disposal site of an aluminum smelting plant. Emphasis in the study was on performance of the methods at low ( 12. Available Cyanide by Flow Injection, Ligand Exchange, and Amperometry, USEPA Method OIA-1677 (Avail-CN). To liberate the cyanide ion from metal-cyanide complexes, ligand exchange reagents are added to the sample prior to analysis. Alternate ligand exchange reagents (10.0 µg/mL tetraethylenepentamine and 5.00 µg/mL diphenylthiocarbazone) were used in lieu of the proprietary reagents prescribed in Method OIA-1677, as they have been shown to give acceptable performance for this procedure (16). The analysis of available cyanide in the samples treated with ligand exchange reagents was performed with a CNSolution Model 3000 Cyanide Analyzer (OI Analytical, College Station, TX). The treated samples are placed on an autosampler, and then 200 µL of each sample is injected into the flow injection manifold system with 0.1 M HCl as the carrier/acid reagent. Hydrogen cyanide is separated from the acidified sample through a 0.1 µm polypropylene gas diffusion membrane and captured in an acceptor stream of 0.1M sodium hydroxide. The cyanide ion is measured amperometrically. This method received approval from USEPA on December 30, 1999 (5) and is currently under evaluation by ASTM International’s Committee D19 on Water. Free Cyanide by Microdiffusion Method, ASTM D 428295 (FCN-MD). Here free cyanide refers to those simple cyanides (HCN, CN-) and/or readily dissociable metalcyanide complexes that yield, at pH 6 and room temperature, HCN which diffuses from the sample, through the enclosed vapor phase, to an NaOH trap. This manual method uses a cylindrical microdiffusion cell with two annular compartments (Conway Diffusion Cell, Bel-Art Products, Inc., Pequannock, NJ). A 3.0 mL sample is placed in the outer ring of the cell, followed by addition of a pH 6 buffer and a dose of dissolved cadmium to precipitate any hexacyanoferrate present (some weak cadmium cyanide complexes may also form). Cadmium can also help remove sulfide, though it is not added specifically for this purpose. The cell is then capped and allowed to stand for 4 h. HCN volatilized from the sample diffuses through enclosed headspace into 1.3 mL of NaOH solution in the center chamber of the microdiffusion cell. Cyanide concentration in the NaOH solution is determined by a colorimetric procedure in which cyanides are converted, by reaction with chloramine-T, to cyanogen chloride which subsequently is reacted with pyridine and barbituric acid to give a red-colored complex. The intensity of red color, which is correlated with cyanide concentration, is evaluated by measuring light absorbance at 578 nm with a Spectronic-21 spectrophotometer (Milton Roy Company, Ivyland, PA). Metal-Cyanides by Ion Chromatography Method, Dionex Corp. (MCN-IC). Since cyanide can form multi-ligand complexes with many metal ions, virtually all complexed-
TABLE 2. Contaminated Water Matrices Evaluated in the Study no.
site/plant name
CW1 manufactured gas plant (MGP) site groundwater, NY CW2 Hanover Park Water Reclamation Plant, Metropolitan Water Reclamation District, Hanover Park, IL CW3 Hanover Park Water Reclamation Plant, Metropolitan Water Reclamation District, Hanover Park, IL CW4 Deer Island Treatment Plant of the Massachusetts Water Resources Authority (MWRA), Boston, MA CW5 aluminum smelting plant, TN a
site/plant type
sample type
sampling date
former MGP site POTW
contaminated groundwater 3/15/99 unchlorinated POTW secondary effluent 4/19/99
POTW
chlorinated POTW secondary effluenta
5/17/99
POTW
POTW primary clarifier effluent
6/14/99
waste disposal site contaminated groundwater
7/14/99
The CW3 sample was dechlorinated after removal of samples for the characterization analyses.
cyanides of interest are anionic and can be separated by anion exchange chromatography and quantified by a suitable detection method, usually by UV absorption (though not all metal cyanide complexes absorb strongly in the UV range of the spectrum). Dionex Corp. (11) developed a method capable of measuring the transition metal cyanide complexes of iron, nickel, copper, cobalt, platinum, palladium, silver and gold (Fe(CN)64-, Ni(CN)42-, Cu(CN)32-, Co(CN)63-, Pt(CN)42-, Pd(CN)42-, Ag(CN)22-, Au(CN)22-, respectively). The anionic cyanometalates are separated on Dionex IonPac anion exchange columns using a sodium perchlorate/sodium hydroxide gradient elution and are then detected by UV absorbance at 215 nm. Under the alkaline conditions of the analysis, ferricyanide (Fe(CN)63-) is reduced to ferrocyanide (Fe(CN)64-) yielding a single analyte peak. For this work, the analysis was conducted using a Dionex DX-500 Ion Chromatography System (Dionex Corp., Sunnyvale, CA) consisting of an IP20 isocratic pump, AD20 UV absorbance detector, GP40 gradient pump, and an LC20 chromatography enclosure. Dionex AG11 guard and AS11 separator columns were employed for separation of metal cyanide complexes. Samples were initially preconcentrated onto a second AG11 column in order to enhance the method for detection in the low ppb cyanide concentration level range. 20 mL sample aliquots were used for preconcentration. (The typical injection size for ion chromatography is 50 mL). The mobile phase consisted of three separate eluents: Eluent 1-150 mM NaCN and 20 mM NaOH; Eluent 2-300 mM NaClO4 and 20 mM NaOH; and Eluent 3-20 mM NaOH. The percentage proportion of Eluent 2 was gradually increased during the analysis run in order to elute the more highly charged metal cyanide complexes. Automated Total Cyanide by Low Power UV Digestion Method (TCN-LPUD). This is an automated analysis method implemented using a Skalar SAN plus segmented flow analyzer (model SA2001) with an SA1050 random access sampler, an SA 5570 in-line distillation unit and an SA 555 UV-B inline digester (Skalar Analytical B.V., The Netherlands). The cyanide as hydrocyanic acid (HCN) is released from cyanide complexes by means of UV digestion and distillation. Cyanides are converted to cyanogen chloride by reaction with chloramine-T. Cyanogen chloride subsequently reacts with pyridine and barbituric acid to give a red colored complex. This method is technically equivalent to the USEPA Method 335.3 (13) which was withdrawn in 1994 for drinking water analysis (14) but still remains approved for reporting cyanide concentrations as required by NPDES permits. Automated Total Cyanide by UV Digestion and Thin Film Distillation Method, ASTM D 4374-93 (TCN-TFD). This method utilizes alkaline UV irradiation, acidification and thin film distillation for cyanide-containing samples in an automated system. The breakdown of the strong metal cyanide complexes, prior to the thin film distillation over a heated plate, is achieved by UV irradiation. Absorption of the liberated HCN gas is carried out using a glass coil and NaOH solution. It also employs the standard colorimetric deter-
mination of the recovered cyanides. The analysis was performed in a thin film distillation unit (Reliance Glass Works, Inc., Bensenville, IL). Weak Acid Dissociable Cyanide, Standard Methods, 20th Edition, Method 4500 CN- I, (WAD-CN). In this manual method, hydrogen cyanide is liberated from a sample acidified with acetic acid to pH 4.5-6.0, which is then heated for digestion and distillation with air purge. HCN in the distillate is collected by passing the distillate gas through an NaOH scrubbing solution. An aliquot of the NaOH solution is then used for color development via the standard colorimetric procedure. The evaluation of this method was performed with a MidiVap (BSI, Co., Oxford, MA) cyanide analysis apparatus which uses 50 mL samples. The method is not yet officially approved by USEPA for NPDES compliance testing but has gained acceptance in several states including Pennsylvania and Texas. The WAD-CN method has been observed (7, 8, 16) to be less prone to interferences than the Cyanide Amenable to Chlorination Method (4, 6). Automated Weak Acid Dissociable Cyanide by Thin Film Distillation Method, Automated Version of Standard Methods 4500 CN- I (WAD-TFD). This is an automated version of the method WAD-CN, employing a pH 4.0 acetate buffer, zinc acetate addition to prevent the recovery of iron cyanides, a thin film distillation unit, and an automated color development procedure.
Experimental Section Waters Tested and Sampling Procedures. The cyanide analytical methods were applied to analysis of test solutions prepared with reagent (deionized) water and five different types of contaminated waters spanning a range of compositions (Table 2). For each contaminated water, a 32-L sample was collected unfiltered. Upon collection, each sample was pretreated, if necessary, to remove sulfide (with powdered lead carbonate) and oxidants (with sodium thiosulfate) following procedures specified in Standard Methods (4). There are other potential interferences in cyanide analyses (e.g., aldehydes, surfactants), but sulfide and oxidizing agents are the most commonly encountered. From the 32 L, 2 L were taken for characterization analyses, and the remaining 30 L were adjusted to pH ) 12.5 with NaOH and then distributed into 1-L brown plastic bottles. The 30 L CW3 water sample (chlorinated POTW secondary effluent) was dechlorinated with thiosulfate prior to splitting for distribution. Groups of five 1-L samples were packed in coolers and shipped to each of the five participating laboratories for next day delivery. Table 3 summarizes the characteristics of these contaminated waters. The characterization analyses also permitted further evaluation of potential interferences for the various methods. The contaminated waters tested had total cyanide concentrations ranging from nondetectable to 182 ppb in the NaOHpreserved, unfiltered samples. The split samples of the contaminated waters were analyzed simultaneously by the participating laboratories using the assigned analytical VOL. 37, NO. 1, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
109
TABLE 3. Characterization Data for Contaminated Watersa contaminated water analyte alkalinity BOD5 chlorine - residual chlorine/oxidants - test strip cyanide - total, manual dist cyanide - WAD nitrogen - ammonia (NH3-N) nitrogen - nitrate (NO3-N) nitrogen - nitrite (NO2-N) nitrogen - total Kjeldahl (TKN) pH sulfate (SO42-) sulfide (S2-) sulfide - test strip thiocyanate (SCN-) total cadmium (Cd) total cobalt (Co) total copper (Cu) total iron (Fe) total lead (Pb) total mercury (Hg) total nickel (Ni) total silver (Ag) total zinc (Zn) total organic carbon (TOC) total suspended solids (TSS)
CW1 547 24.8 0.01 NA 0.182 NA NA
