Chemical Oxygen Demand (COD) is one of the most critical indicators for evaluating organic pollution in water, yet its measurement is highly susceptible to errors arising from sampling, reagent preparation, matrix interferences, digestion, titration, and data processing. This article provides a systematic error analysis of COD determination based on recognized standards such as ISO 6060, ASTM D1252, GB 11914-89, and HJ 828-2017. Understanding these error sources and their control strategies is essential for laboratories, environmental engineers, and quality managers seeking accurate, precise, and legally defensible COD data.
Technical Overview
Chemical Oxygen Demand (COD) is defined as the mass of oxygen, expressed in milligrams per liter, that is chemically equivalent to the amount of dichromate consumed when organic matter and oxidizable inorganic substances in a water sample are digested in a strongly acidic medium, typically with a silver sulfate catalyst. As a composite indicator of reducing pollution load, COD underpins discharge permitting, wastewater treatment process control, and surface water quality classification worldwide. Recognized standardized procedures include the open reflux dichromate titration method (ISO 6060, GB 11914-89 and its successor HJ 828-2017), the dichromate digestion methods of ASTM D1252, the rapid digestion spectrophotometric method (HJ/T 399-2007), and the chlorine-correction method for high-chloride matrices (HJ/T 70-2001).
Despite its apparent operational simplicity, COD is an operationally defined, multi-step determination in which errors propagate cumulatively through sampling, preservation, reagent preparation, digestion, titration or photometry, and data reduction. A single uncontrolled variable, such as chloride interference, drifting titrant concentration, or non-representative sampling, can shift results by tens of percent. Systematic error analysis is therefore indispensable for any laboratory operating under an ISO/IEC 17025 quality framework or reporting data for regulatory compliance. Typical quality targets include spike recoveries of 90–110%, duplicate relative deviations within method-specified limits, and expanded relative measurement uncertainty generally in the range of 5–15%.
Testing Samples
Surface Water and Groundwater
River, lake, reservoir, and groundwater samples usually exhibit low COD levels, often below 30 mg/L. At such low concentrations, even trace contamination from containers, airborne organics, or reagent impurities produces a significant relative error, making rigorous blank control and low-range analytical procedures essential.
Municipal Wastewater
Raw influent typically ranges from 200 to 600 mg/L COD, while biologically treated effluent may fall below 50 mg/L. These samples are heterogeneous, containing suspended solids, fats, oils, and grease, so thorough homogenization before subsampling is critical to obtain representative results.
Industrial Wastewater
Effluents from petrochemical, pharmaceutical, pulp and paper, food processing, fermentation, textile dyeing, and chemical manufacturing industries show extreme variability in COD concentration, from hundreds to tens of thousands of milligrams per liter. Many also contain refractory compounds, surfactants, or high salinity that introduce method-specific interferences.
High-Chloride Water Matrices
Seawater-cooling discharges, brine streams, landfill leachate, and certain chemical wastewaters contain chloride concentrations exceeding 1000 mg/L, which is beyond the direct applicability of routine dichromate methods. These matrices require chloride masking, dilution, or specialized chlorine-correction procedures.
Treated Effluent and Reclaimed Water
Compliance monitoring of final discharges against national discharge standards demands low-level accuracy, since measured values are frequently compared directly against legal limit values where measurement uncertainty can determine pass or fail conclusions.
Testing Items
The primary testing item is COD determined by the dichromate method (CODCr), applicable in unpolluted to heavily polluted waters, generally covering approximately 10–700 mg/L without dilution and higher concentrations through appropriate dilution. The permanganate index (CODMn, GB 11892) is often measured as a complementary item for relatively clean surface water.
Error analysis in COD testing also encompasses a defined set of quality control items evaluated alongside routine samples:
- Method blank values for each analytical batch
- Parallel duplicate determinations for precision assessment
- Matrix spike and recovery tests for accuracy verification
- Certified reference material verification, such as potassium hydrogen phthalate (KHP) based COD standards
- Calibration curve verification for spectrophotometric methods
- Titrant standardization checks for titrimetric methods
Testing Methods
Open Reflux Dichromate Titration
The reference method (ISO 6060, GB 11914-89, HJ 828-2017) digests a measured sample with a known excess of potassium dichromate in concentrated sulfuric acid, using silver sulfate as a catalyst and mercuric sulfate to complex chloride. After approximately two hours of refluxing, the residual dichromate is titrated with standardized ferrous ammonium sulfate using ferroin indicator, and COD is calculated from the oxidant consumed. This method offers the highest accuracy and is the legal reference approach in many jurisdictions.
Closed Reflux Microdigestion Titration
A scaled-down variant (ASTM D1252 Method B) uses sealed digestion tubes with smaller reagent volumes, reducing waste generation while retaining titrimetric quantification. Careful control of digestion temperature uniformity and leak-tight sealing is essential to prevent systematic losses.
Rapid Digestion Spectrophotometric Method
Per HJ/T 399-2007, samples are digested in pre-dosed sealed tubes at 165 °C ± 2 °C for approximately 20 minutes, and the absorbance of the residual hexavalent chromium or generated trivalent chromium is measured photometrically. A high-range system (approximately 100–1000 mg/L) is read near 600 nm, while a low-range system (approximately 15–250 mg/L) is read near 440 nm. This approach offers speed and throughput but depends heavily on calibration transfer and matrix comparability.
Chlorine-Correction Method for High-Chloride Samples
For matrices where chloride exceeds the tolerance of routine methods, HJ/T 70-2001 applies a spectrophotometric correction based on absorbance measurement near 440 nm to quantify the oxygen demand attributable to chloride oxidation, which is then subtracted from the apparent COD to obtain a corrected result.
Method Selection Considerations
- Use reflux titration for reference-level accuracy, regulatory disputes, and high-concentration samples
- Use rapid spectrophotometry for screening, process monitoring, and high-throughput routine testing
- Use the chlorine-correction method when chloride interference cannot be resolved by masking or dilution
- Verify that the selected method range, detection limit, and interference tolerance match the sample matrix and reporting requirements
Error Sources and Analysis
Sampling and Preservation Errors
Non-representative sampling is among the largest and most frequently underestimated error sources. Errors arise from inappropriate sampling locations, failure to homogenize heterogeneous wastewater containing floating oils or settleable solids, and inadequate rinsing of containers. After collection, microbial degradation of organics, volatilization of low-boiling compounds, and adsorption onto container walls alter the true COD. Samples must therefore be collected in clean glass containers, acidified with sulfuric acid to pH ≤ 2, refrigerated at approximately 4 °C, and analyzed promptly, commonly within 48 hours where applicable sampling standards govern hold times.
Reagent Preparation and Stability Errors
Errors in reagent preparation translate directly into systematic bias. Potassium dichromate must be prepared from a primary-standard grade material dried to constant weight and weighed accurately. Ferrous ammonium sulfate titrant is inherently unstable and gradually decreases in concentration through air oxidation, so it must be re-standardized daily against the dichromate standard. Incompletely dissolved silver sulfate–sulfuric acid reagent, degraded ferroin indicator, contaminated acid, or impure water used for blanks all contribute measurable positive or negative shifts. Every reagent should be logged with preparation date, preparer, and verification result.
Interference-Related Errors
Chloride is the dominant interference in COD determination: it is oxidized by dichromate in the acidic digestion medium, consuming oxidant and producing a positive bias that can overwhelm the true organic demand in saline samples. Mercuric sulfate is added at approximately a 10:1 mass ratio relative to chloride for masking, and routine dichromate methods are generally not applicable when chloride exceeds about 1000 mg/L without dilution or correction. Nitrite also consumes oxidant and is typically masked with sulfamic acid. Reduced inorganic species such as ferrous iron and sulfide contribute apparent COD that may or may not be relevant to the monitoring objective. Conversely, certain refractory organics, including pyridine and some aromatic and straight-chain aliphatic compounds, are incompletely oxidized even with the silver catalyst, and volatile organics may be lost from open reflux systems, producing negative bias.
Digestion Process Errors
Complete and reproducible oxidation requires strict control of digestion temperature and time. Insufficient reflux time or a heating mantle that cannot sustain gentle, stable boiling leads to incomplete oxidation and negative bias. In rapid digestion systems, block temperature deviation from 165 °C ± 2 °C, uneven heating across tube positions, or shortened digestion periods degrade comparability. Condenser inefficiency, bumping, or organic contamination of digestion vessels and boiling chips introduces both losses and positive contamination errors. All glassware should be acid-washed and rinsed with organic-free water, and digestion conditions must be validated across the full tube rack.
Titration and Endpoint Errors
Titrimetric quantification is subject to endpoint subjectivity: the ferroin transition from yellow-green to red-brown is judged visually, and inter-operator color perception differences routinely introduce several percent variation. Consistent lighting and background, standardized titration speed with a distinct final drop-wise approach, and titration at the temperature specified by the method all improve repeatability. Burette calibration status, drainage timing, parallax in reading, and failure to keep sample titrant consumption within the optimal window, commonly between one-third and two-thirds of the blank titration volume, amplify relative error. Correcting for the method blank is mandatory.
Photometric and Instrumental Errors
In spectrophotometric COD methods, wavelength accuracy deviations, stray light, detector drift, and absorbance nonlinearity at high concentrations all distort results. Cuvettes must be matched, clean, scratch-free, and consistently oriented, since fingerprints, water droplets, or fogging on sealed digestion tubes measurably alter absorbance. Calibration curves prepared from potassium hydrogen phthalate may not perfectly match the response of real wastewater constituents, creating matrix-mismatch bias, and turbidity or particulates in the digest scatter light. The calibration must be verified with independent standards, and curve linearity, typically requiring a correlation coefficient of 0.999 or better, must be confirmed before use.
Blank, Calibration, and Calculation Errors
The method blank reflects the collective contamination contribution of water, reagents, glassware, and the digestion atmosphere, and every reported COD value is blank-corrected. An elevated or drifting blank invalidates the batch and must be investigated before data release. Additional data-reduction errors include incorrect dilution factors from imprecise pipetting or arithmetic mistakes, use of non-calibrated volumetric glassware, inappropriate rounding, and misapplication of significant figures, particularly when reporting values near regulatory limits or method detection limits.
Human and Environmental Factors
Operator technique differences in sampling, sub-sampling, titration speed, and endpoint recognition remain a persistent component of between-operator variability. Laboratory environmental conditions, such as ambient temperature fluctuations affecting volumetric measurements and titrant molarity, airborne organic contamination near digestion areas, and inadequate documentation or transcription practices, further compound total error. Regular training, standardized operating procedures, and independent data review are the primary countermeasures.
Quality Control and Error Reduction Strategies
An effective error control program addresses the entire measurement chain rather than isolated steps. Laboratories accredited to ISO/IEC 17025 or performing compliance testing for regulators and accredited third-party laboratories should implement, at minimum, the following measures:
- Analyze a method blank with every batch, typically no more than 20 samples per batch, and apply control limits
- Perform parallel duplicate determinations on at least 10% of samples, with acceptance criteria commonly set at relative deviation within 10% for COD at or above 100 mg/L and within 20% below that level
- Verify accuracy through matrix spikes with acceptance limits of 90–110% recovery and through periodic analysis of certified reference materials
- Standardize ferrous ammonium sulfate titrant daily against freshly prepared or verified dichromate standard
- Confirm digestion temperature uniformity across heating blocks and reflux units on a scheduled basis
- Validate calibration curves for photometric methods with independent verification standards before sample analysis
- Control chloride interference through mercury masking, dilution, or the chlorine-correction method, and document the applied strategy
- Enforce sampling SOPs covering container type, acidification to pH ≤ 2, refrigeration, and maximum hold times
- Maintain control charts to detect drift and trends in blank values, recovery, and duplicate precision
- Participate in interlaboratory comparisons and proficiency testing programs to assess total method bias
- Evaluate and report measurement uncertainty in accordance with recognized guidelines such as the GUM framework
By treating COD determination as a system and quantifying the contribution of each error source, laboratories can typically reduce expanded relative uncertainty to well below 10%, ensuring that reported values reliably support compliance decisions and process optimization.
Applications
Error-analyzed COD measurement supports decision-making across the full water management cycle:
- Regulatory compliance monitoring of industrial and municipal discharges against national effluent standards
- Process control of activated sludge, anaerobic digestion, and advanced oxidation systems in wastewater treatment plants
- Surface water and receiving-water quality assessment within environmental monitoring programs
- Treatability studies and technology selection for industrial wastewater, including evaluation of chemical dosing and aeration demand
- Pollution source apportionment and total pollutant load allocation programs
- Verification testing by accredited third-party laboratories for permitting, litigation support, and buyer verification
- Emergency response monitoring following spills of organic pollutants
In every application, the credibility of the COD value depends directly on the rigor of the underlying error analysis and quality control program.
FAQ
Why does high chloride cause a positive error in COD determination?
Chloride ions are oxidized by dichromate in the hot acidic digestion medium, consuming oxidant that should be reserved for organic matter. This makes the sample appear to contain more organic pollution than it actually does. The bias is controlled by complexing chloride with mercuric sulfate at approximately a 10:1 mass ratio, by diluting the sample, or by applying the dedicated chlorine-correction method for highly saline matrices.
What is the acceptable relative deviation for duplicate COD measurements?
Commonly applied acceptance criteria, consistent with typical national standard requirements, are a relative deviation within 10% when COD is at or above 100 mg/L and within 20% at lower concentrations. Duplicates exceeding these limits indicate homogenization, digestion, or titration instability and should trigger investigation and re-analysis of the batch.
How should COD samples be collected and preserved?
Collect samples in clean glass containers after adequate homogenization of the water body or stream, acidify with sulfuric acid to pH ≤ 2 to suppress microbial activity, store at approximately 4 °C in the dark, and analyze as soon as possible, typically within 48 hours where sampling standards specify hold times. Never preserve with acid if the sample must also be tested for other acid-sensitive parameters without splitting.
Which is more accurate: open reflux titration or rapid digestion spectrophotometry?
Open reflux titration is the reference method and generally provides the highest accuracy and broadest applicability, particularly for high-concentration or legally significant samples. Rapid digestion spectrophotometry offers speed, lower reagent consumption, and high throughput, and is well suited to screening and process monitoring, provided calibration is frequently verified and matrix effects are assessed. Results from rapid methods should be periodically cross-checked against reflux titration.
What does a high or drifting method blank indicate?
A high blank signals contamination from water, reagents, glassware, or the laboratory environment, while a drifting blank indicates reagent degradation, most commonly air oxidation of the ferrous ammonium sulfate titrant or deterioration of the dichromate solution. Any batch with an out-of-control blank must be rejected, the contamination source identified and eliminated, and the samples re-analyzed.
What measurement uncertainty can be expected for COD results?
With disciplined quality control, expanded relative measurement uncertainty for COD is typically in the range of 5–15%, dominated by contributions from sampling representativeness, chloride and other interferences, endpoint or photometric precision, and titrant standardization. Laboratories reporting results near regulatory limits should always report uncertainty so that pass-or-fail conclusions can be drawn with stated confidence.