Environmental Assessment: Sampling and Analysis PDF

Environmental Assessment: Sampling and Analysis ERSC-2230H-2025WI accuracy and precision Objectives for today’s lecture 1. accuracy and precision… Wednesday 05 February 2025 reminder: the normal (gaussian) distribution underpins the statistical te...

Environmental Assessment: Sampling and Analysis ERSC-2230H-2025WI accuracy and precision Objectives for today’s lecture 1. accuracy and precision… Wednesday 05 February 2025 reminder: the normal (gaussian) distribution underpins the statistical tests that allows us to evaluate the quality of environmental analytical data! ≈68.3% ± 1σ ≈95.5% ± 2σ ≈99.7% ± 3σ 68.3% of data within ±1σ µ = population mean | 𝑥ҧ = sample mean σ = population standard deviation | s = sample standard deviation the analytical method should be precise and accurate accuracy: how close a measured value is to the ‘true’ value precision: describes how close replicate analysis results on the same sample are to one another ▪ every analysis is wrong; the best you can do is minimize all possible errors ▪ it is important to remember that errors propagate through the entire analytical process always question your results! types of error in environmental analysis 1. systematic (determinate) errors (bias) analytical results all too high or too low. Affects accuracy. 2. random (indeterminate) errors data scattered (approximately) symmetrically about a mean value. Affects precision, can be dealt with statistically. 3. spurious (gross) errors usually obvious as ‘outlier’ results. Detectable by carrying out sufficient replicate measurements. systematic (bias) versus random error truth refers to the Measured ‘true’ value of our analyte. systematic The measured (or (bias) analytical / apparent) value has low accuracy, it is systematically biased to low concentrations, with high random error (lower precision around the measured value). random error How do you quantitively assess accuracy and precision? accuracy is the degree of agreement of a measured value with the true or expected value. Accuracy is measured and expressed as % recovery and calculated according to: % recovery = (measured value / true value) × 100 the true value (concentration) is rarely known for environmental samples. Thus, accuracy is typically determined by spiking a sample with a known quantity of a standard: % recovery on spike = spiked sample value – sample value × 100 spiked add an accurate test should achieve a percentage recovery close to 100%. we will revisit % recovery a little later… precision is the degree of mutual agreement among individual measurements (x1, x2,.. xn) as the result of repeated measurements under the same condition. Precision measures the variation among measurements and may be expressed in terms of relative standard deviation (RSD) and relative percent difference (RPD): RSD = (s/𝑥) ҧ × 100 = CV × 100 RPD = [(A – B)/((A + B)/2)] × 100 where CV denotes coefficient of variation (CV = s/ 𝑥) ҧ and RPD is the difference between the duplicate values (A and B) divided by the average of the duplicate values and multiplied by 100. RSD is used for the evaluation of multiple replicate measurements, whereas RPD is used for measuring precision between two duplicate measurements. we will revisit RSD a little later… quality assurance / quality control (QA/QC) ‘QA/QC is a set of performance characteristics to quantify the quality and test the integrity of analytical data’ involves a number of quality control (assessment) tools field QC samples (e.g., blanks) laboratory QC samples (a) preparation and (b) analysis blanks Note: QC samples are the extra samples collected during the sampling and analytical processes. These QC samples are fundamentally different from the true environmental samples. what is a blank? reagents dust a blank is a sample that is identical to equipment the real sample but contains no measurable you! concentration of the analyte. the purpose of blanks is to assess systematic errors caused by contamination of the reagents, or the sample with the analyte usually during sampling, preservation, or sample pretreatment. quality control procedures during field sampling field (equipment) blanks: field blank are used to detect any contamination from sampling and sampling equipment. These blanks are prepared in the field before sampling begins (collect one per type of sampling device per day). travel (trip) blanks: a travel blank sample undergoes the full handling and shipping process of an actual sample. It is designed to detect sample contamination that can occur during field operation or during shipment (collect one per day) field replicate samples: are field samples obtained from one sampling point, homogenized, divided into separate containers, and treated as separate samples throughout the remaining sample handling and analytical processes. These field replicate samples are used to assess error associated with sample heterogeneity, sample methodology, and analytical procedures (collect ~10% of the samples for duplicate). comparison of analytical results for different types of blanks used during sample collection; remember these do not contain the analyte! 100 deionised water Concentration [mg/L] 1% HNO3 10 trip blank field blank 1 0.1 Al Si Fe Cu Zn Ba quality control procedures during sample preparation significant errors can occur during sample preparation, for example during sample digestion for metals analysis due to (1) cross- contamination, (2) incomplete digestion, or (3) improper procedure. we can control these errors by ‘spiking’ QC samples: spiked sample/matrix spike: a small quantity of a known concentration of analyte stock solution is added to the sample before the sample preparation. This sample then undergoes digestion or extraction and subject to analysis. we can estimate the recovery of analyte from the sample matrix, which helps to estimate accuracy (systematic error). spike prerequisites: spike is completely equilibrated with sample matrix spike has the same physical and chemical form as native analyte spiked concentration is comparable to native concentration estimating % recovery of analyte from a sample accuracy is the degree of agreement of a measured value with the true or expected value. Accuracy is measured and expressed as % recovery and calculated according to: % recovery = (measured value / true value) × 100 the true value (concentration) is rarely known for environmental samples. Thus, accuracy is typically determined by spiking a sample with a known quantity of a standard: % recovery on spike = spiked sample value – sample value × 100 spiked added an accurate test should achieve a % recovery close to 100%; values between 80% and 120% are acceptable. example for the determination of lead (Pb) in soil (a) sample: take 1 g of soil and digest (microwave) with 10 mL of concentrated nitric acid (HNO3); (b) spiked sample: replicate with Pb ‘spike’ of 500 μg added to soil sample. (a) sample (b) spiked sample example for the determination of lead (Pb) in soil (a) sample: take 1 g of soil and digest (microwave) with 10 mL of concentrated nitric acid (HNO3); (b) spiked sample: replicate with Pb ‘spike’ of 500 μg added to soil sample. Measure lead concentration on both samples using ICP-MS. (a) sample (b) spiked sample concentration of spike added to solid Pb in original sample: 500 μg lead sample post-digest is: 10 μg/mL concentration of Pb in ‘spiked’ sample post- digest is: 55 μg/mL example for the determination of lead (Pb) in soil true versus measured (apparent) concentration: take 1 g of soil and digest (microwave) with 10 mL of concentrated acid concentration of lead in sample measured by ICP-MS = 10 μg/mL spike added to soil sample = 500 μg → 500 μg/10 mL = 50 μg/mL concentration of lead measured in spiked sample by ICP-MS = 55 μg/mL recovery = (spiked sample – original sample) / spike added = (55 μg/mL – 10 μg/mL) / 50 μg/mL = 0.9 (or × 100 to get 90% recovery) ‘measured’ concentration in soil = 10 μg/mL × 10 mL / 1 g = 100 μg/g ‘true’ concentration = measured concentration / recovery → = (100 μg/g ÷ 0.90) = 111.11 μg/g soil within groups of three, answer the following (show your work) A 100 mL effluent sample is spiked with 1 mL of 100 mg/L nitrogen standard yielding the following results: the measured concentration of nitrogen in the effluent sample is 0.74 mg/L. The measured concentration in the spiked effluent sample is 1.21 mg/L. what is the recovery? What is the true concentration? remember c1 × v1 = c2 × v2 where c = concentration and v = volume record your answer on a sheet of paper, include the names and student ID of each group member on your page answer: a 100 mL effluent sample is spiked with 1 mL of 100 mg/L nitrogen standard yielding the following results: measured concentration in effluent sample = 0.74 mg/L measured concentration in spiked effluent sample = 1.21 mg/L spike added is 1 mL of 100 mg/L standard into 100 mL sample → using c1 × v1 = c2 × v2 → spike added in 101 mL = 0.99 mg/L recovery = (spiked effluent sample – effluent sample) / spike added = (1.21 - 0.74)/0.99 = 0.4747 or ~0.475 = 47.5% (× 100 to get % recovery) ‘true’ concentration = 0.74 /0.475 = 1.557 mg/L ~ 1.56 mg/L quality control procedures during analysis there are mainly four types of QC samples used for the QA/QC purposes during laboratory analysis: (a) blanks used to assess any potential contamination; (b) spikes used to obtain the percentage recovery and therefore the accuracy; (c) standards for accuracy, and (d) replicates (duplicates) used to determine the analytical precision. blanks are analyte-free water analyzed with the samples. Such blanks will detect any introduction of contaminants and artifacts into samples (false positive results). laboratory control samples (LCS) contain known identities and concentrations of target analytes in clean water. matrix spikes and matrix spike duplicates (MS/MSDs) are field samples spiked in the laboratory prior to sample extraction/digestion and analysis. The percentage recovery is calculated to assess the accuracy, and the relative percent difference is calculated and used to assess analytical precision. quality control procedures during analysis laboratory duplicates are aliquots of the same sample that are prepared and analyzed at the same time but submitted and analyzed as separate samples (see also field duplicates). The analyst does not know that they are duplicates. Discrepancies in duplicate samples indicate poor analytical reproducibility (the analytical precision) and/or the poor homogenization in the field. reference materials are obtained from an independent source with known analytical level(s). calibration standards are the standard solutions used to obtain calibration curves, including a calibration blank and a series of several concentrations (see Laboratory 02!) adding all these QC samples together, the total number of required QC samples should be around 25% of the samples sent out for laboratory analysis. reference materials: definition: a sample of known (‘true’) concentration purpose: independent control of accuracy manufacturer certifies concentration of analyte in sample: self-made consensus result obtained by multiple laboratories prerequisite: should be as similar as possible to the actual samples if unavailable, you can make your own as (inferior) substitute reference materials: laboratory inter-comparison of arsenic in groundwater 35 30 25 arsenic [mg/L] 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 laboratory ID example shows results of analysis of the same sample by 23 different laboratories. Trent is shown in orange.