Introduction To Analytical Chemistry PDF

Introduction to Analytical Chemistry 1  Analytical Chemistry Analytical Chemistry – the science and technology used to determine the chemical composition of samples It tries to answer these questions: What is in the sample? – qualitative analysis How...

Introduction to Analytical Chemistry 1  Analytical Chemistry Analytical Chemistry – the science and technology used to determine the chemical composition of samples It tries to answer these questions: What is in the sample? – qualitative analysis How much of it is there in the sample? – quantitative analysis 2 What is Analytical Chemistry used for? Clinical – tests on blood and other bodily fluids Forensic – CSI Environmental – samples of rocks, soil, sediment, water, air, biota (plant and animal matter), etc. Food – what is in the food and how much is there? Quality control – checking on the manufacturing process (refineries, pharmaceuticals, etc.) Scientific research – geology, anthropology, exploration of planets, etc. Chemistry On Mars 3  What is a sample? Consider that you have to find the concentration of lead in Lake Ontario water. Population – all of the water in Lake Ontario You cannot do measurements on all of the water in Lake Ontario, so you take samples of water from the lake. We infer information about a population from observations made on a subset or sample taken from the population. You might collect three bottles of water from the lake. These individual bottles are called sampling units or sampling increments. 4  Samples The collection of samples that you have are called the gross sample. In taking samples, the goal is always to get a sample that truly reflects the composition of the population (at the time that the sample was collected); that is, we want a representative sample. 5  Samples Very often, we have to be specific about the meaning of the results of our analysis. It may not be as simple as we originally thought. The question we had was: –What is the concentration of lead in Lake Ontario water? 6 Lake Ontario 7  Samples Maybe the question about lead in Lake Ontario water should be more specific. Would you expect the lead concentration to be the same everywhere in the lake? Would it be the same at different depths? Would the results of your lead determination be different if the samples were collected at a different time? Day or night, weekday or weekend, summer or winter. Dissolved lead or total lead? 8  Dissolved or Total Metals Environmental water samples (e.g., water from a river or lake) are very often not perfectly clear. They may have solid particles in them that are not dissolved (silt, sand, sediment, algae, etc.) If there are metals in the water, the metals can be dissolved, or they can be attached to the surface of tiny solid particles suspended in the water (not dissolved). Metals could also be stuck inside tiny particles in the water. 9  Dissolved or Total Metals Total metal in a water sample includes both the metal that is dissolved and the metal that is stuck in or on any undissolved solid particles in the water. You must treat the sample differently if you want to determine dissolved metal than if you want to determine total metal. To find the concentration of dissolved metal, filter the water to remove any undissolved particles, then measure the metal concentration in the filtrate. 10  Dissolved or Total Metals To find the concentration of total metal, digest the water sample (to dissolve the metal which is not dissolved) and then measure the metal concentration in the digestate. Digesting a sample dissolves things in the sample that were not dissolved before (usually by boiling the sample in mineral acid). 11  Samples Samples can be homogeneous or heterogeneous. Homogeneous samples have the same composition (at the microscopic level) throughout, for example a liquid solution (no undissolved particles). Heterogeneous samples have constituent parts that can be distinguished visually or with a microscope. 12  Homogeneity We always want to get a representative sample, and this is easier to do when the population is homogeneous (e.g., water samples). It is harder to do when the population is heterogeneous (e.g., soil samples). 13  Analytical Methods Once the samples come to the laboratory, they are analyzed using one or more analytical methods. An analytical method is a written document that clearly specifies the reagents, apparatus, and procedures needed to do the analysis. This kind of document is often called a Standard Operating Procedure (SOP). It is similar to an experiment in your lab manual at Sheridan, but it is much more detailed. 14  Classes of Analysis Methods can be classified as – wet chemistry methods (e.g., gravimetric methods, volumetric methods) – instrumental methods (e.g., spectroscopy, chromatography) They can also be classified by what is being measured Organic Analysis – the determination of carbon compounds in samples – Most important technique - Chromatography Inorganic Analysis – the determination of elements, ions, and non-carbon compounds – Most important technique – Spectroscopy 15  But first, some terminology Analyte - an analyte is a chemical (element, ion, or compound) that is being measured or identified in a chemical test. It may also be a group of chemicals that belong to the same chemical family, and which are being determined together. Parameter – another name for analyte. 16  But first, some terminology Aliquot – an accurately measured portion of a sample or other liquid. Matrix – the water, soil, sludge, sediment, or whatever that makes up most of the sample. Also called the sample type. The analyte is said to be in the matrix. 17  Samples and Standards In analytical chemistry, you are always doing measurements on samples or standard solutions. A sample is an amount of material in which you are trying to find the concentration of an analyte. The concentration is unknown until the analysis is completed. A standard solution (often just called a standard) is a solution in which the concentration of analyte is known. The concentration is known because the standard solution was prepared by accurately weighing out a pure compound and dissolving it in a precisely known volume in a volumetric flask, or some other method was used to find the concentration accurately. 18  Important Concepts in Analytical Chemistry Accuracy Precision Calibration Detection Limits Linear Dynamic Range Quality Control Quality Assurance 19  Accuracy The agreement between a measurement (or the average of a group of measurements) and the “true” value. Quantitatively expressed as bias Bias = measured - true x 100% true Example: If the true value for a number is 10.0 mg/L and a measurement of it comes out as 9.2 mg/L, what is the bias? Answer: 9.2 - 10.0 x 100% = -8.0 % bias 10.0 Bias can be positive (measured result is too high) or negative (measured result is too low). Measured value is less than the true value ⇒ negative bias or biased low 20  Accuracy Almost always we don’t know the “true” value for the concentration of an analyte in a sample. If we don’t know the true value of the concentration of an analyte in a sample, we can’t say anything about the accuracy of that particular measurement. But we can estimate the accuracy of the method in general. Accuracy is estimated from the recovery of known standards. 21  Recovery of Standards You prepare a standard solution in which you know the concentration of analyte very precisely. This concentration is the “expected value.” You measure that standard solution using your analytical method. You calculate the ratio of the concentration you found with your method to the expected value of concentration, and express it as a percentage. That is called the % recovery. 22  Recovery of Standards 23  Recovery of Standards 24  Recovery of Standards By measuring known standards regularly (several times every day) and keeping a record of the % recovery, an analyst knows what the accuracy is for a method in general. 25  Precision The degree of mutual agreement among replicate measurements. How scattered are the results? For several replicate measurements (>3), precision is usually expressed as the standard deviation (SD). Sometimes expressed as percent relative standard deviation (%RSD). % RSD = SD x 100% mean If results are more precise (that’s what we want), they have a lower standard deviation. 26  Accuracy and Precision Not precise Not precise Not accurate Accurate Precise Precise Not accurate Accurate 27  Detection Limits For every kind of analytical method, there is a detection limit. What do you think “detection limits” means? The smallest concentration that you can detect. The smallest concentration that you can measure. Are these two things the same ? No, they are not. 28  Detection Limits To detect something is to have evidence that some of it present, as opposed to none of it being present. It is a yes/no decision. To measure something is to find a meaningful number for its concentration. It is quantitative. There must be more of a substance present if you want to measure it than to simply detect it. 29  Limits Therefore, in Analytical Chemistry, we speak of Detection Limits and Quantitation Limits. Quantitation Limits are always higher than Detection Limits. Keep in mind that these are measurements at the limits of the capabilities of the analytical methods. 30  Method Detection Limit The formal definition: – Method Detection Limit (MDL): the minimum concentration of a substance that can be measured and reported with 99% confidence that the analyte concentration is greater than zero. Also called Limit of Detection (LOD). What it really means: – The smallest concentration at which you can say there is something (not nothing) there. You can’t reliably say how much. 31  Estimated Quantitation Limit The formal definition: – Estimated Quantitation Limit (EQL): the lowest concentration that can be reliably measured within specified limits of precision and accuracy during routine laboratory operating conditions. Also called Limit of Quantitation (LOQ). What it really means: – The smallest concentration at which you can report a number. The EQL is generally 5 to 10 times the MDL. 32  Reporting Limit The Reporting Limit (RL) is the lowest concentration that a laboratory reports for an analyte. A concentration below that is considered to be “not detected.” Laboratories usually use their EQL as the reporting limit. The reporting limit used by a laboratory must always be above the MDL. 33  Why do MDLs and RLs matter? Why does anyone care how low the concentrations are that a lab can measure? Most of the samples that an environmental lab analyzes need to be measured because Environmental regulations set a limit on how high concentrations of chemicals can be in many types of sample, e.g., rivers, lakes, landfills, etc. For example, Ontario Regulation 169 specifies the maximum allowable concentrations of chemical elements and compounds in our drinking water. These maximum concentrations are called Standards. 34  This is a part of Ontario Regulation 169, showing the highest allowed concentrations of elements and compounds in our drinking water. It shows, for example, that the highest allowable concentration of benzene is 0.001 mg/L, for cadmium, it is 0.005 mg/L. 35  Reporting Limits If a laboratory wants to measure benzene in drinking water in Ontario, it has to be able to reliably measure benzene well below the 0.001 mg/L standard, not just down to the standard. Of course, the same can be said for the determination of all the other parameters in drinking water and also for all the parameters under other regulations. 36  Limits How is this all actually done in practice? For every analytical method, the laboratory determines what the MDL is on a regular basis (typically once a year). The laboratory multiplies the found MDL by a safe margin (about 5x to 10x) to get its EQL/RL. The laboratory publishes its RLs. 37  Method Detection Limit (MDL) Depends on the entire method, not just the measurement step. It is determined from several replicate analyses of a sample in a given matrix containing the analyte. The MDL is calculated from the precision of the measurements. The MDL is not related to the accuracy of the method. 38  How is the MDL determined? 1) Estimate the MDL. 2) Prepare a solution which has a concentration of about 5 to 10 times the estimated MDL. 3) Take 8 portions (aliquots) of that solution through the entire measurement process. 39  How is the MDL determined? 4) Calculate the standard deviation of the 8 results 5) Use Student’s t: MDL = t x SD For n = 8, t = 2.998 For n = 7, t = 3.143 40  Example of MDL Determination The determination of Total Copper in water by Graphite Furnace Atomic Absorption Spectroscopy. 1) Analyst estimates that the MDL is 0.0007 mg/L. 2) Analyst prepares a solution which has 0.005 mg/L of copper. 3) Analyst digests 8 aliquots of the solution. Analyst measures copper in 8 digests on GFAA instrument. 41 Example of MDL Determination 0.0049 0.0049 Mean = 0.0051 mg/L 0.0052 SD = 0.000196 mg/L 0.0054 0.0053 MDL = 2.998 x 0.000196 0.0049 = 0.00059 mg/L 0.0051 = 0.59 µg/L 0.0052 42 What if the 8 numbers were... 0.990 Mean = 0.977 mg/L 1.001 0.977 SD = 0.076 mg/L 0.977 0.809 MDL = 2.998 x 0.076 1.005 = 0.23 mg/L 0.976 But one result looks quite 1.078 different from the others. 43 We might be able to reject the suspect result. Use Dixon’s test for outliers. 0.809 - 0.976 = R 1.078 0.809 - 1.005 1.005 1.001 If R > 0.55, result is an 0.990 outlier and should be 0.977 rejected. 0.977 In this case, R = 0.85 0.976 0.809 should be rejected. 0.809 44 Calculate the MDL using only 7 results 1.078 Mean = 1.001 mg/L 1.005 SD = 0.036 mg/L 1.001 0.990 MDL = 3.143 x 0.036 0.977 = 0.11 mg/L 0.977 0.976 Use t = 3.143 for n = 7 0.809 Instead of t = 2.998 45  Dixon’s Test How do you know if you should do Dixon’s Test? You should do Dixon’s Test if there is any chance at all that one number is different from the other 7. Do Dixon’s Test if one number looks significantly higher or significantly lower than the other 7 numbers. 46  Dixon’s Test Can you use Dixon’s Test twice on one data set, i.e., when two numbers look very different from the other 6? No, Dixon’s Test can be used only once on a set of 8 replicates, when one number looks like it might be different from the other 7 numbers. You cannot reject 2 out of 8 numbers to calculate your MDL based on 6 replicates. You can calculate MDL based on 8 replicates, or 7, but not 6. 47  Why does it matter what concentration is used? At higher concentrations, the SD will be higher. If MDL determination is done at a concentration that is too high (> 10 x MDL), the MDL that is found will be too high - not the true MDL. 48 How often must MDL be determined? It depends on the analytical laboratory’s policy – MDLs must be redetermined every year or every 2 years. MDL must be redetermined if there is a major change to the method or to the instrument. A determination of MDL is a rough estimate - it is very imprecise. 49  Method Detection Limit or Instrument Detection Limit? Sometimes IDL is determined. It is not as useful as MDL. IDL determination is the same as MDL determination except that the variability of the preparation steps is not included. If there is no sample preparation in the method, then IDL is the same as MDL. 50  Linear Dynamic Range Some instruments reach “saturation” as the concentration is increased. The Linear Dynamic Range refers to the range of concentrations that are still on the linear part of the calibration curve. 51  Quality Control (QC) QC is done at the bench level by the analyst as a routine, structured part of analysis. It is done by analyzing QC samples along with real samples within a batch or run. Method Blank (sometimes simply called Blank) Duplicates or Repeats Spiked Blank or QC Standard Matrix Spike or Spiked Sample SRMs and CRMs 52  Method Blank The formal definition: – An aliquot of reagent water or other blank matrix that is treated exactly as a sample including exposure to all glassware, equipment, solvents, reagents, and internal standards, that are used with other samples. The Method Blank is used to determine if method analytes or other interferences are present in the laboratory environment, the reagents, or the apparatus. What it really means: – Water with nothing in it (except the preservatives that are added to normal samples). You should detect nothing when you do a measurement. 53  Field Blank or Travelling Blank The formal definition: – An aliquot of reagent water or other blank matrix that is placed in a sample container in the laboratory and treated as a sample in all respects, including shipment to the sampling site, exposure to sampling site conditions, storage, preservation, and all analytical procedures. The purpose of the FB is to determine if method analytes or other interferences are present in the field environment. What it really means: – Fill a bottle with pure water in the lab; send it to the sampling site and bring it back. 54  What is an acceptable level for a Method Blank? If the Method Blank is measured as less than the Reporting Limit, i.e., “not detected,” this would be acceptable. This would pass QC. If any amount greater than the Reporting Limit is measured in the Method Blank, that is considered to be a QC failure. 55  What happens when there is a QC failure? The analyst must stop doing analyses. The analyst must find and correct the problem that caused the QC failure. Once the problem has been corrected and QC again passes, regular analyses can resume. 56  Laboratory Duplicates The formal definition: – Two aliquots of the same sample taken in the laboratory and analyzed separately using identical procedures. These duplicates will demonstrate the precision associated with laboratory procedures, but not with sample collection, preservation, or storage. What it really means: – Measure one sample twice in a row. 57  Field Duplicates The formal definition: – Two separate samples collected at the same time and placed under identical circumstances and treated exactly the same throughout field and laboratory procedures. They give a measure of the precision associated with sample collection, preservation and storage, as well as with laboratory procedures. What it really means: – Fill two sample bottles at once at the sampling site. 58 How closely do duplicates agree? The Relative Percent Difference (RPD) is calculated. m1 and m2 are the sample and duplicate measurements. RPD = abs(m1 - m2) x 100% (m1 + m2)/2 59  RPD The Total Phosphorus in a sample is measured as 4.7 mg/L. The duplicate is measured as 5.3 mg/L. RPD = abs(4.7 - 5.3) x 100 (4.7 + 5.3)/2 = 0.6 x 100 = 12% RPD 5.0 60  What are acceptable duplicates? Depending on the analysis, in order to pass QC, the RPD must be 15% or less (inorganics) up to 50% or less (organic soils). The acceptable RPD will be specified in the laboratory procedure. 61  Spiked Blank The formal definition: – An aliquot of reagent water or other blank matrix to which known quantities of the method analytes are added in the laboratory. It is analyzed exactly like a sample, and is used to determine whether the method is in control. What it really means: – Make up a solution with an exactly known concentration to check the accuracy of your method. 62  Spiked Blank or QC Standard Organic analysts often call it Spiked Blank, or simply Spike. Inorganic analysts often call it QC Standard. Two different names for the same thing. It is absolutely critical that the Spiked Blank be prepared completely independently from the calibration standards. What do we mean by “prepared independently?” 63  Spiked Blank or QC Standard That means it is prepared from a different bottle of reagent, from a different manufacturer, if possible. This is called using a “second source.” Why is it crucial that the Spiked Blank (QC Standard) be made from a different bottle of reagent than the calibration standards? Hint: What if the bottle contents had partially decomposed, or has contaminants in it? 64 What is an acceptable spiked blank? Acceptance limits are based on historical data. At least 25 points are collected over time. The mean and standard deviation are calculated. A result that is within 3 SD of the mean would be considered acceptable. 65  Matrix Effects Standard solutions prepared in a laboratory are almost always very simple. They usually consist of one, or a few, pure compounds dissolved in RO water. In contrast to this, real samples very often are complex mixtures containing perhaps dozens of components. It is possible that some of the many components in the sample may interfere with your measurement of the one component in the sample that you are trying to determine. 66  Matrix Effects Something in the matrix of the sample may cause your measurement to be too low, i.e., it may suppress your measurement. Alternatively, something in the sample may cause your measurement to be too high, i.e., it may erroneously enhance your measurement. Both of these kinds of interference due to components in the sample matrix are called matrix effects. A laboratory can find out of there is any matrix effect on a sample by analyzing a QC called a Matrix Spike or Spiked Sample. 67  Matrix Spike or Spiked Sample The formal definition: – A sample prepared by adding a known volume of a target analyte to a specified amount of matrix sample for which an independent estimate of the target analyte concentration is available. Matrix spikes (spiked samples) are used to determine the effect of the matrix on a method's recovery efficiency. What it really means: – Make up a solution in which you add an exactly known concentration to a sample instead of to pure water. 68  Matrix Spike or Spiked Sample Example: – You measure Fe in a sample as 11.4 mg/L. – You take another aliquot of the same sample and add 5.0 mg/L Fe to it. You “spiked” the sample. – You measure the spiked sample. – You expect to get 11.4 + 5.0 = 16.4 mg/L. – If you do measure 16.4 mg/L in the spiked sample, there is no matrix effect. You can conclude that the measurement on the original sample is unaffected by its matrix. 69  Matrix Spike or Spiked Sample – If you measure lower than 16.4 mg/L in the spiked sample, that means that something in the matrix of the sample makes it appear lower than it really is; there is a matrix effect, in which the sample matrix suppresses the measurement of Fe. – This means the measurement of the original sample is probably also lower than the true value. – It is also possible that you measure higher than 16.4 mg/L in the spiked sample. That would mean that something in the sample matrix causes the measurement of Fe to be too high. 70 What is an acceptable matrix spike? Acceptance limits are based on historical data. At least 25 points are collected over time. The mean and standard deviation are calculated. A result that is within 3 SD of the mean would be considered acceptable. In this case, it must be expressed as % Recovery. 71  Certified Reference Materials and Standard Reference Materials Certified reference materials (CRMs) are samples, available in different matrices, that have been extensively analyzed by several laboratories and methods and have certified concentration values for specific compounds analyzed. 72 Certified Reference Materials and Standard Reference Materials Standard reference materials (SRMs) are certified reference materials that have been certified by an organization recognized in setting standards, e.g., NRC, NIST. National Institute of Standards and Technology (NIST) SRMs LKSD-1 to LKSD-4 Lake Sediment Samples | Natural Resources Canada 73 Certified Reference Materials and Standard Reference Materials CRMs or SRMs are used by laboratories as a check on the accuracy of their methods. CRMs or SRMs are not available for all analyses. 74  How do all these QC samples fit in with the real samples? Every method has a Run Layout that specifies the order in which real samples and QCs are run within a batch. The Run Layout has strict requirements for frequency of QCs within the run. The analyst is required to follow the Run Layout given in the SOP for the method. 75 Run Layout for Nitrilotriacetic Acid Determination Cal Blank (0 mg/L) Cal Std 0.4 mg/L Cal Std 1.0 mg/L These are calibration standards Cal Std 2.0 mg/L Cal Std 3.0 mg/L Cal Std 5.0 mg/L QC Standard, 2.0 mg/L (second source) Method Blank Sample 1 These 3 are QCs Sample 1 Lab Duplicate Sample 2 Sample 3 Sample 4 Sample 5 76 Run Layout for Nitrilotriacetic Acid Determination Sample 6 Sample 7 Sample 8 Sample 9 Sample 10 Sample 11 Sample 12 Sample 13 Sample 14 Sample 15 QC Standard, 2.0 mg/L (second source) Method Blank Sample 16 These 3 are QCs Sample 16 Lab Duplicate 77  Autosamplers In the modern laboratory, in most methods, samples are rarely run manually, one at a time. Samples are loaded by the analyst onto an autosampler, and then run automatically, one after another, by the instrument. Analyst monitors the progress of the run, collects and assesses data at the end of the run. These can be run overnight as well. 78 79 80  Quality Assurance (QA) QA is the overall system and all of its components to assure quality. QA is done at the management level. QA Manager - person who is responsible for QA system. QA Manager must not be involved in production. QC is a part of QA. 81  QA has a written plan – Quality System Manual (QSM) The QSM defines the quality system. The QSM outlines policies. 82  Laboratory Certification/Accreditation Formal recognition, given by an outside agency, of the competence of a laboratory In Canada: – Canadian Association for Laboratory Accreditation (CALA) – Standards Council of Canada (SCC) In U.S.: – National Environmental Laboratory Accreditation Program (NELAP) 83  Laboratory Accreditation ISO Standard 17025 – “General Requirements for the Competence of Testing and Calibration Laboratories” An international standard You can buy it directly from International Organization for Standardization (ISO) or from SCC ($140) 84  What does a laboratory have to do to get accredited? Accreditation is by parameter and matrix. Performance Evaluation (PE) samples twice every year. Site Assessments once every two years Fees (of course) 85 Standard Operating Procedures (SOPs) Written procedures describing how analyses must be done (Analytical Methods) SOPs specify everything analyst must do – no ambiguity Analysts must follow SOPs exactly. Authorized (signed) Validation data – MDLs, Accuracy, Precision, Linearity Controlled document 86  Control Charts What is a control chart? Why do we use them? How are they made? How are they used? 87  Control Charts A control chart is a graph that shows how a measured result changes over time (days, weeks, and months). It is usually a measurement on a known standard solution. Spiked blank (second source) is usually plotted. Control charts are used to monitor that a process is “in control.” 88  What is a Control Chart ? A typical control chart is a graphical display of a quality characteristic that has been measured or computed from a sample versus the sample number or time. The chart contains: UCL a center line that represents the average value of the quality characteristic corresponding to the in-control state. Mean two other horizontal lines, called the upper control limit(UCL) and and the lower control limit(LCL) are also drawn. LCL Control Chart These control limits are chosen so that if the process is in control, nearly all of the sample points will fall between them. As long as the points plot within the control limits, the process is assumed to be in control, and no action is necessary. 89  What is a Control Chart ? However, a point that plots outside of the control limits is interpreted as evidence that the process is out of UCL control, and investigation and corrective action is required to find and eliminate the assignable Mean causes responsible for this behavior. LCL NOTE: Even if all the points plot inside the control limits, if they behave in a systematic or nonrandom manner, then this is an indication that the process is out of control. 90  Why Do We Use Control Charts? The Use of Control Charts is an Essential Component of Statistical Quality Control (SQC) UCL +3б Mean -3б LCL SQC Is An Expected Quality Element outlined by ISO 17025 91  What is so special about 3 sd and The "68 95 99.7 Rule” A process that is in "control" with a normal distribution has: 68% of the results will fall within 1sd. 95% of the results will fall within 2sd. 99.7% of the results will fall within 3sd. There's ~ three chances in 1,000 (0.3%) that the process will produce a result outside 3sd. 92  How does a Control Chart get started? Before a control chart can be used, it must have the upper and lower control limits (UCL and LCL) established. Where do these limits come from? 93 So How Do we Build A Control Chart ? Step 1: Collect a minimum of 25 data points over a period of at least three weeks. Something 94 Time  Step 2: Calculate the mean of the measured results Calculated Concentration Mean / Average Date of Analysis 95 Step 3 : Calculate and standard deviation (sd) and plot : (mean + 3 standard deviations +3 δ ) and (mean – 3 standard deviations -3 δ ) +3 δ Mean -3 δ 96  How Does a Control Chart Differentiate Between Random and Assignable Causes ? If there is a point outside the limits it may be due to an “intermittent” variation related to an assignable or non-random cause. The Control Chart is saying “Something may be wrong with the measurement. Check it out”. +3 δ Upper Limit Something Out of Control Mean / Point Average Lower -3 δ Limit 97 Sometime  Control Chart Protocols Results from a “second source” standard at a mid-range or lower concentration is obtained each day the instrument is run. Control Charts are updated/plotted and interpreted in “real time”. Out of control situations are addressed immediately (same day) prior to data being reported. Responses to out of control situations are documented. Control charts are audited regularly by QA. 98  Control Chart Protocols For multi-element/compound methods: control charting of a minimum of 10% of the total number of compounds is required. (e.g. Volatiles scan 30 compounds will monitor 3 or greater elements/compounds). 99  Control Charts provide an easily interpreted history of process performance allowing for operators to identify trends and proactively make changes before loss of data due to QC failures. 1 2 Mean 3 1 4 4 5 5 6 6 3 7 7 2 Other trends to be aware of that may signal a process issue include: 7 consecutive points rising or falling and 7 consecutive points on the same side of the mean 100 Common Causes of Violations – Increased Variability Note: Abrupt change in data +3 δ Mean -3 δ Typical Causes Include: variation from SOP poor technique, lack of training etc. 101 Common Causes of Violations – Shift in Mean +3 δ Mean -3 δ Causes Include: incorrect standard or reagent preparation reference material contamination incorrect instrument calibration change in procedure/instrumentation 102 Common Causes of Violations – Downward Trend in Mean +3 δ Mean -3 δ Causes Include: evaporative concentration of calibration material deterioration of reagents 103 Common Causes of Violations: Upward Trend in Mean +3 δ Mean -3 δ Causes Include: deterioration of reference material or reagents evaporative concentration of reference material 104