Chapter 10

QUALITY CONTROL

 

Section 1:  INTRODUCTION TO QUALITY CONTROL

To obtain meaningful data on water quality, the laboratory must first collect a representative sample and deliver it unchanged for analysis.  The analyst must then perform the analysis correctly using the proper method.  Finally, the results of the analyses must be recorded accurately and permanently for future use.  The entire process must be carried out in a consistent, deliberate manner according to an overall plan established by laboratory personnel.  In this chapter, we will discuss how quality assurance and quality control are developed and maintained.

Section 2:  GLOSSARY

Accuracy:  The combination of bias and precision of an analytical procedure which reflects how close the measured value is to the true value.

Bias:  Consistent deviation of measured values from true values caused by systematic errors in a procedure.

Calibration check standard:  The standard used to determine whether or not an instrument is properly calibrated.

Chain of custody:  Procedures and associated documents designed to trace the custody of a sample from the point of origin to final analysis with the intent of legally demonstrating that custody remained intact and that no tampering occurred.

Duplicate:  Performing the analysis of a sample twice to determine reproducibility.  Also, two samples taken at the same time from one location.

Laboratory control standard:  A standard, usually certified by an outside agency, used to measure the bias in a procedure.

Precision:  A measure of the ability to reproduce analytical results.  It is generally determined through the analysis of duplicate samples.

Quality assessment:  The procedure for determining the quality of laboratory measurements using data from internal and external quality control measures.

Quality assurance:  A definitive plan for laboratory operation that specifies the measures used to produce data of known precision and bias.

Quality assurance program:  Written policies and procedures that outline how the laboratory intends to produce data of known and accepted quality.

Quality control:  A set of measures within a sample analysis method to assure that the process is in control.

Section 3:  QUALITY ASSURANCE

Quality assurance (QA) is a set of operating principles that, if strictly followed during sample collection and analysis, will produce data of known and defensible quality.  A quality assurance program must be developed which includes guidelines for lab personnel to follow when collecting and preserving samples, selecting which methods to use for the analyses, and detailed procedures for the actual testing.  The QA program should include training requirements for employees, equipment preventative maintenance procedures, calibration procedures and quality control testing.  Guidelines for data assessment and reporting are also included and may involve the use of charts or tables which document trends and QC sample results.  Lab bench sheets must be maintained that document when the sample was collected, how it was preserved, and what results were obtained.

Section 4:  QUALITY CONTROL

 

Quality Control (QC) includes any testing which is done to prove that the results are reliable.  One of every ten samples analyzed should be a QC check.  This may include duplicate samples, spike samples, reagent blank analyses and known QC samples obtained from outside sources.

For NPDES Lab work, you must (as required by NPDES permit) develop a “Quality Assurance/Quality Control” plan for all parameters on your permit. This requirement includes any contract labs as well as your in house lab. In any event, it is the NPDES permit holders responsibility.

Section 5:  DUPLICATE SAMPLES

Duplicate sample analysis is used to determine reproducibility or consistency in the method.  Duplicates (or split samples) are two samples taken from the same source, digested, distilled or otherwise processed, and then analyzed identically.  The acceptable level or variation between duplicates varies by test and by concentration of analyte.  Table 10-1 show acceptance levels for duplicate samples in water and wastewater.

The percent difference is calculated by dividing the difference between the two results by their average as follows:

% difference = (x - y) / ((x + y)/2)   x 100%

where x = result of first analysis

y = result of second analysis

For example, an effluent sample is tested twice for phosphorous and the results obtained are 0.109 and 0.115 mg/L.

% difference =   (0.115 - 0.109)   /     ((0.115 + 0.109)/2)   x   100%
=   (0.006 / 0.112)   x    100%
= 5%

By checking Table 10-1 we see that a 5% difference for a low level nutrient analysis is acceptable.

Table 10-1

 

Acceptance Limits for Duplicate Samples and Known Additions to Water and Wastewater

                           Recovery of                 Precision of                 Precision of                                    Known Low Level                            High Level
                            Additions                    Duplicates                   Duplicates                            
                                   %                                 %                                 %                              Analysis

Nutrients                 80 - 120                            25                                10
Metals                    80 - 120                            25                                10
Inorganics              80 - 120                            25                                10

Section 6:  SPIKE SAMPLES

 

Spike sample analysis (or known addition) is used to test the method at varying concentrations of analyte.  Known amounts of analyte are added to a sample and the percent recovery is calculated.  If interferences are present in the sample, results may be obtained which are significantly higher or lower than the actual concentration.  The acceptable percent recovery of known additions varies by test and is shown in Table 10-1.

The percent recovery is calculated by dividing the sample result by the expected result as follows:
Expected spike result = sample result + standard conc.

Percent recovery = spike result divided by expected result x spike volume over original volume x 100%

For example, a 100 mL Effluent sample is spiked with 1 mL of 100 mg/L Ammonia nitrogen standard yielding the following results:

Effluent sample result:  0.174 mg/L

Spiked Effluent sample result:  1.21 mg/L

Conc. of 1 mL of 100 mg/L standard in 100 mL:  1.00 mg/L

Expected spike result = sample result + standard conc.

= 0.174 + 1.00

= 1.174

Percent recovery = spike result divided by expected result x spike volume over original volume x 100%

= 1.21 divided by 1.174 x 101/100 x 100%

= 104%

Section 7:  REAGENT BLANKS

Analyze reagent blanks anytime new reagents are used and as often as required by specific methods.  Blanks monitor the purity of reagents, lab distilled water, and the cleanliness of glassware and equipment.  Also, analyze a reagent blank after any sample of high concentration that may cause carryover to the following sample.

Section 8:  EXTERNAL STANDARDS

 

QC samples with known concentrations of analytes can be purchased from a variety of chemical manufacturers.  These should be used to demonstrate that calibration curves are accurate, that the analysis is being properly run, that calibration standards are accurate, and to verify the results of analyses on plant samples.  Results of all QC testing must be recorded.  Sample bench sheets are included in Appendix B.

Section 9:  DATA QUALITY

 

The most important indicators of data quality are its bias and precision which, when combined, express its accuracy.  Only results which have good precision and low bias are accurate.

Bias is a measure of systematic or inherent error.  It may be a result of the method, itself, or the lab’s use or interpretation of the method.  Results that are consistently lower or higher than the true value reflect high bias.

Precision is a measure of the closeness of replicate sample results.  High variability among duplicates reflects low precision.  Consistently close results reflect high precision.  Precision can be determined by calculating the standard deviation of the results.

Accuracy refers to the degree of difference between the measured value and the true value.  Factors affecting accuracy include contaminated glassware, improperly maintained equipment, carelessness in measuring samples or reagents, and failing to follow the method consistently.  Due to increased legislation regulating plant performance and effluent quality, reported values may be used or challenged in court.  Analysts must make every attempt to produce accurate results with low bias and high precision.

Section 10:  SIGNIFICANT FIGURES

To avoid reporting results that are inaccurate or deceiving, “significant figures” are used.  Significant figures give an indication of the reliability of the analytical method used.  Reported values should contain only significant figures.  A value is made up of significant figures when it contains all digits known to be true and one last digit in doubt.  For example, if a value is reported as 18.6 mg/L, the “18” must be a firm value while the “0.6” is somewhat uncertain and may be “7” or “5”.

The number of significant figures for a method is established by testing replicate samples and determining variability.  For example, if an effluent sample is tested 4 times for ammonia and the results are:  0.209, 0.207, 0.211, and 0.206, the average should be reported as 0.21 and not 0.208 since there is variability in the hundredth’s digit.

Section 11:  ROUNDING OFF

 

The rounding off of numbers is a necessary part of analytical reporting.  All instruments, glassware, and methods have limits on significant figures which must be taken into consideration when reporting results.  Numbers that are not significant must be dropped by rounding off.  If the digit 5, 6, 7, 8, or 9 is dropped, round up one unit.  If the digit 0, 1, 2, 3, or 4 is dropped, do not change the preceding digit.  For example, 11.443 is rounded off to 11.44, while 11.447 is rounded off to 11.45.

 

Section 12:  LABORATORY WORK SHEETS

The plant operator has two goals in using a laboratory notebook and work sheets (or bench sheets):  first, to record data, and second, to arrange the data in an orderly manner.  Days of work can be wasted if data are written down on scrap paper which is misplaced or thrown away.  Work sheets and notebooks help prevent error and provide a record of the work performed.  The routine use of work sheets and notebooks is the only way an operator can be sure that all important information for a test is properly recorded.

There is no standard laboratory form.  Most treatment plants usually develop their own work sheets for recording laboratory results and other important plant data.  The work sheets are prepared in a manner that makes it easy to record the data, review it, and recover it when necessary.  These sheets provide the information needed to complete NPDES permit reporting forms.

Each plant may have different needs for collecting and recording data.  At a minimum, work sheets should include the following information:

1.     Date and time of sample collection

2.     Person who collected the sample

3.     Sample preservation

4.     Analyte tested

5.     Raw data

6.     Calculations

7.     Section for recording abnormal conditions

Section 13:  QUALITY CONTROL WORK SHEETS

 

Each laboratory is also required to document proper care of equipment and handling of lab chemicals.  An equipment maintenance sheet should be kept up to date reporting all maintenance performed on lab instruments.  The sheet should include the date, time and technician who performed the work.  Repairs to equipment, changed membranes, cleanings, preventative maintenance checks and observed problems should be recorded.

Every refrigerator, oven, water bath and incubator must have a temperature record indicating on a daily basis both the required temperature and the actual temperature.

A chemical inventory must be maintained that documents date of purchase, expiration date, date of opening, and date when the chemical is used up or discarded.  It is also helpful to record where the chemical was purchased.

A reagent preparation log must be maintained for analysts to record the date of preparation of standards and reagents.  This helps to maintain control over reagent preparation and provides a method to trace the source of error when problems are encountered.

Sample QC bench sheets are included in Appendix B.

 

Answers to Quiz 10.1

1.     Calculate the percent difference between the results; 1.91 and 2.03.

% difference = (x - y) / ((x + y)/2) x 100%
= (2.03 - 1.91)/(2.03 + 1.91)/2 x 100%
= 0.12/1.97  x 100%
= 6%

2.     Why should quality control samples be purchased from outside sources and be tested periodically?

They are used to demonstrate that calibration curves are accurate, that the analysis is being properly run, that calibration standards are accurate, and to verify the results of analyses on plant samples.

3.     At a minimum, what information should be included on lab work sheets?

        1.     Date and time of sample collection

        2.     Person who collected the sample

        3.     Sample preservation

        4.     Analyte tested

        5.     Raw data

        6.     Calculations

        7.     Section for recording abnormal conditions

APPENDIX A

 

References

 

Standard Methods for the Examination of Water and Wastewater, 18th Edition, AWWA, APHA, WPCF; Water Pollution Control Federation, Washington, DC, 1992.

Operation of Wastewater Treatment Plants, Volume II, California State University, Sacramento for the U.S. E.P.A.