Advanced QA/QC for

Wastewater Laboratory Testing

George Bowman
Inorganics Supervisor
sponsored by:

State Laboratory of Hygiene

Rick Mealy
Regional Certification Coordinator
DNR-Laboratory Certification

Disclaimer
Any reference to product or company
names does not constitute endorsement by
the Wisconsin State Laboratory of Hygiene,
the University of Wisconsin, or
the Department of Natural Resources.
 Overview
• QA Manual Kit
• Dealing With Unacceptable LODs
• Spike Calculation Secrets
• Control Limit Tips
• Taking the Correct Corrective Action
• “There’s Something You Should Know
About This Data”
• Documenting Your Documentation

Quality Control is all around us

You often don’t notice it…
but it’s constantly at the root of daily news stories….
…we can see it in the movies or…
…subtley make reference to it in coversation
From the movie Armageddon, listen as Steve Buscemi’s
character makes a witty reference to QC (or lack thereof):
Essential Elements of QA Plan
Schematic Diagram
Sampling Plan (Permit & Process Control)
Sample Handling, Preservation
Analytical Methods
General Lab Quality Control Checks
Quality Control Sample Frequency & Criteria
Corrective Action Plan
Preventive Maintenance Plan
Data Reporting Requirements
List of SOPs

The #1 QA Manual “DON’T”

This is NOT your

QA plan
It’s a tool to help
ACME you develop your
Wastewater Treatment Facility
own QA Plan

ACME
Wastewater Treatment Facility
 Flow Schematic of Beverly Hills 90210 WWTP
__________________
FeCl3
Sample Location and Analyses performed Plant 1 polyelectroly
addition
1. Influent: Flow, BOD, TSS
2. Effluent: BOD, TSS, NH3-N, T. Phos., pH, DO, Cl2 res., Fecal
3. Effluent: Flow Decant
Zone
4 & 5. Aeration Mixed Liquor: Settleability, TSS, VSS, DO

Aeration Basin

Influent Channel
6 & 7. Digestor: Settleability, TSS, VSS, % Solids

Effluent Channel
Page 1
8 & 9. Clarifier: TSS, Sludge blanket Aerobic Final
10 & 11. TSS Digestor Clarifier
12. Sludge Cake: % Solids
13: Digestor: % Solids
14.. Dewatering Supernatant: BOD, TSS, NH3-N,
15. Backwash: BOD, TSS
12 13 Grit Chamber 10 Ferrous
De-Watering
Supernatant chlorid

Grit Chamber 11 additio

36”
Interceptor
14
sewer

Wet Well
Aerobic

Influent Channel
Aeration Basin
Effluent Channel
Digestor Final
comminutors
15 Clarifier
Ter tiary
Filt ers
Cl2 contact
chamber Decant
Aeration
OUTFALL

Chlorine
Zone
Tank
Post

addition
Cl2 contact FeCl3
chamber polyelectrol
Plant 2 addition

Sampling Plan: Permit Monitoring

PERMIT LIMITS
SAMPLE
POINT Weekly Monthly
SAMPLE (Schematic PARAMETERS MONITORING Daily Daily Average Average
LOCATION SAMPLE TYPE Reference) TESTED FREQUENCY Minimum Maximum (mg/L) (mg/L)
Influent Continuous Flow Totalized
1 Daily
Page 2

Influent 24-hr composite

XXflow proportional 1 BOD 3X weekly there should be no limits
time proportional TSS 3X weekly here

Influent Grab

Effluent Continuous Flow Totalized

Outfall 001 2 Daily
Effluent 24-hr composite
Outfall 001 XX flow proportional 2 BOD 3X weekly --- --- 45 30
time proportional TSS 3X weekly --- --- 45 30
(May-October) Ammonia 3X weekly --- --- 3.0 ---
Total Phosphorus 3X weekly --- --- --- 1.0
Acute Toxicity 3X year
Chronic Toxicity 3X year
Effluent Grab
pH Daily 6.0 su 9.0su --- ---
Outfall 001 May-September 2 Chlorine Residual Daily 38 ug/L --- --- ---
May-September Fecal Coliforms weekly --- --- --- 400/100mL
 Sampling Plan: Process Control Monitoring
SAMPLE
POINT
SAMPLE (Schematic PARAMETERS MONITORING
LOCATION SAMPLE TYPE Reference) TESTED FREQUENCY

Aeration Tank Outlet grab 4 & 5 Settleability Daily

TSS Daily
Page 3

VS Daily
Aeration Tank Contents 4 & 5 Dissolved Oxygen Continuous
in-place

Solids
Concentrator
Product-grab 12 & 13 Percent solids As needed

Solids
Concentrator Decant-grab 14 BOD As needed
TSS As needed
Ammonia As needed
Digester ( 30
Contents Grab 6 & 7 Settleability min. ) Daily
%solids Daily
TSS Daily
VSS Daily

Sample Handling
SAMPLE HANDLING TABLE
@
MAXIMUM ANALYTICAL
#
PARAMETER SAMPLE TYPE PRESERVATION CONTAINER HOLDING TIME METHOD
Page 4

BOD 24-hr composite Cool, 4°C Poly 48 hours 5210 B

[flow proportional]
TSS 24-hr composite Cool, 4°C Poly 7 days 2540 D
[flow proportional]
Ammonia 24-hr composite Cool, 4°C; Poly 28 days 4500-NH3 F
[flow proportional] H2SO4 to pH <2
Total Phos 24-hr composite Cool, 4°C; Poly 28 days 4500-P B(5)
[flow proportional] H2SO4 to pH <2 & 4500-P E
pH Grab None Poly Analyze 4500-H+ B
immediately
Chlorine Res. Grab None Poly Analyze 4500-CL G
immediately
Fecal Coliform Grab Cool, 4°C; Poly 6 hours 9222 D
0.008% Na2S2O3
Effluent Toxicity 24-hr composite Cool, 4°C; Poly 28 days subcontract
[flow proportional]

NOTES: Sample Type: (Grab, 24-hour composite [flow or time proportional])

@ From time of completed sampling
(Reference method source) # Standard Methods for the Examination of Water and Wastewater, 19th Ed., 1995
 General Lab Equipment QC
How often should I What am I checking it What if it doesn’t meet
What am I checking? check it? against? specifications?
(Parameter) (Frequency) (Criteria) (Corrective Action)
Sample Refrigerators Daily > 0° C, < 4° C Adjust temp. or
Page 5

Autosamplers Daily > 0° C, < 4° C Adjust temp. or

Balance - 10 mg Weekly 9.7 to 10.3 mg Re-certify weight or
Service balance
Balance - 1.0 g Weekly 0.995 to 1.005 g Re-certify weight or
Service balance
TSS oven Daily 103-105 ° C Adjust temp. or
BOD Incubator Daily 19.0 to 21.0 ° C Adjust temp. or
Desiccators (bowl) Daily Lid lifts easily? Apply silicone grease to rim
TSS Filter screens Daily Pore blockage Clean or replace

BOD-Specific Requirements
Page 1 of 2

Parameter: BOD
Minimum Requirements Facility Requirements
CALIBRATION
Page 6A

Frequency Daily (DO meter) Daily (DO meter)

Calibration levels: 1 1
Evaluation criteria: N/A Set to saturation point
correlation "r" of oxygen in water
Residuals based on temp and pressure
KNOWN STANDARDS Exactly 6 mLs (2%) of a 150 Exactly 6 mLs (2%) of a 150
Composition/True Value mg/L ea. glucose/glutamic acid mg/L ea. glucose/glutamic
Frequency acid 1/20; minimum 1/week Weekly
Evaluation criteria: 198 + 30.5 mg/L 198 + 30.5 mg/L

BLANKS
Frequency Each analysis day (except TSS) Daily
Evaluation criteria: < 0.2 mg/L O2 depletion < 0.2 mg/L O2 depletion
 BOD Requirements
Page 2 of 2
REPLICATES
Frequency 1 per 20 samples per matrix
Influent 1 per 20 samples Every other week (26/yr)
Effluent 1 per 20 samples Every other week (26/yr)
Evaluation criteria:
Influent Range or RPD? 14.5% xxxxxor RPD?
Range
Effluent Range or RPD? 0.75 mg/L Range xxxxxx
or RPD?

SPIKES
Page 6B

Prepared by:
Not Required
Preparation of spikes should not dilute
Adding the sample matrix by more than 10%. mLs
of a Generally, use the same volume of ug/mL standard
to sample in both the spiked and mLs of sample
Final volume= unspiked samples. mLs
Frequency 1 per 20 samples per matrix
Influent 1 per 20 samples
Effluent 1 per 20 samples
Evaluation criteria:
Influent
Effluent
OTHER SPECIFICS
Sample depletion > 2 mg/L residual DO > 1 > 2 mg/L residual DO > 1
Sample pH 6.5 to 7.5 6.5 - 7.5
Residual chlorine Quench if detected Quench if detected
# of dilutions At least 2 At least 3 per sample
Supersaturation Sample DO < saturation
Seed Controls Must treat exactly as samples

Page 7B
Corrective Action
Parameter: BOD
Minimum Requirements Facility Requirements
CALIBRATION
(DO meter) Set to saturation point of
Evaluation criteria: N/A oxygen in water
correlation "r" N/A
Residuals N/A
Re-calibrate if blank DOI > saturation point
Corrective Action If sample DOI > sat. point, bring to room temp & shake

KNOWN STANDARDS
Evaluation criteria: 198 + 30.5 mg/L 198 + 30.5 mg/L
GGA < 167.5: Weak seed (add more)or bad GGA (replace)
Corrective Action GGA > 228.5: Contamination. Identify source and correct
Clean all tubing & glassware
BLANKS
Evaluation criteria: < 0.2 mg/L O2 depletion < 0.2 mg/L O2 depletion
If blanks gain oxygen: Suspect calibration problem
Corrective Action If depletion > 0.2 mg/L Check for contamination. Clean
tubing and still. Obtain new water source.
 BOD Corrective Action-2
REPLICATES
Evaluation criteria:
Influent Range or RPD? 14.5% xxxxx
Range or RPD?
Effluent Range or RPD? 1.3 mg/L xxxxxx
Range or RPD?

Presence of “chunks” in one but not both?….document.

Corrective Action
Qualify data on DMR. Re-evaluate control limits.
SPIKES
Evaluation criteria:
Influent
Effluent

Corrective Action Not Required

OTHER SPECIFICS
Sample depletion > 2 mg/L; residual DO > 1 Do not use result.
Sample pH 6.5 to 7.5 Adjust pH, document, seed.
Residual chlorine Quench if detected Quench, document, seed

Corrective Action
General Concerns
What am I checking it
What am I checking? against? What if it doesn’t meet specifications?
(Parameter) (Criteria) (Corrective Action)
Page 7A

Desiccator seal Adequate? Apply silicon grease to rim

Indicating Drierite Mostly blue Replace/regenerate

TSS Oven Maintains 103-105°C Adjust...repair...replace

(forced air ovens are better)
Balance w/ mg & gm wts Have re-calibrated/ repaired
Class “1” weights Statistical criteria Have re-certified

Thermometers Compare to NIST Apply correction factors

Thermometers for column breaks Replace

Barometer vs. WWW/airport Re-calibrate…replace

If problem continues despite all
efforts, call DNR auditor_________ at ( ___)
 Preventive Maintenance Plan
Each instrument must have a maintenance “history” (record)

pre·ven·tive also pre·ven·ta·tive adj.

Intended or used to prevent or hinder; acting as an obstacle.
Preventing or slowing the course of an illness or disease; prophylactic.

• Clean or replace sample tubing at regular intervals

• Change DO membranes at a specific frequency
• Clean/replace all tubing used in BOD testing regularly
• Regularly scheduled balance maintenance
• Routine verification of thermometer accuracy
• Replace NH3 probe membranes at a specific frequency
• Spectrophotometer: Regularly check λ accuracy
a dilute KMnO4 solution should have peaks at 526 and 546 nm

Page 8
Preventive Maintenance Chart
Preventive Maintenance Procedures
What am I checking? What action am I taking? How often should I do it?
(Equipment/Part) (Action) (Frequency)
Sampler tubing Clean with dilute bleach Every 2 weeks
Sampler tubing Replace Every 6 months
DO membrane Change Every 3 weeks
Electrode filling solution Replace Every week

Corrective Action: Taken to fix a problem

Preventive Maintenance: Taken to prevent Corrective Action
 Lab Equipment Maintenance Log
October 2002

GTB 2.7 5.1 104 20.2 44.6 4,10 735mm 20.8 Refrig adjusted ↓
RGM 1.5 3.8 103 19.8 44.5 4,10 745mm 20.4 Sampler adjusted ↑
GTB 2.3 4.0 67 20.4 44.3 4,10 739mm 19.7 Replaced oven thermometer
RGM 1.5 3.8 103 19.8 44.5 4,10 745mm 20.4
WG 2.2 3.4 108 20.5 44.5 4,10 748mm 21.3 TSS Oven adjusted ↓
WG 2.6 3.7 100 20.3 44.5 4,10 741mm 21.1 TSS Oven adjusted ↑
4 755 mm
4 762 mm
4 770 mm
4 765 mm
4 752 mm
4
4
4
4
4
4

QA Plans - The Bottom Line

Brief - NOT - volumes

Realistic - NOT - marketing "fluff"

Guidance - NOT - Philosophy

Decision trees - NOT - generic options

Reference - NOT - paperweight

Tables - NOT - text

 Detection Limits

LOD vs. LOQ

Limit of Detection (LOD) & Limit of Quantification (LOQ)

LOD
Lowest concentration determined to be significantly
different from a blank
Formerly known as the MDL (Method Detection Limit)

LOQ
Analyte concentration at which one can state with a
specified degree of confidence that an analyte is present
at a specific level in the sample tested.
Defined in code [NR 149.03 (16)] as 10/3 times the
LOD (i.e. 3.33 x LOD)
 Where does the “10/3” come from?
Early MDL theory was based on replicate
measurements (10) of a blank
# of standard deviations [SD] (of 10 measurements)

3X the SD was deemed

10 = LOQ 10X the SD
to be the MDL or LOD 3 LOD deemed to
be the LOQ
Close enough for gov’t work?
The t-value for 7 replicates = 3.143
The t-value for 8 replicates = 2.998 (or 3)

LOD Evaluation: The 5-point check

Spiked at 0.1
(these first 3 are mandatory checks)
LOD= 0.036
1. Is LOD greater than 10% of the spike level? yes
Spiked at 0.1, so LOD should be > 0.01
If LOD < 10% of spike level, re-do at lower spike level

2. Is the spike level greater than the LOD? yes

Common sense: if LOD > spike level, couldn’t detect it

3. Is the LOD below any relevant permit limit? N/A

(if there is one) Permit limit =
 LOD Evaluation: The 5-point check
(additional checks)
Though not specifically required by the EPA method....
these checks help you obtain the best estimate of the LOD.

4. Is the signal-to-noise ratio (S/N) between 2.5 and 10? yes

S/N = Mean/std dev. S/N= 8.69

S/N ratios above 10 suggest that you COULD spike lower

5. Is mean recovery within reasonably expected limits? yes

Mean recovery= mean/spike level x 100 = 98.71%
Expected range is approximately 80 to 120%
Mean recovery above 120 suggests high bias (contamination)
Mean recovery below 80 suggests low bias (less detectable)

What if I cannot meet the

“5-point check”?
Some instruments are simply too precise.
• analyze the replicates over several days (CLP)
• alternatively, intersperse replicates with real samples

If all else fails… …back into it!

a Prepare and analyze a blank spike at a concentration

equal to your calculated LOD/MDL
a Can you quantitate it within 20-30% of expected value?
a If not, repeat this process at a higher concentration until
you achieve a quantitative result (within 20-30% of
target value)
 “Backing into” the LOD
LOD Evaluation: The 5-point check Spiked at 0.5ppm
LOD= 0.009 ppm

1. Is LOD greater than 10% of the spike level? NO!

Spiked at 0.5, so LOD should be > 0.05

STO
P

If LOD < 10% of spike level, re-do at lower spike level

DO NOT prepare LOD replicates: below 0.2 ppm for ammonia

below 0.1 ppm for phosphorus

“Backing into” the LOD

Ammonia Spiked at 0.5ppm LOD= 0.009 ppm

IF... you have repeated the “7 replicates” more than twice

OR... you have reached the lowest recommended spike levels
AND ...you still don’t have a valid LOD...
1. Prepare and analyze a single standard at (or close 0.01 ppm
to) the calculated LOD
2. If you obtain a result within 20-30% of the 0.007 - 0.013
concentration you prepared, then you have
validated your LOD
3. If you DO NOT obtain a value within 20-30% of the
0.000 ppm
prepared concentration then…
4. Prepare another single standard at a concentration 0.02 - 0.05 ppm
slightly higher
0.014 - 0.026 ppm
5. Repeat 2-4 until you obtain a result within 20-30% 0.035 - 0.065 ppm
 Spike
Calculations

Forget What You Know

…it’s just not as simple as

SPIKE - SAMPLE x 100

SPIKE AMOUNT
(although it can be…in some circumstances)

The correct equation...implied, but never said, is:

SPIKE - SAMPLE ( for dilution ) x 100

Corrected

SPIKE AMOUNT ( for dilution )

Corrected
Common Spike Calculation Errors

• Calculating by concentration and

• not accounting for dilution of the sample,
• not accounting for dilution of the spike solution,
• or both.

• Calculating a ratio rather than a recovery.

• Using an incorrect formula.

• The only acceptable formula is
• Spiked Sample - Unspiked sample
Spike Amount

Mass vs. concentration

The to understanding spike calculations is
remembering that you are dealing
with MASS (weight)
rather than CONCENTRATION.
50
With spikes, you are adding a
40 known MASS of analyte.

30 Rather than dealing with the problems of

adding solid material directly, you
20 dissolve a specific salt in reagent water.

10
The volume of reagent water is merely
a vehicle for delivery of the analyte
 Concentration Terms
ppm = parts per
million = 1 part = 1 part
1,000,000 parts 1000 X 1000 parts

ppm = mg/L = milligrams per Liter =

1 mg ⌧ 1 mL x 1 g x 1 L
=
L 1g 1000 mg 1000 mL

ppm = ug/mL = micrograms per milliliter =

= 1 ug x 1 mg x 1g x 1 mL
mL 1000 ug 1000 mg 1g

Concentration basics
Joe analyzes a sample for ammonia.
He takes 50 mLs of sample, adds his buffer solution
After stabilizing, the meter reads 5.0 mg/L.
The concentration value, 5.0 mg/L, means that in a liter of water,
you would find 5.0 mg of ammonia dissolved in it.

But Joe only analyzed 50 mLs of sample; how many mgs of

ammonia were in that 50 mLs?
5.0 mg is the same as 5.0 ug
L mL

5.0 ug 50 mL = 250 ug 1 mg = 0.25 mg

mL 1000 ug

Concentration volume = mass

 A concentration...

X ug × Y mLs
mL ...times a volume

=XY ug = a mass

Diluting to Known Volume

• If you dilute spike (with sample) to a known volume.…
(sample volume used in spike is LESS than that in the unspiked)
sample concentration must be adjusted.

Dilute “to
known volume” Since there is LESS sample-
volume in the SPIKE than there
50 is the SAMPLE, we cannot
mL simply subtract the unspiked
Spike 5 mL sample concentration

Sample Furthermore, because we’ve added

volume from the spike solution, we
45 mL have actually DILUTED the sample
concentration
Corrections: Dilution to Known Volume
You cannot directly subtract background sample
concentration when using the “Dilute to Known
Volume” method.

In this case, rather than

subtracting 2.0 ppm Measured to
Measured to from the spiked sample,
be 2.0 ppm be 6.5 ppm
you must subtract only
45/50 x 2.0 [i.e.,1.8]
Spike 5 mL
to account for the fact
that less sample (and
Sample thus less concentration) Sample
50 mL was in the spike 45 mL

“Adding on top”
• If the spike is added “on top of” the sample
(amount of sample used in spike is SAME as in the unspiked)
the spike concentration must be adjusted.

Analysts are often “tricked” into

thinking that 10 mLs of spike Spiking
solution added to 100 mLs of “on top”
sample represents a 10-fold
dilution of the spiking solution. 50
mL Spike 5 mL

The TRUE spike amount [concentration] Sample

is determined as the ratio of spike
volume to TOTAL sample volume times 50 mL
the spike solution concentration.
 Adding on top - differential effects

If there is no pre-treatment or sample volume

reduction involved (e.g., digestion, distillation):

TWO correction factors are required:

• one for the dilution of sample concentration, and
• another for dilution of spike concentration

If there is a pre-treatment step or sample volume

reduction involved (e.g., digestion, distillation):
No correction factor is required.
Examples: Total phosphorus by hot plate, distilled ammonia

Spiked sample result – Unspiked sample result

Spike Amount
I analyze 50 mLs of sample. It measures 0.1 ppm [ug/mL] ammonia
I take another 50 mLs of sample and add 10 mLs of a 1.0 ppm [ug/mL]
ammonia standard. I analyze this spike and I get 0.2 ppm [ug/mL]

0.0833 58.3
0.2 ppm – 0.1 ppm X 100 = 50% Right?
0.2 ppm
Remember: we are dealing with concentrations here.
If the concentration of unspiked sample was 0.1 ug/mL,
and 50 mLs of sample were used,
then we know 5.0 ug (50 x 0.1) of ammonia came from the sample.

We also know that the total volume of the sample + spike was 60 mLs.
Thus the concentration of the unspiked sample is now 5 ug = 0.0833 ppm
60 mLs
Understanding sample dilution
Here is 50 mLs of sample
with 5 ug of analyte dissolved.
The concentration is 5 ug/50mL
equaling 0.1 ug/mL or 0.1 ppm
If we add 10 more mLs of
reagent water, (the water volume
from a spike)

The concentration of analyte is

now 5 ug per 60 mLs
or 5/60 = 0.08333 ug/ml
= 0.08333 ppm
= 1 ug of
= 1 mL of
sample analyte

Spiked sample result – Unspiked sample result

Spike Amount
I analyze 50 mLs of sample. It measures 0.1 ppm [ug/mL] ammonia
I take another 50 mLs of sample and add 10 mLs of a 1.0 ppm [ug/mL]
ammonia standard. I analyze this spike and get 0.2 ppm [ug/mL]

70.0
0.2 ppm – 0.0833 ppmX 100 = 58.3% Right?
0.2 ppm
0.1667 Yes …70%

Nope! We ALSO have to correct for the dilution of the spike solution.
If 10 mLs of a 1.0 ug/mL spike solution were spiked,
then we know 10 ug of ammonia were added from the spike.
We also know that the total volume of the sample + spike is 60 mLs.
Therefore the concentration of the unspiked sample is 10 ug = 0.1667 ppm
60 mLs
Understanding spike dilution
Take 50 mLs of sample with
ZERO analyte in it (or reagent water)

Add 10 mLs of a spike solution

with a concentration of 1.0
ppm (1.0 ug/mL)

1 ug
1 ug The concentration of the spiked
1 ug sample is NOT 0.2 ppm (10 x 1.0 ppm)
1 ug 50
1 ug 1 ug
There are 10 ug, but
1 ug the total volume is 60 mLs.
1 ug
1 ug 10 ug/60 mLs = 0.1667 ug/mL
1 ug
= 0.1667 ppm
= 1 mL of = 1 mL of
sample 1 ug spike solution

Why phosphorus is different 50 mL

final volume

50 mLs of sample Whether the

plus 5 mLs of 1 final volume
ug/mL spike solution. is 50 mL...
Mixed well.

There’s still only

The sample is 5 ug of analyte
heated and volume
100 mL
reduced to 10-15 final volume
mLs.
to
e
e t
m ilu

The 5 ug of analyte
…or 100
lu d

remain; just in less

vo ou

mL...
al n y

volume of water.
finhe
T
 Spike Recovery Exercise
Calculation of % Recovery
% Recovery = Spiked Sample – Unspiked sample X 100
Amount of spike added
Wastewater Lab operator/analyst “Joe” measures out 50 mLs of
sample, and places the beaker on a stir plate. He then adds 1 mL
of buffer solution. After stabilizing, the meter reads 2.0 mg/L
ammonia.
Unspiked sample 2.0 ug/mL
Unspiked Sample Volume 50 mL
“Joe” then measures out another 50 mLs of sample to prepare a
matrix spike. To the 50 mLs of sample he adds 5 mL of a 25 mg/L
ammonia standard. This beaker is then placed on the stir plate.
He then adds 1 mL of buffer solution. After stabilizing, the meter
reads 4.25 mg/L ammonia.
Spike volume 5 mL Spiked sample 4.25 ug/mL
Spike Conc. 25 ug/mL Total volume 55 mL

What’s the % recovery?

Survey Says...
• about 40% say 90%
• another 25% say 98%
• perhaps 15% say 99%
• less than 10% say 107%,
• we have heard 45% and 53.5%, which
represent a 2X error and answers (A)
and (E)
107%, by the way, is the correct answer.
If you got 90%, you didn't account for dilution of EITHER the sample or the spike
If you got 98% you accounted for dilution of the sample but not for the spike
If you got 99%, you accounted for dilution of the spike, but not for the sample
 Conventional Calculation: “spike “on top”

Matrix Spikes: Ammonia example- adding “on top”

Unspiked sample 2.0 ug/mL Spiked sample 4.25 ug/mL

Unspiked Sample Volume 50 mL Total volume 55 mL

Spike volume 5 mL
Spike Conc. 25 ug/mL

A. Correct the concentration in the unspiked sample = 1.82

2.0 ug/mL X (50/55) mL = 2.0 x 0.91
B. Correct the spike concentration = 2.27
25 ug/mL X (5/55) mL = .091
C. Calculate recovered concentration = 2.43
=(4.25 ug/mL - 1.82 ug/mL
% Recovery = 107.0%
= (C / B) x 100 =(2.43 / 2.27) X 100

Calculating Recovery: A New Way

A. Determine the contribution (ug) from the sample in the spike

1. Subtract the mLs of the spike from the total mLs of sample + spike
2. Multiply the answer from A.1 (above) by the sample concentration

B. Determine the # of ug (of analyte) spiked

Multiply the concentration of the solution used to spike by the # mLs spiked

C. Determine the # of ug (of analyte) in the spiked sample

Multiply the spiked sample concentration by the # mLs of this sample

D. Determine the # of ug (of analyte) recovered

Subtract “A” from “C”

% Recovery = Divide “D” by “B” and multiply by 100

key relationship: mg/L = ppm =ug/mL

 NEW Calculation: “spike “on top”
(use for any analysis that is similar)

Unspiked sample 2.0 ug/mL Spiked sample 4.25 ug/mL

Unspiked Sample Volume 50 mL Total volume 55 mL

Spike volume 5 mL
Spike Conc. 25 ug/mL

A. Contribution (ug) from the sample in the spike = 100

2.0 ug/mL X (55 mL - 5 mL) = 2.0 X 50
B. The # of ug (of analyte) spiked = 125
25 ug/mL X 5 mL
C. The # of ug (of analyte) in the spiked sample = 233.75
4.25 ug/mL X 55 mL
D. The # of ug (of analyte) recovered = 133.75
= C - A = 233.75 - 100
% Recovery = 107.0%
= D / B =(133.75 / 125) X 100

Unspiked sample .246 X 25 = 6.15ug/mL Spiked sample .346x 25= 8.65 ug/mL
Unspiked sample Volume 2mL=>50 mL Total volume 2 mL + 1 mL =>50 mL

Spike volume 1 mL
Spike Conc. 5 ug/mL

A. Contribution (ug) from the sample in the spike = _____

______ ug/mL X (_____ mL - _____ mL) = ______ X ____
B. The # of ug (of analyte) spiked = _____
______ ug/mL X ______ mL
C. The # of ug (of analyte) in the spiked sample = _____
_____ ug/mL X _____ mL
D. The # of ug (of analyte) recovered = _____
______ - ______
% Recovery = ____%
(______ / ______) X 100
Example: Phosphorus- hotplate
Control Limit Reminders

Calculating Control Limits

Matrix spike & RPD Control limits
1. Test the data for and eliminate outliers before proceeding.
2. Calculate the mean and standard deviation of the data.
3. Warning limits = Mean + 2 standard deviations
4. Control limits = Mean + 3 standard deviations
NOTE: RPD is a 1-tailed test, so only Mean +

Range Control limits

1. Test the data for and eliminate outliers before proceeding.
2. Calculate the mean of the data.
3. Warning limits = 2.51 x Mean
4. Control limits = 3.27 x Mean
 What if I don’t have enough data?
If you do QC any less frequently than once
every 2 weeks, you will not have enough data. NR 149.14 (3) (g)

For labs with

• less than 20 quality control results
• within 12 months,
the laboratory may set quality control limits based on
y information given in the authoritative sources,
Standard Methods 18th ed., Table 1020I
• Spikes: 80-120%
• RPD (high concentration): + 10%
• RPD (low concentration): + 25%
y laboratory experience,
Be prepared to defend what you’ve come up with!
y or the experience of other laboratories.
Could use limits from an adjacent facility. Only do this if the
facility has similar processes to your own and they are using
the same procedure as you are.

The Auditor says your limits are too broad

How do they know that?

• Remember: RPD should be about 20% or less
• 20% RPD means a range that is 1/5 of the mean
• Therefore, your upper control limit should be
no more than 20% of your sample mean
• Effluent Range control limit should be no more than 20%
of average effluent concentration
• Example…Effluent BOD averages 7 ppm Your replicate control limit
should not be greater than about 0.2 x 7 = 1.4 ppm

• Influent Range control limit should also be no more than

20% of average influent concentration
 What if I think my control limits are TOO tight?

This is every lab auditor’s dream!

1. Be sure that you are not excluding out-of-control data!

Outliers should be excluded, but
all out-of-control points are not outliers.

2. Include enough significant figures.

• If values are whole numbers (e.g., 10, 89%), you can use
one or even two decimal points to include some variability
(10.2, 89.3%).
• This same problem has been observed when a laboratory
only reported recoveries to the nearest 5%.

Data Reporting vs. Control Limits

Date Sample Replicate Range Sample Replicate Range
1 10 10 0 9.9 9.6 0.3
2 5 4 1 4.6 3.5 1.1
3 10 10 0 9.6 9.6 0 25% increase in control limit
4 10 10 0 9.5 9.9 0.4 • 19 of 20 runs: control limits
5 8 8 0 7.7 7.9 0.2
6 11 10 1 10.6 9.9 0.7 higher without rounding
7 8 7 1 7.5 6.6 0.9 • Sample range 4-10 ppm
8
9
11
6
11
6
0
0
11=10.6 - 11.4
10.7
5.8
10.7
5.5
0
0.3
• 1 ppm max range
• correlation 0.922
10
11
8
10
8
10
0
0
8= 7.5 - 8.5
7.6
9.6
7.7
9.6
0.1
0 • average percent increase
12 11 11 0 10.6 10.8 0.2 was 18%
13 11 10 1 10.6 9.8 0.8
14 9 8 1 8.7 7.6 1.1 • Range was -1% to +44%
15 6 6 0 5.6 5.7 0.1
16 8 8 0 7.6 7.9 0.3
17 11 10 1 10.8 9.9 0.9
Control Limits vs. Data Reporting
18 9 8 1 8.8 7.5 1.3
2.5
19 5 5 0 4.8 4.9 0.1
Reporting 1 decimal

2
20 11 10 1 10.9 9.6 1.3 1.5
place

Sum 178 170 8 172 164 10 0.5

0
Mean 9 9 0.40 9 8 0.51 0 0.5 1 1.5 2 2.5

Warning 2.51 X 1.004 2.51 X 0.505 = 1.26755 Reporting whole numbers

Control Limit 3.27 X 1.308 3.27 X 0.505 = 1.65135

Control limit should be no > 1.74 1.68

 DO NOT plot the data used to create limits
DATE %R
11/01 81
against those same limits!
11/04 87
11/07 90
130
11/10 93
11/13 80 120
11/16 82
11/19 91 110
11/22 94
11/25 83 100
11/28 110
12/01 91 90
12/04 92
12/07 80 80
12/10 88
70
12/13 94
12/16 92 60
12/19 83
12/22 80 50
11/01 11/07 11/13 11/19 11/25 12/01 12/07 12/13 12/19 12/25
12/25 91
12/28 95

N = 20 SD = 7.28 Mean = 88.85 Limits = 67 to 110.7%

Dealing With Outlier Data

There are many statistical tests available for Critical
identifying outliers. One that is relatively easy to N Z
18 2.65
use is the Grubbs test. 19 2.68
20 2.71
Z = mean - questionable data point 21 2.73
22 2.76
SD 23 2.78
24 2.80
Ignore the sign of the “Z” value….is always “ + ” 25 2.82
26 2.84
For replicates, test only the highest value 27 2.86
For spikes, test both the lowest & highest values 28 2.88
29 2.89
Include suspect outlier when calculating mean, SD 30 2.91
If the calculated Z-value > Critical Z value 35 2.98
40 3.04
...for that number of data points, 50 3.13
...then the value is an outlier 60 3.20
 Outlier Test Example?
DATE %R
11/01 81 1. Calculate the mean and SD
11/04 87
11/07 90
Mean = 88.85 SD = 7.278 Limits =67 - 110.7
11/10 93
11/13 80 2. Test the high value (110)
11/16 82
Critical
11/19 91 Z= 110 - 88.85 = 2.9058 N Z
11/22 94
11/25 83
7.278 18 2.65
11/28 110 19 2.68
3. Test the low value (80) 20 2.71
12/01 91
12/04 92 Z= 88.85 - 80 = 1.2159
12/07 80
12/10 88
7.278
12/13 94
12/16 92 4. Discard outliers; re-calculate mean and SD
12/19 83
Since Z110 > criterion, 110 is an outlier
12/22 80
12/25 91 Mean =87.737 SD= 5.4553 Limits = 71 - 104
12/28 95
NOTE: Step 4 may also require a re-check for additional outliers!

Corrective Action
 What IS Corrective Action?

³In a nutshell, Corrective Action is anything

done in response to an out-of-control situation.

³It MUST, however, be designed to identify the

reason for the failure, and then correct it.

³There should also be a plan to quickly verify

that the action taken has the desired effect.

What IS NOT Corrective Action?

Inaction
A limited repertoire of fixes
Addressing the effect vs. the cause
Failure to ensure resolution

Writing “I don’t know why it failed”

Simply considering all matrix spike failures as
“matrix effects”
Checking calculations only
Simply re-running the samples
ACME Labs’ Corrective Action Flowchart

YES DOES IT
DOES IT NO
FUNCTION?
FUNCTION?

DON’T CHANGE IT
YES DID YOU
DID YOU
TRY TO FIX
TRY TO FIX IT?
IT?

DOH !
NO

DOES
ANYBODY YES YES
ARE YOU
KNOW ABOUT UH OH! GOING TO BE IN
IT? TROUBLE?

NO NO
CAN YOU
NO
HIDE IT BLAME SOMEONE
ELSE? PRETEND YOU DON’T
KNOW ABOUT IT.

YES
THEN, THERE IS NO
PROBLEM.
PROBLEM.

You need more than band-aids...

 Getting past the Symptoms;
Determining & Treating the Illness
If you’re not prone to headaches, and suddenly develop a
major migraine...

…how would you feel if the doctor

simply prescribed ...?
Aspirin
Sure… it might work, but as the one e
n
with the problem, YOU don’t want to p hi
waste time with “maybe” solutions or
m

The first step in Corrective Action is to identify

the cause of the problem at hand

Corrective Action =
Continuous Quality Improvement
Ac an
t Pl

k Do
ec
Ch
 Plan
Really think about potential causes
Two heads are better than one…consult others
Maybe check corrective action logs
(we’ll discuss those later)
Identify a plan of action to
address the problem
Collect data or observations
Establish a timetable for
follow-up

Do
Implement the plan--on a small scale
Try to incorporate routine variables
Don’t commit to wholesale changes at
this stage
Most important…do not assume your
plan is the ultimate solution…or even a
viable one
 Think you have it covered?
You've carefully thought out all the angles.

You've done it a thousand times.

It comes naturally to you.

You know what you're doing, its what

you've been trained to do your whole life.

Nothing could possibly go wrong, right ?

Think Again
 Check
Time to evaluate results
Did the action correct the problem?
Did it work under most routine situations?
What did you learn?

Be sure to document
each and every step
along the way

Act

Decision Time!

If your plan seems to have corrected things,

consider full-scale implementation

Alternatively, consider other options

 Corrective Action Form

Corrective Action Form -2

 Corrective Action Form -3

Creating a Corrective Action Plan

Situation Corrective Action
1) Was initial calibration done properly?
BOD: GGA 2) Change in seed source?
failing high 3) Possibility of nitrification?
4) Qualify data on DMR back to last good GGA.

1) Check that membrane is intact; no bubbles.

NH3 electrode
2) Make sure fresh filling solution is used.
slope < -54 mV
3) Is the electrode stablizing normally? too slow?
4) Is the intercept climbing above the LOD?

Phosphorus 1) View plot...does a single standard look funny?

calibration.... 2) Beyond linear range? (about 1 ppm for most)
“r” is <<<0.995 3) Contamination..especially at low level?
 Corrective Action?

Or Taking the
Correct Action?

Corrective Action means

changing roles
Time to STOP being the scientist….
 Summary: Corrective Action

y Identify the source of the problem

y Look beyond the effects…find the cause
y Develop a Game Plan to address the problem
y Implement the plan on a trial basis (Do)
y Evaluate the results of the change (Check)
y Decide whether or not the problem is solved (Act)
y Develop a documentation protocol
(trail of bread crumbs)
y Qualify any affected data

Qualifying Data
DMR: Laboratory QC Comments Box

This box is reserved for comments

SPECIFICALLY related to
laboratory QA/QC problems

Report any Quality Control exceedances here

Very important in assessing data quality

Laboratory QA/QC Comments Box

Failure to report QC exceedances here is a
weakness we are seeing during audits

Historically, reporting anything here has been

perceived as a “black mark” against the facility

Time to change history!

(1) You are required to report this information
(2) If engineers do NOT see information here,
we cannot assist you in resolving
laboratory problems
 The Truth About Qualified Data
Qualified data are NOT necessarily “bad” data.

ified
Qual
Data

ALL this means is that the data user (i.e., the

DNR) needs to take into consideration the
nature of the situation surrounding the
qualification when interpreting the results.

Checking with Mr. Webster

 Qualifying Data - Considerations
• Make sure your comments are meaningful
and understandable to the end user.
• Provide enough information so that the DNR
can assess the data quality.
• Remember to include qualifiers from
subcontract labs!
• Attach comments on a separate sheet if
necessary. Write “see attachment” in the
QC Comments box.

Qualifying Data - Reference Sample Failures

NR 149.13 (4) PROCEDURE FOR CORRECTING

UNACCEPTABLE REFERENCE SAMPLE RESULTS.

(a) All test categories, except category 18– safe drinking

water tests. After 2 consecutive reference sample failures
the laboratory shall…

2. Qualify all test results of the analytes in the test or test

categories which the laboratory has failed to meet
acceptance limits on 2 consecutive reference samples
 Example - Reference Samples

Situation: You have failed your BOD reference

sample for the 2nd consecutive round of testing.
Resolution:
1. Order a 3rd reference sample ASAP
2. Identify and correct the problems
3. Pass the remedial reference sample!
4. Qualify any BOD results on the DMR until you
pass a reference sample

Qualifying Data - QC Failures

NR 149.14 (3)(h) If the results of
• known standards,
• spiked samples,
• method blanks, or
• replicates
exceed the quality control limits,
corrective action shall be taken by the laboratory.
The laboratory shall
• reanalyze the affected samples or
• qualify the results back to the last acceptable
quality control check of the same type
unless the laboratory determines that sample
results are unaffected.
 Qualifying Data - the “HOWs”
Code definition...
NR 149.04 (21m) “Qualify” means to place a written statement
accompanying the test results which identifies anomalies
encountered in generating the data.

Reference Sample failures (2 consecutive)...

NR 149.13 (4)(a) 2. Laboratories shall qualify test results by
placing a statement in their analytical report [i.e. the DMR] stating
that the laboratory has failed 2 consecutive reference samples for
this analyte or analyte group.

QC Exceedances...
NR 149.14 (3)(h) The results are qualified by reporting that the
laboratory analysis was not within the acceptance limits for this test.

QC Examples - Blanks
Situation: Your BOD blank depletions have
been unacceptable for the past week. You traced
the problem to a new bottle of “Cowboy Bob’s”
distilled water.

BOD blank failed.

5/10/01 to 5/17/01 - BOD blank depleted

more than is allowed (0.2 mg/L).
Blank depletions ranged 0.6 to 1.1 mg/L.
Traced to new bottle of water.
 QC Examples - Known Standard
Situation: Your BOD glucose-glutamic acid
(GGA) exceeded acceptance criteria. You used a
new lot of GGA standard the next day and
results were fine.

GGA exceeded acceptance criteria.

5/7/01 - GGA analyzed this day (235

mg/L) exceeded criteria (198 + 30.5).
Repeated GGA with new lot on 5/12/01.
Result was 202 mg/L.

QC Examples - Replicates
Situation: Your influent TSS replicate on
5/17/01 exceeded upper control limit.

Replicate failed for TSS.

5/17/01 - Replicate result (5.5 mg/L) for TSS

on influent exceeded upper control limit (1.9
mg/L). Replicates are done weekly, so data
since 5/10/01 are affected. Heavy rains caused
TSS levels to be 3 times typical levels. Did
another replicate next day and it passed.
 QC Examples - Spikes
Situation: Your phosphorus effluent spike on
5/17/01 exceeded control limits.

Phosphorus spike exceeded control limit

5/17/01 - Spike for phosphorus on final effluent (

35%) exceeded criteria (79-128%).
Final is spiked every two weeks, so data back to
5/3/01 is affected. High phosphorus this day
(1.2 mg/L) and the spike amount was too low
(0.1 mg/L). I raised the spike amount to 0.5 mg/L,
made up a spike the next day and it passed.

Qualifying Data - Final Words

• There is a significant level of QC required in
testing, and thus - statistically speaking- you
are going to exceed something each month.
• Even a lab doing only BOD and TSS 3x/week
can generate up to 24 QC samples/ month.
BOD blanks 1/d x 3d/wk x 4 weeks =12
• Add
BOD in ammonia
known std&(GGA)
phosphorus,
1/wk and
x 4the
weeks = 4
number increases to more than 75/month
BOD eff. replicate 1/2wk x 4 weeks = 2
BOD inf. replicate 1/2wk x 4 weeks = 2
• Consequently, it’s almost an expectation that
TSS eff. replicate 1/2wk
something will be qualified each xmonth.
4 weeks = 2
TSS inf. replicate 1/2wk x 4 weeks = 2
• With qualifiers, “less” is not more.
Documentation

Operating Principles

If you didn’t document

it, you didn’t do it!

If you did the work…

take credit for it!
 Simple Approach to Consistent
General convention:
Reagent Documentation
MM/YY means the
FIRST day
Pre-print labels for new chemicals and reagents
of the month

New Chemical or Reagent Label

Date received: 4/2/01 Date Expires: 6/03

Received by: J. Smith Date Opened: 4/12/01
Required Storage: Room temperature, away from light
Chemical or Reagent: Ascorbic Acid
New Working Standard Label
Standard:Phosphorus, 5 μg/mL Std. Code: 100-3
Date Prepared: 8/5/01 Prepared by: A. Smith
Date Expires: 9/5/01 Storage: 4 oC
Stock Std Code: 50-25

Reagent and Standard

Control Records

• Full traceability of reagents and standards to

the original lot in a logbook
Remember….traceability is in the “eye of the beholder”

• Traceable record of standards and regents used

directly on the analysis record (bench sheet)

• Person preparing the reagents or standards,

preparation and expiration dates on all reagents
and standards
 Reagent and Standard Control
Records Continued

• Storage conditions
• Certification records provided by the
manufacturer with a direct link to the
calibration standards

Remember to record the standard code on the

certificate so there is a link to the standards logbook!

The Not-so-obvious Things that

Need to be Documented
Items that are often overlooked!
Jefferson
• Corrective actions taken City

Remember…you need to be able to show auditors

or courts what was done ….not just tell them!
• Historical QC limits
You may need to defend data 3-4 years old. Could you
tell the courts what the QC limits were when the tests
were done?

• Performance on blind samples and

reference samples
 Paperless laboratories - Reality
or Pipe Dream?
Free Lunch

….and no
such thing as a paperless laboratory!!!

Automating and using electronic recording

keeping process just means….

less Paper less

Recommended Practices when using

Electronic Record-Keeping Processes

• Automated audit trail when ever records are changed

• Limit records access to a few authorized individuals

e.g., Systems Administrator, Lab Director, etc.

• Implement a process to document any changes by the

systems administrator

• Back-up data daily and use a media that will allow

retrieval years later
e.g., Optical storage has a longer life than magnetic media
 Matrix Spike Preparation details
- often overlooked!
What the code says: [NR 149.06 (1)(intro.)]
Records to be retained include but are not limited to ... the following:
(b) Quality control data for spikes, replicates, method blanks, blind standards
reference samples, calibration standards and known standards.
Quality control results shall be traceable to all of the associated sample
results.

What it means (as it relates to spikes):

An auditor must be able to verify spike concentration, which means
• Concentration of the solution used to prepare spikes
• Information necessary to show that spike solution had not expired.
• The volume of spike solution used
• The volume of sample used
• The final volume of sample + spike
• The sample that was used to prepare the spike

Summary
Formulated a gameplan [your QA manual]
Fine-tuned your detection limits
Reviewed spike calculations
Reviewed control limits and outliers
Reviewed PROPER corrective action
Discussed what qualifying data means
Discussed required documentation
Questions?
 George Bowman Rick Mealy
(608) 224-6279 (608) 264-6006
---------------------- -----------------------
Bill Sonzogni, Director
State Laboratory of Hygiene
Wisconsin DNR
2601 Agriculture Drive
PO Box 7921
Madison, WI 53718
Madison, WI 53707
[email protected]
[email protected]

State Lab web address:

http://www.slh.wisc.edu/outreach/
DNR’s LabCert homepage:
http://www.dnr.state.wi.us/org/es/science/lc/