challenges in maintaining drinking water quality at the
TRANSCRIPT
CHALLENGES IN MAINTAININGDRINKING WATER QUALITY AT THE TAP:
CONTAMINATION WITH TOXIC LEAD
Simoni Triantafyllidou
University of ArkansasSeptember 14, 2017
1
A little bit about me
• Bachelor’s in Environmental Technical Engineering University of
Crete, Greece
• M.S. in Environmental Engineering• Ph.D. in Civil Engineering
• Environmental Engineer at USEPA’s Office of Research andDevelopment
Outline • Motivation
- Overview of drinking water distribution challenges in the US
• Research Project Examples - Lead (Pb) in water and children’s blood before Flint, MI - Galvanic corrosion after partial lead service line
replacements - Lead in drinking water of US schools and biokinetic
modeling of children’s blood lead levels - In-building disinfection and unintended consequences
3
Paradigm Shift in Drinking Water Quality
Safe Water Safe Water
Tasty Water
+
+
+Sanitary Water Sanitary Water Sanitary Water
Sanitary Water Safe Water Tasty Water
•Waterborne Disease •DBPs/VOCs • Satisfying Consumer
•Disinfection •More Strict Criteria • Aesthetics (Taste and Odor)
•Water Quantity •Water Quality • Distribution System
Korean Ministry of Environment, modified by Jo, 2006 4
Drinking Water Quality AFTER treatment plant
Water Sanitary Water Source Water Treatment Plant Safe Water
Premise Plumbing
niot m
u erib
tsyt SisD
Tasty Water
Sanitary Water Safe Water
Tasty Water ? 5
Aging Main Distribution Systems in the US
http://www.infrastructurereportcard.org
•Many DS reach or have exceededtheir design lifetime
Clogged Iron Pipe due to corrosion http://www.wrb.ri.gov
Water Main Break NACE, 2010
• Public health implications, resource and financial Implications6
Premise plumbing challenges
Every building is a dead-end • Variety of reactive pipe materials that interact with disinfectant and bacteria
-PVC, PEX, Galvanized, Copper, Brass, Solder, Old Lead
• Variety of plumbing configurations, installation practices (good/bad), and maintenance (good/bad)
• Water use patterns affect Water Age - Flow: Continuous Turbulent Long Stagnation - Temperature, Redox Potential, pH, Disinfectant Residual: Highly Variable
- Microbes: Quantifiable diversity 7modified from Marc Edwards
Premise plumbing challenges
Chemistry of water affects end water quality
• All waters are different in terms of corrosivity and microbial re-growth potential, due to pH?
NOM? Alkalinity?
Cor. Inhibitor?
1) Source water quality 2) Water treatment steps 3) Interaction with distribution system before building
• Water that is “aggressive” for corrosion or microbial growth for certain plumbing materials/configurationsmight be “harmless” to next door plumbing - Variability from building to building - Variability from tap to tap (hot spots) - Variability between hot and cold water from same tap
8
Protect water and public health from plumbing…
Blue Water Due to Copper
Rashes from shower/bathing
Red Water Due to Iron (lead also present)
Blood lead poisoning
…Then protect plumbing from water
Leak in Copper Pipe due to “corrosive” water Mold due to Leaks
Expensive Repairs due to Leaks
9modified from Marc Edwards
Research Interests
Drinking Water
Treatment
Inorganic Aquatic
Chemistry
Corrosion Science
Risk Assessment
Sustainable Drinking
Water Infrastructure
10
Useful research tools
Morphology and Visual & XRF elemental mapping of Tap water collection plumbing particles in faucet aerator inspection through SEM/EDS
Analysis of water • pH, temp, chlorine, TDS • Metals (ICP-MS) • Chloride, sulfate, anions (IC)
Excavated lead pipe Analysis of scale mineralogy Lead scale harvesting with XRD, SEM/EDS
11
Useful research tools
Chemical equilibrium modeling
(e.g., Mineql+)
Blood lead modeling for children (e.g., IEUBK)
12
Much of US water safe, but problems remain
Pb not present in drinking water right after treatment:
BUILDING
– Old Lead Pipe Triantafyllidou and Edwards, 2012
– Old Leaded Solder– Leaded Brass (valves, fittings, faucets, water fountains) 13
Washington DC “Lead-in-Water Crisis”
0 10 20 30 40 50 60 70 80 90
100
90%
'ile
lead
in w
ater
(pp
b)
15 ppb EPA AL
Chloramine Disinfectant 2000
Media Coverage Orthophosphate Corrosion Inhibitor, 2004
Public Health Intervention 2004
1999 2000 2001a 2001b 2002 2003 2004 2005 2006 2007 Year
Environmental Science & Technology, 2009 14
Low
Switch from chlorine to chloramine disinfectant dissolved lead from pipe scales
Blood Lead Level (BLL) Database
Birth Date ZIPCODE Collect Date BLL
1/27/1997 20011 1/27/1999 <3
6/16/1997 20010 9/7/1999 9
10/19/1998 20011 9/7/2000 5
. . . .
. . . .
. . . . 12/24/2005 20011 3/19/2007 <1.0
3/18/2005 20011 8/30/2007 12
N > 28,000 16
in High Pb in Water
2001
Compare Median BLL / All Children / City-wide
Low Pb Water
Low Pb in Water
1999 2003 2005
0.0
0.5
1.0
1.5
2.0
Year
log10
(BLL
)
No Increase in Median BLL
2007
17
Same Data Revisited
• Compare incidence of elevated blood lead (Moore et al., 1977)
% children with BLL > 10 ug/dL
• Perform “neighborhood analysis” (Brown et al., 2001)
Compare “High” vs. “Low” Risk neighborhoods, based on:
1) prevalence of lead pipe
2) elevated lead in water
• Focus on sensitive sub-group (WHO, 2000) Children ≤ 30 months of age
18
Increase in Incidence of
Elevated blood lead
at High Risk
High Pb in Water
Compare % EBL / Zip Code Level / Children ≤ 30 months
0
1
2
3
4
5
6
2000 2002 2004 2006
% C
hild
ren
with
El
evat
ed B
lood
Lea
d
Low Risk Neighborhoods
High Risk Neighborhoods
U.S. Trendline
High Pb in Water
3 X
Year Environmental Science & Technology, 2009 19
Results published in Washington Post ES and T in 2009 followed up
http://www.washingtonpost.com/wp-Best paper of 2009 Award dyn/content/graphic/2009/01/27/GR
Science category, Editor's choice 2009012700721.html 20
Conclusion
• Elevated lead in tap water can contribute or even cause elevated lead in blood of children, in cases of sub-optimal corrosion control at the presence of leaded plumbing
21
Galvanic Corrosion after Simulated Small-Scale Partial Lead Service Line Replacements
22
Partial Lead Service Line Replacement (PLSLR) • 3.3 - 6.4 million US
homes with old lead service lines or connections (Weston and EES, 1990)
• Contribute to 50 – 75 % of the lead in drinking water (Sandvig et al., 2008)
• Partial replacementwith copper pipe mandatory remedial measure to meet water lead regulation
Journal AWWA, 201123
r • Electron flow
Water
OH.
CATHODE Cu Protected
pH 1'
REDUCTION 02+4e·+2H20-)40H-
Galvanic Corrosion
• Electrochemical (galvanic) cell between copper and lead pipe
• Drinking water serves as electrolyte
• Galvanic corrosion may accelerate corrosion of the lead pipe
24
Water chemistry can turn on/off galvanic corrosion: CSMR
English studies first introduced the CSMR as a factor controlling galvanic corrosion in connections of lead solder/copper
Oliphant (1983) and Gregory (1985)
Example calculation:
Chloride to Sulfate Mass Ratio( CSMR) = [Cl - ] 12 mg/L Cl - = = 0.6 -2 -2[SO 4 ] 20 mg/L SO4
0% Cu (i.e. 100% Pb) .. (j tt QCP
50°ToCu
,
- --
83% Cu
· 100% Cu 26
Experimental Results 0% Cu 17% Cu 50% Cu 67% Cu 83% Cu 100% Cu
0
5,000
10,000
15,000
20,000
25,000
30,000
35,000
Pb in
Wat
er (p
pb)
High CSMR Wires Connected
Low CSMR Wires Connected
High CSMR Wires Dis-Connected
High CSMR Wires Re-Connected
Time (Weeks)
Experimental Results
Pb in
wat
er (p
pb) 25000
20000
15000
10000
5000
0
30000
+8x
+10x
+12x
+14x
0 17 50 67 83 100
Experimental, High CMSR (Galvanic or Deposition Corrosion)
% Pb pipe Replaced with Cu
28Journal AWWA, 2011
Conclusions
• Galvanic connections between copper pipeand lead pipe worsened lead release,compared to lead pipe alone
• High CSMR water released much more lead tothe water than did low CSMR: Water chemistry affects the galvanic battery
• High CSMR produced high sustained galvaniccurrents between lead and copper (notpresented here)
Biokinetic modeling to predict blood lead levels from water lead exposure
Water Air Diet Dust Soil Other
Lung GI Tract
Environmental Lead (Pb) Concentration
1. E
XPO
SURE
2. U
PTAK
E
Lung GI Tract
Plasma Extra-Cellular Fluid
Plasma Extra-Cellular Fluid
Red Blood Cells
Exhaled Air
Feces
Hair, sweat, nails
Urine
3. B
IOKI
NET
ICS
Kidney, Liver, Other soft
tissues
Two Bone Compartment
Predict Blood Lead Level(BLL) in children (Treated as a Geometric Mean- GM)
4. PROBABILITY DISTRIBUTION
Predict % of Exposed Children with Elevated Blood Lead (i.e., BLL > Safety
Threshold)
BLL (ug/L)
Safety Threshold GM
Log-normal distribution around GM with fixed GSD
Prob
abili
ty D
ensi
ty
30
Two school districts with water lead problems
31
Elementary Schools Sampled
Seattle Public Schools (SPS)
63 (~3,100 taps)
Los Angeles Unified School District (LAUSD)
601 (~51,000 taps)
Dates Pre: 2004, Post: 2011-12 2008-2009
Range of Lead Detected, µg/L
<1 - 1,600 first-draw <1 – 370 flushed
0.2 - 13,000 first-draw 0.2 - 7,400 flushed
% school taps > 20 µg/L pre
19% of first-draw 3% of flushed (30 sec)
6% of first-draw 1% of flushed (30 sec)
pre / post RA? Yes / No No / No
Model Input: Combined WLL for one Seattle School, pre-remediation
32
Cum
ulat
ive
Perc
entil
e
100 208 g/L 291 g/L 781 g/L
90
80 39 g/L 45 g/L 160 g/L
70
60
50
40 6 g/L 16 g/L 59 g/L
30
20
10
0 1 10 100
First draw Second draw Combined
1000
Water Lead Level (g/L)
Probability Distributions of BLL, 10 ug/dL threshold
0.30 BLL
50%ile WLL WLL (GM, % Elevated Pb % 'ile (µg/L) µg/dL) EBL (%) 0.25
Prob
abili
ty D
ensi
ty
0.20
0.15
CDC level of concern
90%ile WLL
50 90 99
16 45 208
3.3 5.5
15.3
1.0 10.2 81.6
0.10
0.05
99%ile WLL
0.00 0 10 20 30 40
Blood Lead Level (g/dL)
33
Predicted percentage of children with elevated blood lead
Combined WLL (g/L)
Percentile of Combined WLL
50 60 70 80 90 100
Perc
entil
e of
BLL
Exc
eeda
nce
at G
iven
WLL
0
20
40
60
80
100 BLL threshold: 5 g/dL BLL threshold: 10 g/dL
16 25 45 2643317 In
tegr
ated
Per
cent
of E
xpos
edPo
pula
tion
Exc
eedi
ng
50
40
30
20
10
0
25.5% of children exceed BLL threshold of 5 g/dL
4.6% of children exceed BLL threshold of 10 g/dL
0 2 4 6 8 10 12 14 16
Science of the Total Environment, 2014 BLL Threshold (g/dL)
34
Conclusions
• Variability among fountains within a school
• Variability in water lead contamination among schools receiving the same water
• Accounting for variability in water lead levels, variability in children's response and stringent new public health goals predicted blood lead elevations for most sensitive or most exposed children to water lead
Risk?
x
x
35
Hospitals also deserve increased attention
• A 2011 outbreak of hospital-acquired pneumonia in Pittsburg, from waterborne Legionella bacteria, caused
- Several fatalities and lawsuits - Congressional investigation - Extensive press coverage and criticism - Closer look at microorganisms in hospital water
http://www.cnn.com/2012/12/13/health/legionnaires-hospital-water/ 36
In-building disinfection Thermal disinfection Example: ASHRAE Guideline 12-2000
• Water always stored at > 60°C in water heater > 51°C in hot water lines
• Different instructions after outbreaks or for periodic thermal disinfection
Chemical Disinfection • Chlorine • Copper-silver ionization • Chloramine • UV irradiation • Chlorine dioxide • Ozone
Copper-Silver Ionization (CSI) is one option
Flow cells
Controllers
Inside a “Fresh” Flow cell
Haensel, 2012
Good Maintenance Needed
/Silver
Copper/
• Adds copper ions (Cu+2) and silver ions (Ag+) to water biocides • Only a fraction of copper and silver will remain in free ionic form depending on water chemistry
38
Large Hospital in Cincinnati
CSI • Treated surface water - pH 8.6 - Alkalinity
75 mg/L as CaCO3
- Free chlorine 1 mg/L
• A & B are patient buildings supplied with the CSI-treated water • First hospital in Ohio to be regulated under the Safe Drinking
Water Act due to in-building water treatment
Insufficient Cu and Ag levels reaching hospital taps
Cu in
hot
wat
er (µ
g/L)
CSI Softener A A (N=5 taps) CSI Softener A A (N=5 taps)
40
Min. Intended after CSI
300
250
200
150
100
50
0
Ag in
hot
wat
er (µ
g/L)
10/1
4
Min. Intended after CSI
35
30
25
20
15
10
5
0
5/13
7/13
8/13
1/14
2/14
3/14
5/14
7/14
5/13
7/13
9/13
11/1
3
1/14
3/14
5/14
7/14
9/14
11/1
4
Date Date
Submitted to Water Research, 2016
40
Solubility modeling (Mineql+) for Cu
Total Cu Cu(OH)2(s)
Cu+2
CuCO3(aq)
Total Cu = Cu+2 + CuCO3(aq)
41
Solubility modeling (Mineql+) for Ag
Total Ag
Ag+
AgCl(aq)
Total Ag = AgCl(aq) + Ag+
42
Plating of reduced silver onto copper pipes Ag0 dendrites
Cu pipe
Reaction Potential, V Implication
More Noble Ag+ + e-↔ Ag0 +0.799 (Cathodic)
More Active Cu+2 + 2e-↔ Cu0 +0.342 (Anodic)
• Implications on silver disinfecting ability for bulk water and for biofilms • Possibility of deposition corrosion for Cu pipe 43
Aesthetic problems
CSI
• Grey/purple staining consistently observed in bathroom porcelain throughout buildings A and B • XRD analysis identified precipitate as AgCl(s) • Caused temporary inactivation of CSI
Ag+ + Cl- ↔ AgCl(s) K=5.62 x 109 at 25 °C 44
Conclusions
• The cation exchange softener installed in Building A for hot water treatment countered the CSI treatment
•Negative reactions to the staining led the hospital to consider alternatives that would eliminate the staining
• Deposition of metallic silver onto copper pipes after CSI activation was verified for the first time
• Extracting and analyzing pipes hidden inside walls can proactively identify interactions not visible to the naked eye
• Although the primary aspect of CSI is the effect on controlling Legionella and other pathogens in water, non-microbiological implications deserve exploration to holistically evaluate in-building drinking water disinfection
45
Acknowledgments
• Dr. Marc Edwards, Dr. Yanna Lambrinidou, Dr. Daniel
Gallagher and Trung Le (Master’s student) at Virginia
Tech
• Dr. Darren Lytle, Christy Muhlen and Mike Elk at EPA
(hospital project)
• Michael Schock and Dr. Michael DeSantis at EPA
(corrosion studies)