PIPE SCALE ANALYSES FOR CCT EVALUATION

EPA REGION 5 LCR-OCCT WORKGROUP AUGUST 12, 2020

1 What is a Service Line (LSL) pipe scale?

• The combination of lead corrosion products and deposited materials found inside a LSL • Crystalline lead minerals • Deposited materials • Represents decades of reactions and changes in water quality/treatment

2 Why analyze LSLs?

• To evaluate the state of corrosion control treatment (CCT) and understand lead release • To predict impacts of potential changes to CCT or other water treatment processes • To address shortcomings in the ability of theoretical models to predict the lead minerals forming in the field

3 Why analyze LSLs?

• When present, LSLs are the largest source of lead to drinking water • Many decades out from removing all LSLs in the United States • Systems need to know by what mechanism lead release is being controlled

4 Lead minerals contain a lot of lead…

5 Considerations

• Due to the complex nature of many LSL scales there are several practical and technical considerations to take into account

• This presentation will focus on: • How to sample scale from LSLs • LSL scale analysis • Powder X-ray diffraction • Scanning electron microscopy/energy dispersive spectroscopy

• Case study illustrations

6 Visual Characterization

• Detailed description and Outer documentation of the scale Middle • texture, color, physical Inner characteristics etc. • “Macro” and microphotography Drinking Water • Designation of scale layers Lead Pipe Wall

7 Scale Layers Depending on treatment history and distribution system nuances Can be simple… L1 (Pb(II) ) L2 (PbO - litharge)

Or complex…

L1 (amorphous Mn)

L4 (PbO2 - ) L2 (amorphous Fe) L5 (lead phosphate) L3 (amorphous Si,Al) L6 (PbO - litharge)

8 Sampling by layer

• Once the layers are defined the process of carefully subsampling the identified layers can begin • Can be time consuming, L1 depending on complexity of the scale L2 • In some cases may not be able to separate very thin or very L3 friable materials L4

Experience has taught that there can be distinct differences in the mineralogy of the layers, these nuances are lost when the scale is sampled as a whole 9 Resulting Material

• Separated layers are then individually crushed and passed through a 75 µm sieve • One of the drawbacks of subsampling layers will be a low yield (~200 mg or less in some cases) • Some LSLs will provide plenty of material for certain layers whereas other layers can be difficult to obtain enough sample

10 ORD’s main components of scale analysis

Analyses performed on individual layers:

Destructive / Sample amount Analytical Technique Information Nondestructive needed Mineralogy of As little as 10 mg Powder X-ray Diffraction (XRD) crystalline Nondestructive (more is better) material

X-ray Fluorescence (XRF) 200 mg

Elemental Total Carbon/Total Sulfur (TC/TS) Destructive 50 mg composition

Total Inorganic Carbon (TIC) 10-20 mg

Scanning Electron Microscopy Images and spatial (SEM) / distribution of -NA- -NA- Energy Dispersive Spectroscopy elements (EDS) 11 Powder XRD

Powder top-loaded onto silicon holder • Non-destructive method • Ability to analyze small amounts of material

• Provides information Amyl acetate slurry regarding the crystalline onto quartz holder and (to some extent) poorly-crystalline components of the analyzed material

12 Cautions and Caveats • Contamination/Disturbances Trench Sediment Lead Shavings

Cable Tooling

13 More Cautions and Caveats

• Peak overlaps • Pipe scales are rarely composed of a single mineral phase • Auto-ID feature in pattern fitting software should be used with caution • Some practical knowledge of expected (and improbable) minerals in a drinking water environment is useful

14 Some XRD Pattern Complications

Outer layer of ~10μm thick scale • Multiple phases • Peak overlaps • Sediment and lead • Pb phosphates rarely exactly fit database

Outer layer (L1) Pipe scale Inner layer (L2)

Lead pipe wall Lead pipe wall

15 Crystalline

vs.

Amorphous 16 SEM/EDS Analysis

• Preservation of in-situ layer relationships from a representative undisturbed section of the LSL • Cross-sectional view • Useful for understanding complex layering

17 Cross Section equipment

18 Cross Section Preparation

EDS WDS

SEM Chamber

19 Elemental Maps are Useful for General Spatial Relationships

BSD Epoxy L1 L1 L2

L2 L2 L3

Lead Pipe Wall

EDS results from a spectrum taken over an XRF results from subsampling the individual entire area layers Element Weight % Element L1 L2 L3 Al 0.37 Al 1.4 0.8 0.3 Mn 2.05 Mn 7.4 2.7 0.2 Sn 1.05 Sn 2.8 3.1 1.4 Pb 78 Pb 58.6 71.3 83.620 SEM/EDS Interaction Volume

21 Caution Required When Evaluating SEM/EDS Results Common output for SEM/EDS analysis EDS Spectra Elemental Maps Al 0.6 wt% by XRF However, to truly interpret the Si elemental maps the 2.52 wt% associated spectrum by XRF needs to be evaluated

Pb

93 wt% by XRF

While Al appears across the entire map and at a comparable brightness to that of Pb and Si this cannot be interpreted as the elements are present in equal concentrations to one another 22 Case Study Illustrations

• How can scale analysis inform decisions? • Diagnostic/Forensic • Track Results/Changes • Predictive/Preemptive

23 Diagnostic/Forensic Case Study 1: Washington, DC

• Experienced LCR exceedances after switching disinfectants in the year 2000, chlorine to monochloramine • Added orthophosphate starting in August 2004 • LSLs from after the switch to monochloramine and after the addition of orthophosphate were analyzed

24 Scale Analysis Revealed…

After ~ 15 months of 3- PO4 treatment

Discontinuity surface

P, Ca, O, and Al enrichment

Orthophosphate was bypassing the relict Pb(IV) plattnerite layer at the scale water interface and was instead being incorporated into the lower scale layers 25 Washington, DC XRD

Lead phosphate

26 Diagnostic/Forensic Case Study 2: Halifax, NS

History of CCT changes • 1980’s – 1999: Lake Lamont Treatment Facility • orthophosphate at 1 mg/L and 1.5-2.0 mg/L during annual flushing/line cleaning • 1999: Lake Major Water Supply Plant brought on-line

• 1999-2000: pH/alk CCT -- pH 7.6, alkalinity 19 mg/L as CaCO3 • 2001-2002: pH/alk CCT -- pH 7.8, alkalinity 32 mg/L as CaCO3 • 2002-2005: pH/alk CCT – pH 7.0-7.4, alkalinity 8 mg/l as CaCO3 • 2005: phosphate CCT using 75:25 ortho/poly blends – pH 7.4, phosphate dose at 1.5 mg/L, 2 mg/L during annual system flush.

LSLs harvested in 2011 to evaluate why CCT was not lowering Pb levels as much as expected 27 Scale Analysis L3 L4 Revealed…

• Thick Fe/Mn layer

• Minimal development of lead phosphate

• Leadhillite/ – not commonly found in Pb pipe scales L1 L2

Compound/Formula Layer 1 Layer 2 Layer 3 Layer 4 Poorly crystalline Fe & Mn ++++

Hydrocerussite (Pb3(CO3)2(OH)2) ++ +++ ++++ +++

Leadhillite/Susannite (Pb4(CO3)2(SO4) (OH)2) + ++++ +++ +

Hydroxypyromorphite (Pb5(PO4)3OH) + + ++ ++ Litharge (PbO) ND ND ND ++++ 28 Phosphate tied up by Mn/Fe deposit

L1 (thin – Mn-rich layer)

L1 (thick – w/ Mn and Fe sublayers)

L2/L3 L4

Element (wt%) Layer C O Al Si Ca Mn Fe Pb Zn P S L1 1.84 26.00 3.26 1.93 1.30 8.48 11.85 38.94 1.52 3.31 0.09 L2 3.02 10.10 0.88 0.57 0.31 0.20 1.56 81.87 0.20 0.62 1.05 L3 2.95 8.53 0.28 0.20 0.16 0.03 0.19 87.11 0.06 0.50 0.70

1 Tracking Changes Case Study 1: Fall River, MA

• LCR exceedances from 1992-2006 • Due to aging infrastructure there was a minimal chlorine residual within the DS and residual was almost completely depleted within storage tanks • This resulted in a Total Coliform Rule Violation in 2001 • After increasing the chlorine residual there was a Stage 2 Disinfection By-products violation in 2013 • As part of their LSL replacement program, EPA analyzed LSLs from the system annually beginning in 2001

29 30 2001-2013 XRD Evidence Blue boxes show pattern features for hydrocerussite Red dashed boxes show pattern features for plattnerite

31 Scale Analysis Revealed…

• A mix of lead phases governing lead release within the system (Pb(II) and Pb(IV)) • Pb(IV) formation accelerated as a result of water treatment plant improvements (2005) and was found to remain stable into 2006 • LSLs from 2013 show a regrowth of hydrocerussite which indicates deteriorating conditions for maintaining stable Pb(IV) and low Pb release

32 Tracking Changes Case Study 2: Providence, RI • Surface water system (Scituate Reservoir): pH 10.3, alkalinity ~15 mg/L as CaCO3, hardness ~42 mg/L as CaCO3 • Fall 2005 – Spring 2013: pH lowered to 9.7 to optimize dissolved Pb (based theoretical model understanding) • 90% Pb levels increased from 10-15 μg/L to 20-30 μg/L • Pb on iron particulates • LSLs extracted and sent to EPA beginning in 2006 (50+ pipes to date) • Testing orthophosphate CCT • Pipe loop: 2014 - 2016 • Partial system test: 01/2018 – present

• Dose: 3 mg/L as PO4

33 Scale Analysis Revealed…

• Crystalline calcium phosphate phase(s) precipitated on surface of pre-existing lead corrosion scales

• Partial conversion of pre-existing hydrocerussite/plumbonacrite scale to calcium-substituted hydroxypyromorphite (Cax,Pb5-x)(PO4)3OH

Pipe Loop Partial System Test

34 BSD

Epoxy

L1

L2 L3a

Lead Pipe Wall

35 Something unexpected

• PbO2 developed on parts of 3 of 4 LSLs harvested from the partial system test zone • Occurs only on areas underlain by PbO (litharge)

• Previously observed on only 2 pipes of 50 harvested between 2006 and 2014

36 Predictive Case Study: Newport, RI

• Surface water system, pH ~8.3, free chlorine, pH/alkalinity CCT, very reactive organic material • Due to DBP issues (particularly in consecutive systems), was contemplating moving from free chlorine to chloramine

• LCR compliance samples lower than expected for the pH • Targeted LSL water sampling also showed very low Pb levels in overnight standing samples

 Suspected Pb(IV) scales

37 Scale Analysis Revealed… Pb(IV) Plattnerite

Explained Newport’s low 90th percentile values (<15 ppb)

Solution: Newport built a brand new treatment plant to better control DBP precursors and did not switch to chloramine 38 One Final Caution: Chemistry Context and Experience Matters

Outlines = Void Space Higher Pb concentration Pb

250 µm 250 µm

Higher Pb concentration Interpretation A Interpretation B Porous, poorly-adherent scale of non- Void spaces left by dissolution of crystalline material, consistent with highly crystalline Pb (II) other observations in systems with orthophosphate solids, like hydroxy- phosphate and blended phosphate or chloro- treatments 39 Conclusions

• Scale analysis is a complicated undertaking that is labor intensive • Experience, robust methodology, and an attention to detail are needed; otherwise wrong conclusions may be drawn • Results of various analyses are best interpreted holistically and with historic system water quality parameters when possible • Information learned from scale analysis can provide insight into the potential impacts of water quality or CCT changes • Solutions are most effective when predictive studies are conducted, unfortunately this category of study has been uncommon thus far

40 Questions?

[email protected] [email protected] [email protected]

Notice This document has been reviewed by the U.S. Environmental Protection Agency, Office of Research and Development, and approved for publication. The views expressed in this presentation are those of the author(s) and do not necessarily represent the views of the U.S. Environmental Protection Agency. Any mention of trade names, products, or services does not imply an endorsement by the U.S. Government or the U.S. Environmental Protection Agency. The EPA does not endorse any commercial 41 products, services, or enterprises.