Spherulitic Lead Calcium Apatite Minerals in Lead Water Pipes Exposed to Phosphate-Dosed Tap Water

Phosphate dosing is the principle strategy used in the United Kingdom to reduce the concentration of lead in tap waters supplied by lead water pipes. The mechanisms of phosphate-mediated lead control are not fully understood, but solid solutions of lead calcium apatite are thought to play an important role. This study investigated the microstructure of a lead pipe, supplied with high-alkalinity tap water, in which the lead calcium apatite crystals were spherulitic having rounded and dumb-bell-shaped morphologies. XRD, Fourier transform infrared spectroscopy, optical microscopy, Raman spectroscopy, scanning electron microscopy, and energy-dispersive spectroscopy showed that the lead pipe had a well-established inner layer of litharge; a middle layer containing lead calcium apatite spherulites, plumbonacrite, and some hydrocerussite; and an outer layer containing iron, lead, phosphorus, calcium, silicon, and aluminum. It was found that spherulitic lead calcium apatite could be grown in the laboratory by adding hydrocerussite to synthetic soft and hard water-containing phosphate, chloride, and citrate ions at pH 5.5 but not when the citrate was absent. This suggests that dissolved organic molecules might play a role in spherulite formation on lead water pipes. These molecules might inhibit the formation of lead calcium apatite, reducing the effectiveness of phosphate dosing in lead water pipes.

. PHREEQC input files used to create speciation graphs in Figure Figure S1. Change in lead compliance in England and Wales with PbTOT < 10 ppb 1 . In the U.K. compliance is measured by taking random daytime samples from customer properties. Customers are chosen at random and samples are taken between 9am and 5pm on weekdays. Pb 2+ + Cit 3-⇌ PbCitlogk = 7.27 Pb 2+ + 2Cit 3-⇌ PbCit2 4logk = 6.53 Figure S2. PHREEQC input files used to create the speciation graphs shown in Figure S23.    Text S1.
Interpretation of the EDS Maps The purpose of the maps in Figures 2 and 4 of the main text was to show where the elements of Pb, P, Ca and Fe were concentrated. The maps from Figure 2 are presented below. The elements P and Ca are concentrated in the middle layer, Fe is concentrated in the outer layer and Pb is present throughout all layers as well as the Pb metal.
The maps also show minor amounts of P and Ca in the inner layer and Pb metal. However, there were no peaks for P and Ca in the EDS spectra of these regions ( Figure S9) and only background counts were observed. Therefore, the light shades of blue and red in the maps of P and Ca represent the background radiation. A similar explanation can be given for the apparent trace amount of Fe in the middle layer and Pb metal. The EDS spectrum of the outer layer ( Figure S9) shows that Fe is concentrated in the outer layer and isn't detectable in the middle, inner and Pb metal. The fact that the peak is small, even in the outer layer, suggests that it was close to the limit of detection and the small number of counts would increase the error associated with the peak. Therefore, the light shade of yellow / orange in the middle, inner and Pb metal the map of Fe represents the background radiation. Improvements to the maps could have been made by increasing the dwell time on each pixel (increasing the number of counts) and by accounting for and removing the background counts.
It is also possible that low concentrations of P, Ca and Fe in the inner layer and Pb metal was an artefact of the polishing process. In the case of P, this was present in the tap water used to polish the samples, at a concentration of  3 mg/L phosphate and this may have reacted with the freshly polished lead metal surfaces to give a false reading for the inner layer and Pb metal. Alternatively, some Fe, P and Ca containing material removed from the outer and middle layers during polishing may have been deposited on other parts of the sample. The possibility of contamination during polishing is why an argon ion mill cross section polisher was used in a recent study 5 , however this piece of equipment wasn't available in this study. Lastly, in the case of P, the trace amount seen in the inner and Pb metal might be due to the base of the large Pb peak at 2.36 KeV overlapping with the P peak at 2.02 KeV.
Although they have their limitations, the maps shows visually that P and Ca are concentrated in the middle layer and Fe is concentrated in the outer layer.  Figure S15. XRD spectra of laboratory grown spherulites. These were grown by adding hydrocerussite to soft and hard waters containing phosphate and citrate at pH 5.5. The peaks in both spectra index for lead apatite. The same set of peaks can be seen in the XRD spectra for pyromorphite, hydroxylpyromorphite and phosphohedyphane shown in Figure S4.

Soft water
Hard water Figure S16. XRD pattern and secondary electron SEM image of hydrocerussite used to make spherulites. a) XRD pattern showing a single peak at 2 = 34.0° for the {110} planes of hydrocerussite. The single peak was due to preferential orientation in which the large hexagonal (001) faces were parallel to the surface of the silicon sample holder. b) SEM image of the hydrocerussite crystals. b) a) S21 Figure S17. XRD pattern and secondary electron SEM image of crystals from the soft water control expts. a) XRD pattern of crystals in solution after adding hydrocerussite to a pH 5.5 soft water containing 10mg/L phosphate and10mg/L chloride and no citric acid. Samples were taken after 1 day. The XRD pattern indicated that there was a mixture of lead calcium apatite, hydrocerussite and cerussite, which in turn suggested that the original hydrocerussite had mostly dissolved reprecipitating as lead calcium apatite and to a lesser extent cerussite. b) SEM image of the crystals. Spherulites were not observed in this or other SEM images of the sample. Instead, there was a mixture of hexagonal shaped crystals, which were probably partially dissolved hydrocerussite and fine grain nano sized particulates, which could not be resolved. a) b) S22 Figure S18. XRD pattern and secondary electron SEM image of crystals from the hard water control expts. a) XRD pattern of crystals in solution after adding hydrocerussite to a pH 5.5 hard water containing 10mg/L phosphate and10mg/L chloride and no citric acid. Samples were taken after 1 day. The XRD pattern indicated that the sample was mostly lead calcium apatite together with a small amount of hydrocerussite and cerussite. The peak labelled with a * is probably a peal for calcium lead apatite. But it is broader than the other peaks and occurs at the same place as a peak for plumbonacrite and so there is uncertainty about its identity. b) SEM image of the crystals. Spherulites were not observed in this or other SEM images of the sample. The crystals appeared hexagonal in outline with small particulates on their surfaces. The particulates were probably lead calcium apatite and the retention of the hexagonal shape indicates that the replacement may have been psueodomorphic.   Figure S22. Structure of the citrate ion. It has three carboxylate groups, which can bind to Pb 2+ ions and one hydroxide group. Figure S23. Speciation diagrams for 0.5mM citrate in equilibrium with in soft and hard waters containing hydrocerussite. The diagrams were calculated using the stability data in Table S2 and the PHREEQC input files in Figure S2. The graphs are a guide only as the accuracy of the stability data is not known. At pH 5.5, the concentrations of PbTOT (total) in soft water containing citrate is 100 and in hard water it is 57 mg/l. The very high concentrations are due to the stability of the PbCit-ion. Diameter of the square area containing the spheres = 978 µm Diameter of the square area containing the spheres = 1 mm Step 7. Relationship between the volume of water remaining in the pipe and its thickness as a film V = 1.00x10 6 .(Dt-t 2 )L Thickness of the film = 1mm t = 0.001m V = 1.00x10 6 .(0.0095(0.001)-(0.001) 2 )0.1 V = 1.00x10 6 .(9.5x10 -6 -1.0x10 -6 )0.1 V = 2.7 cm3

Conclusions
The calculation shows that the maximum number of phosphohedyphane spheres of diameter 10µm that could have formed in a 10cm length of lead pipe was 9570. This assumes that all the dissolved phosphate in the remaining tap water reacted with lead to form phosphohedyphane and that the amount of lead available was very high. High concentrations might be generated by cracks that form after the pipe was drained. The 9570 spheres equates to approximately 0.032 spheres being observed in a SEM image of dimensions 100µm x 100µm or to 3.2 spheres being observed in a lower magnification SEM image of dimensions 1000µm x 1000µm. If all the spheres occurred together in one location, 1 layer thick, then the calculation shows that the area containing the spheres would have been 1mm across.
Image A (secondary electron plan view) is approximately 100µm x 100µm in dimension and many spherulites are visible, far more than the calculated value of 0.032. Image B (backscattered electron) is a polished cross-sectional view. At least 20 spherulites (8 of them are arrowed) are present. The number of spherulites present cannot be accounted for by the amount of phosphate present in the film of water. Therefore, they must have formed prior to the pipe being drained when it was supplying water to the property. The spherulites were therefore not an artefact of sample preparation.

S32
Image A 0m 500m 1000 Image B