Elsevier

Corrosion Science

Volume 52, Issue 10, October 2010, Pages 3492-3503
Corrosion Science

Increase of the concentration of dissolved copper in drinking water systems due to flow-induced nanoparticle release from surface corrosion by-products

https://doi.org/10.1016/j.corsci.2010.06.027Get rights and content

Abstract

Standard measurements of dissolved copper are made by filtering water samples through 0.45 μm pore-size membranes. However, the surface of corroding metallic surfaces may be covered by topographic features < 0.2 μm and structures that can be detached into the bulk water as nano-sized particles. A SEM, EDX, and AFM characterization of a corroding pipe after flow events revealed surface cavities, detached particles and attached particles with sizes between 0.05 and 0.2 μm. Our findings show that the release of colloidal and nanoparticles of corrosion by-products into the water can result in an increase of the dissolved copper measurements.

Research highlights

► Hydrodynamics determine the release of Cu corrosion by-product nanoparticles ► Detached Cu corrosion by-product nanoparticles were identified as malachite ► Copper release models have to consider hydrodynamics within the pipes

Introduction

For 200 years copper pipes have been used for domestic water services around the world [1]. Nowadays it is the most used piping material in household drinking water systems because its high corrosion resistance [2]. There are considerable experiences and scientific understanding on the properties of this noble material [3]. Since the first copper pipe, there have been a large number of manufacturing developments to obtain better material performances in order to decrease the concentration of copper released into the tap water.

Although copper is an essential metal for the human diet, in some cases the ingestion of copper and long-term overexposure can generate acute and chronic health effects including gastrointestinal diseases and liver damage [4]. The World Health Organization (WHO) recommends 2 mg/L as a maximum concentration value for drinking water [5]. This value is based on gastrointestinal epidemiologic studies conducted on populations under controlled exposures. Several of those studies used copper sulfate salts to add dissolved copper into the drinking water [6], [7], [8]. As a result, over the last decade, research on drinking water supply systems and copper pipeline corrosion has focused on determining and modeling the processes that control the release of copper into the water [9], [10], [11], [12]. In fact, based on experimental measurements, models have been developed to estimate the concentration of soluble copper released into the tap water [13], [14].

In general, the release of soluble copper into the water is controlled by three processes: (1) an electrochemical process that involves two half-reactions, one anodic (metallic copper oxidation) and one cathodic (dissolved oxygen reduction) [15]; (2) a scale formation process associated with thermodynamic equilibrium conditions, affected by pH, dissolved oxygen (DO), temperature, and presence of ions [3], [16], [17]; and (3) a dissolution of solid corrosion by-products and release of dissolved copper into the bulk water [3], [18].

Current knowledge of copper corrosion supports the theory that, for new pipe systems, soluble corrosion by-product release into water is controlled by the solubility of cupric hydroxide (Cu(OH)2(s))[10], [19]. For aged pipe systems, the transition to less soluble and stable phases is catalyzed by the presence of anions in water [20], [21]. In cold and low mineral waters cupric hydroxide ages to tenorite (CuO(s)), however, in water with high dissolved inorganic carbon concentration, ages to malachite (Cu(OH)2·CuCO3(s)) [3]. Malachite dominates the solid phase speciation of Cu(II) for pHs between 5 and 9 [22].

Fluid flow can also affect corrosion by changing chemical and mechanical conditions at the metal–liquid interface. Flow-induced corrosion has been traditionally classified in four different types: mass-transport-controlled corrosion, phase-transport-controlled corrosion, erosion–corrosion, and cavitation corrosion [23], [24], [25]. Mass-transport-controlled corrosion is related to the increased rate of mass-transport due to the flow velocity profile which contributes to increase the amount of corrosive species reaching the metal surface, or alternatively, by enhancing the removal of dissolved corrosion by-products from the solid phase. Phase-transport-controlled corrosion occurs when a liquid phase containing the corrosion agent gets in contact with the metal surface. Erosion–corrosion is associated to the mechanical removal of protective layers from the metal surface by high-velocity turbulent flows through shear stresses applied on solid boundaries. Cavitation corrosion takes place when liquid pressure drops below the vapour pressure, generating an implosion of gaseous bubbles that creates impulsive forces capable of removing material from the solid phase [18], [25], [26]. In addition, there is evidence that mechanical removal of nanoparticles from the metal surface can also occur even for low velocity flows. This process seems to be similar to the so-called erosion–corrosion, but acting at a smaller scale where shear stresses might be capable of sloughing micro and nanoparticles from corrosion by-products [18], [25], [26]. Similarly, fluid flow can enhance concentration gradients that facilitate desorption of labile copper bonded to organic moieties attached to the metallic surface [18]. Therefore, for domestic pipe systems where flow velocities are rather low, it is important to identify the mechanisms that control the release of dissolved and particulate copper into tap water, especially because the health implications of particulate copper are poorly characterized.

According to our current knowledge there is not a study that simulates the dissolution of copper particles in the gastric fluid, however a recent work of lead contamination in drinking water systems shows that gastric fluid can easily dissolve lead particles, with the associated health impact of soluble lead released into the human body [27]. On the other hand, Taylor et al. [28] found through a modeling study that corrosion and dissolution potentials of copper nanoparticles are dependent on the size and shape of the particle. Thus, the typical thermodynamic values calculated for metals and minerals (corrosion by-products) are not necessarily accurate for small particles.

The standard measurement of dissolved copper is made by filtering the water sample through a membrane with a pore-size of 0.45 μm [29]. However, a passivating film of copper carbonate hydroxides, such as malachite, growing over the metallic surface is formed by the aggregation of structures with a size less than 0.2 μm that could be detached into the water due to flow [30]. Thus, the standard definition of dissolved copper includes both soluble cuprous and cupric species, together with particulated copper. To avoid this inaccuracy, operational definitions have been used to describe the size distribution for particulate species in drinking water [31], [32]. McNeill and Edwards [31] divide dissolved copper into colloidal (0.1 μm < [Cu] < 0.45 μm) and soluble copper ([Cu] < 0.1 μm). Using this definition, (nanoparticles) particles < 0.1 μm could be confounded with soluble copper (Table 1). Interestingly, even though the release of particulate copper has been reported [33] and its detachment can be linked with the hydrodynamic conditions of the piping system [18], the effect of flow-stagnation events on the detachment of micro and nanoparticles of copper corrosion by-products has been poorly considered.

Our work focuses on the effect of hydrodynamic conditions on the detachment of copper corrosion by-product nanoparticles under abiotic conditions. This paper presents evidence of the detachment of nano and micro copper carbonate hydroxide structures formed on the inner surface of copper pipes, induced by the shear stress produced by the fluid flow, which increases the concentration of dissolved copper in water.

Section snippets

Materials and methods

Flushing experiments were conducted using a single-pass laboratory system, which consisted of a 1 m long copper pipe with an internal diameter of 1.95 cm and 0.3 L of volume, preceded by a PVC pipe and a Cole-Parmer model N° 7553-75 peristaltic pump connected to a water tank. The copper pipes were preconditioned in a three steps protocol: (1) the pipes were filled with NaOH (0.1 M) to dissolve all oxides present on the inner surface of the pipe. (2) After two minutes with sodium hydroxide, the

Results

The results of the flushing experiments are organized in three sections: (1) copper release measurements, depicted by curves of copper mass versus the volume of water extracted from the tested pipes; (2) surface analyses that include a detailed multi-method characterization of the pipes before and after the flushing experiments and; (3) particle observations, where the results of micro and nanoparticles captured by the sequential filtering procedure are shown.

Discussion

For non-reactive surfaces, a plug-flow analysis would be sufficient to characterize the release of copper during the flushing event. This non-reactive characteristic would be noted as a sudden increase followed by an early stabilization in the mass of copper released (Fig. 2). However, for pipes coated with a reactive film of solid corrosion by-products, the ideal plug-flow assumption is not adequate to describe the release of copper into the water, and the effect of particle detachment must be

Conclusions

This work is the first effort aimed at characterizing the effect of hydrodynamic conditions on the release of copper corrosion by-product nanoparticle into drinking water systems.

Flushing experiments conducted under laminar and transition to turbulent conditions show that even if the wall shear stress produced by the flow is one order of magnitude smaller than the erosion–corrosion threshold values reported by Efird [25] mechanical detachment of nanoparticles could occur. Thus, corrosion

Acknowledgements

This research was funded by CONICYT Grant 24080013/2008 and FONDECYT Project 1080578/2008.

References (50)

  • I.T. Vargas et al.

    Influence of solid corrosion by-products on the consumption of dissolved oxygen in copper pipes

    Corros. Sci.

    (2009)
  • I.T. Vargas et al.

    Empirical model for dissolved oxygen depletion during corrosion of drinking water copper pipes

    Corros. Sci.

    (2010)
  • J.H. Hong et al.

    Ultrafiltration as a tool to study binding of copper to salivary proteins

    Food Chem.

    (2009)
  • R.J. Oliphant

    Causes of Copper Corrosion in Plumbing Systems

    (2003)
  • S.O. Pehkonen et al.

    Effect of specific water quality parameters on copper corrosion

    Corrosion

    (2002)
  • T.H. Merkel et al.

    General corrosion of copper in domestic drinking water installations: scientific background and mechanistic understanding

    Corros. Eng. Sci. Technol.

    (2006)
  • A.M. Dietrich et al.

    Rating method for evaluating distribution-system odors compared with a control

    Water Sci. Technol.

    (2004)
  • WHO

    Guidelines for Drinking-water Quality

    (2004)
  • M. Edwards

    Controlling corrosion in drinking water distribution systems: a grand challenge for the 21st century

    Water Sci. Technol.

    (2004)
  • M.R. Schock, D.A. Lytle, J.A. Clement, Effect of pH, DIC, Orthophosphate and Sulphate on Cuprosolvency, in...
  • B.Y. Shi et al.

    Iron and copper release in drinking-water distribution systems

    J. Environ. Health

    (2007)
  • USEPA, Maximum contaminant level goals and national primary drinking water regulations for lead and copper; final rule,...
  • T.H. Merkel

    Copper corrosion: understanding and modelling general corrosion

    Water Sci. Technol.

    (2004)
  • D.J. Ives et al.

    Copper corrosion III. Electrochemical theory of general corrosion

    J. Electrochem. Soc.

    (1962)
  • M. Edwards et al.

    The role of pipe ageing in copper corrosion by-product release

    Water Sci. Technol. Water Supply

    (2001)
  • Cited by (21)

    • Georgeite: A rare copper mineral with important drinking water implications

      2019, Chemical Engineering Journal
      Citation Excerpt :

      Since the implementation of the LCR, a huge amount of research has been conducted on copper corrosion, solubility, and leaching from materials in drinking water distribution systems. Many of these research findings have improved the water industry’s understanding of how water chemistry affects the solubility of copper minerals found in drinking water distribution systems, the role of stagnation time on copper leaching from copper pipes, and the use of corrosion inhibitors to reduce copper levels at the consumer’s tap [4–9]. Additionally, new information on copper solubility in water has relevance to the drinking water premise plumbing (including buildings) treatment field.

    • Impact of chlorinated disinfection on copper corrosion in hot water systems

      2014, Applied Surface Science
      Citation Excerpt :

      Presence of CuO was also detected on 100 ppm aged samples, but the tenorite amount was very low explaining why it was not detected on XRD diagrams. Tenorite indicates a more advanced corrosion state compared with samples exposed to reference tap water [32–37,55]. The presence of the disinfectant considerably amplifies the degradation through additional cathodic reactions that normally accelerate the pH alkalinisation at the surface vicinity.

    • Modeling MIC copper release from drinking water pipes

      2014, Bioelectrochemistry
      Citation Excerpt :

      This has motivated scientists to study the corrosion on copper pipes to understand the involved mechanisms [4–10]. These studies have been focused primarily on: (1) the influence of water quality parameters (pH, temperature, oxygen, alkalinity, chloride, sulfate, phosphate and organic matter)[5,7,9,10]; (2) operating conditions (flow-stagnation cycles and age of the pipe) [6,11,12]; and (3) the release of copper into the drinking water. Although complex conceptual models have been developed, mathematical implementation of these models has not had major advances.

    View all citing articles on Scopus
    View full text