Black & Veatch's Frank Rogalla looks at minimising bromate formation
Both in Europe and the US, drinking water standards include a 10µg/l limit for bromate. For many plants that have included ozonation in their treatment processes, compliance with the bromate standard can pose a considerable challenge if raw water supplies contain bromide at concentrations approaching 100µg/l or higher. This challenge is accentuated if the ozonation process is designed for or modified to provide inactivation of Cryptosporidium parvum. This article describes research to develop an optimal treatment method to minimise bromate formation when operating ozonation facilities to achieve an internally adopted target of 2-log (99%) Cryptosporidium inactivation. This research culminated in the issuance of a patent that has been placed in the public domain. Las Vegas case study
In Las Vegas, water treatment and distribution to meet the needs of the area's rapidly growing population is provided by the Southern Nevada Water Authority. Approximately 90% of the Las Vegas valley water supply derives from Lake Mead, which receives water from the Colorado River, and the other 10% is groundwater. The raw water has a pH that varies seasonally from 7.5-8.2; an average alkalinity of 130mg/l as CaCO3; an average total hardness of nearly 300mg/l as CaCO3; a turbidity that is usually below 1NTU; a total organic carbon (TOC) concentration from 2.5-3.5mg/l; an average TDS concentration of 600mg/l, with peaks of 800mg/l; and a bromide ion concentration from 60-120µg/l.
The two large drinking water plants in the Las Vegas area are both direct filtration facilities, providing ozonation preceding coagulation, with treatment capacities of 2,280Ml/d and 570Ml/d. The smaller plant was commissioned in 2002 to include ozonation, whereas the larger facility, first started in 1971, was retrofitted with ozonation facilities in 2003.
Ozone is used at both treatment facilities for two principal reasons:
- to enhance filtration performance to consistently maintain a filtered water turbidity less than 0.1NTU from each filter,
- to achieve effective inactivation of Crypto. parvum.
The utility adopted an internal standard of 4-log (99.99%) removal/inactivation of Crypto. with 2-log removal through direct filtration and 2-log inactivation through ozonation. Ozone is generated at both treatment facilities from high-purity oxygen produced on-site by vacuum-pressure swing adsorption (VPSA) air separation units. High-frequency ozone generators produce dosages up to 3.2mg/l. The ozone contacting facilities at both plants consist of multiple parallel ozone contactors with a total detention time of 24min each.
Ozone is applied through porous fine-bubble diffusers in the first cell of each 12-cell contactor. The ozone dose necessary to achieve a 2-log inactivation of Crypto. was determined using the CT concept and an equation derived from data generated during a bench-scale inactivation study of Lake Mead water.BROMATE RESEARCH
With the bromate formation potential of the raw water recently identified and the full-scale retrofit ozone facilities under construction, bromate mitigation methods were evaluated during a three-year pilot research program and most of the laboratory analyses were performed by the Southern Nevada Water Authority laboratory. The pilot plant consists of chlorine contact, ozonation, rapid mixing, four-stage flocculation and dual media filtration. Raw water from Lake Mead is supplied to the pilot plant before any treatment chemicals are added.
The chlorine contactor provides 7min of contact time to simulate detention in the full-scale raw water pipeline. The ozone contactor is made up of 12 cells to provide approximately 24min of contact time at the design flow rate of 6gpm (14m3/h). Ozone can be added through porous diffusers mounted horizontally near the bottom of cells one, two and five. The rapid-mixing, flocculation and filtration steps are scaled to mimic those in the full-scale plant.
The two 25cm diameter stainless steel dual media filters contain 50cm of anthracite and 30cm of sand, identical to the media size in the full-scale plant. The pilot plant is equipped with online meters for measuring turbidity, particle counts, pH, temperature, conductivity and the concentrations of dissolved oxygen, dissolved ozone, feed-gas ozone, and off-gas ozone. All of the online meters are tied into computer software that displays and logs the data.
The first step in the research programme was a screening level evaluation of commonly used bromate mitigation methods, including pH adjustment, ammonia addition and preoxidation to reduce the ozone dose. The pilot plant results showed pH adjustment was effective in reducing the formation of bromate, however, the acid and base doses required to achieve this effect were excessive. The addition of ammonia to act as a 'sink' for the bromide through the formation of bromamines was also effective but would require high chlorine doses to 'breakpoint' the residual ammonia. Chlorine pre-oxidation to reduce bromate formation did not enable compliance with the impending 10µg/l MCL. The most promising results, which warranted further investigation, were obtained by using a combination of chlorine preoxidation and ammonia addition. A matrix of tests was developed to investigate the use of chlorine pre-oxidation, ammonia addition and combinations of the two.
The matrix consisted of more than 100 short-term tests of about 4h each. Chlorine was added to the raw water at doses of 0.25-0.5mg/l and allowed to react for 7min to simulate detention time in the full-scale raw water pipeline. Ammonia at doses of 0.1, 0.3 and 0.5mg/l was added immediately ahead of the first cell of the ozone contactor.
The combination of chlorine preoxidation and ammonia addition was evaluated using doses of 0.25 and 0.50mg/l chlorine and 0.1, 0.3 and 0.5mg/l ammonia to determine whether adding both chemicals produced a synergistic effect. During testing with the combination of chlorine pre-oxidation and ammonia addition, no free chlorine residual was present at the point of ammonia addition. The optimum bromate reduction technique identified by the matrix testing was evaluated further during long-term operation to simulate full-scale operation.
A three-month pilot plant testing period confirmed the effectiveness of the optimal bromate mitigation conditions and enabled researchers to evaluate the capability of biological filtration to convert residual ammonia to nitrate. This was important because converting ammonia to nitrate reduces the amount of chlorine necessary to breakpoint the residual ammonia and provide a free chlorine residual.
Test results were promising, with the combination of chlorine and ammonia pre-treatment yielding the best results. Adding ozone alone without bromate control measures and targeting 2-log inactivation of Crypto. resulted in bromate concentrations of 20-25µg/l. Pre-treatment with a combination of 0.5mg/l of chlorine followed by 0.1mg/l of ammonia reduced the bromate concentration to about 5µg/l at the same ozone dosage used to achieve 2-log Crypto. inactivation with no pre-treatment. Adding 0.5mg/l of ammonia with no preceding chlorine addition reduced the bromate concentration to about 5µg/l, however, the chlorine demand of the filter effluent to breakpoint the residual ammonia and produce a free chlorine residual would be very high. The combination of chlorine and ammonia pretreatment provided the lowest bromate concentrations at the lowest chemical dosages and operating costs. Another promising result observed during the long-term testing was that the biologically active filters began to achieve ammonia nitrification, thereby reducing the chlorine demand of the filtered water.PATENT GRANTED
This innovative approach to controlling the formation of bromate during ozonation was filed as a process patent application. The patent, titled Water Treatment Using Ozone for Reduced Likelihood of Bromate Formation from Bromides Formed in the Water and numbered US 6,602,426 B2, was issued on August 5, 2003. The patent documents the invention with figures and examples and presents 41 claims.
The motivation for seeking a patent was not for profit but rather to protect the authority sponsoring the research as well as other utilities against those who might patent the process and charge for its use. The patent was therefore placed in the public domain, where it is essentially free for use. So, drinking the tap water in Las Vegas is a safer gamble than using the bottle