M.Sc Thesis

M.Sc StudentRadu Irina
SubjectStudy of Pretreatment of Secondary Wastewater in
Reverse Osmosis
DepartmentDepartment of Chemical Engineering
Supervisors PROFESSOR EMERITUS Raphael Semiat
Full Thesis text - in Hebrew Full thesis text - Hebrew Version


Usage of secondary treated sewage for reducing the water shortage problem has increased in Israel and all over the world. There is a growing demand for high quality secondary treated wastewater, purified from organic and microbial contaminations by a UF/MF process and from dissolved salts, by an RO process. One of the problems in the RO process is the presence of phosphate ions in wastewaters. Calcium phosphate is a sparingly soluble salt which can readily precipitate on the membrane during the concentration process accompanying permeate withdrawal. Currently available anti- scalants are not sufficiently reliable to provide effective inhibition of phosphate scales in the RO process.

The specific objective of this research work was to develop a novel simple method for removing phosphate from wastewater fed to the RO process. Calcium phosphate is readily precipitated in an alkaline environment. The principle of the proposed technique is to create an alkaline environment by percolating the feed water through an MgO granular packed bed. Beds of such granules are commercially available and widely used for neutralizing acidic waters. The primary advantage of the proposed technique is in providing a relatively simple process for coping with the calcium phosphate scaling problem.

The goals of the research program were to obtain data on the efficiency of phosphate removal by a magnesia packed bed, to characterize the need for periodic cleaning of the bed and to develop a backwash technique enabling effective regeneration of a clogged bed. The experimental system was designed to enable scale precipitation in continuous flow through two columns connected in series. The feed solution was tap water (Talk = 117.5-170 ppm as CaCO3, Ca = 50-64 ppm) with and without phosphate addition (PO4 = 4-11 ppm).
The dimensions of each column were: diameter of 98 mm and height of 1800 mm, bed height of 500 mm.

The first column contained 4-4.5 kg of granular MgO particles. The strong alkaline environment created by the dissolution of magnesia served to precipitate two scaling components from the feed - calcium phosphate and calcium carbonate. Three kinds of magnesia having different chemical composition and grain sizes were tested:

  • Hydrolit Mg I and Mg II of Akdolit Co. Germany, 70-75% MgO, particle size of 0.5-2.5 mm for Mg I and 2.0-5.0 mm for Mg II.
  • Dead Sea Periclase Co. Israel, 99% MgO, particle size of 2-2.6 mm
  • Corosex of Clack Corporation USA, 97% MgO, mean particle size of 1.4 mm.

The second column contained 4-4.5 kg of limestone particles, 3 to 6 mm average size, of a chemical purity of over 99.5% CaCO3. The purpose of this column was to complete the precipitation process and to filter out precipitated particles. The experiments were carried out at feed flow rates providing residence times of 8.8 - 44 kg ⋅ min/ L in the magnesia column. The main efforts were centered on characterization of the effectiveness of the various types of magnesia particles, on study of the loss of bed activity through decay of the pH, on tests of various bed regeneration techniques and on clarification of possible mechanisms causing vigorous deactivation of the magnesia bed.

The main findings of the research are as follows:

  1. Chemical characterization of the MgO column:
    1. Beaker tests indicated that the magnesia phase governing the solubility of all tested particles is Active Magnesia and not Brucite or Periclase.
    2. The outlet magnesia solubility in continuous flow experiments, expressed bythe saturation ratio index of Active Magnesia SIact, was of the order of 0.11 when feeding distilled water to the bed and of the order of 0.01-0.06 when feeding tap water.
  2. Effectiveness of phosphate and carbonate scale removals:
    1. Fresh magnesia particles were found to be most effective in precipitating calcium phosphate and calcium carbonate scales.
    2. A very high alkaline environment, of pH levels in the range of 10-11, is created by the magnesia bed. Under these conditions, almost all the phosphate is precipitated (residual PO4 concentration below 0.2 ppm), and 40-50% of the CaCO3 is co precipitated with the phosphate.
    3. The limestone column was found to have a negligible contribution to the precipitation process.
  3. Deactivation of the magnesia bed
    1. The potential of calcium phosphate and of calcium carbonate precipitation markedly affected by the pH level of the solution. The effectiveness of the magnesia bed in precipitating the scaling salts was found to deteriorate gradually due to a continuous decay in the pH level generated by the magnesia particles.
    2. An unexpected result of this research was the vigorous deactivation of the magnesia bed that resisted all regeneration techniques used in media filtration beds.
    3. The experimental data indicated that the main parameter governing the decay in bed activity is the total amount of treated solution. The addition of calcium phosphate and calcium carbonate seed crystals to the magnesia bed did not have an effect on the rate of bed deactivation.
  4. Bed deactivation mechanisms.
    1. Scanning Electron Microscope (SEM) photographs of used magnesia particles showed significant coverage of the particles by adherent crystallites.
    2. Energy Dispersive Spectroscopy (EDS) analyses revealed that significant amounts of calcium phosphate and calcium carbonate were present on used magnesia particles.
    3. The failure of vigorous mechanical backwash efforts, by water and air, to dislodge the precipitated scale compounds from the bed, indicates that the crystallites observed on the magnesia particles were attached by strong chemical bonds.
    4. A literature search did not reveal information on the phenomenon of strong chemical bonding between scaling species and MgO particles. This unexpected phenomenon merits elucidation.
  5. The main advantage of phosphate removal by percolating a solution through a magnesia bed is the operational convenience of avoiding the need of a chemical dosage system. The present work has shown the drawback of a strong deactivation of the magnesia particles by adhering scaling compounds. This drawback might be overcome by using a fluidized bed of magnesia powder.
    Kaneko and Nakajima (1988) studied phosphate precipitation in a pilot plant containing a fixed bed of magnesia particles. They were able to obtain a very high phosphate removal for a very long period by controlling the feed pH through dosage of a NaOH solution. The need to incorporate an alkali dosage system to the magnesia bed, detracts from the attraction of the magnesia bed technique.