Etude sur modèle de la désinfection de l'eau par rayonnements ultraviolets Laboratory investigations on the desinfection of water by UV-light

Résumé

Le présent article rapporte les résultats d'une investigation des facteurs géométriques applicables à des réacteurs annulaires à mélange axial complet et appliqués à la désinfection des eaux au moyen de lumière UV de 254 nm. La loi de Bunsen-Roscoe est suivie en prenant en compte la dose corrigée par un facteur géométrique m; dose=mI_ot. Le bactériophage f₂ ne montre pas de phase de latence de mortalité et constitue dès lors un modèle valable. La dose létale à 99 % mesurée est de 470 ± 30 J/m².

Mots-clés:

• UV,
• désinfection,
• élimination des vactériophages,
• évaluation des doses UV,
• coûts de l'UV

Abstract

The f2-bacteriophage is chosen as a test organism to evaluate the geometrical factors which intervene in the disinfection of water by UV-light. This phage is an ARN and single strain cell which has the characteristic of being killed without a lethal lag phase being observed. It is also shown as a representative organism for the estimation of the virucidal action of UV-light.

In this work a cold cathode mercury lamp emitting the 254 nm photon at low intensity, th.e. 14.9 W/m² at the lamp surface, is used. The Lamp has the advantage that its photochemical yield does not depend on the temperature of the water investigated.

The geometrical factor of different annular reactors is investigated by submersing the lamp in water contained in vessels of different diameters. The introduction of an artificial competitor, that is para-hydroxybenzoïc acid, which absorbs part of the 254 nm-photons also allows the impact of the reactor geometry to be evaluated. All the experiments are carried out in batch-type conditions and the water is completely mixed during the experiments. These conditions applied in laboratory investigations are representative of those existing in plug-flow reactors with complete axial mixing of discrete portions of the liquid contained in the annular space between the lamp and the reactor.

In this work all data conform to the Bunsen-Roscoe Law relating the kinetics of first order decay to the irradiation dose : (I.t). No residual resisting organisms or "protected" organisms which could subsist at the end of exhaustive irradiation were observed.

The practical dose, that is the dose at any point of the reactors taking into account the absorption of the light and the increasing surface irradiated at increasing reactor diameter and also the finite dimension of the lamp diameter is accounted for by a single correction factor m : Dose (D) = m.I_o.t, in which, also according to the literature :

       2 r _o (1 - exp [- E (r^- - r_o) ] )

m = ______________________________

            E (r^-2 - r²_o)

where r_o and r^- are respectively the lamp radius and the reactor radius. E is the extinction coefficient (base e).

The data as a whole converge to a 99 % lethal dose of 470 ± 30 J/m² in clear water, either absorbing or not at 254 nm.

In the presence of turbidity, obtained by the introduction of 10 mgL^-1 koalin the efficiency is enhanced by 15 to 20 X, that is the apparent 99 % lethal dose is of 400 ± 30 J/m². By correcting the light intensity transmitted by reflection, this dose is about 420 J/m².

Turbidity in UV-disinfection is thus not necessarily a negative factor in disinfection of liquids with UV-light.

This investigation as a whole establishes an "economical space" between lamp surface and reactor wall of about 4 cm at which the energy consumption for a given level of disinfection attains a sufficiently tow value.

Keywords:

• UV,
• disinfection,
• bacteriophage,
• UV-dose evaluation,
• costs of UV