Wastewater Ultraviolet Disinfection 2

This is Part 2 of the article Wastewater Ultraviolet Disinfection. To read Part 1 of the article, please Click Here

Table of Content

• Wastewater Ultraviolet Disinfection 2
  • Practicalities - Ultraviolet Lamps
    • Low Pressure Low Output Lamps
    • Low Pressure High Output Lamps
    • Medium Pressure Lamps
    • Microwave Lamps
    • Pulsed Radiation Lamps
    • Excimer Lamps
  • Practicalities – Hydraulics
  • Open Channel Level Control
    • Extended Length Weir
    • Counterweighted Gate
    • Automatically Controlled Gate
  • Practicalities – Photoreactivation and Dark Space Repair
    • Practicalities – Fouling
  • Calcium Carbonate Deposits
  • Iron and Aluminium Precipitates
  • Activated Sludge Solids
  • Algae
  • Oil and Grease
  • Practicalities – Health, Safety and Environment
  • Practicalities – Validation
  • References
  • Useful Links
  • Related Articles

Practicalities - Ultraviolet Lamps

There are a number of different lamp types (Giller, 2000):

• Conventional low pressure (LP) mercury lamps;
• Low pressure high output (LPHO) mercury lamps;
• Medium pressure (MP) mercury lamps;
• Microwave lamps;
• Pulsed xenon lamps; and
• Excimer lamps.

The various characteristics of the commercialized lamps are detailed in Table below:

Table 3: Typical operating characteristics for lamps (Used with permission from T. Asano & Metcalf & Eddy, Water Reuse (Metcalf & Eddy, 2007e)). Not to be amended.

McGraw Hill makes no representations or warranties as to the accuracy of this information, including warranties of merchantability or fitness for purpose. McGraw Hill will have no liability to any party for special, incidental, tort or consequential damages arising from the use of this information. Copyright ©2007, McGraw Hill Companies, Inc.

Low Pressure Low Output Lamps

The conventional low pressure (LP) mercury lamps were the original UV source. They consist of a narrow tube (approx 15-20 mm dia. and 0.75 to 1.5 m long) of doped quartz, which contains a small quantity of liquid mercury. At each end there is a filament, to provide heating to vaporize the mercury and between which the electric arc is struck. In the lamp an inert gas’ (usually argon) atoms are excited by collisions with electrons. Excited argon atoms collide with mercury atoms and after a series of hits transfer electrons to the mercury atom to produce an excited mercury state. The excited mercury state then decays, emitting light. There is an optimum argon pressure. The optimum lamp wall temperature is 40 ºC, for optimal mercury vapour pressure. The vapour pressure of mercury is low and the tube diameter small so as to limit reabsorption of the emitted light by the mercury vapour. Most of the light is emitted in the UV band, at a wavelength of 253.7 nm. Other wavelengths of light emitted by such a lamp include; 184.9 and 265.3 nm.  The 254 nm wavelength corresponds to the 6³P₁→6¹S₀ transition. The lamps are essentially monochromatic, with 85% of lamp output at 253.7 nm. These lamps are the most electrically efficient of the mercury lamps. Losses include elastic collision with argon in the lamp and loss of heat from collisions with the lamp wall. Maximum power input is limited to 0.5 W/cm lamp length.

The lamp tube is usually manufactured of doped quartz, to absorb the 185 nm wavelength light, which otherwise would react with oxygen in the air, to produce ozone, reducing UV light emission efficiency. Ozone is also toxic to humans. These doped lamps are termed “Non-ozone Producing Lamps”

Low Pressure High Output Lamps

Low pressure high output (LPHO) lamps are a compromise in design to produce higher UV energy output per lamp, compared with the LP lamps. The lamps are similar in design to the LP lamp, except they operate at a higher temperature and slightly lower electrical efficiency. Mercury is replaced with a mercury amalgam spot, which generates a low mercury vapour pressure at the higher operating temperature. LPHO lamps can be operated with a variable electrical energy input, enabling output to be adjusted to meet demand. Power output can be increased by 2 to 4 times that of a LP lamp, with power production of 2 W/cm lamp length. Efficiency of conversion of electricity to UV power is approximately 30%.

The UV energy output of LP and LPHO lamps reaches a peak at 100 hours (the burn-in period), and then declines slowly to a critical point, where output then falls off very rapidly. LP and LPHO lamps have an operating life of 7000 to 9000 hours with electromagnetic ballasts and 10 000 to 15 000 hours with electronic ballasts. Lamp life is shortened by lamp starting and the default design value is four lamp starts per day.

Medium Pressure Lamps

Medium pressure (MP) lamps operate at high temperatures and produce light at a numerous wavelengths, including ultraviolet and visible wavelengths. The much higher temperature and polychromatic nature of the lamps makes them less electrically efficient. The lamps are however capable of much higher electrical energy input per lamp and greater UV energy output per lamp than the LP and LPHO lamps, being suitable for minimizing lamp numbers and complexity for large systems. All of the mercury is volatilized and the quantity placed in the lamp critical. The lamp life is of the order of 3000 to 8000 hours, although extended lives up to 15 000 hours are now being proven, for some existing lamp systems (Jin, 2007). MP lamp systems can be operated on a variable electrical energy input and thus variable UV energy output. A disadvantage is that the lamps must wait if stopped, until all the mercury condenses before restrike.

Investigations have shown that, when medium pressure UV lamps are employed, radiation at higher (visible light) wavelengths causes damage to other than DNA/RNA, which can be responsible for some DNA/RNA repair.

Operation of the MP lamps is inhibited by voltage surges and sags. Reported by one study was that sags of 25% with electromagnetic ballasts typically result in no significant loss of lamp output. Sags of up to 42% will typically result in reduced UV output. Larger sags will extinguish the lamp (Bimble, 2007). For other details of lamp tolerance to electricity fluctuations please refer to (USEPA, 2006e).

Microwave Lamps

Microwave lamps are similar in design to conventional LP lamps, except that instead of wiring for transferring the electricity into the lamp electrode, the energy is inserted by means of microwaves running down a metallic mesh waveguide into the lamp (Gutierrez 2006). There are no electrodes. The microwave energy excites argon atoms which in turn excite the mercury atoms, before emitting UV radiation. The efficiency of the magnetron, producing the microwave energy makes these lamps slightly less electrically efficient, than LPHO lamps. The big advantage with these lamps is that lamp life is around three times the life of a conventional LPHO lamp. Their life is not affected by lamp starts. Other advantages include a faster warm-up time (12 seconds) lower infrared radiation, greater use of the lamp length for the useful UV emission. There is more flexibility in design (gases and pressures), due to no constraints from electrodes. Microwave lamps are an emerging technology for water treatment and presently are commercialized in one brand of UV disinfection equipment (Severn-Trent Services).

Table 4: Comparison of MP and Microwave Electrodeless Lamps (Gutierrez, 2006)

Figure 17: Drawings showing the arrangement of a microwave lamp and magnetron (provided Courtesy Severn Trent Services. Reprinted with permission)

Pulsed Radiation Lamps

Pulsed radiation xenon lamps produce a broad band UV radiation, including UV, visible, and infrared wavelengths. Electrical energy is stored in a capacitor and then released in a very short pulse (less than a second) to produce an ionized gas or plasma. The plasma reaches temperatures of 10 000 K and then releases a very intense pulse of broad band light energy. The spectrum produced is as broad as is the Sun’s spectrum and 20,000 times as bright as sunlight at sea level (Hunter 1998). One comparative study of a pulsed xenon lamp and a LP lamp was reported by Otaki (2003). The inactivation of bacteria and phages was similar, given the UV output of both lamps. The pulsed xenon lamp, seemed to excel with turbid waters. These lamps are an emerging technology.

Excimer Lamps

Excimer lamps produce an essentially monochromatic UV radiation. To generate the radiation a corona discharge is generated in a gas, in a gap between two dielectric materials (e.g. quartz). The corona discharge produces an excited state of the gas in the lamp where two molecules of gas join together to form an excited dimer (excimer). The unstable excimer molecules revert to their normal state, releasing the light energy. The specific wavelength of light produced depends on the gas mixture. Gas mixtures used include xenon (Xe, 172 nm), xenon chloride (XeCl), xenon bromide (XeBr, 282 nm) krypton (Kr) and  krypton chloride (KrCl)(Hunter, 1998). Use of a phosphor can convert the radiation to the germicidal UV-C band.

Features of this type of lamp include, ability to design the lamp vessel to nearly any shape, high power density, mercury free, instantaneous startup, temperature independent operation and a long life (Giller, 2000) These lamps are an emerging technology (Naunovic, 2007).

Practicalities – Hydraulics

Contributions requested. This section is for hydraulics issues that do not relate to computation of the UV dose by CFD. Include these at “Computational Fluid Dynamics”

Open Channel Level Control

Most open channel UV disinfection systems rely on level control to maintain a near constant top water level. For horizontal lamp UV reactors the top lamp must remain submerged, whilst the submergence must not exceed a maximum distance, which is approximately half of the lamp spacing, from top lamp centreline. Vertical lamp units need to have the submergence remain in a narrow band, as this band will experience fouling, due to alternate submergence and exposure of the lamp zone to air, caused by impurities in the wastewater and the heat of the lamps.

Level control systems can be broken down into the following three types:

·        Fixed weir level systems, such as the extended length weir and the pipe weir;

·        An active level control system called the Counterweighted Gate; and

·        An active level control system called the Automatically Controlled Gate.

Extended Length Weir

The extended length weir consists of a single weir edge at the minimum submergence of the lamps. This is the operating level at zero flow through the channel.

Looking down on the plan view of the weir it looks like a castellated shape or consists of a number of side-by-side V shaped vertical walls, which present a single top edge at the minimum submergence. The maximum submergence of the UV unit is the maximum level above the weir edge, plus allowance for head losses, back up through each bank of the UV unit. The maximum submergence of the UV unit occurs at the maximum flowrate through the channel. If the allowable maximum submergence at the weir is h, then the total length of the weir edge required is l.

Now from the Francis Weir formula (in any consistent set of units):

Equation 28

Where:

            l           = the weir length [m]

            Q         = the maximum channel flowrate [m³/s]

            g          = acceleration due to gravity [9.81 m/s²]

            h          = maximum submergence or head over the weir edge [m]

A pipe weir is an alternative form of weir, which consists of a weir edge made up of a number of vertical pipes cut off at the same level at the minimum submergence of the channel. The pipes are welded to a diagonally aligned plate which rises from the channel floor to the maximum submergence of the weir. Walls at the two side walls of the channel enclose the sides of the weir to maximum submergence. Applying the above formula, the total top edge length of the pipes is l [m]. Given a minimum n number of pipes of inside diameter d [m] then (or any consistent set of units):

Equation 29

Counterweighted Gate

The counterweighted gate is a top hinged, bottom opening, rectangular gate across the cross section of the channel at the outlet. The channel level is maintained automatically, given that the geometry of the gate meets certain requirements

Trojan UV uses these types of gates for their larger UV3000 and UV3000Plus, in-channel units.

These gates need to have well lubricated hinges, so as to ensure minimal hunting of water level. Realistically these gates may not meet the fine tolerance of h on water surface level. The gates need to have adjustment of the counterweights on the weight arm, and angle of the weight to the gate, for fine adjustment of the water surface level.

Automatically Controlled Gate

This system consists of a slide gate across the outlet of the channel, sliding up and down, with either a top opening or bottom opening rectangular opening. The gate is driven by a level sensor, inputting to a PID level control loop, which actuates the gate through an actuator.

Wedeco uses this form of system for TAK UV units, using a downwards opening slide gate. Wedeco further can compute the flowrate through the channel, by the water level and gate edge height according to the Modified Francis Weir formula.

Practicalities – Photoreactivation and Dark Space Repair

There are mechanisms in cells which repair damage to the DNA/RNA. These can be divided into light catalysed repair, called photoreactivation, and dark space repair mechanisms.

Photoreactivation is the cleavage of the thymine-thymine (cytosine or uracil) double bonds initiated by enzymes and visible (blue to UV - 350 to 450 nm) light wavelengths (Sancar, 1994). The lower the UV dose administered to the organisms (the less the DNA damage) and the higher the visible light irradiation time the greater is the repair effect of photoreactivation (Lindenauer, 1993).

Harm (Harm, 1980) suggested a two step mechanism for photoreactivation as being:

Step 1: A photoreactivating enzyme (PRE) binds to the Pyrimidine dimer. This process does not require light, but is dependent on temperature pH and ionic strength. The step is reversible, but formation kinetics are favored.

Step 2: Repair of the DNA and release of the PRE. This step is catalysed by light and dependent on light intensity and associated reaction kinetics. The step is normally completed in less than a millisecond. The freed up PRE is free to form another dimer complex.

Photoreactivation of E. coli (Chan, 1995), thermotolerant coliforms, Faecal streptococci and Salmonella (Baron, 1997) Aeromonas salmonicida and Vibrio anguillarum (Liltveld, 2000) was investigated under saline conditions. Photoreactivation commenced, but could not be sustained, and counts fell back to initially inactivated levels after a few hours. Concentrations of seawater of 5000 mg NaCl/L and upwards were responsible. Dose and bacteria type were variables that affected photoreactivation performance. Free somatic coliphages did not repair at all outside a host cell, as they themselves do not possess any repair enzymes.

Photoreactivation of Aeromonas salmonicida, Vibrio anguillarum and Yersinia ruckeri was investigated (Liltveld 1996). The photoreactivation of A. salmonicidia and V. anguillarum was similar to that of E. coli, and requiring 8 hours exposure for completion. Addition of caffeine or yeast extract was shown to be inhibitory to photoreactivation, as for E. coli.

Zimmer (2002), (Zimmer, 2003) investigated photoreactivation of E. coli (ATCC 11229) and Cryptosporidium parvum (Iowa isolate), under UV irradiation under both LP and MP lamps, using a weighted dose calculation for the MP lamp and doses of 5, 8 and 10 mJ/cm² and 1 and 3 mJ/cm² for the two microorganisms respectively. The photoreactivation of the MP lamp irradiated E. coli did not demonstrate any recovery, whilst the LP lamp samples demonstrated 0.7, 2.6 and 2.8 log recovery at the 5, 8 and 10 mJ/cm² doses respectively. There was some photoreactivation of C. parvum at 1 mJ/cm² for the LP UV lamp only. This suggests that, under the right conditions, LP lamps allow photoreactivation, whilst MP lamps do sufficient damage to inactivate photoreactivation. There was essentially no photoreactivation for the MP lamp for a dose of 1 mJ/cm² and 3 mJ/cm² for either UV lamp type.

However one may alternately conclude that the MP lamp does the same damage at a lower weighted dose than that of the LP lamp (Shin 2009).

Oguma (Oguma, 2002) developed the Endonuclease Site Survey (ESS), to count dimers or breaks in the DNA, on E. coli (K12) as a means to determine the number of dimers in an organism and proved that the numbers of ESS was essentially linear with UV dose and corresponded inversely with survival ratio. Legionella pneumophila was found to be almost completely photoreactivated after 3 log reduction, by UV irradiation, by both LP and MP lamps (Oguma, 2004a) By comparison photoreactivation of E. coli (K12) was of a smaller magnitude and was reduced under irradiation by MP lamps, compared with LP lamps. An endonuclease sensitive site (ESS) assay, was used to confirm numbers of dimers and the above photoreactivation response. Oguma et al. went on to use MP lamps with filters restricting the UV light irradiating E. coli (K12) (Oguma, 2005). Only the MP lamp with full spectrum gave little photoreactivation, demonstrating that the full MP UV spectrum inactivates parts of the cell essential for photoreactivation, and it was the broad wavelengths, including those above 300 nm, and not those at either 230, 254 or 300 nm, which took a major part of this role.

Oguma (Oguma, 2004b) found that, under LP UV light, yeast extract, of MW greater than 1000 and less than 3500 Dalton, did not hinder ESS repair, but hindered the colony forming ability to recover.

Kashimada (Kashmida, 1996) looked at E. coli and coliform photoreactivation and hypothesized an exponential recovery relationship. The maximum final survival depends on the ratio of inactivation slope of the photoreactivated maximum survival curve to the slope of the UV irradiation only survival versus dose curve.

Tosa (Tosa, 1999) hypothesized a different relationship, when studying Enterohemorrhagic E. coli. Only slight photoreactivation was observed for EHEC O2O157:H7, but substantial reactivation was observed with EHEC O26.

Nebot Sanz (Nebot Sanz, 2007) developed the Tosa Relationship from first principles. They showed that an S-shaped curve was suitable for representing the photoreactivation process. They produced a family of photoreactivation curves for total and thermotolerant coliforms and Streptococcus faecalis, for tests undertaken in unfiltered secondary effluent. The treatment plant was a conventional activated sludge plant, treating ordinary municipal sewage, using LPHO UV lamps. Photoreactivation was carried out using a Phillips TLD (3.7 W), UV-A lamp; wavelength range 310 to 420 nm with a peak at 360 nm.

The form of the photoreactivation relationship, in terms of survival is:

Equation 30

Integrated this becomes:

Equation 31

Alternatively this can be expressed as an S shaped curve:

Equation 32

Where:

            S          = Survival ratio during the course of photoreactivation 

            S_m        = Survival ratio after infinite photoreactivation 

            k          = Inactivation constant
            t           = photoreactivation irradiation time (minutes)

Nebot Sanz (Nebot Sanz, 2007) determined that dark space repair accounts for a much lower fractional repair of organisms. They proposed an additional mortality constant and term:

Equation 33

Or the complete Dark Space Repair equation is:

Equation 34

Log-linear regression of S_m, k and M series for each of the equations, for photoreactivation and for dark space repair, as a function of the UV dose, yields an essentially linear response, for each type of bacteria and photoreactivation and dark space repair separately. Plots of the photoreactivation, as regressed by the Nebot Sanz equations are presented below:

Figure 18:

Figure 19

Figure 20

Bohrerova (Bohrerova, 2007) investigated the use of four different fluorescent lamps and sunlight for photoreactivating E. coli strain 11229. They found two groups of inactivation rate behavior for the four lamps as a function of the reactivation time. They used the spectra of the photoreactivating lamps, photoreactivation wavelength sensitivity and wavelength specific solution absorbance. They found that, when the above parameters are taken into account, the photoreactivation repair rates, for the different lamps, corresponded, p=0.25.

Photorepair under Durham sunlight was completed in only 15 minutes under the far more intense sunlight. When the light intensity, typical solar spectrum for this period, wavelength specific photoreactivation sensitivity and wavelength were taken into account, the solar photoreactivation repair rates for E. coli were not significantly different from those of the fluorescent lamps (p=0.88, inactivation rate constant =-0.0016 cm²/mJ).

The study demonstrated that completely different photo repair rates are achieved by different lamps or spectra. When the photoreactivating illumination rates are converted to a weighted fluence the different spectra and irradiation intensities enable a closer comparison of photo-reactivation. They recommended that experimenters’ provide full details of the Blue to UV-A spectrum when reporting results. Consideration should also be included, when using sunlight or reactivation light sources containing UV-B radiation, that the lethal effects of sunlight may impact on the viability of microorganisms.

Practicalities – Fouling

Fouling of quartz sleeves or fluoropolymer tubes is a physical or chemical build-up of a substance on the sleeves or tubes, which reduces the transmission of UV light. Fouling of fluoropolymer tubes occurs less rapidly.

The definition of the Fouling Factor is as follows:

Equation 35

The UV intensity measurements need to be taken at the same (water quality) UV% transmittance and UV power input.  The factor is less than unity. Typical values of the fouling factor, used in designing a UV disinfection unit, are:

Table 5: Typical Design Values for the Quartz Sleeve Fouling Factor

 Fouling factors are highly site and condition specific. The operator should know the design fouling factor (F_t) and measure the fouling factor as part of the cleaning procedure, using the built-in UV intensity sensor. If the observed fouling factor (F_t) is less than the design value, the operator needs to clean the unit more frequently. The rate of fouling may change with operation, or season.

What substances can cause fouling:

·        Calcium or magnesium carbonate precipitated on the sleeves or tubes;

·        Iron/alum phosphate/hydroxide precipitation or attachment of same particles, if iron/alum salts are dosed upstream.

·        Activated sludge solids settling out on the sleeves or fluoropolymer tubes at low flowrates;

·        Algae detached from the clarifiers or transfer channels, being caught around or algal cells attaching to the lamp sleeves or tubes;

·        Oil and grease deposits and scum forming solids;

The quartz surfaces have on them negatively charged silanol groups, which can act as sites for adsorption, complexation and precipitation. A study has however shown that the quantity of material adhered generally greatly exceeds that which would correspond to the theoretical density of silanol groups (Lin, 1999a).

Fouling layer mass accumulation increases at zero order, or approximately linearly with time. The different fractions of metals and anions will generally be similar with variation in time.  It was observed, in one case, where lamps were parallel to the flow, that fouling is highest on top surfaces, lowest on underside surfaces and intermediate on side surfaces. This agrees with the hypothesis that there is some gravitational settling of some foulants on top surfaces (Lin, 1999b).

Lin (1999c) postulated that thermal effects were important and that low velocities past tubes would accelerate fouling.

Assuming that fouling rates are sleeve temperature dependent, then fouling rates for different lamp types, in the same water quality, will be in the order:

MP lamps    >>    LPHO lamps    >    LP lamps    >    Microwave lamps

Gehr (Gehr, 2000) carried out the most detailed studies; modeling for UV disinfection of chemically-assisted primary treatment effluents. These effluents are useful starting points for determining mechanisms for secondary effluents. The key study is published by Nessim and Gehr (Nessim, 2006) The fouling mechanisms are complex, and it is worthwhile to repeat the conclusions of the paper here:

They measured both the mass accumulation and UV intensity reduction. The mass rate of fouling did not correlate with the UV intensity reduction. Thus foulants of different composition have differing UV absorbances. The presence of iron resulted in the highest UV absorbances. The presence of iron, or iron and organics resulted in the second highest rate of UV intensity reduction. Calcium and organics came third in terms of UV intensity reduction.

At UV dosages above 35 mJ/cm² three fouling mechanisms were considered to be involved:

(a) Precipitation of Fe(OH)₃, when iron, perhaps with a small quantity of calcium present, when or iron and organics are present;

(b) precipitation of CaCO₃ when only calcium and organics were present; and

(c) ion replacement of calcium-organic complexes to iron-organic complexes, when the wastewater contained calcium, iron and organics.

The presence of low concentrations of iron, even two orders of magnitude lower than calcium, still resulted in predominately iron precipitation. This indicates that the kinetics for precipitation of iron is much stronger than precipitation of calcium. When calcium is moderate (10 mg/L) and iron and phosphate around 1 mg/L, calcium phosphate complexes are formed. High concentrations of calcium result in calcium carbonate and other complexes. When both iron and calcium are both absent, the rate of fouling is much lower.

The authors proposed that the presence or absence of UV light affects the precipitation mechanisms. The authors proposed that the UV energy causes the dissociation of calcium and iron organics complexes. These complexes reform on the quartz sleeve walls, preferentially as ferric hydroxide and iron-organics complexes. Fe(OH)₃ precipitates as a continuous film around quartz sleeves, however the accumulation of CaCO₃ tends to be concentrated at the top, which suggests a sedimentation mechanism for the latter.

High temperatures and high UV doses resulted in higher rates of fouling. This was hypothesized to be due to: (a) precipitation of compounds with inverse solubility, such as CaCO₃ and FePO₄; and (b) UV induced formation of calcium and iron-organic complexes. High flowrates through the UV reactor reduce fouling rates, by reducing temperature increases, as postulated by Lin.

The presence of phosphorus, reduces precipitation in the irradiated section of the reactor. Perhaps high phosphorus and UV together inhibit the precipitation of phosphorus complexes.

Peng and Gehr (Peng, 2005) studied fouling of wastewater UV disinfection units, which employed mechanical and chemical-mechanical wipers. The chemical-mechanical wipers were consistently able to maintain fouling factors at or above 94%. The mechanical wipers resulted in long-term fouling factors around 28-30%. SEM analysis of mechanical wipers revealed scratches on the surface of the quartz sleeve and a tendency to smear the fouling. Fouling commenced at peaks and troughs on the quartz surface. Chemical-mechanical cleaning resulted in some scratching of the surface, but good removal of fouling deposits. SEM-EDX and SEM-XRF revealed the foulants to be aluminium, iron, calcium, magnesium, phosphorus, carbon, and sulfur, sodium, potassium and chlorine. Chemical cleaning by immersion in sodium hydroxide, sodium bisulfite, or phosphoric acid solutions required greater than 10 to 24 hours immersion to remove the foulants. Phosphoric acid was the most effective solution for removing the foulants.

Calcium Carbonate Deposits

Calcium and magnesium ions are present as hardness. Carbonate species are saturated in the effluent from dissolved CO₂ from the activated sludge process. The exact form of carbonate is dependent on the effluent pH;

CO₃^2- + 2H⁺ Û HCO₃^- + H⁺ Û H₂CO₃          Reaction 1

In Reaction 1 – the reaction moves to the left if the H+ is removed from the system (neutralized by hydroxide to form water) if the pH is high. So formation of carbonate is favoured by high pH.

Equation 36

Equation 37

Equation 38

Now K_a1 = 10^-6.42         and K_a2 = 10^-10.43          at 15 ºC (Metcalf & Eddy, 2003b)

Ca²⁺/Mg²⁺ + CO₃^2- Û Ca/MgCO₃                  

Reaction 2

If the above chemistry drives fouling, then:

Formation of the carbonate precipitate is favoured by high Ca, or Mg concentrations (Hardness) and high carbonate ions (thus high pH). As pH is a log₁₀ scale, the carbonate species will increase by approximately 15 to 20 times for every 1.0 pH unit increase.

Reaction 2 is an endothermic (ΔH > 0) reaction so, by Le Chatelier’s principle, high water temperature favours precipitation. This is referred to as inverse solubility. Higher water temperatures are caused by high ambient temperatures and low or zero flowrates with lamps still operating and producing heat.

In fluoropolymer tube units, less heat is conducted to the tubes and the heat is taken away by the air. Thus the fluoropolymer tubes do not get as hot. Thus carbonate-hardness precipitation does not tend to occur.

Calcium carbonate fouling is a, white to fawn (due to solids and metal impurities), cloudy, powdery accumulation on the sleeves, when viewed dry. It is best removed with phosphoric acid, or citric acid. Citric acid is less effective.

Iron and Aluminium Precipitates

Iron phosphate and/or hydroxides particles can settle out on sleeves and tubes, at low flowrates, in a similar manner to activated sludge solids. It is expected that iron (III) precipitates are high UV absorbers.

Excess aluminium and iron ions, as carbonate, phosphate or hydroxide (if used in the process) can be precipitated on the surfaces of quartz sleeves.

These deposits are best removed with acid cleaning. Other stronger acids may be required to shift heavy metallic deposits, such as iron and copper.

Activated Sludge Solids

Activated sludge solids, such as fine particulate solids can settle out on the tops of sleeves or inside fluoropolymer tubes at low or zero flowrates. Optimisation of the activated sludge process may reduce entrained solids.

Fluoropolymer tubes can readily be flushed or brushed to eliminate accumulated solids. A jet nozzle is the best manner to clean these tubes. Quartz sleeves tend to get warmer and this may encourage the deposits to stick to the sleeve.

If these deposits, are present on quartz sleeves, they will tend to be associated with carbonate deposits and so acid is the preferred cleaning method (see above). Otherwise brushing will be necessary to remove solids deposits.

Algae

Algae tends to build-up inside and around the outside walls of clarifier launders, clarifier walls and effluent transfer channels. It is encouraged by warmer temperatures and increased sunlight. The best method of eliminating algae build-up is to keep the launder inside and outside walls clear of algae, by brushing, scraping or hosing, or by covering the areas where the algae to grow, thus preventing visible light getting to the water column.

Strands of algae may wrap around the UV unit and accumulate, inhibiting transmission of UV light to the water. Algae particles can stick to quartz sleeves at low flowrates and higher temperatures. This is not such a problem for fluoropolymer tube units.

Brushing and hosing the sleeves and supports is best for removal of strands of algae and algae accumulation. Otherwise try acid cleaning the sleeves, if associated with carbonate deposits.

Oil and Grease

Oil and grease may arise from carryover of foaming organisms. The lamps may need frequent cleaning with detergent if there is significant solids accumulation on the lamps. Frequency would depend on the rate of solids build-up, remembering the usual fouling design factor is of the order of 0.7 to 0.8.

Practicalities – Health, Safety and Environment

The ultraviolet radiation emitted by UV lamps will burn skin and eyes. Looking at a low pressure lamp, through a face shield and glasses, the visible radiation seems minor, however the invisible UV radiation emitted by this type of lamp is ten times stronger than the visible light emitted. LPHO and MP lamps produce 3 to 50 times more intense UV radiation. The UV light emitted is attenuated by many metres in air or a number of centimeters in water.

Exposure of the skin to a UV lamp without any protection will burn the skin in seconds. The mild case of this is erythma or sunburn. Exposure sensitivity maxima occur at wavelengths of < 240 to 255 and 290 to 310 nm (Masschelien 2002b). Longer exposure can cause blistering of the skin and even bleeding. Exposure of the eyes to a UV lamp will cause “welding flashes” or “arc eye” within seconds. This is burns to the delicate skin just inside the eyeball, which rubs against the eye on blinking. He damage feels like someone has put a shovelful of sand in your eye. A doctor can prescribe cocaine eye drops to provide some relief. Secondary damage can include retinal lesions, yellowing of the lens and cataracts.

The other safety issue is the presence of dangerous electrical high voltages used to start UV lamps and the presence of water that can cause electrocution.

The following precautions are suggested, as a safety standard. Adoption of suggestions depend on the organization’s approach to safety for their site:

• Ensure UV disinfection systems include provision of earth leakage detection (ground fault interrupt) or other electrical protection devices, in            addition to the usual circuit breakers, etc.
• Install lockable UV opaque covers over open channel UV disinfection units, to prevent access by the public.
• Place signs “DANGER - UV RADIATION - WILL BURN EXPOSED SKIN AND EYES” around the unit.
• Do not open closed in-pipe UV units whilst the lamps are on, unless exposure precautions are taken.
• When working around UV channels with automatic level control, and covers are removed, wear a pair of UV absorbing clear plastic glasses or UV absorbing dark glasses. This will protect your eyes if the automatic level control fails and the lamps are exposed.
• Do not remove UV modules from channels whilst still operating, unless precautions are taken to protect skin and eyes. Usually they are interlocked to prevent operation when removed.
•  If it is necessary to remove a UV module from the channel, and allow it to operate (e.g. to carry out intensity or fouling measurements) then cover the skin from head to toe (pull up socks, wear long pants, place rubber gloves, which come up the wrists, on the hands and wear a long sleeved         shirt or jumper, wear a calico hood, UV absorbing glasses AND a face shield to cover the head and face) to protect yourself from UV exposure. Also place warning signs on any approaches to the UV disinfection area “Exposed UV lamps – DO NOT ENTER”.
•  If exposure occurs, obtain medical advice as soon as practicable.
• Lamps and sleeves produce sharp edges when broken. Conventional UV lamps contain mercury (wear rubber gloves when handling broken           lamps).
• Disposing lamps to landfill will contaminate the landfill. Please return your used lamps and sleeves to the lamp or sleeve supplier or a recycler, for  recycling of the quartz and mercury.

The Ozone problem particularly affects the fluoropolymer tube units. You need to specify “Non Ozone Producing” lamps when purchasing lamps. Ozone will be produced, from oxygen in the air, by lamps that are not designed as such, and this will absorb UV radiation. Ozone has a characteristic – “electrical discharge” odour and is a free radical, so it can affect bodily tissue (e.g. the lungs). The peak TWA (in Australia) is 0.1 ppm (0.2 mg/m3 @ 25 ºC). Drager have a tube for measuring Ozone levels, if it arises.

Practicalities – Validation

Contributions requested

N.B.  This article is divided into 2 parts. Please Click Here to read Part 1

References

This is part 2 of Wastewater Ultraviolet Disinfection.  All references to this article relate to both part 1 and  2. For a full list of references , please Click Here

Useful Links

Related Articles

References

Wastewater Ultraviolet Disinfection 1

Wastewater Disinfection

Wastewater Lagoon Disinfection

Wastewater Chlorine Disinfection

Wastewater Ozonation

Wastewater Chlorine Dioxide Disinfection

Other Wastewater Disinfectants

Featured Articles