Microwave Uv: a New Wave of Tertiary Disinfection

WEFTEC®.06

MICROWAVE UV: A NEW WAVE OF TERTIARY DISINFECTION

Richard L. Gutierrez1, Keith N. Bourgeous1, Andrew Salveson2, Jeremy Meir3, and Allan Slater3

1Carollo Engineers 2500 Venture Oaks Way Sacramento, CA 95833

2Carollo Engineers, Walnut Creek, CA

3Quay Technologies Ltd. Cookham Dean, Berks, UK ABSTRACT

Currently UV disinfection technologies available for use in drinking water and water reuse applications typically utilize one of three types of mercury UV lamps. These three types of mercury UV lamps include low pressure, low-pressure high-output, and medium-pressure lamps. All of these lamps contain electrodes that facilitate in the generation of UV radiation. These electrodes are of delicate construction and their deterioration is the primary source of failure in UV disinfection systems. An emerging UV disinfection technology that eliminates the need for electrodes is the microwave powered electrodeless mercury UV lamp. The objectives of this study were to review available literature on commercially available UV disinfection technologies used in drinking water and water reuse applications, and to provide a detailed explanation of the lamp properties and characteristics for comparison with the microwave powered electrodeless UV lamp. A description of traditional mercury UV lamps, the technology and application of microwave powered electrodeless mercury UV lamps, the advantages and disadvantages of the use of the microwave powered electrodeless mercury UV lamps, and a commercially available Microwave UV system that is currently undergoing validation testing as per the NWRI/AwwaRF 2003 Ultraviolet Disinfection Guidelines will be presented in this paper.

KEY WORDS

UV Disinfection, microwave UV, electrodeless UV lamps

INTRODUCTION

The first application of Ultraviolet (UV) radiation for disinfection dates back to the 1890’s with the development of the mercury vapor lamp and quartz tube (Phillips, 1983). However, the development of this technology was not realized with the success of chlorination which was a cheaper, more reliable, disinfection source at the time (Hijnen 2006). In recent years, the use of UV disinfection systems at wastewater treatment plants for the disinfection of tertiary treated effluent has become more common and the technology has reemerged. The impetuses for the increase in the use of UV disinfection systems at wastewater treatment plants are and have typically been 1) stringent new discharge permit limits for trihalomethanes and other chlorine generated disinfection byproducts, and 2) safety concerns over the use of gaseous chlorine. UV disinfection systems have proven to be a safe and effective way to disinfect wastewater without producing the traditional chlorine generated disinfection byproducts. UV radiation disinfects water by damaging the DNA of pathogens such as viruses, bacteria, and protozoa, so that they

can no longer reproduce and are thus incapable of infecting a host. UV disinfection systems typically utilize one of three types of mercury UV lamps. These three types of mercury UV lamps include low pressure, low-pressure high-output, and medium-pressure lamps. All three of these mercury UV lamps contain electrodes that facilitate in the generation of UV radiation. The following topics will be discussed in this paper. First, a description of traditional mercury UV lamps which utilize electrodes and are currently being used in UV disinfection systems will be presented. Second, the technology and application of microwave powered electrodeless mercury UV lamps will be discussed. Third, a description of a commercially available Microwave UV system that is currently undergoing validation testing as per the NWRI/AwwaRF 2003 Ultraviolet Disinfection Guidelines will be presented. Fourth, the advantages and disadvantages of the use of the microwave electrodeless mercury UV lamps will be presented.

MERCURY UV LAMP CHARACTERISTICS AND OPERATION

Most UV lamps, including all lamps used for disinfection, generate radiation through gas discharge. They contain a filling composed of mercury and an inert gas, typically argon, enclosed in UV transparent (no phosphorus) envelopes. An electrical field of one up to 100 volts per centimeter (V/cm) is used to accelerate electrons and cause them to collide with the mercury and argon atoms (Heering 2004). Initially, few collisions occur between the electrons and mercury atoms due to the low vapor pressure of mercury at ambient temperatures. The argon, which has a high ionization energy, is readily excited to a metastable (non-radiative) state and therefore many collisions occur between the electrons and the argon atoms. Argon atoms in the metastable state can only loose energy through collision with other atoms. When metastable argon atoms with high ionization energy collide with mercury atoms, ionization of the mercury atoms occurs. UV radiation is emitted by the ionized mercury atoms as they return to their lower energy state (Phillips 1983).

Mercury is the ideal element for this application for several reasons as described by Phillips (1983): • It has a low ionization energy, enabling the domino effect that occurs between the molecules to happen more readily. • It is the most volatile metal and has sufficient vapor pressure at ambient temperature to provide the optimum conditions for producing resonance radiation. • Its energy state is at a level that produces resonance radiation at useful wavelengths for disinfection (200 to 300 nanometers (nm)). • It is chemically inert to the other materials used in the lamp assembly.

The purpose of the argon is to aid in startup of the lamps by reducing the starting voltage required through the atomic interactions discussed above. Argon is the most commonly used inert gas in the fill material for UV lamps because it is the least expensive (Phillips 1983).

A critical component in the operation of mercury UV lamps is the pressure of the mercury. Operation at low pressures results in insufficient collisions between electrons and mercury atoms, which makes starting the lamps difficult and reduces efficiency. At higher pressures, collisions are increased but efficiency is reduced due to increased elastic collision losses and reabsorption of the UV radiation within the lamp. Reabsorption occurs because the process through which a photon is emitted from an excited mercury atom is reversible and the photon can therefore be reabsorbed by other atoms. The phenomenon makes the selection of the lamp

diameter important since photons emitted by atoms at the center of the lamp have a greater likelihood of being reabsorbed by other atoms as the diameter increases (Phillips 1983).

ELECTRODED UV LAMPS

All of the UV disinfection systems that are listed in the Treatment Technology Report for Recycled Water (December, 2005) that have now or at one time received conditional acceptance utilize electroded mercury UV lamps. There are three types of mercury UV lamps that are conditionally accepted for use in tertiary UV disinfection systems. These three types are the low- pressure (LP) lamp, the amalgam or low-pressure high output (LPHO) lamp, and the medium- pressure (MP) lamp. Although all these lamps use electrodes (see Figure 2a for a picture of an electrode in a LPHO lamp), they differ in several important aspects including their emission spectrum, mercury pressure in the lamp, operating temperature, power requirements, efficiency (defined as the amount of input electrical power that is converted into UV light emitted in the germicidal spectrum range), and lamp life. A summary of typical characteristics for each of these lamps is presented in Table 1.

In all electroded UV lamp systems, electricity from an external power source flows through the electrodes and is conducted directly into the gas discharge. UV lamp electrodes are a delicate construction composed of a coil of fine tungsten wire. Between the coils, there is a mixture of earth oxides (calcium, barium, and strontium), which aid in emission of the electrons. The lamp is configured with an electrode at each end, which take turns acting as a cathode, each with a single connection to the power supply. When the lamp is switched on, the electrodes heat up rapidly and transition from a “glow discharge” to what is referred to as an “arc discharge”. The arc is formed when heating of the cathode results in ejection of electrons through thermionic emission. The lamps require a high starting voltage to initially heat the electrodes (Phillips 1983).

Electrode mercury lamps have negative resistance and thus require a ballast for stable operation. There are two major types of ballasts currently used in UV disinfection systems, the electromagnetic ballast and the electronic ballast. The electromagnetic ballast is the more conventional technology that has been available longer, is more durable, and has a longer rated life. Electronic ballasts, coming into more frequent use, are more energy efficient, allowing UV operation at a wider range of ballast power (and thus lamp intensity). Ballasts can have significant impacts on the power input requirements of electroded UV lamps (Dussert 2005).

Electroded Lamp Life

Electrode deterioration is the most common method of failure for electroded lamps. In normal tungsten filament light bulbs, the mode of failure is usually catastrophic and occurs when the filament breaks and the bulb can no longer produce light. In contrast, with electroded UV lamps, failure is usually caused through deterioration of the electrode through loss of electrode material. There is a gradual loss of output until the lamp is no longer capable of producing sufficient radiation for its intended purpose, in this case disinfection (Phillips 1983). Deterioration is aggravated by frequent on-off switching of the lamps due to a high cathode fall during the “glow discharge” phase when the lamps are switched on. During the glow phase spattering of electrode materials occurs, resulting in deterioration (Phillips 1983).

Table 1 - Summary of Typical Lamp Characteristics for Available Electroded UV Lamps 1 Lamp Type Low Pressure Low Pressure High Medium Pressure Output Hg Pressure (atm) 0.01 0.01 1-2 Amount of Hg (mg) 5-50 35-100 40-400 Operating Temperature (oC) 40 150-200 590-900 Electrical Power (W) 80 200-300 3,000-5,000 Germicidal Power (W) 27 80 300-450 Efficiency (%) 32-40 30-36 12-16 Rated Life (hrs) 8,000-12,000 8,000-15,000 3,000-9,000 1 information presented in this table has been taken from Dussert, 2005; and Phillips, 1983

Description of Low Pressure Mercury UV Lamps

The Low Pressure Mercury (LP) UV Lamp is the most efficient lamp for radiating UV light at the wavelengths required for disinfection. In LP lamps, the mercury is inserted into the filling as a single drop and remains in liquid state during lamp operation. The operating temperature is around 40oC and the mercury is at its vapor pressure, which provides optimum efficiency for production of resonance radiation. Typically, with LP lamps operated at these conditions, about 60 percent of the power input results in resonance radiation at wavelengths of 185 nm and 254 nm, with three percent at other wavelengths, so the radiation output is considered to be relatively monochromatic (Heering 2004). The remaining 37 percent of the input power is lost through elastic collision with argon molecules and with the walls of the lamp. Most of the 185 nm radiation is absorbed by the lamp walls and does not contribute to disinfection. Thus the 185 nm to 254 nm output ratio is of significant interest in optimizing the operation of LP lamps and has been investigated by many researchers (Barnes 1960, Read 1964, Johnson 1971).

The wavelengths of radiation emitted are dependant on the energy state transition of the mercury 3 molecules. Resonance radiations at 254 nm and 185 nm are produced from the transitions 6 P1 1 1 1 → 6 S0 and 6 P1 → 6 S0, respectively. At the operating pressures of LP lamps, the mercury 1 atoms are rarely excited above the 6 P1 energy state, so radiation at other wavelengths is minimal. The 185 nm to 254 nm ratio has been found to increase with current and temperature, and decrease with respect to lamp diameter and the atomic weight and pressure of the inert gas used. The argon pressure is important for several other reasons as well. Low argon pressures will impair starting and cause greater electrode deterioration. High argon pressures will increase elastic collisions between electrons and argon atoms and thus reduce efficiency.

A disadvantage of the LP lamp is that the achievable radiation intensity is low since the optimum discharge conditions can only be maintained at power loads below 0.5 watts per centimeter (W/cm). Low-pressure high output (LPHO) and amalgam lamps use an amalgam and decreased tube diameter, which allows the mercury pressure to remain the same with increased power input of up to 2 W/cm. This allows the specific radiant power to be increased by up to three times that of the LP lamps (Heering 2004).

Description of Medium Pressure Mercury UV Lamps

For mercury lamps to efficiently convert electrical energy to radiation, they require either high electron temperatures or high mercury atom temperatures. At the pressures where LP lamps operate, electron temperature is relatively high. Medium Pressure (MP) UV lamps operate at a pressure region of 1 to 2 atm (Dussert, 2005). In this region, the electron and mercury atom temperatures are the same and thus in a state of “local thermal equilibrium” where the concept of “plasma temperature” applies. The concept of local thermal equilibrium is theoretical and in actuality the system is continually losing energy as the electrons, which are in a slightly higher energy state, transfer energy to the mercury atoms. The plasma temperature for MP lamps is 5000 to 7000K (Phillips 1983). MP lamps have high radiative efficiencies and good spectral efficiency (Heering 2004).

The gas fill in the MP lamp is the same as for the LP lamp, however the mode of operation differs considerably. As discussed previously, in the LP lamp operation the mercury remains in the liquid state and the pressure is controlled by the ambient temperature. In turn, the MP lamp operates with all the mercury evaporated and the pressure of the lamp is therefore controlled by the amount of mercury in the lamp. (Phillips 1983) For this reason, the amount of mercury inserted in the MP lamp is critical and must be carefully calculated. To maintain the required pressures, the coolest part of the lamp must have a temperature in excess of 400oC. The lamps are typically operated at temperatures in the range of 590 to 900 oC (Dussert 2005). The operating voltage for the MP lamp is dependant on the lamp length, the quantity of mercury per unit length of lamp, the lamp diameter, and the power consumption of the lamp. MP lamps are typically 10 to 30 mm in diameter and similar in construction to LP lamps, with electrodes at each end. The lamp tube material is vitreous silica.

A disadvantage of the MP lamp is that although the spectral efficiency is high, much of the radiation it emits is at wavelengths outside the germicidal range. There are many peaks at different wavelengths for the MP lamp in the 200 to 600 nm range. Therefore, its germicidal efficiency is only around 12 percent, much less than for the LP lamps. Lamp life is also much reduced, ranging from 2000 to 8000 hours (Cekic 2004). Another disadvantage is that the MP lamp must be cooled down after operation before it can be reignited so that the evaporated mercury can condense and return to liquid form. This is referred to as “restrike time” which can be as long as several minutes and is determined by the lamp cooling method, lamp length, and the available ignition voltage (Phillips 1983).

Comparison of the Disinfection Efficacy of LP and MP Lamps

As mentioned above, LP and MP lamps vary considerably with respect to their emission spectrums. LP and LPHO lamps are monochromatic, with the majority of the UV radiation emitted at a wavelength of 253.7 nm. MP lamps are polychromatic and emit UV radiation at wavelengths in the 200 to 600 nm range. Lamp output spectra recorded for a LP and MP lamps used in a collimated beam test as part of AwwaRF 2004 at the Metropolitan Water District of Southern California are presented in Figure 1. The germicidal UV spectrum range is documented to be wavelengths from 220 to 300 nm.

Figure 1 – Lamp Output Spectra Recorded for Low Pressure and Medium Pressure Lamps (Modified from Malley 2004).

ELECTRODELESS MICROWAVE UV LAMPS

Microwave powered electrodeless mercury UV lamps are actually not a new source of UV light but were conceptualized over forty years ago for industrial processing applications (Cekic 2004). Microwave UV lamp curing systems for inks and coatings have been very successful due to their efficient production of UV light with less infrared heating than other conventional UV sources. The market for microwave UV lamps used in curing is estimated to be close to $100 million annually (Osepchuk 2002). One of the reasons for the commercial success of these lamps is its rapid “restrike” capability and the lack of influence of on-off switching on the life of the lamp. However, only recently has this technology been applied for the disinfection of water.

Microwave powered electrodeless UV lamps differ from the electroded lamps in that there is “no direct electron flow from the external driving circuitry through the discharge plasma” (Cekic 2004). Instead of utilizing electrodes, microwave energy is generated by magnetrons and directed through wave guides into the quartz lamp sleeves containing the gas filling. The directed microwave energy excites the argon atoms, which in turn excite the mercury atoms to produce radiation as they return from excited states to states of lower energy, as is the case with other mercury UV lamps. Electrodeless lamps operate at higher pressures than MP lamps, in the range of 5 to 20 atm (Phillips 1983) compared to 1 to 2 atm for MP lamps (Dussert 2005). A comparison of the physical characteristics of the MP lamp and electrodeless lamp are provided in Table 2.

The irradiance of the UV light emitted can be either polychromatic or monochromatic, because the spectrum emitted can vary depending on the filling material pressures (which are a function of the energy input into the lamp) and on the filling material used (as discussed above). Microwave UV lamps allow greater flexibility for variations in parameters such as lamp diameter, operating pressures, and fill materials due to the absence of electrodes. This allows for greater optimization of radiation at specific wavelength regions (Fassler 2004). The intensity of the radiation increases when the applied microwave power is increased (Horikoshi 2002). The spectral output of a commercially available electrodeless microwave UV lamp is similar to that shown in Figure 1 for the LP lamp, with maximum intensity occurring at the 254 nm wavelength (personal communication with Allan Slater, Quay Technologies Ltd.).

Table 2 – Summary of Typical Lamp Characteristics for MP and Electrodeless Lamps. Lamp Type Medium Pressure Microwave Electroded Lamp Electrodeless Lamp Hg Pressure (atm) 1-2 5-20 Amount of Hg (mg) 40-400 10-15 Operating Temperature (oC) 590-900 40-60 Electrical Power (W) 3,000-5,000 1,000 Germicidal Power (W) 300-450 200-300 Efficiency (%) 12-16 20 Rated Life (hrs) 3,000-9,000 27,000

A picture of an electrodeless lamp of a commercially available UV disinfection system (Quay Technologies, Limited) that is used for tertiary disinfection is presented in Figure 2B. Four of the electrodeless lamps are placed together and held in the proper position by the white lamp brackets shown in Figure 3A. A wave guide (see Figure 3B) that directs the microwave energy around the lamp bundle but allows the UV light to escape, is placed inside the quartz sleeve with this four lamp bundle (see Figure 3C). As can be seen in Figure 4A, four quartz sleeves each with a wave guide and a four lamp bundle are mounted vertically on a rack. This rack is then positioned in the UV channel. After the rack has been placed in the UV channel, a 1000 W magnetron is mounted on top of each quartz sleeve assembly (see Figure 4B). The magnetron serves to supply the necessary microwave energy to cause the UV light generation in each four- lamp bundle. The number of racks that are included in each UV channel is sight specific and is a function of the flow, UV transmittance and the required delivered dose. A top view of the path that the microwaves follow in the quartz sleeve is presented in Figure 4C. A picture of the lamps in operation with the quartz sleeves only partially submerged is presented in Figure 4D.

Electrode A

Figure 2. View of an A) traditional low pressure high-output lamp with an electrode and B) an electrodeless low pressure high-output lamp used on a Microwave UV system.

A B C

Figure 3. View of an A) 4 lamp bundle, B), a wave guide and C) quartz sleeve of the microwave UV system with the wave guide and 4 lamp bundle inside.

A B)

Magnetrons

4 electrodeless lamps

Figure 4. View of an A) 4 quartz sleeve rack, B) rack in channel with magnetrons C) top view of lamps in each quartz sleeve, and D) partially submerged microwave UV electrodeless lamps in operation.

Magnetron Technology

All microwave powered UV lamps depend on the magnetron as the microwave power source. Critical criteria for a power source are efficiency, low cost, power, and reliability. Although its reliability has been a concern in the past (Phillips 1983), the magnetron is currently the best available technology for supplying microwave power. The number of available sizes and types of magnetrons is very limited. The most common type is “cooker magnetron”, which was the breakthrough that lead to the microwave-oven revolution. It is available for powers of approximately 0.5 to 5.8 kW at 2.45 GHz frequency (Osepchuk 2002). Microwave powered UV lamps utilize magnetrons based on the cooker magnetron, which operate at 900W. The efficiency of the 2M244 – 900W magnetron, utilized by Quay Technologies Ltd., is 70 to 72 percent once the input power is raised to above 30 percent of the total power (Slater 2006). The magnetrons weigh approximately 2 pounds each and employ inexpensive ferrite magnets, a simple filter box around electrode leads, and have a sales price of approximately $10.00 U.S. Suppliers exist worldwide including China, Japan, Korea, Thailand, and Russia, with production of cooker type magnetrons estimated to be on the order of 30 to 40 million per year worldwide (Osepchuk 2002).

Advantages and Disadvantages of Microwave Powered Electrodeless UV Lamps

Phillips (1983) has listed the advantages of using a microwave powered electrodeless UV lamps over conventional electroded lamps as the following:

• The lamps warm up quickly and are capable of disinfection within 12 seconds compared to start up times of 20 seconds to up to three minutes for electrode lamps. It takes 10 seconds to warm up the magnetron filaments and an additional two between ignition and full radiation output. If the magnetrons are kept warm in standby mode, the ignition time is only about two seconds.

• Eliminating the electrodes from the lamp, eliminates the primary deterioration process associated with UV lamps, resulting in a lamp life approximately 3 times that of electroded lamps.

• Eliminating the electrodes allows for narrower lamps which reduces the amount of reabsorption as well as the heat capacity and infrared radiation generated.

• The lamp has very low residual radiation of energy and thus almost instant shut off capability. This prevents overheating of heat sensitive materials near the lamps.

• There is no energy loss associated with electrodes.

• The gas discharge fills the entire lamp column in an electrodeless lamp as opposed to an electroded lamp which requires space at the top and bottom for the electrodes. Therefore radiation is produced through the entire length of the lamp.

Disadvantages of using the microwave powered electrodeless UV lamps as listed by Phillips (1983) include the following:

• The electrodeless lamp system has more components than the conventional electroded system, including the magnetron, wave guides, and cooling fans for the magnetrons.

• The wire mesh results in attenuation of some of the radiation emitted by the lamps. Approximately 90 percent of the area of the mesh screen is open area.

• Magnetron life is limited and requires replacement. Magnetrons have been warranted for up to 10,000 hours of operation by Quay Technologies Ltd. (Slater 2006).

REFERENCES

Barnes, B.T. (1960) Journal of Applied Physics, Vol. 31, 852-854.\

Bergmann, H.; Iourtchouk, T.; Schops, K.; Bouzek, K. (2002) New UV Irradiation and Direct Electrolysis - Promising Methods for Water Disinfection. Chemical Engineering Journal, Vol. 85, 2002, 111-117. Cekic, M.; Ruckman, M. (2004) Physics of Electrodeless UV Lamps and Applications of UV Radiation. The Physics of Ionized Gases: 22nd Summer School and International Symposium, CP740. Dussert, B.W. (2005) Essential Criteria for Selecting an Ultraviolet Disinfection System. AWWA Journal, July 2005, Vol.97, No.7, 52-58. Fassler D.; Drewitz A.; Thomas Ch.; Henne, M.; Wipprich W.; Meyer, A.; Johne, S. (2004) New Sources for the Short UV - Design, Emission, Actinometry. UV-Karlsruhe 2004 Conference, Germany. Heering, W. (2004) UV Sources - Basics, Properties and Applications. UV-Karlsruhe 2004 Conference, Germany. Hijnen, W.A.M.; Beerendonk, E.F.; Medema, G.J. (2006) Inactivation credit of UV radiation for viruses, bacteria and protozoan (oo)cysts in water: A review. Water Research, 40 (2006), 3-22. Horikoshi, S.; Hidaka, H. (2002) Environmental Remediation by an Integrated Microwave/UV Illumination Technique. 3. A Microwave Powered Plasma Light Source and Photoreactor To Degrade Pollutants in Aqueous Dispersions of TiO2 Illuminated by Emitted UV/Visible Radiation. Environ. Sci. Technol. 2002, 36, 5229-5237. Johnson, P.D. (1971) Appl. Phys. Lett. 18, 381-382. Malley, J.P.; Ballester N.A.; Margolin, A.B.; Linden, K.G.; Mofidi, A.; Bolton, J.R.; Crozes, G.; Laine, J.M.; Janex M.L. (2004) Inactivation of Pathogens with Innovative UV Technologies. Ammerican Research Foundation and American Water Works Association, 2004. Osepchuk, J.M. (2002) Microwave Power Applications. IEEE Transactions on Microwave Theory and Techniques, Vol. 50, No.3, March 2002. Phillips, Roger (1983) Sources and Applications of Ultraviolet Radiation, Academic Press, London. Read, T.B. (1964) British Journal of Applied Physics. Vol. 15, 837-841.

Slater, A. (2006) Personal communication, Quay Technologies Ltd., Cookham Dean, Berks, UK

Treatment Technology Report For Recycled Water (2005) State of California Division of Drinking Water and Environmental Management. Ultraviolet Disinfection Guidelines for Drinking Water and Water Reuse (2003) Second Edition, National Water Research Institute and the American Water Works Association Research Foundation.