01 March 2018

Kitchen Exhaust Filtration Design

Commercial Kitchen

Complete 2 part series on kitchen exhaust design

Part 1 of 2-part series on kitchen exhaust design

This article is split between two consecutive issues. Part 1 (Ecolibrium March 2018) looks at kitchen exhaust filtration and discusses current Australian Standards compared with current benchmark practices within the United States. Using first principle analysis, critical design parameters are provided in this article for kitchen exhaust filtration design in Australia.

Part 2 (Ecolibrium April 2018) reviews the likely performance of common filtration methods for kitchen exhaust, and aims to uncover economic and effective choices for kitchen exhaust applications that will achieve the design performance parameters established in part 1.

The requirement to effectively filter kitchen exhaust discharge to a non-objectionable effluent is increasing due to health concerns and odour complaints. Research out of the USA has shown that in the past emissions from barbeque grills in New York have been the sole attributor to 400 deaths annually [1]. Filtering kitchen exhaust could no longer be solely an aesthetic or comfort issue. But how do we approach kitchen exhaust filtration?

Why Filter Kitchen Exhaust?

Within Australia, there are standards that govern kitchen ventilation including council and state regulations, local, state and federal fire codes and most significantly AS/NZS 1668.1 [2] and AS 1668.2 [3] – all of which are referred to by the NCC (National Construction Code) for deemed to satisfy solutions.

AS 1668.2:2012 identifies all kitchen exhaust over 1,000 L/s as “Type B Objectionable Effluent” and in addition to the exhaust discharge requirements for other exhausts, requires that this is then:

  • Discharged vertically
  • Discharged at no less than 5 m/s
  • Discharged at least 6 metres away from boundaries (including street boundaries), outdoor air intakes and natural ventilation openings.

Failure to satisfy all of these requirements will result in a system that does not comply with AS 1668.2:2012.

These requirements aren’t always practical or cost effective though, particularly when working with existing buildings. For these situations the NCC does make a performance solution pathway available to us, with a verification process prescribed for soundly engineered designs. AS 1668.2 offers some non-mandatory commentary to guide designers toward potential solutions.

The NCC and AS 1668.2:2012 also require designers to consider situations where exhaust may pose a danger or nuisance. Exhaust filtration systems can help manage these risks.

Whilst one of Australia’s current motives for kitchen exhaust filtration is to avoid nuisance complaints and provide odour free air discharge, it is of interest that in the USA (where kitchen exhaust regulation and technology is typically more advanced), certain places enforce the filtration of kitchen exhaust to improve ambient air quality and therefore the health of the population.

A study from Department of Health and Mental Hygiene in New York estimated that emissions from char broilers (grills) in New York contributed to more than 12.5% of PM2.5 attributable deaths annually in the period 2005-2007. This equates to 400 deaths per year [1]. If all char broilers had effective kitchen exhaust filtration installed and maintained, it is estimated nearly 350 of these deaths could have been prevented [1]

Another study completed at a similar time in New York suggested that 20% of all PM2.5 particles in the air are from commercial cooking, this is more than the amount attributed to on road vehicles at 17% [4].

emissions sources diagram

Figure 1: PM2.5 Emissions Sources in New York

Separately, researchers at the University of California estimated that “the average diesel-engine truck on the road today would have to drive for 10 miles (16km) on the freeway to put out the same mass of particles as a single charbroiled (grilled) hamburger patty. [5]”

This in turn led to an amendment to New York law, effective as of May 6, 2016, that prohibits the operation of any new commercial char broiler (grill) and any existing chain driven commercial char broiler (grill) to cook more than 875 pounds (400kg) of meat per week unless it has an emissions control device that meets the requirements established by the Commissioner of the Department of Environmental Protection (DEP) [1].

What is Suitable Filtration for an Engineered Solution?

Under AS1668.2 the only “deemed to satisfy” solution is to provide an exhaust that complies with a type B objectionable effluent as described above, in addition to all the requirements for non-objectionable discharges.

However due to increasing urban density, and available space for mechanical services within buildings, exhausting vertically from the roof more than 6 metres away from a boundary, it is often not possible.

To overcome this, an engineered solution might include a filtration system to clean the air to a “required” level. The definition of this “required” level is unclear, whilst some recommendations are present as commentary after clause 3.10.3 in AS1668.2, this still doesn’t give us a clear deemed-to-satisfy solution. It also doesn’t necessarily give us a concrete, repeatable engineered solution that will always be acceptable to neighbours, councils and building certifiers, especially given the variability in contaminant level of different commercial kitchen exhausts.

As an industry, we need to define a clearer picture of what the requirements for kitchen exhaust filtration are in Australia. Are we just concerned about odour, the visual eyesore of smoke, or the health of the general population (and PM2.5 related deaths)?

As we shall see, the most cost-effective method can actually address all three concerns!

The Science behind Kitchen Exhaust Filtration

Kitchen exhaust filtration can be broken up into two clear aspects; particulate filtration – which in this case includes grease, smoke and free moisture, and gaseous filtration which includes odours.

The first rule for kitchen exhaust filtration is that effective and sustained gas phase/odour filtration cannot be achieved without an adequate level of prior effective particulate filtration.

Additionally, particles (such as smoke) are also odourous. Therefore, to ensure that odours are removed to a reasonable level, a significant level of particle filtration also needs to be in place.

It turns out that the level of particle filtration required to produce cost effective odour filtration, is also the level required to eliminate smoke as a visual pollutant, and significantly reduce PM2.5 emissions.

The empirically successful number used in the USA and Asia with success, is a minimum particle filtration efficiency of 95% at 0.3μm (micron).

Why 0.3μm?

The rating of the performance of filters for kitchen exhaust filtration at the particle size of 0.3μm is for two reasons:

  • 0.3μm is known as the most penetrating particle size (MPPS)
  • 0.3μm is a common particle size for smoke

Most Penetrating Particle Size (MPPS)

The particle size of 0.3μm (MPPS) is typically selected as the test point for rating filtration efficiency. This is because particles above and below this size (0.3μm) are generally easier to capture [6].

The science behind the capture of particles on a traditional HVAC filter illustrates why 0.3μm is the most penetrating particle size. There are three main filtration mechanisms – impaction, interception and diffusion.filtration mechanisms models
Figure 2: Filtration Mechanisms

The filtration mechanisms are best explained using a HEPA filter as an example. Despite having a filtration efficiency of 99.97%+ at 0.3μm, HEPA filters don’t operate as a sieve where only particles smaller than the smallest gap will pass through. Instead they comprise of a mesh of fibres with gaps between fibres over 1μm not uncommon [7], [8].  In a simplistic way, instead of acting as a sieve the HEPA relies on the probability that as a particle passes through the fibres, they will either impact onto a fibre (A) or come within 1 particle diameter of a fibre and be drawn onto a fibre through interception(B).

Naturally as the particle size decreases, the probability of a particle passing through the filter without impaction or interception increases. However, as the particle size decreases below 1μm, the particle undergoes diffusion (C), otherwise known as Brownian motion [6].

Brownian motion is where the particle is becoming closer in size and mass to gaseous molecules in the air. This causes the effect of collisions and interactions of the particle with gas molecules to increase, causing the particle to move radially to the direction of bulk airflow. This results in a significant increase in residence time and randomness of path for the particle though the filter; and therefore, far greater opportunity for the particle to be captured in the filtration process.

Figure 3 shows the overall minima in fractional efficiency, which is between 0.1-0.4μm. This is where the improvements in filtration due to diffusion are yet to offset the reduction in interception and impaction due to smaller particle sizes.

Most Penetrating Particle Size-Fractional Efficiency by Particle Size
3: The impact of diffusion, interception and impaction on overall filtration efficiency.

Similarly, for an electrostatic precipitator (ESP), the filtration efficiency decreases when the particle size decreases as the amount of charge held by the particle decreases. However, when the particle size decreases past 1μm, this reduction in charge held is offset by the increased residence time in the ESP from the effects of diffusion and the minimum at 0.3μm remains.

We therefore cannot assume that an efficiency specified at 0.01μm indicates a higher efficiency for all particle sizes larger than 0.01μm. We want to ensure we are testing products at their worst performing point – 0.3μm – and this principle may be applied to many types of particle filtration technologies.  If you want to compare apples to apples, always compare system efficiency at a particle size of 0.3μm.

The EN1822:5 (2009) [9] is widely recognised as the best standard for testing a product fit for purpose at 0.3μm.  The ASHRAE 52.2 (2017) [10] or EN779 (2012) [11] test products over the range of 0.3-10μm, so extracting the performance of the product at 0.3 micron from these tests is also reasonable.

The ASHRAE 52.1 (1992) [12] standard is an obsolete standard, ASHRAE have since released various versions of ASHRAE 52.2 to supersede 52.1 which reflects the growing understanding of filter testing methodology. ASHRAE 52.1 (1992) dust spot efficiency tested the filters performance at approximately 0.7-0.8μm [13]. It has been shown mathematically that a filter with an ASHRAE 52.1 dust spot efficiency of 90-95% would test 65% efficient at 0.3μm [13], a critical difference given the importance of removing particles of this size for kitchen exhaust filtration.

It is also important to check that the velocity and/or airflow in the filtration system designed is the same as the test report, as most filtration products, especially ESP’s, drop significantly in filtration performance as the velocity of air increases.

Smoke particles

Smoke particles are normally 0.3-1μm in size [6], [14], and is another significant reason why 0.3μm is chosen to compare system efficiency within kitchen exhaust filtration systems.  A system that delivers poor efficiency at 0.3μm will lead to poor filtration of smoke, a common visual and odourous pollutant present in kitchen exhaust. Smoke can also be a significant health risk when present in ambient air.

Figure 4 shows the mass of particles generated when cooking bacon on a gas stove. There is a significant mass of particles, mostly smoke, generated at 0.3μm, which is particularly noticeable when the food is cooked at a high heat. The 2nd peak of particles at 7-10μm is likely comprised of fine grease aerosols [15].

particle mass distribution during cooking
Figure 4: Particle mass distributions during the grilling of 50g of bacon on a gas stove [14]

Why 95%?

The rating of the performance of filters for kitchen exhaust filtration of 95% at the particle size of 0.3μm is done for three reasons:

  • 95% filtration of 0.3μm particles ensures the air is mostly clear of smoke and almost entirely clear of particles of a smaller and larger size as discussed previously with impaction, interception and diffusion.
  • Common odour control/gaseous filtration mechanisms don’t perform effectively or economically when there are too many particles present in the air.
  • 99.97% at 0.3μm whilst better for performance, poses economic and footprint issues. 95% is regarded as the sweet spot for a high level of filtration before extensive design changes are required.

Gaseous filtration and particles

Below is a brief description of 2 common methods of gaseous filtration used to control odour, and how poor prior particle filtration will render them ineffective. Part 2 in the series on Kitchen Exhaust Filtration Design will go into further detail on these technologies.

Adsorbent/chemisorbent media

Adsorbent media, commonly activated carbon, consists of micropores within larger granules. Small odorous gas molecules find their way inside these pores where various forces adsorb the gas molecules to the media, therefore removing them from the airstream. If the media granule is subject to particle contamination, the outer surface of granule will become coated creating a seal. This results in a poor performance due to the unused pores being inaccessible and going to waste.

diagram of absorption of gas molecules
Figure 5: The adsorption of gas molecules within the pores of activated carbon


Ozone works by breaking down molecules into smaller and smaller molecules with oxidation until eventually simple, common odourless gases are left such as CO2, N2, H2O and O2 etc.

When relatively large particles such as grease and smoke are present in the air which are 100 to 1,000,000+ times larger [15] than odorous gas molecules, the ability of the ozone to break down the small odour molecules can be reduced. This is because the oxidative power of ozone will be absorbed by large particles which may not be fully broken down to a simple odourless gas, particularly given the short residence time in kitchen exhaust filtration and available concentration of ozone. The tiny odorous molecules are therefore effectively “hidden” and can remain in the airstream along with the particles which are not fully broken down.


Nuisance odours are not the only factor when considering options to improve the ambient air quality from discharged cooking fumes. The health of the general population is now also being taken into consideration. By understanding the first principles of particle and gaseous filtration, we can make better decisions regarding kitchen exhaust filtration applications to form an economic and effective solution.

The World Health Organisation estimates that approximately 3 million deaths are attributable solely to ambient (outdoor) air pollution each year [16]. Kitchen exhaust has been identified in poor ambient air quality. Increasing awareness of the effect of poor ambient air quality on human health could see filtration of kitchen exhaust discharges becoming increasingly important in the future.

Part 2 in next month’s Ecolibrium uses the first principles of filtration reviewed here to analyse some common kitchen exhaust filtration technologies including hood filters, ultra-violet (UV), multi-stage filter packs, electrostatic precipitators (ESP’s), activated carbon and ozone. The focus will be placed on how we can economically and effectively achieve a particulate filtration efficiency of 95% at 0.3μm with sufficient downstream odour control.

Written by:

Jonathan Bunge, National Product Manager, Airepure Australia Pty Ltd

Technical review by:

Brett Fairweather, Mechanical Engineering Consultant, It’s Engineered Pty Ltd

Peter Mathieson, Technical Director, Aurecon Australia Pty Ltd


Jonathan Bunge is a graduate of Melbourne University with a Masters in Chemical Engineering, and has gained engineering experience within petrochemical facilities, the coal industry and the gaseous filtration industry. His role with Airepure Australia enables him to specialise in the control of gaseous emissions, hazardous and pungent odours, laboratory fume exhaust solutions as well as commercial cooking and industrial process exhaust solutions.

Brett Fairweather is a Mechanical Engineering Consultant with 20 years of mechanical services experience in the design and delivery of projects. He is committed to training and support of other practitioners in the industry and volunteers his time to many committees responsible for the ongoing development of Standards, Codes and Regulations relating to building services.

Peter Mathieson is a Mechanical Engineer with over 30 years of mechanical services design experience and over that period has designed many kitchen exhaust installations, many with discharge filtration. His design of kitchen ventilation has improved with experience and is with reference to the current ASHRAE research.


[1] Department of Enviromental Protection, Section 1403 (c) of the New York Charter and Sections 24-105 and 24-149.4 of the New York City Administrative Code., New York: Department of Enviromental Protection, 2016.
[2] Standards Australia, AS/NZS 1668.1, SAI Global, 2015.
[3] Standards Australia, AS1668.2, SAI Global, 2012.
[4] N. Y. Health, “New York City Community Air Survey,” New York, 2015.
[5] L. Bose, “Controlling the Emissions of Charbroiled Burgers,” 18 September 2012. [Online]. Available: https://ucrtoday.ucr.edu/8896. [Accessed 11 May 2017].
[6] M. K. Owen and D. S. Densor, “Airborne particle sizes and sources found in indoor air.,” Atmospheric Environment., vol. Part A., no. General Topics, 26(12), pp. 2149-2162., 1992.
[7] T. F. Scientific, “Fibremetric – SEM image of hepa filter,” Thermo Fisher Scientific, [Online]. Available: https://www.phenom-world.com/software/fibremetric. [Accessed 29 January 2018].
[8] A. Mason, S. Wylie, A. Thomas, H. Keele, A. Shaw and A. Al-Shamma’a, “HEPA Filter Material Load Detection Using a Microwave Cavity Sensor,” International Journal on Smart Sensing and Intelligent Systems, vol. 3, no. 3, pp. 322-337, 2010.
[9] BSI, BS EN 1822-5:2009 High efficiency air filters (EPA, HEPA and ULPA). Determining the efficiency of filter elements, BSI, 2009.
[10] ASHRAE, ASHRAE 52.2:2017 Method Of Testing General Ventilation Air-Cleaning Devices For Removal Efficiency By Particle Size, ASHRAE, 2017.
[11] BSI, BS EN 779:2012 Particulate air filters for general ventilation. Determination of the filtration performance, BSI, 2012.
[12] ASHRAE, Standard 52.1-1992 — Gravimetric and Dust-Spot Procedures for Testing Air-Cleaning Devices Used in General Ventilation for Removing Particulate Matter, ASHRAE, 1992.
[13] Camfil Farr, ASHRAE Testing for HVAC Air Filtration – A Review of Standards 52.1-1992 & 52.2-1999, Camfil Farr, 2002.
[14] G. Buonanno, L. Morawska and L. Stabile, “Particle emission factors during cooking activities,” Atmospheric Environment 43, vol. 43, p. 3235–3242, 2009.
[15] The Australian Institute of Refrigeration, Air Conditioning and Heating, Fire safety – Kitchen hood exhaust systems Understanding and addressing the special fire risks inherent in commercial kitchen ventilation systems, AIRAH, 2016.
[16] N. Mehio, D. Sheng and J. De-en, “Quantum mechanical basis for kinetic diameters of small gaseous molecules,” The Journal of Physical Chemistry, vol. 118, no. 6, pp. 1150-1154, 2014.
[17] W. H. Organization, “Ambient air pollution: A global assessment of exposure and burden of disease.,” 2016.

Part 2 of 2 part series on kitchen exhaust design

Part 2 in this series analyses various filtration methods seen for kitchen exhaust as supported by scientific analysis. This article aims to uncover what are the economic and effective choices for any given kitchen exhaust application to achieve the proposed design parameters from Part 1.

Different cuisine types produce varying amounts of moisture, grease, smoke and odour and the resulting cooking fumes comprise a combination of solid particles, liquid droplets and vapour / gaseous phase contaminants. Various kitchen exhaust treatment technologies are available to choose from – each with pros and cons affecting cost and performance.

The technologies analysed in this article include kitchen exhaust hoods, hood filters, ultra-violet (UV), multi-stage filter packs, electrostatic precipitators (ESP’s), gas filtration media and ozone.

Kitchen Exhaust Hood

The major function of a kitchen exhaust hood is to ensure all the cooking fumes and heat generated by cooking processes are captured and removed by the hood.

AS1668.2 prescribes a minimum exhaust airflow rate through kitchen hoods depending on the size of the hood, the type of hood and the appliances under the hood. Alternatively, AS1668.2 allows a more detailed calculation based on VDI 2052, DIN18869 or other tested engineered solutions [1]. These are often based on the actual amount of heat and plume volume generated from specific kitchen appliances. These tailored calculations can significantly reduce the airflow required to be exhausted and therefore the capital and operating costs of kitchen exhaust systems.

Effective treatment of kitchen exhaust using any method requires air temperature in in the duct to be under 50°C. At higher temperatures particulates can pass through the filtration system in the gas phase, overpower gas filtration systems and condense back into a particulate matter when exhausted to the colder outside air. This is normally not a problem if the hood is designed as per AS1668.2 or any other allowable standard. This is due to the amount of dilution air required reducing the temperature in the exhaust to below 50°C. However, care should be taken to consider the actual expected temperature in the exhaust of any design.

If the temperature in the exhaust duct where treatment is required is expected to exceed 50°C, (common if there is a solid fuel application) additional measures should be taken. A high temperature is one of the most likely reason for failure of kitchen exhaust filtration. Solutions include the incorporation of a cold-water mist injection hood, in-duct wet scrubber or any other form of airstream cooling prior to filtration.

 Hood filtration

The first and often the most underutilized part of a successful kitchen exhaust filtration system is effective particle load reduction at the hood. Removing a significant load of particles at the hood, will reduce the maintenance costs on equipment downstream.

Metal Mesh/Baffle Filter

As grease laden air passes through metal mesh/baffle filters with various obstructions to the airflow, the air constantly changes direction. The resultant change in inertia causes the grease to be separated from the air and onto the metal surface. All metal hood filters vary in the way they obstruct the airflow, but generally the greater the obstruction, the greater the pressure drop and efficiency. Baffle filters can have an advantage as they are often self-draining and rigid as opposed to honeycomb/mesh filters which collect and hold grease in place and can be harder to clean.

aerosol particle size graph[9]

Figure 1 shows common hood filters and their efficiency between 1-10μm. A similar diagram showing the efficiency of hood filtration between 1-10μm should be available and requested from suppliers. Whilst hood filters don’t offer much filtration at 0.3μm (smoke), they can significantly reduce the load on downstream filtration and the frequency of duct cleaning required.

Another important aspect of the hood filter is fire safety. According to AS/NZS1668.1:2015 clause 6.2.9 [2], if you have kitchen exhaust ductwork that is longer than 10 metres inside the building and you have a flame under the hood, either from solid fuel or gas, you must have a UL1046 rated filter in your hood.

The UL1046 test determines the abilities of hood filters to “Limit the projection of flames downstream when subjected to flames on the upstream face, after having been loaded with grease in a manner representative of cooking that produces a grease-rich exhaust” [3]. Therefore, a mesh/honeycomb filter which works by holding and capturing grease in place may not be effective, as this grease can be used as a medium for the fire to pass through. Instead look for a self-draining filter with plenty of obstructions to the airflow, as these obstructions can also limit the passage of flames. Please check with your hood filter supplier and ask for a UL1046 certificate if required to ensure compliance.

Despite baffle filters being more expensive and having a higher pressure drop, they should be considered for fire safety, ease of cleaning and their higher efficiency which lowers overall operating costs of the kitchen filtration system. Maintenance documentation for these designs should clearly identify the purpose of the selected filters, avoiding future substitution with inadequate alternatives.

Ultra Violet Light (UV)

UV technology uses the energy from the UV light to undertake a process called photolysis, which is the decomposition or separation of molecules by the light. The aim is to decompose and breakdown grease, smoke and odour until eventually simple, common odourless gases are left such as CO2, N2, H2O and O2. However as discussed previously in Part 1, the grease and smoke particles are often simply too large and the concentration too high to break down to simple odourless gases given the short residence time in kitchen exhaust filtration and the power of UV lights. Figure 1 illustrates this. Additionally, UV lights can become coated in grease and smoke which shields its ability to impact on particles present in the kitchen exhaust air. UV also generates ozone which is harmful to human health and needs to be controlled. As with ozone filtration, UV lights may require downstream carbon filtration to remove ozone prior to discharge. UV lights need to be automatically switched off when hood filters are removed for maintenance to prevent possible exposure of human eyes to UV light.

Other Hood Technologies

Other hood filtration technologies such as cold-water misting hoods and hood mounted electrostatic precipitators may also form part of a performance based alternative to a deemed to satisfy design, however make sure performance at 1-10μm is checked if it considered a load reducing mechanism. If it is designed to clean the air of all particulates, check the filtration efficiency at 0.3 micron to a relevant standard as discussed in Part 1.

Particle filtration

After the particulate load has been reduced at the hood, there are three types of common particle filtration designed to achieve 95% efficiency at 0.3μm. The multi-stage filter pack, a manual clean electrostatic precipitator (ESP) and an automatic wash ESP.

All filtration systems perform best with a straight duct run of 2-4 duct diameters prior and after the system, this allows the contaminants and airflow to be spread evenly over the filters.

Multi-Stage Filter Pack

These systems typically offer a low capital cost solution with medium operating costs from the filter change outs.

multi stage filter pack
Figure 2: A typical multi-stage filter pack

These systems are typically designed with four (4) stages of filtration consisting of three (3) sequential stages of particulate filtration and one (1) final stage of odour control (covered later).

  • Pre-filtersacrificial: Main contaminant filtered is coarse grease. Typically a low cost G4 filter, with the shortest lifespan.
  • Extended surface filter – high capacity: Main contaminant filtered is fine grease. Typically a F7-F9 filter with a medium lifespan.
  • Final filterefficiency: Main contaminant filtered is smoke. Typically an E11 (HEPA) filter with a minimum efficiency of 95% at 0.3μm with a long lifespan.

Filters are chosen to be resistant to moisture and grease. The material of choice for the second filter and final filter is therefore often fibreglass, whilst the pre-filter is best constructed with a metal frame and rigid, thick, well supported media.

The velocity of air through filter pack systems should be below 1.85m/s. A higher velocity leads to an excessive pressure drop, poor contaminant to filter surface area ratio, less stability of moist filter mediums and likely poorer performance of downstream odour control systems.

A multi-stage filter pack filtering 2,500l/s of air has a surprisingly large media surface area of 100m2. This allows the system to filter a large amount of grease and smoke before replacement of filters is required. However, if the expected load of grease and smoke is high, an automatic wash ESP should be the preferred choice due to a potentially lower life cycle cost.

Electrostatic Precipitator (ESP)

ESP’s use electrostatic forces to charge particles initially with a given polarity before collection on plates downstream of the opposite polarity. Or sometimes the collection plates are grounded with opposing plates of the same polarity as the initial charge that repel the particle to the grounded plate.

filter stages
Figure 3: Electronic air cleaning with an electrostatic precipitator

Particles with a larger size can hold a larger charge relative to their mass, and therefore migrate faster towards the collection plate [4]. The speed of radial movement towards the collection plates is known as the migration velocity, with faster migration velocities resulting in a higher filtration efficiency. However, like mechanical filtration discussed in Part 1 of this series, once particles are smaller than 1μm, the effect of the direction of bulk air flow on the particle is negated due to diffusion and the particles can move easier radially to bulk flow [4] ensuring more residence time in the filter. In fact, particles of 0.01μm are as easy to catch as particles of 100μm (Figure 4), further highlighting the need to compare products at 0.3μm to understand how a product really performs.

particle size graph[4]

ESP’s have been maximised with respect to the laws of physics for 60+ years, and we cannot simply “turn up” the power to achieve better filtration through higher levels of charging. If we do turn the power up, the voltage difference between two plates of different charge becomes too high, resulting in an excessive voltage gradient between the plates.

For typical air in kitchen exhaust, if a voltage gradient of 1kVdc/mm is exceeded this will result in the breakdown of air [5], and therefore immediate transfer of current from the charged plate to the ground plate. This is known as “arcing” and is responsible for the ticking noise that may be heard from malfunctioning ESP’s. When arcing occurs, the whole system becomes grounded and the systems efficiency immediately becomes 0%. This will have a severe effect on downstream odour filtration systems.

Therefore, the main differences between various ESP’s is with the thickness and type of construction materials and whether the unit has an automatic wash system. Due to physics being maximised, there is limited scope for a differing filtration performance for 0.3μm particles at a set velocity.

The last piece of the puzzle with ESP’s is to determine a velocity through the system which is slow enough to ensure that 95% of particles of size 0.3μm (which have the slowest migration velocity) will have enough residence time in the filter to be captured.

Studies have been completed by Dalmon and Lowe [6] which show the migration velocity as a function of axial velocity through the electrostatic filter.

gas velocity graph[6]

From Figure 5 we can see that the optimal axial velocity through an ESP that will generate the highest migration velocity is 2m/s. The migration velocity relative to axial velocity is largely affected by a parameter called particle mobility. At low velocities the effective migration velocity is reduced due to the large amount of turbulence that is generated by an energized field with low airflow, whilst at high velocities the net migration velocity is reduced due to re-entrainment and scouring of particles [7].

Single cell ESP’s are therefore normally designed to be deep enough to offer the sufficient residence time to achieve 95% efficiency at 0.3μm, whilst taking advantage of a higher migration velocity around 1.8-2.0m/s.

Whilst there is a lot more to it, such as the “stickiness” and resistivity of particles, if the volume of throughput air is doubled for the same size filter, the effective migration velocity is lowered significantly. This can result in the length of the unit needing to be up to 6 times longer to capture the same particle due to a net migration velocity that is 3 times slower (Figure 6) – an inefficient use of materials and space.

horizontal distance travelled by particle[6]

Therefore, double pass ESP’s (two single cell ESP’s arranged in a series) do not necessarily allow you to double the airflow and maintain 95% efficiency at 0.3μm. Because of this, it is always advised to increase the face area of the filter to reduce the velocity to below 2.0m/s, preferably 1.85m/s, instead of putting single cell ESP’s in series.

It is possible to increase the velocity above 1.85m/s and achieve 95% at 0.3μm by placing a HVAC particulate filter after a single stage ESP. The reduction in efficiency from the ESP due to an increased velocity will increase the loading on the HVAC filter, and the HVAC filter will require more frequent replacement and it introduces extra static into the system.

If the ESP is doing its job and correctly sized – there will be extensive contamination trapped on the plates of the ESP. These plates must be cleaned / regenerated periodically to remove grease and smoke particles. An ESP can clean air of all contaminant levels; however, the required frequency of cell cleaning directly relates to the contaminant load; a higher contaminant load means more frequent washing of the cells to maintain performance of the system.

grease buildup on esp
Figure 7: Grease build up on an ESP due to poor maintenance to the extent that the grease directly connected the charged and the ground plates – grounding the whole unit.

ESP’s will require a wash anywhere from daily to weekly, making a manual clean system prohibitive due to labour costs and the automatic wash system a must. If the plates are not cleaned, the gap between the plates becomes smaller and smaller which increases the voltage gradient. Once the voltage gradient is above 1kVdc/mm, arcing occurs, and the ESP efficiency goes to 0%.

 Automatic Wash Electrostatic Precipitator (ESP)

The automatic wash ESP is a high capital cost solution but has the lowest operating costs due to the self-cleaning mechanism. Despite the automatic wash system, a manual clean will still be required every 3-24 months depending on the cooking load. Additional infrastructure is required including a hot water supply and a drain to a grease trap.

automatic wash electrostatic precipitator
Figure 8: A typical automatic wash electrostatic precipitator

Large scale, self-washing ESP systems may include the following options:

  • Pre-filter – typically a mesh filter to capture large particles which may short circuit the ESP cell. It also assists with providing laminar flow over the ESP.
  • ESP Cell – removes smoke and grease particulates, and requires regular cleaning to maintain performance.
  • Auto-wash function – typically a programmable automatic wash and fan dry function that sprays the ESP cell with a hot water and detergent mix to keep it running effectively and extend manual cleaning intervals.
  • Final filter – typically a F7/8 extended surface filter used to temporarily protect the odour removal systems from poorly maintained ESPs or increase filtration performance.

Manual Clean Electrostatic Precipitator (ESP)

The manual clean ESP typically offers a medium capital cost solution with high operating costs due the labour required to manually service the system. As the efficiency degrades with loading, the frequency of cleaning required to keep an ESP operational is every 1-7 days depending on grease and smoke load. As the capital cost is higher than a filter pack, and the operating costs are also likely higher given the same quantity of grease and smoke filtered, it is seen an undesirable choice for kitchen exhaust filtration.

Gaseous Odour Filtration

Gaseous odours in kitchen exhaust consists of many different compounds; the composition and quantity vary depending on the type of cuisine. Cooking fumes consist of n-alkanes and fatty acids with n-alkanols, aldehydes, ketones , sterols, polycyclic aromatic hydrocarbons (PAHs) and aromatic amines (AA’s) amongst many others also present [8] [9] [10] . The ability of the two most common methods of odour control being adsorbent/chemisorbent media and ozone are covered in this article.

As discussed in part 1, both gaseous odour control mechanisms require the air to be filtered to 95% at 0.3μm prior to gaseous odour filtration.

Adsorbent/chemisorbent media

The most common gaseous odour adsorbent media is activated carbon, however there are various types of media including impregnated activated alumina known as a chemisorbent media. Chemisorbent media reacts with the contaminants ensuring they are not released, whilst adsorbent media can re-release trapped contaminants if a preferred contaminant is present or the temperature is too high.

The various types of medias and impregnations focus on effective removal of particular contaminants – whereas plain activated carbon is seen as a general odour adsorbent media. Often plain activated carbon is initially used, however, if certain odours from the cooking are not consistently trapped to a desirable level, it is replaced with a mixture of activated carbon and activated alumina impregnated with sodium or potassium permanganate.

Gaseous odour filtering media is a mature technology, but there are still many key design parameters that need to be addressed to ensure cost effective efficient performance. The empty bed residence time, the quality of the media and the pressure drop are the three components of an effective system. Care must also be taken to ensure that the air temperature is not above 50 as the adsorption/chemisorption of contaminants at these temperatures is severely inhibited.

The empty bed residence time (EBRT) is the volume of media in the filtration system divided by the volumetric flow rate of air. A minimum value of 0.06 seconds is recommended with 0.08 seconds preferred. This results in an initial contact efficiency above 99%, which providing the filter is well sealed, will result in an initial contact/removal efficiency of 99%+. When the active sites on the media are consumed by gaseous odours, the removal efficiency drops below the contact efficiency, as the active sites in the media now may be already occupied when contact is made. Eventually too many active sites are full, and gaseous odours begin to pass through, this is when the media needs to be replaced.

relationship between ERBT and contact efficiency

There are various configurations of media that can achieve an ERBT of 0.08 seconds, the most common being 25mm thick, 450mm deep v-bank arrangement with a duct velocity of 1.85m/s. Cost and pressure drop are obvious considerations and specifying a minimum ERBT is the simplest way to ensure you are getting a high efficiency with gas filtration media.


Ozone works by breaking down molecules into smaller and smaller molecules with oxidation until eventually simple, common odourless gases are left such as CO2, N2, H2O and O2. This process is known as ozonolysis. There are two commonly used methods to generate ozone:

  • Corona discharge
  • Ultraviolet (UV) Radiation

Both methods involve drawing air through a modular unit attached onto the side of a duct that is under negative pressure. The negative pressure in the duct creates a flow of outside air through a unit containing high voltage plates (corona discharge) or 185nm UV lamps (ultraviolet radiation) generating ozone in the airflow that enters the main duct.

A unit is designed to treat a volumetric amount of air, and generally a minimum of 2-5 seconds residence time is specified to allow the ozone to clean the specified volume of air with a moderate gaseous odour loading before exhausting outside. The exact science between ozone concentration, residence time, gaseous odour concentration, and filtration efficiency leaves a lot to be desired, with rules of thumb of unknown origin including the 2-5 second residence time dominating designs.

There is a major concern relating to residual amounts of ozone that is exhausted to occupied spaces. Ozone exposure in humans has been associated with dizziness, insomnia, coughing, chest pain, reduction in lung function and irreversible obstructive airway diseases [11]. In particular, ozone exposure is damaging to the un-developed lungs of children [12]. Ozone has been found to be injurious to health at levels consistently above 50 ppb [13], whilst Safe Work Australia enforces a 100 ppb peak limit for human exposure [14]. AS1668.2 recommends that no residual ozone beyond ambient levels remains in the final exhaust air.

An assurance that a system is engineered to ensure no excess ozone will ever be exhausted should be considered a must to obtain compliance for an engineered solution.

To prevent residual ozone discharge, the first option is to install an ozone detector just prior to the building kitchen exhaust point; the ozone level detected here can control the amount of ozone generated upstream. As the odours from the cooking load increase and decrease, the amount of ozone consumed varies and therefore the amount required to be generated to ensure no excess ozone is discharged also changes. Care needs to be taken to ensure the prolonged reliability and calibration of these sensors over time – as failure generates a serious health risk.

The second option is to install a carbon trap at the discharge point to capture remaining ozone in the stream. Again, care must be taken to replace media as required to prevent any possible discharge of ozone to external locations.


We can significantly improve the effectiveness of kitchen exhaust filtration using the proposed design principles from part 1 of:

  • Particle filtration should always occur prior to gaseous odour filtration.
  • Particle filtration performance should be based on performance at 0.3μm (within the most penetrating particle size range of 0.1-0.4um).
  • A particle filtration minimum performance of 95% at 0.3μm should be targeted.

When carefully considered, there are many filtration technologies that effectively and economically meet the proposed design principles and the requirements of an engineered performance solution required by the National Construction Code (NCC). The laws of innovation suggest that this list will continue to evolve and grow.

By adopting these principles, and diligently testing and auditing products in the market, we can improve the ambient air quality, reduce visual pollution and reduce odour complaints.

Written by:

Jonathan Bunge, National Product Manager, Airepure Australia Pty Ltd

Technically reviewed by:

Brett Fairweather, Mechanical Engineering Consultant, It’s Engineered Pty Ltd

Peter Mathieson, Technical Director, Aurecon Australia Pty Ltd



[1] Standards Australia, AS1668.2, SAI Global, 2012.
[2] Standards Australia, AS/NZS 1668.1, SAI Global, 2015.
[3] UL1046 ED.4, Underwriters Laboratories, 2010.
[4] A. Mizuno, “Electrostatic Precipitation,” IEEE Transactions on Dielectrics and Electrical Insulation, vol. 7, no. 5, pp. 615-24, 2000.
[5] P. A. Tipler and G. Mosca, Physics for scientists and engineers, Macmillan, 2007.
[6] J. Dalmon and H. J. Lowe, “Experimental Investigations into the Performance of Electrostatic Precipitators for P.F. Power Stations,” in Coll. , Int. la Physique des Forces Electrostatiques et Leurs Application Centre National de la Recherche Scientifique, 1961.
[7] K. Parker, “PDF Title: 3. Factors Impinging on Design and Performance,” in Electrical operation of electrostatic precipitators, Institute of Engineering and Technology, 2003, pp. 39-62.
[8] J. H. Cheng, Y. S. Lee and K. S. Chen, “Carbonyl compounds in dining areas, kitchens and exhaust streams in restaurants with varying cooking methods in Kaohsiung, Taiwan,” Journal of Environmental Sciences, 2015.
[9] W. To, Y. Lau and L. Yeung, “Emission of carcinogenic components from commercial kitchens in Hong Kong,” Indoor and Built Environment, vol. 16, no. 1, pp. 29-38, 2007.
[10] Y. Zhao, M. Hu, S. Slanina and Y. Zhang, “Chemical compositions of fine particulate organic matter emitted from Chinese cooking,” Environmental science & technology, vol. 41, no. 1, pp. 99-105, 2007.
[11] M. Lippmann, “Health effects of ozone a critical review.,” Japca, vol. 39, no. 5, pp. 672-95, 1989.
[12] Netcen, “Guidance on the Control of Odour and Noise from Commercial Kitchen Exhaust Systems.,” Department for Environment, Food and Rural Affairs , London, 2005.
[13] J. R. Stedman, H. R. Anderson, R. W. Atkinson and R. L. Maynard, “Emergency hospital admissions for respiratory disorders attributable to summer time ozone episodes in Great Britain.,” Thorax, vol. 11, no. 52, pp. 958-63, 1997.
[14] Safe Work Australia, Workplace Exposure Standards for Airborne Contaminants, Safe Work Australia, 2013.


About the Author

Jonathan Bunge

Jonathan Bunge
(M.Eng Chemical)
Product Development Manager
Airepure Australia

Request PDF

  • This field is for validation purposes and should be left unchanged.