Health benefit/burden, PM2 .5 removal effectiveness, and power consumption based comparison of common residential air‐cleaning technologies in the United States

Abstract This modeling study compared the common air cleaners in U.S. residences based on averted disability‐adjusted life years (DALYs) related to indoor PM2.5 concentration reduction and the DALYs resulted from carbon‐di‐oxide (CO2) emissions from power consumption. The technologies compared include mechanical fibrous filters, electret fibrous filters, and electronic air cleaners. For DALYs estimation, the indoor PM2.5 concentration and power consumption were first calculated and compared. These were then multiplied by the respective health damage factors. Air cleaners were compared under several indoor particle size distributions scenarios. A methodology was developed to evaluate the influence of the aging of air cleaners on the selected comparison criteria. The results suggest that the averted DALYs from indoor PM2.5 concentration reduction far supersedes the indirect DALYs associated with the operational power consumption of the air cleaners. Hence, the DALY‐based ranking of the air cleaners considered was the same as that of their effectiveness to reduce indoor PM2.5 concentrations. However, the result should be taken with care as only the use‐phase of air cleaners was considered. For future study, a complete life‐cycle assessment is recommended. Considering aging can change the ranking of the air cleaners and is thus advised to be incorporated in further studies.

impacts. 11 Overall, an ideal air purifier should be able to reduce indoor particles to safe levels with low power consumption and thus result in overall health benefits.
Different indoor air-cleaning technologies are available in U.S. residential buildings. Broad categorization includes fibrous filters (FFs), electret fibrous filters (EFFs), and electronic air cleaners (EACs). 12 The FFs are known to collect the dust particles from inflow air by mechanisms including diffusion, interception, and inertial impaction. 13 In EFFs, the fibers carry an electric charge and, hence, additionally collect particles by electrostatic attraction. 14 The commercial EAC usually consists of an electrostatic precipitator (ESP) with a pre-filter mesh for removing large particles and a post-filter for odor removal. 12 In ESP, the particles are passed through a strong electric field and afterward collected on alternatively charged or grounded collection plates. 13 All these air cleaners have their advantages and disadvantages. In general, FFs can achieve high filtration efficiency, but at the cost of high-pressure drop which leads to large power consumption. 15 Both filtration efficiency and pressure drop tend to increase with the accumulation of dust over the filter with usage/ loading. 12 New EFF can achieve higher filtration efficiency compared to the uncharged FF having the same filter parameters. 14 It is due to the additional electrostatic attraction of particles toward the charged fibers. However, with loading, the charges can be shielded by the collected particles, which reduces the in-use filtration efficiency initially. After a substantial build-up of the dust cake, filtration efficiency starts increasing similar to other uncharged FFs. 12 EACs generally have lower pressure drop compared to FFs and EFFs. 12,15 However, additional power is required to ionize the gas molecules. With use, EAC's filtration efficiency may decrease, and pressure drop may not change much. 12 Several modeling studies have been done in the past to compare residential air cleaners in different scenarios. 6,[16][17][18] The comparison was usually conducted based on minimum efficiency reporting value (MERV). El Orch et al 6  particle loading rates of HVAC filters in different scenarios. The filtration efficiency was assumed to be the same as from new filters even after dust loading. These studies either neglected the impact of loading on filtration efficiency and pressure drop or ignored the power consumption of these technologies. Also, none of these studies focused on the health benefits/burden of using air cleaners.
Several studies tried to incorporate indoor air quality into lifecycle assessment (LCA). 4,[20][21][22] The main purpose of installing air cleaners in the house is to provide a healthy environment to individuals. 10 However, the indirect health impacts associated with power consumption due to carbon-di-oxide (CO 2 ) emissions are not estimated. 1 kg equivalent of CO 2 emissions (customary100-year global warming potential values) is associated with a damage factor of 2 to 6.2 × 10 −7 Disability-Adjusted Life Years (DALYs). 11  The figure of merit (FOM), representing the ratio between filtration efficiency and pressure drop, is often used to compare air cleaners as higher FOM indicates better performance air cleaner. 23 However, FOM may not be sufficient to evaluate the suitability of a device in a particular environment. Sources and strengths of residential indoor PM 2.5 vary substantially from one indoor environment to another. 6,19 A high filtration efficiency device but having highpressure drop may be necessary for scenarios having high indoor PM 2.5 , to bring the concentration to a safe level. However, in some lower concentration scenarios, the same high filtration efficiency filter may not be desirable and would unnecessarily lead to high power consumption, and hence, may have an overall adverse health impact.
For this, DALY offers more flexibility as it varies for different indoor PM 2.5 concentration scenarios.
This study compares the averted DALYs due to a decrease in indoor PM 2.5 concentration and DALYs increase resulting from power consumption among three residential air-cleaning technologies commonly used in U.S. residences, namely EAC, EFF, and FF. For this, the study first estimated and compared indoor PM 2.5 concentration and power consumption. The evaluation used different indoor PM 2.5 sources and a range of air exchange rates, particle deposition loss, and penetration factors.

| Residential building model
A completely mixed one-box model was assumed as a representative of residential building envelope ( Figure 1). N in and N o were the indoor and outdoor size-resolved particle number concentration Practical Implications: • The results suggest important factors to be considered while comparing the air cleaning technologies for a particular indoor environment.
• The method developed allows comparing air cleaners without the burden-shifting from indoor air pollution reduction by air cleaners to their power consumption.
• These associations persisted only for disinfecting wipes and were no longer observed for green and home-made products, when considering the co-use of irritants and sprayed products at home.  except for the absence of the building mechanical ventilation air intake system, as in most residential buildings, the fresh air enters only through natural ventilation via doors or windows, or through infiltration (cracks in building envelope). 19 Major assumptions were-no concentration gradient near the source of indoor emission and no re-suspension or coagulation of particles. The consideration of re-suspension would be important for coarse particles. 18

| Particle size distributions (PSDs)
Several indoor PM 2.5 concentration scenarios were formed. Indoor emissions were assumed to be from either one or both cooking and smoking. The study included both urban and rural location PSDs.
The scenarios and their explanation were provided in  windows were open to a large extent (high window opening) and 80% of the time, to a low extent (low window opening). 6 In total, there were 16 PSDs.
Air exchange rate (AER, λ i , 1/h), particle size-resolved penetration factor (P), and size-based deposition loss rate (K dep , 1/h) were A tri-modal log-normal distribution was fitted to the penetration factor curve for closed window case given in Figure 3 (a) of El Orch et al. 6 The obtained distribution parameters can be found in Table A.4 F I G U R E 1 One box model of a residential building having the air-cleaning device installed. It is representing the fate and transport of pollutants in the building envelope. The implication of clean air and power consumption by air cleaners is also indicated 3) was used. The deposition loss rate and penetration factors were given for particles of size less than 10 μm. The same curves were extended for larger diameter particles as given in The air re-circulation rate (λ R , h −1 ) across the air cleaner was cal- The selected air cleaners were renamed using the type of technology followed by the MERV rating. For example, "NS" an FF with MERV 6 was renamed as FF6. As two EFFs ("NM" and "FUA") have MERV rating 12, these were renamed as EFF12.1 and EFF12.2, respectively. Table 1 shows the pressure drop and the modified names.
The filtration efficiency and pressure drop data were at the flow rate drop data for aged device case were given in Table A.5 in SI A.1.3.

| Indoor PM 2.5 concentration
To obtain indoor PM 2.5 concentration, first the indoor size-resolved particle number distribution N in was calculated by Equation 1.
Outdoor particles can enter indoor depending upon penetration factor (P) and AER (λ i , 1/h The above steps were repeated assuming size-resolved densities in place of constant unit density. This was to check whether the ranking of performance of air cleaners was affected by the density assumption or not. For particles of size lower than 10 μm, same size-resolved densities were used as that used by Azimi et al 16 (Table 2). For particles of size larger than 10 μm, density of 2.5 g/cm 3 was assumed. 25

| Power consumption
For FF and EFF, during operational phase, power consumption was only to overcome pressure drop, termed as fan power (P f , W). It was calculated as P f = Q r ∆p f /(η fan .3600), where Q r was the air flow rate through the air-cleaning device in m 3 /h, ∆p f was the pressure drop experienced by air-cleaning device [Pa], and η fan was the fan efficiency. The factor 3600 was to convert time in h to s. η fan was taken as 0.5 in this study and Q r was the same as mentioned in The total power consumption by EAC was the sum of the device power and fan power (P eff = P device + P f ). The fan power calculation for EAC was the same as that for FF and EFF. The power consumption of an aged device also depended upon the loading. Hence, time-weighted power consumption was calculated using the times calculated in Section 2.3.

| DALYs from indoor PM 2.5
The DALYs associated with indoor PM 2.5 concentration were calcu- was to convert C PM2.5 from μg/m 3 to kg/m 3 .   (3)

| Indoor PM 2.5 concentration
TA B L E 2 Size-resolved particle densities used for particles of size less than 10 μm having a significantly lower MERV rating of 7. The reason is, for particles of size lower than 0.3 μm, the size-resolved filtration efficiency of EFF7 is higher than that of EFF12.1 ( Figure A.6 (a) in SI A.2.2). Note that the MERV rating is based only on the filtration efficiency for particles of sizes greater than 0.3 μm. 16 Hence, the results suggest the need to incorporate the filtration efficiency of lower diameter particles while rating the commercial filters. Same is stated by Hecker and Hofacre, 12 and Azimi et al. 16 The results are somewhat biased as only very low MERV FF was taken for comparison. However, Hecker and Hofacre 12 reported that only the low MERV FFs are common in U.S. residences.
For the aged device case, as represented by translucent brown diamonds in Figure 2, the PM 2.5 removal effectiveness is again highest in case of EAC14. This is even after selecting the worst filtration efficiency EAC among the three EACs reported in Hecker and Hofacre 12 (details in SI A.1.2). However, among EFF12.1 and EFF7, the ranking is reversed compared to new filters. The overall effectiveness of aged EFF7 is lower than that of EFF12.1. Thus, only considering the filtration efficiency of new filters is insufficient to judge the air cleaner in a real-life situation.
The indoor PM 2.5 concentration after changing the unit density assumption to that mentioned in Sec. 2.3 is given in Figure

| Power consumption
The power consumption of all new devices and aged EAC14 is given in Figure 3.

| Disability-adjusted life years
Simplifying whereas, the power consumption ranges between 17 to 183 Watts (Section 3.2). In no situation, the health benefit through PM 2.5 reduction is lesser than the health burden related to CO 2 emission from power consumption. Thus, the ranking based on the DALY criterion is similar to what is shown in Section 3.1. A common way to evaluate the air cleaner performance is by the figure of merit (FOM).

| CON CLUS ION
The criterion DALY was used to measure direct and indirect health benefits/burdens from air cleaners. This study revealed that the reduction in indoor PM 2.5 concentration improved the DALYs far more than the indirect DALYs associated with the operational power consumption of the air cleaners in United States. However, the results should be taken with care as only the use-phase was considered.
Future research considering the entire life-cycle including end-oflife treatment phase of the air cleaners is needed.
For new devices (constant filtration efficiency and pressure drop over time) and aged devices (changing filtration efficiency and pressure drop over time), the effectiveness to remove indoor PM 2.5 was the best for EAC14 followed by EFFs. No clear deduction could be made for power consumption.
It was observed that the aging of a filter can change the ranking of the air cleaners. Aged EFF7 removed PM 2.5 more effectively than aged EFF12.1. The results were the opposite for the new device case. Hence, it is recommended for modeling studies to consider the aging effect.
The current MERV rating is assigned by considering filtration efficiency for particles of diameter greater than 0.3 μm. It was observed that air cleaners with a higher MERV rating may not necessarily be more effective in reducing indoor PM 2.5 compared to lower MERV rating cleaners. Thus, it would be recommended to adjust the MERV rating to include the filtration efficiency for particles of size less than 0.3 μm.
The study did not consider the harmful byproducts from the air cleaners. ESP technology is known for ozone emissions. Thus, the extension of this study should include the impact of byproducts in DALY calculation.
The study considers only the air cleaners that are commonly used in U.S. residential buildings. Thus, the evaluated performance ranking of the cleaners in this study is not a given. For office buildings, where high-efficiency particulate absorbing (HEPA) filters are common, the ranking can be different. However, the methodology developed in this study is generalizable and can be extended to other situations.

AUTH O R CO NTR I B UTI O N S
Not applicable, as number of authors are only two.

ACK N OWLED G M ENT
The work was partially supported by Center for Filtration Research at University of Minnesota with the Subaward No. W530672710.

Open access funding provided by Eidgenossische Technische
Hochschule Zurich.

FU N D I N G I N FO R M ATI O N
The work was partially supported by Center for Filtration Research at University of Minnesota with the Subaward No. W530672710.

CO N FLI C T O F I NTE R E S T
No, there is no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analyzed during this study are included in this published article (and its supplementary information files).