Accepted Manuscript Performance of sorption-based portable air cleaners in formaldehyde removal: Laboratory tests and field verification Xiaoyue Zhu, Mengqiang Lv, Xudong Yang PII:
S0360-1323(18)30165-3
DOI:
10.1016/j.buildenv.2018.03.030
Reference:
BAE 5364
To appear in:
Building and Environment
Received Date: 14 November 2017 Revised Date:
19 March 2018
Accepted Date: 20 March 2018
Please cite this article as: Zhu X, Lv M, Yang X, Performance of sorption-based portable air cleaners in formaldehyde removal: Laboratory tests and field verification, Building and Environment (2018), doi: 10.1016/j.buildenv.2018.03.030. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1
Performance of sorption-based portable air cleaners in
2
formaldehyde
3
verification
4
Xiaoyue Zhu, Mengqiang Lv, Xudong Yang*
5
Beijing Key Laboratory of Indoor Air Quality Evaluation and Control, Department of
6
Building Science, Tsinghua University, Beijing 100084, P. R. China
tests
and
8
*
9
Dr. Xudong Yang
Corresponding author:
M AN U
7
Department of Building Science, Tsinghua University, Beijing 100084, China
11
Tel: +86 (10) 62788845
12
Fax: +86(10) 62773461
13
Email:
[email protected]
AC C
EP
TE D
10
14
1
field
RI PT
laboratory
SC
removal:
ACCEPTED MANUSCRIPT Abstract
16
Formaldehyde is one of the primary indoor air contaminants that widely exists in
17
construction materials and household consumable products. Acute exposure to
18
formaldehyde causes irritation and dermal allergies, while chronic exposure can result
19
in DNA and chromosomal damage. However, only a handful studies have evaluated
20
formaldehyde removal capacities of portable air cleaners (PACs) both in the field and
21
in the laboratory. The laboratory performance of PACs has not been statistically
22
compared with their field performance. This study evaluated the initial formaldehyde
23
removal capabilities of several relatively popular commercial sorption-based PACs in
24
Chinese market by measuring their clean air delivery rates (CADRs) in an 8 m3
25
environmental chamber. The modified ‘pull-down’ method was applied in this study,
26
and the total operation time of the tested PACs was 1.5 h. The laboratory results
27
showed wide variations in the CADRs (13.8 m3/h to 75.6 m3/h), which was in
28
agreement with the CADRs reported in previous studies. A single-zone field test under
29
natural ventilation was also conducted in a bedroom with an area of 25 m2 and a
30
volume of 67.5 m3 using the best performing PAC. The results were statistically
31
analyzed for any significant difference between the laboratory and field data. The
32
difference between the laboratory and field performance of the tested PAC was
33
insignificant at a confidence level of 95%.
35 36
SC
M AN U
TE D
EP AC C
34
RI PT
15
Key words: Portable air cleaner; Formaldehyde; Clean air delivery rate; Field test
2
ACCEPTED MANUSCRIPT 1. Introduction
38
With the rising living standards and growing health awareness, indoor air quality has
39
become increasingly important. Formaldehyde is one of the primary indoor air
40
contaminants that widely exists in construction materials and household consumable
41
products. Acute exposure to formaldehyde can result in mucus membrane irritation,
42
dermal allergies, and allergic asthma. Chronic exposure to high concentrations of
43
formaldehyde has severe adverse impacts on health such as causing neurotoxicity
44
syndromes, pulmonary function damage, DNA and chromosomal damage[1].
45
Guideline value of indoor formaldehyde concentration stipulated in the Chinese
46
national standard is 0.10 mg/m3[2]. However, in poorly ventilated new and renovated
47
homes, formaldehyde concentration can surge up to two to four times of the upper
48
limit[3]. Effective measures should be taken to remove indoor formaldehyde.
49
Indoor air contaminants removal is mainly achieved by three means: 1) source control;
50
2) ventilation; 3) air purification. However, occupants today are faced with an
51
intractable dilemma. On the one hand, elimination of indoor source is difficult
52
because it calls for joint effort of manufacturing, enforcement, and legislation. On the
53
other hand, intensified airtightness of modern buildings and frequent haze limits
54
natural ventilation in China. Consequently, air purification is indispensable for
55
removing air pollutants.
56
Portable air cleaners (PACs) have been readily received by the market. According to a
57
survey carried out in 2005, three out of every ten households own at least one type of
58
air cleaning device in the United States[4]. However, some researchers reviewed the
59
literature on air cleaners and noted that studies focusing on the removal of volatile
60
organic compounds (VOCs) were inadequate[5], not to mention studies about
61
formaldehyde.
62
With regard to laboratory tests, a large portion of the studies focused on in-duct air
AC C
EP
TE D
M AN U
SC
RI PT
37
3
ACCEPTED MANUSCRIPT cleaners (or filters) [6-11]. Filters in the air channels were examined and evaluated
64
according to the Korean or Japanese standards. These standards designate the
65
single-pass efficiency, removal efficiency, and effectiveness as primary indicators of
66
PAC performance. However, the same filter behaves differently when installed as
67
in-duct air cleaners or stand-alone air cleaners. Furthermore, due to their limited
68
single-pass removal efficiency, in-duct cleaning devices for VOCs are rare, hence
69
investigating filters alone is not sufficient. Studies on PACs are rare and most of them
70
were reported several years ago. In 1989, Daisey and Hodgson compared four air
71
cleaners based on their initial clean air delivery rates (CADRs) for nitrogen dioxide,
72
dichloromethane,
73
hexanal[12]. Later in the 1990s, several researchers evaluated the effectiveness of air
74
cleaners in the removal of diverse pollutants by measuring the CADRs. Tests were
75
conducted in varisized environmental chambers. Nelson et al. developed a method
76
that utilizes the real-time concentration to calculate the CADRs for different analytes
77
in environmental tobacco smoke and tested three air cleaners (chamber volume: 18
78
m3)[13]. Shaughnessy et al. tested the effectiveness of a dozen commercial air
79
cleaners for dust particulates, environmental tobacco smoke, bio-contaminants, and
80
formaldehyde in a 24.8 m3 chamber[14]. Niu et al. extended the Association of Home
81
Appliance Manufacturers (AHAM) method to estimate the initial effective cleaning
82
rate of nineteen air cleaners for toluene removal in a 6.4 m3 chamber [15]. According
83
to these studies, sorption-based air cleaners were only marginally effective in
84
removing gaseous contaminants. In 2005, Chen et al. used a mixture of sixteen VOCs
85
to evaluate the VOC removal ability of fifteen air cleaners with different technologies
86
in a 54.4 m3 chamber[16]. The results of this study suggested that sorption-type filter
87
was the most effective off-the-shelf commercial technology for VOC removal,
88
however, the removal efficiencies varied significantly with filter design. Cleaning
89
technologies advance quickly. It is necessary to examine if the performance of today’s
90
air cleaners exceed the performance of their predecessors.
n-heptane,
toluene,
tetrachloroethylene,
and
AC C
EP
TE D
M AN U
2-butanone,
SC
RI PT
63
4
ACCEPTED MANUSCRIPT 91
Laboratory tests are usually performed without field verification. It remains
92
ambiguous whether laboratory data can predict the behavior of air cleaners in real
93
situations. If this is feasible, significant efforts could be saved because field test is
94
rather expensive and time-consuming. Most of the previous field tests employed a
95
‘steady-state’ method[13,
96
correct partitioning of the different stages of pollutant removal. However, field test
97
should avoid elevating concentration to laboratorial level. Laboratory tests aim at
98
evaluating PACs quickly. Therefore, devices are usually challenged with 10 times or
99
even higher concentrations of contaminants compared to typical indoor environments.
100
However, field tests aim to predict the performance of PACs close to real
101
concentration levels. Unfortunately, portable devices that can accurately measure
102
formaldehyde at low concentration are usually expensive and not very popularized.
103
Moreover, the ‘steady-state’ method requires constant release intensity to calculate the
104
CADR. Thus, it uses standard gases of the targeted contaminants and precise gas
105
generator, which makes field tests even more laborious.
106
Although diverse technologies for formaldehyde removal are available, many of them
107
are for industrial use. In terms of technologies used in residential environment, where
108
concentrations are low, the most commonly used technologies are 1) adsorption; 2)
109
ozone/ion oxidation; 3) ultraviolet photo-catalytic oxidation. Previous studies noted
110
that the technologies that require energy input often have problems related to the
111
generation of hazardous byproducts and ozone exposure[19]. This has hindered wide
112
acceptance of such technologies. Adsorption is still frequently used by many cleaning
113
devices. Therefore, our study focused on sorption-based PACs.
114
This study evaluated the initial formaldehyde removal capabilities of several
115
commercial sorption-based PACs by measuring their CADRs in an 8 m3
116
environmental chamber. The modified ‘pull-down’ method was applied, and the total
117
operation time of the tested PACs was 1.5 h. A single-zone field test under natural
AC C
EP
TE D
M AN U
SC
RI PT
17, 18]. Real-time concentration data is required for the
5
ACCEPTED MANUSCRIPT ventilation was also conducted in a bedroom with an area of 25 m2 and a volume of
119
67.5 m3 using the best performing PAC.
120
The Mann Whitney U test was used to compare results from laboratory and field
121
tests. The U test is a widely-used nonparametric test which is sufficient in comparing
122
two independent samples from populations of skewed distributions and different
123
variances[20]. Moreover, this method is robust even when sample sizes are small and
124
different [21].
125
2. Experimental methods
126
2.1 Tested air cleaners
127
The PAC market in China has experienced rapid growth over the past several years.
128
Nowadays, customers can choose from hundreds of brands. To ensure the selection of
129
representative PAC specimens, market research was conducted prior to the formal
130
experiments. Consumer-oriented questionnaires were randomly circulated on
131
professional online platforms. Respondents were asked about their age, gender, and
132
whether they possessed a PAC. For people who had PACs, they were asked about the
133
brand names used. Then respondents had to choose the most effective technology in
134
their opinions from the options provided (shown in Fig.1). The online platform
135
recorded IP locations and time used for finishing the survey to exclude repeated and
136
irresponsive answers. Totally 360 valid responses were collected. Fig.1 summarizes
137
the favorability and usage rate (derived from a statistical analysis of the 85 most
138
salable PACs) of the most common technologies in China. The favorability of a
139
technology is the ratio of the number of respondents that considered the technology
140
attractive to the total number of respondents. Usage rate of a technology is the ratio of
141
the number of PACs utilizing the technology to the total number of investigated PACs.
142
For instance, the usage rate of ‘High Efficiency Particulate Air filter (HEPA)+
143
Activated Carbon (AC)’ technology was 97.6%. Results indicated that well developed
AC C
EP
TE D
M AN U
SC
RI PT
118
6
ACCEPTED MANUSCRIPT and handy technologies such as ‘HEPA+AC’ were more likely to be adopted as a
145
standard configuration, while advanced technologies played a secondary role to
146
enhance cleaning effect. However, even though all the PACs tested used activated
147
carbon, the adsorption effect varied widely with the specific filter design. We carried
148
out further studies on these PACs with the objective to select some best-selling brands
149
according to sales record in 2016 from the major online shopping websites of China,
150
major manufacturing areas, and common filter forms. Table 1 lists the general
151
information about the PACs used in this study.
TE D
M AN U
SC
RI PT
144
152
Fig. 1 Usage rate and favorability of different cleaning technologies in China: HEPA:
154
High efficiency particulate air filter; AC: Activated carbon; PCO: Photo-catalytic
155
oxidation; ESP: Electrostatic precipitator. Usage rate: the ratio of the number of PACs
156
utilizing the technology to the total number of investigated PACs; Favorability: the
157
number of respondents that considered the technology attractive to the total number of
158
respondents.
AC C
EP
153
159 160
Table. 1 Information of tested portable air cleaners 7
ACCEPTED MANUSCRIPT PAC Code
C
>1500
306.5
Similar to A, but with double-face organization (350x350x40 mm, 0.85 kg)
455
90
Np
320
200
>1500
SC
RI PT
4 filters: Pre-filter + Coarse granular carbon filter (433x235x10 mm, 0.85 kg, deodorization) + Chemical-impregnated filter paper (433x235x10 mm, 0.80 kg, for formaldehyde) + HEPA (433x235x34 mm, 1.00 kg)
170
2 filters: Pre-filter + Composite filter (363x278x45 mm, 0.80 kg, fine granular activated carbon is sandwiched between the HEPA filter pleats)
>1500
EP
D
95
(mg)
300
M AN U
B
CCM* (Claimed by Manufacturer)
Pre-filter + 'Smokestop' filter (456x365x45 mm, 1.00 kg, polyester fiber woven with coconut shell granular activated carbon) +HEPA
TE D
A
Airflow Rate (m3/h)
Type of Filter
Nominal CADR for Formaldehyde (m3/h)
*CCM: cumulate clean mass. It is a performance index defined by China national standard which
162
refers to the total mass of the cleaned contaminant when CADR drops to 50% of the initial value;
163
Np: Not provided
164
2.2 Materials
165
Formaldehyde gas was generated by evaporating liquid formalin (37% solution;
166
impurity content < 0.38%; Tianjin Yongda Chemical Reagent Company, China). The
AC C
161
8
ACCEPTED MANUSCRIPT 167
dosage was determined according to desired initial concentration and pilot
168
experiments. The theoretical dosage of the solution Vs (ml) was calculated as:
169
Vs =
170
C0 (g/m3) is the initial concentration; V (m3) is the chamber volume; Cs (%) is the
171
mass concentration of the solution (37%); ρs (g/ml) is the density of the solution (1.08
172
g/ml). Subsequently, pilot experiments were performed to check whether the dosage
173
was appropriate and adjustments were made accordingly until the difference between
174
the tested initial concentration and the target value fell below 0.5 mg/m3.
175
Formaldehyde solution was transferred into light-proof vials using a Dragon-lab top
176
pipette (100-1000 µl, ± 2.5%). The vials were then placed on an electrical heater, near
177
the oscillating fan so that the formaldehyde evaporated could swiftly enter the mixing
178
zone.
179
This study used a spectrophotometric method with 3-methyl-2-benzothiazolinone
180
hydrazine (MBTH) reagent to measure the concentration of formaldehyde in air. This
181
method is recommended by the Chinese Standard GB/T 18204.26-2000 and rivals the
182
high-performance liquid chromatography (HPLC)[22]. Fig. 2 is the comparison curve
183
of the MBTH and HPLC methods. The concentration range was 0.1-6.0 mg/m3. A
184
1 mol/L solution of chromogenic reagent was prepared from ammonium ferric sulfate
185
dodecahydrate (purity > 99%), and a spectrophotometer (WF J 7200, 325-1000 nm, ±
186
2.0 nm) was used to measure the absorbance.
C0 ⋅ V Cs ⋅ ρs
AC C
EP
TE D
M AN U
SC
RI PT
(1)
9
SC
RI PT
ACCEPTED MANUSCRIPT
M AN U
187
Fig. 2 Comparison curve of the MBTH and HPLC methods
189
2.3 Facility setup
190
The air-tight environmental chamber that was used for the laboratory tests was made
191
of stainless-steel to minimize the adsorption effect. The natural decay rate of
192
formaldehyde inside the chamber was below 0.05 h-1, and the infiltration rate
193
measured with CO2 was 0.0002 h-1. The dimensions of this chamber were 2.0 m
194
(length) x 2.0 m (width) x 2.0 m (height), with a total volume of 8.0 m3. This volume
195
was considered because it was big enough to house the needed testing device while
196
still small enough to achieve a uniform pollutant distribution with the air from the
197
small fans installed in the environmental chamber. Before each test, the inner surfaces
198
of the chamber were scrubbed with deionized water and the chamber was flushed with
199
clean air for at least 12 h. The clean air was directly drawn from outdoor atmosphere.
200
We monitored the concentration of formaldehyde in the cleansing air, which was 1%
201
to 5% of the concentration inside the environmental chamber during a test.
202
Additionally, we took empty samples to measure the background concentration before
203
the targeted contaminant was released. The background concentration was subtracted
AC C
EP
TE D
188
10
ACCEPTED MANUSCRIPT when calculating the CADR, thus the remaining formaldehyde was not a major
205
problem either. To monitor the properties of air, an automatic temperature and
206
humidity recording device (WSZY-1, Tian Jian Hua Yi Co., China) was fixed on the
207
wall of the chamber. The use of a portable electrothermal oven and desktop humidifier
208
were sufficient to regulate the temperature and relative humidity inside the chamber
209
because of its small volume. The power of the equipment was closely regulated until
210
the air properties reached the set point (temperature: 23 ± 0.5 ˚C, relative humidity: 50
211
± 10 %) and readings taken by the automatic temperature and humidity recording
212
device were constant for at least 10 min. Thermal insulation materials and high levels
213
of airtightness helped maintain stable temperature and relative humidity in the
214
chamber. In a typical test, fluctuations in temperature and relative humidity were less
215
than 1% and 5%, respectively.
216
Besides the pre-mounted ceiling fan, which constantly stirred the upper layer of air, a
217
stand-alone oscillating fan was also installed to homogenize the lower layer of air
218
throughout the process. Although uniformity verification test was not conducted
219
because of sampling condition restrictions, the fans used in this study were as
220
powerful as those used in the 30 m3 chamber, where uniform field could be created.
221
Therefore, it was safe to assume a well-mixed field in the 8 m3 chamber used in this
222
study.
223
Considering that the orifice of the prewired sampling tube was near the center of
224
volume of the chamber, the PAC should be placed away from the middle of the
225
chamber, otherwise the purified air stream would directly flush the orifice. Meanwhile,
226
an adequate distance (0.5 m) between the PAC and the wall was maintained to avoid
227
hindering airflow as illustrated in Fig. 3.
AC C
EP
TE D
M AN U
SC
RI PT
204
228
11
SC
RI PT
ACCEPTED MANUSCRIPT
229
Fig. 3 Schematic of the environmental chamber used for the laboratory test
231
A single-zone field test was performed in the research bedroom of a flat located on the
232
third floor of a laboratory building in Changzhou, Jiangsu Province. Fig. 4 shows the
233
layout of the flat. The bedroom was about 5.0 m in length and width and 2.7 m in
234
height (Volume: 67.5 m3). It had an outer door and two external windows, which
235
separated the room from a corridor that was exposed to the outside. The internal door,
236
connecting the bedroom and the adjoining living room, was tightly sealed to prevent
237
cross-impact. As demonstrated in Fig. 5, the room was only primitively painted and
238
had no furniture. In fact, the bedroom was painted more than a year before the
239
research started. The background concentration of formaldehyde in the bedroom was
240
lower than 0.05 mg/m3 (0.035 ppm). Three oscillating stand-alone fans were arranged
241
to overcome the heterogeneous distribution of air pollutants in the room. The air
242
cleaner to be tested was placed in the middle of the room.
TE D
EP
AC C
243
M AN U
230
12
RI PT
ACCEPTED MANUSCRIPT
244
SC
Fig. 4 Layout of the flat where field test of the air cleaners was conducted
246
TE D
M AN U
245
Fig. 5 Photograph of the research bedroom in the flat
248
To simulate a real situation, the room temperature and relative humidity were not
249
manually controlled, but were continuously monitored by at least three automatic
250
recording devices. Like in the laboratory test, formalin solution was heated to generate
251
formaldehyde gas. The location of the electric heater is marked with a triangle in Fig. 3.
252
Considering the different using habits of people, the PAC was tested under three natural
253
ventilation modes: 1) infiltration: only one of the external windows was slightly opened
254
(2.0 cm); 2) moderate ventilation: an external window was opened (one-third of its
255
area); 3) full ventilation: both the external windows were fully opened. The PAC was
256
tested under each mode, which was repeated at least twice.
AC C
EP
247
13
ACCEPTED MANUSCRIPT 257
2.4 Calculation and measurement of the CADR
258
The clean air delivery rate (CADR) is the most commonly used index when evaluating
259
the initial performance of PACs. The mass balance model from which the CADR was
260
derived is established below (Fig. 6).
SC
RI PT
261
M AN U
262 263
Fig. 6 Schematic of the mass balance model of indoor air pollutants. Q: fresh air
264
volume (m3/h); ks: loss ratio due to chamber characteristics (h-1); E: total source
265
intensity (mg/m3∙h); Cout: outdoor formaldehyde concentration (mg/m3);
266
V
TE D
dC Q + ks + CADR = − dt QCout + E V C− Q + ks + CADR
(2)
(3)
EP
267
dC = Q (Cout − C ) + E − ( k s + CADR )C dt
QCout + E Q + ks = B, = Kn Q + k s + CADR V
Let
269
Q stands for fresh air volume (m3/h); E refers to total source intensity (mg/m3∙h); ks is
270
the loss ratio due to chamber characteristics; Kn is the decline rate of the pollutant
271
when there is no PAC in the room. Obviously, CADR can be calculated from the
272
difference of the decay rate of the contaminant between experiments with and without
273
the PAC. For this purpose, the ‘pull-down’ method was used. It is recommended in
AC C
268
14
ACCEPTED MANUSCRIPT China national standard for air cleaners[23] and was applied in several prior
275
studies[12-16].
276
Firstly, certain amount of the targeted contaminant was injected into the experimental
277
space. After the concentration reached the peak and stabilized for 10 min, the PAC
278
was turned on. The concentration was measured at a certain interval during the decay
279
period, and the time when the PAC started to operate was defined as time zero. Since
280
the concentration established in the chamber was much higher than that in the outdoor
281
environment, Cout can be approximated to zero. Source intensity E also approximated
282
zero during the decay period. Whereupon, Equation (3) can be simplified to Equation
283
(4). Thus, by employing a linear regression analysis of ln(
284
obtained.
285
C CADR ln = − K n + t V C0
286
Although many researchers used chambers that had volumes bigger than 30 m3 in
287
their experiments[15, 16], chambers smaller than 10 m3 (this study used an 8 m3
288
chamber) were also efficient as long as test conditions were properly modified. First
289
of all, the test duration and the interval between samplings were shortened. Because
290
with the same PAC operating, concentration of the contaminant decays faster in the
291
smaller chamber. Shortened test duration insured that the concentration did not drop
292
below the detection limit of the gas inspection method. On the other hand, Chen et al.
293
noticed[16] that CADR was not constant when the test duration became excessively
294
long. In other words, the ‘pull-down’ method which assumes a constant CADR is
295
inappropriate for long-term performance evaluation[16]. The CADR was nearly
296
constant at the expense of limiting the evaluation to only an initial performance
297
evaluation. Additionally, we increased the amount of the injected contaminant to
C ) , the CADR can be C0
(4)
AC C
EP
TE D
M AN U
SC
RI PT
274
15
ACCEPTED MANUSCRIPT obtain reasonable initial concentrations. This insured the amount of adsorbed
299
formaldehyde in each sample after the sampling time was shortened.
300
Consequently, questions came down to finding out the relationship between C0 and
301
the maximum measurable CADR (CADRmax ) . If the minimum detectable
302
formaldehyde content of the MBTH method and the sample volume are known, the
303
minimum allowable final concentration (Cmin) can be calculated. Obviously, the
304
maximum measurable CADR reduces the concentration to its inferior limit within the
305
test time. Based on the mathematic model mentioned above, the following inequations
306
can be established.
307
Ct ≥ C min
308
CADRmax − kn + V
309
(6)
310
Here, t refers to the test duration and Ct refers to the concentration at the end of the
311
test. We suppose that the chamber volume is 8 m3 and C0 varies from 1 mg/m3 to 15
312
mg/m3. Fig. 7 shows the correlation between the CADRmax and test duration at
313
different C0 levels. As illustrated by the curve, when C0 reaches the value stipulated
314
by the Chinese national standard test protocol for PACs, which is 1.0 mg/m3, the test
315
duration should be reduced to 10 min if devices with a CADR greater than 100 m3/h
316
are to be examined. However, this is impractical because sufficient data points are
317
required to improve the quality of the regression model. On the other hand, if samples
318
are drawn every 10 min to obtain 7 data points, the CADRmax will be restricted to 50
319
m3/h even when the C0 is increased to 15 mg/m3. In addition, although higher C0
320
allows higher CADRmax, this benefit decreases with increasing C0. To sum up, we
321
conducted some preliminary tests to obtain a general estimation of the capacity of
322
PACs, according to which the final test conditions were determined (see Table. 2).
(5)
M AN U
SC
RI PT
298
AC C
EP
TE D
Ct Cmin ⋅ t = ln C ≥ ln C 0 0
16
ACCEPTED MANUSCRIPT 323
Our results showed that the data quality improved significantly when the modified
324
method was applied.
M AN U
SC
RI PT
325
326
Fig. 7 Correlation between the CADRmax and test duration at different C0 levels: a
328
higher CADRmax can be measured within the same period of time with a higher C0, but
329
this boosting effect decreases with increase in C0
TE D
327
330
Table 2 Comparison of standard and modified test method
332
Volume
Duration
(m3)
(min)
AC C
Method
EP
331
Sampling
Sampling
interval
duration
(min)
(min)
Flow rate (L/min)
Initial Concentration (mg/m3)
Standard [23]
30
60
10
5
0.5
1.0-1.2
Modified
8
30
5
2
0.5
2.5-3.5
333
2.5 Sampling and analyzing the samples
334
Samples were drawn by an air pump (Beijing Municipal Institute of Labor Protection, 17
ACCEPTED MANUSCRIPT QC-2 atmospheric sampler, China) through a stainless-steel tube located at the center
336
of the chamber and a section of Teflon pipe. The flow rate was set at 0.50 L/min by a
337
float flowmeter in order to provide intensive contact between the air sample and the
338
absorbent i.e., MBTH (0.05 mg/ml, 5 ml). The sample volume was calibrated for
339
temperature and pressure differences between the test condition and the standard
340
condition (equation 7):
341
Qc = Q ⋅
342
Qc is the calibrated sample volume; Q is the original sample volume; Ts and Ps are the
343
standard temperature (273.15 K) and the standard atmospheric pressure (101.35 kPa),
344
respectively; T and P are the real temperature and atmospheric pressure, respectively.
345
The first sample, based on which the initial concentration of formaldehyde was
346
calculated, should be drawn about 10 min after the release of formaldehyde gas. The
347
air cleaner to be tested was turned on after the first sample was drawn and the rest of
348
the samples were drawn at an interval of 5 min for a period of 30 min. The entire
349
procedure is illustrated in Fig. 8. Before releasing the contaminant, a sample, which
350
served as the empty sample, was taken to remove the impact of background
351
formaldehyde concentration.
RI PT
335
TS P ⋅ T PS
352
AC C
EP
TE D
M AN U
SC
(7)
353
Fig. 8 Sampling process of the PAC test in the environmental chamber
354
The basic procedures used in the field test were similar to those used in the laboratory. 18
ACCEPTED MANUSCRIPT However, a subtle distinction should be noted; real rooms usually have a much higher
356
permeability than the environmental chamber used in the laboratory, which means that
357
the air exchange rate in real rooms is strongly affected by complex factors such as
358
temperature and pressure difference. Therefore, each air cleaner test should be
359
coupled with a blank test to determine the natural decay rate (kn). The interval
360
between these two tests should be as short as possible. Meanwhile, CO2 and SF6 were
361
released as tracer gas to measure the ventilation rate. This was done to check if the
362
saltation in fresh air volume radically changes the natural decay rate. CO2 was
363
monitored by TES-1370 portable CO2 detectors, and SF6 was sampled by
364
INNOVA-1303 multipoint sampler and analyzed by INNOVA-1412i photoacoustic
365
gas monitor. These devices did not release or adsorb formaldehyde. Meanwhile, the
366
reagent used for formaldehyde measurement did not adsorb CO2 or SF6.
367
The samples were promptly taken to the laboratory after the test and analyzed using a
368
spectrophotometer at a wavelength of 630 nm. Before spectrophotometric analysis,
369
400 µL of ferric sulfate solution (0.10 mol/L) was added to the samples and allowed
370
to react for at least 15 min to develop color. Hence, absorbance of the sample was
371
measured 20 min after the chromogenic reagent was added. Every absorbance value
372
corresponded to a particular value of formaldehyde content on the standard curve
373
plotted beforehand; the formaldehyde concentration was calculated by dividing the
374
formaldehyde content by the sample volume, which was calibrated according to
375
temperature. By taking the natural logarithm of formaldehyde concentrations and
376
performing linear regression, the slope of the fitting line that indicates the total decay
377
rate was determined. The product of the slope and space volume is numerically equal
378
to the CADR. Only the lines with an R-squared value greater than 0.90 were
379
considered.
AC C
EP
TE D
M AN U
SC
RI PT
355
19
ACCEPTED MANUSCRIPT 3. Results and Discussion
381
3.1 Measured CADRs of PACs
382
Fig. 9 provides a visual comparison of the average CADR (of repetitive tests) of the
383
four PACs. The number inside each column refers to the average CADR, and the
384
highest point of the error bar is the maximum while the lowest point is the minimum
385
CADR of repetitive tests.
386
TE D
M AN U
SC
RI PT
380
Fig. 9 Histogram with error bars of the averaged CADR of the four PACs
388
Filters were changed after they had been used for three times. In fact, the adsorbed
389
mass of formaldehyde in each test was around 15 mg, and the cumulate cleaned mass
390
(CCM) of the PACs, which refers to the total mass of the cleaned contaminant when
391
CADR drops to 50% of the initial value, was at least 1500 mg. The CADRs of PAC-A
392
and PAC-C appear to be very small for the real rooms. Considering a regular bedroom
393
with an area of 20 m2 and a volume of about 60 m3, PAC-A can only provide an
394
equivalent ventilation rate of 0.23 h-1 for formaldehyde while PAC-C can provide an
395
equivalent ventilation rate of 0.41 h-1 for formaldehyde. For instance, if the initial
396
formaldehyde concentration is 0.6 mg/m3, which is usually the case in newly
AC C
EP
387
20
ACCEPTED MANUSCRIPT furnished residential rooms in Beijing [24], the use of PAC-A can only lower the
398
formaldehyde concentration to 0.12 mg/m3 (approximately the upper limit specified in
399
the Chinese national Standard) after 7 h.
400
PAC-D outperformed all the other PACs with an average CADR of 75.6 m3/h. This
401
suggests its ability to provide an equivalent ventilation rate of 1.26 h-1 in a regular
402
bedroom, which rivals mechanical methods of ventilation. Analyzing reasons for
403
PAC-D’s outstanding performance was beyond the scope of this paper. Nevertheless, a
404
reasonable hypothesis could be made by inspecting the mass transfer process. The
405
adsorption of VOCs involves three main stages: 1) mass transfer from bulk to sorbent
406
surface; 2) diffusion inside the solid; 3) adsorption and desorption onto the pore
407
surfaces [25]. Firstly, in order to be compacted into interlayers of the HEPA filter,
408
activated carbon should be processed into super fine particles; the activated carbon
409
granules incorporated into PAC-D were the smallest in size. It has been proved that
410
finer particle size is advantageous for the mass transfer process[26] because it
411
promotes sufficient contact. Meanwhile, the density of the HEPA filter was increased
412
and therefore the volume of air that could pass through them was reduced.
413
Performance of the filter in the subsequent stages largely depends on the
414
characteristics of the activated carbon incorporated into the filter.
415
Among the PACs tested, PAC-B was the only one to use chemisorption as a dominant
416
mechanism for formaldehyde removal. As mentioned earlier, PAC-B was mounted
417
with a cardboard-like filter shown in Fig. 10, and claimed to be a specialized
418
formaldehyde scrubber. Chemisorption is rarely used alone; porous materials that
419
promote physical adsorption are often used as media to provide reaction sites. The
420
most common porous material used is modified activated carbon; a number of studies
421
have proved that it enhances chemisorption[27, 28]. Oddly, it was observed that the
422
media of the ‘formaldehyde removal filter’ of PAC-B could not develop a porous
AC C
EP
TE D
M AN U
SC
RI PT
397
21
ACCEPTED MANUSCRIPT structure. Compared to activated carbon with a specific surface area that usually
424
exceeds 1000 m2/g, the pleats designed to increase contact surface in the
425
cardboard-like filter was rather insufficient. Since physical adsorption on this filter is
426
constrained, it needs to form abundant functional groups on its surface to let
427
chemisorption play the dominant role. Previous studies have reported that the
428
potential key parameters that affect chemisorption include relative humidity,
429
significant change in concentration of the targeted contaminant, and the media
430
incorporated into the filter[29]. In our tests, the concentration only had minor changes,
431
hence it did not have a strong effect on the CADR of PAC-B.
SC
RI PT
423
TE D
M AN U
432
433
Fig. 10 Structure and size information of the ‘formaldehyde removal filter’ used in
435
PAC-B.
436
Unfortunately, all the PACs that were tested had a CADR that was lower than their
437
nominal values. Manufacturers and certifying authorities have a somewhat casual
438
attitude regarding the VOC removal ability of PACs because the market demand for
439
PACs with particulate matter removal ability is higher. Consequently, the metrics
440
related to particulate matter are cautiously inspected and has a more reliable nominal
441
value. We tested the CADR of PAC-B for removal of particulate matter as well. An
442
average CADR of 282 m3/h was observed, which is close to its nominal value (307
AC C
EP
434
22
ACCEPTED MANUSCRIPT m3/h). Customers are readily satisfied by the efficient removal of particulate matter by
444
PACs, and usually do not care about the comparatively insufficient VOC removal
445
capacities. To make explicit, the applicable area is calculated with reference to the
446
CADR for particulate matter so that it is likely to be oversized for VOC removal.
447
Additionally, occupants usually close windows when a PAC is being operated, thus
448
creating a condition for VOCs to accumulate.
449
However, there was some relevance between the actual CADR and the nominal value;
450
the performance of PACs with a higher nominal CADR was better than the
451
performance of PACs with a lower nominal CADR. Another comforting news is that
452
in terms of formaldehyde removal, the overall performance of the PACs tested in this
453
study was remarkably better than the performance of products that were reported in
454
previous studies: the highest CADR of the 18 household air cleaners tested by Kim et
455
al. did not exceed 24 m3/h[30]; the CADR of the best performing PAC tested by Chen
456
et al. was merely 10.9 m3/h[16].
457
One limitation was that the ‘pull-down’ method can only test the initial performance
458
of PACs. Evaluation based on initial performance cannot be arbitrarily extended to
459
long-term performance. Because in addition to the rate of mass transfer process and
460
adsorption, adsorption capacity also has significant influence on long-term
461
performance. A PAC showing satisfactory initial performance does not necessarily
462
perform well in long term especially when the filter saturates easily. Therefore, the
463
initial performance is not enough to estimate a PAC. Long-term performance deserves
464
further study as well.
SC
M AN U
TE D
EP
AC C
465
RI PT
443
466
3.2 Comparison between laboratory and field tests
467
Proper attention was paid to the comparability between the laboratory and field test 23
ACCEPTED MANUSCRIPT before analyzing the results. The influences of test conditions including temperature,
469
relative humidity, and initial concentration need to be specified because these
470
parameters were stably maintained in the environmental chamber but fluctuated in
471
real situations. Therefore, PAC-D was further examined under different temperatures,
472
relative humidity (RH), and initial concentrations (C0) to cover their variation range in
473
the field test. Table 3 lists all the test conditions and Fig. 11 provides the visual
474
comparison of the CADR of PAC-D under different test conditions. Although from the
475
perspective of adsorption mechanism, temperature is associated with several physical
476
parameters and water molecules compete with formaldehyde for active adsorption
477
sites. The values of both the parameters are far from extreme in normal rooms and
478
fluctuate within a relatively narrow range; PAC-D seemed robust enough to resist
479
these fluctuations. No significant change in the results was observed when the initial
480
concentration rose from 1.1 mg/m3 to 2.8 mg/m3. However, it should be noted that this
481
was only a minor change in initial concentration; different results could be obtained
482
when the gap becomes wider. Additionally, sensitivity to these test conditions varied
483
with different adsorbents. The influence may be significant for other adsorbents[31].
SC
M AN U
TE D AC C
EP
484
RI PT
468
485 24
ACCEPTED MANUSCRIPT 486
Fig. 11 CADR of PAC-D under varying (within a limited range) temperatures, relative
487
humidity, and initial formaldehyde concentrations. Parameter specifications of “low”
488
and “high” can be seen in Table 3.
489
Table 3. Test conditions for the studies on the influence of variations (within limited
491
ranges) in relative humidity, temperature, and initial concentration on PAC
492
performance
PAC-D
High RH (%)
Temperature (˚C)
C0 (mg/m3)
47.7
75.1
23.0
2.5
Low Temperature (˚C)
High Temperature (˚C)
RH (%)
C0 (mg/m3)
18.6 Low C0 (mg/m3) 1.1
493
SC
Low RH (%)
M AN U
PAC code
RI PT
490
29.4
51.5
2.7
High C0 (mg/m3)
Temperature (˚C)
RH (%)
2.8
25.0
50.3
Field test is challenging mainly because of the following reasons. First, the sink and
495
source effect in real situations is complicated and sometimes cannot be overlooked
496
even in unfurnished rooms. However, the universal ‘pull-down’ method is based on a
497
model that does not consider this effect (emission rate = 0), hence it presumably
498
brings about poor regression quality. Second, real rooms have much higher air
499
exchange rates than environmental chambers. Subtracting such high air exchange
500
rates from the total decay rate to obtain the CADR causes significant errors. The air
501
exchange rate depends on multiple factors such as temperature and air velocity, and is
502
likely to fluctuate during a test. This fluctuation can alter the sink and source
503
characteristics indoors. Lastly, real rooms are not mounted with necessary equipment
504
and do not have any specialized sampling apertures. Since formaldehyde usually
505
cannot be accurately measured by online equipment, frequent sampling, which
AC C
EP
TE D
494
25
ACCEPTED MANUSCRIPT unavoidably interferes with the conditions in the room, is required. Therefore, perfect
507
conformity between the conditions in the field and laboratory test is unrealistic; a
508
more sensible method is conducting a divergence analysis based on descriptive
509
statistical data.
510
Only part of the laboratory data was selected for comparison. Tests conducted under
511
similar conditions, i.e. temperature, relative humidity, and initial concentrations, were
512
selected for comparison. Table 4 summarizes all the test results and the corresponding
513
test conditions. Category ‘Field-A’, ‘Field-B’, and ‘Field-C’ refers to the field test
514
with infiltration, moderate ventilation, and full ventilation respectively. The average
515
values of the CADR (field test: 72.1 m3/h; laboratory: 74.5 m3/h) were very close, but
516
discrepancy between their maximum and minimum values was observed. This was
517
likely because of the scattered characteristics of the field test data; the standard
518
deviation of field test data was 12.1. By contrast, data from the laboratory tests had a
519
standard deviation of 2.3, which suggests better stability.
M AN U
SC
RI PT
506
520
Table 4. Conditions and results of the laboratory and field tests. T is the temperature,
522
RH is the relative humidity, C0 is the initial concentration, ACH is the air exchange
523
rate
TE D
521
RH (%)
C0 (mg/m3)
ACH (h-1)
CADR (m3/h)
26.7
62.6
1.04
0.59
91.2
Field-A
26.7
48.9
1.08
0.58
69.7
Field-A
31.4
50.6
1.04
0.56
77.0
Field-B
26.9
75.0
1.05
0.58
74.5
Field-B
31.3
49.0
1.13
0.65
62.8
AC C
Field-A
EP
T( )
Category
26
ACCEPTED MANUSCRIPT 23.2
33.0
1.23
0.61
56.2
Field-B
24.2
56.4
1.04
0.72
84.6
Field-C
27.1
69.6
1.08
0.96
60.9
Laboratory
18.6
50.5
2.37
<0.01
77.1
Laboratory
26.8
50.1
2.80
Laboratory
29.6
50.3
2.95
Laboratory
23.2
50.6
2.38
<0.01
76.2
Laboratory
22.9
47.7
2.49
<0.01
72.3
Laboratory
23.3
Laboratory
22.7
RI PT
Field-B
73.1
<0.01
76.3
M AN U
SC
<0.01
65.2
1.06
<0.01
71.0
75.1
1.32
<0.01
75.3
Field-A: field test with infiltration only (windows closed); Field-B: field test with moderate
525
natural ventilation (windows partially open); Field-C: field test with full natural ventilation
526
(windows fully open)
527
We used IBM SPSS Statistics V22.0 to check whether there is a significant difference
528
between the field and laboratory data. The output is presented in Table 5. The
529
significance of the null hypothesis that assumes equal variances is 0.006, which is
530
much smaller than 0.05, therefore unequal variances should be accepted. The bottom
531
section of the table shows the indicator of difference analysis i.e., sig, which is 0.613
532
for this case. This value is significantly higher than 0.05, suggesting that the
533
difference is insignificant at a confidence level of 95%.
AC C
EP
TE D
524
534 535
Table 5. Independent samples U test output
27
ACCEPTED MANUSCRIPT Laboratory
Field
N
7
8
Mean
74.5
72.1
Standard Deviation
2.32
12.08
F
10.648
Levene's Test for Equality of Variances Sig.
Sig.
0.006
0.613
SC
Independent Samples-Mann-Whitney U Test
RI PT
CADR
536
Our results indicate that the behavior of PAC-D in the field was similar to that in the
538
environmental chamber. However, it is important to keep in mind that the effect will
539
be completely different for the field and laboratory because it depends on the relative
540
magnitude of the CADR and volume of the space. The performance of PAC-D may be
541
satisfactory for a 60 m3 room but it can hardly meet the demands of a 120 m3 room.
542
TE D
M AN U
537
4. Conclusion
544
This study estimated the formaldehyde removal abilities of four sorption-based
545
portable air cleaners (PACs). Tests were conducted both in the laboratory and field in
546
order to compare the performance of PACs in the well-controlled laboratory
547
environment to that in real situation. The following conclusions could be drawn from
548
this study:
549
1) The CADRs of the four PACs tested in the laboratory had great discrepancies.
550
Values ranged from 13.8 m3/h to 75.6 m3/h. The best performing PAC used a
551
composite filter with ultra-fine activated carbon sandwiched between the
AC C
EP
543
28
ACCEPTED MANUSCRIPT 552
interlayers of the HEPA filter. This cut down airflow bypass and promoted
553
sufficient contact between the air to be treated and the composite filter.
554 555
2) Difference between laboratory and field results was insignificant at a confidence level of 95%.
RI PT
556
Acknowledgements s
558
This study was supported by the National Key Basic Research and Development
559
Program of China (Grant No. 2016YFC0700500) and the Innovative Research Groups
560
of the National Natural Science Foundation of China (Grant No. 51521005).
M AN U
SC
557
561 562
References
563
[1]. K.H. Kim, S.A. Jahan, J.T. Lee. Exposure to Formaldehyde and Its Potential Human Health Hazards. Journal of Environmental Science and Health, Part C
565
2011, 29: 277-299
568 569 570 571 572
2002 (in Chinese)
EP
567
[2]. GB/T 18883-2002, Indoor Air Quality Standard, Ministry of Health, Beijing,
[3]. T. Salthammer, S. Mentese, R. Marutzky. Formaldehyde in The Indoor Environment. Chemical Reviews 2010, 110: 2536-2572
AC C
566
TE D
564
[4]. R. J. Shaughnessy, R. G. Sextro. What Is an Effective Portable Air Cleaning Device? A Review. Journal of Occupational and Environmental Hygiene 2006, 3: 169-181
573
[5]. Y. P. Zhang, J. H. Mo, J. Sundell, P. Wargocki, J. Zhang, J. C. Little, R. Corsi, Q.
574
H. Deng, M. H. K. Leung, L. Fang, W. H. Chen, J. G. Li and Y. X. Sun. Can
575
Commonly-Used Fan-Driven Air Cleaning Technologies Improve Indoor Air 29
ACCEPTED MANUSCRIPT 576 577
Quality? A Literature Review. Atmospheric Environment 2011, 45: 4329-4343 [6]. A. Bastani, C. S. Lee, F. Haghighat, C. Flaherty and N. Lakdawala. Assessing the
578
Performance of Air Cleaning Devices – A Full-Scale Test Method. Building and
579
Environment 2010, 45:143-149 [7]. M. A. Sidheswaran, H. Destaillats, D. P. Sullivan, S. Cohn and W. J. Fisk. Energy
RI PT
580 581
Efficient Indoor VOC Air Cleaning with Activated Carbon Fiber (ACF) Filters.
582
Building and Environment 2012, 47:357-367
583
[8]. C. S. Lee, F. Haghighat. Experimental Evaluation of the Performance of
Gas-Phase Air Filters Using a Dynamic Closed-Loop Test System. ASHRAE
585
Transactions, Volume 112, Part 2
[9]. H. Matsumoto, M. Shimizu and H. Sato. The Contaminant Removal Efficiency of
M AN U
586
SC
584
587
an Air Cleaner Using the Adsorption/Desorption Effect. Building and
588
Environment 2009, 44:1371-1377
589
[10].
A. T. Hodgson, H. Destaillats, D. P. Sullivan and W. J. Fisk. Performance of
Ultraviolet Photocatalytic Oxidation for Indoor Air Cleaning Applications. Indoor
591
Air 2007, 17: 305-316
TE D
590
592 593 594
[11]. C. Q. Yang, X. D. Yang, B. Zhao. Person to person droplets transmission characteristics in unidirectional ventilated protective isolation room: The impact of initial droplet size. Building Simulation, 2016. 9: 597-606
595
[12].
J.M. Daisey, A.T. Hodgson. Initial Efficiencies of Air Cleaners for the
Removal of Nitrogen Dioxide and Volatile Organic Compounds. Atmospheric
597
Environment 1989, 23: 1885-1892
599 600 601 602 603 604
AC C
598
EP
596
[13].
P. R. Nelson, S. B. Sears and D. L. Heavner. Application of Methods for
Evaluating Air Cleaner Performance. Indoor Environ 1993, 2: 111-117
[14].
R. J. Shaughnessy, E. Levetin, J. Blocker and K. L. Sublette. Effectiveness of
Portable Indoor Air Cleaners: Sensory Testing Results. Indoor Air 1994, 4: 179-188
[15].
J. L. Niu, T. C. W. Tung and V. W. Y. Chui. Using Large Environmental
Chamber Technique for Gaseous Contaminant Removal Equipment Test. 30
ACCEPTED MANUSCRIPT 605 606
ASHRAE Transactions 1998, 104 [16].
W. Chen, J. S. Zhang and Z. Zhang. Performance of Air Cleaners for
607
Removing Multiple Volatile Organic Compounds in Indoor Air. ASHRAE
608
Transactions 2005, 111(1): 1101-1114
610 611
[17].
C. H. Reed, S. J. Nabinger and S. J. Emmerich. Characterizing Gaseous Air
RI PT
609
Cleaner Performance in the Field. Building and Environment 2008, 43: 368-377 [18].
C. H. Reed, V. Henzel, S. J. Nabinger and A. K. Persily. Development of a
Field Test Method to Evaluate Gaseous Air Cleaner Performance in a Multizone
613
Building. Journal of the Air & Waste Management Association 2008, 58:
614
7,919-927 [19].
A. Persily, C. H. Reed, S. Watson, N. Martys, E.J. Garboczi and H. Davis.
M AN U
615
SC
612
616
NISTIR 7525: Standards Development for Gas Phase Air Cleaning Equipment in
617
Buildings. U.S. National Institute of Standards and Technology, 2008
618
[20].
N. Nachar. The Mann Whitney U: A Test for Assessing Whether Two
Independent Samples Come from the Same Distribution. Tutorials in Quantitative
620
Methods for Psychology 2008, vol. 4: 13-20
621 622 623
[21].
TE D
619
J. M. Stonehouse, G. J. Forrester. Robustness of the t and U tests under
combined assumption violations. Journal of Applied Statistics 1998, 25: 63-74 [22].
W. H. Liang, S. Yang and X. D. Yang. Long-Term Formaldehyde Emissions
from Medium-Density Fiberboard in a Full-Scale Experimental Room: Emission
625
Characteristics and the Effects of Temperature and Humidity. Environmental
627 628 629 630 631 632
AC C
626
EP
624
Science & Technology 2015, 49: 10349–10356.
[23].
GB/T 18801-2015, Air Cleaner, General Administration of Quality
Supervision, Inspection and Quarantine of the People's Republic of China & Standardization Administration of the People's Republic of China, Beijing, 2015 (in Chinese)
[24].
D. Q. Xu, X. Y. Dong, G. F. Wang, Z. G. Tang and Y. P. Li. Investigation of
Indoor Air Organic Pollution in Decorated House in Beijing. Chinese Journal of 31
ACCEPTED MANUSCRIPT 633 634
Health Laboratory Technology 2006, 16: 1281-1283 (in Chinese) [25].
T. B. Cheng, Y. Jiang, Y. P. Zhang and S. Q. Liu. Prediction of Breakthrough
635
Curves for Adsorption on Activated Carbon Fibers in a Fixed Bed. Carbon 2004,
636
42: 3081-3085
638 639
[26].
D. Das, V. Gaur and N. Verma. Removal of Volatile Organic Compound by
Activated Carbon Fiber. Carbon 2004, 42: 2949-2962 [27].
RI PT
637
F. Lin, G. Q. Zhu, Y. N. Shen, Z. Y. Zhang and B. Dong. Study on the
Modified Montmorillonite for Adsorbing Formaldehyde. Applied Surface Science
641
2015, 356:150-156
642
[28].
SC
640
S. Saeung, V. Boonamnuayvitaya. Adsorption of Formaldehyde Vapor by
Amine-Functionalized Mesoporous Silica Materials. Journal of Environmental
644
Sciences 2008, 20: 379-384
645
[29].
M AN U
643
J. J. Pei, J. S. S. Zhang. On the Performance and Mechanisms of
646
Formaldehyde Removal by Chemi-sorbents. Chemical Engineering Journal 2011,
647
167:59-66 [30].
H.J. Kim, B. Han, Y.J. Kim, Y.H. Yoon, T. Oda. Efficient Test Method for
TE D
648
Evaluating Gas Removal Performance of Room Air Cleaners Using FTIR
650
Measurement and CADR Calculation. Building and Environment 2012, 47:
651
385-393
653 654 655 656
[31].
B. Yu, N. Li, W. He, J. Ji, S. Zhang, H. Chen. Multifunctional Solar Wall for
Dehumidification, Heating and Removal of Formaldehyde: Part 1. System
AC C
652
EP
649
Description, Preparation and Performance of SiO2/TiO2 adsorbent. Building and Environment 2016, 100: 203-214
32
ACCEPTED MANUSCRIPT
Research Highlights Formaldehyde removal by portable room air cleaners were tested in the laboratory and the field Clean air delivery rates of tested air cleaners varied from 13.8 m3/h to 75.6 m3/h
RI PT
The laboratory and field performances of the tested air cleaner matched
AC C
EP
TE D
M AN U
SC
reasonably well