Long-Term Formaldehyde Emissions from Medium-DensityFiberboard in a Full-Scale Experimental Room: EmissionCharacteristics and the Effects of Temperature and HumidityWeihui Liang, Shen Yang, and Xudong Yang*

Department of Building Science, Tsinghua University, Beijing 100084, P. R. China

*S Supporting Information

ABSTRACT: We studied formaldehyde emissions from themedium-density fiberboard (MDF) in a full-scale experimentalroom to approximate emissions in actual buildings. Detailedindoor formaldehyde concentrations and temperature andhumidity data were obtained for about 29 months. Temperature,relative humidity (RH), and absolute humidity (AH) ranged over−10.9−31.4 °C, 46.5−83.6%, and 1.1−23.1 g/kgair, respectively.Annual cyclical seasonal variations were observed for indoorformaldehyde concentrations and emission rates, exhibitingentirely different characteristics than those in an environmentalchamber under constant environmental conditions. The max-imum concentration occurred in summer rather than at initialintroduction of the material. The concentrations in summercould be a few up to 20 times higher than that in winter, depending on the indoor temperature and humidity conditions.Concentrations decreased by 20−65% in corresponding months of the second year. Indoor formaldehyde concentrations werepositively correlated with temperature and AH but were poorly correlated with RH. The combined effects of temperature andAH on formaldehyde emissions from MDF in actual buildings were verified. These detailed long-term experimental results couldbe used with environmental chamber measurement data to scale up and validate emission models from chambers held at constantconditions to actual buildings.

■ INTRODUCTIONFormaldehyde is one of the most ubiquitous gaseous pollutantsindoors.1 Acute exposure to formaldehyde could cause nose,eye, and throat irritation;2−4 chronic exposure could increaserespiratory illness and even cancer.5−7 The World HealthOrganization has recommended that formaldehyde concen-tration in air should be below 0.1 mg/m3 as a 30 min average.8

This value is also adopted by the Chinese national standard forthe 1-h exposure limit.9 However, indoor formaldehydeconcentration in 70% of the newly built or remodeled housesexceeded the national standard in China,10 with a mean value of0.238 mg/m3. This was much higher than that in developedcountries, which were 0.05 mg/m3 in Germany and 0.037 mg/m3 in Denmark.1 In actual buildings, indoor temperature andhumidity are generally not constant11,12 and vary with outdoorenvironmental conditions as well as occupant behavior.13,14

Consequently, formaldehyde is emitted under variable temper-ature and humidity conditions. Understanding the emissioncharacteristics of building materials in a realistic indoorenvironment is needed to model indoor exposures and mitigatehealth risks.Emissions from building materials have previously been

studied under conditions quite different from those in actualbuildings. The most common approach is the use of anenvironmental chamber or field and laboratory emission cell in

which emission measurements are conducted at constanttemperature, relative humidity (RH), and ventilation rate.15,16

Generally speaking, when a material is tested in a ventilatedenvironmental chamber, the concentration in the air of thechamber typically increases to a maximum value in the first fewhours and then continuously decreases, with the highestemission rate at the beginning of the experiment. Empiricalmodels17 and diffusion mass transfer models18−20 have beendeveloped to describe the emission behaviors of materials basedon these studies. However, under the variable environmentalconditions in actual buildings, emission characteristics ofbuilding materials may be different. Thus, emissions underrealistic conditions may better replicate emission characteristicsin the built environment.More realistic conditions are typically evaluated through field

studies, which are time-consuming, expensive, and moredifficult to carry out than studies using environmentalchambers. Most field studies conduct short-term (days toweeks) measurements, with few evaluating long-term (morethan one year) concentrations as would be found in actual

Received: May 11, 2015Revised: August 5, 2015Accepted: August 11, 2015

Article

pubs.acs.org/est

© XXXX American Chemical Society A DOI: 10.1021/acs.est.5b02217Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Dow

nloa

ded

by U

NIV

VE

RA

CR

UZ

AN

A o

n A

ugus

t 26,

201

5 | h

ttp://

pubs

.acs

.org

P

ublic

atio

n D

ate

(Web

): A

ugus

t 21,

201

5 | d

oi: 1

0.10

21/a

cs.e

st.5

b022

17

buildings. Wolkoff et al.21 measured formaldehyde concen-trations in two semi-detached twin apartments for one year.They found that formaldehyde concentrations were higher insummer in the vacant apartment, while decreased and thenincreased during autumn in the occupied one. Pollutanttransport occurred between these two apartments, whichcomplicated the situation. Brown22 measured formaldehydeconcentrations four times over eight months in an occupieddwelling. Formaldehyde concentrations were found to decaydouble exponentially. However, the above studies could notdirectly characterize emissions from materials because humanactivities may also have contributed to the observed form-aldehyde concentrations, and multiple emission sources mayhave been present in the studied spaces. Crump et al.23

measured formaldehyde in four unoccupied test houses for 24months to investigate the relationship between sources andindoor concentrations. As multiple materials were present inthe house, it was difficult to distinguish emission from eachindividual source. Thus, long-term measurements conducted inactual houses without interference by human behavior andmultiple sources are needed to understand realistic materialsemission characteristics. Existing field studies also have notcontinuously measured environmental conditions. The fre-quency of concentration measurements and the lengths of thestudies may also have been insufficient considering the long-term effects of formaldehyde emissions on indoor air quality.The main objectives of this study are to investigate the long-

term formaldehyde emission characteristics of buildingmaterials under variable temperature and humidity conditions,to explore the extent to which formaldehyde emissions vary byseason, to identify the effects of temperature variations onemissions, and also to explore the importance of humidity onformaldehyde emissions. We measured formaldehyde emissionsfrom a medium-density fiberboard (MDF) for more than 29months using a full-scale experimental room. Indoor form-aldehyde concentrations and air ventilation rates weremeasured frequently throughout the study. Temperature andhumidity were monitored continuously over the entire timeperiod.

■ MATERIALS AND METHODS

The Full-Scale Experimental Room. Dimension of thefull-scale experimental room was 4 m × 3 m × 3 m (length ×width × height). It was located in the corner of an experimentalbase in a rural district of Beijing. The total area of theexperimental base was 5000 m2, and a 2.5 m high enclosure wasbuilt around it. There was only one door in the south wall ofthe room, which remained closed during the entire measure-ment period. Figure 1 gives the digital images and schematicillustration of the experimental room and its wall assembly. Theroom layout and interior surface materials shared similarfeatures as a typical room in China, but the room possessedseveral key features to better serve the goal of this particularstudy, as described below.

(1) The room was unoccupied. Thus, all human-relatedinterferences to emissions could be eliminated. Anotheradvantage was that measurements could be conducted atany time without disturbing occupants or requiringpermission.

(2) The construction of the room was completed two yearsbefore the experiment begun, and only latex wall paintand ceramic floor tiles were used as interior construction

materials. There were no formaldehyde emissions fromthe ceramic floor tiles, and emissions from the latex wallpaint were negligible by the time the experiment began.Thus, the only formaldehyde emission source in thisroom was the MDF introduced. Formaldehyde is a smallmolecule compound whose adsorption capacity is not asstrong as other VOCs and aldehydes.24 Previousexperimental studies indicated that sorption of form-aldehyde by wall paint was insignificant.25

(3) Temperature and humidity were allowed to varynaturally in the room, without the use of heating or airconditioning equipment. In this way, temperature andhumidity ranges were much larger than that commonlyobserved in actual homes. This would allow betterinvestigation of the possible impact of temperature andhumidity on material emissions.

(4) A fan was operated to ensure good mixing of air in theroom, with the same design purpose as the fan in anenvironmental chamber.26

Studied Material. The studied material was a typical MDFdirectly purchased from the manufacturer. It was transportedand placed inside the room immediately after being received.Hence, it could be considered as a new material at thebeginning. The boards were leaned against the interior walls atan angle. Both the front and back surfaces of the boards emittedformaldehyde, which could be simplified to a one-dimensionalmaterial with half thickness and double emission area to theoriginal board. Key features of the studied MDF are listedbelow.

(1) Dimension of the original full-size MDF is 2.44 m × 1.2m × 0.012 m (length × width × thickness). Three full-size boards together with a small piece (dimension of0.18 m × 1.2 m × 0.012 m) were placed in the room,resulting in a loading ratio of 0.5 m2/m3. This value wascomparable to that of densely furnished spaces such as asmall room or kitchen but higher than most living orbedrooms in actual residential buildings.27 A relativelylarge loading ratio was selected to allow higher roomconcentrations over a long period of time to minimizethe error caused by measurement uncertainty.

Figure 1. Digital images of the interior and exterior of the full-scaleexperimental room and the schematic illustration of the wall assembly:(a) interior image, (b) exterior image, and (c) wall assembly.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b02217Environ. Sci. Technol. XXXX, XXX, XXX−XXX

B

Dow

nloa

ded

by U

NIV

VE

RA

CR

UZ

AN

A o

n A

ugus

t 26,

201

5 | h

ttp://

pubs

.acs

.org

P

ublic

atio

n D

ate

(Web

): A

ugus

t 21,

201

5 | d

oi: 1

0.10

21/a

cs.e

st.5

b022

17

(2) The MDF was directly exposed to air different from itstypical use as chairs surrounded by foam or tables andcountertops laminated by a plastic cover. This couldsimplify the emission process and better fit the mainpurpose of this study. More complicated and realisticfeatures could be added later after the fundamentalemission characteristics of the material were understood.

(3) Headspace analysis of the MDF indicated that form-aldehyde was the dominant emitted compound. Head-space concentrations of other aldehydes and ketoneswere 1 or 2 orders of magnitude lower than form-aldehyde, and the sum of other aldehydes and ketoneswas less than 5% of the formaldehyde mass. The totalVOC concentration in the MDF headspace was as low as15.7 μg/m3. Thus, potential generation of formaldehydefrom chemical oxidation of terpenes, alkenes, orhydrocarbons could be eliminated.28,29

The above features of the experimental room and studiedmaterial will benefit the understanding of long-term form-aldehyde emission characteristics as well as the quantification oftemperature and humidity effects. On the other hand,formaldehyde concentrations in the experimental room weremuch higher than that in actual homes due to large loadingratio, elimination of the retard resistance from the coversurrounding the MDF, and extreme summertime environ-mental conditions. Concentration differences in differentseasons could also be more pronounced due to larger indoortemperature and humidity ranges across seasons. The form-aldehyde concentrations measured in this experimental roomshould not be considered as typical in actual homes.Formaldehyde Measurements. Entry into the exper-

imental room could cause air exchange between outdoor andindoor air, affecting the indoor formaldehyde concentration. Toeliminate human interference in the concentrations, a samplingportal was placed near the door frame so that samples could becollected from outside the room without opening the door andentering. The portal was located 1.2 m above the floor.Formaldehyde was measured using the 3-methyl-2-benzothia-zolinone hydrazone (MBTH) spectrophotometric method, inaccordance with the Chinese national standard.30 The accuracyof this method was evaluated and is described in Section S1 ofSupporting Information. To collect formaldehyde samples, astainless steel tube was extended into the center of the roomthrough the sampling portal. Indoor air samples were pumpedthrough the stainless steel tube into a glass sampling tube usinga QC-2 pump (Beijing Institute of Labor Protection) at a flowrate of 100−200 mL/min. Formaldehyde in the air wasabsorbed into a solution in the glass tube containing 5.0 mL of50 μg/mL MBTH. The sampling time was 5−30 min,depending on the estimated formaldehyde concentration inthe air, with less volume collected at higher concentrations. Toanalyze the formaldehyde concentrations, 0.4 mL of 10 g/Lferric ammonium sulfate solution was added to the samplingtube. The tube was shaken and then held for 15 min, duringwhich formaldehyde was converted into a blue cationic dye bythe MBTH. The light absorbance was measured using aspectrophotometer at 630 nm (Unic 7200, China). Qualitycontrol and assurance of the method is detailed in Section S2 ofthe Supporting Information.Frequent sampling (ranging from a few days to a few weeks)

was conducted to evaluate variations and trends over time.Three duplicate samples were taken for each measurement. The

measurement period was 9 October 2012 to the end ofFebruary 2015. The indoor background formaldehyde concen-tration was also measured before introducing MDF. Outdoorformaldehyde concentrations were also occasionally measured.

Environmental Measurements. Indoor temperature andRH were measured by an automatic data logger (WSZY-1sensor, Beijing, China) placed in the center of the room. Theinstrument measurement ranges were from −40 to 100 °C fortemperature and 0−100% for RH with measurement errors of0.5 °C for temperature and 3% for RH. The sampling intervalwas 10 min. The ventilation rate was measured monthly in2013 using the CO2 decay method31 to analyze annualvariations.

■ RESULTS AND DISCUSSIONFormaldehyde Concentration. The measured indoor

formaldehyde concentrations are illustrated in Figure 2.

Relative standard deviations (RSD) were within 10% formost measurements, indicating good repeatability of the results.The mean value for each test was used for data analysis. Theyear was divided into spring (March−May), summer (June−August), autumn (September−November), and winter (De-cember−February) seasons. Indoor formaldehyde concentra-tion ranges and means for each season are listed in Table 1.Indoor formaldehyde concentrations did not decrease con-tinuously, instead having different trends in different seasons. Amonotonic increase in the concentration occurred in spring,reaching a maximum value of 4.78 mg/m3 in summer 2013 and3.19 mg/m3 in summer 2014. Concentrations then decreasedthroughout midlate summer and autumn and reached aminimum in winter. Both the levels and ranges differed indifferent seasons, with summer > spring > autumn > winter.The maximum concentration was 30−50 times the minimum,exhibiting significant differences in the same year. The meanconcentration in summer was 20 times that in winter.Therefore, indoor formaldehyde concentrations and trendswere strongly seasonal.Both the formaldehyde concentrations and ranges decreased

from one year to the next. The mean concentration was 3.59mg/m3 in the summer of 2013 and decreased to 2.52 mg/m3 inthe summer of 2014. Similar concentration decreases weremeasured in other seasons from 2013 to 2014. Generallyspeaking, the formaldehyde concentration decreased 20−65%

Figure 2. Indoor formaldehyde concentration and emissions rateprofiles in the experimental room. Aut., Win., Spri., and Sum. areabbreviations for autumn, winter, spring, and summer, respectively.Error bars are the standard deviations of the formaldehydeconcentration results.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b02217Environ. Sci. Technol. XXXX, XXX, XXX−XXX

C

Dow

nloa

ded

by U

NIV

VE

RA

CR

UZ

AN

A o

n A

ugus

t 26,

201

5 | h

ttp://

pubs

.acs

.org

P

ublic

atio

n D

ate

(Web

): A

ugus

t 21,

201

5 | d

oi: 1

0.10

21/a

cs.e

st.5

b022

17

in corresponding months of the second year. Concentrationranges also narrowed, for example, from 1.32−4.78 mg/m3 inthe summer of 2013 to 2.17−3.19 mg/m3 in the summer of2014. The above results indicate that formaldehyde emissionsdecreased annually within the seasonal trends noted above.

Formaldehyde Emission Characteristics. Emission RateCalculation. In the time-scale of several hours or a day, indoortemperature and humidity varied in a small range. Indoorenvironmental conditions were relatively stable. Consequently,a short-term steady state emission rate could be assumed. Theformaldehyde emission rate from MDF at a certain time t wasestimated based on the following equation:

= × −t C t C LER( ) VR ( ( ) )/amb (1)

where ER(t) is the emission rate (mg/m2·h) at time t, VR is theventilation rate of the experimental room (h−1), C(t) is themeasured indoor formaldehyde concentration (mg/m3) at timet, Camb is the outdoor formaldehyde concentration (mg/m3),and L is the loading ratio of the source material (0.5 m2/m3).The background formaldehyde concentration in the

experimental room was 0.021 mg/m3. The outdoor form-aldehyde concentration varied only slightly within 0.012 ±0.010 mg/m3 during the entire study period. Thus, there wasno significant difference between the outdoor and backgroundformaldehyde concentration, and the assumption that therewere no other formaldehyde emission sources in theexperimental room was confirmed. Additionally, no seasonaltrends were observed in the outdoor formaldehyde concen-tration (Camb).The monthly measured ventilation rates for 2013 are

illustrated in Figure 3. It varied randomly between 0.21−0.35

h−1. The relatively stable ventilation rate was ideal for directcomparison purposes. The main reason for the small variationwas that no heating or cooling equipment was operated in theexperimental room; thus, indoor−outdoor temperature differ-ences were small during all of the seasons. Moreover, the effectof outdoor wind pressure on the ventilation rate was minimizedby an enclosure around the room. No seasonal variations inventilation rate would be expected. The mean and standarddeviation (SD) of the ventilation rate throughout the year were0.29 h−1 and 0.046 h−1, respectively. Ventilation conditionsover the entire test period were assumed to follow the patternobserved in 2013.Thus, the loading ratio (L) was constant, and outdoor

formaldehyde concentrations (Camb) were much lower thanTable

1.Indo

orFo

rmaldehyde

Con

centration

s,Fo

rmaldehyde

Emission

sRates,Tem

perature,RelativeHum

idity(R

H),andAbsoluteHum

idity(A

H)in

DifferentSeason

s

concentration(m

g/m

3 )em

ission

rate

(mg/m

2 ·h)

temperature

(°C)

RH

(%)

AH

(g/kg a

ir)

timeperio

dseason

mean

min

max

mean

min

max

mean

min

max

mean

min

max

mean

min

max

9/10/2012−

30/11/2012

autumna

1.00

0.33

1.77

0.57

0.18

1.02

8.7

−0.3

19.3

63.5

55.3

71.2

4.6

2.3

8.2

1/12/2012−

28/2/2013

winter

0.19

0.09

0.38

0.10

0.04

0.23

−4.18

−10.9

4.3

66.3

58.7

73.6

1.9

1.1

3.4

1/3/2013−31/5/2013

sprin

g1.38

0.28

3.18

0.80

0.16

1.84

14.7

2.5

26.9

60.1

49.1

69.1

6.7

2.9

13.5

1/6/2013−31/8/2013

summer

3.59

1.32

4.78

2.07

0.76

2.76

27.0

19.1

3175.2

55.1

83.6

17.2

10.2

23.1

1/9/2013−30/11/2013

autumn

0.79

0.15

2.4

0.45

0.08

1.38

14.2

024.7

67.3

49.9

78.2

7.8

1.9

14.7

1/12/2013−

28/2/2014

winter

0.11

0.1

0.12

0.06

0.05

0.06

−1.3

−5.6

4.5

58.6

50.9

68.3

2.1

1.3

3.4

1/3/2014−31/5/2014

sprin

g1.07

0.17

2.21

0.61

0.09

1.28

17.1

3.2

3055.3

46.5

62.8

7.2

2.7

14.5

1/6/2014−31/8/2014

summer

2.52

2.17

3.19

1.46

1.25

1.84

28.1

22.7

31.4

64.3

51.4

72.9

15.6

10.8

21.1

1/9/2014−30/11/2014

autumn

0.66

0.1

1.33

0.38

0.05

0.77

14.0

2.6

26.1

66.7

55.8

76.1

7.4

2.8

15.1

1/12/2014−

28/2/2015

winter

0.07

0.05

0.08

0.03

0.02

0.04

−1.2

−4.0

4.8

54.8

49.4

59.9

1.9

1.5

2.8

aMeasurementdata

for2012

onlypartially

coveredtheautumnseason

(9October−30

Decem

ber).T

hese

data

werenotused

forcomparison

with

data

forautumn2013

and2014.

Figure 3. Ventilation rate results for the experimental room in 2013.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b02217Environ. Sci. Technol. XXXX, XXX, XXX−XXX

D

Dow

nloa

ded

by U

NIV

VE

RA

CR

UZ

AN

A o

n A

ugus

t 26,

201

5 | h

ttp://

pubs

.acs

.org

P

ublic

atio

n D

ate

(Web

): A

ugus

t 21,

201

5 | d

oi: 1

0.10

21/a

cs.e

st.5

b022

17

indoor formaldehyde concentrations (C(t)). Uncertainty or theRSD of the emission rate for each measurement may be largerthan that for the concentration considering the variations in theventilation rate (VR). However, the overall trends in emissionrates would not be affected by this factor. Thus, the meanemission rate for each measurement would vary in a similarmanner as did the indoor formaldehyde concentration, asillustrated in the right-hand y-axis of Figure 2. Means andranges for emission rates in the different seasons are listed inTable 1.Comparison of Emission Characteristics between Environ-

mental Chambers and the Experimental Room. Theemission rate from MDF decreased from the beginning ofthe measurement until 1 December 2012, consistent with thatobserved in an environmental chamber with constant environ-mental conditions. One might draw the conclusion that MDFemission in actual buildings would be similar to those in thechambers based on such data. However, this trend did notcontinue over time, and an entirely different emission patternoccurred in the experimental room compared to that in theenvironmental chamber, as discussed below.

(1) Material emission rates in an environmental chambershould decrease continuously.32 However, in theexperimental room, an apparent increase in the emissionrates occurred in spring and summer. Annual cyclicalseasonal variations were observed.

(2) In an environmental chamber, the highest emission ratesusually occur at the beginning when materials areintroduced.17 However, in the experimental room,maximum emissions occurred in summer instead of atthe beginning of the study. For example, the highestemission rates in the summer of 2013 and 2014 were2.76 and 1.84 mg/m2·h, respectively, much higher thanthe 0.93 mg/m2·h initial emission rate. These resultssuggest that environmental conditions should beconsidered in estimating formaldehyde emission ratesfrom MDF.

(3) When materials are tested in an environmental chamber,a “steady state” could be achieved with prolongedemission time.33 The emission rates change little afterthat.34 However, when the MDF was placed in theexperimental room, no such “steady state” could beassumed for the long time scale.

Due to apparent differences in emission characteristicsobserved between environmental chambers and the exper-imental room, emission models developed based on constantenvironmental conditions should not be directly applied toactual buildings. Adjustments to the model or modelparameters are required to represent the more complexsituation in actual buildings.Influencing Factors. The main differences between the

experimental room and environmental chambers were thetemperature and humidity conditions. The impacts of these twofactors are examined below.Influence of Temperature. The overall trends in form-

aldehyde concentration and temperature were consistent(Figure 4). Obvious seasonal variations were also observed inindoor temperature. Indoor formaldehyde concentrationsincreased in spring and decreased in autumn, as did thetemperature. The same patterns in concentration and temper-ature were observed in summer and winter. On the basis of thisqualitative analysis, temperature was likely one of the key

factors inducing seasonal variations in formaldehyde concen-trations. The Pearson correlation coefficient between theformaldehyde concentration and temperature was 0.84,indicating a strong positive correlation between these twoparameters.Temperature ranges for each season are listed in Table 1.

The annual variations were from −10.9 to 31 °C in 2013 and−5.6 to 31.4 °C in 2014, respectively. This broad range goesbeyond all possible temperature situations in actual buildingsand inclusive. In particular, it is interesting to compareemissions at high- and low-temperature extremes. Themagnitude of indoor temperature variations in summer werenearly the same as or smaller than those in winter (19.1 to 31°C vs −10.9 to 4.3 °C); however, the formaldehydeconcentration ranges were much larger in summer than inwinter (1.32 to 4.78 mg/m3 vs 0.09 to 0.38 mg/m3). Thus,formaldehyde emissions were more sensitive to temperaturechanges at higher temperatures.Temperatures and ranges over the next year did not change

substantially from the first year (Table 1, Figure 4). However,formaldehyde concentrations and ranges decreased significantly(0.09 to −4.78 mg/m3 vs 0.10 to 3.19 mg/m3). The samecorrelations between temperature and concentration in eachseason were observed. These results demonstrate the reducedeffect of temperature on formaldehyde emissions over the longterm, which was not revealed in previous studies usingenvironmental chambers.34−36

Influence of Humidity. Ranges for RH in each season arepresented in Table 1, and indoor RH and formaldehydeconcentration profiles are illustrated in Figure 5a. Indoor RHmainly varied between 55 and 70%. RH began to increase at theend of May 2013 and reached a maximum of 83.6% in themiddle of August 2013. This increase in RH was due to rainyweather during that period. Indoor RH decreased and returnedto almost the same level as before after 15 October 2013.Unlike formaldehyde concentration and temperature, RH wasrelatively stable without strong seasonal variations. The Pearsoncorrelation coefficient between the indoor formaldehydeconcentration and RH was 0.33, indicating that seasonalvariations in the formaldehyde concentration were not stronglyassociated with RH.The absolute humidity (AH) profile, on the other hand,

exhibited clear seasonal variations (Figure 5b). The maximumAH occurred in summer and the minimum in winter, with a

Figure 4. Indoor formaldehyde concentration and temperature profilesfor the experimental room.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b02217Environ. Sci. Technol. XXXX, XXX, XXX−XXX

E

Dow

nloa

ded

by U

NIV

VE

RA

CR

UZ

AN

A o

n A

ugus

t 26,

201

5 | h

ttp://

pubs

.acs

.org

P

ublic

atio

n D

ate

(Web

): A

ugus

t 21,

201

5 | d

oi: 1

0.10

21/a

cs.e

st.5

b022

17

range from 1.1 to 23.1 g/kgair. The Pearson correlationcoefficient between the formaldehyde concentration and AHwas 0.89, indicating a strongly positive correlation betweenthem.Previous studies on the influence of humidity on form-

aldehyde emissions were mainly conducted in environmentalchambers and focused on analysis of RH. The influence ofhumidity on formaldehyde emissions has also been neglected infield studies.21−23 However, Parthasarathy et al.35 found thatthe emission rate increased when RH increased from 50% to85%. Similar conclusions were obtained in other studies.36,37

Given the apparent discrepancies between the results of thisand previous studies, it is important to consider thefundamental mechanism of formaldehyde emissions fromMDF. A common explanation about the mechanism ofhumidity effect on formaldehyde emissions is that freeformaldehyde would increase due to urea−formaldehydedepolymerization and hydrolysis in the presence of freewater.36−38 Competition between pollutant and water mole-cules for free adsorption sites of the building material is anotherpossible mechanism.39−41 However, it is the molar amount offree water that determines the amount of hydrolysis. Theabsolute quantity of free water in the air (AH) is a betterapproximation of the molar amount of free water than the RH.Moreover, RH changes with temperature even when the molaramount of free water remains the same in the air. Thus, from

the standpoint of the hydrolysis reaction mechanism, AH is amore appropriate metric for evaluating the influence ofhumidity on formaldehyde emissions than RH.

Verification of the Effects of Temperature and AH onFormaldehyde Emissions through Small EnvironmentalChamber Tests. On the basis of the above results, the trendsin formaldehyde concentration, temperature, and AH werequite similar. To further verify temperature and humidityeffects, additional small-scale environmental chamber tests wereconducted by varying these two parameters individually.Small environmental chamber conditions and procedures are

detailed in Section S3 of Supporting Information. Time trendsfor formaldehyde concentrations in the chamber air are shownin Figure 6. The maximum concentration occurred in the first

few hours and then decreased continuously, unlike the annualseasonal variations observed in the experimental room. Bycomparing the results for Cases 1 and 2, it was found that at thesame temperature, formaldehyde emissions increased withincreasing AH. A similar trend was observed by comparing theresults for Cases 1 and 3, demonstrating that increasingtemperature also increases emissions.

Comparison with Other Studies. The measurementsettings and results of this and several other studies are brieflycompared in Table 2. The unique features and advantages ofthis study are summarized as follows.

(1) Human-related emissions, as well as disturbances by thetesting personnel, were eliminated in this unoccupiedexperimental room. Other formaldehyde sources andinteractions were not present, so that the measuredconcentrations were solely due to emissions by the MDFstudied. In other studies, the measurement results mayrepresent the cumulative contributions of several factors.

(2) The overall study period using the experimental roomwas 29 months, far longer than previous controlledlaboratory or field studies. Formaldehyde concentrations,as well as temperature, humidity, and ventilation ratedata, were all obtained over this long study period.Annual cyclical seasonal variations in formaldehydeconcentrations and emission rates were observed alongwith annual attenuation. Positive correlations betweenthe formaldehyde concentration and temperature and theformaldehyde concentration and AH were also clearlyidentified based on the detailed long-term data. Due to

Figure 5. Indoor formaldehyde concentration and humidity profilesfor the experimental room: (a) formaldehyde concentration andrelative humidity (RH) and (b) formaldehyde concentration andabsolute humidity (AH).

Figure 6. Formaldehyde results of the small-scale environmentalchamber tests.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b02217Environ. Sci. Technol. XXXX, XXX, XXX−XXX

F

Dow

nloa

ded

by U

NIV

VE

RA

CR

UZ

AN

A o

n A

ugus

t 26,

201

5 | h

ttp://

pubs

.acs

.org

P

ublic

atio

n D

ate

(Web

): A

ugus

t 21,

201

5 | d

oi: 1

0.10

21/a

cs.e

st.5

b022

17

the shorter durations and sometimes confounding data inother studies, these interesting variations and correlationswere not reported.

(3) Temperature and humidity varied naturally, and theranges for these variables were deliberately allowed to bemuch larger than typically occur in occupied buildings.Thus, the effects of temperature and humidity onformaldehyde emissions could be fully represented andanalyzed. In particular, the effects of low extremes intemperature and AH were analyzed in the field for thefirst time.

Thus, this study had several advantages, includingelimination of multiple interferences, long-term detailedmeasurement of data, and incorporation of broader temper-ature and humidity ranges. The results provide a meaningfulcontribution to characterization of long-term formaldehydeemissions and can be used to guide factor analysis as well asemissions model validation for actual buildings.

■ ASSOCIATED CONTENT

*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.est.5b02217.

Additional details on the accuracy of the MBTH methodand comparison results with HPLC (Page S2, Figure S1).Quality control and assurance of the MBTH method(Page S2−S3). Environmental settings and procedures ofthe small environmental chamber test (Page S3−S4,Table S1). (PDF)

■ AUTHOR INFORMATION

Corresponding Author*Telephone: (86)10-62788845. Fax: (86)10-62773461. E-mail:xyang@tsinghua.edu.cn.

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTS

This study is supported by the National Natural ScienceFoundation of China Project No. 51178237 and ShenzhenInstitute of Building Research.

■ REFERENCES(1) Salthammer, T.; Mentese, S.; Marutzky, R. Formaldehyde in theindoor environment. Chem. Rev. 2010, 110 (4), 2536−2572.(2) Main, D. M.; Hogan, T. J. Health effects of low-level exposure toformaldehyde. J. Occup. Environ. Med. 1983, 25 (12), 896−900.(3) Yang, X.; Zhang, Y.; Chen, D.; Chen, W.; Wang, R. Eye irritationcaused by formaldehyde as an indoor air pollution-a controlled humanexposure experiment. Biomed. Environ. Sci. 2001, 14 (3), 229−236.(4) Lang, I.; Bruckner, T.; Triebig, G. Formaldehyde andchemosensory irritation in humans: a controlled human exposurestudy. Regul. Toxicol. Pharmacol. 2008, 50 (1), 23−36.(5) Franklin, P.; Dingle, P.; Stick, S. Raised exhaled nitric oxide inhealthy children is associated with domestic formaldehyde levels. Am.J. Respir. Crit. Care Med. 2000, 161 (5), 1757−1759.(6) Rumchev, K. B.; Spickett, J. T.; Bulsara, M. K.; Phillips, M. R.;Stick, S. M. Domestic exposure to formaldehyde significantly increasesthe risk of asthma in young children. Eur. Respir. J. 2002, 20 (2), 403−408.(7) Formaldehyde, 2-Butoxyethanol and 1-tert-Butoxypropan-2-ol;IARC Monographs on the Evaluation of Carcinogenic Risk toHumans; World Health Organization, International Agency forResearch on Cancer, 2006; Vol. 88.(8) Development of WHO Guidelines for Indoor Air Quality WorldHealth Organization; WHO-ROE, Regional Office of Europe, 2006.(9) Indoor Air Quality Standard; GB/T18883-2002; Beijing:Administration of Quality Supervision, Inspection and Quarantine,2000 (in Chinese).(10) Tang, X.; Bai, Y.; Duong, A.; Smith, M. T.; Li, L.; Zhang, L.Formaldehyde in China: Production, consumption, exposure levels,and health effects. Environ. Int. 2009, 35 (8), 1210−1224.(11) Korjenic, A.; Teblick, H.; Bednar, T. Increasing the indoorhumidity levels in buildings with ventilation systems: simulation aideddesign in case of passive houses. Build. Simul. 2010, 3 (4), 295−310.(12) Soutullo, S.; Enríquez, R.; Jimenez, M. J.; Heras, M. R. Thermalcomfort evaluation in a mechanically ventilated office building locatedin a continental climate. Energy Build. 2014, 81, 424−429.(13) Saeki, K.; Obayashi, K.; Iwamoto, J.; Tone, N.; Okamoto, N.;Tomioka, K.; Kurumatani, N. The relationship between indoor,outdoor and ambient temperatures and morning BP surges from inter-seasonally repeated measurements. J. Hum. Hypertens. 2014, 28 (8),482−488.(14) Fabi, V.; Andersen, R. V.; Corgnati, S. P.; Olesen, B. W. Amethodology for modelling energy-related human behaviour:Application to window opening behaviour in residential buildings.Build. Simul. 2013, 6 (4), 415−427.(15) Bohm, M.; Salem, M. Z.; Srba, J. Formaldehyde emissionmonitoring from a variety of solid wood, plywood, blockboard and

Table 2. Summary of Measurement Conditions and Results of This Study with Those of Previous Studies

studies Wolkoff et al.21 Brown22 Crump et al.23 this study

house occupancy occupied unoccupied occupied unoccupied unoccupiedtesting person position indoor indoor indoor indoor outdoorbuilding materials number multiple multiple multiple multiple only MDFother VOCs existences yes (21)a yes (21)a yes (27)a yes (>5)a nonetime duration 12 months 12 months 8 months 24 months 29 monthstotal number of concentration tests 10 times 10 times 4 times 28 times 71 timesduplicated samples no duplicate no duplicate two duplicates no duplicate three duplicatestemperature interval not reported not reported not reported not reported 10 minRH interval not reported not reported not reported not reported 10 mintemperature range (°C) 14.9−23.9 17−33 21−29 20−30 −10.9−31.4RH range (%) 40−56 24−67 32−54 33−40 46.5−83.6AH range (g/kgair) not reported not reported 6.6−10.8 not reported 1.1−23.1ventilation measurement period March to May March to May not measured not reported a whole yearventilation measurement frequency not reported not reported not measured one time monthly

aNumbers in parentheses are the reported number of VOCs in the corresponding references.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b02217Environ. Sci. Technol. XXXX, XXX, XXX−XXX

G

Dow

nloa

ded

by U

NIV

VE

RA

CR

UZ

AN

A o

n A

ugus

t 26,

201

5 | h

ttp://

pubs

.acs

.org

P

ublic

atio

n D

ate

(Web

): A

ugus

t 21,

201

5 | d

oi: 1

0.10

21/a

cs.e

st.5

b022

17

flooring products manufactured for building and furnishing materials.J. Hazard. Mater. 2012, 221-222, 68−79.(16) Kim, K. W.; Kim, S.; Kim, H. J.; Park, J. C. Formaldehyde andTVOC emission behaviors according to finishing treatment withsurface materials using 20L chamber and FLEC. J. Hazard. Mater.2010, 177 (1), 90−94.(17) Brown, S. K. Chamber assessment of formaldehyde and VOCemissions from wood-based panels. Indoor Air 1999, 9 (3), 209−215.(18) Yang, X.; Chen, Q.; Zhang, J. S.; Magee, R.; Zeng, J.; Shaw, C. Y.Numerical simulation of VOC emissions from dry materials. Build.Environ. 2001, 36 (10), 1099−1107.(19) Huang, H.; Haghighat, F. Modelling of volatile organiccompounds emission from dry building materials. Build. Environ.2002, 37 (12), 1349−1360.(20) Xu, Y.; Zhang, Y. An improved mass transfer based model foranalyzing VOC emissions from building materials. Atmos. Environ.2003, 37 (18), 2497−2505.(21) Wolkoff, P.; Clausen, P. A.; Nielsen, P. A.; Molhave, L. TheDanish twin apartment study; part I: formaldehyde and long-termVOC measurements. Indoor Air 1991, 1 (4), 478−490.(22) Brown, S. K. Air toxics in a new Australian dwelling over an 8-month period. Indoor Built Environ. 2001, 10 (3−4), 160−166.(23) Crump, D. R.; Squire, R. W.; Yu, C. W. F. Sources andconcentrations of formaldehyde and other volatile organic compoundsin the indoor air of four newly built unoccupied test houses. IndoorBuilt Environ. 1997, 6 (1), 45−55.(24) Petitjean, M.; Hantal, G.; Chauvin, C.; Mirabel, P.; Le Calve, S.;Hoang, P.; Picaud, S.; Jedlovszky, P. Adsorption of benzaldehyde atthe surface of ice, studied by experimental method and computersimulation. Langmuir 2010, 26 (12), 9596−9606.(25) Salthammer, T.; Fuhrmann, F. Photocatalytic surface reactionson wall paint. Environ. Sci. Technol. 2007, 41 (18), 6573−6578.(26) ASTM D 6007-02. Standard Test Method for DeterminingFormaldehyde Concentration in Air from Wood Products Using a SmallScale Chamber; American Society for Testing and Materials: WestConshohocken, PA, 2008.(27) Yao, Y. Research on Some Key Problems of Furniture VOCEmission Labeling System. Ph.D. Thesis, Tsinghua University, 2011.(28) Nazaroff, W. W.; Weschler, C. J. Cleaning products and airfresheners: exposure to primary and secondary air pollutants. Atmos.Environ. 2004, 38 (18), 2841−2865.(29) Nørgaard, A. W.; Kofoed-Sørensen, V.; Mandin, C.; Ventura, G.;Mabilia, R.; Perreca, E.; Cattaneo, A.; Spinazze, A.; Mihucz, V. G.;Szigeti, T.; Kluizenaar, Y.; Cornelissen, H. J. M.; Trantallidi, M.;Carrer, P.; Sakellaris, I.; Bartzis, J.; Wolkoff, P. Ozone-initiated terpenereaction products in five European offices: replacement of a floorcleaning agent. Environ. Sci. Technol. 2014, 48 (22), 13331−13339.(30) Methods for Determination of Formaldehyde in Air of Public Places;GB/T18204.26-2000; Beijing: Administration of Quality Supervision,Inspection and Quarantine, 2000 (in Chinese).(31) Van Hooff, T.; Blocken, B. CFD evaluation of naturalventilation of indoor environments by the concentration decaymethod: CO2 gas dispersion from a semi-enclosed stadium. Build.Environ. 2013, 61, 1−17.(32) Cox, S. S.; Little, J. C.; Hodgson, A. T. Predicting the emissionrate of volatile organic compounds from vinyl flooring. Environ. Sci.Technol. 2002, 36 (4), 709−714.(33) Qian, K.; Zhang, Y.; Little, J. C.; Wang, X. Dimensionlesscorrelations to predict VOC emissions from dry building materials.Atmos. Environ. 2007, 41 (2), 352−359.(34) Xiong, J.; Wei, W.; Huang, S.; Zhang, Y. Association betweenthe emission rate and temperature for chemical pollutants in buildingmaterials: general correlation and understanding. Environ. Sci. Technol.2013, 47 (15), 8540−8547.(35) Parthasarathy, S.; Maddalena, R. L.; Russell, M. L.; Apte, M. G.Effect of temperature and humidity on formaldehyde emissions intemporary housing units. J. Air Waste Manage. Assoc. 2011, 61 (6),689−695.

(36) Frihart, C. R.; Wescott, J. M.; Chaffee, T. L.; Gonner, K. M.Formaldehyde emissions from urea-formaldehyde-and no-added-formaldehyde-bonded particleboard as influenced by temperatureand relative humidity. For. Prod. J. 2012, 62 (7), 551−558.(37) Sidheswaran, M.; Chen, W.; Chang, A.; Miller, R.; Cohn, S.;Sullivan, D.; Fisk, W. J.; Kumagai, K.; Destaillats, H. Formaldehydeemissions from ventilation filters under different relative humidityconditions. Environ. Sci. Technol. 2013, 47 (10), 5336−5343.(38) Myers, G. E. The effects of temperature and humidity onformaldehyde emission from UF-bonded boards-A literature critique.Forest Prod. J. 1985, 35 (9), 20−31.(39) Lin, C. C.; Yu, K. P.; Zhao, P.; Lee, G. W. M. Evaluation ofimpact factors on VOC emissions and concentrations from woodenflooring based on chamber tests. Build. Environ. 2009, 44, 525−533.(40) Huang, H.; Haghighat, F.; Blondeau, P. Volatile organiccompound (VOC) adsorption on material: influence of gas phaseconcentration, relative humidity and VOC type. Indoor Air 2006, 16(3), 236−247.(41) Suzuki, M.; Akitsu, H.; Miyamoto, K.; Tohmura, S.; Inoue, A.Effects of time, temperature, and humidity on acetaldehyde emissionfrom wood-based materials. J. Wood Sci. 2014, 60 (3), 207−214.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.5b02217Environ. Sci. Technol. XXXX, XXX, XXX−XXX

H

Dow

nloa

ded

by U

NIV

VE

RA

CR

UZ

AN

A o

n A

ugus

t 26,

201

5 | h

ttp://

pubs

.acs

.org

P

ublic

atio

n D

ate

(Web

): A

ugus

t 21,

201

5 | d

oi: 1

0.10

21/a

cs.e

st.5

b022

17