RR-0803: Volatile Organic Compound Concentrations and Emission Rates in New Manufactured and Site-Built Houses

Introduction

Indoor sources of volatile organic compounds (VOCs) are a determinant of indoor air quality in houses. Many materials used to construct and finish the interiors of new houses emit VOCs including formaldehyde. These VOC emissions can result in substantial contamination of indoor air. It is recognized that such contamination is a probable cause of acute health effects and discomfort among occupants (Andersson et al., 1997).

Data on the impacts of VOC sources in houses currently being built in North America are needed, as historical data may be outmoded. Formaldehyde concentrations in manufactured houses serve as an example. Concentrations in excess of 0.1 ppm were frequently encountered in manufactured houses built a decade or more ago (National Research Council, 1981; Sexton et al., 1986 and 1989). Today, less plywood paneling is used to finish the interiors of these houses. Also, the emissions of formaldehyde from wood products are generally lower (Kelly et al., 1999). Consequently, formaldehyde concentrations in contemporary manufactured houses may be lower than historical values. There have been other changes in the products and materials used in the construction of houses of all types, which may impact VOC concentrations. Examples include the increased use of engineered wood products and adhesives and the reformulation of paints.

Ventilation is another determinant of indoor air quality in houses. Ventilation serves as the primary mechanism for removal of gaseous contaminants generated indoors. Thus, higher contaminant concentrations are expected at lower ventilation rates given constant emission rates. The trend in new construction, supported by governmental programs, is to make house envelopes tighter. This practice improves energy efficiency by decreasing the infiltration of unconditioned outdoor air. Consequently, ventilation rates in new houses without supplemental ventilation can be relatively low with a related potential for degraded indoor air quality.

This study was conducted with the primary objective of characterizing and comparing the airborne concentrations and the emission rates of total VOCs and selected individual VOCs, including formaldehyde, among a limited number of new manufactured and site-built houses.

Secondarily, the study attempted to identify the sources of specific VOCs and evaluate the effectiveness of several source substitution treatments for reducing concentrations of VOCs.

Methods

Descriptions of Study Houses

Four manufactured and seven site-built houses were included in the study. They were located in the eastern and southeastern United States in hot-humid and mixed-humid climates.

The manufactured houses were produced by a single facility in Florida. In their construction and level of finish, they were generally representative of better-quality two-section houses sold in that state. They were manufactured in July 1997 and, within three weeks, were installed at a nearby sales center. Air sampling was initially conducted on 9/16/97. Subsequent air samples were collected on 11/19/97 and 5/1/98.

The manufactured houses differed with respect to size and finish materials (Table 1). Construction techniques and basic materials were identical. The floor assembly was plywood attached to wood joists that were mounted on a steel frame. The exterior and interior walls were framed with wood studs. There were small vented attic spaces. The interior walls and ceilings were covered with gypsum board panels. Many of the wall panels were pre finished with vinyl wallpaper. Floors in the main living areas were carpeted. In House M1, finish floors for the kitchen, baths and utility area were ceramic tile. In the other houses, sheet vinyl was used for these areas. Each house was equipped with a heat pump heating, ventilating and air-conditioning (HVAC) system. House M2 had a ducted fresh air inlet and utilized a fan re-cycling strategy to provide supplemental ventilation (Rudd, 1996). The control device periodically turned on the HVAC system fan to entrain outdoor air if the thermostat did not call for heating and cooling over an extended period (set to 20 minutes off followed by 10 minutes on). The other houses had a manually switched exhaust fan mounted in the living area ceiling that was designed for intermittent duty.

Several VOC source substitution treatments were evaluated. Latex paints with VOC contents of 0-1 g L^-1 were applied in House M2. A carpet system, consisting of a good quality nylon pile carpet and a synthetic fiber carpet cushion, was installed in Houses M2 and M4. These two materials have been shown to have low emissions of VOCs (Schaeffer et al., 1996; Hodgson, 1999). The other houses had conventional latex paint and a lower grade carpet installed over a bonded urethane carpet cushion. All of the houses were decorated and furnished. Decorations variously consisted of painted wainscoting, wallpaper, and built-in cabinetry. The houses were unoccupied, but used as sales models. The HVAC systems were operated on a consistent basis.

The site-built houses were small, detached single-family dwellings designed as entry-level housing (Table 2). They were located in three geographically separate projects. Air samples were collected in House A on 12/10/97 within one month of its completion. There were three demonstration houses in Project B that differed primarily with respect to exterior wall construction. Air samples were collected on 4/7/98, about two months after completion. Three more demonstration houses in Project C also differed primarily with respect to exterior wall construction. Air samples were collected on 11/4/98, about one month after completion. Additional measurements were made the following month in House C3 at typical and reduced ventilation rates.

The site-built houses were all built on cement slab floors. House A was the only two-story structure. The exterior walls were made of various materials. Houses A, B1 and C1 were conventionally framed with wood studs. Houses B2 and B3 had two different types of insulating concrete form walls, consisting of expanded polystyrene sections filled with a concrete core. The exterior walls of House C2 were formed with standard concrete blocks. House C3 was constructed with a lightweight insulating block system. All of the houses had vented attic spaces. The inner walls in all houses were framed with wood studs. The interior wall and ceiling surfaces were painted gypsum board. Finish floors consisted of carpeting and sheet vinyl flooring. A heat pump HVAC system was installed in each house. Houses A and C3 had supplemental ventilation systems; a balanced heat-recovery ventilation (HRV) system was installed in House A, and a ducted fresh air inlet and a fan re-cycling control were installed in House C3. At the time of sampling, all houses were completely finished, including cabinetry, but they were not furnished or occupied.

Air Sampling and Analysis

Indoor sampling was conducted during periods in which house ventilation rates were relatively constant and at near steady-state conditions. Air samples were collected in the main living area of the houses, typically a combined living/dining room. Sampler inlets were ~1.5 m above the floor. Air samples were also collected at outdoor locations adjacent to the houses. Sample flow rates were regulated with electronic mass-flow controllers.

VOCs were collected on sorbent samplers containing Tenax-TA (P/N 16251, Chrompack, The Netherlands). Sample flow rates were ~0.1 L min^-1. Typical sample volumes were ~1 L. Samples for formaldehyde and acetaldehyde were concurrently collected on silica cartridges impregnated with 2,4-dinitrophenylhydrazine (XPoSure Aldehyde Sampler, Waters Corp.). Sample flow rates were ~1 L min-1. Sample volumes were ~30 L.

VOCs were quantitatively analyzed by gas chromatography/mass spectrometry (GC/MS) following U.S. Environmental Protection Agency (EPA) Method TO-1 (Winberry et al., 1990). Samples were thermally desorbed with a cryogenic inletting system (Model CP-4020 TCT, Chrompack, The Netherlands).

Fifty-two target VOCs were selected for analysis. These were representative of the major chemical classes of compounds that occur in indoor air, indicative of specific indoor sources, or of interest because they have low odor thresholds or are potent sensory irritants. Many are on the list of 60 target VOCs recommended to be included in an analysis of TVOC (ECA-IAQ, 1997). Multi-point, internal standard calibrations were prepared using pure compounds.

The relative precision of the sampling and analytical method (as measured by a coefficient of variation) has been determined to be about ±10% for many compounds and about ±35% for ethylene and propylene glycols (Hodgson, 1999). The uncertainty in the measurement of acetic acid was greater than ±35%.

The GC/MS method for TVOC has been described (Wallace et al., 1991; Hodgson, 1995). The total-ion current (TIC) chromatogram for a sample was integrated over a retention-time range bounded by n-heptane and n-heptadecane using parameters that captured most of the chromatographic response. The mass of the compounds represented by the sum of the TIC area was calculated relative to the amount and area response of the internal standard. This method was calibrated with a mixture of ten common alkane and aromatic hydrocarbons. The uncertainty in the method when applied across a range of buildings with different VOC sources is estimated to be about ±40% (Wallace et al., 1991). Aldehyde samplers were analyzed for formaldehyde and acetaldehyde by high-performance liquid chromatography following U.S. EPA Method TO-11 (Winberry et al., 1990).

Measurement of Ventilation Rates

Ventilation rates were measured concurrently with the collection of air samples. HVAC systems were operated during sampling and for at least the prior day to maintain interior temperatures. Supplemental ventilation was used in the three houses with this system. Manually switched exhaust fans were off. Exterior doors and windows were closed. A tracer gas, sulfur hexafluoride (SF6), was introduced throughout a house by syringe. SF6 concentrations were measured with a photoacoustic infrared analyzer (Model 1302, Brüel & Kjær Instruments, Denmark). After sufficient time for mixing, the ventilation rate in air changes per hour (h^-1) was calculated as the slope of the least squares linear regression of the natural log concentration of SF6 versus time.

Measurement of VOC Emission Rates From Materials

A specimen of plywood flooring (softwood, phenol-formaldehyde resin) was collected from the manufactured house assembly line. Emission rates of VOCs and formaldehyde were determined following the American Society for Testing and Materials Standard Guide D-5116-97 (ASTM, 1997). The specimen was held in a 10.5-L stainless-steel chamber at 23 ± 1°C and 50 ± 10% relative humidity. The exposed surface area was 0.0074 m². Nitrogen was introduced at 1.0 ± 0.05 L min^-1. Gas samples were periodically collected from the exhaust over 72 hours.

Data Analysis

Emission rates of the target compounds were calculated for the houses and the chamber study assuming that the houses and the chamber were ideal continuously-stirred tank reactors (CSTRs) operating at near steady-state conditions. Losses of compounds due to factors other than ventilation (i.e., sink effects) were ignored; consequently, the calculated values were net rates. The steady-state form of the mass-balance model for CSTRs was used (ASTM, 1997). Quasisteady state, area-specific emission rates (ER) in μg m^-2 h^-1 were calculated as:

Where V is the ventilated volume (m³); a is the ventilation or air change rate (h^-1); C is the air concentration of the compound in the house or chamber (μg m^-3); and C0 is the outdoor air concentration or the chamber inlet air concentration of the compound (μg m^-3). For the houses, A is the floor area (m²), and for the chamber study, A is the exposed surface area of the material (m²). Normalization by floor area facilitates comparisons among houses of different sizes.

Results

Ventilation Rates

The ventilation rates measured at the time of air sampling are shown in Figure 1. For the manufactured houses, the American Society of Heating, Refrigeration and Air-conditioning Engineers (ASHRAE) minimum recommended "by volume" rate of 0.35 h^-1 is applicable (ASHRAE, 1989). The measured rates in the four houses were all in excess of this value. The ventilation rates in Houses A (with an HRV system), C1 and C2 were measurably below the ASHRAE minimum recommended "by occupant" rates (occupancy = 1 person per bedroom + 1) shown in Table 2.

Concentrations of TVOC and Target Compounds

Table 3 compares the concentrations of 28 target compounds and TVOC in the manufactured and site-built houses. The target compounds with the highest concentrations were selected for presentation. The GM concentrations for the manufactured houses were calculated using all three sets of data obtained over the study period. The GM concentrations for the site-built houses are for single data sets. In this and subsequent tables, compounds are first ordered by chemical class and then by decreasing volatility within each class. For compounds with concentrations below the lower limit of quantitation, a concentration equal to one-half of the limit was used in the calculation.

The GM concentrations of TVOC in the manufactured and site-built houses were 1.5 and 2.7 mg m^-3, respectively. The GSDs for the two house types were similar. The individual house data are shown in Figure 2. Houses C1 and C2 had the highest TVOC concentrations of 4.9 and 3.7 mg m^-3, respectively. Outdoor values were typically less than 0.15 mg m^-3 (GM outdoor concentration = 0.09 mg m^-3 for all data sets).

There was substantial similarity in the concentrations of individual VOCs measured in the two house types. In the manufactured houses, the predominant compounds based on their GM concentrations were a-pinene, ethylene glycol, formaldehyde, acetaldehyde, hexanal, and acetic acid. These compounds were also predominant in the site-built houses. The ranges in their values and in their distributions, as indicated by the GSDs, were similar between the two data sets. Additional compounds with relatively high GM concentrations in the site-built houses were toluene, b-pinene, and 1-butanol.

Because of the observed similarities in concentrations between the two house types, the data for the individual compounds were combined into a single set and summary statistics were calculated. The 52 target VOCs plus formaldehyde and acetaldehyde are categorized by their GM concentrations for the 11 houses in Table 4. Eleven VOCs had GM concentrations of

Formaldehyde was one of the most abundant compounds (Figure 3). The GM concentration for the entire data set was 40 ppb, with a GSD of 1.6. With the exceptions of Houses C1 and C2, the concentrations were all less than 50 ppb. Formaldehyde concentrations within each manufactured house were relatively constant over time. Outdoor values were 6 ppb, or less (GM outdoor concentration = 2 ppb for all data sets). . .

Download complete report here.