Following the short-term testing, the building managers identified several potential sources of indoor contaminants including composite wood paneling, painted gypsum board walls and high density plastic barrels used as hydroponic containers for the bio-filtration-based air cleaning technology. Both new and aged plywood wall paneling was present in the building and the paneling was coated with a clear polish. The goal of this study was to measure material specific emission factors for VOCs and carbonyls and characterize the potential influence of the polish coating on the wood panel material. All building materials that were tested for emissions were harvested from the PBC building, double wrapped in foil and shipped directly to Lawrence Berkeley National Laboratory for testing. A description of each of the material samples is provided in Table 1. All samples except for the hydroponic drum were cut to 0.023 m 2 . In the laboratory, the sides and backs of the material were sealed with aluminum tape and stainless steel backing plates, respectively, to leave only the front face of the material exposed for testing. Each sample was placed individually in 6-liter stainless steel conditioning chamber as illustrated in Figure 1. The conditioning chambers were closed with Teflon lined lids and held at approximately 22 ˚C and 50% relative humidity to precondition the materials prior to sampling. For new materials, samples are typically preconditioned to allow the emissions to drop to a more relevant value for estimating long-term emission rates. For materials that are allowed to age in the environment, the conditioning period is important to allow chemicals that have partitioned into the material from the environment, e.g.,drainage planter pot chemicals that are not indigenous to the material, to off gas so that the measured emission rates are relevant to the material being tested.
The emission testing generally followed California Specification 01350 and ASTM Standard Guide D-6007-02 using small emission chambers. The approach has been used for a wide range of materials measuring both VOCs and carbonyls as described previously and as summarized below. Four emission chambers installed in a controlled environment oven provide an isolated environment with constant temperature and humidity. The constant humidity was maintained by splitting the flow of dry carbon/HEPA filter air with a portion of the air bubbling through a water bath then re-mixed to achieve the desired humidity for air flowing through each chamber. The chambers, shown in Figure 2, are made of stainless steel and all interior surfaces are coated with Sulfinert® coating to minimize chemical interaction with chamber walls. The chambers are 10.75 L and are operated with an approximate ventilation rate of 1 liter per minute equivalent to 5.6 air changes per hour , or 2.6 m3 [air]/m2 [exposed surface area]/hour. The standard tests are operated at 25 °C and 50% RH. After being pre-conditioned, each test material was transferred to the test chamber and placed on Sulfinert® treated screens resting slightly below the midpoint of the chamber. Each material was allowed to equilibrate in the test chamber for at least 30 minutes after being transferred from the conditioning chamber before testing. Once equilibrated, the air samples were collected directly from the test chamber and analyzed for VOC and Aldehydes as described below. The materials were first tested after 24 hours of conditioning, and then again after at least seven days of conditioning. The first sampling period was used to get information on upper bound emission rates and allow for the identification of the mix of chemicals in the emission stream.
The second sampling period, after seven days conditioning, provides the emission factors that are more relevant to the long-term emission pattern. Additional measurements were collected for the new wood with new polish to further understand how the polish affects the emissions from the material. VOC samples were collected and analyzed following the U.S. Environmental Protection Agency Method TO-17. VOC air samples were collected directly from the chambers by drawing chamber air through multi-sorbent tubes with a primary bed of TenaxTA® sorbent backed with a section of Carbosieve®. A peristaltic pump was used to pull the air through the sample tubes at a rate of approximately 100 mL/min for 1 hour. The flow was measured using a DryCal gas flow meter and was recorded at the beginning and the end of the sampling period. Before subjected to chemical analysis, each sample was spiked with 120ng of gas-phase 1-Bromo-3 Fluoro-Benzene , which was used as the internal standard in the quantification method. Analytes were thermally desorbed from the sampling tubes using a thermodesorption auto-sampler , a thermo-desorption oven , and a cooled injection system . Desorption was performed in splitless mode where the desorbed analytes were refocused on the cooled injection system prior to injection. Desorption temperature for the TDS started at 30 ˚C with a 0.5 minute delay followed by a 60 ˚C ramp to 250 ˚C and a 4 minute hold time. The cooled injection system was fitted with a Tenaxpacked glass liner that was held at -10 ˚C throughout desorption and then heated within 0.2 minutes to 270 ˚C followed by a 3-minute hold time. Compounds were resolved on a GC equipped with a 30 meter HP- 1701 14% Cyanopropyl Phenyl Methyl capillary column with helium flow of 1.2 mL/min.
The initial temperature of the oven was -10 ˚C held for 0.5 minutes then ramped at 5 ˚C/min to 40 ˚C then 3 ˚C/min to 140 ˚C and finally at 10 ˚C/min to 250 ˚C and held for 10 minutes. The resolved analytes were quantified using electron impact mass spectrometry, , with mass to charge ratio limits of 44.0 m/z and 450.0 m/z. The MS was operated in full scan mode with a solvent delay of 3.00 minutes. Compounds were initially identified using NIST mass spectral search program for the NIST/EPA/NIH mass spectral library with identity confirmed and quantified using pure standards. When pure standards were not available, the analyte was reported in terms of toluene equivalence by comparing the instrument response for the total ion chromatogram of the chemical to a multi-point calibration of TIC response for toluene.The volatile carbonyls including formaldehyde, acetaldehyde and acetone are quantified using USEPA Method TO-15. As with the VOC samples, the air was drawn directly from the chambers during sampling. The sampling rate was maintained at less than 80% of the total air flow through the chamber to prevent back flow of unfiltered air into the chamber during testing. The sample air passed through silica gel cartridges coated with 2,4-dinitrophenyl-hydrazine, which quantitatively reacts with the carbonyl functional group effectively trapping the aldehydes and other low molecular weight carbonyl compounds. A peristaltic pump was used to pull the air through the cartridge at a rate of approximately at 800 mL/min for 1 hour. The flow was measured using a DryCal gas flow meter and recorded at the beginning and the end of the sampling period. Prior to analysis, sample cartridges were eluted with 2ml of high purity acetonitrile and the effluent was brought to a final volume of 2 ml.The HPLC was fitted with a C18 reverse phase column and run with 65:35 H2O: Acetonitrile mobile phase at 0.35 mL/minute and UV detection at 360 nm. Multi-point calibrations were prepared for the target analytes using commercially available hydrazone derivatives of formaldehyde,plant pot with drainage acetaldehyde and acetone.All materials were initially tested after only 24-hours of conditioning time. Prior to conditioning, the materials had been tightly wrapped in foil and packaged individually in resealable plastic bags during shipping so the initial emissions were expected to be elevated. The purpose of this in initial testing was to identify the chemicals in the emission stream. A total of forty chemicals were identified in the emission stream from the six materials tested. All chemicals are listed in Table 2 along with steady state concentrations measured after 24-hours of conditioning. All values are listed in Table 2 for comparision but values above the typical method limit of quantification of 0.5 µg/m3 are listed in bold text. The plastic material from the hydroponic drum was tested as received without sealing the back and sides so the exposed area was approximately double that of the other materials. The initial measurements found that except for hexadecane and tetradecane, most of the VOCs from the plastic material, including the aldehdyes, were near or below the minimum detection limit. Therfore, the plastic is not likely a source of indoor contaminants in the indoor environment of the PBC.
The drywall material produced a number of elevated VOCs with two that exceeded the linear range of the analytical method. Drywall is typically a low VOC material although fresh coatings such as paint or plaster can emitt VOCs during curing. The drywall samples tested in this study appear to be freshly painted because the edges were sealed with paint. This might explain the elevated propylene glycol and benzyl alcohol. The new wood paneling had very high levels of formaldehyde both with and without polish although the polished panel had the high test levels of formaldehyde overall. However, the unpolished new wood panel produced a wider variety and higher levels of VOC in the emissions. The old wood paneling produced much lower levels of formaldehyde and the levels of VOCs in general were similar both with new and old polish. After the initial tests were completed to identify the target chemicals in the emission stream, the materials were returned to the conditioning chambers for approximately six more days before measuring the emission factors. Concentrations for the plastic material from the hydroponic drum remained low in the second test so emission factors are not reported for the plastic hydroponic drum material. The standard emission factors determined for the wood paneling and drywall materials are reported in Table 3. The formaldehyde emissions from new wood with new polish were still significantly elevated after 7 days but we suspected that the combination of polish and storage may have increased the time needed for the emission factor to drop to a relativily constant level. To address this, we continued to condition the new wood with new polish for an additional week and re-tested. The additional time needed to condition the new wood with new polish may have been due to a higher capacity of the polish coating for accumulating formaldehyde during storage. This possibility was tested and is discussed further below. The standard emission factors for the materials from the PBC are summarized in Table 3. Several of the chemicals that were initially detected in the materials were no longer detectable in the emission stream after a week of conditioning and are therfore not listed in Table 3. The painted drywall continued to have extremely high levels of benzyl alcohol and propylene glycol as well as quantifiable levels of several other aldehydes , alcohols and esters that may be related to the coating material and/or sorbed into the drywall matrix from the environment. The wood paneling material presented a mix of VOCs depending on if the polish and/or wood were new or old as illustrated in Figure 3. Figure 3 lists the sum of all emission factors for VOC presented as stacked colums with the largest overall emission factors listed in decreasing order from bottom to top on the figure legend. Emission factors listed in Table 3 that are below the approximate limit of quantification of 1.65 µg/m2 /h are not included in Figure 3. Overall the drywall material had the highest sum of individual emission factors with the paneling material emitting 134, 129, 33 and 7 for the new wood no applied polish, old wood new polish, new wood new polish and old wood old polish, respectivily. Formaldehyde emissions for the old wood paneling with new and old polish, and the drywall were all similar ranging from 10 µg/m2 /h to 22 µg/m2 /h . For the new wood, the formaldehyde emissions were approximately an order of magnitude higher than the other materials for both the polished and unfinished surfaces. The emission results for formaldehyde are illustrated in Figure 4 showing that the polish coating does not seem to significantly change the measured emission factors when the age of the wood paneling is taken into consideration.