IA-Quest: Indoor air quality emission simulation tool - frequently asked questions

Answer to question 1:

The software calculates the concentrations of contaminants that would occur in a ventilated indoor space due to emissions from materials contained within that space. The calculation of concentrations assumes a simple single-zone mixing model. The emission characteristics of materials are obtained from the database packaged within the program. These emission characteristics are derived from laboratory tests of materials.

In short, the program allows users to:

• conduct a simulation to assess the VOC emission impact of selected materials based on the amount of materials used and the ventilation rates
• browse a database of measured emission characteristics for various building materials
• query and search the database
• generate reports on emission characteristics and simulation results
• browse properties and health effects, if any, of the modeled chemicals

Answer to question 2:

Different types of users are expected to benefit from the package, including manufacturers of building materials interesting in evaluation/development of products; building designers, renovators and managers interested in creating low-VOC environments for their occupants; researchers.

Answer to question 3:

A user needs to have the following information to simulate indoor air concentrations of chemicals from single or multiple building materials:

• volume of a space
• ventilation schedule (can be imported if it is complex)
• area, in-time and out-time of materials
• simulation duration
• output time interval, i.e., time step for calculation

Building materials can be selected from the database. Chemicals emitted from the selected building materials will be automatically loaded to the calculation section such that a user can select all or some of the chemicals for simulation.

Answer to question 4:

With the cursor positioned in the graphing region, click and drag a box (not a line) from lower right to upper left and release. The axes should be restored to the original display.

Answer to question 5:

The file "RoomSimDB.mdb" created in the subdirectory "Database" contains the information of the simulation cases that you have created (room conditions, material selections and case descriptions). You should periodically create backups of this file (optionally with a new filename which includes the date on which it is saved).

Answer to question 6:

At this point, there is no mechanism for a user to add his or her data to the software. The research team at the NRC is aware of the need to combine data generated from different labs. Prior to the efforts, producers for, and documentation of, specimen collection, handling, conditioning, and chamber characterization/operation should be in place. If you're interested in the inter-laboratory collaboration, please contact Doyun Won (doyun.won@nrc-cnrc.gc.ca or 613-993-9538).

Answer to question 7:

The software calculates indoor air concentrations using a mass balance approach as shown in the following equation.

where:

• C: Air concentration in a space ( mg/m³ )
• V: Volume of a space ( m³ )
• Q: Airflow (ventilation) rate ( m³/h )
• C_in: Air concentration in ventilation air ( mg/m³ )
• A_i: Surface area of the i^th source ( m² )
• EF_i: Emission factor of the i^th source ( mg/m²/h )
• n: number of sources

The emission factor is defined in the database by the coefficients of a pre-selected source model. The Runge-Kutta method is used to solve the governing equation.

Answer to question 8:

The material emission database currently consists of about 2,300 data sets of emission characteristics based on emission testing of 69 building materials commonly used in Canada. The emission testing was done by the NRC in accordance with ASTM D 5116. The emission characteristics are specified as coefficients of source models.

The database has the following information:

• coefficients of source models for "abundant" volatile organic compounds (VOC), "Target" VOCs and Total VOC (TVOC) from 69 building materials
• R² value as a result of curve-fitting to get source models
• measured emission factors (maximum and nominal at 24 hour)
  Test record (test date, temperature, air change rate, RH, air velocity, turbulence level and/or loading ratio)
• specimen record (generic description of a test specimen, specimen dimension - width, length, & thickness, and/or weight)
• chemical information (physical properties, health effect data, and/or synonyms of a chemical name)

Answer to question 9:

The material emission database contains emission characteristics of volatile organic compounds coming from building materials. The following information is included for VOCs:

Properties: molecular weight, vapor pressure and boiling point

Health effect data:

• odour detection threshold ¹
• occupational exposure limit (USA, Denmark) ¹
• mucous membrane irritation threshold)¹
• non-cancer chronic reference exposure level by California ²
• permissible exposure limit by OSHA ³

The references of the information include:

¹ Database with properties of 808 VOCs, B.Jensen, P.Wolkoff, Nat. Inst. Occup. Health, Denmark 1996.
² Non-caner chronic reference exposure level, Office of Env. Health Hazard Assessment, California EPA, available at http://oehha.ca.gov/air/allrels.html.
³ Permissible exposure level set by the Occupational Safety and Health Administration (OSHA), US Department of Labor to protect workers against the health effects of exposure to hazardous substances. PELs are regulatory limits on the amount or conc. of a substance in the air and are based on an 8-hour time weighted average (TWA) exposure. ( C ) denotes a ceiling limit. Available at the following; TABLE Z-1 / /TABLE Z-2

Answer to question 10:

GC/MS chromatographs from an emission test were analyzed to identify and quantify "Target" VOCs (see Question 11), "abundant" VOCs (see Question 12) and TVOC (see Question 13). In addition, HPLC analysis of many of the tests was performed to determine the emission rates of carbonyl compounds (including formaldehyde) from many of the materials tested.

DNPH cartridge samples were collected at 24 hour intervals for the limited test specimens reported in the database (CRP7, CRP7a, CT1, CT2, HWF1, LAM1, LAM2, LAM3, LIN1, LIN2, MDF2, VWB1, OSB4, OSB5a, OSB5b, OSB5c, OSB5d, OSB6a, OSB6b, OSB7a and OSB7b). For many tests, DNPH sampling was not performed at sufficient frequency to enable accurate calculation of emission coefficients. With the exception of the tests noted above, the absence of formaldehyde/carbonyl compound emission data does not necessarily mean that these compounds were not emitted from the materials.

Answer to question 11:

To aid the chemical identification and quantification, a target VOC list was assembled. The 90 "Target" VOCs were selected from 11 published lists by national and international agencies such as the World Health Organization and Health Canada coupled with experience gained from emission tests at the NRC. The target VOC list for material emissions was intended to include chemicals that were: 1) known or suspected to have health or irritation concerns (health criterion); 2) known to be emitted from the building materials (building material); 3) often found in indoor air (indoor air), and 4) suitable for sorbent sampling and analysis with GC/MS or carbonyl analysis with HPLC (analysis). Presence of "Target" VOCs were checked in all samples and they were quantified against their own standards (pure compounds).

Answer to question 12:

In addition to "Target" VOCs, GC/MS chromatographs were analyzed for additional "Abundant" VOCs outside of the target list. "Abundant" VOCs were defined as compounds whose level is more than 1% of TVOC at 24. Toluene was used as a calibration standard for quantification of "abundant" VOCs.

Answer to question 13:

TVOC (Total Volatile Organic Compound) was calculated from the GC/MS chromatogram (total ion count) by adding the areas of all peaks of C6 to C16 hydrocarbons with the retention time between 5 and 40 min. It should be noted that the TVOC value can be lower than the sum of all individual VOCs due to the difference in the calibration method. In this study, toluene was used as a calibration standard for quantification of TVOC, while their own authentic standards (pure compounds) were used for individual VOCs on the "Target" VOC list. Also, it is important to note that carbonyl compounds that are not identified with GC/MS, but quantified by HPLC, are not included in the TVOC value. For example, formaldehyde is not included in the TVOC value.

Answer to question 14:

The "nominal" emission factor is defined as the emission factor of a chemical at 24 hours after a specimen is placed in the dynamic chamber. The value was calculated using the measured chamber concentration at 24 hour, air change rate (N, 1/h) and material loading ratio (L, m²/m³ ):

The "nominal factor" can used to compare contaminant emissions within the same type of products/materials.

Answer to question 15:

MasterFormat™ is a system of numbers and titles for organizing construction information into a standard order or sequence. MasterFormat™ is produced jointly by the Construction Specifications Institute (CSI) and Construction Specifications Canada (CSC). Recently MasterFormat™ expanded from 16 Divisions to more than 40 Divisions. Within each division are sections.

Emission testing was conducted by the NRC in accordance with ASTM D 5116-90. The testing period ranged from 72 to 362 hours for dry materials and 78 to 440 hours for wet materials to capture the concentration decay portion properly. Therefore, the number of chamber air samples also varied from 6 to 22 for dry materials and from 16 to 40 for wet materials. Exceptions are associated with 2 OSB specimens whose tests lasted for 1 year to investigate long-term emission behaviors. For long-term tests, a total of 40 samples were taken for each test.

The main chemical analysis method involves air sampling on multi-layered sorbent tubes and the thermal desorption/GC/MS (gas chromatography and mass spectrometry) analysis. The HPLC (high performance liquid chromatography) analysis in combination with DNPH (dinitrophenylhydrazine) cartridge sampling was added for carbonyls such as acetaldehyde, butanal, formaldehyde, hexanal, pentanal and acetone. However, DNPH cartridge samples were collected at 24 hour intervals for the limited test specimens reported in the database (CRP7, CRP7a, CT1, CT2, HWF1, LAM1, LAM2, LAM3, LIN1, LIN2, MDF2, VWB1, OSB4, OSB5a, OSB5b, OSB5c, OSB5d, OSB6a, OSB6b, OSB7a and OSB7b). For many tests, DNPH sampling was not performed at sufficient frequency to enable accurate calculation of emission coefficients. With the exception of the tests noted above, the absence of formaldehyde/carbonyl compound emission data does not necessarily mean that these compounds were not emitted from the materials.

Answer to question 17:

With the advice of the Consortium's Technical Advisory Committee, and approval from the Steering Committee, materials were selected to represent the broad cross section of commonly used materials that go into building construction and furnishing, and that through exposed surfaces or contact with air that reaches the indoor environment of the building, are expected to have impact on indoor levels of organic chemicals. Materials typically used in both residential and commercial construction were selected. Where possible, specimens were obtained directly from manufacturers. When this was not possible, materials were obtained from large retail outlets. The "Test Notes" field in the "Emission Test Record Details" window specifies the source of each specimen. In all cases, great care was taken to preserve specimen integrity and representativeness.

Answer to question 18:

Specimen collection and handling procedures depended on the type of material collected. For "wet" materials (e.g. paints, stains, caulks, adhesives, etc.) materials were simply obtained in their packaged form and stored at 23 oC for at least 48 hours prior to testing. In some cases (e.g. paints), a single specimen (can) was purchased and immediately mixed and transferred to several clean containers to obtain multiple specimens for repetitive testing. Each individual specimen was thus treated equally, and the loss of highly volatile components through repetitive specimen opening/handling was avoided. For "dry" materials (e.g. plywood, carpeting, acoustical panels, flooring materials, etc.), specimen collection was conducted by one of two methods. For specimens collected directly from manufacturing sites, detailed instructions for specimen selection and handling were provided to the collection personnel together with clean specimen bags shipped in packing cases. On receipt of the specimens at the lab, they were immediately cut to their final test dimensions, repackaged in clean specimen bags, flushed 3 times with ultra-zero compressed air, resealed, and stored for a minimum 48 hours at 23 oC prior to testing. A clean, air filled bag was also shipped with the case and returned with the specimens as a check for possible contamination during transport. For dry materials collected from retail suppliers, the process was similar, except in the case of panel materials (OSB, MDF) where 3 panels from the centre of an intact stack were removed and clamped together. They were returned to the lab in this fashion to preserve the integrity of the central specimen from which the test material was cut. For boxed dry materials (ceiling tiles, flooring materials, etc.), a full, unopened box was obtained, and the test specimen cut from a central piece.

Answer to question 19:

It was observed that the emission factor versus time data could be classified into three groups. The following 3 equations were used as source models that specify how the emission rates (factors) change over time.

Power-law decay model:

Peak model:

Constant model:

where:

• EF: Emission factor ( mg/m²/h )
• t: time ( h )
• a, b, x_o: constants

The equations were fitted to the measured emission factor versus time data to determine the constants (a, b and xo), which were packaged into a material emission database.

Answer to question 20:

Concentration versus time (C-t) profiles from a chamber test were converted to emission factor versus time (EF-t) profiles, which were used to determine model coefficients of empirical source models through curve fitting.

The following equation, which is based on a mass balance in a test chamber, was used to calculate emission factors from measured chamber concentration.

where:

• EF: Emission factor calculated based on measured concentrations ( mg/m²/h )
• N: Chamber air change rate (l/ h ) as a ratio of Q (air flow, m³/h ) and V (chamber volume, m³ )
• L: Material loading ratio ( m²/m³ ) as a ratio of A (specimen area, m² ) and V.

As recommended in ASTM Standard D5116, this is a method to calculate emission factors when emission rate is not constant and/or the chamber has not reached steady-state (ASTM, 1997).

When there are not enough data points available and the chamber has reached steady-state, Equation 1 can be simplified to Equation 2.

ASTM Standard D5116 recommends a method to calculate the slope of ΔC/Δt based on 3 data points. Since the method leads to n−1 emission factors from n+1 data points, a new method based on two data points was used to calculate ΔC/Δt in this research (Equation 3). Although the new method saves only one more data point, it sometimes makes a difference when there are not many data points available.

Therefore, a new set of concentration versus time data was also generated by Equations 4 and 5.

The emission factor versus time (EF-t) data were used to calculate the coefficients of source models via curve-fitting. It was observed that the EF-t profiles generally follow a power-law or peak type of decay functions (see Question 19). Curve fitting was performed using SigmaPlot 2001 for Windows Version 7.0, which uses a non-linear regression method for a non-linear model. Microsoft Excel was not used for curve fitting in this research since it uses transformed regression model for logarithmic, power, and exponential trend lines. The model coefficients and R2 value were packaged into the database.

Answer to question 21:

The research team at the NRC is aware that the emission data can have a large amount of uncertainty due to specimen, experimental, and environmental variability. Any given material may have both multiple manufacturing sources and specific production types/classes. In addition, a single product may demonstrate significant variability in emission characteristics due to transient production factors (changing feedstock materials, environmental and process conditions during manufacture, etc.).

The research team used an oriented strand board as a case study to quantify the uncertainty of VOC emission factors from a single material likely due to the non-homogeneous nature of raw material ingredients, manufacturing processes, and handling/storing processes. A series of samples of oriented strand board (OSB) were collected and subjected to chamber tests for VOC emissions under standardized conditions (23oC, 50% RH, 1 air change per hour, 0.4 m2/m3 loading). Specimens were collected directly from the mill sites of 3 different manufacturers. Repeat samples were also collected from the same retail outlet on 3 separate occasions (same manufacturer, 3 different production dates), from separate panels produced on the same production date, and from multiple locations within the same panel. Variability in the VOC emissions from these samples was found to exceed the analytical uncertainty by an order of magnitude in some cases.

However, the software currently does not provide any information on the uncertainty levels associated with emission factors or indoor air concentrations. Considering that the uncertainty levels are different for different chemicals and materials, the variability test results based on OBS could not be applied to other materials in the database.

Answer to question 22:

Emission coefficients are based on tests conducted at test conditions of 23°C and 50% and a period of ~120 hours on average (the testing period ranged from 72 to 362 hours for dry materials and 78 to 440 hours for wet materials to capture the concentration decay portion properly). Extrapolating results to longer time periods or to other environmental conditions will affect the reliability of the results and should be done with caution.

Answer to question 23:

The simulation algorithms in the software were verified with an independent computer code written in Fortran.

Answer to question 24:

In its current state, the software considers building materials only as sources of volatile organic compounds. It does not provide estimation of the "sink effect" by building materials. Therefore, the potential for reductions in peak VOC concentrations and subsequent re-emission of VOCs over prolonged periods of time is not considered.

Answer to question 25:

The research team at the NRC monitored indoor air concentrations in a research house for approximately a year while and after it was constructed. Building/furnishing materials collected from the construction site were also subjected to emissions testing and the results used to create a separate database (not included in the publicly released database as they do not meet the strict specimen collection protocol). The measured concentrations were compared with the simulation results generated by the software based on 10 sources (judged to be major potential VOC sources in the newly constructed house). The results showed that the software could predict the TVOC level reasonably well in spite of the limited number of building products simulated. For individual VOCs, the concentration measurements in the house were generally under-estimated by the software. Under-estimation was more prevalent for long-term emissions. Un-modeled sources are likely to be the cause of this discrepancy. The importance of identifying sources and estimating the source areas accurately was emphasized through the exercise.

Answer to question 26:

Yes. As additional materials are tested by the NRC or by other test labs (according to the strict test protocols and quality requirements established by the NRC as minimum requirements for database input) it is the intent to add such results. Manufacturers and/or other agencies interested in submitting specimens for testing or data from accepted labs may contact Bob Magee (Robert.Magee@nrc-cnrc.gc.ca and 613-963-9931) for additional information.

Answer to question 27:

Some of data were re-analyzed in Version 1.1.

A portion of the data used to determine coefficients of the source models was eliminated for "wet" materials (e.g. paints, stains, caulks, adhesives, etc) in Version 1.1. For Version 1.0, all experimental data were used to determine source model coefficients. It was found later that the source models tend to put the emphasis on the high concentrations before 12 hour and provide less desirable prediction capabilities for long-term (t > 12 hour) emissions compared to short-term (t < 12 hour) emissions. Therefore, the coefficients based on the experimental data after 12 hour were selected for "wet" materials in Version 1.1. Using the experimental data after 12 hour for curve-fitting (Question 20) assures that the long-term emissions can be predicted better. This is to reflect our judgment that occupants tend to avoid the sources when the emissions are strong (e.g., right after a painting activity) and, therefore, predicting the long-term emissions well is more important from the occupant exposure perspective.

The source models and the curve-fitting method are identical to those explained in Question 19 and 20, respectively. The re-analyzed materials for Version 1.1 include adhesives (AD3, AD6, AD10, Crp7a, Lin2), caulks (CK2, CK5, CK9), paints (PT5, PT7, PT8), polyurethanes (UR3, UR5, UR8) and waxes (WX2, WX4, WX6). Additionally, peak models for UP1 in Version 1.0 were replaced with power-law decay models in Version 1.1 to improve the predictability of long-term emissions.

27.2 Two bugs in Version 1.0 were fixed in Version 1.1:

• the function related to "C_in" of the "Calculation" tab did not take the user's input in Version 1.0. The function now works with Version 1.1
• in Version 1.0, simulation cases with an identical 'case name' could be generated. In Version 1.1, a user cannot generate simulation cases with a duplicate or blank 'case name'