How a Laboratory Can Help You Identify Problem Chinese Drywall
by Alyson Fortune, Air Quality Scientist; Michael Tuday, Director of R&D; Nicole Pannone, Air Service Specialist
For the past four years, the U.S. Consumer Product Safety Commission (CPSC) has been receiving complaints from homeowners regarding corrosion and odors in their homes linked to imported drywall. The problem drywall, which was installed in homes between 2004 and 2007 and is commonly referred to as “Chinese drywall,” has resulted in more than 2,500 complaints to the CPSC. The complaints originated from homeowners primarily in the southeastern part of the United States, but have since been reported throughout the country.
Homeowners have linked their Chinese drywall to corrosion in their air conditioner coils, corrosion in copper wiring, and emission of foul odors. The odors have been described as smelling like rotten eggs, burnt matches, and other sulfurous smells.
Columbia Analytical has been studying this issue and testing both foreign and domestic drywall samples since February 2008. Laboratory tests have been developed to aid in the identification of defective drywall products. These tests may be used to verify visual home inspections and determine if corrosion effects are from drywall and not from other household items, such as carpets, cleaners, paints, or personal care products.
This article presents a chronology of how Columbia Analytical established their test methods for determining problem drywall and how each of the issues that arose was resolved with a laboratory solution.
Laboratory Testing to Determine Corrosion
Before laboratory testing of drywall can begin, it is necessary to answer the fundamental question of whether the drywall caused the corrosion that was observed in residential and commercial structures. To investigate whether or not a causal relationship existed, laboratory experiments were conducted using copper tubing that was exposed to 3×3 inch samples of drywall that had been collected from homes where corrosion was present and from homes where it was not. The tests were conducted in 1L glass chambers at 37º C for 21 days.
The results of this simple experiment were conclusive: in the absence of any other contributing factors, the copper tubing exposed to the warm, humid environment in the presence of the “defective” drywall blackened and corroded, while the tubing exposed to the same conditions in the presence of the non-defective drywall did not.
Differentiating Corrosive and Noncorrosive Drywall
In order to evaluate what might be different about the composition of the drywall that caused corrosion and drywall that did not, a sample of each type was analyzed by direct thermal desorption.
To prepare the samples, a sample of each kind of drywall was finely crushed and placed directly into a thermal desorption tube. Each sample was thermally desorbed, followed by analysis by gas chromatography/mass spectrometry (GC/MS).
In the sample that caused corrosion, a very large amorphous peak was present at the end of the chromatographic run (see image). This unique peak was not found in the chromatographic results from the non-corrosive drywall samples.
Based on its mass spectrum, the peak in question was determined to be elemental sulfur.
Specifically, the peak represents a particular allotrope of elemental sulfur: orthorhombic cyclooctasulfur, a molecule comprised of eight sulfur atoms in a ring. The chemical “structural finger-print” (i.e., mass spectrum) of this compound is unique, leading to a definitive identification. The presence of this compound allows for clear differentiation between corrosive and non-corrosive drywall.
Method Development for the Analysis of Elemental Sulfur
After the initial identification of elemental sulfur in corrosive drywall, an optimized solvent extraction approach was developed to more effectively resolve and quantify this compound. This approach yielded a much cleaner peak, simplifying the interpretation and identification, providing a result in mg per kg (ppm) of elemental sulfur for both the affected and unaffected drywall samples.
Subsequently, an alternate method for quantifying elemental sulfur was developed using gas chromatography with electron capture detection (GC/ECD). This alternate method offers comparable sensitivity to the GC/MS method, but allows for more rapid turnaround time at a substantially lower cost.
Method Development for the Analysis of Reduced Sulfur Compounds
In preliminary chamber testing, two compounds, in particular, were present in the off-gassed samples: carbonyl sulfide and carbon disulfide. However, in the mechanisms that had been considered or proposed for the corrosion and blackening of the copper elements, it seemed unlikely that the presence of these two compounds would produce the extent of the observed corrosion. The presence of hydrogen sulfide was suspected. To test this hypothesis, 25g of drywall were placed in the chamber, with 500mL of humidified zero-grade air. The sample was incubated for 72 hours at 37º C.
After the incubation period, the headspace was analyzed by gas chromatography/sulfur chemiluminescence detection (GS/SCD) for 20 target analytes, including hydrogen sulfide (H2S), carbonyl sulfide (COS) and carbon disulfide (CS2). The reporting limit for each compound ranged from 2-5 ppbV in air, and the results were then converted to micrograms per kilogram of drywall sample.
Summary of Data Collected for Elemental Sulfur and Hydrogen Sulfide
In 2009, using the two aforementioned analytical techniques for measuring elemental sulfur (S8) and hydrogen sulfide (H2S), several hundred drywall samples (primarily from homes in Florida, Louisiana, and Virginia) were analyzed.
The following graphs summarize the range of concentrations detected in these drywall samples.
- For the S8 analysis, the population was 176 samples (n=176).
- For the H2S analysis, the population was 253 (n=253).
- Units for the S8 analysis are mg/kg.
- Units for the H2S analysis are µg/kg.
- In each graph, the pink diamonds indicate the mean.
- In some cases, no manufacturer labeling was visible, so it was not possible to identify the samples and sort the data by manufacturer.
- From the labeling that was available, it was evident that the drywall did not originate from one particular manufacturer.
Elemental Sulfur (S8) (left graph): Range of Detected Concentrations (mg/kg).
Hydrogen Sulfide (right graph): Range of Detected Concentrations (µg/kg)
S8 – Hydrogen Sulfide (H2S) Relationship
The relationship between S8 and H2S was examined further by considering samples for which both tests were performed.
- Sample population was 228 drywall samples (n=228).
- 146 samples tested positive for S8, 82 tested negative.
- For the population that tested positive for S8, 97% also tested positive for H2S, indicating a strong correlation between the two tests.
- Similarly, for the samples that tested negative for S8, 93% also tested negative for H2S.
S8 – Carbonyl Sulfide (COS) Relationship
The relationship between S8 and carbonyl sulfide was also studied, and while evident, it was not as strong as the relationship between H2S and S8.
- Sample population was 148 drywall samples (n=148).
- 72% of the samples that tested positive for elemental sulfur also had carbonyl sulfide.
- 87% of the samples that were negative for S8 were also negative for carbonyl sulfide.
S8 – Carbon Disulfide (CS2) Relationship
Finally, the relationship between S8 and carbon disulfide was studied.
- Sample population was 148 drywall samples (n=148).
- In samples where S8 was detected, 99% tested positive for CS2
- However, in samples where S8 was not detected (i.e. non-corrosive drywall), CS2 was still detected 44% of the time.
This last finding suggests that CS2 may also be present in non-corrosive drywall samples.
Association Between S8, H2S, and Blackening/Corrosion
There appeared to be a strong association between the three indicator tests for corrosive drywall – S8, H2S, and the Jar Blackening/Corrosion test – as demonstrated by the following visual.
For a sample population of 143 samples:
- 87 were positive for S8
- Of those 87 samples, 95% were both positive for the H2S and Blackening/Corrosion tests.
- 56 samples that were negative for S8
- 87% of those samples were also negative for both the H2S and Blackening/Corrosion tests.
These results support the claim that elemental sulfur (S8) is a unique and reliable indicator of corrosive drywall.
Further Study: Evaluation of Indicator Tests on a Representative Subset
Having established a unique and reliable indicator for the corrosive drywall, Columbia Analytical designed an experiment to evaluate all three indicator tests (S8, H2S, and jar test for blackening/corrosion) on a set of representative corrosive and non-corrosive samples.
Three samples of non-corrosive drywall were selected (collected from homes where no corrosion was observed), and four samples of corrosive drywall were selected. Results from this limited sample set are representative of results from hundreds of samples that have been analyzed over the past several years.
- Three samples of noncorrosive drywall
- Domestic (2 brands – one purchased in CA, one from an unaffected FL home)
- Imported from Mexico (from another unaffected FL home)
- Four samples of corrosive drywall
- Elemental Sulfur
- Reduced Sulfur Gases
- Jar Test: Blackening/Corrosion
After performing the three analyses on the seven samples, a strong correlation between the two subsets was observed.
- The samples that were identified as non-corrosive tested negative for hydrogen sulfide, carbonyl sulfide, and carbon disulfide. They also tested negative for elemental sulfur, and they did not produce any observable blackening or corrosion in the Jar Test.
- The samples that had previously been identified as corrosive tested positive for all of the above-mentioned parameters.
The donut graph below illustrates the chemical profiles of each of the four samples of corrosive drywall. Each of the four samples is represented by a ring in the donut and has a similar yet unique signature, with varying concentrations of the key parameters (S8, H2S, COS, and CS2). Abundance of each parameter has been normalized for graphing purposes.
Comparison of Four Corrosive Drywall Samples
Determination of Other Potentially Harmful Chemicals Emitted by Corrosive Drywall
Having observed a variety of sulfur compounds emitted from corrosive drywall samples, it was desirable to evaluate what other potentially harmful chemicals might be off-gassing from corrosive drywall.
To study this, static chamber tests were performed. A static test was selected rather than a dynamic one in order to capture all of the potential events and to avoid sample dilution over time.
For this study, off-gas samples were collected after 72 hours for the following analyses on the specified media:
- Volatile Fatty Acids were collected on a sodium hydroxide-treated silica gel sorbent tube and analyzed by gas chromatography/mass spectrometry (GC/MS) via an established, validated in-house method. (Refer to Graph 3.1)
- Aldehydes were collected on silica gel sorbent tubes treated with 2,4-Dinitrophenylhydrazine (DNPH) and analyzed by high performance liquid chromatography with UV detection (HPLC/UV) via EPA Method TO-11A. (Refer to Graphs 3.2 and 3.3)
- Volatile Organic Compounds (VOCs) were collected on thermal desorption tubes and analyzed by thermal desorption/GC/MS via EPA Method TO-17 (Refer to Graph 3.4).
Graph 3.1 – Volatile Fatty Acid Data Summary
This graph focuses on the detected concentrations of acetic and formic acid, two compounds that have been identified as possible contributors to the corrosion phenomenon.
- No clear pattern emerged regarding formic acid concentrations.
- Overall, higher concentrations of acetic acid were detected in the corrosive drywall samples than in the non-corrosive drywall samples.
Graph 3.2 – Aldehyde Data Summary
This graph shows the abundance of each of the detected aldehydes (total ng/sample) on the Y-axis.
- The red lines represent the 4 corrosive samples.
- The blue lines represent the 3 noncorrosive samples plus a duplicate.
- Some levels of aldehydes were higher in noncorrosive samples.
Levels of formaldehyde and acetaldehyde (see Graph 3.3) were similar in both corrosive and non-corrosive drywall.
Graph 3.3 – Formaldehyde & Acetaldehyde Data Summary
This graph summarizes the findings of formaldehyde and acetaldehyde, the two aldehydes of main concern with regards to indoor air quality and drywall emissions.
Levels of formaldehyde and acetaldehyde were similar and/or higher in non-corrosive samples relative to the corrosive samples.
Graph 3.4 – VOC Data Summary: Total VOC (TVOC as Toluene)
Each of the seven samples is shown on the x-axis with its total VOC (TVOC) concentration quantified as toluene in µg/kg.
- The red dashed line shows the average TVOC concentration for the 4 corrosive samples
- The purple dashed line shows the average TVOC concentration for the 3 non-corrosive samples.
As with the aldehydes, the TVOC levels were similar if not higher in the non-corrosive drywall samples than the corrosive drywall samples.
The average corrosive TVOC value was approximately two times less than the average non-corrosive TVOC value.
In addition to determining total VOC concentrations in the samples, the compounds present in the off-gas were evaluated as tentatively identified compounds (TICs) to further characterize the corrosive and non-corrosive samples.
The compounds detected in the corrosive subset were very chemically similar, even though they originated from different homes in different states. Low levels of other organic acids than acetic acid, such as isobutyric acid, propionic acid, and hexanoic acid, were also detected.
The following list of compounds was unique to the corrosive drywall samples:
• Isobutyric acid
• Propionic acid
• Hexanoic acid
• Sesquiterpene hydrocarbons
• Propylene glycol
Non-corrosive domestic drywall samples were also chemically very similar within the subset.
The following list of compounds was unique to the non-corrosive drywall samples:
• C8H12 Alkylbenzenes
• C10H14 Alkylbenzenes
|• p-Chloro(trifluoromethyl) benzene
• Diethylene glycol
• Propyl propionate
The following compounds were seen in both corrosive and non-corrosive drywall:
• Acetic Acid
• Texanol Isobutyrate
Odorous Compounds Detected in Corrosive Drywall
Corrosive drywall has a very complex odor character. In its dry form, the smell has often been described as a burnt match or burnt firecracker odor. In its wet form, it can be reminiscent of driving by a petroleum refinery, with a sour smell.
From the various tests that were performed on the corrosive samples, a wide range of odorous compounds have been identified:
|• Acetic acid
• Propionic acid
• Isobutyric acid
• Hydrogen sulfide
• Dimethyl sulfide
• Carbon disulfide
• Ethyl isopropyl disulfide
• bis (1-isopropyl) Disulfide
• Ethyl isobutyl disulfide
These odorous compounds were detected via EPA TO-17 and thermal desorption GC/MS in all four corrosive samples. They were not present in any of the non-corrosive samples. Many of these odorous compounds were also seen in a similar 2008 study conducted by Burdack-Freitag et al1.
Causes of Blackening and Corrosion of Copper and Other Metals
Columbia Analytical conducted a series of jar blackening/corrosion tests in an attempt to pinpoint the key chemicals that cause blackening and corroding of copper and other metals.
In each of the jar tests, a piece of clean copper tubing was placed in the glass chamber and a warm, humid environment was simulated. Then, chemicals or compounds that had been identified as likely present in the drywall or in the home environment were introduced to the environment in the chamber.
- Elemental sulfur
- Elemental sulfur added with carbon monoxide
- Hydrogen sulfide
- Iron disulfide
- Sulfur dioxide
- Carbonyl sulfide
- Carbon disulfide
- Formic acid and acetic acid
- Sulfuric acid
The accompanying image shows that each of the chemical combinations yielded slightly different blackening/corrosion results. Based on the results of these tests, and supported by additional research, the primary culprits in this phenomenon appear to be hydrogen sulfide, carbonyl sulfide, and acetic acid.
Based on the preceding body of laboratory research, the following conclusions have been drawn:
- The test for Elemental Sulfur (S8) can be used as a reliable marker for corrosive drywall.
- Blackening of copper is consistent with sulfide attack.
- Corrosion appears occur by a two-pronged mechanism: sulfide attack and formicary (ant’s nest) corrosion.
- Hydrogen sulfide, carbonyl sulfide, and acetic acid appear to be the main contributors to the corrosion phenomenon.
For guidance on what to do after positive identification of corrosive drywall, visit the Florida Department of Health website for Case Definition, or the Consumer Product Safety Commission Drywall Information Center.
1. Burdack-Freitag, A., Mayer, F., and Breuer, K. Identification of Odor-Active Organic Sulfur Compounds in Gypsum Products. Clean, 37 (6), 459 – 465 (2009).
–> Learn more about Chinese Drywall Testing…
–> Learn more about Chinese Drywall Testing Protocol…
Tags: acetic acid, allotrope, amorphous peak, analysis, Blackening, carbon disulfide, carbonyl sulfide, Chinese Drywall, Consumer Product Safety Commission, copper tubing, Corrosion, Corrosive, Corrosive Drywall, COS, CS2, CSPC, drywall, elemental sulfur, Florida Department of Health, Gas chromatography, glass chamber, GS/SCD, gypsum, H2S, headspace, humid, hydrogen sulfide, identify, Laboratory, mass spectrum, Method Development, Noncorrosive, Odorous Compounds, orthorhombic cyclooctasulfur, problem drywall, Reduced Sulfur Compounds, sulfur chemiluminescence detection, testing