Winter Newsletter 2011
Light Bulbs Were Falling: Application of
FT-IR, 1H-NMR and GC/MS in a Failure
Investigation of a Polycarbonate
A light fixture manufacturer received complaints from customers that the bulbs were falling from the fixtures. Several thousand fixtures had already been installed, each holding multiple light bulbs. The sockets, known as “tombstones,” showed visible cracks. These failures were showing up about a year after the bulbs were installed.
Prior to replacing the “tombstones,” the manufacturer needed to know the cause of the failure to prevent future issues. They contacted Chemir to perform a comparative analysis between a control and a failed sample. The first step was an initial characterization or “screening” of the samples, to inspect the failures with optical microscopy and identify the base polymer.
 |
| Figure 1: Image of the Unused and Used “Tombstones” |
Optical Microscopy
Each failed socket was observed to contain random fractures. In addition, there seemed to be slight yellowing of the surface of the “tombstones,” in and around the cracked areas. Upon close inspection, the “tombstones” and surrounding areas were observed to be coated with a fine white residue. Figure 1 shows the unused and used “tombstones” as well as an expanded view of a fracture (insert).
Horizontal Attenuated Total Reflectance (HATR) and Proton Nuclear Magnetic Resonance (1H NMR)
The first step of the general strategy for understanding this polymer failure was to identify the bulk material. It was thought that identification of the bulk material could yield clues as to possible failure routes. Figure 2 exhibits the HATR spectrum and computer-generated spectrum match for polycarbonate. Figure 3 exhibits the 1H NMR spectrum, which was observed to be consistent with polycarbonate.
 |
| Figure 2: HATR Spectrum of Control “Tombstone” |
 |
| Figure 3: 1H NMR Spectrum of Control “Tombstone” |
In addition, a comparison was made between a control sample and a failed sample. Comparison between these two samples showed that there are chemical differences. The failed samples contained an additional IR absorbance band at approximately 1730cm-1, indicating the possible presence of an amide stretch, see figure 4 below.
 |
| Figure 4: Image of the HATR spectrum for Control and Failed “Tombstone.” |
Fourier Transform Infrared Spectroscopy (FT-IR) of Residue
After careful observation of the area surrounding the failures, a white residue was observed. The white residue was analyzed and to be consistent with dicyclohexylamine by a computer-generated spectral match. Figure 5 below shows the FT-IR match.
 |
| Figure 5: FT-IR Spectrum of Unknown “White Residue” |
Conclusion
The polymer used in the lamp holders was a common amorphous thermoplastic, polycarbonate. We observed the presence of an amide signal in the failed sample, which was not present in the control sample. In addition, we found a white residue present on the light fixture, consistent with dicyclohexylamine. The results suggest that a chemical reaction was happening between the amine and the polycarbonate. Essentially, it was shown that an “aminolysis” of the polymer (effectively an “unzipping” of the polymeric structure) was likely occurring in the failed samples. The next step in the analysis was to extract and quantify the amount of dicyclohexylamine to prove the correlation between amount of exposure and failure. Table 1 summarizes the results of the quantitation.
 |
| Figure 6: GC/MS Spectrum of Unknown “White Residue” |
Table 1: Results of the quantitation for Dicyclohexylamine (DCH)
Sample Type |
Concentration DCH (μg/g) |
Tombstone – Unused |
0.97 |
Tombstone – New |
3.1 |
Tombstone – Intact, observed white residue |
35 |
Tombstone – Failed, removed from service |
59 |
The strong correlation between the failed sample and the amount of DCH present suggested that the initial assessment of aminolysis was likely a major contributing factor in the failure of the sample. To prove that this was the case, a forced degradation study was completed. The samples were exposed to known amounts of DCH for various amounts of time and at different temperatures inside sealed jars. This information was useful as it provided the client with time limits for failure for the currently installed fixtures. Table 2 exhibits the results of the forced degradation study. To control for temperature effects (if present) on the polymer, a set of samples were held at the specified temperature without DCH present in the jar. The control group did not have any failures. As may be observed, as the concentration and temperature increase, so did the rate of failure.
Table 2: Results of the Forced Degradation Study
Temperature |
Average Concentration in Headspace (μg/g) |
Days to First Observed Failure |
Control Group Failure (Y/N) |
23°C |
0.06 |
56 |
N |
35°C |
0.10 |
8 |
N |
53°C |
0.29 |
7 |
N |
65°C |
0.39 |
4 |
N |
To show the extent of the failure, the samples were removed from the sealed jars and an image was recorded. Figure 7 shows the control and failed samples after 90 days of exposure. The insert shows the jar with duplicate “tombstones” hung to be exposed to the headspace.
 |
| Figure 7: Forced Degradation Study Results (Insert – Experimental set up of Jar) |
Chemir honors the confidentiality of our clients. This case study is representative of the types of projects performed by Chemir. Any similarity to a specific project is purely coincidental.
