By Ramin Farnood, Mehrrad Saadatmand & Hooman Foroughi
Natural gas is odorless and odorants are added to alarm the consumer in case of a gas leak. Odor fading, i.e. the removal and loss of odorant from gas, could pose a potential risk to consumers. In this paper, the results of experiments conducted to investigate possible chemical and physical mechanisms responsible for odor fading are summarized. Evidence of chemisorption, adsorption and desorption of tertiary butyl mercaptan (TBM) as an odorant on the iron oxide inside pipes were observed. It was found that by increasing pressure, rusted surface of pipes, temperature, and also by decreasing the gas flow rates, TBM removal was increased. It is reasonable to consider that the above parameters could also affect odor at the natural gas delivery point. This manuscript is the condensed version of the original paper entitled “Odor Fading in Natural Gas Distribution Systems” published in the journal of Process Safety and Environmental Protection in March 2015.
Keywords: Odor fading, Natural gas, Pipeline, Mercaptan, TBM
A continuous gas flow laboratory apparatus consisting of a removable sample holder was used to mimic the gas pipeline continuous flow. Different types of samples were used in this study including polyethylene pipe material, stainless steel pipe material, and analytical grade iron oxides. Pure methane with 51 ppm of TBM was used to represent the odorized natural gas. In each experiment, the odorized gas was directed over the sample and the concentration of mercaptan at the inlet and outlet were measured using gas chromatography (GC).
In addition, samples were analyzed using X-Ray Photoelectron Spectroscopy (XPS) to examine the possible interactions of TBM with the pipe material.
2. Results and discussion
The odor fading process is the net result of the exchange of mercaptan between the bulk of the gas and the surface of the pipe and is comprised of a series of linked steps. In the first step, the odorant comes in contact with the inner surface of pipe walls through mass transfer. In the second step, the odorant interacts with the surface through chemical reactions and/or adsorption. In the third step, some of the odorant could be desorbed from the pipe surface back to the natural gas. In the following sections the possible adsorption or chemisorption of TBM to the inner pipe wall for different pipe materials is discussed. Also, effects ofparameters that are involved in the transfer of TBM from natural gas to the pipe wall are investigated.
2.1 Surface analysis experiments
Nano-powder iron oxide, polished steel, stainless steel, and polyethylene samples were used to check the possible reactions of TBM with pipe surfaces. As shown in Figure 1, in the XPS analysis results, three different types of iron oxides could be recognized on the rusted surface of steel pipes: yellow iron oxide (α-FeO(OH)), red iron oxide (γ- FeO(OH)), and black iron oxide (FeO. Fe2O3). The XPS measurements showed no trace of sulfur on tempered stainless steel samples. After short exposure to TBM, the polished steel pipe sample showed a small trace of sulfur but it was not discernible from the background noise. Contrary to stainless steel and polished steel pipe samples, the results of experiments on all rusted steel pipe samples showed significant traces of sulfur. XPS results also showed significant sulfur peaks at the binding energies corresponding to sulfurcontaining compounds, proving the chemical interactions (chemisorption) between mercaptan and these samples.
By increasing the exposure time to TBM, evidence of sulfurcontaining compound was eventually detected on the polished stainless steel sample. However, the rate of interaction of polished steel with mercaptan was significantly smaller than those of rusted steel pipe and iron oxide.
Moreover, according to the XPS results, the amount of the sulfur detected on the surface of black iron oxide sample after short time (1 to 4 hours) exposure to TMB was 0.26% (atomic percent) and it did not change by prolonged exposure (about 20 hours) to TBM. The red iron oxide showed a higher value of sulfur on its surface, i.e. 0.32% for short exposure and 0.48% for long exposure experiments. Lastly, the yellow iron oxide had the highest concentration of sulfur atoms on its surface among the three different types of iron oxides. The short exposure experiment for yellow iron oxide resulted in 0.42% and the long exposure experiment showed 0.58% of sulfur. Two clear conclusions can be made from these results: 1) chemisorption of TBM by the steel pipe is mainly due to the formation of iron oxide (rust) on the pipe surface, and 2) chemi-sorption mostly occurs within the first few hours.
The XPS measurements of polymer pipe also showed traces of sulfur on the sample surface. This suggests possible chemical interactions between TBM and polypropylene pipes.
2.2. TBM removal
Several breakthrough tests were conducted to ascertain the amount of TBM removal from the odorized gas by the samples. Here, the breakthrough is defined as the outlet concentration of TBM over its initial concentration. In other words, breakthrough is the concentration of TBM at the reactor outlet which is not reacted with iron oxide. These runs showed breakthrough of TBM within the first 50 min of the experiments. As the iron oxide samples became saturated with TBM, the concentration of TBM gradually increased and reached the inlet concentration. In the following sections, effects of various operating parameters on the TBM removal are discussed.
2.2.1 Effect of sample loading
It is expected that the onset of breakthrough to be dependent on the amount of iron oxide. The results of sample loading experiments consistently showed that higher loads of iron oxide needed longer breakthrough times. For 0.018 g of iron oxide powder, the breakthrough time was approximately 1 min, while for the 0.1 g of the sample, the breakthrough happened after about 20 min. These results suggest that the extent to which a gas pipeline is rusted significantly affects the amount of mercaptan required for odorization and the duration of the pre-odorization of the pipeline.
2.2.2 Effect of flow rate
The gas flow rate can also influence the breakthrough of mercaptan. By increasing the flow rate, the breakthrough occurred earlier. This means that at higher flow rates, the target mercaptan concentration at the tail end of the pipe can be reached after a shorter period of time. This is consistent with the field experience where odor fading has been often associated with conditions where the gas flow rate was either very low or zero.
2.2.3Effect of pressure
Increasing the pressure increases the adsorptive capacity and therefore it took longer for the exit gas concentration to reach its plateau level. It is known that under isothermal conditions, higher pressure results in greater adsorption of odorant molecules. By increasing the pressure, the partial pressure of odorant in the gas phase increases, and hence the adsorption rate will also increase. This is the likely reason for the slightly lower breakthrough rate at elevated pressures. The distribution system of natural gas usually operates at pressures below 60 psig, which is supplied via a high-pressure pipeline typically operating around 220 psig. These results highlight the importance of considering the pressure effect on the odorant loss in the gas distribution pipelines.
2.2.4 Effect of temperature
Since the effective diffusivity in gases is directly related to temperature, increasing the temperature will increase the TBM diffusivity resulting in a greater TBM removal and therefore a delay in the breakthrough time. It should be noted that increasing the temperature also increases desorption of TBM from iron oxide and decreases the net adsorption rate. This finding underscores the importance of the seasonal change in the amount of mercaptan required for odorization. For example, at the depth of around 2 meters (a typical depth for gas pipelines) the average temperature could vary from 4°C to 21°C in Ottawa, Ontario. Such a large temperature change will have a great impact on the amount of mercaptan required for the pre-odorization of new gas pipelines in the winter time compared to the mercaptan needed in the summer.
2.3. Evidence of desorption
To check the possible desorption of TBM from iron oxide samples, the gas feed was changed from odorized gas to mercaptan-free pure methane. The iron oxide sample, which was initially exposed to 50 ppm TBM for about 70 minutes, was then exposed to pure methane for about 200 minutes for desorption. It was found that the sorption of TBM on the iron oxide was reversible to a great extent. The experimental data showed that during the first phase of this experiment (i.e. 70 min exposure to TBM), 5.2 mg of TBM was absorbed or adsorbed per each gram of iron oxide sample. However, during desorption process, only about 1.5 mg of TBM was removed per gram of iron oxide after 200 min. This indicates that desorption rate was considerably slower than the adsorption process. Also, from this experiment, it is not clear what fraction of TBM could be desorbed by prolonged exposure to pure methane.
This study confirms that mercaptans could interact with both stainless steel and polypropylene pipe material via chemical reaction (chemi-sorption). In addition, rusted steel pipes appear to be more prone to TBM removal by chemi-sorption. Breakthrough studies showed that TBM removal is a function of the amount of iron oxide (rust) in the pipe and the operating parameters of the system including pressure, temperature, gas flow rate, and the odorant concentration.
RAMIN FARNOOD Professor Ramin Farnood is Associate Chair of Research and a Principal Investigator in the Department of Chemical Engineering & Applied Chemistry in University of Toronto. His research interests are in the areas of Bio-based materials and bioenergy, Environmental Engineering and Microstructure Characterization. E: firstname.lastname@example.org
MEHRRAD SAADATMAND Mehrrad Saadatmand, a recent graduate of University of Toronto, has an extensive research and publication record on a wide range of topics related to heavy oil, natural gas and fluid mechanics. He is currently an adjunct professor in the New York City College of Technology. E: email@example.com Twitter: @smssaadat
HOOMAN FOROUGHI Hooman was a postdoctoral fellow in Professor Ramin Farnood’s research group at the University of Toronto in 2012. Hooman’s research interests include mathematical modeling, simulation, and optimization of water and wastewater treatment processes and polymer recycling. E: firstname.lastname@example.org