Analysis and Identification of Contaminants in Diesel Fuel Filtration and Storage Systems

Size: px

Start display at page:

Download "Analysis and Identification of Contaminants in Diesel Fuel Filtration and Storage Systems"

Francis Floyd
5 years ago
Views:

Paper No. IFC10-020 Analysis and Identification of Contaminants in Diesel Fuel Filtration and Storage Systems Steven R.

1 Paper No. IFC Analysis and Identification of Contaminants in Diesel Fuel Filtration and Storage Systems Steven R. Westbrook Southwest Research Institute James Doyle Philip Johnson Donaldson Company ABSTRACT This paper presents the results of analysis of contaminants found in diesel fuel storage tanks and filtration systems. The analyses show that the contaminants are not of the type commonly found in diesel fuel. Numerous analytical techniques were employed including electron microscopy, x-ray diffraction, energy dispersive spectroscopy, ultraviolet/visible spectroscopy, infrared spectroscopy, and gas chromatography-mass spectrometry. INTRODUCTION On new Tier 4 diesel engines some fuel injector manufacturers are targeting minimum fuel cleanliness levels of ISO 17/16/13 or better at all times, to maintain engine warranty. This is required so that the on vehicle filtration will have a hope of meeting the injector requirements of ISO 12/9/6 on engine post final filtration. This level of cleanliness on engine is needed to prevent damage to the new high pressure fuel injector s surfaces. This cleanliness level, 17/16/13, is met only occasionally in today s delivered fuels and cleanliness levels much dirtier are routine. Oil companies and fuel distributors acknowledge fuels are routinely in the 22/21/18 range upon delivery (that s about 16X dirtier than allowed by equipment warranty). Yes, there are supplies in the 16/15/13 range but they are not the norm and leave little room for error only taxing the on board filtration to its limits of performance. To have any certainty of cleaning a load of dirty fuel to those levels to maintain fuel injector warranty on the operating equipment, single pass filtration of the highest efficiency is nearly mandatory. The potential of kidney loop filtration with lower efficiency filtration is not feasible in many cases due to time, product volumes and system plumbing issues. Modern transportation fuels are subject to contamination throughout their journey from production to use. Water, dirt, corrosion products, and microbial growth are all common fuel contaminants. However, because of the complexities of modern diesel fuel, fuel distribution and storage, and fuel use, additional contaminants are becoming more common. The analysis and identification of some of these contaminants is presented herein. Contaminant A Contaminant A was recovered from a fuel clean up filter used on fuel in a storage tank. The fuel was owned by a large fleet operating in the upper mid-western United States. The fuel contained 2 volume % biodiesel. The fuel was causing premature plugging of dispensing filters. The problems were occurring during winter operation. A sample of the contaminant is shown in Figure 1. SwRI analyzed the contaminant by gas chromatography-mass spectrometry and obtained the chromatogram in Figure 2. The identities of the peaks labeled in Figure 2 are given in Table 1. The sample was also analyzed using infrared spectroscopy (FTIR). The FTIR spectrum is given in Figure 3. Of note is the peak at about 1730 cm -1. This peak is attributable to carbonyl bonds. That type of bond is to be expected since this tank contained biodiesel. However, the typical carbonyl peak for biodiesel comes at 1740 cm -1 due to the ester linkage. The fact that this peak comes slightly lower indicates that the sample contains carbonyl groups other than just esters. As shown in Table 1, Contaminant A was composed primarily of mono- and di-glycerides, by products of the production of biodiesel. In some cases, these compounds are present at sufficiently high concentration that they will precipitate at 10th International Filtration Conference 150

2 temperatures above the cloud point of the biodiesel blend. Typically, they will not redissolve until the fuel temperature reaches 20 to 25 C. In this instance, the filters that plugged were housed in a building kept at 15 C, well above the cloud point of the fuel. Table 1. Identification of Peaks Shown in Figure 2 Peak Number Compound(s) 1 C 18 methyl esters 2 mono-olein mono-palmitin and monostearin 3 4 mono-stearins Figure 1. Contaminant A: Collected from a Fuel Dispensing Filter 5 sitosterol 6 di-glycerides 2 3 } Minutes Figure 2. GC-MS Chromatogram of Contaminant A 10th International Filtration Conference 151

FTIR Spectrum 60 50 Transmittance, percent 40 30 20 10 4000 3500 3000 2500 2000 1500 1000 Wavenumber, cm -1 Figure 3.

This fleet had several underground storage tanks at numerous locations.

3 FTIR Spectrum Transmittance, percent Wavenumber, cm -1 Figure 3. FTIR Spectrum of Contaminant A Contaminant B Contaminant B came from a dispensing filter operated by a major fleet in the eastern part of the United States. This fleet had several underground storage tanks at numerous locations. Some tanks had dispensing filters rated at 2- micron porosity; some locations had dispensing filters with larger porosity ratings. Only the 2-micron filters were being plugged, even though all tanks received the same fuel. Figure 4 is a picture of a section of one of the plugged filters. SwRI collected material from the plugged filter but were unable to find a suitable solvent to allow a GC-MS analysis. Figure 5. SEM Photomicrograph of Contaminant B Figure 5 is a scanning electron microscope (SEM) photomicrograph of Contaminant B. Note that the particles are actually agglomerations of spherical particles that are less than 1 micron in diameter. SwRI analyzed the particles by x-ray diffraction, the resulting spectrum is Figure 4. Fuel Dispensing Filter Plugged with Contaminant B 10th International Filtration Conference 152

4 shown in Figure 6. As shown in Figure 6, the spectrum matched spectra for carboxylate salts. These salts are formed by the reaction of acid-type corrosion inhibitor and lubricity improver additives in the fuel with contaminants found in the fuel delivery and storage system. The resulting salts are surface active and are generally insoluble in both fuel and water. As mentioned earlier, the particles were plugging the 2-micron dispensing filters but leaving other filters in the system apparently unaffected. This is attributed to a combination of the porosity of the filter (no other filters in the system had such a small porosity) and the chemical attraction of the particles to the fibers of the filter (polypropylene fibers). The attraction of the particles to the fibers is demonstrated by the optical microscope image, Figure 7, showing the particles adsorbed to the surface of filter fibers from a previously unused filter. (Material collected from the used filter was mixed with diesel fuel. The mixture was then filtered through filters from an unused filter. Figure 6. X-ray Diffraction Spectrum of Contaminant B Figure 7. Contaminant B Particles Adhering to the Surface of Polypropylene Filter Fibers 10th International Filtration Conference 153

Contaminant C Figure 8 shows a bulk fuel farm at a small contract earth moving company based in the Midwest.

Bulk Fuel Storage Tanks This contractor company runs a small fleet of earth moving equipment and has some long term contract business with locked in pricing constraints.

Over the winter periods much of their work tails off and ideally they needed to be able to store fuels without fear of degradation.

5 Contaminant C Figure 8 shows a bulk fuel farm at a small contract earth moving company based in the Midwest. Their system is comprised of four twelve thousand gallon storage tanks and a couple of day tanks holding approximately two thousand gallons each. Figure 8. Bulk Fuel Storage Tanks This contractor company runs a small fleet of earth moving equipment and has some long term contract business with locked in pricing constraints. To control costs the contractor buys fuel at spot market in order to gain the best possible fuel pricing they can. Over the winter periods much of their work tails off and ideally they needed to be able to store fuels without fear of degradation. Historically, this customer had experienced some bacterial contamination problems in one of their tanks and they had also experienced filter plugging on their mobile equipment. Donaldson was called in by the company owners to address water and contamination issues during fuel storage and dispensing. Contaminated tanks were drained, opened, and cleaned prior to filling. The filter program installed for this customer was comprised of a main inlet filter and a smaller safety filter on the outlet side of their system. Donaldson also installed specialized water absorbing air breathers along with a dry air purge system to prevent water moisture ingression, and to assist with the drying of fuel while in storage. The theory being, by keeping the fuel clean and dry, the stored fuel should remain useable for longer periods. Figure 9 shows patch tests taken from the fuel before and after it had passed through the inlet filter into the main storage tanks, demonstrating that the fuel in the storage tank was clean. Donaldson estimated the cleanliness of the stored fuel at approximately ISO 14/13/11. Figure 9. Pre-Filtration and Post-Filtration Filter Patch Tests Showing Filtered Fuel Was Clean In late August 2009 the company filled the majority of their storage facility with No.2 ultra-low sulfur diesel. Approximately 40,000 gallons of fuel passed through the inlet filter into storage. During filling of the storage tanks the differential pressure experienced across the inlet filters did not detectably increase at a flow rate of 250 gallons per minute. As mentioned earlier; Donaldson installed a dry air purge across one of the fuel storage tanks and monitored the relative humidity of the air inside the tank above the fuel and the relative water saturation of the fuel. Figure 10 depicts the relative humidity of head space air above the fuel in storage over time. The red line is a tank without dry air purge showing high moisture content similar to outside atmospheric conditions. The pink line shows relative humidity of the air over the tank with 1 cfm dry air purge. The blue line depicts the tank head space air temperature and it closely follows ambient conditions outside the tanks from August through January. 10th International Filtration Conference 154

6 Figure 10. Relative Humidity Data for Tanks With and Without Purging Fuel Moisture Content (% Saturation) Top Surface Water Content Versus Time /14 10/24 11/3 11/13 11/23 12/3 12/13 12/23 1/2 Figure 11. Water Content of Top-Level Fuel Sample Fuel Moisture Content (% Saturation) Average Water Content Versus Time /14 10/24 11/3 11/13 11/23 12/3 12/13 12/23 1/2 Figure 12. Water Content of All-Level Fuel Sample 10th International Filtration Conference 155

7 Figures 11 and 12 show fuel surface water content and fuel average (through the vertical column of fuel) water content over time in storage. The light blue solid line is for the tank with no air purge. The dark blue dashed line is for the tank with air purge. Over time it can be seen that the moisture level in purged tank decreases as compared to the tank with no purge. From the data we believe it can be concluded that the fuel in storage remained comparatively clean and dry during its storage. In April the contractor started to use some of the stored fuel in the dry air purged tank. After pumping approximately 300 gallons of previously cleaned and dry fuel through identical elements to those used on the inlet filter, the outlet filter became completely blocked and stopped the discharge flow. The filters on the outlet filter were removed and returned to Donaldson for examination and determination of the cause of the blockage. Figure 13 is a photomicrograph of what was discovered in one of the layers of the filtration media. Through examination of the filter media Donaldson found contaminant in a similar layer in the outlet filter elements. No such levels of this contaminant were found on the similar inlet filters, even though they had been used to pre-clean the exact same fuel and many thousands of gallons had passed through them. Throughout the fuel storage period, Donaldson had collected fuel samples from the storage tank in order to monitor the moisture level of the fuel. Donaldson also performed a scan of each fuel using a UV-Vis spectrometer. Figure 14 shows the resulting scans (360 to 400 nanometers) for samples from August 09 to May of These scans show what Donaldson believes is the disappearance of fuel additive as it reacts in the fuel tank. As you can, see the additive peak at about 378nm decreases readily with time. The other 2 lines are refinery barrel diesel samples from Donaldson s corporate tech source at different times of the year. One sample shows a similar peak the other does not. Figure 13. SEM Photomicrograph of Contaminant C 10th International Filtration Conference 156

Figure 14. UV-Vis Spectra for Fuel Samples from Contractor Storage Tank A section of the outlet filter was also sent to SwRI for additional analysis to confirm the composition of the contaminant.

8 Figure 14. UV-Vis Spectra for Fuel Samples from Contractor Storage Tank A section of the outlet filter was also sent to SwRI for additional analysis to confirm the composition of the contaminant. The section of outlet filter was placed in a beaker with a non-polar solvent. The beaker was put in an ultrasonic cleaner for 5 minutes to dislodge particles trapped in the filter. Those particles were collected on a Whatman GF-F, 0.7 micron pore size, glass-fiber filter. The material collected from the filter section was analyzed using SEM, FTIR, EDS, and XRD. The amount of material collected from the section of outlet filter was insufficient to obtain an accurate XRD result; the SEM and FTIR results are presented in Figures The SEM photomicrographs, Figures 15 and 16, are essentially equivalent to the results obtained by Donaldson. The larger particles visible in the figures are agglomerations of much smaller particles that are less than 1 micron in diameter. The film that covers much of the surface appears to be formed by particles that are only barely visible as individual, spherical particles in Figure 16. The FTIR spectrum in Figure 17 has a distinct carbonyl peak at 1736 wavenumbers. The peaks between 1400 and 1600 wavenumbers are weak, owing to the small sample size; but, they are indicative of the presence of carboxylate ions in the sample. The EDS spectrum in Figure 18 shows that the most abundant element in the contaminant was carbon. The silicon and oxygen are primarily associated with the glass in the glass fiber laboratory filter; however, some of the oxygen can be attributed to carboxyl groups in the contaminant. Sodium, calcium, and potassium in the sample are most likely present as the cations in the carboxylate salts. Iron is likely due to corrosion products. The other elements are present as very minor contaminants. The SwRI findings support the Donaldson findings that the material collected from the outlet filters of the contractors fuel storage tank contained carbonyl compounds. Donaldson was able to confirm the cleanliness of the subject fuel and fuel storage tanks. Donaldson was further able to demonstrate the lack of moisture in the fuel / tank. One possible explanation for these compounds is that the reactions to form them occurred prior to the fuel entering the storage tanks. At first the particles were too small to be removed in any large amount by the inlet filters. Once in the storage tanks, the smaller particles agglomerated to form particles that were now sufficiently large to cause plugging of the outlet filters. In essence, the contaminant formed in the contractors fuel storage tank although the necessary reactions occurred earlier in the distribution chain. An alternative explanation is that the particles formed inside the tank due to an interaction with the metal storage tank materials. 10th International Filtration Conference 157

9 Figure 15. Photomicrograph of Contaminant C Figure 16. Photomicrograph of Contaminant C 10th International Filtration Conference 158

10 Figure 17. FTIR Spectrum of Contaminant C Figure 18. EDS Spectrum of Contaminant C 10th International Filtration Conference 159

11 SUMMARY The three contaminants discussed in this paper were not the typical contaminants found in diesel fuel storage tanks. They required more detailed analysis in order to confirm their composition and source. They all represent a new challenge to fuel filtration systems since in each case the fuel was apparently free of contaminants at the time of final delivery. Contaminant A was a precipitate formed at temperatures above the cloud point of the biodiesel fuel blend. Both fuels used in the blend met the prevailing specifications. [Since this incident, the biodiesel specification, D6751, has been revised to address this situation.] Even though the dispensing filters were kept at 15 C, the contaminant did not redissolve in the fuel. Contaminant B was carboxylate ions that remained sufficiently small that they only plugged very tight porosity (2 microns or smaller) fuel filters. This meant that the ions could potentially travel all the way to the engine and possibly cause deposits in the engine fuel system. Contaminant C may have started to form prior to introduction to the fuel storage tank. They passed through filters meant to exclude contaminants from the tank, only to agglomerate to larger particles during storage. This, in turn, caused significant filter plugging when the fuel was dispensed from the storage tank. Contaminant C may also have formed in the fuel storage tanks. Like Contaminant B, it is also reasonable to assume that some of the Contaminant C particles passed through the outlet filters and made their way downstream. ACKNOWLEDGEMENTS The authors wish to thank Southwest Research Institute and Donaldson Company for their respective support of the work presented here. As previously unencountered contaminants appear in diesel fuel distribution and storage systems, the ability to obtain and maintain clean fuel will become increasingly difficult. Fuel suppliers, filter manufacturers, and fuel users will be required to work even more closely than before in order to meet these challenges. 10th International Filtration Conference 160

Article: The Formation & Testing of Sludge in Bunker Fuels By Dr Sunil Kumar Laboratory Manager VPS Fujairah 15th January 2018

Article: The Formation & Testing of Sludge in Bunker Fuels By Dr Sunil Kumar Laboratory Manager VPS Fujairah 15th January 2018 Introduction Sludge formation in bunker fuel is the source of major operational