- Email: [email protected]
Saturated monoglyceride effects on low-temperature performance of biodiesel blends
Fuel Processing Technology 118 (2014) 302–309
Contents lists available at ScienceDirect
Fuel Processing Technology journal homepage: www.elsevier.com/locate/fuproc
Saturated monoglyceride effects on low-temperature performance of biodiesel blends☆ G.M. Chupka a,⁎, L. Fouts a, J.A. Lennon b, T.L. Alleman a, D.A. Daniels b, R.L. McCormick a,⁎ a b
National Renewable Energy Laboratory, 15013 Denver West Parkway/MS1634, Golden, CO 80401, USA Innospec Fuel Specialties LLC, 200 Executive Drive, Newark, DE 19702, USA
a r t i c l e
i n f o
Article history: Received 1 February 2013 Accepted 4 October 2013 Available online 1 November 2013 Keywords: Biodiesel blends Low-temperature Cloud point Saturated monoglyceride
a b s t r a c t The effect of saturated monoglyceride (SMG) content of four B100s on the cloud point (CP) of blends with four diesel fuels was examined. Detecting CP with a more sensitive light-scattering method allowed observation of an early (higher temperature) CP in blends containing approximately 0.01 wt.% to 0.03 wt.% SMG. Blend samples with SMG content in this range may be particularly prone to unexpected filter clogging above the measured CP. Results for a 140 blend sample matrix revealed that SMG content had a larger effect on CP than other blend properties. An increase of 0.01 wt.% SMG in a biodiesel blend increased CP by as much as 4 °C. At a constant SMG level, increasing biodiesel content lowered CP, as did increasing the diesel fuel aromatic content, by improving the solubility of SMG. This implies that lowering the SMG content of a B100 allows preparation of higher biodiesel content blends having the same or lower CP. Increasing the unsaturated monoglyceride-toSMG ratio by blending in monoolein lowered CP, presumably because monoolein inhibits nucleation of SMG. In most blends with SMG content above 0.01 wt.%, polymorphic phase transformation of crystallized SMG (converting from the metastable α-form to the less soluble, stable β-form) was observed. © 2013 The Authors. Published by Elsevier B.V. All rights reserved.
1. Introduction Problems with the low-temperature operability performance of biodiesel in blends with petroleum diesel are infrequent, but continue to limit the use of biodiesel in winter months. A troubling aspect of this problem is that in some cases precipitates above the blend cloud point (CP) have been detected and have led to plugging of fuel filters and subsequent vehicle stalling, as well as plugging of fuel dispenser filters [1]. CP is considered a conservative indicator of the lowtemperature performance of a fuel [2]. The Coordinating Research Council found that for B100 fuels with a high cold soak filtration time, low-temperature operability issues were observed at and above the CP for B20 blends [3]. Implementation of a cold soak filtration time limit in the B100 standard (ASTM D6751) led to a reduction in the frequency of low-temperature operability issues. However, these issues still occur, and in many cases are attributed to the precipitation of saturated monoglycerides (SMG), an impurity in B100.
☆ This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial-No Derivative Works License, which permits non-commercial use, distribution, and reproduction in any medium, provided the original author and source are credited. ⁎ Corresponding authors. Tel.: +1 303 275-4475. E-mail address: [email protected] (G.M. Chupka).
Van Gerpen and coworkers [4] found that SMG were a main component of a “creamy paste-like” material that was found on a plugged fuel filter from a vehicle operating on B20. It was also noted that diglycerides, especially saturated diglycerides, were also present on the plugged fuel filters. Selvidge and coworkers also found that a blend with 2.5 vol.% soy-derived B100 containing above about 0.07 wt.% SMG could block dispenser filters as temperatures dropped to −18 °C [5]. Poirier and coworkers concluded that SMG could be the cause of filter plugging during heavy-duty vehicle testing at low temperature [6], while Brewer showed that SMG were responsible for storage tank deposits and dispenser filter clogging [7]. Recently, we investigated the effect of SMG and other impurities on precipitate formation at cold temperatures in B100s [8,9]. The studies included measurement of CP (cooling the sample) and final melting temperature (FMT), measured by heating the sample. B100 samples were spiked with SMG, water, and steryl glucosides (SG) at various levels. The main component that affects both the CP and FMT was SMG. Water and SG did not show a significant effect in any of the B100. The polymorphic phase behavior [10] of SMG can explain their impact on the FMT measurements. In some samples, FMT was much higher than CP, indicating that the SMG were undergoing a solventmediated polymorphic phase transition [8,9]. The increase in CP and FMT was observed above a concentration level of about 0.2 wt.% to 0.3 wt.% SMG. The rate at which the samples were heated during the test also had an effect on the measured FMT. For some samples, polymorphic behavior of the SMG was only revealed at relatively slow
0378-3820/$ – see front matter © 2013 The Authors. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.fuproc.2013.10.002
G.M. Chupka et al. / Fuel Processing Technology 118 (2014) 302–309
(0.1 °C/min) heating rates for laboratory bench tests. Notably, a heating rate of 0.1 °C/min or 6 °C/h may be considered fast for the diurnal heating and cooling cycles experienced by a parked vehicle or in an aboveground storage tank. Investigating the polymorphic phases of SMG in B100 by differential scanning calorimetry revealed that a meta-stable α-gel is initially formed upon cooling the sample [9,11]. This metastable phase then converts to the most stable β phase, which is much less soluble and has a higher melting temperature. Measurement of the solubility of the stable β phase versus the meta-stable α phase in B100 shows that they may differ by as much as an order of magnitude at 5 °C. Polymorphs are also known to occur for hydrocarbon waxes that precipitate from diesel fuel [12]. However, the difference in solubility for the various forms is relatively small. Because biodiesel is typically blended with petroleum-derived diesel fuel, it is important to understand the low-temperature phase behavior in these blends. To this end, we measured CP and FMT when blending spiked and unspiked B100 into various diesel fuels to create various levels of SMG in the finished blends. The B100 were blended into the diesel fuel at three different blend levels; B5, B10, and B20. Four diesel fuels with a wide range of aromatic content and CP were employed. 2. Materials and methods 2.1. Fuels Diesel fuels used for blending included a Tier 2 certification diesel (Fuel A), a high-aromatic (37%) diesel (Fuel B), a local (Colorado) No. 2 diesel (Fuel C), and a low CP No. 1 diesel from Minnesota (Fuel D). Additionally, we prepared some blends with a Canadian fuel designated CP43 (Fuel E), and a second low-CP No. 1 diesel (Fuel F). All petroleum diesel fuels were ultra-low sulfur diesel meeting the requirements of ASTM D975 for their respective grades. The primary B100 fuels used for this study were the same as those used in our previous investigation [9]. They consisted of a canola (B100 I) containing 6.7 wt.% saturated fatty acid methyl esters (FAME), soy-derived (B100 II) 15 wt.% saturated FAME, and two mixed feedstock B100s (B100 III and B100 IV) with 24 wt.% and 36 wt.% saturated FAME, respectively. They all met the requirements of ASTM D6751 as received. In addition, five B100 that were obtained as part of a nationwide B100 quality survey were also utilized. Details of their characterization are reported elsewhere [13]. The B100 were manipulated to have different levels of impurities prior to blending, including SMG, monoolein (MO), deionized water, and SG. For monoglyceride spiking, a custom monoglyceride mix of MO, monopalmitin, and monostearin was made for each feedstock. Because MO was added with the SMG to preserve the unsaturated: saturated monoglyceride ratio, the high spike level samples for B100 II and IV failed the ASTM D6751 limit for total glycerin, as did the spiked B100 I sample. Monoglycerides were obtained from NuChek Prep and were 99% pure or greater. SG was obtained from Matreya LLC. The Table 1 B100s used for preparation of biodiesel blends. B100s used to prepare diesel fuel blends B100 I B100 II
B100 III
B100 IV
As-received (0.053 wt.% SMG) Spiked to contain 0.1 wt.% SMG As-received (SMG below detection) Spiked to contain 0.05 wt.% SMG Spiked to contain 0.145 wt.% SMG (at total glycerin specification limit) Spiked to contain 0.23 wt.% SMG, 1000 ppm water, 40 ppm SG As-received (0.045 wt.% SMG) Spiked to contain 0.23 wt.% SMG (at total glycerin specification limit) Spiked to contain 0.32 wt.% SMG, 1000 ppm water, 40 ppm SG As-received (0.152 wt.% SMG) Spiked to contain 0.33 wt.% SMG Spiked to contain 0.43 wt.% SMG, 1000 ppm water, 40 ppm SG
303
spiked B100 were heated and sonicated briefly until they were visually bright and clear prior to blending. Table 1 lists the B100 used to prepare biodiesel blends. Blends containing 5, 10, and 20 vol.% biodiesel were prepared by volume using a pipette resulting in 5 mL samples. This provided a total of 140 blend samples.
2.2. Fuel property tests As received, B100 and diesel fuels were characterized using ASTM methods as noted earlier. FMT was measured using the Phase Technology 70X Analyzer. FMT is defined as the temperature at which the light scattering signal returns to baseline during warming because all of the crystals have re-dissolved into solution. The experiment is normally conducted at 1.5 °C/min heating rate; however, in some cases slower heating rates were used as noted. The CP of diesel blends was measured using one or more CP methods, including D5771, D5773, and D7689. Comparison of the three instruments and methods used shows definite differences in their configurations, cooling rates, volumes utilized, and the manner in which they detect the formation of crystals. All three of the CP methods employed were automated and intended to produce results equivalent to the ASTM referee method, D2500, which is conducted manually with visual observation of crystal formation. D5771 bears the most physical resemblance to D2500. However, it functions by transmitting light towards the center of a bottom-mirrored test jar. An optical detector measures the intensity of the transmitted light in a 45-mL sample. Adequate levels of crystal formation reduce the intensity of the transmitted light, signaling that CP has been reached. This method uses a stepped cooling protocol rather than a constant (linear) cooling rate. D7689 also functions by measuring the loss of transmitted light. The light passes through a relatively thin cross-section of the 0.5-mL sample, parallel to the bottom of the container as the sample is cooled at a constant rate of 1.5 °C/min. Method D5773 uses light scattering to detect the presence of crystals and employs the Phase Technology 70X Analyzer [14]. A 150 μL sample is placed into a mirrored-bottom cup with a Peltier temperature control device. Light is then reflected off of the mirrored surface while the sample is cooled at 1.5 °C/min. When crystals are present, the light is scattered, registering a signal at the detector indicating the presence of solid particles. The detection method employed in D5773 is much more sensitive than the “loss of transmission” approach used in the other CP methods [15]. This method gives very close agreement with both the referee method, ASTM D2500 (relative bias of 0.03 °C), and the other methods employed here (D5771 and D7689). D5771 and D5773 have been shown to be equivalent to D2500 (the referee CP method) for petroleum-derived fuels, while D7689 is equivalent if corrected for a small bias. Thus, they should in theory give equivalent results. However, relative rates of cooling (heat transfer) can dramatically affect the size, shape, and number of crystals formed at a given temperature. Both D5773 and D7689 have significantly larger surface-area-to-volume ratios compared to D5771. Additionally, while the cooling rates of D7689 and D5773 are well defined, the cooling rate in D5771 is not defined. D5771 has a significant area of air space between the cooling jacket and the sample. Heat transfer across this air space will not be as efficient as in D5773 and D7689 because of the close proximity of the sample to the cooling surface(s) in these methods.
3. Results Table 2 lists the properties of the diesel fuels used for blending. Detailed B100 properties have been previously reported [9]; however, Table 3 contains information on the properties most relevant to this work. Tables containing all of the CP and FMT data for blends are included in the supplementary data.
304
G.M. Chupka et al. / Fuel Processing Technology 118 (2014) 302–309
Table 2 Diesel fuel properties. Diesel
Aromatic content (%)
CP (°C)
FMT (°C)
T90 (°C)
Method Fuel A Fuel B Fuel C Fuel D Fuel E Fuel F
D1319 9 37 25 15 18 21
D5773 −28.1 −27.2 −16.5 −45.7 −56.3 −45.9
In house −22.0 −22.9 −14.0 −42.5 −40.5 −40.2
D86 302 297 321 244 279 249
Light Scattering Signal
0.01 wt% SMG
Late CP
Distilled B100
Early CP
3.1. Cloud point measurement by light scattering For CP measured by ASTM D5773, we found that some biodiesel blends exhibit what appears to be two CP results: an early signal increase at higher temperature followed by a more obvious sharp increase later in the run. We refer to these features as early CP and late CP, respectively. Fig. 1 shows example D5773 light scattering plots depicting results for a sample with two CP results, as well as a sample that shows a single CP. The sample exhibiting two CP results was B100 II in Fuel C spiked to contain a total of 0.01 wt.% SMG in the B5 blend. The second sample was B100 II as received (residual SMG in this sample was negligible) in Fuel C, also as a B5 blend. The addition of just a small amount of SMG causes the appearance of the early CP. Depending on which point is chosen for the CP, there can be a large impact on the results. In this example, the difference between the early and late CP is almost 8 °C. We have not observed this effect for neat petroleum diesel. B5 and B20 blends of the quality survey B100 [13] with diesel E and diesel F were analyzed using all the three CP methods (D5771, D5773, and D7689). Examination of the D5773 results reveals that most samples exhibited the early and late CPs. Fig. 2 compares the results between the three methods when using D5773 with early CP and late CP. In some cases, D5773 with early CP results are significantly higher than those measured with either D7689 or D5771. The D5773 late CP results are much more in line with results from the other methods, and by extension, with the referee method, D2500. All three methods used here agree closely with D2500 when testing standard petroleumderived diesel. Distilled B100 spiked with SMG also show D5773 CPs closer to those measured with D5771 and D7689 when the late CP is used. However, once the SMG level is between about 0.01 wt.% and 0.03 wt.% in the blend, the early CP becomes significant enough that it may also be detected with D7689. The amount of SMG present in the sample can be indicative of whether or not an early CP is observed when using D5773. For blend samples prepared with a distilled B100 (B100 II and III, as received), an early CP is not seen in any of the four diesel samples (Fuels A–D) at the B5, B10, or B20 level. For cases where there is approximately 0.01 wt.% SMG or greater in the blend, an early CP is noted. As shown in Fig. 3, as the SMG level is increased, the size of the early CP signal also increases. Once the amount of SMG is above about 0.03 wt.% in the blend, the early CP becomes dominant enough that this is where CP is determined by D5773 when using the automated method. Because this early CP may be missed by other methods, samples with SMG
-20
-18
-16
-14
-12
-10
-8
-6
-4
-2
0
Temperature, oC Fig. 1. Light scattering plots for B5 blends prepared from Fuel C with B100 II containing 0.01% SMG in the blend and B100 II as received (distilled B100).
content in this range may be particularly prone to unexpected filter clogging above the measured CP. 3.2. Effect of SMG on biodiesel blend cloud point The 140 blends prepared from both unspiked and spiked B100 I through IV were tested for CP using the D5773 method. CP is defined as “the temperature of a liquid specimen when the smallest observable cluster of hydrocarbon crystals first occurs upon cooling under prescribed conditions” [16]. For biodiesel blends, we take this to mean not just hydrocarbons, but saturated FAME, SMG, or any other crystallizing species. We have therefore used the early CP in cases where there was an early and a late CP. This affected the results for about 15 out of the 140 samples. Preliminary analysis showed no effect of the water or SG content of B100 on CP for the subset of samples containing these impurities. Linear regression analysis was used to evaluate several factors affecting the blend CP. The variables analyzed and the hypothesized effects were: • Base diesel CP: defines the temperature where petroleum wax will crystallize. • Percent aromatic content and percent biodiesel in the blend: both the presence of aromatics and biodiesel may increase the solubility of relatively polar materials such as SMG. • Percent saturated FAME and percent SMG in the blend: these species have the lowest solubility at low temperature and, along with petroleum wax, are responsible for crystal formation at CP. Table 4 lists the coefficient and the p-value for each factor analyzed. All five factors considered were significant at 95% confidence or greater. The R2 for the multiple regression analysis was 0.72, which indicates that the model captures about 70% of the variation in CP for these samples. Repeating the analysis using only late CP values did not change
Table 3 B100 properties as received. B100
Saturated FAME (%)
CP (°C)
FMT (°C)
CSFT (s)
SMG (wt.%)
Total MG (wt.%)
UMG/SMG
Total glycerin (wt.%)
Method B100 I B100 II B100 III B100 IV
GC 6.7 15 24 36
D5773 −1.2 −1.4 4.4 9.5
In house 6.8 5.6 8.4 13.1
D7501 110 91 100 78
D6584 0.053 bLOQa 0.045 0.152
D6584 0.557 bLOQa 0.148 0.393
D6584 9.5 4.9b 2.3 1.6
D6584 0.170 0.005 0.041 0.120
CSFT = Cold Soak Filtration Time GC = gas chromatograph MG = monoglycerides. a Limit of Quantitation, 0.020 wt.%. b Typical of soy-derived biodiesel.
G.M. Chupka et al. / Fuel Processing Technology 118 (2014) 302–309
10
D5771 or D7689 CP, oC
Table 4 Linear regression data for factors affecting blend CP.
B5 D7689 B5 D5771 B20 D5771 B20 D7689
0
-10
-20
-40
-40
-30
-20
-10
0
10
D5773 Early CP, oC 0
B5 D7689 B5 D5771 B20 D5771 B20 D7689
D5771 or D7689 CP, oC
-10
Variable
Coefficients
p-Value
Variable range
ΔCP for 10% of range change
Diesel CP Aromatic content % Biodiesel % SMG % Sat FAME
0.15 −0.13 −0.54 404 1.4
b0.001 0.026 b0.001 b0.001 0.003
−45 °C to −16.5 °C 7% to 35% 0% to 20% 0% to 0.085% 0% to 7.2%
0.3 °C −0.3 °C −1 °C 3.6 °C 0.7 °C
SMG content had a larger effect on CP than any of the other variables. Increasing blend SMG by just 0.01 wt.% would increase CP by over 4 °C on average. Fig. 4 shows the effect of increasing amounts of SMG in B100 II in Fuel C and Fuel D (the highest and lowest CP base fuels, respectively). Clearly, SMG content of a biodiesel blend can have a significant impact on CP. This trend is similar for the other fuels tested and points to SMG being the most significant contributor to the increase in CP typically observed for biodiesel blends. Fig. 5 examines the same data, and additional data for the other B100, plotted as a function of SMG content in the blend. For all fuels, SMG content has a dramatic effect on CP, however, there is also a clear effect of biodiesel blend level. Except for B100 I, CP is very similar for SMG content up to 0.01 wt.% to 0.015 wt.% regardless of biodiesel blend level or saturated FAME content. Above this level, the solvency effect of biodiesel is evident, with higher biodiesel content fuels exhibiting lower CP. For Fuel C blends, there also appears to be an effect of biodiesel saturated FAME content at the B20 level. While this effect is also present in the other fuels, it is less distinct. Fig. 5 also shows that the CP effect of a 0.01 wt.% change in SMG can be much larger than 4 °C at low SMG levels. Several of the B100 samples used to make blends at
-30
-50 -50
305
-20
-30
0
Fuel C
-40
-50 -50
-40
-30
-20
-10
0
D5773 Late CP, oC Fig. 2. Comparison of D5773 early CP and late CP results with results from methods D5771 and D7689.
Cloud Point, oC
-5
-10
-15 Distilled 0.048% SMG 0.148% SMG 0.229% SMG
-20
R2, but produced pb0.05 for only three out of the five parameters (diesel CP, percent biodiesel, and percent SMG); coefficient values were not significantly different.
-25 0
5
10
15
20
25
Percent Biodiesel
Light Scattering Signal
0
Fuel D
-10
Cloud Point, oC
0.0304 wt% SMG
0.0152 wt% SMG
-20
-30 Distilled 0.048% SMG 0.148% SMG 0.229% SMG
-40
0.0076 wt% SMG
-50 -25
-20
-15
-10
-5
0
5
Temperature, oC Fig. 3. Effect of SMG content on the early CP for B100 IV blended with Fuel A.
0
5
10
15
20
25
Percent Biodiesel Fig. 4. Effect of SMG content in B100 II on CP in Fuels C and D (SMG is percent in B100).
306
G.M. Chupka et al. / Fuel Processing Technology 118 (2014) 302–309
10
20 Fuel A
Fuel B
5 10
Cloud Point, oC
Cloud Point, oC
0 0 B5 B100 I B10 B100 I B20 B100 I B5 B100 II B10 B100 II B20 B100 II B5 B100 III B10 B100 III B20 B100 III B5 B100 IV B10 B100 IV B20 B100 IV
15% Saturated FAME
-10
6.7% Saturated FAME
-20
0.02
0.04
0.06
0.08
-10
B5 B100 I B10 B100 I B20 B100 I B5 B100 II B10 B100 II B20 B100 II B5 B100 III B10 B100 III B20 B100 III B5 B100 IV B10 B100 IV B20 B100 IV
15% Saturated FAME
-15 -20
6.7% Saturated FAME
-25
-30 0.00
-5
-30
0.10
0.00
0.02
Percent SMG
0.04
0.06
0.08
0.10
Percent SMG 20
15 Fuel C
10
Fuel D
10
5
Cloud Point, oC
Cloud Point, oC
36% Saturated FAME 24% Saturated FAME
0
B5 B100 I B10 B100 I B20 B100 I B5 B100 II B10 B100 II B20 B100 II B5 B100 III B10 B100 III B20 B100 III B5 B100 IV B10 B100 IV B20 B100 IV
-5 15% Saturated FAME
-10
6.7% Saturated FAME
-15
0.02
0.04
0.06
0.08
15% Saturated FAME B5 B100 I B10 B100 I B20 B100 I B5 B100 II B10 B100 II B20 B100 II B5 B100 III B10 B100 III B20 B100 III B5 B100 IV B10 B100 IV B20 B100 IV
-10 6.7% Saturated FAME
-20 -30
-20 0.00
0
0.10
-40 0.00
0.02
Percent SMG
0.04
0.06
0.08
0.10
Percent SMG
Fig. 5. Effect of blend SMG content on CP for B100 blended into Fuels A–D.
the highest SMG levels failed the total glycerin limit in ASTM D6751. Removal of these data points does not alter these trends. The slightly different results for B100 I may be caused by its low saturated FAME content, or by other factors as discussed in Section 3.3. Because B100 II as received does not contain SMG at detectable levels, its blends can be used to more clearly demonstrate the effect of saturated FAME content on CP. Fig. 6 shows CP for this B100 blended into all four base diesel fuels at the B5, B10, and B20 levels. In the highest CP base fuel (Fuel C), there is little effect of blending this distilled biodiesel on CP. Apparently, petroleum wax is the first species to precipitate at temperatures greater than or equal to −16.5 °C (the CP of the base diesel fuel) in this fuel for blends with saturated FAME
content up to 3 wt.%. For the lower CP diesel fuels, saturated FAME does increase CP, likely because of lower solubility at the lower temperatures of these CP experiments. Fuels A and B have almost the same CP, but are respectively the lowest and highest aromatic content fuels examined. A comparison of the CP of B5 blends in these fuels as a function of SMG content is shown in Fig. 7. In B5 blends, the effects of other factors such as saturated FAME content and biodiesel content should be minimized. For SMG content above about 0.005 wt.%, blends with the higher aromatic content diesel have a consistently lower CP, indicating that
10 -10 5 -15
Cloud Point, oC
0
Cloud Point, oC
-20 -25 -30 -35
-5 -10 -15 -20
Fuel A Fuel B Fuel C Fuel D
-40 -45
9.4% Aromatic (Fuel A) 35% Aromatic (Fuel B)
-25 -30 0.000
-50 0.0
0.5
1.0
1.5
2.0
2.5
3.0
0.005
0.010
0.015
0.020
0.025
Percent SMG
Percent Saturated FAME Fig. 6. Effect of percent saturated FAME (B100 II, as received) on the CP in diesel fuels A–D.
Fig. 7. Effect of diesel fuel percent aromatic content on CP for B5 blends as a function of SMG content.
G.M. Chupka et al. / Fuel Processing Technology 118 (2014) 302–309
aromatics can have a solvency effect on SMG and lower CP, consistent with the regression analysis model. 3.3. Effect of unsaturated monoglyceride on cloud point B100 I is a canola-derived B100 and contains a significantly higher ratio of unsaturated:saturated monoglyceride (UMG/SMG) than the other B100 used in this study (see Table 3). The CP of blends for B100 I were generally lower than expected based on trends shown in Fig. 5 for the other three B100 blends. We have observed that addition of MO, an unsaturated monoglyceride, can in some cases reduce the CP of a B100 blend. For example, in neat B100 I, the measured CP of a sample containing 0.28 wt.% SMG was 9.4 °C. Addition of 1.8 wt.% MO, for a total of 2.3 wt.% UMG, to the same sample lowered the CP to 2.9 °C. Fig. 8 shows the effect of SMG content on CP for blends with B100 I having UMG/SMG = 9.5 and 5.2. Data for B100 II with UMG/SMG = 4.9 are also included. Based on the trends shown in Fig. 5 for B100 II, III, and IV, the CP versus SMG curves for B100 I blends would be expected to be similar to those observed for B100 II, especially at low SMG and biodiesel content. However, CP for B100 I blends with UMG/SMG = 9.5 falls well below the B100 II curves for SMG greater than about 0.005 wt.%. Note that this UMG/SMG ratio is typical of canola-derived biodiesel. When B100 I was spiked with SMG only, such that UMG/SMG = 5.2, the measured CPs were very close to the CP curves observed for B100 II blends having UMG/SMG = 4.9. When additional MO was added to the B100 II/Fuel D blends to increase UMG/SMG up to 9.5, the CP decreased to the same level observed for the B100 I blends at this UMG/SMG ratio. Clearly, the presence of UMG, if the UMG/SMG ratio is high enough, can depress the CP of B100 and blends. 5
Fuel A
Cloud Point, oC
0 -5 -10 -15
B100 I B5 High U/S B100 I B10 High U/S B100 I B20 High U/S B100 I B5 Low U/S B100 I B10 Low U/S B100 I B20 Low U/S B100 II B5 B100 II B10 B100 II B20
-20 -25 -30 0.00
0.01
0.02
0.03
0.04
0.05
307
UMG could possibly interact with SMG to prevent or slow nucleation, similar to compounds such as water and stearic acid [17]. There are numerous industrial crystallization applications where additives are used to prevent or enhance the rate of nucleation and crystal growth [18]. It may be possible that MO, with polarity similar to SMG but significantly greater solubility at cold temperatures, could be interacting with the polar head groups of the SMG to inhibit nucleation and reduce CP on the time scale of the CP experiment. The most effective additives for inhibiting nucleation generally have similar structure to the nucleating species [19,20]. 3.4. SMG polymorphic phase transformation in biodiesel blends In previous work [8,9], we have shown that in B100, SMG precipitates as a metastable gel that can transform into the stable β polymorph over time or upon heating via solvent-mediated polymorphic phase transformation [21]. We proposed that this transformation could account for many of the unanticipated occurrences of fuel filter clogging by precipitate formation above the CP in biodiesel blends. The polymorphic transformation can manifest itself as a large signal increase during warming of the samples at 1.5 °C/min in light scattering plots, and the FMT is significantly higher than the CP. This signal increase was correlated to the polymorphic conversion of the SMG in microscope experiments, confirmed by X-ray powder diffraction, and differential scanning calorimetry. In many cases, however, the 1.5 °C/min heating rate is too fast to observe the transformation, and it is only revealed by lowering the heating rate to 0.5 °C/min or lower [9]. The polymorphic phase conversion can also occur in biodiesel blends. Fig. 9 shows two examples of light scattering plots for the FMT experiment for B100 II in Fuel C. For distilled B100 II blended at 5%, no polymorphism is noted and the signal returns rapidly to the baseline. For a B10 blend spiked to contain 0.023 wt.% SMG, there is a large signal increase and FMT increased by about 20 °C. Fig. 10 shows FMT and CP as a function of SMG content for blends of all four B100 in Fuel D. Plots for Fuels A, B, and C are similar and are provided in the Supplementary data. In all cases for B100 I, and above about 0.01 wt.% SMG for B100 II and III, the FMT-CP difference increases to 20°C or higher, indicating the transformation of α-SMG into the most stable and lower solubility β form. At lower concentrations, apparently all of the SMG dissolves before undergoing transformation at this heating rate. B100 I may exhibit transformation at lower concentration because of the CP suppression caused by the high UMG/SMG ratio, which lowers the starting temperature for the FMT experiment. This increases the time available for transformation in this experiment and may imply that transformation could be observed at concentrations below 0.01 wt.% SMG for B100 II and III at slower heating rates.
Percent SMG 0
B5 Distilled B100 B10 0.023 wt% SMG
Cloud Point, oC
-10 -15 B100 I B5 High U/S B100 I B10 High U/S B100 I B20 High U/S B100 I B5 Low U/S B100 I B10 Low U/S B100 I B20 Low U/S B100 II B5 B100 II B10 B100 II B20 B100 II B5 High U/S B100 II B10 High U/S B100 II B20 High U/S
-20 -25 -30 -35
Light Scattering Signal
Fuel D
-5
-40 0.00
0.01
0.02
0.03
0.04
0.05
Percent SMG Fig. 8. Effect of SMG content and UMG/SMG (U/S in figure) ratio on CP. High U/S = 9.5, low U/S = 4.9 or 5.2.
-20
-15
-10
-5
0
5
10
15
20
Temperature, oC Fig. 9. FMT experiment light scattering plots of B100 II blended into Fuel C demonstrating polymorphic phase conversion when SMG is present.
308
G.M. Chupka et al. / Fuel Processing Technology 118 (2014) 302–309
20
30
10 0 -10 -20 -30 -40
B100 II/Fuel D
Cloud Point or Final Melting Temperature, oC
Cloud Point or Final Melting Temperature, oC
B100 I/Fuel D
CP FMT at 1.5oC/min
20 10 0 -10 -20 -30 -40
CP FMT at 1.5oC/min
-50
-50 0.00
0.01
0.02
0.03
0.04
0.05
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
Percent SMG
Percent SMG 40
B100 III/Fuel D
20 10 0 -10 -20 -30 -40
CP FMT at 1.5oC/min
B100 IV/Fuel D
Cloud Point or Final Melting Temperature, oC
Cloud Point or Final Melting Temperature, oC
30
30 20 10 0 -10 -20 CP FMT at 1.5oC/min FMT at 0.3oC/min
-30 -40 -50
-50 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
Percent SMG
Percent SMG Fig. 10. CP and FMT as a function of SMG content for B100 in Fuel D.
For B100 IV, transformation is not observed at the 1.5°C/min heating rate. However, lowering the heating rate to 0.3 °C/min reveals a significant increase in FMT. The composition of B100 IV is apparently such that the phase transformation is slower and requires a slower heating rate to be revealed. Because real-world vehicle or storage tank heating rates could potentially be in the 1 °C to 2 °C/h range (0.015 °C to 0.03 °C/min), polymorphic phase transformation of the SMG could cause unanticipated fuel filter clogging for blends of all of these B100 samples; it may not be predictable by higher heating rate laboratory tests. B100 IV as received contained much higher levels of impurities than did B100 II or III. B100 II was produced by vacuum distillation and contains extremely low impurity levels. B100 III was also distilled or partially distilled and also had very low impurities. Thus, B100 IV may contain many impurities common to animal fats such as sterols, sterol oxides, fatty acids, or fatty acid oxidation products, or residual to the biodiesel production process (soaps and glycerol, in addition to monoand diglycerides) that are not present in B100 II or III. It is well established that impurities or additives can alter the rate of polymorphic phase transformation [18,22] and possibly, such an effect is active here. Because of the very low UMG/SMG ratio in B100 IV, we tested the hypothesis that adding MO to increase this ratio could increase the rate of phase transformation as evidenced by increasing FMT. The presence of monoglycerides has been shown to increase the rate of polymorphic phase transformation for palm oil crystallization [19]. However, adding MO was not successful at increasing FMT at the 1.5 °C/min heating rate. 4. Conclusions This study has shown two potential causes of precipitate formation above CP. For biodiesel blends containing approximately 0.01 wt.% to
0.03 wt.% SMG some CP measurement methods may not be adequately sensitive to detect the temperature where crystals first occur upon cooling. Samples with SMG content in this range may be particularly prone to unexpected filter clogging above the measured CP. Second, blends containing SMG may experience solvent mediated polymorphic phase transformation of the SMG, where the metastable α-form crystallizes initially but converts to the less soluble, stable β-form upon heating. Additionally, our results show that SMG content is the single largest factor affecting the CP of biodiesel blends. While the presence of saturated FAME will also lead to an increased CP, the SMG effect occurs at concentrations well below 0.1 wt.%. A significant effect due to increasing saturated FAME content requires an increase of more than 5 wt.%. At a constant SMG level increasing biodiesel content lowered CP, as did increasing the diesel fuel aromatic content, by improving the solubility of SMG and saturated FAME. This implies that by choosing a B100 containing a lower level of SMG a biodiesel blender can increase biodiesel blend level while maintaining or even reducing the CP of the blended product. In a surprising observation, increasing the UMG/SMG ratio could significantly reduce the CP of blends, presumably by inhibiting nucleation of the SMG. While these conclusions are based on the use of the D5773 early CP, it is important to note that this affected the CP result for only 15 out of the 140 samples. Use of the late CP for these samples in the data analysis did not significantly change the results. The results indicate that cold weather operability problems might be avoided by measurement of CP with the most sensitive methods available, by careful control of the blend SMG content, and potentially by engineering biodiesel to have a higher UMG/SMG content than would normally occur. The degree of FAME saturation is a secondary factor.
G.M. Chupka et al. / Fuel Processing Technology 118 (2014) 302–309
Acknowledgments This work was supported by the U.S. Department of Energy, Office of Vehicle Technologies, Fuels and Lubricants Technologies Program under Contract no. DEAC36-99GO10337 with the National Renewable Energy Laboratory. Funding was also provided by the National Biodiesel Board. The authors wish to thank Dr. Gordon Chiu of Phase Technology for comments on the manuscript.
Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.fuproc.2013.10.002.
References [1] R.O. Dunn, Effects of minor constituents on cold flow properties and performance of biodiesel, Progress in Energy and Combustion Science 35 (2009) 481–489. [2] J.E. Chandler, Comparison of all weather chassis dynamometer low temperature operability limits for heavy and light duty trucks with standard laboratory test methods, SAE Technical Paper No. 962197, 1996. [3] R.L. McCormick, J. Chandler, R. Gault, S. Howell, M. Nikanjam, A. Buczynsky, J. Jetter, R. Barenescu, D. Arters, D. Forester, K. Mitchell, Biodiesel blend low temperature performance validation, CRC Project: DP-2a-07June 2008. (http://www.crcao.org). [4] J.H. Van Gerpen, E.G. Hammond, L. Yu, A. Monyem, Determining the influence of contaminants on biodiesel properties, SAE Technical Paper No. 971685, 1997. [5] C. Selvidge, S. Blumenshine, K. Campbell, C. Dowell, J. Stolis, Effect of biodiesel impurities on filterability and phase separation from biodiesel and biodiesel blends, Proc. of the 10th International Conference on Stability, Handling, and Use of Liquid Fuels, Oct 7–11, 2007; Tucson, AZ, 2007, (http://iash.net/Archives/Archives_Proceed_ 2007.html). [6] M.-A. Poirier, P. Lai, L. Lawlor, The effect of biodiesel fuels on the low temperature operability of North American heavy duty diesel trucks, SAE Technical Paper No. 2008-01-2380, 2008.
309
[7] M. Brewer, Understanding low temperature biodiesel issues. Impact of impurities. Case studies of recent industry incidents in EU, 2nd International Congress on Biodiesel, Munich, November 17–19, 2009, 2009. [8] G.M. Chupka, J. Yanowitz, G. Chiu, T.L. Alleman, R.L. McCormick, Effect of saturated monoglyceride polymorphism on low-temperature performance of biodiesel, Energy & Fuels 25 (2011) 398–405. [9] G.M. Chupka, L. Fouts, R.L. McCormick, Effect of low-level impurities on lowtemperature performance properties of biodiesel, Energy & Environmental Science 5 (2012) 8734–8742. [10] E.S. Lutton, The phases of saturated 1-monoglycerides C14–C22, Journal of the American Oil Chemists' Society 48 (1971) 778–781. [11] W.G. Morley, G.J.T. Tiddy, Phase behaviour of monoglyceride/water systems. J.C.S, Faraday Transactions 89 (1993) 2323–2331. [12] E.B. Sirota, H.E. King Jr., D.M. Singer, H.H. Shao, Rotator phases of the normal alkanes: an X-ray scattering study, Journal of Chemical Physics 98 (1993) 5809–5824. [13] T.L. Alleman, L. Fouts, G.M. Chupka, Quality parameters and chemical analysis for biodiesel produced in the United States in 2011, NREL/TP-5400-57662, 2013. (http://www.nrel.gov/docs/fy13osti/57662.pdf). [14] ASTM Standard D5773-10, Standard Test Method for Cloud Point of Petroleum Products (Constant Cooling Rate Method), Section 13.4.3, ASTM International, West Conshohocken, PA, 2010. [15] A.J. Wright, S.S. Narine, A.G. Marangoni, Comparison of experimental techniques used in lipid crystallization studies, Journal of the American Oil Chemists' Society 77 (2000) 1239–1242. [16] ASTM Standard D2500-11, Standard Test Method for Cloud Point of Petroleum Products, ASTM International, West Conshohocken, PA, 2011. [17] C.H. Chen, E.M. Terentjev, Aging and metastability of monoglycerides in hydrophobic solutions, Langmuir 25 (2009) 6717–6724. [18] K.W. Smith, K. Ghaggan, G. Talbot, K.F. van Malssen, Crystallization of fats: influence of minor components and additives, Journal of the American Oil Chemists' Society 88 (2011) 1085–1101. [19] E. Fredrick, I. Foubert, J. Van De Sype, K. Dewettinck, Influence of monoglycerides on the crystallization behavior of palm oil, Crystal Growth and Design 8 (2008) 1833–1839. [20] B.A. Hendrikson, D.J.W. Grant, The effect of structurally related substances on the nucleation kinetics of paracetamol (acetaminophen), Journal of Crystal Growth 156 (1995) 252–260. [21] D. Croker, B.K. Hodnett, Mechanistic features of polymorphic transformations: the role of surfaces, Crystal Growth and Design 10 (2010) 2806–2816. [22] C.-H. Gu, K. Chatterjee, V. Young Jr., D.J.W. Grant, Stabilization of a metastable polymorph of sulfamerazine by structurally related additives, Journal of Crystal Growth 235 (2002) 471–481.
Contents lists available at ScienceDirect
Fuel Processing Technology journal homepage: www.elsevier.com/locate/fuproc
Saturated monoglyceride effects on low-temperature performance of biodiesel blends☆ G.M. Chupka a,⁎, L. Fouts a, J.A. Lennon b, T.L. Alleman a, D.A. Daniels b, R.L. McCormick a,⁎ a b
National Renewable Energy Laboratory, 15013 Denver West Parkway/MS1634, Golden, CO 80401, USA Innospec Fuel Specialties LLC, 200 Executive Drive, Newark, DE 19702, USA
a r t i c l e
i n f o
Article history: Received 1 February 2013 Accepted 4 October 2013 Available online 1 November 2013 Keywords: Biodiesel blends Low-temperature Cloud point Saturated monoglyceride
a b s t r a c t The effect of saturated monoglyceride (SMG) content of four B100s on the cloud point (CP) of blends with four diesel fuels was examined. Detecting CP with a more sensitive light-scattering method allowed observation of an early (higher temperature) CP in blends containing approximately 0.01 wt.% to 0.03 wt.% SMG. Blend samples with SMG content in this range may be particularly prone to unexpected filter clogging above the measured CP. Results for a 140 blend sample matrix revealed that SMG content had a larger effect on CP than other blend properties. An increase of 0.01 wt.% SMG in a biodiesel blend increased CP by as much as 4 °C. At a constant SMG level, increasing biodiesel content lowered CP, as did increasing the diesel fuel aromatic content, by improving the solubility of SMG. This implies that lowering the SMG content of a B100 allows preparation of higher biodiesel content blends having the same or lower CP. Increasing the unsaturated monoglyceride-toSMG ratio by blending in monoolein lowered CP, presumably because monoolein inhibits nucleation of SMG. In most blends with SMG content above 0.01 wt.%, polymorphic phase transformation of crystallized SMG (converting from the metastable α-form to the less soluble, stable β-form) was observed. © 2013 The Authors. Published by Elsevier B.V. All rights reserved.
1. Introduction Problems with the low-temperature operability performance of biodiesel in blends with petroleum diesel are infrequent, but continue to limit the use of biodiesel in winter months. A troubling aspect of this problem is that in some cases precipitates above the blend cloud point (CP) have been detected and have led to plugging of fuel filters and subsequent vehicle stalling, as well as plugging of fuel dispenser filters [1]. CP is considered a conservative indicator of the lowtemperature performance of a fuel [2]. The Coordinating Research Council found that for B100 fuels with a high cold soak filtration time, low-temperature operability issues were observed at and above the CP for B20 blends [3]. Implementation of a cold soak filtration time limit in the B100 standard (ASTM D6751) led to a reduction in the frequency of low-temperature operability issues. However, these issues still occur, and in many cases are attributed to the precipitation of saturated monoglycerides (SMG), an impurity in B100.
☆ This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial-No Derivative Works License, which permits non-commercial use, distribution, and reproduction in any medium, provided the original author and source are credited. ⁎ Corresponding authors. Tel.: +1 303 275-4475. E-mail address: [email protected] (G.M. Chupka).
Van Gerpen and coworkers [4] found that SMG were a main component of a “creamy paste-like” material that was found on a plugged fuel filter from a vehicle operating on B20. It was also noted that diglycerides, especially saturated diglycerides, were also present on the plugged fuel filters. Selvidge and coworkers also found that a blend with 2.5 vol.% soy-derived B100 containing above about 0.07 wt.% SMG could block dispenser filters as temperatures dropped to −18 °C [5]. Poirier and coworkers concluded that SMG could be the cause of filter plugging during heavy-duty vehicle testing at low temperature [6], while Brewer showed that SMG were responsible for storage tank deposits and dispenser filter clogging [7]. Recently, we investigated the effect of SMG and other impurities on precipitate formation at cold temperatures in B100s [8,9]. The studies included measurement of CP (cooling the sample) and final melting temperature (FMT), measured by heating the sample. B100 samples were spiked with SMG, water, and steryl glucosides (SG) at various levels. The main component that affects both the CP and FMT was SMG. Water and SG did not show a significant effect in any of the B100. The polymorphic phase behavior [10] of SMG can explain their impact on the FMT measurements. In some samples, FMT was much higher than CP, indicating that the SMG were undergoing a solventmediated polymorphic phase transition [8,9]. The increase in CP and FMT was observed above a concentration level of about 0.2 wt.% to 0.3 wt.% SMG. The rate at which the samples were heated during the test also had an effect on the measured FMT. For some samples, polymorphic behavior of the SMG was only revealed at relatively slow
0378-3820/$ – see front matter © 2013 The Authors. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.fuproc.2013.10.002
G.M. Chupka et al. / Fuel Processing Technology 118 (2014) 302–309
(0.1 °C/min) heating rates for laboratory bench tests. Notably, a heating rate of 0.1 °C/min or 6 °C/h may be considered fast for the diurnal heating and cooling cycles experienced by a parked vehicle or in an aboveground storage tank. Investigating the polymorphic phases of SMG in B100 by differential scanning calorimetry revealed that a meta-stable α-gel is initially formed upon cooling the sample [9,11]. This metastable phase then converts to the most stable β phase, which is much less soluble and has a higher melting temperature. Measurement of the solubility of the stable β phase versus the meta-stable α phase in B100 shows that they may differ by as much as an order of magnitude at 5 °C. Polymorphs are also known to occur for hydrocarbon waxes that precipitate from diesel fuel [12]. However, the difference in solubility for the various forms is relatively small. Because biodiesel is typically blended with petroleum-derived diesel fuel, it is important to understand the low-temperature phase behavior in these blends. To this end, we measured CP and FMT when blending spiked and unspiked B100 into various diesel fuels to create various levels of SMG in the finished blends. The B100 were blended into the diesel fuel at three different blend levels; B5, B10, and B20. Four diesel fuels with a wide range of aromatic content and CP were employed. 2. Materials and methods 2.1. Fuels Diesel fuels used for blending included a Tier 2 certification diesel (Fuel A), a high-aromatic (37%) diesel (Fuel B), a local (Colorado) No. 2 diesel (Fuel C), and a low CP No. 1 diesel from Minnesota (Fuel D). Additionally, we prepared some blends with a Canadian fuel designated CP43 (Fuel E), and a second low-CP No. 1 diesel (Fuel F). All petroleum diesel fuels were ultra-low sulfur diesel meeting the requirements of ASTM D975 for their respective grades. The primary B100 fuels used for this study were the same as those used in our previous investigation [9]. They consisted of a canola (B100 I) containing 6.7 wt.% saturated fatty acid methyl esters (FAME), soy-derived (B100 II) 15 wt.% saturated FAME, and two mixed feedstock B100s (B100 III and B100 IV) with 24 wt.% and 36 wt.% saturated FAME, respectively. They all met the requirements of ASTM D6751 as received. In addition, five B100 that were obtained as part of a nationwide B100 quality survey were also utilized. Details of their characterization are reported elsewhere [13]. The B100 were manipulated to have different levels of impurities prior to blending, including SMG, monoolein (MO), deionized water, and SG. For monoglyceride spiking, a custom monoglyceride mix of MO, monopalmitin, and monostearin was made for each feedstock. Because MO was added with the SMG to preserve the unsaturated: saturated monoglyceride ratio, the high spike level samples for B100 II and IV failed the ASTM D6751 limit for total glycerin, as did the spiked B100 I sample. Monoglycerides were obtained from NuChek Prep and were 99% pure or greater. SG was obtained from Matreya LLC. The Table 1 B100s used for preparation of biodiesel blends. B100s used to prepare diesel fuel blends B100 I B100 II
B100 III
B100 IV
As-received (0.053 wt.% SMG) Spiked to contain 0.1 wt.% SMG As-received (SMG below detection) Spiked to contain 0.05 wt.% SMG Spiked to contain 0.145 wt.% SMG (at total glycerin specification limit) Spiked to contain 0.23 wt.% SMG, 1000 ppm water, 40 ppm SG As-received (0.045 wt.% SMG) Spiked to contain 0.23 wt.% SMG (at total glycerin specification limit) Spiked to contain 0.32 wt.% SMG, 1000 ppm water, 40 ppm SG As-received (0.152 wt.% SMG) Spiked to contain 0.33 wt.% SMG Spiked to contain 0.43 wt.% SMG, 1000 ppm water, 40 ppm SG
303
spiked B100 were heated and sonicated briefly until they were visually bright and clear prior to blending. Table 1 lists the B100 used to prepare biodiesel blends. Blends containing 5, 10, and 20 vol.% biodiesel were prepared by volume using a pipette resulting in 5 mL samples. This provided a total of 140 blend samples.
2.2. Fuel property tests As received, B100 and diesel fuels were characterized using ASTM methods as noted earlier. FMT was measured using the Phase Technology 70X Analyzer. FMT is defined as the temperature at which the light scattering signal returns to baseline during warming because all of the crystals have re-dissolved into solution. The experiment is normally conducted at 1.5 °C/min heating rate; however, in some cases slower heating rates were used as noted. The CP of diesel blends was measured using one or more CP methods, including D5771, D5773, and D7689. Comparison of the three instruments and methods used shows definite differences in their configurations, cooling rates, volumes utilized, and the manner in which they detect the formation of crystals. All three of the CP methods employed were automated and intended to produce results equivalent to the ASTM referee method, D2500, which is conducted manually with visual observation of crystal formation. D5771 bears the most physical resemblance to D2500. However, it functions by transmitting light towards the center of a bottom-mirrored test jar. An optical detector measures the intensity of the transmitted light in a 45-mL sample. Adequate levels of crystal formation reduce the intensity of the transmitted light, signaling that CP has been reached. This method uses a stepped cooling protocol rather than a constant (linear) cooling rate. D7689 also functions by measuring the loss of transmitted light. The light passes through a relatively thin cross-section of the 0.5-mL sample, parallel to the bottom of the container as the sample is cooled at a constant rate of 1.5 °C/min. Method D5773 uses light scattering to detect the presence of crystals and employs the Phase Technology 70X Analyzer [14]. A 150 μL sample is placed into a mirrored-bottom cup with a Peltier temperature control device. Light is then reflected off of the mirrored surface while the sample is cooled at 1.5 °C/min. When crystals are present, the light is scattered, registering a signal at the detector indicating the presence of solid particles. The detection method employed in D5773 is much more sensitive than the “loss of transmission” approach used in the other CP methods [15]. This method gives very close agreement with both the referee method, ASTM D2500 (relative bias of 0.03 °C), and the other methods employed here (D5771 and D7689). D5771 and D5773 have been shown to be equivalent to D2500 (the referee CP method) for petroleum-derived fuels, while D7689 is equivalent if corrected for a small bias. Thus, they should in theory give equivalent results. However, relative rates of cooling (heat transfer) can dramatically affect the size, shape, and number of crystals formed at a given temperature. Both D5773 and D7689 have significantly larger surface-area-to-volume ratios compared to D5771. Additionally, while the cooling rates of D7689 and D5773 are well defined, the cooling rate in D5771 is not defined. D5771 has a significant area of air space between the cooling jacket and the sample. Heat transfer across this air space will not be as efficient as in D5773 and D7689 because of the close proximity of the sample to the cooling surface(s) in these methods.
3. Results Table 2 lists the properties of the diesel fuels used for blending. Detailed B100 properties have been previously reported [9]; however, Table 3 contains information on the properties most relevant to this work. Tables containing all of the CP and FMT data for blends are included in the supplementary data.
304
G.M. Chupka et al. / Fuel Processing Technology 118 (2014) 302–309
Table 2 Diesel fuel properties. Diesel
Aromatic content (%)
CP (°C)
FMT (°C)
T90 (°C)
Method Fuel A Fuel B Fuel C Fuel D Fuel E Fuel F
D1319 9 37 25 15 18 21
D5773 −28.1 −27.2 −16.5 −45.7 −56.3 −45.9
In house −22.0 −22.9 −14.0 −42.5 −40.5 −40.2
D86 302 297 321 244 279 249
Light Scattering Signal
0.01 wt% SMG
Late CP
Distilled B100
Early CP
3.1. Cloud point measurement by light scattering For CP measured by ASTM D5773, we found that some biodiesel blends exhibit what appears to be two CP results: an early signal increase at higher temperature followed by a more obvious sharp increase later in the run. We refer to these features as early CP and late CP, respectively. Fig. 1 shows example D5773 light scattering plots depicting results for a sample with two CP results, as well as a sample that shows a single CP. The sample exhibiting two CP results was B100 II in Fuel C spiked to contain a total of 0.01 wt.% SMG in the B5 blend. The second sample was B100 II as received (residual SMG in this sample was negligible) in Fuel C, also as a B5 blend. The addition of just a small amount of SMG causes the appearance of the early CP. Depending on which point is chosen for the CP, there can be a large impact on the results. In this example, the difference between the early and late CP is almost 8 °C. We have not observed this effect for neat petroleum diesel. B5 and B20 blends of the quality survey B100 [13] with diesel E and diesel F were analyzed using all the three CP methods (D5771, D5773, and D7689). Examination of the D5773 results reveals that most samples exhibited the early and late CPs. Fig. 2 compares the results between the three methods when using D5773 with early CP and late CP. In some cases, D5773 with early CP results are significantly higher than those measured with either D7689 or D5771. The D5773 late CP results are much more in line with results from the other methods, and by extension, with the referee method, D2500. All three methods used here agree closely with D2500 when testing standard petroleumderived diesel. Distilled B100 spiked with SMG also show D5773 CPs closer to those measured with D5771 and D7689 when the late CP is used. However, once the SMG level is between about 0.01 wt.% and 0.03 wt.% in the blend, the early CP becomes significant enough that it may also be detected with D7689. The amount of SMG present in the sample can be indicative of whether or not an early CP is observed when using D5773. For blend samples prepared with a distilled B100 (B100 II and III, as received), an early CP is not seen in any of the four diesel samples (Fuels A–D) at the B5, B10, or B20 level. For cases where there is approximately 0.01 wt.% SMG or greater in the blend, an early CP is noted. As shown in Fig. 3, as the SMG level is increased, the size of the early CP signal also increases. Once the amount of SMG is above about 0.03 wt.% in the blend, the early CP becomes dominant enough that this is where CP is determined by D5773 when using the automated method. Because this early CP may be missed by other methods, samples with SMG
-20
-18
-16
-14
-12
-10
-8
-6
-4
-2
0
Temperature, oC Fig. 1. Light scattering plots for B5 blends prepared from Fuel C with B100 II containing 0.01% SMG in the blend and B100 II as received (distilled B100).
content in this range may be particularly prone to unexpected filter clogging above the measured CP. 3.2. Effect of SMG on biodiesel blend cloud point The 140 blends prepared from both unspiked and spiked B100 I through IV were tested for CP using the D5773 method. CP is defined as “the temperature of a liquid specimen when the smallest observable cluster of hydrocarbon crystals first occurs upon cooling under prescribed conditions” [16]. For biodiesel blends, we take this to mean not just hydrocarbons, but saturated FAME, SMG, or any other crystallizing species. We have therefore used the early CP in cases where there was an early and a late CP. This affected the results for about 15 out of the 140 samples. Preliminary analysis showed no effect of the water or SG content of B100 on CP for the subset of samples containing these impurities. Linear regression analysis was used to evaluate several factors affecting the blend CP. The variables analyzed and the hypothesized effects were: • Base diesel CP: defines the temperature where petroleum wax will crystallize. • Percent aromatic content and percent biodiesel in the blend: both the presence of aromatics and biodiesel may increase the solubility of relatively polar materials such as SMG. • Percent saturated FAME and percent SMG in the blend: these species have the lowest solubility at low temperature and, along with petroleum wax, are responsible for crystal formation at CP. Table 4 lists the coefficient and the p-value for each factor analyzed. All five factors considered were significant at 95% confidence or greater. The R2 for the multiple regression analysis was 0.72, which indicates that the model captures about 70% of the variation in CP for these samples. Repeating the analysis using only late CP values did not change
Table 3 B100 properties as received. B100
Saturated FAME (%)
CP (°C)
FMT (°C)
CSFT (s)
SMG (wt.%)
Total MG (wt.%)
UMG/SMG
Total glycerin (wt.%)
Method B100 I B100 II B100 III B100 IV
GC 6.7 15 24 36
D5773 −1.2 −1.4 4.4 9.5
In house 6.8 5.6 8.4 13.1
D7501 110 91 100 78
D6584 0.053 bLOQa 0.045 0.152
D6584 0.557 bLOQa 0.148 0.393
D6584 9.5 4.9b 2.3 1.6
D6584 0.170 0.005 0.041 0.120
CSFT = Cold Soak Filtration Time GC = gas chromatograph MG = monoglycerides. a Limit of Quantitation, 0.020 wt.%. b Typical of soy-derived biodiesel.
G.M. Chupka et al. / Fuel Processing Technology 118 (2014) 302–309
10
D5771 or D7689 CP, oC
Table 4 Linear regression data for factors affecting blend CP.
B5 D7689 B5 D5771 B20 D5771 B20 D7689
0
-10
-20
-40
-40
-30
-20
-10
0
10
D5773 Early CP, oC 0
B5 D7689 B5 D5771 B20 D5771 B20 D7689
D5771 or D7689 CP, oC
-10
Variable
Coefficients
p-Value
Variable range
ΔCP for 10% of range change
Diesel CP Aromatic content % Biodiesel % SMG % Sat FAME
0.15 −0.13 −0.54 404 1.4
b0.001 0.026 b0.001 b0.001 0.003
−45 °C to −16.5 °C 7% to 35% 0% to 20% 0% to 0.085% 0% to 7.2%
0.3 °C −0.3 °C −1 °C 3.6 °C 0.7 °C
SMG content had a larger effect on CP than any of the other variables. Increasing blend SMG by just 0.01 wt.% would increase CP by over 4 °C on average. Fig. 4 shows the effect of increasing amounts of SMG in B100 II in Fuel C and Fuel D (the highest and lowest CP base fuels, respectively). Clearly, SMG content of a biodiesel blend can have a significant impact on CP. This trend is similar for the other fuels tested and points to SMG being the most significant contributor to the increase in CP typically observed for biodiesel blends. Fig. 5 examines the same data, and additional data for the other B100, plotted as a function of SMG content in the blend. For all fuels, SMG content has a dramatic effect on CP, however, there is also a clear effect of biodiesel blend level. Except for B100 I, CP is very similar for SMG content up to 0.01 wt.% to 0.015 wt.% regardless of biodiesel blend level or saturated FAME content. Above this level, the solvency effect of biodiesel is evident, with higher biodiesel content fuels exhibiting lower CP. For Fuel C blends, there also appears to be an effect of biodiesel saturated FAME content at the B20 level. While this effect is also present in the other fuels, it is less distinct. Fig. 5 also shows that the CP effect of a 0.01 wt.% change in SMG can be much larger than 4 °C at low SMG levels. Several of the B100 samples used to make blends at
-30
-50 -50
305
-20
-30
0
Fuel C
-40
-50 -50
-40
-30
-20
-10
0
D5773 Late CP, oC Fig. 2. Comparison of D5773 early CP and late CP results with results from methods D5771 and D7689.
Cloud Point, oC
-5
-10
-15 Distilled 0.048% SMG 0.148% SMG 0.229% SMG
-20
R2, but produced pb0.05 for only three out of the five parameters (diesel CP, percent biodiesel, and percent SMG); coefficient values were not significantly different.
-25 0
5
10
15
20
25
Percent Biodiesel
Light Scattering Signal
0
Fuel D
-10
Cloud Point, oC
0.0304 wt% SMG
0.0152 wt% SMG
-20
-30 Distilled 0.048% SMG 0.148% SMG 0.229% SMG
-40
0.0076 wt% SMG
-50 -25
-20
-15
-10
-5
0
5
Temperature, oC Fig. 3. Effect of SMG content on the early CP for B100 IV blended with Fuel A.
0
5
10
15
20
25
Percent Biodiesel Fig. 4. Effect of SMG content in B100 II on CP in Fuels C and D (SMG is percent in B100).
306
G.M. Chupka et al. / Fuel Processing Technology 118 (2014) 302–309
10
20 Fuel A
Fuel B
5 10
Cloud Point, oC
Cloud Point, oC
0 0 B5 B100 I B10 B100 I B20 B100 I B5 B100 II B10 B100 II B20 B100 II B5 B100 III B10 B100 III B20 B100 III B5 B100 IV B10 B100 IV B20 B100 IV
15% Saturated FAME
-10
6.7% Saturated FAME
-20
0.02
0.04
0.06
0.08
-10
B5 B100 I B10 B100 I B20 B100 I B5 B100 II B10 B100 II B20 B100 II B5 B100 III B10 B100 III B20 B100 III B5 B100 IV B10 B100 IV B20 B100 IV
15% Saturated FAME
-15 -20
6.7% Saturated FAME
-25
-30 0.00
-5
-30
0.10
0.00
0.02
Percent SMG
0.04
0.06
0.08
0.10
Percent SMG 20
15 Fuel C
10
Fuel D
10
5
Cloud Point, oC
Cloud Point, oC
36% Saturated FAME 24% Saturated FAME
0
B5 B100 I B10 B100 I B20 B100 I B5 B100 II B10 B100 II B20 B100 II B5 B100 III B10 B100 III B20 B100 III B5 B100 IV B10 B100 IV B20 B100 IV
-5 15% Saturated FAME
-10
6.7% Saturated FAME
-15
0.02
0.04
0.06
0.08
15% Saturated FAME B5 B100 I B10 B100 I B20 B100 I B5 B100 II B10 B100 II B20 B100 II B5 B100 III B10 B100 III B20 B100 III B5 B100 IV B10 B100 IV B20 B100 IV
-10 6.7% Saturated FAME
-20 -30
-20 0.00
0
0.10
-40 0.00
0.02
Percent SMG
0.04
0.06
0.08
0.10
Percent SMG
Fig. 5. Effect of blend SMG content on CP for B100 blended into Fuels A–D.
the highest SMG levels failed the total glycerin limit in ASTM D6751. Removal of these data points does not alter these trends. The slightly different results for B100 I may be caused by its low saturated FAME content, or by other factors as discussed in Section 3.3. Because B100 II as received does not contain SMG at detectable levels, its blends can be used to more clearly demonstrate the effect of saturated FAME content on CP. Fig. 6 shows CP for this B100 blended into all four base diesel fuels at the B5, B10, and B20 levels. In the highest CP base fuel (Fuel C), there is little effect of blending this distilled biodiesel on CP. Apparently, petroleum wax is the first species to precipitate at temperatures greater than or equal to −16.5 °C (the CP of the base diesel fuel) in this fuel for blends with saturated FAME
content up to 3 wt.%. For the lower CP diesel fuels, saturated FAME does increase CP, likely because of lower solubility at the lower temperatures of these CP experiments. Fuels A and B have almost the same CP, but are respectively the lowest and highest aromatic content fuels examined. A comparison of the CP of B5 blends in these fuels as a function of SMG content is shown in Fig. 7. In B5 blends, the effects of other factors such as saturated FAME content and biodiesel content should be minimized. For SMG content above about 0.005 wt.%, blends with the higher aromatic content diesel have a consistently lower CP, indicating that
10 -10 5 -15
Cloud Point, oC
0
Cloud Point, oC
-20 -25 -30 -35
-5 -10 -15 -20
Fuel A Fuel B Fuel C Fuel D
-40 -45
9.4% Aromatic (Fuel A) 35% Aromatic (Fuel B)
-25 -30 0.000
-50 0.0
0.5
1.0
1.5
2.0
2.5
3.0
0.005
0.010
0.015
0.020
0.025
Percent SMG
Percent Saturated FAME Fig. 6. Effect of percent saturated FAME (B100 II, as received) on the CP in diesel fuels A–D.
Fig. 7. Effect of diesel fuel percent aromatic content on CP for B5 blends as a function of SMG content.
G.M. Chupka et al. / Fuel Processing Technology 118 (2014) 302–309
aromatics can have a solvency effect on SMG and lower CP, consistent with the regression analysis model. 3.3. Effect of unsaturated monoglyceride on cloud point B100 I is a canola-derived B100 and contains a significantly higher ratio of unsaturated:saturated monoglyceride (UMG/SMG) than the other B100 used in this study (see Table 3). The CP of blends for B100 I were generally lower than expected based on trends shown in Fig. 5 for the other three B100 blends. We have observed that addition of MO, an unsaturated monoglyceride, can in some cases reduce the CP of a B100 blend. For example, in neat B100 I, the measured CP of a sample containing 0.28 wt.% SMG was 9.4 °C. Addition of 1.8 wt.% MO, for a total of 2.3 wt.% UMG, to the same sample lowered the CP to 2.9 °C. Fig. 8 shows the effect of SMG content on CP for blends with B100 I having UMG/SMG = 9.5 and 5.2. Data for B100 II with UMG/SMG = 4.9 are also included. Based on the trends shown in Fig. 5 for B100 II, III, and IV, the CP versus SMG curves for B100 I blends would be expected to be similar to those observed for B100 II, especially at low SMG and biodiesel content. However, CP for B100 I blends with UMG/SMG = 9.5 falls well below the B100 II curves for SMG greater than about 0.005 wt.%. Note that this UMG/SMG ratio is typical of canola-derived biodiesel. When B100 I was spiked with SMG only, such that UMG/SMG = 5.2, the measured CPs were very close to the CP curves observed for B100 II blends having UMG/SMG = 4.9. When additional MO was added to the B100 II/Fuel D blends to increase UMG/SMG up to 9.5, the CP decreased to the same level observed for the B100 I blends at this UMG/SMG ratio. Clearly, the presence of UMG, if the UMG/SMG ratio is high enough, can depress the CP of B100 and blends. 5
Fuel A
Cloud Point, oC
0 -5 -10 -15
B100 I B5 High U/S B100 I B10 High U/S B100 I B20 High U/S B100 I B5 Low U/S B100 I B10 Low U/S B100 I B20 Low U/S B100 II B5 B100 II B10 B100 II B20
-20 -25 -30 0.00
0.01
0.02
0.03
0.04
0.05
307
UMG could possibly interact with SMG to prevent or slow nucleation, similar to compounds such as water and stearic acid [17]. There are numerous industrial crystallization applications where additives are used to prevent or enhance the rate of nucleation and crystal growth [18]. It may be possible that MO, with polarity similar to SMG but significantly greater solubility at cold temperatures, could be interacting with the polar head groups of the SMG to inhibit nucleation and reduce CP on the time scale of the CP experiment. The most effective additives for inhibiting nucleation generally have similar structure to the nucleating species [19,20]. 3.4. SMG polymorphic phase transformation in biodiesel blends In previous work [8,9], we have shown that in B100, SMG precipitates as a metastable gel that can transform into the stable β polymorph over time or upon heating via solvent-mediated polymorphic phase transformation [21]. We proposed that this transformation could account for many of the unanticipated occurrences of fuel filter clogging by precipitate formation above the CP in biodiesel blends. The polymorphic transformation can manifest itself as a large signal increase during warming of the samples at 1.5 °C/min in light scattering plots, and the FMT is significantly higher than the CP. This signal increase was correlated to the polymorphic conversion of the SMG in microscope experiments, confirmed by X-ray powder diffraction, and differential scanning calorimetry. In many cases, however, the 1.5 °C/min heating rate is too fast to observe the transformation, and it is only revealed by lowering the heating rate to 0.5 °C/min or lower [9]. The polymorphic phase conversion can also occur in biodiesel blends. Fig. 9 shows two examples of light scattering plots for the FMT experiment for B100 II in Fuel C. For distilled B100 II blended at 5%, no polymorphism is noted and the signal returns rapidly to the baseline. For a B10 blend spiked to contain 0.023 wt.% SMG, there is a large signal increase and FMT increased by about 20 °C. Fig. 10 shows FMT and CP as a function of SMG content for blends of all four B100 in Fuel D. Plots for Fuels A, B, and C are similar and are provided in the Supplementary data. In all cases for B100 I, and above about 0.01 wt.% SMG for B100 II and III, the FMT-CP difference increases to 20°C or higher, indicating the transformation of α-SMG into the most stable and lower solubility β form. At lower concentrations, apparently all of the SMG dissolves before undergoing transformation at this heating rate. B100 I may exhibit transformation at lower concentration because of the CP suppression caused by the high UMG/SMG ratio, which lowers the starting temperature for the FMT experiment. This increases the time available for transformation in this experiment and may imply that transformation could be observed at concentrations below 0.01 wt.% SMG for B100 II and III at slower heating rates.
Percent SMG 0
B5 Distilled B100 B10 0.023 wt% SMG
Cloud Point, oC
-10 -15 B100 I B5 High U/S B100 I B10 High U/S B100 I B20 High U/S B100 I B5 Low U/S B100 I B10 Low U/S B100 I B20 Low U/S B100 II B5 B100 II B10 B100 II B20 B100 II B5 High U/S B100 II B10 High U/S B100 II B20 High U/S
-20 -25 -30 -35
Light Scattering Signal
Fuel D
-5
-40 0.00
0.01
0.02
0.03
0.04
0.05
Percent SMG Fig. 8. Effect of SMG content and UMG/SMG (U/S in figure) ratio on CP. High U/S = 9.5, low U/S = 4.9 or 5.2.
-20
-15
-10
-5
0
5
10
15
20
Temperature, oC Fig. 9. FMT experiment light scattering plots of B100 II blended into Fuel C demonstrating polymorphic phase conversion when SMG is present.
308
G.M. Chupka et al. / Fuel Processing Technology 118 (2014) 302–309
20
30
10 0 -10 -20 -30 -40
B100 II/Fuel D
Cloud Point or Final Melting Temperature, oC
Cloud Point or Final Melting Temperature, oC
B100 I/Fuel D
CP FMT at 1.5oC/min
20 10 0 -10 -20 -30 -40
CP FMT at 1.5oC/min
-50
-50 0.00
0.01
0.02
0.03
0.04
0.05
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
Percent SMG
Percent SMG 40
B100 III/Fuel D
20 10 0 -10 -20 -30 -40
CP FMT at 1.5oC/min
B100 IV/Fuel D
Cloud Point or Final Melting Temperature, oC
Cloud Point or Final Melting Temperature, oC
30
30 20 10 0 -10 -20 CP FMT at 1.5oC/min FMT at 0.3oC/min
-30 -40 -50
-50 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
Percent SMG
Percent SMG Fig. 10. CP and FMT as a function of SMG content for B100 in Fuel D.
For B100 IV, transformation is not observed at the 1.5°C/min heating rate. However, lowering the heating rate to 0.3 °C/min reveals a significant increase in FMT. The composition of B100 IV is apparently such that the phase transformation is slower and requires a slower heating rate to be revealed. Because real-world vehicle or storage tank heating rates could potentially be in the 1 °C to 2 °C/h range (0.015 °C to 0.03 °C/min), polymorphic phase transformation of the SMG could cause unanticipated fuel filter clogging for blends of all of these B100 samples; it may not be predictable by higher heating rate laboratory tests. B100 IV as received contained much higher levels of impurities than did B100 II or III. B100 II was produced by vacuum distillation and contains extremely low impurity levels. B100 III was also distilled or partially distilled and also had very low impurities. Thus, B100 IV may contain many impurities common to animal fats such as sterols, sterol oxides, fatty acids, or fatty acid oxidation products, or residual to the biodiesel production process (soaps and glycerol, in addition to monoand diglycerides) that are not present in B100 II or III. It is well established that impurities or additives can alter the rate of polymorphic phase transformation [18,22] and possibly, such an effect is active here. Because of the very low UMG/SMG ratio in B100 IV, we tested the hypothesis that adding MO to increase this ratio could increase the rate of phase transformation as evidenced by increasing FMT. The presence of monoglycerides has been shown to increase the rate of polymorphic phase transformation for palm oil crystallization [19]. However, adding MO was not successful at increasing FMT at the 1.5 °C/min heating rate. 4. Conclusions This study has shown two potential causes of precipitate formation above CP. For biodiesel blends containing approximately 0.01 wt.% to
0.03 wt.% SMG some CP measurement methods may not be adequately sensitive to detect the temperature where crystals first occur upon cooling. Samples with SMG content in this range may be particularly prone to unexpected filter clogging above the measured CP. Second, blends containing SMG may experience solvent mediated polymorphic phase transformation of the SMG, where the metastable α-form crystallizes initially but converts to the less soluble, stable β-form upon heating. Additionally, our results show that SMG content is the single largest factor affecting the CP of biodiesel blends. While the presence of saturated FAME will also lead to an increased CP, the SMG effect occurs at concentrations well below 0.1 wt.%. A significant effect due to increasing saturated FAME content requires an increase of more than 5 wt.%. At a constant SMG level increasing biodiesel content lowered CP, as did increasing the diesel fuel aromatic content, by improving the solubility of SMG and saturated FAME. This implies that by choosing a B100 containing a lower level of SMG a biodiesel blender can increase biodiesel blend level while maintaining or even reducing the CP of the blended product. In a surprising observation, increasing the UMG/SMG ratio could significantly reduce the CP of blends, presumably by inhibiting nucleation of the SMG. While these conclusions are based on the use of the D5773 early CP, it is important to note that this affected the CP result for only 15 out of the 140 samples. Use of the late CP for these samples in the data analysis did not significantly change the results. The results indicate that cold weather operability problems might be avoided by measurement of CP with the most sensitive methods available, by careful control of the blend SMG content, and potentially by engineering biodiesel to have a higher UMG/SMG content than would normally occur. The degree of FAME saturation is a secondary factor.
G.M. Chupka et al. / Fuel Processing Technology 118 (2014) 302–309
Acknowledgments This work was supported by the U.S. Department of Energy, Office of Vehicle Technologies, Fuels and Lubricants Technologies Program under Contract no. DEAC36-99GO10337 with the National Renewable Energy Laboratory. Funding was also provided by the National Biodiesel Board. The authors wish to thank Dr. Gordon Chiu of Phase Technology for comments on the manuscript.
Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.fuproc.2013.10.002.
References [1] R.O. Dunn, Effects of minor constituents on cold flow properties and performance of biodiesel, Progress in Energy and Combustion Science 35 (2009) 481–489. [2] J.E. Chandler, Comparison of all weather chassis dynamometer low temperature operability limits for heavy and light duty trucks with standard laboratory test methods, SAE Technical Paper No. 962197, 1996. [3] R.L. McCormick, J. Chandler, R. Gault, S. Howell, M. Nikanjam, A. Buczynsky, J. Jetter, R. Barenescu, D. Arters, D. Forester, K. Mitchell, Biodiesel blend low temperature performance validation, CRC Project: DP-2a-07June 2008. (http://www.crcao.org). [4] J.H. Van Gerpen, E.G. Hammond, L. Yu, A. Monyem, Determining the influence of contaminants on biodiesel properties, SAE Technical Paper No. 971685, 1997. [5] C. Selvidge, S. Blumenshine, K. Campbell, C. Dowell, J. Stolis, Effect of biodiesel impurities on filterability and phase separation from biodiesel and biodiesel blends, Proc. of the 10th International Conference on Stability, Handling, and Use of Liquid Fuels, Oct 7–11, 2007; Tucson, AZ, 2007, (http://iash.net/Archives/Archives_Proceed_ 2007.html). [6] M.-A. Poirier, P. Lai, L. Lawlor, The effect of biodiesel fuels on the low temperature operability of North American heavy duty diesel trucks, SAE Technical Paper No. 2008-01-2380, 2008.
309
[7] M. Brewer, Understanding low temperature biodiesel issues. Impact of impurities. Case studies of recent industry incidents in EU, 2nd International Congress on Biodiesel, Munich, November 17–19, 2009, 2009. [8] G.M. Chupka, J. Yanowitz, G. Chiu, T.L. Alleman, R.L. McCormick, Effect of saturated monoglyceride polymorphism on low-temperature performance of biodiesel, Energy & Fuels 25 (2011) 398–405. [9] G.M. Chupka, L. Fouts, R.L. McCormick, Effect of low-level impurities on lowtemperature performance properties of biodiesel, Energy & Environmental Science 5 (2012) 8734–8742. [10] E.S. Lutton, The phases of saturated 1-monoglycerides C14–C22, Journal of the American Oil Chemists' Society 48 (1971) 778–781. [11] W.G. Morley, G.J.T. Tiddy, Phase behaviour of monoglyceride/water systems. J.C.S, Faraday Transactions 89 (1993) 2323–2331. [12] E.B. Sirota, H.E. King Jr., D.M. Singer, H.H. Shao, Rotator phases of the normal alkanes: an X-ray scattering study, Journal of Chemical Physics 98 (1993) 5809–5824. [13] T.L. Alleman, L. Fouts, G.M. Chupka, Quality parameters and chemical analysis for biodiesel produced in the United States in 2011, NREL/TP-5400-57662, 2013. (http://www.nrel.gov/docs/fy13osti/57662.pdf). [14] ASTM Standard D5773-10, Standard Test Method for Cloud Point of Petroleum Products (Constant Cooling Rate Method), Section 13.4.3, ASTM International, West Conshohocken, PA, 2010. [15] A.J. Wright, S.S. Narine, A.G. Marangoni, Comparison of experimental techniques used in lipid crystallization studies, Journal of the American Oil Chemists' Society 77 (2000) 1239–1242. [16] ASTM Standard D2500-11, Standard Test Method for Cloud Point of Petroleum Products, ASTM International, West Conshohocken, PA, 2011. [17] C.H. Chen, E.M. Terentjev, Aging and metastability of monoglycerides in hydrophobic solutions, Langmuir 25 (2009) 6717–6724. [18] K.W. Smith, K. Ghaggan, G. Talbot, K.F. van Malssen, Crystallization of fats: influence of minor components and additives, Journal of the American Oil Chemists' Society 88 (2011) 1085–1101. [19] E. Fredrick, I. Foubert, J. Van De Sype, K. Dewettinck, Influence of monoglycerides on the crystallization behavior of palm oil, Crystal Growth and Design 8 (2008) 1833–1839. [20] B.A. Hendrikson, D.J.W. Grant, The effect of structurally related substances on the nucleation kinetics of paracetamol (acetaminophen), Journal of Crystal Growth 156 (1995) 252–260. [21] D. Croker, B.K. Hodnett, Mechanistic features of polymorphic transformations: the role of surfaces, Crystal Growth and Design 10 (2010) 2806–2816. [22] C.-H. Gu, K. Chatterjee, V. Young Jr., D.J.W. Grant, Stabilization of a metastable polymorph of sulfamerazine by structurally related additives, Journal of Crystal Growth 235 (2002) 471–481.