VOLUME I: ANOSTRUCTURED NITROGEN-CONTAINING CARBON
4.2.1 Determination of Acceptable Treatment Parameters using TGA
The active ORR catalysts studied in previous section were prepared using carbon as a support for an Fe salt, since carbon is conductive. Using carbon as a support is more practical for use in an actual fuel cell; however, alumina is better suited for theoretical studies for several reasons. First, alumina can be made in the lab with well-defined characteristics and high purity, unlike commercial carbons that will contribute Fe contamination and a plethora of ill-defined and potentially reactive surface carbon-oxide species. Second, the alumina itself cannot be made active for the ORR since it is not conductive. Third, it is possible to selectively remove the alumina after the pyrolysis by treatment with a strong acid (HF) leaving behind only the nitrogen-containing carbon formed during the pyrolysis, allowing for easier characterization of the pyrolysis product.
In this section pure alumina, and alumina supported Fe and Ni are examined as supports for acetonitrile decomposition. Determining the temperature necessary to initiate carbon growth is the first step for catalyst preparation. TGA experiments were run to see what temperature carbon deposition takes off on the various supports as the temperature is increased. Table 8 shows the temperature at which sample weight began to increase in the presence of acetonitrile humidified N2. Light-off temperatures were slightly higher than for the Vulcan Carbon supported catalysts in the previous study, although again all samples are within +/- 100oC of 600oC. Metal containing samples began to increase their mass at a lower temperature than pure alumina, with Ni having the
most favorable effect.
Sample T(oC) pure Al2O3 684 2% Fe / Al2O3 627 reduced 2% Fe / Al2O3 606 2% Ni / Al2O3 576 reduced 2% Ni / Al2O3 545
Table 8: Acetonitrile decomposition light-off temperatures for alumina supported samples.
Reducing the samples prior to the acetonitrile treatment lowered the take-off temperature 21oC and 31oC for the Fe and Ni samples respectively. Figure 35 and Figure 36 compare the alumina supported Fe and Ni sample growth profiles during the temperature ramp with and without pre-reduction. Because the ability of the alumina support to adsorb water, the samples undergo a weight loss during the beginning of the temperature ramp. Since the pre-reduced sample was already subjected to a temperature of 650oC, its weigh loss is less significant. However, after the weight increase initiates, the rate of weight gain is essentially identical for the two samples. The alumina supported Fe sample also had the same trend in growth rate with or without pre- reduction. Because of the negligible difference in fiber growth rates, further prepared catalyst samples were not reduced before the acetonitrile treatment. Additionally, the carbon-supported samples were not pre-reduced with hydrogen (because the carbon could gasify), therefore, for comparison it is beneficial that the treatment parameters match exactly. Furthermore, activity testing also showed negligible differences between the pre-reduced and non-reduced samples.
A second series of experiments carried out on the precursors with TGA was necessary to determine carbon yields during sample preparation. The TGA of the initial weight loss during the temperature ramp is shown for all the different Fe-doped supports of this study in Figure 37. The differences in the TGA profile are dependant on how the Fe is introduced to the sample. For instance, the 2% Fe and 10% Fe samples both had Fe acetate introduced through wet impregnation, then they were dried at 100oC overnight.
Therefore, both samples have similar profiles with little weight loss below 100oC, and a total weight loss of about 24% once the temperature reaches 800oC. Most of the mass loss is believed to be coming from water desorbing from the alumina surface. There are two possible modifications to the preparation procedure that effect this profile. First, incipient wetness catalysts were not dried, so they contain more water, and lose more mass at lower temperatures. However, the overall weight loss is only slightly higher.
Second, the 57Fe enriched samples were also consequently enriched with the acetic acid used to dissolve the 57Fe. Therefore these samples undergo more of a mass loss at higher temperatures presumably from the acid decomposing. The total mass lost is much higher for these samples. The total mass lost for each support was accounted for when the weight changes were calculated, and the term “dry support weight” reflects that the weight lost during the temperature ramp has been applied.
The next TGA experiments carried out sought to determine what an acceptable treatment time would be for catalyst preparation. In these experiments the temperature was raised to 800oC, and left for 12 hours to observe the weight increase trend. The alumina had much higher carbon uptake than the other samples (including the Vulcan Carbon samples in the previous study), although the rate did slow after 6 hours of
treatment, as Figure 38 shows. The figure also demonstrates that adding a metal to alumina will cause the weight gain to level out after 2 hours of treatment. Further characterization will attempt offer some clues to help explain this trend.
Despite some small differences, the decomposition of acetonitrile over an alumina support generally seems to be a similar process to the decomposition over Vulcan Carbon. In the Vulcan Carbon study, the most active catalysts resulted from treatments at 900oC for 2 hours. Therefore, for comparison sake, samples were prepared using the three different supports (pure alumina, 2% Fe on alumina, and 2% Ni on alumina) under the same conditions. Additionally, a pure alumina support was treated for 12 hours, since TGA indicated the support was capable of weight increases for extended treatments.
Figure 35: TGA profile for 2-wt% Fe/Al2O3 during temperature ramp in acetonitrile with and without pre-reduction.
Figure 36: TGA profile for 2-wt% Ni/Al2O3 during temperature ramp in acetonitrile with and without pre-reduction.
Figure 37: TGA profile of weight loss during temperature ramp in inert atmosphere for some various Fe-doped alumina precursors.
Figure 38: TGA profiles of acetonitrile decomposition at 800oC over alumina supports.