D. E. Resasco

School of Chemical Engineering. University of Oklahoma

100 E. Boyd St, Norman OK 73019

(405) 325-4370, fax (405) 325-5813, e-mail resasco@ou.edu

web page: www.ou.edu/catalysis

Palladium catalysts are usually not selective for the reduction of NO with methane because, in the presence of excess oxygen, they are very active for the methane combustion reaction, leaving NO unconverted. However, when supported on H-ZSM-5 Pd becomes highly selective for NO reduction. In previous work, we have shown that other acidic supports such as H-MOR, sulfated zirconia and SiO₂-Al₂O₃ are also able to promote the selectivity enhancement. Comparing one of the most active Pd catalyst (0.3% Pd/H-ZSM-5) to a typical Co-ZSM-5 catalyst, we found that the Pd catalyst exhibits a higher selectivity and activity under clean conditions. One of the typical limitations of zeolitic supports for industrial SCR applications has been its low tolerance to the presence of water in the feed, which at the high temperatures of the reaction may cause structutural deterioration. It is important to investigate the possibility of using non-zeolitic supports such as the sulfated zirconia and tungstated zirconia to determine whether they can resist the presence of these impurities. In this contribution we investigated the effects of different pretreatment histories, and the presence of water and SO₂ on several Pd catalysts.

Activity tests were conducted by adding 10% of H₂O or 75 ppm SO₂ to a feed stream consists of 7,500 ppm CH₄, 1.9% O₂, with or without 3500 ppm of NO while maintaining a constant GHSV. One of the most interesting results found in this study was the suppression of methane combustion under SCR conditions. The combustion suppression was most significant in catalysts that yielded the maximum NO conversion. Reaction cycles on a 0.3% Pd/H-ZSM-5 catalyst were conducted by turning NO on-and-off in the presence of CH₄/O₂. In every case, increases in NO conversion accompanied the decrease in methane combustion every time the NO was turned on. Upon removal of NO from the feed, the combustion activity regained almost immediately. EXAFS analysis indicate that the suppression of methane combustion in the presence of NO is attributed to the conversion of PdO clusters to the isolated Pd²⁺ ions. This redispersion of Pd by the presence of NO is believed to occur via a volatile PdO(NO) species, which would diffuse through the zeolite pores and get anchored at a protonic site. It was found that this reversibility was only observed on some supports. For example, on Pd/HZSM-5 and Pd/H-Mor, the same conversion levels were obtained after each cycle. By contrast, on Pd/SZ catalyst, the NO conversion and methane combustion strongly depended on the catalyst history. For example, after exposing a 0.1% Pd/SZ catalyst to CH₄+O₂ for 2h, the NO conversion only regained about 50% of its original value. The state of Pd and density of acid sites after exposure to various conditions have been characterized by EXAFS, CO chemisorption, microcalorimetry, and TPD and FTIR of adsorbed NO.

In addition to the effects of pretreatment history, we have examined the activity of several Pd based catalysts in the presence of water vapor and SO₂. Water, similar to NO, is a Lewis base that might compete with NO on the same adsorption sites. At the same time, since acid sites are needed for the stabilization of the selective Pd²⁺ species, the presence of water might inhibit the selectivity of Pd catalysts. In fact, it was observed that, in the presence of water, the NO conversion of 0.1% Pd/SZ catalyst was suppressed by about 30%. However, removing water from the feed after 2h showed that the decrease in activity was reversible. One question that needs to be addressed here is whether the water blocks the Pd species or the acid sites under SCR. In our previous studies, we have shown that physical mixtures of Pd supported on an inert support (SiO₂) and bare sulfated zirconia are active for SCR, although both materials by themselves did not have a high NO reduction activity. This effect may be linked to the redispersion of Pd in the presence of NO from SiO₂ to sulfated zirconia, similar to that proposed for H-ZSM-5.

The NO conversion was measured in two separate experiments on a 1:1 physical mixture of 0.1% Pd/SiO₂ and sulfated zirconia. For the first run, the mixture was pretreated in pure He, for the second it was pretreated in the presence of 10% of water. During the first two hour under CH₄+O₂, the catalyst pretreated under dry He exhibited a high initial methane combustion, while a same sample pretreated under wet environment had a low initial methane combustion of ~ 10%, but greatly increased after 30 min on stream, reaching the same level of conversion as the first one. This behavior indicates that water is most probably blocking the Pd on the Pd/SiO₂. On an inert support, we have shown that majority of the Pd species are in the form of PdO clusters. Water may interact with these PdO clusters, forming Pd(OH)₂ species and decreasing the rate of combustion. However, after 30 min on stream, the OH species that adsorbed on PdO clusters during the pretreatment was completely desorbed, results in a sharp increase in methane combustion rate. Although the steady state methane combustion remains the same for wet or dry pretreatment, as soon as NO was turned on after the two hour under CH₄+O₂, the NO conversion for the sample pretreated under wet conditions was less than half of that pretreated in dry He. This behavior indicates that during the pretreatment under water, the acid sites responsible for the anchoring of the volatile Pd species become blocked and lose their ability to stabilize the Pd²⁺ species needed for SCR.

We may conclude that the NO reduction activity on Pd catalysts could be sensitive to its history, depends on the type of support used. Under SCR conditions, water most probably attacks the acid sites, however, the effects are reversible as soon as it was removed from the feed.