REVIEW pubs.acs.org/IECR
Impact of Sulfur Poisoning on the Catalytic Partial Oxidation of Methane on Rhodium-Based Catalysts Stefano Cimino* and Luciana Lisi Istituto Ricerche sulla Combustione CNR, P.le V. Tecchio 80, 80125, Napoli, Italy ABSTRACT: Catalytic partial oxidation (CPO) of natural gas into syngas (CO and H2) at short contact time over precious-metalstructured reactors has received increasing attention in industrial applications. However, only recently, sulfur poisoning of the bestperforming and costly rhodium catalysts has been recognized as a serious issue of this technology. This work critically reviews and integrates recent experimental results obtained from our group and others, both at steady state and during transient operation of the CPO reactor, particularly with regard to poisoning/regeneration cycles and low-temperature light-off phase. Furthermore, the origin of sulfur poisoning of Rh active sites during the CPO of methane has been studied via in situ diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy, using CO as a probe molecule, and an attempt is made to identify sites responsible for the loss in the catalytic activity by comparing the results with the activity data. The effect of the type of alumina support and the partial substitution of Rh with either Pt or Pd were investigated to enhance the sulfur tolerance of the catalyst.
1. INTRODUCTION Depletion of the worldwide reserves of crude oil has resulted in an increased interest in the use of natural gas to meet the world’s energy demand. Even though the world has large deposits of natural gas, most of them are located in remote areas; consequently, it must be transported across vast distances to reach its markets.1 The catalytic partial oxidation (CPO) of methane over preciousmetal catalysts has been shown to be an attractive way to obtain syngas (CO and H2). In contrast to steam reforming, CPO has the advantage of being mildly exothermic and can be carried out at extremely short contact times (102104 s) with autothermal reactor configurations and >90% selectivity to H2 and CO.2,3 An additional advantage of CPO is that it produces a H2/CO ratio of 2, which is needed for the downstream processing to liquid fuels.3 Therefore, CPO has emerged as a suitable reaction path for onboard reforming of gaseous or liquid fuels for automotive and residential fuel cell systems and has been explored for industrial syngas production from natural gas for FischerTropsch process or the synthesis of high-value chemicals.4,5 Furthermore, the beneficial role of H2 in fuel combustion, in terms of emissions control and stability enhancement, leads to increased opportunity for CPO in industrial power generation technology.6,7 In particular, CPO of various hydrocarbons has been proposed as an integrated preliminary conversion stage for ultralow NOx hybrid catalytic burners for gas turbines or boilers.811 Although CPO activity has been promising from laboratoryscale experiments, its application in industrial processes still requires further development to identify and mitigate potential technical risks in several aspects, including its compatibility with impurities tolerance, particularly sulfur species in the process fuel stream.12 Desulfurization units can be used to significantly reduce the sulfur content in the feed, but its inclusion increases the complexity, size, and cost of the fuel processing system. Therefore, it will be more desirable to develop catalysts that are intrinsically sulfur tolerant and are not readily poisoned by the r 2011 American Chemical Society
amounts of sulfur commonly found in fuels such as natural gas (typically 810 ppmv in pipeline gas). Many catalysts have been tested for the CPO of hydrocarbons from methane up to diesel and jet fuels, and Rh-based catalysts have shown the highest activity and selectivity to syngas.1 CPO proceeds through an exo-endothermic sequence of reactions: by means of spatially resolved composition and temperature profiles, it was shown that O2 is rapidly depleted,3,1315 and a significant portion of CH4 is consumed to produce a mixture of H2, H2O, CO, and CO2 within a short front section of the catalyst bed (oxidation zone). In the remaining section, syngas production occurs mainly through the steam reforming reaction of the remaining CH4 fuel,13,14 and the equilibrium water-gas shift (WGS) reaction occurs at comparable rates, depending on H2O and CO2 enrichment.3 It was reported that the higher syngas yield on rhodium, with respect to platinum, is strictly associated with its higher activity for methane steam reforming, whose contribution is significant, even at very short contact times.13 The adverse impact of sulfur compounds on catalytic performance is well-known16,17 and is the subject of much current research, particularly in the case of (autothermal) steam reforming of liquid fuels1824 and automotive catalysts.25 However, studies on the effect of sulfur-bearing compounds naturally occurring in the fuel, or added as odorants to pipe-line natural gas during the CPO of methane over Rh-based catalysts have, so far, remained scarce. As recently pointed out by the group of L. D. Schmidt,26 poisoning by CH3SH on RhCe-coated foam monoliths operated at short contact time has a marked negative effect on CH4 conversion and H2 selectivity, even at small concentrations of sulfur, mainly Special Issue: Russo Issue Received: July 28, 2011 Accepted: December 15, 2011 Revised: December 14, 2011 Published: December 15, 2011 7459
dx.doi.org/10.1021/ie201648e | Ind. Eng. Chem. Res. 2012, 51, 7459–7466
Industrial & Engineering Chemistry Research
REVIEW
attributed to partially reversible inhibition of the steam reforming reaction path. Furthermore, under elevated reactor pressure (up to 160 psi), sulfur did not cause a permanent poisoning effect, and a RhCe-supported foam catalyst demonstrated stable long-term performance, with a total of 8 ppm of sulfur in the natural gas stream.12 Moreover, it was shown that platinum has greater tolerance than RhCe to sulfur poisoning,26 which, in turn, seems to be unaffected by the type of sulfur-bearing compound.27 Accordingly, we set out to investigate the impact of sulfur addition during the CPO of methane at high temperatures (>700 °C) and short contact times over rhodium catalysts supported on either La2O3γ-Al2O3- or SiO2γ-Al2O3-coated honeycomb monoliths. We also studied the enhancement in sulfur tolerance of rhodium-based catalyst by partially substituting rhodium with either platinum or palladium, which will be highly economical, because of the high cost of rhodium metal. This work critically reviews and integrates recent experimental results obtained both at steady state and during transient operation of the CPO reactor, particularly with regard to poisoning/regeneration cycles and low-temperature light-off phase. Furthermore, the origin of sulfur poisoning of Rh-active sites during the CPO of methane was studied via in situ diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy, using CO as a probe molecule, and an attempt was made to identify sites responsible for the loss in the catalytic activity by comparing the results with the activity data.
2. EXPERIMENTAL SECTION 2.1. Catalyst Preparation and Characterization. CPO of methane to syngas was investigated over monometallic 1 wt % Rh and bimetallic RhPt and RhPd catalysts (0.50.5 wt %) prepared by sequential impregnations on commercial γ-Al2O3 stabilized with either 3% La2O3 or 10% SiO2 wt % (respectively, type SCFa140-L3 and type Siralox 10-360, from Sasol), and anchored as a washcoat layer by a dip-coating procedure onto honeycomb monoliths with straight and parallel channels of roughly square section (cordierite, 600 cpsi by NGK, D L = 17 mm 10 mm). Further experimental details on catalysts preparation and characterization (via inductively coupled plasma mass spectroscopy (ICP-MS), scanning electron microscopy coupled with energy-dispersive X-ray analysis (SEM-EDAX), BrunauerEmmettTeller (BET) surface area analysis, CO chemisorption, temperature-programmed reduction in a hydrogen atmosphere (H2-TPR), temperature-programmed desorption in N2 after aging in SO2/air mix (SO2-TPD), in situ DRIFT with CO as probe molecule) can be found elsewhere.2830 2.2. Catalyst testing. The impact of sulfur addition (258 ppmv H2S or SO2) was studied under pseudo-adiabatic operation of the CPO reactor at CH4/O2 feed ratios in the range of 1.62, under both transient and steady-state conditions, using air as an oxidant, or oxygen with N2 added to obtain a 10 or 20 vol % dilution of the feed (gas hourly space velocity (GHSV) of 5 1048 104 h1 under standard conditions, based on the volume of the catalytic honeycomb; fixed preheating of 250 °C). Catalytic light-off temperatures were determined by rampingup the external furnace from 200 °C to the light-off temperature at a rate of ∼5 °C/min under transient conditions, using methane air mixtures at fixed feed ratio (CH4/O2 = 1.8), in the absence and presence of up to 20 ppm H2S. Reactor temperatures were measured by means of Type K thermocouples (d = 0.5 mm) placed in the middle of the central channel of the catalysts, in contact with the solid.
Figure 1. Transient response to H2S addition (8, 18, or 37 ppm) and removal on (a) methane conversion and (b) catalyst temperature. Panel (c) depicts the steady-state impact on methane conversion and yields to CO and H2. Rh on Laγ-Al2O3 catalyst; CH4/O2 feed ratio = 2, gas hourly space velocity (GHSV) = 6.7 104 h1, N2 = 20 vol %.
The fate of sulfur was determined by gas chromatography (GC) employing a specific high-sensitivity dual-plasma sulfur chemiluminescence detector (SCD, by Agilent) and a gas-pro column. Methane conversion, yields, and selectivities to CO and H2 were calculated according to the following definitions: ! CH4 OUT xCH4 ¼ 100 1 CHOUT þ COOUT þ COOUT 4 2 YCO ¼ 100
CHOUT 4
COOUT þ COOUT þ COOUT 2
!
! SCO ¼ 100
Y H2
100 ¼ 2
YCO xCH4
CHOUT 4
H2 OUT þ COOUT þ COOUT 2
!
! SH2 ¼ 100 2
Y H2 xCH4
These definitions are based on the exit dry-gas mole fractions of CO, CO2, CH4, and H2 independently measured by a continuous analyzer with cross sensitivity correction (ABB). O2 was always 7460
dx.doi.org/10.1021/ie201648e |Ind. Eng. Chem. Res. 2012, 51, 7459–7466
Industrial & Engineering Chemistry Research
REVIEW
Table 1. Impact of 8 ppm H2S Addition on the Steady-State Methane CPO on Rh/La-γ-Al2O3 Monolitha H2S differenceb, Δ (80)
CH4/O2 = 2
0 ppm
8 ppm
XCH4
86.1%
81.9%
4.2
YH2
84.7%
78.1%
13.2
YCO YH2O
80.6% 1.4%
76.1% 3.8%
4.5 +4.8
YCO2
5.5%
5.8%
+0.3
Tcat
752 °C
839 °C
+87 °C
a Feed CH4/O2 = 2, GHSV= 6.7 104 h1, N2 = 20 vol %. b Differences in molar flows of products per 100 CH4 feed.
completely converted under steady-state operation (TCD-GC analysis) after lightoff had occurred; H2O production was calculated from the oxygen balance. The carbon and hydrogen balances were always closed, within (1.5% and (3.5%, respectively.
3. RESULTS AND DISCUSSION Figure 1 shows that sulfur poisoning on Rh/La2O3γ-Al2O3 catalysts is rapid and completely reversible, and it is dependent on sulfur concentration. The step addition of sulfur to the feed to CPO reactor resulted in a rapid decrease in CH4 conversion with a corresponding sharp increase in the catalyst temperature. At each sulfur level, CPO performance reached a steady state within roughly 10 min; however, upon removing the H2S from the reaction feed, the activity of the catalyst was found to increase immediately and the initial CPO activity/selectivity was regained on a time scale that is comparable to that of the poisoning process. The presence of cerium was reported to slow the recovery of CPO activity of Rh/CeO2γ-Al2O3 foam catalyst.26 The complete reversibility of the sulfur poisoning effect under the studied reaction condition was also confirmed by the identical values of the temperature measured in the catalyst before the addition of sulfur and after its removal from the reaction feed. Moreover, the extent of catalyst deactivation was dependent on sulfur concentration in the feed but not on the type of sulfurbearing compound (as H2S or SO2). In fact, GC analysis of the product stream confirmed that, under the typical CPO operating conditions (i.e., high H2 partial pressure and high temperatures (>800 °C)), all of the sulfur is transformed to H2S (with only traces of COS), in agreement with thermodynamic predictions.27 As shown in Figure 1c, at steady state, sulfur poisoning impacts more severely on the H2 yield, since both methane conversion and H2 selectivity progressively decrease with the addition of sulfur, whereas the CO selectivity is almost unaffected also at the highest H2S concentrations. On the other hand, the oxygen conversion was always 100% and was not affected by sulfur poisoning, in good agreement with results of Bitsch-Larsen et al.,26 who reported that the rates of O2 consumption (i.e., O2 spatial profiles along the reactor) and the initial CH4 conversion are independent of the sulfur level. This is most likely due to the kinetics of the oxidation reactions, which remains much faster than the transport phenomena, even in the presence of sulfur. Careful analysis of the differences in products distribution induced by sulfur addition,26,28 exemplified in Table 1, coupled
Figure 2. Impact of the addition of sulfur (058 ppm) during steadystate CPO of methane over Rh on La2O3γ-Al2O3 honeycomb catalyst at two N2 dilution levels of the feed (20 and 55.6 vol % = air as an oxidant); GHSV = 6.7 104 and 8.1 104 h1 respectively; CH4/O2 = 2: (a) increase in catalyst temperature, as a function of the variation of integral methane conversion; (b) catalyst temperature, as a function of sulfur level in the feed; and (c) thermodynamic activity of chemisorbed sulfur on rhodium, defined as the ratio of H2S to H2 at the exit of the reactor, as a function of the inverse of the temperature recorded on the catalyst.
with the increase in the temperature level of the catalyst,12,26,28 shown in Figure 2, strongly suggest that the poisoning effect of sulfur is related to its ability to inhibit the endothermic steam reforming (SR) reaction path described by reaction 1, CH4 þ H2 O T CO þ 3H2
ΔHr 25°C ¼ þ 206 kJ=mol
ð1Þ which mainly occurs under kinetic control in the second region of the catalyst bed between the unconverted methane and water produced in the first oxidation zone of the reactor.3,13,15 Indeed, rhodium is reported to be very active for the steam reforming of methane, particularly when dispersed on γ-Al2O3 supports,3 even at extremely short contact time (