nh3 storage on a zeolite scr catalyst measured using a microwave based method reduced sensitivity for more strongly held nh3

We demonstrate that a comparison of the frequency changes of the resonant modes can help determine variations in the spatial profile of the NH3 storage along the SCR axis.. Analysis of d

Procedia Engineering 168 ( 2016 ) 11 – 14

1877-7058 © 2016 The Authors Published by Elsevier Ltd This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of the organizing committee of the 30th Eurosensors Conference

doi: 10.1016/j.proeng.2016.11.114

ScienceDirect

30th Eurosensors Conference, EUROSENSORS 2016

D Kubinskia*, A Bognerb

a Ford Research and Innovation Center, 2101 Village Rd., Dearborn, Michigan, 48121, USA b

Department of Functional Materials, University of Bayrueth, Bayreuth 95440, Germany

Abstract

One method for reduction of nitrogen oxides from the exhaust of modern diesel vehicles is the selective catalytic reduction (SCR) of NOx by NH3 injected onto a zeolite SCR catalyst Knowledge of the amount of NH3 stored on these catalysts is useful for optimal conversion Here, measurements of the NH3 storage on a Cu-chabazite zeolite SCR catalyst were done using the microwave-frequency resonator cavity method Shown are results for NH3 loading

on a previously empty catalyst and for the subsequent desorption of the more weakly held NH3 Temperature was held constant at 250 °C We demonstrate that a comparison of the frequency changes of the resonant modes can help determine variations in the spatial profile of the NH3 storage along the SCR axis Analysis of desorption data shows that the sensitivity of the resonant frequency change per mass of NH3 stored on the catalyst is lower for that more strongly held

© 2016 The Authors Published by Elsevier Ltd

Peer-review under responsibility of the organizing committee of the 30th Eurosensors Conference

Keywords: selective catalytic reduction (SCR); Cu-chabazite; ammonia storage; resonant cavity; microwave

1 Introduction

One of the more common after treatment methods used for NOx reduction in today’s diesel vehicles is the selective catalytic reduction (SCR) of NOx by NH3 In these systems, urea containing solution injected into the exhaust gas stream is converted to NH3 which selectively reacts with the NOx on the SCR catalyst to produce N2 and

* Corresponding author Tel.: +1-313-322-3810

E-mail address: dkubinsk@ford.com

© 2016 The Authors Published by Elsevier Ltd This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of the organizing committee of the 30th Eurosensors Conference

water The SCR catalyst is often a zeolite material and conversion of NOx is optimized when excess NH3 is stored

on its acidic sites Information about the amount of NH3 stored on the catalyst is desired, as too little can result in

poor NOx conversion and too much in NH3 emissions In a series of papers, it has been demonstrated that it is

possible to measure this quantity directly and in-situ using the microwave resonator cavity method [1-3] Resonant

frequency changes were demonstrated to correlate with the mass of NH3 stored on the catalyst A temperature

dependency was noted Here we demonstrate an additional factor affecting the sensitivity of this technique We

report on the changes in sensitivity measured as the more weakly held NH3 desorbs from the catalyst at constant

temperature

2 Experimental

In this study the SCR catalyst was a commercial Cu-chabazite zeolite used in serial application Figure 1 shows

the configuration of the cylindrical microwave resonant cavity used to measure the NH3 storage on the catalyst The

stainless-steel cavity was 17.1 cm long and ~4.4 cm in diameter with metal stainless screens welded on each end

The Cu-chabazite SCR catalyst, cored from a larger piece, was 7.6 cm long and ~4.2 mm in diameter and was

mounted approximately in the center of the cavity A single antenna positioned between the catalyst and back screen

enabled measurement of the S11 reflection parameter, with microwave signals generated and detected by an Agilent

E5071C network analyzer A 60 liter/min gas flow was heated via an in-line heater upstream of the cavity, with

temperature probes placed near the center of the metal screens at the ends of the cavity The composition of the

outlet gas was monitored by a FTIR analyzer (MIDAC) Additionally, automotive NOx sensors, which are also

sensitive to NH3, were placed both upstream and downstream of the cavity All measurements were made in 5%

accompanying O2 and 4% water The gas temperature measured a few mm upstream of the front screen was kept at

250 °C Temperature just downstream the cavity was ~238 °C The mass of NH3 stored on the catalyst was

calculated from differences in the gas concentrations following the procedure described in Ref 2

In general, the resonant frequency, fo, of the cavity of volume V c will alter due to a change in dielectric

permittivity, 'H, of the material within For the TE resonant modes this is given by [4]:

dV E

dV E V f

f

C C

V

V o

2 2

|

|

|

| ) (

&

&

³

³ '



|

Overlaid in Fig 1 are the squares of the calculated electric field strengths along the SCR axis for the TE111 and TE112

resonant modes For the TE112, the transverse electric field magnitude is zero midway along the axis of the catalyst,

and Eq 1 predicts that this mode will not respond to changes in the dielectric permittivity occurring there

Comparatively, the electric field for the TE111 mode is non-zero in that same location and will register a response

These differences will be used to determine information about variations in the NH3 storage along the SCR catalyst

axis

Fig 1 Schematic of the cylindrical resonant cavity and Cu-chabazite SCR catalyst placed within Overlaid are plots of the square of the electric

field variation along the cavity axis for the TE 111 and TE 112 resonant modes

3 Results

Figure 2 shows the measured correlation between the estimated mass of NH3 stored on the SCR and the corresponding resonant frequency changes for the TE111 and TE112 modes at 250 °C Identified in the figure are the intervals corresponding to the loading of the 400 ppm feed-gas NH3 on the previously empty catalyst (between 0 and

3500 sec) and the desorption of the more weakly held NH3 beginning after the feed-gas was set back to 0 ppm Note that for the loading interval when NH3 was in the feed-gas, both the resonant frequencies and mass of stored NH3

saturate after ~1000 sec The most notable difference between the responses for the two resonant modes during this period is the reduced sensitivity observed for the TE112 mode near ~500 sec The magnitude of 'f/fo decreased for both modes during the desorption interval Very little NH3 was measured downstream of the SCR catalyst at the end

of desorption at 11,000 sec At that time the SCR still contained NH3, about 50% full, which was held on average more strongly [1] This was reacted off with NO added to the gas stream, bringing the frequencies back to their original values This final step is not shown here, but the data are similar to that previously reported [1-3]

To better elucidate differences in the NH3 adsorption and desorption behaviors, the frequency change of the

TE112 mode is plotted relative to the corresponding change for the TE111 in Fig 3 Data are shown for both the NH3

adsorption and desorption intervals Differences in the responses for these modes during the loading interval (blue

curve from A o B) are a consequence of the NH3 storage moving through the catalyst with time, filling it from front

to back, and encountering different electric field profiles as the NH3 storage “front” moves along the SCR axis The region with reduced slope in the middle of the adsorption curve is a consequence of the loss of sensitivity of the

TE112 mode at the cavity center where its electric field is zero For the desorption interval we observed a very

different behavior, with Fig 3 showing a straight line response (red line from B o C) Since there are obvious

differences in the electric field distribution across the SCR axis for the two modes, this indicates little spatial variation along the SCR’s axis in the relative rate of NH3 desorption during this interval

4 Discussion

Fig 2 Top plot shows the feed-gas and downstream NH 3

concentrations Middle plot shows the mass of NH 3 estimated on

the catalyst with the loading and desorption intervals identified

Lower plot shows the corresponding change in resonant

frequencies for the TE 111 and TE 112 modes Data at 250 °C with

O 2 = 5% and water = 4% Space velocity = 30,000 hr -1

.

Fig 3 A plot of the fractional frequency change of the TE 112

mode versus that for the TE 111 during the NH 3 loading (blue) and

NH 3 desorption (red) intervals Point A denotes the empty catalyst, point B the fully loaded by the 400 ppm NH3 , and point

C the end of the desorption interval where only the more strongly

held NH 3 remain These same points are noted in Fig 2

We define the sensitivity, S, at each moment as the magnitude of the percent change in resonant frequency

(relative to the unloaded state) per gram of NH 3 stored on the catalyst Careful analysis of the data for both modes in Fig 2 shows a reduced sensitivity at the end desorption interval compared to that at its start when the catalyst was

full (points C and B in Fig 2, respectively) Figure 4 shows this in detail, plotting the fractional change in sensitivity

for the two modes during the desorption interval The sensitivity decreased for both resonant modes as the NH3

desorbed, dropping at the end to less than 0.6 of its starting value The nearly equal values for the fractional change

in S observed for the two modes further demonstrate the relative desorption of NH3 during this interval being independent of axial position along the catalyst

The constant temperature desorption removes the more weakly held NH3, with only the more strongly held

remaining on the SCR at point C [1] The observed sensitivity decrease for the more strongly held NH3 may be related to a corresponding reduction in the proton conductivity on the zeolite surface [5] Following the argument given elsewhere, we expect the more weakly held NH3 on the zeolite surface, possibly a (NH3)nNH4 complex, to have a lower activation energy for mobility and interact with the microwave electric fields more easily (i.e., higher sensitivity) than is the case for that more strongly held, most likely as NH4 [5, 6]

5 Conclusions

Using the microwave resonant method we found that although loading of NH3 on the SCR Cu-chabazite catalyst proceeded in a front to back direction with time, its subsequent desorption after it was fully loaded occurred uniformly along the catalyst axis We also demonstrated that the sensitivity of the microwave resonant cavity method (defined as frequency change per gram of stored NH3) decreased as the more weakly held NH3 desorbed at constant temperature The more strongly held NH3 remaining on the catalyst at the end of the desorption interval showed only ~60% the sensitivity value measured when it was fully loaded

References

[1] D Rauch, D Kubinski, G Cavataio, D Upadhyay, R Moos, Ammonia loading detection of zeolite SCR catalysts using a radio frequency

based method, SAE Int J Eng 8(3) (2015) 1126-1135.

[2] D Rauch, D Kubinski, U Simon, R Moos, Detection of the ammonia loading of a Cu Chabazite SCR catalyst by a radio frequency-based method, Sens and Act B: Chemical 205 (2014) 88-93

[3] R Moos, D Rauch1, M Votsmeier, D Kubinski, Review on radio frequency based monitoring of SCR and three way catalysts, Top Catal

59 (2016) 961–969

[4] L.F Chen C.K Ong, C.P Neo, V.V Varadan, V.K Varadan, Microwave Electronics, first ed., John Wiley and Sons, West Sussex, 2004 [5] U Simon, U Flesch, W Maunz, R Muller, C Plog, The effect of NH 3 on the ionic conductivity of dehydrated zeolites Na beta and H beta, Microporous and Mesoporous Materials 21 (1998) 111–116

[6] L Rodríguez-González, E Rodríguez-Castellón, A Jiménez-López, U Simon, Correlation of TPD and impedance measurements on the

Figure 4: Comparison of the relative RF sensitivity (S = 'freq/mass NH3 ) for the TE 111 and TE 112 modes during the NH 3 desorption interval Data are plotted as a function of the mass of NH 3 stored on the catalyst The path B o C is as identified in Fig 2 The values have been normalized relative to the sensitivity a point B, SB , the value for the fully loaded catalyst