8/11/2019 assets_14.-IMCS-2012_1.1.5-IMCS2012 (1)

http://slidepdf.com/reader/full/assets14-imcs-2012115-imcs2012-1 1/4

8/11/2019 assets_14.-IMCS-2012_1.1.5-IMCS2012 (1)

http://slidepdf.com/reader/full/assets14-imcs-2012115-imcs2012-1 2/4

 

Fig. 1. Resonance spectra for the coke-free (gray) and the coke loaded (about 10 g C  /100 g cat , black) fixed-bed.In the upper graph, the reflection parameter |S11| is shown over frequency. The lower graph displays the

transmission parameter |S12|.

be observed. Further details of this also calledcavity perturbation method can be found in [10].The coke deposits change the electricalproperties of the resonator filling, especially theconductivity [11]. These alterations lead tonoticeable changes of the resonant frequenciesand the attenuation of the waves.

Measurement results

To validate the resonance behavior of thereactor without catalyst, a simulation model hasbeen created using COMSOL Multiphysics.Therewith, the S-parameters are computed.The obtained spectra suit the measurementresults quite well. By this way, it is also possibleto investigate the influences of variousgeometric aspects on the measurementtechnique. The simulation results also matchanalytical calculations.

Static measurements of the reactor reveal an

extensive temperature dependence of thismethod. The resonance frequencies shift by 1.5to 2 % due to a temperature difference of540 °C.

Figure 1 shows the measured values of thereflection parameter |S11| and the transmissionparameter |S12| in the reactor over frequency inthe range from 3.5 to 5.1 GHz. The gray curveswere taken at the beginning of themeasurement, when the catalyst particles arecoke free. The black curves show the state after48 hours of coking, when the particles have

reached a carbon load of about 10 gC/100 gcat.Both parameters, |S11| and |S12|, reveal anobvious dependence on the coke loading. The

reflection parameter curve in the coke-free stateshows several peaks, each for a distinct mode.When the particles are fully coke loaded, onlyone peak is clearly visible. The transmissionparameter does not exhibit the separateresonance frequencies distinctly, but theparameter values decrease by approximately45 dB with coke loading.

To illustrate the effects of the coke deposits on|S11| in detail, figure 2 shows one peak of thespectrum for different coking times. In the firstfour hours no changes in the spectrum wereobserved. Thereafter, the resonance frequencynoticeably shifts towards lower values and theamplitude decreases with increasing time andaccordingly increasing coke loading. The lastcurve shown was taken after 24 hours of cokingwhich corresponds to a coke load ofapproximately 5 gC/100 gcat. Thereafter, nofurther change of the resonance frequency can

be observed.

The influence of the coke loading on thetransmission parameter |S12| is examinedparticularly as well. Therefore single peaks of|S12| are tracked over the coking time. Infigure 3 the attenuation of the peak occurring at5.05 GHz is plotted over the coking time. Afteran initial time of 5 hours, during which theamount of coke on the catalyst is too small tobe acquired with this measurement technique,the attenuation starts to increase. After42 hours of coking, the final value of -68 dB is

reached. Thereafter, the signal is overlaid withnoise, probably caused by the experimentalsetup.

DOI 10.5162/IMCS2012/1.1.5

IMCS 2012 – The 14th International Meeting on Chemical Sensors 77

8/11/2019 assets_14.-IMCS-2012_1.1.5-IMCS2012 (1)

http://slidepdf.com/reader/full/assets14-imcs-2012115-imcs2012-1 3/4

Hence,the coklonger ticoke creflectiowhy thmonitor 

electricalimit abothe fixed

The decarbonacan beregener the resorder. TcoincideThis fincomplet

using the tr ing procesime and than be det  paramete  level ofd to a ce

l conductivitve which no-bed is pos

ctivation oceous depo

burnt offtion of thenance behhe spectra

with theing provesly.

Fig. 2. A si 

Fig. 3. Att 

nsmissioncan be

rewith a laected, co  |S11|. Anoke loadinrtain amou

y of the cat wave propible.

the catalits is rever with oxygeoked fixed-vior chang

at the endnes beforethat the co

gle resonanc 

nuation of a

parameter |bserved foger amountpared tobvious rea  can only

nt is that

lyst exceedgation thro

st due toible. The cn. Duringbed in thises in rever of the burncoke loadi

ke is remo

e mode of |S

ingle peak o

12|r a

oftheonbethe

s agh

theke

theayed

-offng.ed

A

ThGfinre

R

[1]

[2]

[3]

1| for differen

 |S12 | over th

knowledge

e authors Arman Resancial supppectively.

ferences

C. Kern, A. JheterogenenTechnik  doi: 10.1002

N. Müller,detection ofparticles ansensors, Ap262 (2010);

S. OrdonThermogravideposits o

catalysts usThermochimdoi: 10.1016

 coking times

 coking time.

ments

.J. and R.Mearch Fouort (Je 257

ess, VerkokuKatalysator 

78, 8,cite.2006000

. Kern, R.coking and

fixed bedlied Catalysi oi: 10.1016/j.

z, H.metric det  alumina-su

d as hydrodca ActaS0040-6031(

 

.

. are indebtndation (D/12-2, Mo

ng und Koksen, Chemie

1033-104851

Moos, A. Jeregenerationreactors by A: General ,

.apcata.2010.

astre, F.rminationpported no

echlorination379, 25-3401)00598-6

ed to theFG) for060/5-2)

bbrand inIngenieur

(2006);

ss, Directof singleelectrical

382, 254-5.001

. Diez,of cokele metal

catalysts,(2001);

DOI 10.5162/IMCS2012/1.1.5

IMCS 2012 – The 14th International Meeting on Chemical Sensors 78

8/11/2019 assets_14.-IMCS-2012_1.1.5-IMCS2012 (1)

http://slidepdf.com/reader/full/assets14-imcs-2012115-imcs2012-1 4/4

[4] R. Moos, M. Wedemann, M. Spörl, S. Reiß,G. Fischerauer, Direct Catalyst Monitoring byElectrical Means: An Overview on PromisingNovel Principles, Topics in Catalysis, 52, 2035-2040 (2009); doi: 10.1007/s11244-009-9399-6

[5] S. Reiß, D. Schönauer, G. Hagen, G. Fischer-

auer, R. Moos, Monitoring the ammonia loadingof zeolite-based ammonia SCR catalysts by amicrowave method, Chemical Engineering andTechnology , 34, 791-796 (2011);doi: 10.1002/ceat.201000546

[6] G. Fischerauer, M. Förster, R. Moos, Sensing theSoot Load in Automotive Diesel Particulate Filtersby Microwave Methods, Measurement Scienceand Technology , 21, 035108 (2010);doi:10.1088/0957-0233/21/3/035108

[7] P. Fremerey, S. Reiß, A. Geupel, G. Fischerauer,R. Moos, Determination of the NOx Loading of an Automotive Lean NOx Trap by Directly Monitoringthe Electrical Properties of the Catalyst MaterialItself, Sensors, 11, 8261-8280 (2011);doi: 10.3390/s110908261

[8] R. Moos, Catalysts as Sensors - A PromisingNovel Approach in Automotive Exhaust Gas Aftertreatment, Sensors, 10, 6773-6787 (2010);doi: 10.3390/s100706773

[9] N. Müller, A. Jess, R. Moos, Direct detection ofcoke deposits on fixed bed catalysts by electricalsensors, Sensors and Actuators B: Chemical ,144, 437-442 (2010);doi: 10.1016/j.snb.2009.03.008

[10] G. Fischerauer, M. Spörl, A. Gollwitzer, M.

Wedemann, R. Moos, Catalyst State Observationvia the Perturbation of a Microwave CavityResonator, Frequenz , 62, 180-184 (2008);doi: 10.1515/FREQ.2008.62.7-8.180

[11] N. Müller, R. Moos, A. Jess, In-situ monitoring ofcoke deposits during coking and regeneration ofsolid catalysts by electrical impedance-basedsensors, Chemical Engineering and Technology ,33, 103-112 (2010); doi: 10.1002/ceat.200900380

DOI 10.5162/IMCS2012/1.1.5

IMCS 2012 – The 14th International Meeting on Chemical Sensors 79