Atmospheric non-thermal plasma has been recognized as a successful tool for the treatment of exhaust andwaste air for abatement of pollutants, air cleaning, and odor removal. To improve the efficiency of plasmacleaning processes a combination of plasma with adsorption filters and or catalysts is used. The fundamentalinteraction mechanisms of the surface, plasma, and gas remain the subject of further investigations of thedetails. A quantitative FTIR-spectroscopic study of the effect of surface barrier discharge on the adsorptionand desorption of ethanol at activated carbon is presented. First of all the adsorption and desorption werecharacterized for different input concentrations and different carrier gas flows without operating the plasma.After that the interaction of the plasma and the gas mixture was analyzed. Finally the activated carbon filter wasreinstalled behind the plasma stage and the loading and unloading of the activated carbon filter element wereinvestigated while burning the plasma. It is shown that the plasma stage initiates an additional decompositionof ethanol at the surface of the activated carbon because of the interaction of long-lived O3 with physicaladsorbed C2H5OH. The total amount of filtered C2H5OH does not change but the portion of decomposedC2H5OH that is not rinsed out increases. The plasma caused effect at the surface reaches 68% of the effect ofplasma in the gas phase.
The deodorization of air and the reduction of emissions fromwaste air and exhaust gas are becoming increasingly important issuesin industrialized society. Due to its impact on air, soil and waterexhaust pollutions affect the whole environment and thus humanhealth. Therefore environmental norms and standards are constantlyincreased by national and international authorities. The possibility touse atmospheric non-thermal plasmas for air-pollution control is wellknown [1–6]. Plasmas contain active and highly reactive species, inparticular electrons, ions, atoms, molecules, and radicals, which candecompose pollutant molecules and particulate matter in the gasphase (aerosols) and on surfaces. In most applications the problemis so complex that plasma technology has to be combined with classicfilter methods like adsorption and scrubbing to get satisfactoryresults. However, the fundamental interaction mechanisms of thesurface, plasma, and gas remain subject of further investigations ofthe details. This paper intends to contribute to clarifying aspects ofthe effect of non-thermal plasma on gas adsorption on surfaces. Inparticular a quantitative FTIR-spectroscopic study of the effect ofatmospheric surface barrier discharge on ethanol in the gas phaseand its adsorption and desorption at activated carbon (ac) will bepresented. Since this study intends to the inaction of active carbon
: +49 3834 554301.Brandenburg).
rights reserved.
and plasma species a VOC was selected which is known for purephysisorption. Ethanol fulfills this requirement and is easy to handlein time consuming experiments. Furthermore, ethanol is the mostabundant VOC emitted from corn silage and thus a major contributorto tropospheric ozone from agricultural sources [7]. The occupationalexposure limit of ethanol in Germany is 500 ppm [8].
2. Experiment
Fig. 1 shows a sketch of the experimental setup including ac filterelement, gas handling, power supply and measuring systems ofplasma power, temperature and humidity of gas and its composition.Synthetic air (N2/O2=80/20) enriched with ethanol was used ascarrier gas. This experiment was specifically designed to study theeffect of atmospheric pressure plasma on adsorption and desorptionof ethanol at activated carbon.
The setup consists of a box made of polymethyl methacrylate(PMMA) (volume: 13×43×13.5 cm³=7.5 L) containing a plate withseveral holes for gas distribution, the atmospheric pressure plasmastage and behind that the activated carbon filter element (ac filterelement). Atmospheric pressure plasma stage serves a surface barrierdischarge consisting of four dielectric plates (17×13 cm), each coveredwith a mesh electrode from both sides.
The ac filter element is a 5 cm long PMMA-tube with an innerdiameter of 1.1 cm. It is filled with 1.65 g of activated carbon particlesof cylindrical geometry (AFA-2-1200). The gas composition (if not
noted otherwise) was measured with a Fourier transform infraredspectrometer (Gasmet CR-2000 FTIR, ANSYCO GmbH) with a resolu-tion of 8 cm−1. The temperature and the humidity of the carrier gaswere measured using commercial sensors placed in front of theplasma stage and behind the ac filter element. The added bypassenables the direct measurement of the gas composition at the inletof the filter. All experiments were performed at room temperatureunder normal atmospheric pressure.
Firstly the flow characteristics of the filter were determined bymeasuring the breakthrough curves with an empty filter element for
time [ min ]
rela
tive
etha
nol c
once
ntra
tion
a
c
Fig. 2. Relative breakthrough curves (c(t)/cbypass(t)) for an input of ethanol concentrationa) without activated carbon/without plasma, b) with activated carbon/without plasma, c) w
different input concentrations of ethanol (500–2000 ppm) and differ-ent carrier gas flows (300–1000 L h−1) without operating the plasma.These experiments give the standards for the total amount of C2H5OHthat passed through the system to enable comparison with the subse-quentmeasurements. Fig. 2a shows the typical behavior for the emptyfilter element without plasma. The integral over time represents thetotal amount of C2H5OH that flowed through the system.
Then the ac filter element was then filled with activated carbon tostudy adsorption and desorption of ethanol without operating plasmafor the same concentrations and flows (Fig. 2b). The integral over
time [ min ]
b
d
of 1000 ppm, a volume flow of 300 L h−1, and a specific energy density of 46 J L−1
ithout activated carbon/with plasma, and d) with activated carbon/with plasma.
128 R. Basner et al. / Surface & Coatings Technology 234 (2013) 126–131
time represents the total amount of C2H5OH that passed the system.The difference between the integrated curves Fig. 2a and b shown inFig. 3b gives the total amount of C2H5OH adsorbed at the activatedcarbon inside the ac filter element.
Each experiment with the filled filter element was performedusing freshly activated carbon. The results agree with typical break-through curves expected for this type of classic filter with a type ILangmuir-isotherm characteristic for micro porous solids and onemonolayer of physical adsorbed gas [9,10]. However, the ac filterelement used in this study is extremely undersized which results inbreakthrough appearing immediately. The filter size was chosen toenable reasonable duration of loading and unloading (1–2 h). Theadsorbed mass of ethanol equals the desorbed mass if the loaded acfilter element is rinsed out with synthetic air. Therefore any chemicalreaction between ethanol and activated carbon can be excluded. Thefiltration process is therefore a purely physical way of attachment.
The experimentswithout operating plasma showed the best stabil-ity and reproducibility for a flow rate of 300 L h−1 and an inputconcentration of ethanol of 1000 ppm. Therefore this combination ofparameters was selected to study the plasma impact. A loading timeof 49 min was kept constant to compare the different measurementsregarding the total amount of C2H5OH.
The effect that the atmospheric pressure surface plasma stage hadon the breakthrough curves of ethanol was determined both with(Fig. 2d) and without (Fig. 2c) activated carbon. The resulting areasrepresenting the amount of filtered C2H5OH are plotted in Fig. 3cand d respectively.
Furthermore the products of chemical reactions were identifiedand their concentration was measured under plasma operation. Thefollowing four different cases were considered: 1. interaction of plas-ma with synthetic air, 2. interaction with synthetic air and activatedcarbon, 3. interactionwith synthetic air and ethanol, and 4. interaction
time [ min ]
re
lative
eth
an
ol co
nce
ntr
atio
n
d
b
Fig. 3. Differences of the break through curves (s. Fig. 2) representing the amount of ethanolactivated carbon and decomposed by the plasma.
with synthetic air, ethanol, and activated carbon. In each case thespecific energy density SED (plasma power per flow rate) was varied.
As shown in Fig. 1 the high voltage power supply consists of afrequency converter (CHROMA 61603) and a transformer (BREMERT40B). The current was measured via a shunt (50 Ω) while the highvoltage was measured by means of high-voltage probe (TektronixP6015A). Both signals were analyzed with an oscilloscope (TektronixTDS 3052) and the multiplication of averaged current and voltageslope divided by the shunt value gives the power dissipated into theplasma P. The values were carefully checked with the method ofso-called Lissajous-Figure (Voltage–Charge plots). To realize corre-sponding P in the range of 0.3 to 3.9 W in this setup, the frequency ofthe discharge voltage needs to be varied from 100 to 500 kHz sincethe amplitude can be varied up to 10 kVpp. The plasma power givenabove corresponds to a range of SED from 3.6 to 47 J L−1.
3. Results and discussion
Because of inelastic collisions of free electrons with gas moleculesthe plasma generates ions and reactive species which can initiatechemical reactions in the gas phase and/or at surrounding surfaces.Using pure synthetic air N2 and O2 can be excited, dissociated orionized followed by the production of nitrogen oxides (NxOy) andozone (O3) [11,12].
Since atomicNandOare not detectablewith FTIR technique themainproduct that was found is O3. Furthermore additional measurementsusing a higher resolution spectrometer (0.5 cm−1) (FTIR-SpectrometerThermo-Fisher-Antaris IGS) yield nitrous oxide (N2O) [12] with aconcentration of 14 ppm at fixed parameters: flux=300 L h−1, f=100 Hz, Upp=10 kV, SED=46 J L−1.
For the first case, interaction of the plasma with synthetic air,furnishes the concentration of ozone given in Fig. 4 as a function of
time [ min ]
c
: b) adsorbed at the activated carbon, c) decomposed by the plasma, d) adsorbed at the
SED, which is available for the interaction with ethanol and/or acti-vated carbon under the respective combination of parameters. Theconcentration increases linearly with the SED independent from theway of its variation.
After filling the filter element with activated carbon (case 2) theinteraction of the plasma generated O3 and N2O with the surface ofthe activated carbon was analyzed. The N2O passes the activatedcarbon without any interaction. Its concentrations before and behindthe ac filter element are identical. Behind the ac filter element carbonmonoxide (CO, 40 ppm at SED=46 J L−1) and carbon dioxide (CO2,50 ppm at SED=46 J L−1) were detected. An immediate increase ofthe CO and CO2 signals was measured behind the ac filter elementafter the ignition of plasma accompanied with O3 production. Incontrast the concentration of O3 behind the ac filter element increasesslowly with time. The CO and CO2 are a result of the reaction betweenO3 and the carbon [13]. Since it is known that activated carbon acts asa filter for ozone the temporal behavior of the O3 concentration at theoutlet represents its breakthrough curve of the used undersized acfilter element.
The next case (case 3) includes the interaction of the plasma stagewith the mixture of synthetic air, ethanol, and an empty filter element.The plasma interacts with ethanol directly inside the plasma gap and atthe surface of the electrodes/dielectric barriers. We note that additionalspecies other than ozone can be formed. These include atomic oxygen(O*) and hydroxyl radicals (OH*) [5]. Although the decomposition ofethanol by these species cannot be excluded, these cannot be discussedas the experiment (FTIR-analysis) is not designed to allow theirmeasurement. However using the FTIR-Spectrometer Thermo-Fisher-Antaris IGS shows that the addition of ethanol (CC2H5OH=1000 ppm,other parameters see before) leads to the formation of acetaldehyde(C2H4O, about 100 ppm at SED=46 J L−1) as a new reaction productbehind the plasma stage. It is a by-product of the incomplete oxidationof ethanol with ozone [14]. The dehydrogenization from hydrocarbons
0 10 20 30 40 50
200
400
600
800
1000
ozon
e co
ncen
trai
on [
ppm
]
SED [ Jl-1 ]
af = 500 HzUss = 8, 9, 10 kV
0 10 20 30 40 50
200
400
600
800
1000
ozon
e co
ncen
trat
ion
[ ppm
]
SED [ Jl-1]
bf = 100, 300, 500 HzU
ss = 10 kV
Fig. 4. Concentration of ozone as a function of SED a) constant frequency, b) constantamplitude.
by means of OH* or O* is known to result in the oxidation of aldehydes,and less frequently acetic acid [14]. However such or other stableby-products whichmight result from the reaction of ozone with decom-position products, e.g. higher oxidation stages of nitrous oxide specieswere not detected by FTIR. CO and CO2 are also generatedwith a concen-tration of about 50 ppm each in the presence of ethanol. Water contentwas also observed however this could not be quantified due to thepresence of a large unstable background signal. Direct decompositionof C2H5OH is believed to occur as switching on the plasma stage resultsin the concentration of C2H5OH increasing by 40%. This result correlateswith the observed increase of C2H4O, CO, CO2 and the decrease of O3 to60% of the initial concentration of 700 ppm.
We conclude that the decomposition of C2H5OH occurs essentiallyby the reaction with O3. Intermediate products of an incompletedecomposition (C2H4O) as well as final products of a completedecomposition (CO, CO2, H2O) were found. It is noted that the com-plete decomposition corresponds to a complete non-thermal oxida-tion by the plasma.
Fig. 5 summarizes quantitatively the results of the plasma effect onthe decomposition of C2H5OH as a function of SED (plasma—blackcurves). The mass of decomposed C2H5OH increases linearly (noteerror bars) with SED in the considered range independent from theway of the variation of the SED. The other curves (plasma+ac—red bul-lets, plasma+ac—blue stars) will be discussed later on in the paper.
The final experiments include the entire system consisting ofplasma stage and ac filter element in a mixture of synthetic air withethanol (case 4). FTIR measurements made using the higher resolu-tion present no new products at the filter outlet. In the space betweenthe plasma stage and the filter element comparable concentration ofC2H5OH, O3, C2H4O, CO, CO2 and N2O was detected as in the case ofan empty filter element. However, behind the activated carbon theconcentration of C2H4O is doubled immediately after switching onthe plasma. This indicates that the physically adsorbed C2H5OH at
0 10 20 30 40 500.00
0.05
0.10
0.15
0.20
plasma plasma + ac / measured plasma + ac / calculated
b
b
a
f = 100, 300, 500 HzU
ss = 10 kV,
cons
umed
eth
anol
mas
s [ g
]
SED [ Jl-1 ]
0 10 20 30 40 500.00
0.05
0.10
0.15
0.20
plasma plasma + ac / measured plasma + ac / calculated
af = 500 Hz, U
ss = 8, 9, 10 kV
cons
umed
eth
anol
mas
s [ g
]
SED [ Jl-1 ]
Fig. 5. Decomposed and filtered mass of ethanol (measured: black, red bullets; calcu-lated: blue stars) as a function of SED a) at constant frequency and b) at constantamplitude.
130 R. Basner et al. / Surface & Coatings Technology 234 (2013) 126–131
the surface of the activated carbon is also decomposed by the longlived ozone. After the breakthrough the concentration of C2H5OH atthe outlet of the filter increases rapidly to 60% of the initial concentra-tion at the beginning of the loading (Fig. 2d). At this time the O3
concentration of 73 ppm is 11% of the O3 generated in the initialplasma generated case and 19% of the input value for the ac filterelement respectively.
The O3 input concentration to the ac filter element differs from theinitial plasma generated value by the amount that has already beenconsumed by the decomposition of C2H5OHbefore the acfilter element.In order to demonstrate the consumption of O3 the breakthrough curveof ozone was analyzed 35 and 75 min after the breakthrough for aSED=46 J L−1.
The results regarding the portions of the entire amount of O3 arepresented in Fig. 6. Also, about 35% of the produced O3 are consumedfor the decomposition of C2H5OH before the filter element (O3/C2H5OH)after 35 and 70 min. The interaction of plasma/ozone with C2H5OHhere is independent of the load condition of the ac filter element.
In contrast the portion directly consumed at the activated carbon(O3/ac) increases over the course of the experiment and the portion ofthe slippage increases according to the breakthrough curve. Therefore,a second indirect consumption of O3 (O3/ac/C2H5OH) exists besidesO3/ac. This represents the interaction of O3 with the adsorbed ethanol.The portion (O3/ac/C2H5OH) increases in the course of the experiment.That is caused by the increased amount of physical adsorbed C2H5OH atthe surface of the activated carbon.
The sum of direct and indirect consumption of O3 decreases from50% to 44%. That will reduce until the saturation of the ac filterelement is reached due to the limited filtering effect.
All these considerations require a constant chemical adsorption ofO3 at the surface of the activated carbon independent of the admix-ture of C2H5OH. The measurement in case 4 cannot distinguishbetween O3/ac and O3/ac/C2H5OH. Therefore the presented portionof direct consumption represents an upper limit and the portion ofindirect consumption a lower limit.
The analysis of the breakthrough curves of C2H5OH for the com-plete filter (Fig. 2, d) yields the total amount of consumed C2H5OH(Fig. 3d) consisting of the 3 portions: 1. decomposed by plasma/O3,
3 2 5
47.1 % (O 3/ac)
2.9 %
O3/ac/C2H5OH
a
O3/ac/C2 5OHH
35.3 % O 3/C2H5OH)
38.2 % (O 3/ac)
5.9 %
20.6 % slippage35.3 % O 3 2 5
b
35.3 % O 3 2 5 14.7 % slippage35.3 % O 3/C2H5OH)
Fig. 6. Ozone balance after: a) 35 min and b) 70 min load time at SED=46 J L−1.
2. adsorbed by activated carbon, and 3. decomposed by O3 at thesurface of the activated carbon. Fig. 5 shows the results as a functionof SED together with the effect using only the plasma. The measuredtotal amount of consumed C2H5OH as well as the pure plasma effectincrease with the SED independent from the way of its variationwith nearly the same slope. The improvement of the filtering effectwith increasing SED seems to be caused by the increased impact ofthe plasma on C2H5OH in the gas phase.
The integral values of the differences of breakthrough curves ofethanol with ac filter element but without plasma (example givenin Fig. 3b) give the portion of physically adsorbed C2H5OH. Thisportion was added with the pure plasma effect (example given inFig. 3c). These calculated values are also presented in Fig. 5 andare in reasonable agreement with the measured sum (examplegiven in Fig. 3d). In contradiction to the results regarding the bal-ance of O3 (Fig. 6) there is no indication of the decomposition ofC2H5OH at the surface of the activated carbon. To answer this ques-tion a complete balance of C2H5OH was drawn. The results arepresented in Fig. 7 for two operating states of the plasma stage forthe complete system until reaching the termination criterion(49 min loading).
The explanation and theway of calculation of the pillars representingthe different portions of consumption of C2H5OH are indicated in thelegend. The blue and magenta pillars give again the results presentedin Fig. 5. The main results shown in Fig. 7 are:
1. Without plasma 24% (II green) of the C2H5OH input value for thecomplete system (I red) is physically adsorbed at the activatedcarbon after 49 min.
2. The pure plasma/O3 effect of decomposition (III blue) increaseswith increasing SED linearly up to 12% (also shown in Fig. 5).
3. The ethanol mass reduced by plasma/O3 and filtered by activatedcarbon (IV magenta) increases with increasing SED linearly asalready discussed (compare Fig. 5).
4. The C2H5OH input value for the ac filter element (V cyan) decreaseslinearly with increasing SED by the amount of decomposed C2H5OHbecause of the plasma/O3 effect alone (difference: I− III).
5. The amount of physically adsorbed C2H5OH (VI brown) decreasedlinearly with increasing SED due to the decreased C2H5OH inputvalue (item 4).
6. The O3 effect of decomposition at the surface of the activated carbon(VIII yellow) increases with increasing SED up to 8% (difference:VII−VI).
4. Conclusions
We present a quantitative study of the effect of an atmospheric sur-face barrier discharge plasma on the adsorption and desorption of eth-anol at activated carbon (AFA-2-1200), analyzed by FTIR-spectroscopicmeasurements. Plasma generated ozone and ethanol as the modelcontaminant (VOC) were found to be the dominant reactants. Theconsumption/decomposition pathways were determined. The sig-nificantly new result is that the plasma stage initiates an additionaldecomposition of ethanol at the surface of the activated carbon be-cause of the interaction of long-lived O3 with physically adsorbedC2H5OH. This confirms the results concerning the balance of O3 andgives an explanation for the distinct measured increase of acetalde-hyde (C2H4O) concentration behind the ac filter element. The pro-cess is accompanied by an increase of the amount of physicallyadsorbed C2H5OH. The total amount of filtered C2H5OH does notchange but the portion of decomposed C2H5OH that is not rinsedout increases. The plasma generated long-lived O3 at the surfacereaches 68% of the effect of plasma in the gas phase at a SED of46 J L−1.
500Hz/8kV 500Hz/9kV 500Hz/10kV0.0
0.1
0.2
0.3
0.4
0.5a
Eth
anol
mas
s [ g
]
3.6 25.9 47 SED Jl-1
100Hz/10kV 300Hz/10kV 500Hz/10kV0.0
0.1
0.2
0.3
0.4
0.5b
Eth
anol
mas
s [ g
]
9 27 47 SED [Jl-1 ]
I C2H5OH total input
II physical adsorbed of I
III pure plasma/O3 of I
IV plasma/O3 and ac of I
V
VI
VII physical adsorbed + O3
effect at ac(IV -III)
VIII O3 effect at ac (VII - VI)
V C2H
5OH input for ac (I -III)
VI physical adsorbed of V
Fig. 7. Distribution of ethanol mass: a) at constant frequency and b) at constant amplitude.
Partly supported by the transnational project “PlasTEP—Dissemination and fostering of plasma based technological innovationfor environment protection in the Baltic Sea region” (part-financed bythe European Union–European Regional Development Fund).
References
[1] J.S. Chang, Plasma Sources Sci. Technol. 17 (2008) 045004 (6pp.).[2] H.-H. Kim, Plasma Process. Polym. 1 (2004) 91.[3] T. Hammer, Contrib. Plasma Phys. 39 (5) (1999) 441.[4] A.M. Vandenbroucke, R. Morent, N. De Geyter, C. Leys, J. Hazard. Mater. 195 (2011) 30.[5] S. Schmid, M.C. Jecklin, R. Zenobi, Chemosphere 79 (2010) 124.
[6] J. van Durme, J. Dewulf, W. Sysmans, C. Leys, H. van Langenhove, Chemosphere 68(2007) 1821.
[7] F. Montes, S.D. Hafner, C.A. Rotz, F. Mitloehner, Atmos. Environ. 44 (2010) 1987.[8] Technische Regeln für Gefahrstoffe 900 (Technical guidelines for hazardous com-
pounds), 2010. (GMBI 2010 Nr. 3).[9] Compendium of Chemical Terminology. IUPAC RECOMMENDATIONS, 2 ed.,
Blackwell Science, Oxford u.a, 1997.[10] F. Rouquerol, J. Rouquerol, K. Sing, Adsorption by Powders & Porous Solids:
Principles, Methodology and Applications, 1 ed. Academic Press, 1999.[11] U. Kogelschatz, Plasma Chem. Plasma Process. 23 (1) (2003) 1.[12] U. Kogelschatz, B. Eliasson, M. Hirth, Ozone Sci. Eng. 10 (1988) 367.[13] J. Grundmann, S. Müller, R.-J. Zahn, Plasma Chem. Plasma Process. 25 (5) (2005) 455.[14] E. Rischbieter, Ozonierung von Alkenen in Alkoholen als Lösungsmittel. Technische
![[2009]Fast Removal of Malachite Green by Adsorption on Rice Husk Activated Carbon](https://static.documents.pub/doc/80x56/55cf9da4550346d033ae87d5/2009fast-removal-of-malachite-green-by-adsorption-on-rice-husk-activated.jpg)
