2. $E%$m EJ, chemical hydrides ?S%%R@486&P~'JA@'E%T@ ? (20%)

3. $E#@&2i@ EJ, chemical hydrides fris#g$@@?%!&!@F& ? (20%)'

4. 2 0 I H Polypyrrole %I BTEX 2FdrnbR'%f3l'FJ ? M%M%@H9BP%@? I

5. (20% ] $29 IGC $PS@EWH Polypyrrole ZSE$ EBffEZFdrnMB %@ ? @@%M%rn%4@BP%l'FJ?

6. (10% ] 11,6fi@?j?Y@@$ (3 C6-C9 benzene, toluene, ethylbenzene, and o-, m-, p- xylene %ISE$ EBdEtFdWM%E@ ?

Inlcmalianal Journal of Hydmgen Energy 33 (2W8) 360-365 - ww-+&&icrbdo;uuijhydene .

Wf+- Chemical hydrides: A solution to high capacity hydrogen storage and supply

I I Rajesh B. Bin iwa lea** , S. Rayalu", S. Devottaa, M. Ichikawab ,

'Nalio,zol E!tvirnnmer!lnl Engi,neering Rerearth Inrlilulc. Ntlaru Mary. Nogpur 440320, Itrdia b ~ a ~ a l y r i r Rereach Center, Holikido U,tib,erriry. Supporn. Japan

~Lceivul 19 June ZW7: accspld 9 Jvly 2007 ' Available online 12 Ssplembcr 2007

I Abslract

Cycloalkanes are good candidate media for hydrogen slorage (6.5 wt% and 6062 kg~?/m3). A novel approach for h e supply of hydrogen, lhrough liqaid organic hydrides (LOH) using catalytic reaction pair of dehydrogenation of cycloalkanes and hydrogenation of corresponding amnutics is a useful process for supply of hydrogen. Hydrogenation of amnutics is relvively well-established prows. However. h e efions z e needed to develop efficient catalyst for dehydmgenalion of cycloalkanes. In chis paper we review h e dehydrogenation of cycloalkanes as useful reaction for storage of hydmgen in chemieal hydnda. 0 2007 InlemaLional Association for Hydrogen Energy. Published by Elsevier Ltd. All righu reserved.

i Kqvordr: H y d m g ~ storage; Liquid organic hydridcs; Dcl ivq to fuelling swlion; Cyclodkmes

1. Introduction

Onboard storage and supply of hydrogen for fuel cell is an important issue in the successful application of PEMFC. Also the transportation of Hz from production facilities to the fu- elling station needs to be considered for successful application of hydrogen economy L1.21. Hydrogen being very flammable gas its tmnsport involves several safety issues. The safety is- sues are related to the lower and higher explosion limits for Hz concentration in air, low ignition energy in air and requirement of high pressure (typically 30C-S00psi) storage or some times cryogenic temperatures, if it is tmnsported in liquid form. These problems can beovercome if the hydrogen is either adsorbed on some materials such as carbon nano-tubes or chemically stabi- lized such as in the case of metal hydrides or alanates. Several studies report the development of hydrogen storage materials such as metal hydrides (3-61, Mg-based alloys [7,81, carbon materials [P-161, boron compounds [17], chemical hydrides [IB], etc. While developing such hydrogen storage materials, capacity of the material in terms of weight and volume basis is an imponant factor to be considered. W ~ t h a limited capacity it would result in weight penalty and C02 emissions associated

- Canupauding author. Tel.: +919822745768. E-mail oddrerc rb-bidwde@nccri.rrrin (R.B. Binimle).

with transportation. Also the adsorption and desorption kinet- ics is to be sufficiently fast to provide continuous Hz supply. Another important requisite is to trmpo? the hydrogen con- taining media at close to atmospheric temperature and pressure conditions m d possibly using the existing infrastructure such as lonies.

Hydrogen containing chemicals which are useful for storage of hydrogen include methanol, ammonia and cycloalkanes. At STP all these are in liquid phase and therefore provide advan- tage of possibility of using existing infrastructure being used for gasoline. The hydrogen slorage capacities of these chem- ical hydrides may range in the scale ofI6-8wt%. Earlier it was thought that since the chemical storage method is non- reversible, the compounds cannot be used in cycles. In view of this supply of hydrogen through liquid organic hydrides GOHI using cataly tic reaction pais of dehydrogenation of cycloalkanes such as methylcyclohexane, cyclohexane and decalin; and hy- drogenation of corresponding aromatics is, a useful process for supply of hydrogen to PEMFC. This is one of the most promis- ing methods to store. transport and supply with in situ genem- tion of hydrogen due to several advantages associated with this system which include; COX free hydrogen at fuelling stations. reversible catalytic reactions, recyclable reactants and products and relatively high hydrogen contents (6-8% on weight ba- sis and about 6 0 6 2 kg ~ 2 / m ] on volume basis). Particularly,

0360-31991S-see Imnt mvler O 1007 lntcmaiond AsrocuJLin Ior Hydmgen doi:10.1016lj.ijhydcnc2W7.07.0~

Energy. Published by Elrcvicr LUI. All righU reserved.

Hydrogen Generation , b*lHz 30 I

I

Fig. I. Schematic of hydrogen rtomge a d supply system.

the appmach is most useful to transport the hydrogen fmm cen- nalized genemtion facilities to fuelling stations. A schematic of catalytic reaction pair of hydrogenation of toluene and d e hydrogenation of methylcyclohexane for hydrogen storage and supply is depicted in Fig. 1. Several studies have been reported on cycloalkanes as cattier for hydrogen and particulady de- velopment of dehydrogenation catalysts. The impottant issues being addressed by these repons are dehydrogenation temper- ature, dehydrogenation rates for various cycloalkanes, hydro- gen evolution rates and reactor type for efficient heat transfer. Considering the development in this important field addressing hydrogen storage which is of current interest, this paperwas envisaged to review state-of-art and to identify the future scope of research.

2. Cycloakanes for hydrogen storage

Seveml cycloalkanes may be used as hydrogen cattier as LOH include cyclohexane, methylcyclohexane, teralin, decalin. cyclohexylbenzene, bicyclohexyl. I-methyldecalin, etc. A few reactions of dehydrogenation of cycloalkanes are depicted in Fig. 2. One mole of cycloalkane has potential to transport 3 4 m o l of hydrogen. The endothermic energy requirement for these reactions is in the range of 6469 id per mole of H2 this is much lower than energy that could be obtained by oxidation of Hz. 248kUmol. Hydrogen stomge capacities of cycloalka- nes and other storage media are compared in the Table I along with boiling and melting points. Due to high boiling points of cycloalkanes, the present infrastructure such as oil tankers and tank lorries can be used for. the long-term stomge and long- distance transpoltation of hydrogen in the form of LOH.

3. Catalysts for dehydrogenation of cycloalkanes

Fig. 2. Various polenlivl cycloalhes with hyGmgen vtonge capacily and endolhmnic energy requirement.

h b l e I Hydrogen rwngc capacity of various nledh m d heir physical pmpenies

Storage media Hydmgcn conlenl 1 Boling Melting ' point ('C) point ('Cl

~ 1 % mom:

Cyclohuac Muhylcyclohcxmc Tcvalin cirdeealin rmnrDccalin cyclohuylbsrvc~ie bicyclohexyl cir.~yra-I-melhyldecalin rmru-onri-I-mclh~lddin LiH + Hz0 L i B b +4H10

I A 6.2 - - Noble metal catalysts particularly Pt and bimetallic catalysts

MmN4~Alo.uCoo.uH1.r 1.2 4.8 - - Pt-M (M = second metal) are well reported for highly se- (wilh lIW I ) - -32.9 - lective dehydrogenation of cycloalkanes. Dehydrogenation of

methylcyclohexane over a Pt-Rdalumina catalyst in the pres- ence of added hydrogen has been reported [19]. Pt and Pt-Re catalyst has been reporled for dehydrogenation of methylcyclo- hexane to produce hydrogen [201. Dehydrogenation of methyl- cyclohexane was reporled by Newson et al. [21I, n u b e and Taube [221, Klvana et al. [23] and Griinenfelder and Schucan [24] for seasonal storage of energy through storage and supply of hydrogen using Pt-based catalysts. Various supports used for Ptcatalystsinclude alumina, activatedcarbon and anodized alu- minum. Catalytic decalin dehydrogeno-aromatization for long- term storage and long-distance transportation of hydrogen us- ing catalyst WC. Pt-IrIC and Pt-W/C [25,26]. The addition of second metals such as Mo, W, Re, Rh, Ir and Pd on the carbon-supported Pt catalysts reported to result in enhanced dehydrogenation rate due to the promotion of C-H bond cleav- age and/or desorption of aromatic products. A physical mixture of PUAC and PdIAC catalysts exhibits higher activities than the monometallic WAC catalyst owing to the synergistic ef- fects of spillover, migration, and recombination of hydrogen over Pt and Pd catalysts [27]. In an attempt of designing cata- lysts using less Pt or non-noble metal catalysts Ni-based cata- lysts have been reported. These include NiC. RulC and Ni-Ru [28] and Ni-Pt catalysts for dehydrogenation reactions. Partic- ularly Ni-PUAC catalysts has shown highly selective dehydro- genation of cyclohexane with minimum use of Pt Selectivity

I towards hydrogenolysis was observed to bp only 0.8-1.2% with addition of O.Swt% Pt in 20wt% Ni [2?]. Dehydrogenation of methylcyclohexane has been reported gn K-Pt/Al2O3 cata- lyst at relatively higher temperature of 32,0°C [30]. It appears from the literature available that the most active and selective catalyst for dehydrogenation of cycloalkpes is Pt. However, the bimetallic catalysts containing a small amount of Pt have promising potential for cost-effective catalysts.

The dehydrogenation is not a shape selective reaction and it is reported that when well-dispersed metal.is present as catalyst then is the prevailing reaction as compared to hydrogenolysis. Activated carbon with high surface area is good support for well-dispersed metal catalysts. Funher, absence of any acid sites on carbon support prevent cleavage of C-H bonds thus reduees the formation of carbon and in tun impmves catalyst stability.

4. Comparison of various dehydrogenation catalysts

Hydrogen evolution rates reported over various catalysts un- der different reaction conditions and temperature are compared in Table 2. Most of the catalysts reported are Pt-based catalysts and reactions are canied out at temperatures ranging £ram 210-350°C. Activated carbon in granule and cloth Corm, alu- mina and alumite are the supports used Tor catalysts. The most

Table ? Hvdmccn evolution ntui remaled for vvious cvcloalbcs ova different wmlyru

ReacLlnt Cdyru Temp. (OC) H, ewlution nlc Rwclor symem' Ref. (mmal/g-/ min)

Cyclohuvle 5 wt% WAC 82 0.034 huh-wire [Jll Cyclohzane 3.82~1% PUAC 300 1800 Batch-wire Cross-ref of [32] Cyclahuvle Zwt%PtlAlumina 300 910 Balch-wise Cmrs-ref of 1321 Mclhyleyclohuue 3.82w1% WAC 300 17W Bauh-wise Cmss-ref of [32] Decnlin 3.82~1% WAC 300 610 Barch-wise Cmss-ref of 1321 Cyclohunnc 2wl%PUAluminr 315 Z9 Flaw-system I ~271 Cyclohuvle 10 wl% WAC clalh 260 98 Flow-system ' L321 Cyclohuvle 10 wl% WAC clalh 330 510 Flow-rynem 1321 Muhylcyclohuvlc IOwt%PUAC clalh 298 520 Flow-system 1321 Decalin 1 0 ~ 1 % WAC cloth 320 460 Flow-system 1321 Cycl~huvlc PLldumile (P13g/m3) 300 1600 Flaw-sybtcm 1321 Cyclohuu!c PUalumL (P13g/m11 350 3800 Flaw-ryrtcm 1321 Cyclohuan I?wt%Pt-Rh/AC clolh 331 520 Flow-system 1321 Cycloh-e I I wl% Pt-ReIAC cloth 328 550 Fluw-ryvcm 1321 Cyclohunne 5 wt% PUAC 235 98 Spnypulud system 1291 Cyclohunnc lOwl%NilAC 3W 7.1 Spmy-pulud syslcm 1291 Cyclohuue 20wt%Ni/AC 300 8 5 Spny-pulwd synem 1291 Cyclohuane 40w146NUAC 300 6.8 Spny-pulrcd system 1291 Cyclohuanc 20 wt%Nif0.5wt%R/AC 300 13.1 Spny-pulsed ry&lcm 1291 Teualin 5 wt% PUAC 210 103 Supcrhwlcd liquid film 1281 Tcudin 5 wt% PUAC 240 3 0 Supcrhwted liquid film 1281 Decalin 5 wl%PUC 210 1235 Liquid hlm 1251 D d i n PcW/C (5 wt% metal) 210 4231 Liquid hlm Decalin Pt-LrlC (5 wt%mclal) 210 25.2 Liquid film 1251 Teualin Pt-LrIC (5 w t 8 mc~al) 210 19.6 Liquid film 1251 Decalin 5wt46PUC 210 555 Liquid film [%I Decalin 5wt46PUC 210 . 833 Liquid film 1x1 Decalin 5wt%WC 210 160 Liquid film 1251 Dsill in Swt%WC 210 441.4 Liquid hlm 1251 Mclhylcyclahu~nc 0.1wt%Kf0.6wl%R/AI~O~ 320 744 (mmol/L,/ m i d Axed bed 1301 -

I KB. BUIIIvale et aL/lmrrnot8onol Journal

active dehydrogenation system reportedis in flow system under spray-pulsed injection of cyclohexane over Pt/alumite catalysts with hydrogen evolution rate of 3800mmol/g/min. Whereas, in batch mode hydrogen evolution rate of 1800 mmol/g/min is re- ported for 3.82wt% Pt/ AC at 300°C. Under the spray-pulsed continuous reactor system again WAC is having enhanced hy- drogen evolution mte of 98 mmol/g/min as compared to NiAC caialysts. However, from the economic point of view it would be advantageous to use Ni-based catalysts. Reaction of decalin under the liquid film condition over 5 wt%Pt/C has exhibited hydrogen evolution rate of 444mmoVdmin which is relatively high at temperature of 210°C. When Pt-IrIC (Pr/Ir = 4) or Pt-WIC (Pt/W = I) catalysts were used under the same liq- uid film conditions it has shown hydrogen evolution of 25.2 and 42.31 mmol/g/min at 210°C [25]. More recently consid- ering higher rate of dehydrogenation, tetralin was used as hy- drogen storage medium and dehydrogenation was carried out over 5 wt% W C under superheated liquid film condition with hydrogen evolution rates of 103 and 25Ommol/g/min at 210 and 240°C, respectively [281. Another recent publication on K-PtfAlz03 catalyst has reported hydrogen evolution rate of 744mmol/ m i n / L , using fixed bed system [30].

Position - b

Ax. 3. Thermal pmfils of h e camlysl surfact under (a) wwd and ib) dry condilions.

5. Reaction conditions for effective dehydrogenation

Due to the endothermic nature of the dehydrogenation of cyclic hydrocarbons chemical equilibrium is favored at higher temperature; the reactions are performed at relatively high tem- perature under steady-state operations in gas phase. Hodoshima et al. proposed a liquid-film state condition and conducted de- hydrogenation of decalin over carbon suppoaed Pt-based cat- alysts [25,26.28]. Kameyama and coworkers have used a non- steady spray-pulse mode for dehydrogenation of 2-propanol on Pt/a-A1203 [31]. Ichikawa and coworkers have reported high hydrogen production rates using wet-dry multiphase con- ditions in batch mode L27.321 and using spray-pulse reactor during dehydrogenation of cyclohexane and decalin over car- bon supported Pt. In another publication same group has re- ported highly efficient production of hydrogen without COz emission in the dehydrogenation of cyclic hydrocarbons under a non-steady spray-pulse operation over supported Pr and Pt-M (?d = Re, Rh, Pd) catalysts. Cyclohexane, methylcyclohexane, tetlalin and decalin were dehydrogenated by the Pt-containing catalysts supported on thin active carbon cloth sheets and alu- mite (anodized aluminum) plates. Reduction rate of hydrogen under the spray-pulse mode is reported to be higher than the conventional batch-type liquid phase reaction and the steady state gas phase reaction in the flow system. Production rate of hydrogen was dependent on the rate of reactant feed, the re- action temperature and the support. Retardation by products adsorbed on the catalysts was negligible under the spray-pulse operation.

Hodoshima et al. [28] reported efficient hydrogen supply from tetralin with superheated liquid-film-type catalysis and using catalysts namely NiC, Ru/C and Ni-Ru for operating fuel cells.

The heat transfer to the catalyst and providing required en- dothermic energy is a key issue for higher catalytic activity in terms of higher hydrogen evolution mtes. Biniwale et al. has studied thermal profile of catalysts surface under spray-pulsed injection of cyclohexane over Pr catalysts supported on acti- vated carbon and alumite. Spny-pulsed injection of reactant re- sulted in higher tempentures periodica!ly on the catalysts sur- face as revealed from the temperature profiles shown in Fig. 3 This helped in maintaining higher catalytic activity.

6. Patents available

Several patents have appeared claiming use of chemical hydrides including various solid hydrides, liquid hydrides and panicularly LOH. Solid hydrides such as NaB&. LiB&, KB&. RbB& are reported for storage of hydrogen effec- tively [33]. The chemical hydride hydrdgen generatian system here uses the solute, which is selected from the group consist- ine of NaBH4. L i H 4 . KBH4, RbBH4. A similar method of

and storing~hydro&n bas been disclosed in world patent WO 01/5l410 usini a chemical hydride solution, such

NaB& [34]. As described in this chemical hydride reacts with water in the presence of a catalyst to generate hydrogen. The drawback of these methods is relatively low hydrogen content as compared to possible storage in cycloalka- nes and irreversible dehydrogenation leading to no-recycling of the hydrides.

Another class of chemical hydrides reported in various prior arts is liquid organic hydride. Hydrogen &olution during dehy- drogenation of methylcyclohexane, decalin, dicyclohexyl and cyclohexane to toluene, naphthalene, biphenyl and benzene,

respectively, in the presence of iddium-based molecular com- plex catalyst is disclosed [35].

The storage and release of hydrogen by means of a sub- stantially reversible catalytic hydrogenation of extended pi- conjugated substrates is described in US Patent 20050002857 [361. The pi-conjugated substrates include large polycyclic aromatic hydrocarbons, polycyclic aromatic hydrocarbons with nitrogen heteroatoms, polycyclic aromatic hydrocarbons with oxygen heteroatoms, polycyclic aromalic hydrocarbons with alkyl, alkoxy, nitrile, ketone, ether or polyether sub- stituenu, pi-conjugated molecules comprising five membered rings, pi-conjugated molecules comprising six and five mem- bered rings with nitrogen or oxygen heleroatoms, and extended pi-conjugated organic polymers. The hydrogen storage capac- ity obtained was in the range of 14.7% by weight which is relatively low for economical use of these chemicals.

Use of benzene. toluene, xylene, mesitylene, naphthalene, anthracene, biphenyl, phenanthrene and their alkyl derivatives as possible aromatic substrates for producing hydrogen for fuel cells bas been reported in Japanese palenu [37,38]. Whereas, in another Japanese patent [391 hydrogen storage and supply system is described with catalysl conlaining at least one metal selected fmm Ni, Pd, Pt, Rh, Ir. Ru, Mo, Re, W, V, Os, Cr, Co and Fe having good activity for both hydrogenation of hydrogen slorage body comprising an aromatic compound and dehydrogenation of 'a hydrogen supply body comprising +e hydrogenated derivative of aromatic compound. A Japanese patent P2002134141 describes the equipment used for &hy- drogenation of LOH with arrangement for intermiltent supply of prescribed amount of liquid organic hydride and a product separator for separating hydrogen fmm other producu.

It is apparent, from increasing patent applications in the field of chemical hydrides for storage of hydrogen, that the method has potential for practical applications.

7. Challenges in development of chemical hydrides

Chemical hydrides are definitely potential candidate as hy- drogen storage and supply material. However, development of catalyst for selective and efficient dehydrogenation, particularly in case of cycloalhes, needs to be pursued. catalyst develop- ment also requited to target catalytic-reactor configuration to overcome heat tmnsfer limitations for this strongly endother- mic reaction. Recycling of chemical hydrides as in case of aromatic-cycloalhes pairs would be advantageous for reuse of materials; however, the economics has to be worked out for transportation cosJ in both ways and the weight penalty. Alter- natively at the place of delivery of hydrogen, aromatics can be supplied to local chemical industries as solvenu.

8. Conclusions

LOH, aromatics-cycloalhes pairs are potential candidate for high capacity hydrogen storage. Since these can be trans- posed at near ambient conditions may be most useful ior long- distance transportation of hydrogen. The endothermic reaction of dehydrogenationcan be carriedout effectively with advanced

reactor systems such as spray-pulsed injection forcreating alter- nate wet and dry conditions over catalyst surface. Development of efficient dehydrogenation catalyst is important particularly considering the cost of Pt-based catalysts. At present the ef- forts are apparent from recent publications for using minimum amounl of Pt for dehydrogenation of cycloalkanes. As LOH can be recycled or altemalively aromatics can be supplied as solvent to chemical industries. this method of hydrogen storage and supply would be economically feasible.

References

I11 Ananthachar V, Duffy 11. Efficiencies of hydrogen storage syswnr onboard fuel cell vehicles. Solar Eacrgy 2005;78:687.

121 David 6 An ovcrview of advanced malerials for hydrogen sloragc I M a w Rocerr Techno1 2005;1623:169.

131 S i g h BK. Singh AK. Srivasuva ON. Impmvcd hydrogen sorption characteristics in FeIll,Mm material. In1 I Hydrogen Energy 1996; 21(2):111.

141 Lavoche M. SmcNTal and thermodynamic pmpenies of metallic hydddu used for cncrgy slorage. I Phys Chcm Solids U)W,65:517.

151 K. Kogurc Y, N a h o r i Y, Orimo S. Reversible hydddiog and dehydriding reactions of pemvskilc-tw hyddde NaMgH3. Scr Matcrialia 2005:53(31:319.

161 Muthukumar P. kt&h Mdya M, Srinivass Mlnhy S. Expcrimcnu on a metal hydridc.blsed hydrogen norage device. l a J Hydrogcn E n c w

~ - .~ 2005; 30 (15): 1569.

[7J Singh AK. Singh AK, Srivasuva ON. On thc synth- of the MglNi alloy by mechanical alloying. I Alloys Compd 1995;227:63.

181 Vijay R, Sundensan R Maiya MP. Srinivasa Murthy S. Campamivc evaluation of Mg-Ni hydrogen absorbing maledalr prepared by mechanical alloying. In1 I Hydrogcn Energy 2005;30501.

191 Sandmck G. A panadmic ovcrvicw of hydrogcn sloragc alloys fmm a gas rcadoo p i n t of view. J Alloys Compd 1999,293-5:877.

[LO] Simonyan VV, lohmoo JK. Hydrogen slotage in sarbon nanolukc and graphitic nanofibm. J Alloys Compd ZWZ: 330-2659.

[Ill Hirssher M. Bechrr M. Ha& M. Quintel A, S W o v a V. Choi YM el al. Hydrogen slorage in carboo nanormcrurcr. I Alloys Campd2002: 330-2654.

1121 Darkrim FL. Malbrvnot P. T m g t i a GP. Rc)icw of hydrogcn n o n g c b) jdrarplion in sybon nulowkc. In1 J H)dmgen Encrg) 2002:27:193.

It31 Sanl;uanM.LWakirdvanA.G-ac R VenuvanalinaamP. ViruanalhanB. . . H e m m substituted carbon nanotubes: Can 1h;y k the astivating cenlcrs for hydrogcn adsorption. Bull Cat Soc India 2M)2;L:167.

1141 Wswanathan B. Sadman M, Scibioh MA. Carbon nanomalerials--ax thcv aoomoriaE candidates for hvdroecn rtoraecl Bull Cat Soc India . " zoi3;i:iz.'

"

1151 Becher M, Hauka M, Huschsr M. Quinlcl A. Skakalova V, Weglikowrka UD s t al. Hydrogcn storage in carbon nanotuks. C R Phys 2M)3:4:1055.

1161 Fujii H, Shin 0. Hydrogcn smragc pmpeniv, in nanc-smclvrd magncrium- and carbomrclaLd materials. Physica B:Condenrcd Matter 2003;328:77.

[I71 Fakioglu E, Yurvm Y, Vwkoglu R.J. A review of hydrogcn noragc syslcms based on boron and irr compounds. Int I Hydrogen Eaergy 20W;29:1371.

1181 Kojima Y, Kawai Y. N h i r h i H. Marrumom S. Campresscd hydrogen generation using chcmical hydride. I Power Sourscr 2004;135(1-2):36.

1191 loull~imurugeran K. Bhatia S, Srivasuva RD. Kinetics of &hydro. genatioo of methylcyclohuanc over a platinUm-menium-alumina calalysl in the p w o c c of added hydrogch Ind Eng Chcm Fundam 1985;24(4):433.

1201 Coughlin RW. KawaLami K. H- A. Activiry, yield paucms, and coking khavior of Fl and FlRt. caulysrr dwing dehydrogenation of mcthylcyclohuane: I. In the absence of sulfur. J Catll 1984;88:150.

:?I1 Ali IK. Newson El. Rippin DWT. Exceeding equilibrium conversion with a catalytic membrane reactor for the dehydrogenation of mcthylcyclohuanc. Chcm Eng Sci 1994,49(13):2129.

'06:Z6Z!SWZ U'D V 140 lddv 'sllaa l a y S ' u p d o JOJ ? r i m odil-uqg-pmb!~ palaqladns q~!m u!,enr molj qddnr uoS'qiq ~)uo!ag~yla 'S n q n v q 'N wm!H 'S em!qmPoH l8zl

~ a : c ~ ! m o z u a ~ v piq [ddv :.suop!puoa anqdplnm .+la.%, lapun w w a a ~ p ~ e uo puoddns n r i p u a tupy~lom->d n ~ o v o l a i a pmb![ molj u a a q i q jo uop!ma loa!am .w eflTqa1 'V eqorynd 'N dw ILz1

' S~ I :~Z :EOOZ finus s a 8 q i n I IY 'slam" IFPW IOJ m o o m a q i q e rn ~d m p m a q i q a m p q q d t d p g ~ m % q i L p p T=Q ?~.Weg 'S n m w ~ 'L wags 'V a- 's rqrlsopoe lgzl

aslnd Kq m p n p pue o m q o l a i a Bu!sn n o p m p d uaS'qiq l m ~ g l s 'L61:8z!mOz h q 'J=%U~KH rial .uaS'qKq JO uo!nvodnren 'w e m q q 'A 0109 s~ olobenqe~ eme%qn 'y q o r y n ~ 'N e i m RE] inmsp-auol pm ~ ~ Z W S ~ l - a u o l 19 u o p e q e m o r e a m ~ q i q a p '61EI:PF6661 !4 8 q M q 3 n o p e o a 8 q i q a ~ q y pmb!~ IOJ uopendo qlenp s g K ~ e a adL~-m~y-p!nb!i 'S nmxnwh 'v a s o w 'S em!rlsopo~ l n l aspd iwdr e jo podold v .H m i q 'y 010mmh 'I Wseieqq Ilcl

'8KI:IE!900Z i m o 3 m 8 q i ~ 1 lul poqlam ~ q i q ppnp a~nre&o u! uoIIpJaua% uaaorpiq IOJ l s i p ~ m u o p ~ a a q i q ~ JO l m w d a p ~ a a 'S runjxs!t.l 'S opohn'3 aqemeM 's !Ig 'O ! ~ q s o h lor1

'Es:(z-llsol:sooz 117 Peg ld jo ionom IPS e jo uopgpe i q i ~ ! ~ p a n q luomaaueqm pue 101anzu paslnd imdr Su!m uoqm pzle~pre uo pauoddns nsKpp= paseq !N nno aueraqol~h 30 u o p u a ~ q k q a a .I nreww 'y o v q o ~ 'm apm!ma 1621

Investigation of the relationship between surface thermodynamics of the chemically synthesized polypyrrole films and their gas-sensing

responses to BTEX compounds C.W. Lin*, Y.L. Liu, R. Thangamuthu

D w r ~ ~ ~ o f ~ M E n ~ n ~ ~ ~ & ( c N a i i i d Yvnlin Unlver r i ryu f& i rmemd~~hnu I~~ Irullln. fiiwon. ROC

Rsdced 18 Osbbrr 2002; accepld 18 F c b q 2001

Abrknct

Cl--dopedpolypynule (PQCI) 6 h as senringmaterials were preppred chemically uing F a , asan oxidanland wcre uedto detect benzenq tolumq ethylbazme,and wlmc (BTEX) cnmwunds iu theirvaporphnrcz. vhich were fovnd to be abla to cnbanw the doping l ~ e l of PA.Q and hmce i n a d the mduetiritv 0iPA.Q ma e&n& to them k isomers of xvlmc. i.c anho 10). ie ta (m). md p& @), wcre also investigVcd for comp&on. Tb;c r&itiviti& of PPya s-r exposed to BM c o i p o ~ wilh &ious concmmtiam were m w e d and found lo Lio iu the mgr of 0.08mlUppm (bclumc)-OBmWppm (o.xylcne). An adrorpU~ model bard M tho hpmuir i s o h correlated w d l with the qcrimmral remlrs and w s used to interpret tho scming behavior of P f i a sensor towards B% vama The adsomtion equilibrium c&wanrs. K.. wera thereiore able to be determine acwrdim to this madel A iuthcr study of h i & between the Lrcs te i B 7 M compound and&= P4.a suriace was c&ed out by a method oiim- gss chmmatognpby (IGC) with PfiCI-coaled gLus beds pack4 in a column as 'bo SUUMW phue and the inhrted BTEX compound as the mobils phase. The refention volumes for these compovnda were meanved to be iu the iollowing order: o-xylms > rn-xylene z D-xvleno clhvlbmzcnc > toluene. This madtude s & r i c e war melatedwell with that o i the e-kibrium &tantskL. %he free

1. Introduction

It is well know that electmn<ondueting polymen can be used as chemical senson by measuring the conductivity changes as a function of secondary doping or undoping of detected s p i e s [l-G]. Polypymle (P@) was initially used to measnrethererponrelo ammoniavapor by Nylanderet al. [I] in 1983. Lam, the similar device was used to measure t h e q o w s lo othu gases [ 2 4 ] under suitable conditions; p r d l y these gases reacted with P@ by oxidizing and reducing the P@, respectively. In order to improve the characteristics of PPy film, several PPy-based composiles

of a monolayer, the thiohess of the sensing film, the ad- sorption equilibrium cannant as well as the change of site redsrance. We utilized this miaoxopic adrarptim model to explain the sensing behaviors of the conducting polyum s-rs and also endeavored tn make correlation between sensing parameters and the experimental results to give an inlerpreiation for the different q o n s e s .

Recently, the attention of various air pollution hntrol agencies has been inmasingly focused on the cone01 of volatile organic compounds (VOCs). Considerable quanti- ties of VOCs are p r o d u d from indlLFtrial sources such as printing and coating ficilities, foundries, electronics, and

as gas semors had been studied [7-12j. H m g el a. 1131 paint &ufacturin&u. A group of VOC2, including aro- have propored a microscopic w s e n s i n ~ model indicating matic hydrocatbans such as benzene, toluene, ethylbenzene. that the iasitivity of the s&& depends& the site number and xyiene (BTEX), are widely used in ind&.&and

serious advase e f f a u on the quality of air. 'lhe utiliiation -

' -ding a u h TeI.: +886-5541-2601x4613; tn +886-5431-2071. E-nu78 d h s : l i p ~ ~ ~ . o d u l w (C.W. lh).

. . of conducting polymers as gas sensors for detecting these highly health-threatening VOCs may be a promising ap- poach We have previously presented the resulu of sensing

0915-IOOIIO3R - rco Lion, rmlur O 2W3 UrMcr ~ ; r m c e B.V. All "ghu rcuneb ~ o ~ : I O . I O I ~ ~ ~ L ( ~ ~ O ~ ~ O ~ ) W ~ ~ ~ ~ ~

m p m e s of elecuochemically prepared PeCI04 senrm cxposed to BTFX compounds [14]. Tbe dect of adsorp tion f i i t y and the undoping efficiency could be distin- euished and calculated in terms of adsomtian muilibrium " constaut We obtained a good correlation between the ad- sorption equilibrium co-t and the measured sensitivity and concluded that the ffiniw of tho interested commund to the sensing layer dominateh the sensing response & our work.

Although the development of gas sensors based on the or- ganic conducting polymers is plausible, the exact nature of the inleiactions between the analyie molecules and the con- duc t i ngpo lymcr su~ remain l&ly unknom. Identifica- tionof the intaactim undemone at the sensor-ms interlice is w i d forthe design ands~lectionoffuture s&or materi- als. In the past various technisues 115-221 have beenused to investigate &ces of c~du&~-ma te r i~s . Amongst these techniques, inverse gas chromatography (IGC) appem to be particularly useful for determining the sorption properties of molecules on swfice Chehimi et al. [21,22] investigated surfaces of conducting materials using IGC with packed column and repaned that the surface to be amphotcric, i.e. mlar molecules w s acids and bases) adsarbed to a &ter degree of the s m l h than to non-pblar molecules, and the preparation condition of the polymer played a vilal role in the surface thermodynamics of the polymer.

The oresent work is a continuation ofour earlier studies on -~~

the development of gas sensors based an organic conducting polymer films, and is devoted to investigalo tho gas-sensing behaviors of chemically prepared C1--doped polypyrrole ( P ~ C O sensor exposed to BTEX compounds. BGde thiE, we also intend to present some preliminary results obtained using IGC technique subjected & the relati&ship of the sur- fice interaction force and the s e h g response.

2. Experimental

2.1. Pmpnratim of PPy sensing l q r

An interdigirated gold electrode saeen-printed onto the d a c e of analumina substrate was used. its smcture and dimensinns were the same as in the earlia repons [I 1,121. Chaniresistors were fibrieated by deposirian of these elec- modes with PPv thin film. PolvmerLation of wrrolo was carried out by a chemical oxidation using ~ e ~ i i a s an ox- idant Tbe alumina subsuate with interdigitated Au elec- mode was immersed into the FeCb solnlion for 1.0 h and then immersed in pynole solution for polymerization for 3Omin.

2.2, Memremenr of resirfance change of P e C I

Tbe resistance chan~e ofPPvcI film due to tho exmsure

the cment-voltage (1-0 melu and an automatic data ac- suisitian svstem Resistancemeasurements were laken when the response was maximum and constant AU BTEX; sam- ples were obtained h m Aldrich (HPLC grade) and used without any further purificatim.

2.3. Inverse gas chmmnfogrqhy

2.3.1. Pm~armion ofstafionmv ohare afIGC Glass b k s (60mhes) we<& t o h y &out0.5 g were

ulhasonic cleaued in a water bath conlaircine mn-ianic sur- - fictaut This weight was suitable for packing a 2 m long col- umn with 118 in. in diameter. The pre-cleaned glass pads were then i m m d in 14mM Feel3 oxidant solutio? and were slightly &ed for about 1 h to make FcCl3 be ad- sorbed homogeneously on the glass beads. Tbe glass beads were then m f e r r e d into a beaker containing pre-distilled pynole monoma kept at a constant tempa- ature of 6°C and were gently stirred for 12h for chemi- cal oxidation in solution Tbe PPy-coated glass beads were then d e d with methanol to remove cxcess oxidant and vacuum-dried at 45°C for 24 h

23.2. Measuremm~s of IGC dafa Measurements were carried out with a China Chromatoe- -

raphy 8900 gas chromatograph fined with a flame ionization detector. High purity grade niuogen was used as the car- rier gas. Tho flow ravi was 62mlmin-t as measured by a soapbubble flow meter. Oven tempemre was varied h m 423 to 443 K. The carrier gas flow ralo was adjusted at each temperauro to maintain a-constant flow mte bver the tem- perature range sudied The column inlet pressure (Pi) was 1.60 bar and the outlet pressure (Po) was atmosphere. The injector and detector zones were kept at 463 K to minimize temperature gradients in these mas. The stalionary phase was condirioned for 8 h prior to chromatorrraohic measure- - . ments. Robes were injected manually at least in triplicate by a @+tight Hamilton syringe. To achieve exmeme dilu- tion of the probes, the syringe w u purged as many times as necessary. Tbe signals were recorded with a digilal recorder and the retenrim times were calculated graphically [23] and m t e d the dead time to for the column using air as an non-inkractin~ m a r k

Nan-polar samples for a dispersion interactian refepce slate included C6-G n-alkanes were obtained h m Aldrich (HPLC grado) i d u&d without any further purification. All other BTEX samoles are same as those d in the sensitiviw measurements.

U3. Data annlysis In IGC, mcaslrements were carried out at infinite probe

dilution, i.e. appmximately lao sorbart surface coverage. w e that theierults ob&ed h m the experiments apply exclusively to adsorbate-adsorbent interactions. The reten-

of BTFX compounds-(with ~2 mixhue with level 99.99%) were measured using admice mainly consisting of

tion time i f the probe, I,, the time it takes the probe to uavel through the column, can be convened lo the net retenlion

38 CH! Lin u ~/./&JZJM md

volume VN, by [241

VN =JF(t, - to) = JFtn

where F is the earrier gas flow rate. I, the net retention time, and J the correction factor of pressure gradient for mobile phase

The value of J was calculated lo be 0.7625 in the present smdy. The net retention volume, VN is related to the Gee energy of adsorption, AG. (Jmol-') by [25]

where R is the gas constant, T the column tempemture, and C, a constant which takes inlo account the weight and spe- cific sudace area of the packing material and the standard states of the probes in the mobile and the adsorbed states [26]. Manipulation of AG, (or RTln(v~)) data for the \rari- ous probes leads to the evaluation of y,d, the dispersive con- tribution to the surface energy of the materials.

2.3.4. Dispersive corrhibulior~ lo llre swjoce energy A standard method to evaluate y,d, the dispersive compo-

nent of the surface energy, relies on the retention data of the n-alkane series [27]. Practically for these probes, AG, or simply RTln(v~) wlues are plotted against nc, the number of carbon atoms in the n-alkane molecules, and the slope of the linear correlation refers to A G ' ~ ~ , the h e en& of adsorption of a methylene group.

where Nis the Avogadro number, al is the cross-sectional area of an adsorbed CH2 group (6A2), and yCHl is the surface energy of the solid containing only methylene groups such as polyethylene (yCH1= 36.8 - 0.0581 (OC)) [27].

2.3.5. Spec& inlemclions If both dispersive and specific interactions are operative

at the gdsolid interface, it is assumed that they conhbute to the tad AGp in a additive manner:

AG, = A G ~ + AG: (5)

where 'd' and 's' are rererred to dispersive and specific in- teractions, respectively. Several approaches were proposed to distinguish between AG; and AG: [28]. In the present investigation, we use the kaditionat approach of Brook- man and Sawyer [29] in which AGn or RTln(v~) values are related to the boiling point of the injected probes. The n-alkanes series lead to a reference linear correlation, and for polar probes interacting specifically with stationnry phase, or RTln(VN) wlues deviate from the linear correlation de- fined by the n-alkanes. For a given polar probe, the veitical distance Gum the rererence h e yields:

-AG: = R T ~ ~ ( V N / V N , ~ ~ ) (6)

$4H : F~~ Auuarorr B 94 (2003) 3&15 i

where VN,rer is the net retention voluhe of a hypothetical n-alkane that boils at the same temperature as the rest polar probe.

2.3.6. Applicolion of solvolion equolions lo IGC dolo From earlier research, solvation equations using sample,

or solute parameters as descriptors ein be applied to the eorrelation and predietion of adsorptibn phenomena The phenomena may inelude speeifie retention volume, log(Vg) [30,31], or gas-liquid partition coefficients [32-341 on vat- ious adsorbents. The general linear solvation energy rela- tionship (LSER) equation can be described as follows:

log(SP) = c + rR2 +SIT? + o$ + bp? + I log(~I6) (7)

where SP is the experimental data, R2 t+e solute excess mo- larrehction that reflects the ability of solute to interaet with sorbent thmugh rr- and 11-eleckon pairs; 4 is the solute dipolatity-polarisability, a measure of the dipole-dipole or dipole-induced dipole interdctions; a: and the eKective hydrogen-bond acidity and basicity of the solute and L16 the solute gas-lied partition coefficient of hexadecane at 298Y a measure of dispersion interactions. All the param- eters, except R2, are derived ffom thermodynamical data of partitioning andlor complexation equilibria The conshnts c, r, s, o, b, aud 1 are found by multiple linear regression anal- ysis. Eq. (7) allows the s u r k to be described in terms of specific chemical propeliies. Consequently, slmightfonvard interpretation of the IGC data can be applied in terms of the surface chemisky of PPyCl. The statistical analysis and the LSER descriptors calculations were canied out with minitab sofhvare.

3. Results and discussion

. 3.1. lnlernclion mechnnirms of elechvn-donolinggnses with PPyCl

P@ is a n-type semiconductor when doped with CI-. Therefore, the exposure of elecbon-donating gases such as ammonia and alcohol, to P@CI causes an increase in con- ductance, i.e. a decrease in resistance. The recovery of the conducting polymer to its initial oxidized state is ascribed to desorption, by flushing with N2, of the nucleophiles. 'The interaction between PPyCl and eleckon-donating compound is generally considered to be a doping effect. Thus. p-type dopants can increase the doping level of the polymer chain by enhancing the effect of the original dopanL

3.2. Responrer ofPPyCl sensor lo BTEYcompoundr

o-Xylene is an elemon-donating compound, when ex- posed which causes an increase in the conductivity of the chemically prepared P@Cl sensor film As a gas-sensing test, we used o w f f cycles at 67ppm concentration of o-xylene to study the gas-sensing behavior of PPyCl film.

I 0-xylms roorenlratlon - 67 ppm

1 0 1 2 3 I

T h o x 10: beel

Fig 1. Several o w 6 ~ycler lo study the repmducibility of thc scnring behavior of PPyQ film la o-xylens vapor at 67ppm

As can be seen in Fig. 1, a reasonable reproducibility was obtained. However, a minor drift of baseline was ob- served. As shown in Fig. 2, the resistance change (AR) becomes greater as the o-xylene concentration increased in steps fiom 20 to 60ppm in N2. Also at a-xylene con- centration of M)ppm, the sensor responded as quickly as in about 60s. The response times were about 200 to 400s for concentrations lower thyl GOppm In con!xast, vapors of higher concentration (60ppm) took longer time (585 s) to recover fiom the oxidized stale. A plot of resishnce changes of PPyCl sensor vs. o-xylene concentration was given in Fig. 3. It is clear &om the plot that the resistance change is linearly proportional to the o-xylene concentra- tion The slope of the linear plot represents the sensitivity of the sensor characterized as the resistance change per unit o-xylene concentration The sensitivity of the PPyCI sensor to o-xylene vapor was 0.8 mf2lppm

Ethylbenzene is another eleclrondqnating compound which also causes an increase in the cpnductivity of the PPyCI sensor. As a gas-sensing test we used on-off cy- cles at 90ppm concentration of ethylbenzene to study the gas-sensing behavior of PPyCI film As can be seen in Fig. 4, a reasonable reproducibility can be noticed How- ever, a small drift of baseline was obsenled. As shown in Fig. 5. the signal of resistance change becomes greater as the ethylbenzene concentration increased in steps h m 20 to 90ppm The response times for ethylbenzene were ranged fiom 180 to 460s. Compared to'o-xylene (Fig. 2), t the response speed of ethylbenzene (Fig.,S) was very slow. At the same time, compared to recovej speeds of ethyl- benzene, desorption of o-xylene fiom PPyCI film was much slower manifesting the interaction between the o-xylene was stronger than that with ethylbenzene. This observa- tion may imply that o-xylene had a better affinity towards PPyCI film The resistance changes of PPyCI sensor plot- ted against ethylbenzene concentmtion are shown in Fig. 6. It can be seen that resistance change varies linearly with ethylbenzene concentration. A sensitivity of 0.4 mfllppm was obtained fiom the slope of he linear plot.

With PPyCI sensor, a similar procedure was followed for the detection of vapors of other two xylene isomers tumely m-xylene and pxylene. As shown in Fig. 7, he resistance changes are linearly proportional to concenmions of these two isomers. The sensitivities obtained h m the linear plots were 0.6 mUppm for both m-xylene andp-xylene. Similarly, the sensitivity obtained fmm linear plot of AR vs. concentra- tion for benzene vapor (not shown in Fig. 7) in the concen- tration range of 200-500ppm was 0.08 mf21ppm The sen- sitivities obtained with PPyCI sensor for BTEX compounds

-0. on I I I I I 1 0 400 800 1200 1600 2000

Time, [sec] Fig. 2 Change in I&- AR c a d by increaring o-xylcn. c~nscnmtion in stcpr fmm 20 to 60ppm in N2.

0 - 2 0 - g, -0.02

3 5 3 a m CI -0.04 V) .r(

V) s

o-xy lene

-

-

gns off

Fig. 3. Change in ds~g .ce , AR ofthe P P N sauor plotted against o-xylens uuuenkation

are in Lhe following order:

o-xylene > m-xylene

9 s.01 2 p-xylene > ethylbenzene > toluene > benzene

3.3. Calculations of the parameters for the gar-senring model

ethylhenme concentrsUon = 90 pprn

.O.M Hwang et al. [I31 have previously proposed a microscopic

o I I 3 4 s gas-sensing model to explain the behaviors of PPy-based sensors to electrondonating compounds. The overall resis-

~ l r n o x lO'[srr] tance of the composite film ean be regarded as the paralleling Fig. 4. S N ~ o w f f cycler to study the reproducibility of the sensing of sevelal pseudo-monolayers and each layer is composed behavior ofPPyCl film to sthylbmzcnc vapor at mppm. of several resistors in series. In this model, R, r, n, and m

Time, C.s.e~l Fig. 5. Change in resistance, AR caused by hcmshg etbylbarrme cnnccnkadon in steps fmm 20 to 9Oppm in NI.

I I CU! Lln ol al./Smw.q md Acrvolorr B 94 (2003) 36-45 1 41

SLllliS~b -OW04 olm&n

ethylbuurnc

0 m-xylem

-120 o 20 40 60 80 loo o on1 o.m onu acn aos a06

Ethylbparpne chnmthlh, W CAB t p ~ ' l Fig 6. -gc in rerisranC4 AR Or Pw vnrrrr PIou Fig. 8. I/AR vr. 1/Cm fmPPyCI thin film upoac? to tolucns, slhylbcn- slhylbmrme UrnCUiilation zens, 0-aylene, and m-aylene n p m .

represents resistance of a monolayer, resistance of an active site, thickness of the thin film, and the number of active sites on a pseudo-monolayer, respectively. It manifests that the plot of reciprocal of the resistance change against the recip- rocal of gas concentration is a l i n w relationship according to the following equation based on the proposed model:

where AR is the resistance difference after and before gas sorption, Km the adsorption equilibrium cons tan^ CAO the concentration of the detected gas, rl and ro are the site resistance as the site is vacant and occupied, respectively. The value of [m(rl-ro)]ln can be detedned kom the re- ciprocal of the intercept and Km can be obtained by di- viding the intercept by the slope. It had been previously reported that this model interprets well the behaviors of PPy-PEO [1.1] and PPy-PVA [I?] composite sensors ex- posed to ethanol vapors by comparing the experimental re- sults. An increase in the polymerization charge usually led to a thicker sensing film, i.e. a greater n-value, and there-

Ti. 7. Change in rerktancq A.7 of lhc P e e l wruar planed against the cancortnfion of m-aylme andpxylcnc

fore decreased [m(rl-rp)]ln values. However, the thickness (determined by the total electrical charge) was kept con- stant in this investigation. Fig. 8 shows the plot of IlAR vs. ~/CAO for the PPyCl thin film exposed to toluene, ethylben- zene, o-xylene, and nl-xylene wpor, respectively. The values of [m(r~-ro)]ln and Km are given in Table I for baKene, toluene, ethylbenzene, o-xylene, m-xylene andp-xylene. Km value was dependent on the affinity of the detected eam- pound to the sensing material (PPyCl) and found to fol- low the order. o-xylene(l.97 x lo-') s m-xylene (1.07 x lo-') s p-xylene(7.45 x lo4) s ethylba~ene(6.50 x lo4) s toluene(Z.50 x lo4) s benzene(l.72 x As the values of m (number of active sites on a monolayer) and n (number of layers proportional to the thickness of the film) were ideally assumed to be identical (is. polymeriza- tion conditions and poly&tion period), therefore, the affinity of vapor to the conducting material film determine the value of the sorption equilibrium constant Km. Accord- ingly, o-xylene was the one having the highest &ty to conducting PPyCI film among the investigated BTEX com- pounds. The results of response time and recovery speedalso indicate similar phenomena that o-xylene vapor had higher affinity. It can also be noticed that there is a eorrelation be- tween the Km value and the sensitivity. With increasing Km value, the sensitivity increases linwly in the lowerKm value region and at higherK, values, sensitivity increased sharply

Table 1 lie valuer of sensitivity, K. and m(r1- royn for BTEX vapors detected by PPyCl smsor

Vapm Scnritivity K, m(n - rayn (n) WPP)

Bmrene 8 x lo-' 1.72 x lo-' -0.44 Toluens 2 x lo4 2.50 x 104 -0.91 Ethyltmzne 4 x 104 6.50 x I@ -0.70 0-Xylene 8 x lo4 1.97 x 10-1 -0.44 m-Xylarc 6 x lo4 1.07 x 10-1 -0.54 P X Y I ~ 6 x lo4 7.45 x 104 -0.71

I I 42 CII! Lin u aL/Smom ond Ac~arors B 94 (2003) 36-43 !

0 5 6 7 8 9 10

number of carbon $tom#

Fib 9. Plot o i K , ( x 1 ~ ) vs. smsitivity ( x LO4) br BTF.X wpms dclcctcd Fig. 10. RTW'N) vs. number of carbon atom piof for Aorptim or by PPyCl rcnsor. n-alkanes on PPy.

as shown in Fig. 9. The sensing abiliry of the conducting polymer film depended on both the secondary doping (or undoping) level by the detected compound and the affinity of the compound to the sensing material. Since there was no regular change in [m(rl -ra)]ln wlues with sensitiviry, it can be concluded that in the present case the sensing abiliry of the conducting PPyCl polymer frLn is dominated by the affinity of the compound to the sensing material rather thnn the secondary doping (or undoping) level by the detected compound In order to prove his, the rest of work deals with the study of affinity in lerms of the thermodynamic parame- ters between the chcmically preparedPPyC1 film and BTEX compounds using IGC technique.

3.4. Inverse gas ci~mmofography

RT~(VN) values for n-allme homologous series ad- sorbed onPPyCl at three temperatures are given in Tzble 2. Fig. 10 shows plob of RT~(VN) vs. nc, the numbcr of carbon atoms in the alkane molecule at 423, 433, and 443K Excellent linear plols are generated showing &at the adsorption energy for the n - ahuw increases with the number of carbon atoms in the chain. The slope of linear

Table 2 Mrarvrcd due ofRTln(t'~) lm n-abxu A d e d onto P@CI at lhm tempaatlmr

Tmperahlrs (K) Robes R T l n ( v ~ ) &J/-l)

443 n-He-e 3.168 n-Heptans 6393 n-Wane 9.821 n-Nonane 12.479

433 n-Hudllc 4.789 n-Hcpm 7.740 n-Wane 11.086 n-Nonanc 14.403

421 n-Heme 6386 n-Hcpm 9327 +&tans 12.687 n-Nonane 16300

plot is refer to hGCH1, the h e energy of adsorption of a methylene group (see Section 2.3.3). It can be seen fiom the following Eqs. (9K11) that in the present study the AG"~ obtained at three temperatures are varied fiom 3.14 to 3.31 d m m 2 whichclosely agree with thereported values of 2.6-5.01dm-~ [35].

423K: RTln(v~) = 3.310311~ - 13.653 (11)

The AGCH~ values can be used to calculate yp values. The results of this calculation are given in, Table 3. values are decreased with increase in temperature. This is due to increase in the kinetic e n e m of the adsorbed molecules on the surhce which eause deaeae in the surface energy. y,d value of 74.731dm-~ for PPyCl at 423K shows &at the PPyCi is a high surface-energy material compared to conventional polymers [35]. Thus, PPyCl have dispersive properties which lie between those for conventional poly- mers and high surface-energy materials such as metals and graphite. It can obviously be expected a strong interactions between the high surface-energy PPyCl adsorbent and polar adsorbate. Therefore such high surface-energy conducting polymer coatings can advantageously be used to fabricate sensors with better sensitivity.

Using the following thermodynamic relation, from the slope of linear plot between vs. T, the dispersive compo- nent of enthalpy change AH, and entmpy change AS, can be caleulated

Tale 3 Dispersion componmt of s d e e -y lor PPy de&cd at thrse temperatm

Tmperabm (K)

443 433 . 423

Y? (dim1) 69.96 72.10 74.73

I I C% Lin or ol . /Smw.~ ond Acruaon B 94 (2003) 36-42 43

. .

4 A bcnzsne ~ i g . 12. plot of K, (XI@) vl. ~ p o i f i o ahmption mugy for BTEX

vapors. -Dirpedn

0 bxlinc Table 5

310 360 4W 440 Valuer o f various probes panme+.cn [I61

B o b g tcmpcrnturc, 1x1 prober log(V~) ~1 n E:a EB lo&'')

Fig. 11. Plot ofRTln(V~) vr. the boiling point for n - h r and BTU: prober adsorbed onto PPyCl at 443 K .

In the pment study, the value of enthalpy change, AH. is 175.45 mJ m-l and entropy change, AS. is 0.2383 m-l K-I.

3.5. Adro?ptionfme energies ofthe pmbes

Fig. I I demonstrates RTL~(VN) vs. boiling point plot for PPyCl at 443 K. The n-alkane series lead to a linear plot which constitutes a reference for London dispersive intkmc- tions. The marker corresponding to the BTEX compounds nre located off the baseline defined by the n-alkanes, thus indicating that PPyCl behaves amphotically. As mentioned above, the separation of the marker From the baseline was referred to the specific interaction between the interested BTEX compound and the PPyCl adsorbent and the dis- tance corresponds to -AGi. The adsorption Free energies of BTEX compounds are given in Table 4. As can be seen, the ncgative valucs of AG. show an order of o-xylene > m-xylene > p-xylene E ethylbenme > toluene. It can also be noticed that the dispersive component contributed largely to the total h e energy of adsorption On the other hand, the con(ribution of specifie intenetion is not signifi- cant. Howevcr, the specifie tcrm also showed a similar or- der of o-xylene > rn-xylcne > ethylbcnzene Z p-xylene > toluene.

Fig. 12 illustrates the variation of adsorption equihbrium constant, Km with -AG:. It can be noticed an excellent lin- ear correlation between Km and -AG: i.e. specific interac-

B u u a e Toluae my- *Xylcne m-Xylmc P X Y ~ W C6 CI C8 C9 M r h M l E h o l Watcr

tion between the BTEX compouods andPPyClsurface. This obswdtion is an experimental pmof to our earlier conclu- sion from the resistance measurements lhat the a5inity (i.e. the coverage) of the dctected vapor to the sensing material dominated the sensitivity of the gas sensor in the present study.

3.6. Application ofsolvation quatiom to IGC ohta

Retention volumes of 13 probes have been obtained by the IGC method. Using retention volume and LSER parameters of solutes, a multiple linear regression analysis was carried our The dab are given in Table 5, and the LSER equation is given below.

Table 4 Valuer of -AG.. - A G ~ . and -AG: faf BTEX pmber adsorbed onto PPyCi

Frcc ayy of adsorption (Wlmol) k c n e Tolucnc Ethylbuvene o-Xylcne mxylene P X Y I ~

-AG. 5.12 8.79 11.93 13.61 1268 12.02 -AG: 4.42 7.94 10.94 11.75 11.21 11.06 -AG: 0.70 0.85 0.99 1.86 1.47 0.94

I 44 CN! Lin or ol./Senrw.~ and

Table 6 log(VN) valuer or bmenc obtained from 1GC uperimcnt and LSER

Predictions can be made regarding the activity of the PPyCl surface by examining theconstanh in the above equa- tion The largest coefficient is (hat for n the dipolarity- polarisability of the solute, a m w e of the ability of a molecule to stabilize a neighboring charge or dipole. This observation indicates that the presence of nibogen andlor chlorine in the polymer may lead to the formation of dipoles within the PPy layer, with which solute molecules may in- teract Due to this polar nature of PPyCl surface, non-polar benzene vapor gives low sensitivity (0.008 1uQ1ppm) com- pared to high sensitivity (0.08 mWppm) of polar o-xylene vapor.

The constant for u indicates that the PPyCl surface ex- hibie largely basic nature. The negative constant for p I& inforces this, denoting that solutes with a hydrogen-bond basic nature would experience repulsive f o ~ ~ e s . Such basic nature would be expected from the electron-rich stluctuml features of PPy due to conjugated backbone of carbon and niimgen Agaii a negative value for coefficient R2 indicates a repulsive force experienced by the solutes with large n- and n-elecuonic characters.

In order to test the d d i t y of the LSER treamenf the calculated value h m this beament was compared with the experimental value. It can be seen h m Table 6 that retention volume of benzene obtained from experiment is 0.604 which is comparable with 0.666 calculated using LSER There is about 9.4% differences between the two values showing the validity of LSER approach in the present investigation.

4. Conclusions

The sensing responses of the C1-doped PPy sensors towards BTEX compounds were investigated. i t was found that the PPyCl thin film deposited on a pair of suitable in- terdigitated elemodes could be used to detect BTEX vapors reliably by measuring the resistance change of the film The sensitivity of PPyCl sensor for BTEX compounds are found to lie in the range of 0.08mWppm (benzene)-0.8mWppm (0-xylene). A linear relationship of 11AR plotted against 11C.40 was obtained from the theoretical considerations and was correlated well with the experimental resulh. The adsorption equilibrium constants, K,, calculated accord- ing to the adsorption theory, were in a magnitude order of o-xylene(l.97 x 10-3 s m-xylene(l.07 x s p-xylene(7.45 x lo4) s ethylbenzene(6.50 x lo4) s toluene (2.50 x lo4) s benzene (1.72 x lo4). A good correlation between the sensitivity and K, values was ob- tained. The interactions between the BTEX compounds and

the sensing material were M e r evaluated using IGC tech- nique. A good cmlation between the adsorption equilib- rium consmh, K,, and specific component of kite energy of adsorption, -AGl, substantiate the hot that the affinity (is. the coverage) of the detected BTEX vapor to the sens- ing material dominated the sensitivity of the gas sensor in the present work. From the detailed analysis of the interac- tion forces by LSERs, it was found that the surface of the PPyCl was basic in nature.

Acknowledgements i i

This work was financially suppoed by the National Science Council of the Republic of China, Taiwan (Grant NSC-90-2214-E-224-007) and Minisny of Education, Tai- wan (Grant EX-9 l-E-FA 09-5-4).

References

[I] C. Nylandq M Armgat4 L lundrmrm, in: Seiyama et al. (U.), An Ammonia DNnor Based on a Conducting P o l p u ka Chem i d S a m , P m x 4 Q . r or the lnPematiooal Meeting on Ckmid S m o n , FuLw4 J a p q Ekcvtr, Kodmiu, 1983. pp. 20W.07.

p] PN. W e g S.K. LingCbung, Smr. Achraton 19 (1989) 141. [3] PN. BMlcg S.K. Ling-Cauog, S m . Achraton 20 (1989) 287. [4] PN. Baden. P.BM Arrher, S.K. LingChuog, Sens. Achraton 19

(1989) 125. [q L LahdesmaLi, A Invcnstam, A lvarkq Wanta 43 (1996) 125. [q N M Ratshia; AoaL Chim Acla 239 (1990) 257. [A J.P. Blanc, G. Blasquez, LP. GBmain. A Larbi, C Maleyrsan. H.

Robert S m . Actuators 14 (1988) 143. [8] MD. ~aoli, RJ. Wallnwn, Ak. D& 1. Bargo& 1. Chcm. Sac Chcm.

Carnun, (19M) 1015. [91 G. Hailin. L Yucben~. Smr Armatan B 21 (1994) 57.

[ioj P. B.U& F. c a c i l l i ~ N&. S- & A32 (19921 313. [I l l CW. Lin, CR LeZ BJ. Hwng, Maw Qm. Phys. 55 (1998) 139. [I21 CW. Lin, BJ. Hwng, C.R Lec, 1. AppL Polym Sci 73 (1999)

2079. [I31 BJ. H w a a J.Y. Yng, C.W. Lin, 1. Elcchwhu. Sw. 146 (1999)

1231. [I41 C.W. Lin, S.S. Liu, BJ. Hwang, 1. Apd. Polym Sci 82 (2001) 954. [15] ML Abel, MU Chcbimi. S y n h Met 66 (1994) 225. [ l a K.L. TM, B.T.G. TM, K. Kan& KG. Neoh, Y.K. Ox, Phyr. Rev.

8: C n k Matter 42 (1990) 7561. [I71 M Jmwicz, 1. Janata, K. Ashley, S p m , Anal. Chnn 59 (1987)

253. 1181 1. Li. E. Wanc. M. Green P.E Wcrt. Svntb Met 74 11995) 127. i19j P.R ~easdale:~. Ge, K Gilrnme, G.& Wallace, ~ol~m.lnd. 29

(1992) 299. 1201 H. Ge, aa. wallace. I. chmmat 588 119911 z. i2lj M U ~ h e h i m i , E. ~ i o g i r ~ m a u , M dh, 1. Chim Phyr

89 (1992) 1173. 1221 MU Chehimi. M L Akl. E P i o E i s - M m u . M D e h . Svnlh . . - - .

~ e t 60 (1993) IS. [23] JK Can&, CL Young, Physico;hcmid M c a r ~ m m t by Gar

Chmmatogdpby, Why, Chirchester, 1979. I241 D K Lm4 T.C. Ward, H.P. Schrsihr (W), h a r e Gas C%m

mawgaphy: chacmia t i on or Polymar and 0th- Materials. h ceedinp or me ACS Symporiw Suits 391, Amnican Ck&d Society, Washinglm, DC, 1989.