Food

A study of the factors affecting the diffusion of chlorinated hydrocarbons into polyisobutylene and polyethylene-co-propylene for evanescent wave sensing

Description
A study of the factors affecting the diffusion of chlorinated hydrocarbons into polyisobutylene and polyethylene-co-propylene for evanescent wave sensing
Categories
Published
of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  A study of the factors affecting the diffusion of chlorinatedhydrocarbons into polyisobutylene andpolyethylene-co-propylene for evanescent wave sensing R. Howley a , B.D. MacCraith b , K. O’Dwyer b , P. Kirwan a , P. McLoughlin a,* a Waterford Institute of Technology, Cork Road, Waterford City, Ireland  b  NCSR, Dublin City University, Glasnevin, Dublin 9, Ireland  Received 17 July 2002; received in revised form 23 January 2003; accepted 29 January 2003 Abstract Two polymers, polyisobutylene (PIB) and polyethylene-co-propylene (60% ethylene) (E/Pco), have been considered aspreconcentration media for the detection of chlorinated compounds using mid-IR evanescent wave spectroscopy. In order tooptimizeandpredict sensor response factorsaffectingthe diffusion ofanalytes intothe polymerfilmshasbeenexaminedusingasilverhalide sensing fibercoupled toa FTIRspectrometer.AFickian diffusionmodel was usedto quantifythe diffusion process.The diffusion model calculated a diffusion coefficient based on such parameters as refractive index of the polymer cladding andlight guiding core, the polymer cladding thickness and the principal analyte wavelength of detection.The diffusion of analyte isomers was investigated. The diffusion of chlorinated benzene isomers was employed to determinewhether the position of chlorines on the ring effects diffusion. Physical steric and molecular size effects were found to dominatethe diffusion of these compounds.Solution composition was a fundamental issue in sensor design. The polarity of the sample stream was another diffusiondeterminant. The polarity of the matrix solution was altered by the addition of organic solvents. Rate of diffusion was seen toincrease with decreasing matrix polarity. E/Pco was found to be less susceptible to polarity modification.The results obtained can be employed as a basis for the selection of a polymer for particular industrial applications and also asa means of improving the speed of sensor response. # 2003 Elsevier Science B.V. All rights reserved. Keywords:  Evanescent wave spectroscopy; Optical fibers; Polymer films; Structural isomers; Polarity 1. Introduction Chlorinated hydrocarbons (CHCs) are a class of organic pollutants widely used in industry which havetheir srcins in crude oil and petroleum. In addition totheir individual toxicities these compounds contributeto global environmental problems such as the green-house effect, depletion of the ozone layer and acidrain. Due to increasing environmental regulations thein situ monitoring of CHCs has become an importantsensor application.The use of mid-IR fiber-optic evanescent wavespectroscopy (FEWS) has become widespread as abasis for sensing chlorinated hydrocarbons [1–7]. Pre- viousworksuchastheresearchofKrskaetal.[3,4]andGo¨beletal.[7],isconcentratedonexperimentaldesign Vibrational Spectroscopy 31 (2003) 271–278 * Corresponding author. Tel.:  þ 353-51-302056;fax:  þ 353-51-302679. E-mail address:  pmcloughlin@wit.ie (P. McLoughlin).0924-2031/03/$ – see front matter # 2003 Elsevier Science B.V. All rights reserved.doi:10.1016/S0924-2031(03)00020-1  and increasing the sensitivity of the measurement.Polyisobutylene (PIB), and polyethylene-co-propylene(60% ethylene) (E/Pco), have been previously emplo-yed as sensor claddings both for fiber-optic [7] andATR studies [8–12]. They are both amorphous poly-mers with an optical window in the mid-infrared(MIR) region, where chlorinated hydrocarbons canbe selectively detected below 1500 cm  1 .This work concentrates on the identification of astrategy for sensor optimization based on a detailedknowledge of the diffusion process in the polymercladding. Artificial creation, investigation and com-prehension of factors affecting diffusion contribute tothe design of a functional commercial sensor, whichis capable of dealing with real environments. Theemphasis of this paper is on isomeric selectivity of the polymers claddings and also on effects of back-ground matrix polarity on analyte diffusion.Diffusion can be quantified by calculation of adiffusion coefficient, which permits evaluation of polymer enriching properties. Go¨bel et al. [9] workedin stopped-flow mode to investigate diffusion beha-viour of three CHCs in polymer films using a boxmodel algorithm. Using the box model they calculateda diffusion coefficient for MCB of 5  10  9 and2 : 3  10  9 into E/Pco and PIB respectively, tempera-ture/polymer film thickness for these measurementswas not specified. This model was compared by Go¨bel[11] to a method based on the analytical solution of Ficks diffusion equation which is a faster and easierway of determining the diffusion coefficient. Thecalculation of the diffusion coefficient in this work is based on a Fickian diffusion model. Fickian diffu-sion describes the relationship of the mass flux of apenetrant to the concentration gradient present. Aconstant surface concentration, one-dimensionalFick-ian diffusion model was proposed by Fieldson andBarbari [13]. The concentration profile of a compound diffusingintoapolymerisintegratedwiththeintensityof the evanescent wave. A diffusion curve is simulatedbased on parameters such as, refractive index of thepolymer and the fibercore,the principal wavelength atwhich the analyte absorbs and the polymer film thick-ness. A diffusion coefficient is calculated based onbest fit with the experimental results.A polymer-coated silver halide (AgCl) fiber is usedas the sensing element in this work. AgCl fibers aresuitable for fiber-optic evanescent wave spectroscopy,as they are non-brittle, rather flexible and non-cladded.More importantly, such fibers are transparent over awidespectralrangeparticularlythefingerprintregionof themid-IR(600–1500 cm  1 ).Thepolymersstudiedarechosen due to their low glass transition ( T  g ) tempera-tures. In general, the lower the  T  g  the higher the freevolume between the polymer chains, thus enhancingthe diffusion of organic species within the polymer [8].ItwaspreviouslyfoundthatPIBandE/Pcohave super-ior enriching properties when compared to LDPE,2-polybutadiene, oxidized polyethylene, chlorinatedpolyethylene, ethylene/vinyl acetate copolymer andpoly(4-methyl-1-pentene)[8].AccordingtoBharadwaj and Boyd [14] PIB has a fractional free volume of 0.355 and E/Pco can be assumed to have a freevolumeof 0.42 comparing it to thevalue for polyethylene [14].Initially, the diffusion model is used to examinediffusion into the polymers for isomeric selectivity.Diffusion behaviour of chlorobenzenes is investigatedto determine whether the position of the chlorines onthe benzene ring affects the diffusion of the compoundinto the polymer. Dichlorobenzenes (DCBs) arewidely used in industry and in domestic products suchas odour-masking agents, chemical dyestuffs and pes-ticides. Trichlorobenzenes (TCBs) are used mainly aschemical intermediates and solvents.Links between the trends in diffusion and analyteproperties such as molar volume, water solubility, etc.,are examined. Such trends would facilitate the predic-tion of diffusion behaviour for other series of isomericcompounds and enable distinction between isomers.The effect of matrix polarity on the diffusion of tetrachloroethylene (TeCE) into the two polymers isalso explored. In waste streams there is variability inpolarity in the background matrix. The consequenceof variation in polarity on the diffusion of analytesinto the polymer cladding must be fully quantifiedwhen analyzing results. A polymer cladding could beselected based on its isomeric selectivity and ability towithstand variations in matrix polarity. 2. Experimental 2.1. Reagents Methanol(spectroscopicgrade),ethanol(spectrosco-pic grade), tetrachloroethylene, 1,3-DCB, 1,2-DCB, 272  R. Howley et al./Vibrational Spectroscopy 31 (2003) 271–278  1,4-DCB, 1,2,4-TCB, 1,2,3-TCB and 1,3,5-TCBwere obtained from Sigma–Aldrich (Ireland) and wereused without further purification. Polyisobutylene[–CH 2 C(CH 3 ) 2 –] n ,  M  v  4,7000,000, density 0.915,refractive index 1.5045,  T  g   64  8 C and polyethy-lene-co-propylene (60% ethylene) (–CH 2 CH 2 –)  x -[–CH 2 CH(CH 3 )–]  y ,  M  v  170,000, density 0.860, refrac-tive index 1.48,  T  g   50  8 C were both supplied bySigma–Aldrich (Ireland). 2.2. Instrumentation Analyses were carried out using a Bomem MB120FTIR spectrometer (JVA Analytical Ltd. (Ireland)).Data was acquired and processed using BomemGrams software. The operating temperature for allexperimentswas22  8 C.Eachspectrumconsistedof32scans, with each spectrum taking 46 s to complete. Atypicalruntimewas25–50 min.TheFTIRemployedaliquid nitrogen cooled MCT detector together with apre-amplifier (JVA Analytical Ltd.). Eight hundredsixty micron unclad silver halide (AgCl) sensing fiberwas connected via 1000  m m core/clad AgCl fiber(Ceram Optec (FRG)). The lead in/out fiber was‘butt-coupled’ to the detector. The flow cell was heldin place using a  xyz  positioner, see Fig. 1. Previouswork  [15] has shown significant enhancement of theevanescent wave interaction with the polymer film byinjecting light into the fiber at an angle of 20 8  to thefiber axis.Sample delivery system. A Gilson Minipuls 3peristaltic pump was obtained from AGB Scientific(Ireland). Three millimeter i.d. Tygon 1 tubing, 2 mmi.d. stainless steel tubing and Teflon 1 tubing connec-tors were used to deliver the analyte solution to a10 cm/4 cm 3 glass flow cell (AGB Scientific), whichhoused the sensing fiber. The solution was delivered atan optimized rate of 10 ml min  1 . 2.3. Preparation of the polymer  The polymer was dissolved (2% w/v) in decahy-dronaphthalin (Dekalin), (Merck-Schuchardt (FRG)),2% (w/v) and heated to 70  8 C. It was left to stirmagnetically for 48 h without heating. 2.4. Fiber coating The 860  m m AgCl fiber was prepared by repeatedrinsing with methanol and final washing with acetone.It was dried in air for 10 min to remove residualsolvent. The fiber was then weighed. A dip coater(Chemat Technologies Inc., USA; model 201) wasused to coat the fiber at a speed of 4 mm s  1 .The fiber was then clamped on a holder and dried inair for 48 h and then reweighed. Film thickness wascalculated based on mass difference. The fiber washeld in place in the flow cell with Teflon 1 septa. 2.5. Preparation of analyte solutions The appropriate quantity of DI water was measuredinto a clean 500 ml Duran bottle. Using a glass pipettethe desired quantity of methanol/ethanol was added. Fig. 1. Experimental design used in this work.  R. Howley et al./Vibrational Spectroscopy 31 (2003) 271–278  273  Using 0.5–10 and 10–100  m l Labpette micropipettes(AGB Scientific) the appropriate quantity of the ana-lyte was measured into the bottle. Solid samples weredissolved in methanol. The samples were then sealedimmediately and magnetically stirred for 20 min priorto analysis. 2.6. Analysis of samples DI water was allowed to flow through the flow cellfor 20 min prior to analysis. A background spectrumof DI water flowing through the cell was taken. Theinlet tubing was then introduced to the sample con-tainer and sealed using a rubber bung. The samplewasstirred during analysis. Measurement was initiatedwhen the sample liquid reached the flow cell. Absor-bance values were based on peak heights by selectingtwo points on either side of the peak and applying anaverage baseline. 3. Results/discussion 3.1. Isomeric effects PIB and E/Pco have previously been used in thefingerprint region of the IR spectrum to detect CHCs[8–12]. The chlorinated compounds examined inthis study all absorb in the fingerprint region, Fig. 2.Three of the compounds—1,4-DCB, 1,2,3-TCB and1,3,5-TCB—are solids. They were dissolved in 40 mlmethanol prior to dilution with water; the liquid ana-lytes were spiked with 40 ml methanol for comparableresults. The methanol peak at 1016 cm  1 does notinterfere with the detection of the analytes as shownin Fig. 2. Table 1 lists absorption bands used for the detection of each of the analytes. These bands do notoverlap and thus there is a capability for multicom-ponent analysis without the need for complex datamanipulation. Each isomer can be detected selectively.Partial least squares was used as the basis for such amanipulation technique used by Go¨bel [11]; however,flexibility is sacrificed as an initial calibration step isrequired.Diffusion coefficients were calculated with analytedata obtained from the experimental results. Thediffusion curves were correlated with the Fickiandiffusion model employing the parameters outlinedin Table 2. Fig. 3 presents the correlation of  100 mg l  1 1,4-DCB diffusing into a 2.4  m m PIBand a 2.5  m m E/Pco film, with the binary diffusionmodel. The quality of the correlation gives confidenceto the use of the model to calculate accurate diffusioncoefficients.The calculated diffusion coefficients for the CHCcompounds diffusing into both PIB and E/Pco areshown in Table 3. Initial examination of the trendsinPIBshowthatthediffusioncoefficientsarethesame Fig. 2. Overlaid absorbance spectra of five chlorinated benzenes studied. Note: methanol peak at 1016 cm  1 .Table 1Absorption bands at which the analytes were detectedCompound Wave number (cm  1 )1,2-DCB 7491,3-DCB 6741,4-DCB 8181,2,3-TCB 7721,3,5-TCB 1097274  R. Howley et al./Vibrational Spectroscopy 31 (2003) 271–278  for 1,3-DCB and 1,4-DCB, i.e. 0.908E–10    7E–13and 0.889E  10    8E  13 cm 2 s  1 , respectively. Thediffusion coefficient of 1,2-DCB is lower at0.549E  10 cm 2 s  1 . Review of the trend in polarityvalues for the compounds as shown in Table 4 [16]displays no correlation between analyte relative watersolubility and analyte diffusion coefficients. Thissuggests that the difference in diffusion coefficientisduetosteric/molecularsizeeffects.Stericeffectsarefound to dominate, followed by molecular size andthen polarity effects. Saleem et al. [17] studied thediffusion of positional isomers of xylene into low-density polyethylene (LDPE). They found that shapeof the diffusant influences the diffusion behaviour. Itwas observed that  o -xylene has the lowest diffusioncoefficient in spite of its smaller size (based onmolecular volume), compared to its counter parts.This was considered to indicate that its two adjacentmethyl groups distort its symmetry and thus make itless mobile. This effect was found to be less marked in m -xylene and  p -xylene where molecular symmetry isrestored. Saleemetal. [17]examinedthe diffusionofanumber of compounds in order to assess the relation-ship between penetrant size and diffusion coefficient.It was concluded that diffusion coefficients decreasewith increasing molecular size. They also found that,as the diffusant size becomes larger, the decreaseof   D  with increasing molar volume becomes lesspronounced. We also see these trends in the diffusionof the chlorinated benzenes into PIB. 1,2-DCB has alower diffusion coefficient than 1,3-DCB or 1,4-DCB. Table 2System parameters employed in the binary diffusion model [4]Parameter ValueRefractive index fiber core 2.1Refractive index polymer cladding E/Pco 1.48, PIB 2.4Angle at ATR interface 86.44Polymer film thickness E/Pco 2.5  m m, PIB 2.4  m mWavelength of CHC absorption CHC dependent (Table 1)Fig. 3. Correlation of 100 mg l  1 1,4-DCB diffusing at 818 cm  1 into a 2.4  m m PIB and a 2.5  m m E/Pco film, with the binary diffusion model.  D PIB  ¼  0.889E  10    0.008E  10,  D E/Pco  ¼  1.47E  10    0.01E  10 cm 2 s  1 at 22  8 C. Error bars are based on   1 S.D. based on triplicateresults. Experimental data points shown are the average of triplicate results.Table 3Diffusion coefficients (cm 2 s  1 ) for the analytes diffusing into PIBand E/Pco calculated by correlation with the binary diffusion modelat 22  8 C, flow rate 10 ml min  1 Compound PIB E/Pco1,2-DCB 0.549E  10    7E  13 3.40E  10    2E  111,3-DCB 0.908E  10    7E  13 1.57E  10    1E  111,4-DCB 0.889E  10    8E  13 1.47E  10    1E  111,2,3-TCB 0.543E  10    1E  12 1.67E  10    01,3,5-TCB 1.18E  10    1E  11 1.47E  10    1E  11All analyses performed in duplicate.Table 4Water solubility and molar volume values at 25  8 C [15]Compound Water solubility(mg l  1 )Molar volume(   10 6 m 3 mol  1 )1,2-DCB 156 1121,3-DCB 125 1141,4-DCB 81.3 –1,2,3-TCB 18 –1,3,5-TCB 6.01 –  R. Howley et al./Vibrational Spectroscopy 31 (2003) 271–278  275
Search
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x