Paintings & Photography

A Distributed Reactivity Model for Sorption by Soils and Sediments. 11. Slow Concentration-Dependent Sorption Rates

Description
A Distributed Reactivity Model for Sorption by Soils and Sediments. 11. Slow Concentration-Dependent Sorption Rates
Published
of 6
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 Distributed Reactivity Model forSorption by Soils and Sediments. 8.Sorbent Organic Domains: Discoveryof a Humic Acid Glass Transitionand an Argument for aPolymer-Based Model E U G E N E J . L E B O E U F A N D W A L T E R J . W E B E R , J R . * Environmental and Water Resources Engineering Program,Department of Civil and Environmental Engineering,The University of Michigan, Ann Arbor, Michigan 48109-2125  Analysisofahumic acidbydifferentialscanningcalorimetryhas revealed the existence of a glass transition point.Glasstransitiontemperatures, T   g ,ofwater-wetanddesiccator-dryspecimens were foundto range from43 ° Cfor water-wet humic acid to 62 ° C for dry samples. Phenanthrenesorptionisotherms for theseandother natural andsyntheticorganic matrices having known glass transition tempera-tures were determined to exhibit linearity and nonlinearitycorrespondingrespectivelytotherubbery(expanded)andglassy(condensed)statesofthesorbent. Invokingalimitingcaseof thedistributedreactivitymodel basedonpolymersorption theory, we explain the observed sorptionbehavior as comprised in each case by a linear phase-partitioning component and a Langmuir-like nonlinear ad-sorptioncomponent. Weconcludethatpolymersorptiontheoryprovides auseful contextinwhichtoassess sorptionphenomena associated with soil and sediment organicmatter, providingmore accurate projections of contaminantbehavior in environmental systems and better informedspecifications of appropriate remediation measures. Introduction In this paper, we experimentally reinforce our view of soiland sediment organic matter (SOM) as a mixture of naturalmacromoleculesconsistingofrubbery(expanded,fluid-like)andglassy(condensed,relativelyrigid)componentsspanning a range of glass transition temperatures,  T  g  ; i.e., the tem-peraturethatseparatestheglassystatefromtherubberystate.Severalinvestigatorshavesimilarlylikenedsoilorganicmattersorptionbehaviortothatofsyntheticpolymers( 1 - 7  ),butnoclear physical evidence linking natural macromolecularbehavior in soil systems to synthetic polymers has beenreporteduntilnow. Toestablishsuchevidence,weemployeddifferential scanning calorimetry (DSC) to identify the exist-ence of a glass transition point for a humic acid, a naturalcomponent of soils and sediments. We then performedsorption experiments with this material and other naturalmacromoleculeshavingknownglasstransitiontemperaturesand compared their sorption behaviors to that of rubbery and glassy synthetic polymers under similar experimentalconditions. Wethenbringouranalogybetweensoilorganicmatter and synthetic polymers full circle by showing theapplicability of a polymer-based limiting case of the Dis-tributedReactivityModel(DRM)forsorbentorganicdomains;i.e., domains II and III ( 8  ). Background Soil organic matter is a relatively complex, heterogeneouscomposite of partially or completely degraded biomoleculesofplantandanimalsrcin. Comprisedfirstascarbohydrates,lipids, and proteins, these biomolecules undergo gradualdecompositionthroughvariousdegreesofdiagenesistoformfulvic and humic acids, humin, and eventually kerogen, theprimary organic component of shales and coals ( 9, 10  ). Although the final composition of soil organic matter iscomplex, we believe that it retains a semblance of themacromolecularcharactersuggestedbyproposedstructuralmodels for fulvic and humic acids ( 11 - 13  ) and for huminsand kerogens ( 9, 10, 14  ).Synthetic polymers, in general, have relatively homoge-neous structures with clearly defined chemical and physicalcharacteristics. They are characterized in part by   T  g   and by their solubility parameter,  σ  p , a measure of their intermo-lecular bonding energy.  T  g   marks a second-order phasetransition in which there is continuity of the free energy function and its first partial derivatives with respect to statevariables such as temperature or pressure, but there is adiscontinuityinthesecondpartialderivativesoffreeenergy.Thereis,therefore,continuityinenthalpy,entropy,orvolumeatthetransitiontemperaturebutnotintheconstant-pressureheat capacity,  Q  ° H  ( 15  ). Hence, measurements of changes in Q  ° H  with increasing temperature yield information about  T  g  as well as about the magnitude of change in  Q  ° H  that occursin the transition from glassy state to rubbery state. Becausetherubberystateallowsgreatermolecularmotion,itexhibitsa greater ability to disperse heat and thus manifests acorrespondinglyhigher Q  ° H . Suchmeasurementscanbemadeusing DSC techniques.Because  T  g   is a function of macromolecular mobility, any changes to the macromolecular structure that increase ordecrease this mobility will have similar effects on  T  g  . Forexample, increased cross-linking restricts chain mobility of larger macromolecular segments, while increased attractiveforces between molecules (as measured by the solubility parameter)requiremorethermalenergytoproducemolecularmotion. Thus, T  g   willgenerallyincreasewithincreasedcross-linking and increased σ  p . In addition, an increase in the freevolumeofthemacromolecule(i.e.,thatvolumenotoccupiedbythecomponentmoleculesthemselves)allowsmoreroomfor molecular movement and thus yields an accompanying reduction in  T  g  . Swelling of a macromolecular sorbent by thermodynamicallycompatiblesolutes(i.e.,thosepossessing similar  σ  p  values) will therefore tend to increase the freevolume and lower  T  g   ( 16  ). With respect to glass transitions in natural systems, wedraw an analogy to what we have earlier termed “soft” and“hard” carbons (e.g., refs 5, 8, 17, and 18). That is, we view diagenesis as a process in which relatively young, expanded,lightly cross-linked, “rubbery” organic matter is convertedinto more condensed, highly cross-linked, more aromatic,“glassy” structures having reduced molecular mobility andcorrespondingincreasedglasstransitiontemperatures. Giventhis perspective, we expect diagenetically less mature (e.g.,soft carbon) soil organic matter such as humic and fulvicacids to possess lower glass transition temperatures thandiageneticallymoremature(e.g.,hardcarbon)organicmatter,such as that comprising kerogens. Coal, a form of kerogen,hasbeenshowntoundergoglasstransitionsattemperatures * Corresponding author: e-mail: wjwjr@engin.umich.edu;telephone: 313-763-1464; fax: 313-763-2275. Environ. Sci. Technol.  1997 ,  31,  1697 - 1702 S0013-936X(96)00626-8 CCC: $14.00  󰂩  1997 American Chemical Society VOL. 31, NO. 6, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY  9 1697  rangingfrom307to359 ° C( 19  ). Additionally,inthepresenceofpyridine,athermodynamicallycompatiblesolventforcoal, T  g   decreases by more than 200  ° C ( 19, 20  ). Biopolymers,such as cellulose, have also been shown to undergo a glasstransitionandhaveexhibitedsimilarreductionsintheirvaluesinthepresenceofwater,a“good”swellingsolventforcellulose( 21 ). Experimental Section Sorbents.  Three natural organic sorbents spanning a rangeof diagenetic alteration and one synthetic organic sorbentsimilar in solubility parameter and permachor values to thathypothesized for SOM ( 4  ) were selected for study. The foursorbentsincludedcellulose,humicacid,coal,andasyntheticpolymer. Highly purified cellulose was obtained fromScientific Polymer Products, Inc. (100  µ m average particlesize) and used as received. The humic acid was obtained inpowderformfromAldrichChemicalCompany,Inc.andwaspurified prior to use employing the technique of Kilduff and Weber ( 22  ). Illinois No. 6 coal was obtained from ArgonneNational Laboratories Premium Coal Sample Program. Thecoal was extracted extensively with pyridine before use in amannersimilartotheprocedureofHallandLarsen( 20  );i.e.,Soxhlet extraction with pyridine (99 + %, Aldrich ChemicalCompany, Inc.) for approximately 6 days until there is nofurther noticeable discoloration of the pyridine. The coal wasthenextractedwithacetone(HPLCgrade,Mallinckrodt)for 24 h to assist in the removal of the pyridine and placedin an oven at 105  ° C for 24 h to allow for volatilization of residual pyridine and acetone. Following baking, the coal wascrushedandsievedtoretainthe38 - 63  µ msizefraction.Poly(isobutylmethacrylate)(PIMA)wasobtainedinbeadformfrom Scientific Polymer Products, Inc. The polymer wascleansed of polymerization artifacts prior to use employing the manufacturer suggested technique of sequential solventflushing. Thistechniqueinvolvesplacementofapproximately 100 g of sample in a borosillicate glass column and flushing  with 18 L of double-distilled, deionized, filtered water(Nanopure, Barnsted Corp.), followed by 1 L of HPLC grademethanol (Mallinckrodt Chemical), followed once more by 18 L of Nanopure water. The cleaned polymer was thenfreeze-dried for 24 h and placed in a desiccator prior to use.N 2  (99.9995%, BOC Gases) based BET surface areas foreach sample were determined using Micromeritics Acceler-ated Surface Area and Porosimitry (ASAP) Model 2010 witha liquid nitrogen (77.35 K) bath. Surface areas and sorbentelemental analysis are summarized in Table 1. Differential Scanning Calorimetry.  Calorimetric mea-surements were performed on humic acid and PIMA using a Perkin-Elmer Series 7 differential scanning calorimeter inthe scanning mode (from 0 to 110  ° C). This instrumentdirectly measures the difference in power applied to sampleand reference cells in milliWatts (mW) as it scans a presettemperature range. Samples of desiccator-dry and water-equilibrated (7 days) samples were weighed into sealablealuminum pans having one small pin-hole punctured in thetoptoallowvolatilizationofwater. Thenon-wethumicacidsamplewasdriedfor30minintheDSCcellunderN 2 (99.998%,BOCGases)at110 ° C;allothersampleswereanalyzedwithoutfurther preparation. Each sample was cooled to 0  ° C, andcalorimetric analyses were performed to a temperature of 110 ° C. Confirmationofnofurthervolatilizationofthesampleor of physically-sorbed water after drying under N 2  at 110 ° Cfor 30 min was determined through independent thermalgravimetric studies (TA Instruments Hi-Res 2950 thermalgravimetric analyzer) scanning a temperature range of 25 - 110  ° C. Additional DSC experiments confirmed that the  T  g  ofthewater-wethumicacidsamplesreturnedtotheirsrcinaldry  T  g  ( ( 4 ° C)aftervolatilizationofallphysically-sorbedwater. Isotherms. (A) Chemicals.  Spectrophotometric gradephenanthrene (98%, Aldrich Chemical Co., Inc.) was used asthe target or probe solute for the sorption isotherm measure-ments. Stock solutions of phenanthrene were prepared asnoted in ref 8 and stored at  - 5  ° C in glass bottles withaluminumcrimpcapscontainingTeflon-linedsiliconesepta.Solute solutions for the isotherms consisted of appropriateamountsofphenanthrenestocksolutionaddedtoabufferedsolution of double-distilled, filtered water (Nanopure, Barn-stedCorp.)containing0.005MCaCl 2 and100mg/LNaN 3 (forbiologicalcontrol),bufferedatpH7withNaHCO 3 . Methanolconcentrations within these solutions were maintained atlessthan0.2%byvolumeinallexperimentstoreducepossibleco-solvency effects ( 25  ). (B)SorptionExperiments.  Establishedprocedures( 5,8,26  )employingabottle-point,fixed-sorbentdosagetechnique were utilized for conducting all sorption experiments incompletely-mixed batch reactors (CMBRs). The CMBRsconsistedof40-mLglasscentrifugetubes(forcellulose,PIMA,and humic acid) and 125-mL glass bottles (for coal), eachsealedwithscrewcaps,Teflon-linedsiliconesepta,andsilverfoiltominimizesystemlossestotheTeflonliner. Eachreactorcontainedenoughsorbenttoensure35 - 70%sorbateuptake.The sorbents were pre-equilibrated with 5 mL of bufferedaqueous solution (described above) in the reactors for 72 hprior to addition of the phenanthrene sorbate to ensurethoroughwettingofsorbentsurfaces. Allreactorswereplacedinrotatingtumblersand,basedonpreliminarysorptionrateand equilibrium studies, equilibrated at 5 or 45 ° C ( ( 0.5 ° C)for 4 weeks, except those for the humic acid, which wereequilibrated for 2 weeks. Solids were separated by centrifu-gation at 2000 rpm for 10 min for the 40-mL tubes and by sedimentationfor1hforthe125-mLbottles. Controlreactorscontaining no sorbent were prepared and operated in thesame manner as described for the sorbent-containing reac-tors. After centrifugation, an approximately 2.0-mL sample of supernatant was immediately withdrawn from each of thereactors and placed in 4-mL glass vials containing 2.0 mL of methanol. Phenanthreneconcentrationsinthesupernatant/methanol mixtures were analyzed using a Hewlett-PackardModel 1050 HPLC in the manner described in ref 8. Averagesystem losses to control reactors were consistently less than TABLE 1. Sorbent Characteristics otherimportantpropertiesandcharacteristicselementalanalyses(%) a  sorbent C H N S Oash(%)particlediametere(  µ m)N 2 -BETsurfacearea(m 2 /g)dryglasstransitiontemp T  g ( ° C) b  wetglasstransitiontemp T  g ( ° C) c  cellulose d  26.0 4.0 0.0 0.0 70.0 ND e  100(av) f  2.72 225 g  - 45 g  humicacid h  50.7 4.5 1.2 0.2 31.4 7.5 38 - 180 3.95 62 43PIMA i  55.0 8.3 0.0 0.0 36.7 ND e  300 - 355 0.02 55 50Illinois No.6coal  j  65.7 4.2 1.2 4.8 8.6 15.5 38 - 63 5.64 355 k  ND e  a Determinedonamasspercentagebasis.  b  Desiccatororoven-dried.  c  Equilibratedwithwaterfor7days.  d   Theoreticalelementalanalysisvalues. e  Notdetermined.  f  Manufacturer data.  g  Ref 21.  h  Elemental analyses fromref 23.  i   Theoretical values for elemental analysis.  j  Elemental analysesfrom ref 24.  k  Approximate from ref 19. 1698  9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 6, 1997  3% of initial concentration, thus no corrections in isothermcalculationswererequired. IsothermmodelparameterswereestimatedusingSYSTATsoftware(Version5.2.1,SYSTAT,Inc.).It is important to note that subsequent extended time or“aging” sorption studies have revealed a very small butcontinualincreaseinphenanthreneuptakeintheglassycoaland PIMA in time periods extending beyond three months.True thermodynamic equilibrium thus may not be reachedin the case of these sorbents, especially at 5 ° C. Equilibriummeasurements in this study are therefore more accurately referred to as “practical” isotherms. Results and Discussion In Figures 1 and 2, we present results from our calorimetricinvestigationofglasstransitionsfordryandwater-wetsamplesof our reference synthetic polymer, poly(isobutyl methacry-late), and the purified humic acid. As shown in Figure 1A,the measured  T  g   of the dry PIMA agrees closely with themanufacturer-specified value of 55 ° C. Equilibration of thePIMA with water, however, results in a reduction of ap-proximately 5  ° C in the  T  g   to a value of 50  ° C (Figure 1B).Glass transitions for humic acid are presented in Figure2. As evident by a comparison of Figures 1A and 2A, the dry humic acid shows a less sharp phase transition than that of thesyntheticpolymer. Webelievethespreadingoftheglasstransition over a greater temperature range is the result of the heterogeneous nature of the humic acid as compared tothe relatively homogeneous synthetic polymer, with respectto both chemical structure and molecular weight. As il-lustrated by comparing panels A and B of Figure 2, equilibra-tion of the humic acid with water brought about a reductionin  T  g   from approximately 62 to 43  ° C, as well as an almostcomplete elimination of an endothermic peak immediately following the phase transition.Thediscrepancybetweenthe5 ° Cdropin T  g  forthewater- wet synthetic polymer and the observed 19  ° C drop for the water-wet humic acid relative to their respective dry statescan be attributed to different degrees of interaction of watermoleculeswiththerespectivesorbingmatrices. Water,witha solubility parameter ( σ  p ) of 23.4 (cal/cm 3 ) 0.5 ( 27  ), interactsmore favorably with humic acid, which has a  σ  p  value of approximately 11.5 (cal/cm 3 ) 0.5 ( 28  ), than it does with PIMA, which has a  σ  p  value of 8.63 (cal/cm 3 ) 0.5 ( 29  ). This increasedinteraction results in greater water uptake (we observedapproximately75%greateruptakeofwaterbythehumicacidas compared to PIMA on a mass of water/mass of sorbentbasis),causinggreaterswellingwithinthehumicacidmatrix and a greater resultant lowering of   T  g  . Additional work withawater-wetpoly(methylmethacrylate), σ  p ) 10.5(cal/cm 3 ) 0.5 ( 29  ) has shown a  T  g   lowering of approximately 15  ° C.The environmental remediation implications of phasetransitionphenomenainnaturalorganicmatterbecomemostapparent when placed in the context of synthetic polymersorption theory. Rubbery polymers, with their relative easeofmolecularmotion,behavesimilarlytofluids,withinwhichsimple Brownian motion and Fickian diffusion of solutemolecules occurs. This allows successful modeling of thesorption/desorption behavior of such materials in terms of phase partitioning theory ( 30, 31 ). Glassy polymers, withdecreased molecular mobility, are described as containing fixedfree-volumemicrovoidswithinwhichsorbingmoleculesare adsorbed and immobilized, resulting in nonlinear ad-sorptionbehavior. Ifthemicrovoidcharacteristicsaremoreor less homogeneous throughout the glassy state matrix, thenonlinearadsorptionwillexhibitaLangmuir-typecharacter.Diffusionwithinthesematricesisoftennon-Fickianinnature, with diffusion coefficients depending upon both the con- AB FIGURE1. Calorimetricanalysisoftheglasstransitionsofpoly-(isobutylmethacrylate).(A)Desiccator-driedspecimenwithnormalendothermicover-relaxationpeakandobservedglasstransitionof 55 ° C.(B)Water-wetspecimen(equilibratedfor7days)withreducedover-relaxationresponseandloweredglasstransitionof50 ° C.Note the large endothermic response near 100 ° C due to thevolatilizationofwater. AB FIGURE2. Calorimetricanalysisoftheglasstransitionsofhumicacid.(A)Desiccator-driedhumicacidwithsignificantlyreducedover-relaxationresponseandbroadglasstransitionpeakat62 ° C.(B)Water-wetspecimen(equilibratedatpH7.0for7days)withlittleover-relaxationresponseandloweredglasstransitionof43 ° C. VOL. 31, NO. 6, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY  9 1699  centration of the solute and the relaxation of the polymermatrix. In such cases, solute diffusion coefficients aregenerally 2 - 3 orders of magnitude smaller than their cor-responding values for polymers in their rubbery state ( 32  ).In the first paper in this series ( 17  ), we proposed a modeldesignedtoaccountformultiplesorptiondomainsofdifferentreactivity at the soil - sediment particle scale. This model,termedtheDistributedReactivityModel(DRM),hastheform where q  e,T , q  e,L ,and q  e,NL refertothetotal,linearcontribution,and nonlinear contribution solid-phase concentrations,respectively(inunitsof   µ g/g; K  D,L isthesumofthedistributioncoefficientsforalllinearlysorbingsoil - sedimentcomponents(L/g); C  e istheresidualsolution-phaseconcentrationofsoluteat equilibrium (  µ g/L);  K  F  is the Freundlich capacity factor[(mg/g)(L/mg) n  ];and n  istheFreundlichexponent(unitless),a joint index of the cumulative magnitude and diversity of energies associated with the adsorption of solute by thecomponents of a heterogeneous condensed phase ( 33  ) anda readily evident measure of the degree of linearity of asorption isotherm. As we noted in the srcinal DRM paper( 17  ) and have since discussed in more detail ( 33  ), theFreundlich term in eq 1 represents a summation of severaldistinct Langmuir-type (i.e., capacity-limited and relatively constant energy) nonlinear adsorptions at different sites ina heterogeneous matrix. Vieth and Sladek ( 34  ) earlier proposed a more restrictivetwo-domainmodelfordescribingthecombinedrubberyandglassy state sorption behavior of chemically homogeneouspolymers. That model, which included a linear phasepartitioningcomponentandasingle,limited-siteLangmuir-type isotherm component, represents a limiting case of theDRM; i.e. where  q  e,T ,  q  e,L ,  q  e,NL , and  C  e  are as defined previously;  K  D,C  isthe distribution coefficient of the linear component of thelimiting case DRM;  Q  ° a  represents the adsorbed phase soluteconcentration that corresponds to saturation of a relatively homogeneous site-limited sorption domain (  µ g/g); and  b   isacoefficientrelatedtotheenthalpyofsorptioninthatdomain(L/  µ g). Thismodelhasbeensuccessfullyappliedinanumberof different applications involving relatively homogeneouspolymer structures ( 35  - 37  ), including descriptions of solutesorption and transport in molecular sieves, dye transport intextilefibers( 38  ),andmostrecentlysorptionofhydrophobicorganic chemicals to soils and sediments ( 7  ).The simplifications of the DRM that lead to the polymer-based expression given in eq 2 are 2-fold. The first simpli-fication assumes that a single partitioning reaction into ahomogeneous rubbery phase accounts for all of the linearcomponent, whereas the more robust DRM addresses linearsorptionreactionsontomineralsurfacesaswellaspartitioning into highly amorphous organic matter. Most significantly,thesecondsimplificationtreatssorptionbyaglassypolymerphaseasasingularsite-limitedandrelativelyconstantenergy process, i.e., a Langmuir-type adsorption. The DRM treatsthe nonlinear component of adsorption as a set of multiplereactions involving different sites of different energy, thusmanifesting Freundlich-type behavior, i.e., a summation of several Langmuir-type adsorptions ( 17, 33  ).Based on our finding of a glass transition point for a soilderived humic acid, we here draw an analogy between SOMsorptionbehaviorandthatofsyntheticpolymersandadvancethe limiting case form of the DRM given in eq 2, the DualReactiveDomainModel(DRDM),forcharacterizingsoilandsediment domains II and III ( 8  ).In Figure 3, the Freundlich term on the right-hand sideofeq1isusedtoemphasizetheoveralllinearity/nonlinearity of aqueous-phase sorption of phenanthrene [ σ  p  )  9.8 (cal/cm 3 ) 0.5 , ref 27] by three samples of natural organic matterand PIMA. An overall linear sorption would be reflected by an n  value of unity, while values of  n  lower than unity signify increasing heterogeneity of sites and increased sorptionenergies ( 33  ). Based on polymer sorption theory and eachspecimen’s water-wet  T  g  , we would expect that at 5  ° C allsorbents except cellulose will exhibit nonlinear sorptionbehavior. ThisexpectationisborneoutinFigure3A,inwhichitisclearlyevidentthattheFreundlich n  valueissignificantly lessthan1forallsamplesexceptcellulose. AsnotedinFigure3B, however, only the coal, which has a relatively high  T  g  ,continuestodemonstratesignificantlynonlinearsorptionat45  ° C. All other samples, including the PIMA and humicacid,whichhavewater-wet T  g  valuesclosetotheexperimentaltemperature, show linear or nearlinear sorption behavior.Table 2 summarizes the Freundlich model parameters.BecausethehumicacidsamplestudiedandthePIMAareboth expected to have relatively homogeneous organicmatrices, we can apply the DRDM introduced above toevaluatetheirrespectivesorptiondata. Theresults,presentedinFigure4,illustratelargenonlinearcontributionstosorptionfor both sorbents at 5 ° C, while at 45 ° C (i.e., at or near their water-wet T  g  values)thesenonlinearcontributionsdisappear q  e,T ) ∑ i  ) 1 k  ( q  e,L ) i  + ∑ i  ) 1 k  ( q  e,NL ) i  ) K  D,L C  e + K  F C  e n  (1) q  e,T ) q  e,L + q  e,NL ) K  D,C C  e + Q  ° a bC  e 1 + bC  e (2) FIGURE 3. Apparent equilibriumisotherms for aqueous-phasephenanthrenesorptiononIllinoisno.6coal[ T  g ) 355 ° C( 19  ), σ  p ) 10.6( 29  )],humicacid[ T  g )  62 ° Cdry,43 ° Cwater-wet, σ  p )  11.5( 28  )],PIMA[ T  g )  55 ° Cdry,50 ° Cwater-wet, σ  p )  8.63( 29  )],andcellulose[ T  g )  225 ° Cdry, - 45 ° Cwater-wet( 21 ), σ  p )  13.8( 29  )].(A)Freundlichmodelfitwith95%confidenceintervalforsorptionat5 ° C.(B)Freundlichmodelfitwith95%confidenceintervalforsorptionat45 ° C. 1700  9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 6, 1997  for the PIMA and are several orders of magnitude less thanthe linear contributions for the humic acid, giving rise toalmost complete partitioning behavior, consistent with exist-ence of a predominantly rubbery state. Although we readily admitthatcomparisonofthesimplelinearityornonlinearity of isotherms may not in itself constitute a basis for drawing definitiveconclusionswithrespecttotheexistenceofseparaterubberyorglassystateswithinsoilorganicmatter,webelievethatthesedata,whenplacedinthecontextofourdifferentialscanning calorimetry results, provide strong evidence tosupport the hypothesis.Consideringthisnewevidence,wefindgeneralagreementbetween the model that treats soil organic matter as twodistinctly different domains ( 8  ) and the rubbery and glassy state behavior of synthetic polymers. This lends support toour hypothesis that the phenomenon of aging can beattributed principally to slow sorption into, and correspond-inglyslowdesorptionoutof,thecondensedorglassyfractionsoforganicmatter. Asthemosthighlycondensedsoilorganicfractionsareaccessedoverextendedtimeframes,thediffusionprocessbecomesincreasinglyslower( 8  ). Asnotedpreviously,diffusion into glassy regions of synthetic polymers is severalorders of magnitude slower than diffusion into rubbery matrices. A related phenomenon commonly observed forpolymersusedasionexchangeresinsistheincreaseddifficulty in regeneration that occurs as the polymer becomes morefully exhausted over extended periods of operation, i.e., asthe diffusing exchange ion migrates into increasingly morehighly cross-linked regions of the polymer ( 39  ). This “aging effect”, reported in the polymer literature almost 40 yearsago ( 40  ), is attributed to the slow relaxation of glassy macromolecules, eventually providing for a reduction inresistancetosolutemigrationandthustofurthermovementofthecontaminantintotheorganicmatrix. Theprocesscanoccur over periods of days, weeks, months, or even years.Theenvironmentalremediationpracticeofactivelydesorbing contaminantsovershorttimeperiodsthuslikelyimpactsonly thosecontaminantscontainedinthemoreexpanded,rubbery fractionsofsoilorganicmatter. Masstransferoutofthemorecondensed fractions is hampered not only by slow diffusionout of those matrices but also by continued diffusion intothem, i.e., true sorption equilibrium has not likely beenachieved. The longer the contaminant is able to diffuse intoa condensed organic matter matrix, the more difficult it is toreverse its flow and, thus, to completely desorb from thatmatrix. Failuretorecognizeandaccountfornonequilibriumand/or nonlinear sorption and desorption behavior of morecondensedorglassyorganiccarbonassociatedwithsoilsandsedimentscanthusresultinlargeerrorsincontaminantfateand transport modeling ( 41 ). Acknowledgments  We thank Brett Bolan and Dr. Albert Yee of the MaterialsScience Engineering Department at The University of Michi-gan for informative discussions on DSC and for the use of their DSC system and David Peevers, Colleen O’Brien, andTina Katopodes, undergraduate research assistants in ourlaboratories, for their assistance in the experimental aspectsofthiswork. Wealsothanktheanonymousreviewersofourmanuscriipt for their detailed comments and Dr. Alok BhandariandDr.TanjuKaranfil,bothpost-doctoralresearchassociates in our program, for their helpful suggestions andfor providing the humic acid sample. This research wasfundedinpartbytheU.S.EnvironmentalProtectionAgency,Office of Research and Development, Great Lakes and Mid- AtlanticCenter(GLMAC)forHazardousSubstanceResearchR2D2 Program and in part by The University of MichiganthroughaUniversityofMichiganRegentsFellowshiptoE.J.L.Partial funding of the research activities of GLMAC and thusof this research was also provided by the State of MichiganDepartment of Environmental Quality. Literature Cited (1) Pignatello, J. J.  Reactions and Movement of Organic Chemicals in Soils  ; Sawhney, B. L., Brown, K., Eds.; Soil Science Society of  America,Inc.andAmericanSocietyofAgronomy,Inc.: Madison, WI, 1989; Chapter 3.(2) van Hoof, P.; Andren, A. W.  Organic Substances and Sediments inWater  ,;Baker,R.A.,Ed.;LewisPublishers: Chelsea,MI,1991;Chapter 8. TABLE 2. Freundlich Model Parameters 5 ° C 45 ° Csorbent  K  F a  n R  2 N  b  K  F a  n R  2 N  b  cellulose 1.20 ( 0.77 0.995 ( 0.050 0.999 16 0.592 ( 1.33 0.977 ( 0.050 0.998 24humicacid 51.00 ( 0.62 0.758 ( 0.025 1.000 18 9.91 ( 1.15 0.951 ( 0.031 1.000 34PIMA 3.06 ( 0.80 0.766 ( 0.052 0.999 20 11.22 ( 2.57 1.038 ( 0.084 1.000 22Illinois No.6coal 839 ( 244 0.635 ( 0.066 1.000 16 582.7 ( 1.2 0.640 ( 0.029 1.000 24 a Units: (  µ g/g)/(  µ g/L) n  .  b  Number of observations. FIGURE 4. Practical equilibriumisotherms for aqueous-phasephenanthrenesorptiononPIMAandhumic acid. (A)DRDMforsorption at 5  ° C. (B) DRDM for sorption at 45  ° C. Note thedisappearanceof theLangmuir contributiontothe45 ° CPIMAisothermandthesubsequentcoincidenceofthelinearandDRDMsorptionmodels. VOL. 31, NO. 6, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY  9 1701
Search
Similar documents
View more...
Tags
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