| Til bund | | Forside |

Miljøprojekt nr. 1194, 2007

Kemisk oxidation af sediment- og grundvandsforureningen på depotet ved Høfde 42 - fase 1: Projektbeskrivelse og forundersøgelser

Indholdsfortegnelse

Forord

Sammenfatning og konklusioner

• Vidensindsamling
• Laboratorieforsøg
• Forslag til fuldskalaoprensning
• Forslag til pilotforsøg

Summary and Conclusions

• Gathering knowledge
• Laboratory tests
• Proposal for a full scale clean up
• Proposal for a pilot study

1 Indledning

• 1.1 Baggrund
• 1.2 Formål

2 Høfde 42 - Beskrivelse af lokaliteten

• 2.1 Geologi og hydrogeologi
• 2.2 Forureningskarakteristik
  • 2.2.1 Fri fase
• 2.3 Jordforurening
• 2.4 Grundvandsforurening
• 2.5 Forureningsmasse
• 2.6 Indspunsning af depotet i 2006
• 2.7 Indsatsområde for afværgeforanstaltninger

3 Vidensindsamling om kemisk oxidation

• 3.1 Litteraturgennemgang
• 3.2 Nøgleegenskaber af organofosforpesticider
• 3.3 in-situ kemisk oxidation
  • 3.3.1 Oxidationspotentiale
  • 3.3.2 Effekt over for forureningskomponenter
• 3.4 Beskrivelse af metoder
  • 3.4.1 Fenton's reagens
  • 3.4.2 Ozon eller ozon med brintperoxid (avanceret oxidation)
  • 3.4.3 Permanganat
  • 3.4.4 Persulfat
• 3.5 Betragtninger ved in-situ afværge
  • 3.5.1 Tilsætning af iltningsmiddel
  • 3.5.2 Geologi
  • 3.5.3 Fri fase (DNAPL)
  • 3.5.4 Designbetragtninger
  • 3.5.5 Fordele og ulemper ved metoderne
• 3.6 Reaktion med organofosforpesticider
  • 3.6.1 Fentons reagens
  • 3.6.2 Ozon og ozon/brintperoxid
  • 3.6.3 Permanganat
  • 3.6.4 Persulfat
• 3.7 Anbefaling af metoder til laboratorieforsøg

4 Laboratorieforsøg

• 4.1 Formål
• 4.2 Prøvemateriale
• 4.3 Metoder
• 4.4 Beskrivelse af forsøg
  • 4.4.1 Forsøg med ozon
  • 4.4.2 Forsøg med Fenton
  • 4.4.3 Forsøg med Permanganat
• 4.5 Forsøgsresultater
  • 4.5.1 Analyse af vandprøven
  • 4.5.2 Ozon
  • 4.5.3 Fenton
  • 4.5.4 Permanganat
• 4.6 Vurdering
• 4.7 Anbefaling til pilotprojekt

5 Vurdering af anvendelse af kemisk oxidation i fuldskala

• 5.1 Formål
• 5.2 Forslag til afværgestrategi
  • 5.2.1 Indsatsområde
  • 5.2.2 Oxidationsmiddel
• 5.3 Konceptuel design
  • 5.3.1 Injektionsmetoder
  • 5.3.2 Oxidationsmiddel
• 5.4 Vurdering af metodens begrænsninger
• 5.5 Krav til spunsvæg og topmembran
• 5.6 Krav til grundvandssænkning, energiforsyning mm.
• 5.7 Miljøpåvirkninger
• 5.8 Vejrforhold og metodens anvendelighed
• 5.9 Omkostninger
• 5.10 Vurdering af nødvendig oprensningstid

6 Forslag til pilotforsøg

• 6.1 Formål
• 6.2 Forundersøgelser
• 6.3 Testområde
• 6.4 Projektbeskrivelse
• 6.5 Drift og kontrol
• 6.6 Monitering
• 6.7 Sikkerhed og sundhed
• 6.8 Miljøpåvirkninger
• 6.9 Økonomi
• 6.10 Tidsplan for drift og monitering

Referenceliste

Bilag 1 Lokaliteten

Bilag 2 Litteraturstudie

Bilag 3 Laboratorieforsøg

Bilag 3.1 Fentons og ozon/brintperoxid

Bilag 3.2 Permanganat

Forord

Denne rapport vedrører oprensning ved kemisk oxidation af sediment- og grundvandsforureningen på depotet ved høfde 42 – fase 1: Projektbeskrivelse og forundersøgelser. Rapporten indeholder en vurdering af egnetheden af kemisk oxidation til oprensning af forureningen på Høfde 42 samt forslag til et pilotforsøg og fuldskalaoprensning.

Projektet er udbudt af Ringkjøbing Amt og Miljøstyrelsen den 21. november 2005. Arbejdet er udført inden for rammerne af Teknologiudviklingsprogrammet for jord og grundvandsforurening.

Rapporten er udført af et konsortium bestående af COWI, GeoSyntec Consultant, Dr. Pignatello og ISOTEC med COWI som kontraktholder og projektleder, se nedenstående tabel. Sammensætningen af konsortiet har sikret, at den bedst mulige ekspertise på såvel de teoretiske som de praktiske områder har været til rådighed for løsningen af opgaven.

Herudover har firmaet “Laboratory of Applied Process Technologies - APT” (www.aptwater.com) i Californien udført laboratorieforsøgene med avanceret oxidation med ozon og brintperoxid. APT er et firma, som er specialiseret i avanceret vandbehandling af bl.a. grundvand og spildevand, fx med ozon og brintperoxid.

Miljøstyrelsen har nedsat en styregruppe til at følge arbejdet. Styregruppen har bestået af:

• Ole Kiilerich (formand), Miljøstyrelsen
• Børge Hvidberg, Ringkøbing Amt
• Henrik Aktor, Aktor Innovation Aps

Sammenfatning og konklusioner

• Vidensindsamling
• Laboratorieforsøg
• Forslag til fuldskalaoprensning
• Forslag til pilotforsøg

Denne rapport vedrører vurdering af mulighed for oprensning af sediment- og grundvandsforureningen på depotet ved høfde 42 med kemisk oxidation - fase 1. Rapporten omfatter følgende dele:

1. Vidensindsamling om anvendelse af in-situ kemisk oxidation
2. Begrænsede laboratorieforsøg til at belyse effekten af kemisk oxidation
3. Forslag til fuldskalaoprensning med kemisk oxidation
4. Forslag til pilotforsøg med kemisk oxidation

In situ kemisk oxidation (ISCO) er en aggressiv afværgeteknologi for jord- og grundvandsforurening. I denne rapport er der gennemgået de mest relevante metoder inden for in-situ jord- og grundvandsoprensning:

• Fentons reagens
• Ozon og ozon/brintperoxid (avanceret oxidation – AOP)
• Permanganat
• Persulfat

Vidensindsamling

Litteraturstudiet viser, at der ikke er nogen erfaringer med kemisk oxidation på den specifikke forureningssammensætning på Høfde 42 hverken i Danmark eller i udlandet. Der er dog enkelte studier med især parathion og malathion, men ikke inden for in-situ jord- og grundvandsoprensning. På baggrund af litteraturstudiet blev det vurderet, at Fentons reagens og avanceret oxidation med ozon og brintperoxid umiddelbart ser ud til at være de mest effektive metoder. Da permanganat er meget stabil i grundvandszonen, har dette stof nogle praktiske fordele ved feltoprensning i forhold til de andre iltningsmidler. På baggrund af vidensindsamlingen blev det anbefalet, at udføre indledende laboratorieforsøg med Fentons reagens, ozon/brintperoxid og permanganat.

Laboratorieforsøg

Der er udført begrænsede laboratorieforsøg på grundvandsprøver fra Høfde 42 med Fentons reagens, ozon/brintperoxid og permanganat. Laboratorieforsøgene viste, at alle de anvendte metoder giver en hurtig nedbrydning af pesticiderne, men ved oxidation med ozon (alene) og permanganat opstår der giftige oxoner som nedbrydningsprodukter. Oxidation med ozon og brintperoxid (AOP) vurderes som den mest effektive ved de gennemførte forsøg. Fenton oxidation viser også meget gode resultater dog med et enkelt problem. Ved processen genereres MP2-syre, hvilket kan skyldes stor dosering af phosphat i den tilsatte jernkatalysator. MP2-syre er dog ikke særlig giftig.

På baggrund af laboratorieanalyserne blev det vurderet, at kemisk oxidation er en meget relevant in-situ afværgemetode på Høfde 42. Fenton synes umiddelbart at være den mest attraktive metode for en fuldskalaløsning, set ud fra økonomiske, miljømæssige og praktiske vurderinger. Fenton oxidation er i praksis enkel at gennemføre uden de store risici og miljøproblemer. Doseringen er relativ enkel og metoden kan nedbryde pesticider, stofforureninger og nedbrydningsprodukter. Dannelsen af MP2-syre forventes at kunne undgås ved at anvende en anden jernforbindelse end ved laboratorieforsøgene. Det vurderes, at avanceret oxidation og evt. permanganat også er potentielle metoder. Det må bero på en samlet vurdering af fordele og ulemper ved brug af de to metoder på lokaliteten.

Forslag til fuldskalaoprensning

Der er opstillet konceptuel design af fuldskalaoprensning i det øvre magasin. Det vurderes, at de samlede udgifter hertil udgør i størrelsesordenen 67 mio. kr. excl. moms og at et projekt kan udføres indenfor en periode på 3 - 5 år. Det forventes, at en fuldskala oprensnings væsentlig vil reducere miljørisikoen fra depotet.

Metoden kan også anvendes i det nedre magasin, men det vurdere vanskeligere at anvende metoden i det lavpermeable adskillende lerlag.

Forslag til pilotforsøg

Der er udarbejdet forslag til pilotforsøg med Fentons reagens i det øvre magasin. Det overordnede formål med pilot forsøg med Fentons reagens er at vurdere om metoden er en cost-effektiv metode til oprensning af forureningen på Høfde 42. Forslag til testområdet omfatter et område på 100 m² med 4 injektionspunkter og ca. 5 moniteringsboringer. Testcellen placeres i et område med relativ kraftig jord- og grundvandsforurening samt mindre mængde af fri fase. Der injiceres Fentons reagens 3-5 gange med en injektionsmængde på 1,9 m³ væske i hver injektionspunkt pr. injektionsrunde. Der foreslås som udgangspunkt anvendt en 15 % opløsning. Perioden mellem hver injektion er ca. 6-8 uger. Udgifterne til pilotprojektet er i størrelsesordenen 1,1 - 1,75 mio. kr excl. moms.

Det foreslås, at der forud for pilotforsøget udføres supplerende laboratorieforsøg mhp. at forbedre designgrundlaget for pilotforsøget. Der foreslås forsøg med (1) bestemmelse af jordens oxidationsbehov, (2) forsøg med traditionel Fentons reagens (3) forsøg med at reducere dannelsen af MP2-syre, (4) forsøg med at optimere dosering af Fentons reagens og (5) forsøg med behandling af fri fase.

Summary and Conclusions

• Gathering knowledge
• Laboratory tests
• Proposal for a full scale clean up
• Proposal for a pilot study

This report is an assessment of the possibility of cleaning up the sediment and groundwater contamination at Høfde 42 using chemical oxidation - phase 1. The report comprises the following parts:

1. Gathering knowledge of the use of in situ chemical oxidation
2. Limited laboratory tests to illustrate the effect of chemical oxidation
3. Proposal for a full scale clean up using chemical oxidation
4. Proposal for a pilot study using chemical oxidation

In situ chemical oxidation (ISCO) is an aggressive remediation technology for soil and groundwater contamination. This report will go through the most relevant methods in the area of in situ soil and groundwater contamination clean up, which are:

• Fenton's reagent
• Ozone and ozone/hydrogen peroxide (advanced oxidation - AOP)
• Permanganate
• Persulphate

Gathering knowledge

A study of the literature shows that no experience with chemical oxidation for the specific contamination composition at Høfde 42 is available, neither in Denmark nor abroad. However, individual studies of especially parathion and malathion exist, but not in the area of in situ soil and groundwater clean up. On the basis of the literature study, it was assessed that Fenton's reagent and advanced oxidation with ozone and hydrogen peroxide seem to be the most effective methods. As permanganate is very stable in the groundwater zone, this substance has practical advantages during field clean up compared to the other oxidants. On the basis of the knowledge gathered, it was recommended to carry out a preliminary laboratory test using Fenton's reagent, ozone/hydrogen peroxide and permanganate.

Laboratory tests

Limited laboratory tests have been carried out with groundwater samples form Høfde 42 using Fenton's reagent, ozone/hydrogen peroxide and permanganate. The laboratory tests showed that all the methods used result in a fast breakdown of the pesticides, but using oxidation with ozone (only) and permanganate creates hazardous oxones as degradation products. From the tests carried out, oxidation with ozone and hydrogen peroxide (AOP) is assessed to be the most effective. Fenton oxidation also shows good results, but with a single problem; during the process, MP2 acid is generated, which might be due to a too large dosage of phosphate in the added iron catalyst. However, MP2 acid is not particularly hazardous.

On the basis of the laboratory analyses, it was assessed that chemical oxidation is a very relevant in situ remediation method for Høfde 42. Fenton seems to be the most attractive method for a full scale solution from an economical, environmental and practical point of view. Fenton oxidation is in practice simple to carry out without major risks and environmental problems. It has a large range, dosing is simple and it can degrade pesticides, substance contamination and degradation products. It is expected to be possible to avoid the production of MP2 acid by using another iron compound than in the laboratory tests. It is assessed that advanced oxidation and possibly permanganate are also potential methods. This must depend on an overall assessment of advantages and disadvantage in the use of the two methods on the specific site.

Proposal for a full scale clean up

A conceptual design of a full scale clean up in the upper aquifer has been set up. It is assessed that the total expenses for this constitute approximately DKK 67 million excl. VAT, and that a project can be carried out within a period of three to five years. It is expected that a full scale clean up will reduce the environmental risk from the depot substantially.

The method can also be used in the lower aquifer, but it is assessed to be more difficult to use the method in the low permeable separating stratum of clay.

Proposal for a pilot study

A proposal for a pilot study using Fenton's reagent in the upper aquifer has been prepared. The primary objective of the pilot study using Fenton's reagent is to assess whether the method is a cost effective method for cleaning up the contamination at Høfde 42. The proposal for a test area comprises a 100 m² area with four injection points and approx. five monitoring borings. The test cell is placed in an area with relatively heavy soil and groundwater contamination and minor volume free phase. Fenton's reagent is injected three to five times with an injection volume of 1.9 m³ liquid in each injection point per round of injection. Using a 15 % solution is proposed as starting point. The period between every injection is approx. six to eight weeks. The cost of the pilot project is in the order of DKK 1.1 - 1.75 million.

It is proposed that prior to the pilot project, a supplementary laboratory test is carried out in order to prepare the design basis for the pilot project. The following tests are proposed: (1) determination of the oxidation requirement of the soil, (2) tests with traditional Fenton's reagent, (3) test to reduce the production of MP2 acid, (4) tests to optimise the dosing of Fenton's reagent, and (5) test for treating free phase.

1 Indledning

• 1.1 Baggrund
• 1.2 Formål

1.1 Baggrund

Ringkjøbing Amt og Miljøstyrelsen har besluttet at igangsætte en systematisk vurdering af forskellige afværgeteknologier over for sediment- og grundvandsforureningen på depotet ved Høfde 42 på Harboøre Tange. Dette arbejde udføres inden for rammerne af Teknologiudviklingsprogrammet for jord og grundvandsforurening.

Denne rapport vedrører oprensning ved kemisk oxidation af sediment- og grundvandsforureningen på depotet ved høfde 42 – fase 1: Projektbeskrivelse og forundersøgelser. Projektet tager udgangspunkt i COWI’s tilbud af 12. december 2006.

Områdets afgrænsning fremgår af nedenstående figur 1.1.

Figur 1.1 Oversigtskort /7/

1.2 Formål

Projektets overordnede formål er at tilvejebringe den tilstrækkelige og nødvendige viden for valget af den tekniske/økonomisk optimale løsning til fjernelse af forureningen i sediment og grundvand ved Høfde 42.

I fase 1 er formålet at få belyst forskellige tekniske muligheder for en oprensning baseret på kendt viden og eventuelt begrænsede feltanalyser og/eller laboratorieforsøg. Målet er således, at de forskellige projekter i fase 1 giver grundlag for at udvælge et mindre antal oprensningsteknologier til pilotskalaforsøg. Endvidere skal fase-1-projekterne afklare eventuelle specielle krav til et pilotforsøg, herunder udformningen og størrelsen af forsøgsceller.

Det konkrete formål med undersøgelserne i denne rapport er:

• indsamling af relevant viden (litteraturstudie) om anvendelse af in-situ kemisk oxidation på forureningen på Høfde 42 og herudfra udpege de mest relevante metoder til efterfølgende laboratorieforsøg. Litteraturstudiet skal beskrive:
  • rensningsprocesser (princip og mekanismer)
  • oprensning i felten (udformning og udstyr)
  • renseeffektivitet, restforurening, biprodukter og energiforbrug
  • pris
  • arbejdsmiljø og påvirkning af følsomme naturarealer
  • referencer på tidligere anvendelser

• udføre begrænsede laboratorieforsøg til at belyse effekten af kemisk oxidation på den specifikke forureningssammensætning på Høfde 42 og herudfra vurdere den mest relevante metode for et pilotprojekt
• beskrive forslag til et evt. pilotforsøg på Høfde 42 med kemisk oxidation, herunder økonomi, ressourceforbrug og aktiviteter
• give et overordnet forslag til fuldskalaoprensning herunder:
  • effekt og krav til spunsvæg og topmembran
  • miljøpåvirkning i relation til de omgivende naturarealer
  • anvendelighed ved de barske klimabetingelser på Vestkysten
  • evt. krav til grundvandssænkning, energiforsyning og vejanlæg

2 Høfde 42 – Beskrivelse af lokaliteten

• 2.1 Geologi og hydrogeologi
• 2.2 Forureningskarakteristik
  • 2.2.1 Fri fase
• 2.3 Jordforurening
• 2.4 Grundvandsforurening
• 2.5 Forureningsmasse
• 2.6 Indspunsning af depotet i 2006
• 2.7 Indsatsområde for afværgeforanstaltninger

2.1 Geologi og hydrogeologi

Fra terræn træffes enten vind-/vandaflejret finkornet strandsand eller tilkørt/indpumpet finkornet sand. Herunder træffes overvejende fin- til mellemkornet sand stedvist siltet og med indslag af tørv ned til det såkaldte ”indskudte ler/gytjelag”. Områdets øvre sekundære grundvandsmagasin er knyttet til sandlagene over det indskudte lerlag og har centralt i deponeringsområdet en mægtighed på ca. 3-4 meter, dog markant varierende med nedbørsmængder og vandstanden i Vesterhavet. Under det indskudte lerlag er der ned til mellem ca. kote -8,5 og -10,2 DVR90 truffet en nedadfinende sekvens af overvejende finkornet sand med indslag af silt og indlejrede sandede og stedvist lerede siltlag. Områdets nedre sekundære grundvandsmagasin er knyttet til den sandede del af jordlagene under det indskudte lerlag, og ud fra feltobservationer er det kun de øverste ca. 2,0 meter under det indskudte lerlag, der er tilskrevet en egentlig vandføringsevne. Under denne sandede/siltede sekvens er der truffet en markant overgang til siltet ler (den såkaldte ”fjordler”) i kote ca. -8,5 til -10,2 DVR90. Princip af geologisk opbygning fremgår af den konceptuelle model å figur 2.1.

Den hydrauliske ledningsevne i det øvre sekundære magasin er beregnet til ca. 3 × 10^-4 m/s /4/ og den hydrauliske ledningsevne i det nedre sekundære magasin er beregnet til 4,2 × 10^-5 m/s /4/.

2.2 Forureningskarakteristik

En konceptuel model af forureningssituationen på Høfde 42 fremgår af figur 2.1 /1/. Det fremgår at forureningen findes i den mættede zone, og at den kraftigste forurening findes i det øvre sekundære magasin og i et indskudt lerlag ned til det nedre sekundære magasin. Der er dog stedvis også fundet kraftig forurening i det nedre sekundære magasin.

Forureningen ved Høfde 42 er en meget kompleks blandingsforurening, som er sammensat af mere end 100 forskellige kemiske forbindelser. I forbindelse med risikovurdering af forureningen er der identificeret en række stoffer, som er vurderet til at udgøre den største risiko for vandkvaliteten i Vesterhavet /2/:

• Parathion (EP3)
• Methyl-parathion (MP3)
• Ethyl-sulfotep
• Fyfanon (malathion)
• Kviksølv
• EP1
• EP2-syre

Figur 2.1 Konceptuel model /1/

2.2.1 Fri fase

Der forekommer fri fase typisk i det øvre sekundære magasin, mens forekomsten i det nedre sekundære magasin er væsentligt mere begrænset. Parathion og methylparathion udgør den væsentligste del af den fri fase (ca. 70 % /2/). De mest forurenede sandprøver har haft et indhold af fri fase på ca. 2 vægtprocent. Det bemærkes i /2/, at forekomsten (vertikalt og lateralt) af fri organisk fase i det egentlige hotspotområde ikke er endelig afklaret. I bilag 1.3 er der kort, som viser udbredelse af fri fase på Høfde 42.

2.3 Jordforurening

Den væsentligste forureningsmasse findes som jordforurening bundet til sediment. Parathion og methylparathion er de væsentligste stoffer. Der er målt op til 33.000 mg/kg af parathion og 13.000 mg/kg af methylparathion. Bilag 1.1 viser jordforureningens fordeling.

2.4 Grundvandsforurening

Tabel 2.1 viser gennemsnitlige koncentrationer af udvalgte forureningskomponenter i grundvandet centralt i høfdedepotet, og tabel 3.2 viser gennemsnitskoncentrationer i fanen for en hel række af forureningskomponenter. I bilag 1.2 er der vist en figur med forureningsudbredelse af parathion i det øvre magasin. Den horisontale forureningsfordeling er vist på figur 2.2.

Tabel 2.1 Vurderede gennemsni77skoncentrationer i grundvandet centralt i høfdedepotet (mg/l) /4/

Figur 2.2. Horisontal afgrænsning af forureningen. Det røde område angiver kildeområdet, dvs. areal med høje jordkoncentrationer, medens det blå område angiver vandbåren forurening. Fra /7/. Røde prikker angiver boringer.

2.5 Forureningsmasse

Tabel 2.2 viser et groft overslag over massen af udvalgte forureningskomponenter fordelt på sedimentbundet, fri fase og opløst i grundvand. Det fremgår at den væsentligste forureningsmasse er sedimentbundet (ca. 95%). Desuden findes den altdominerende forureningsmasse i det øvre magasin.

I 2 er området inddelt i 4 delområder:

1. depot-/nesivningszone
2. indre randzone
3. ydre randzone

Den væsentligste forureningsmasse findes i område 1 (omkring 85 % af den samlede masse). Områdeinddeling fremgår af bilag 1.4.

Tabel 2.2 Den skønnede mængde af de væsentligste forureningskomponenter i høfde 42, fordelt på sedimentbundet, organisk fri fase og opløst i grundvandet /2/

i.b.: ikke bestemt

2.6 Indspunsning af depotet i 2006

I forbindelse med et ønske om at reducere udsivning af miljøfremmede stoffer fra Høfde 42 området, har Ringkjøbing Amt ladet udarbejde et projekt, om indebærer nedsætning af en spunsvæg omkring det kraftigst forurenede område.

Den afgrænsende spuns omkring høfdedepotet har en udstrækning på ca. 580 lbm og består af en fri stålspunsindfatning med tætnede spunslåse. Spunsprofil Acelor AZ 13 10/10, stålkvalitet 240 Mpa. Spunsen har en topkote på +3,00 m og en spidskote på -10,80 m. Den afgrænsende spuns er mod Vesterhavet beskyttet af en stenkastning.

Det område, der spunses, er vist på nedenstående figur 2.3..

Figur 2.3. Figuren viser det fremtidige indspunsede område. Spuns er angivet med rødt.

2.7 Indsatsområde for afværgeforanstaltninger

På baggrund af forureningssituationen, som den kendes nu, er følgende væsentlige forudsætninger gjort i forbindelse med udarbejdelse af redegørelsen:

• Oprensningsareal er 19.000 m² svarende til det indspunsede område
• Tykkelsen af den umættede zone er ca. 2-6 m
• Tykkelsen af den mættede zone er 3-4 m, svarende til det øvre sekundære magasin ned til det tynde lerlag, der adskiller det øvre fra det nedre magsin.
• Volumen af indsatsområde i den mættede zone er ca. 76.000 m³
• Areal af område med fri fase er ca. 6.200 m². Den gennemsnitlige tykkelse af fri fase er ca. 0,05 m (0,01 – 0,25 m). Volumen af sand med fri fase er ca. 310 m³.

3 Vidensindsamling om kemisk oxidation

• 3.1 Litteraturgennemgang
• 3.2 Nøgleegenskaber af organofosforpesticider
• 3.3 in-situ kemisk oxidation
  • 3.3.1 Oxidationspotentiale
  • 3.3.2 Effekt over for forureningskomponenter
• 3.4 Beskrivelse af metoder
  • 3.4.1 Fenton's reagens
  • 3.4.2 Ozon eller ozon med brintperoxid (avanceret oxidation)
  • 3.4.3 Permanganat
  • 3.4.4 Persulfat
• 3.5 Betragtninger ved in-situ afværge
  • 3.5.1 Tilsætning af iltningsmiddel
  • 3.5.2 Geologi
  • 3.5.3 Fri fase (DNAPL)
  • 3.5.4 Designbetragtninger
  • 3.5.5 Fordele og ulemper ved metoderne
• 3.6 Reaktion med organofosforpesticider
  • 3.6.1 Fentons reagens
  • 3.6.2 Ozon og ozon/brintperoxid
  • 3.6.3 Permanganat
  • 3.6.4 Persulfat
• 3.7 Anbefaling af metoder til laboratorieforsøg

Dette afsnit beskriver den viden, som findes om behandling af organofosfor-pesticider ved kemisk oxidation med fokus på parathion, methyl parathion, malathion og aminoparathion. Der gennemgås relevante metoder med Fenton’s reagens, avanceret oxidation med ozon og ozon og brintperoxid (AOP), permanganat og persulfat. Der er kun gennemgået metoder, som er relevante ved in-situ jord- og grundvandsforureninger (fx er photooxidation ikke gennemgået).

Der er udført en systematisk opsamling af den væsentligste litteratur på verdensplan, herunder tekniske manualer, peer-reviewed tidsskrifter, fagblade og konferenceartikler.

I dette afsnit gives en oversigt over de væsentligste forhold i litteraturstudiet. For en mere detaljeret gennemgang henvises til det samlede litteraturstudie i bilag 2.

3.1 Litteraturgennemgang

Resultaterne af litteraturgennemgangen er opdelt i følgende punkter:

• Beskrivelse af de væsentlige fysiske kemiske egenskaber af organofosforpesticider
• Gennemgang af metoder med kemisk oxidation, herunder kemiske reaktioner, vurdering af design og etablering af in-situ behandling, fordele/ulemper ved metoden samt forhold omkring sikkerhed og økonomi
• Reaktionskemi med organofosforpesticider med de enkelte metoder med kemisk oxidation
• Beskrivelse af udvalgte cases med Fenton’s og ozon/brintperoxid
• Tabel med oversigt over den indsamlede litteratur med nøgleinformationer om forfatter, forureningskomponent, iltningsmiddel, laboratorium/feltundersøgelser, effektivitet, nedbrydningsprodukter, geologi mm.

3.2 Nøgleegenskaber af organofosforpesticider

De væsentligste fysiske og kemiske egenskaber af organofosforpesticiderne er beskrevet i bilag 2 – tabel 2.

• Da massefylden af de rene stoffer er større end 1, vil stofferne være tungere end vand. De frie stoffer betegnes også DNAPL (Dense Non-Aquous Phase Liquids).
• De har en relativ lav opløselighed i vand
• Damptrykket er også lavt, hvilket medfører en lille flygtighed
• De kan nedbrydes ved hydrolyse, men dette kræver høj pH. Ved de nuværende pH forhold (<5) er hydrolysen meget lille
• Stofferne kan nedbrydes både under iltede og reducerede forhold, dog er mange af nedbrydningsprodukterne også meget giftige. Med den lave pH på lokaliteten er nedbrydningen langsom

3.3 in-situ kemisk oxidation

In situ kemisk oxidation (ISCO) er en agressiv afværgeteknologi for jord- og grundvandsforurening, som pt. kun er anvendt på relativt få projekter i Danmark, men som inden for især de sidste 10 år har vundet stor anvendelse i USA. Metoden har dog været kendt i mange år inden for eksempelvis spildevandsrensning. I denne rapport gennemgås følgende metoder:

• Fentons reagens
• Ozon og ozon/brintperoxid (avanceret oxidation – AOP)
• Permanganat
• Persulfat

3.3.1 Oxidationspotentiale

Renseeffekten af de forskellige kemiske oxidanter varierer afhængig af forskellige faktorer, herunder oxidationspotentialet (E⁰) af oxidanten og reaktionen mod en given forureningskomponent. Oxidationspotentialet for de enkelte kemiske oxidanter er vist i tabel 3.1. Jo større oxidationspotentiale jo stærkere oxidant er stoffet. Det ses, at Fentons reagens har det største oxidationspotentiale og permanganat det svageste.

Tabel 3.1. Redox potentiale af almindelig anvendte kemiske oxidanter (ITRC, 2005)

3.3.2 Effekt over for forureningskomponenter

Tabel 3.2 viser en oversigt over hvilke forureningskomponenter, som kan behandles med de respektive metoder. Det fremgår, at Fentons reagens, ozon og AOP er de mest bredspektrede oxidanter.

Permanganat har primært vist sig at være effektiv til oxidation af klorerede ethener, men ikke klorethaner eller benzinkomponeter. I kontrast hertil har Fenton vist sig at være effektiv til oxidation af benzinkomponenter, PAH, de fleste typer af klorerede opløsningsmidler og nogle phospor-pesticider (methyl parathion).

Tabel 3.2. Effektivitet af forskellige kemiske oxidanter. Modiferet fra ITRC, 2005.

3.4 Beskrivelse af metoder

3.4.1 Fenton’s reagens

Fenton reaktionen er veldokumenteret både inden for grundvands- og spildevandsområdet. Ved Fenton metoden anvendes brintperoxid sammen med en katalysator (jernsalte). Herved dannes hydroxylradikaler, som med et oxidationpotentiale på 2,8 V er et meget kraftigt oxidationsmiddel. Processerne er komplicerede, men det hele starter med, at jern(II)-salt reagerer med brintperoxid under dannelse af hydroxylradikaler efter ligningen:

Fe⁺² + H₂O₂ → Fe⁺³ + OH^- + HO•

Først reagerer jern(II) med brintperoxid og danner hydroxylradikal (OH·), hvorefter der sættes en kædereaktion, som vi ikke her skal komme nærmere ind på. Det skal dog nævnes, at nogle processer vil accelerere dannelsen af hydroxylradikaler, mens andre vil stoppe dannelsen. Jern(III) vil kunne omdannes til jern(II), og det åbner for dannelsen af nye hydroxylradikaler.

Hvis jern(III) udfælder som Fe(OH)₃ - og det starter allerede så småt ved pH = 2,5 - kan man risikere at miste så meget jern, at kædeprocessen stopper. Derfor gælder det om at holde en lav pH (typisk 3-5).

Alternativt kan man dosere jern som en kompleks forbindelse, der holder jern(III) opløst. Man kalder ofte denne blanding af jernsalt og kompleksdanner for en Fenton katalysator. Som kompleksdanner anvendes fx NTA, EDTA eller citronsyre. Det er nødvendigt at anvende en Fenton katalysator, når oxidationen skal foregå i neutral eller basisk miljø for at begrænse udfældning af ferrihydroxid. Denne metode kaldes ”Modificet Fenton” og er den metode, som er anvendt ved laboratorieforsøget beskrevet i afsnit 4.

Der er imidlertid risiko for, at den organiske kompleksdanner efterhånden selv oxideres. Derfor kan det undertiden være nødvendigt at dosere mere Fenton katalysator undervejs i oxidationsprocessen.

Hydroxylradikaler eksisterer kun i brøkdele af et sekund, og når de dannes, skal der derfor være de stoffer til stede, som skal oxideres. Holdbarheden af Fenton kemikalier i grundvandszonen er derfor kort, og ”influensradius” fra en injektionsboring er derfor også relativ lille (få meter).

Det skal bemærkes at Fenton processen er en varmedannende proces (exoterm).

3.4.2 Ozon eller ozon med brintperoxid (avanceret oxidation)

Ozon alene eller ozon i kombination med brintperoxid er et meget stærkt oxidationsmiddel. Ozon kan nedbryde organisk stof ved direkte oxidation eller ved dannelse af hydroxylradikaler, som er endnu stærkere oxiderende end ozon selv. Ved oxidationen bruges det ene iltatom i ozon, mens de to andre iltatomer afgives som luftarten ilt, hvilket betyder en kraftig luftudvikling ved ozonering.

Ved anvendelse af ozon sammen med brintperoxid dannes der hydroxylradikaler, hvilket betyder, at denne blanding et meget kraftigt oxidationsmiddel. Processer, hvor der dannes hydroxylradikaler som mellemprodukter, kaldes ofte for AOP (Advanced Oxidation Processes). Ozon + brintperoxid vil derfor være et kraftig iltningsmiddel end ozon alene.

De fleste felterfaringer med ozon eller ozon/brintperoxid er fra avanceret vandbehandling, men der er kun få erfaringer med metoden til in-situ grundvandsrensning. Metoden er som Fenton meget bredspektret, så de fleste kendte forureningskomponenter kan nedbrydes med metoden. Ozon kan også anvendes i den umættede zone i modsætning til de andre metoder, som er mest egnet i den mættede zone.

3.4.3 Permanganat

Permanganat er et kraftigt iltningsmiddel med et oxidationspotentiale på 1,7 V - dvs. lidt svagere end ozon (2,2 V) og noget svagere end hydroxylradikaler (2,8 V). Permanganat er stabilt og har derfor en langtidsvirkning ved injektion i grundvand, hvilket gør det særligt interessant ved feltoprensning.

Permanganat er effektiv overfor nedbrydning af kulstof dobbeltbindinger (C=C), aldehydgrupper eller hydroxylgrupper. Eksempelvis er permanganat meget effektiv over for klorerede ethener (fx tetraklorethylen med kulstof dobbeltbindinger), men ikke klorerede ethaner, som ikke har nogen dobbeltbinding.

Ved pH < 3,5 omdannes permanganat til Mn⁺², hvilket svarer til en ændring i mangans iltningstrin fra 7 til 2.

MnO₄^- + 8H⁺ + 5e^- → Mn⁺² + 4H₂O

Ved pH mellem 3,5 og 12 omdannes permanganat fortrinsvis til brunsten (MnO₂), hvilket svarer til en ændring i iltningstrin fra 7 til 4.

MnO₄^- + 2H₂O + 3e^- → MnO₂ (s) + 4OH^-

Det mindste forbrug af permanganat fås derfor ved pH < 3,5, hvor forbruget af permanganat kun er 3/5 af forbruget ved pH mellem 3,5 og 12.

Permanganat kan tilsættes som enten kalium- eller natriumsalt afhængig af ønsker og behov. Natriumsaltet er mere opløseligt i vand, men kaliumsaltet er billigere.

3.4.4 Persulfat

Persulfat S₂O₈^2- er et relativt nyt iltningsmiddel ved jord- og grundvandsoprensninger. Natriumsaltet (Na₂S₂O₈) er det mest anvendte idet det har stor opløselighed i vand (730 g/l ved 25 grader celcius). Der findes også kalium og ammoniumsalte men disse anvendes ikke så ofte. Persulfationen er et kraftigere iltningsmiddel end brintperoxid.

Persulfationen kan ved temperaturer omkring 15-20 grader virke som iltningsmiddel:

S₂O₈^-2 + 2e^- → 2SO₄^-2 (1)

Anvendelse af persulfat som iltningsmiddel er dog ikke særlig effektiv ved normale grundvandstemperaturer. Thermisk spaltning (ligning 2) af persulfat eller reaktion med en egnet reduktant, fx ferroioner (ligning 3-4), danner sulfatradikaler, som dramatisk forøger oxidationsstyrken af persulfat. Mangan eller kobber kan også anvendes som reduktant til dannelse af sulfatradikale

varme + S₂O₈^2- → 2 SO₄^{- •} (2)

Fe²⁺ + S₂O₈²^- → Fe³⁺ + SO₄^{- •} + SO₄²^- (3)

Fe²⁺ + SO₄^{- ·} → Fe³⁺ + SO₄²^{- •} (4)

SO₄^{- •} + H₂O → HSO₄^- + OH^• (5)

3.5 Betragtninger ved in-situ afværge

3.5.1 Tilsætning af iltningsmiddel

Generelt bliver iltningsmidlerne tilsat til undergrunden gennem en række injektionsboringer enten kontinuert eller som enkeltinjektioner. Som boringer kan anvendes både direct push (fx geoprobe) eller traditionelle boringer. Antal af injektionsboringer vil især afhænge af de geologiske forhold, mængde af oxidationsstof, og om der anvendes recirkulationsløsning eller passive systemer med enkeltinjektioner. Principskitse fremgår af figur 3.1.

Figur 3.1 Typisk anvendelse af kemisk oxidation

3.5.2 Geologi

Som med andre in situ metoder kræves det, at man kommer i kontakt med forureningen. ISCO er derfor mest velegnet i permeable aflejringer. Metoden kan også anvendes i lavpermeable aflejringer, men her må påregnes længere oprensningstid og vanskeligheder med restforurening. Især metoderne med Fentons reagens og ozon er problematiske mht. lavpermeable aflejringer pga. den korte levetid af iltningsmidlerne.

3.5.3 Fri fase (DNAPL)

ISCO kan anvendes både ved lave og høje opløste koncentrationer. Selve omdannelsen af forureningen sker i den opløste fase. Fri fase behandles ikke direkte men sker ved løbende opløsning til vandfasen. Fjernelse af DNAPL er således begrænset af hastigheden af opløsningen fra den fri fase til den opløste fase. Flere laboratoriestudier har vist, at kemisk oxidation kan forøge hastigheden af opløsning af den fri fase med faktor 2 – 30 gange (fx Kim and Gurol 2005). Fjernelse af fri fase er primært afhængig af overfladearealet, hvor behandlingen sker, og reaktionsraten. Andre faktorer, der har betydning for oprensning af fri fase, er bl.a. levering af iltningsmiddel til behandlingsområdet og opløseligheden af forureningskomponenterne.

3.5.4 Designbetragtninger

Tabel 3.3 opsummerer nogle af de vigtigste forhold, som man skal være opmærksom på ved brug af kemisk oxidation. Det fremgår, at Fentons reagens eren ustabil varmedannende (eksoterm) behandlingsmetode, mens permanganat er en stabil proces, som ikke medfører nogen varmedannelse. Halveringstiden for Fenton’s reaktioner er i størrelsesordenen sekunder til minutter, mens halveringstider for permanganat typisk er fra dage til måneder. Ozon er også meget reaktivt og ustabilt med en kort halveringstid for forureningskomponenter. Teknikker der anvender Fentons reagens i DNAPL områder, vil typisk bestå af 2 – 4 injektioner over en periode på 2 – 9 måneder. I modsætning hertil kan brug af permanganat eller persulfat udføres med færre injektioner (holder sig typisk 3-6 måneder inden de er fuldstændig nedbrudt). Permanganat kan også anvendes til recirkulation gennem behandlingsområdet.

Tabel 3.3. Betragninger ved anvendelse af ISCO (modificeret fra /9/).

3.5.5 Fordele og ulemper ved metoderne

Der er nogle generelle fordele og ulemper ved kemisk oxidation. Fordele ved kemisk oxidation er at hvis iltningsmidlet kommer i kontakt med forureningen sker nedbrydningen meget hurtigt. Ulemper er bl.a., at der kan ske mobilisering af tungmetaller og at der er tale om aggressive metoder overfor grundvandsmiljøet, som bl.a. kan reducere den mikrobiologiske aktivitet. Tabel 3.4 viser specifikke fordele og ulemper for den enkelte metode. Der henvises i øvrigt til bilag 2, afsnit 3.2.

Tabel 3.4 Fordele og ulemper ved metoderne

3.6 Reaktion med organofosforpesticider

I bilag 2 – afsnit 6 er der en detaljeret gennemgang af den viden, der findes om nedbrydning af organofosforpesticider ved kemisk oxidation. Generelt er der ingen studier med samme forureningssammensætning som på Høfde 42. Der er dog enkelte studier med især parathion og malathion. Det er karakteristisk, at studierne typisk er på laboratorieniveau, og at studierne ikke er målrettet mod in-situ grundvandsrensning.

3.6.1 Fentons reagens

I litteraturen ses flere studier med organofosforpesticider. Der ses også flere studier med nedbrydning af 4-nitrophenol, dimethylfosfat og p-Nitropenol som er beslægtede stoffer. Der er typisk tale om laboratoriestudier.

Forsøgene viser, at man ved Fentons oxidation kan nedbryde de pågældende stoffer. Der er dog ingen studier på lignende forureningssammensætning som på Cheminova. Det vides derfor ikke, om Fenton på forhånd vil være effektiv over for forureningen på Høfde 42.

Der er ingen studier af nedbrydning af eksempelvis oxoner, som dannes ved iltning af parathion og malathion, og som er meget giftige.

3.6.2 Ozon og ozon/brintperoxid

De eneste studier af ozonering af organofosforpesticiderne er gennemført med parathion og 4-nitrophenol. Studierne viser, at parathion kan nedbrydes, men at der kan dannes ethyl- og methylparaoxon, som er meget giftig. Det vides ikke, om disse oxoner vil blive fuldstændig nedbrudt med ozon. Der er ingen studier af nedbrydning med en blanding af ozon og brintperoxid, men da dette er et stærkere oxidationsmiddel, må der forventes mindst den samme effekt som med ozon alene.

3.6.3 Permanganat

Der er kun få studier af iltning med permanganat på nogle af forureningskomponenterne på Høfde 42.

I et studie /8/ blev malathion omdannet til malaoxon og methyl-parathion blev omdannet til methyl-paraoxon inden for 5 min. Nitrophenoler ser også ud til at blive iltet af permanganat. Herudover er der ingen data for de øvrige forureningskomponenter på lokaliteten.

Sammenfattende vurderes det, at organofosforpesticiderne bliver iltet af permanganat, men kendskabet til slutprodukterne er mangelfuldt.

3.6.4 Persulfat

Der er ikke specifikke studier af organofosforpesticiderne på Høfde 42. Det vides, at persulfat kan angribe aromatiske forbindelser og sandsynligvis vil reagere med parathion og p-nitrophenol. Der er dog kun få data, og det forventes, at reaktionen ved normale grundvandstemperaurer vil være langsom. Det vil sandsynligvis være nødvendigt at varme grundvandszonen op eller anvende katalysator.

3.7 Anbefaling af metoder til laboratorieforsøg

Fentons reagens er det eneste iltningsmiddel blandt in-situ oxidationsmetoderne, som er påvist at kunne nedbryde organofosforpesticiderne på Høfde 42. Det anbefales derfor, at der udføres indledende laboratorieforsøg med grundvand fra Høfde 42 for at undersøge, om Fenton oxidation kan anvendes til komplet nedbrydning af den specifikke forureningssammensætning, der findes på Høfde 42.

Det anbefales, at der også udføres indledende forsøg med ozon samt en kombination af ozon og brintperoxid for at undersøge, om der kan ske en fuldstændig nedbrydning af forureningskomponenterne fra Høfde 42. Ozon og ozon/brintperoxid er meget kraftige oxidationsmidler, og det må på forhånd forventes, at også oxidationsprocesser kan give en væsentlig nedbrydning af de organiske stoffer.

Permanganat er muligvis for svagt et iltningsmiddel til at nedbryde stofferne fra Høfde 42 fuldstændigt. Der er dog indikationer på, at organofosforpesticiderne vil blive nedbrudt, men nedbrydningsprodukterne kendes ikke. Da permanganat er meget stabil i grundvandszonen, har dette stof nogle praktiske fordele ved feltoprensning i forhold til de andre iltningsmidler. Det anbefales derfor, at der udføres indledende forsøg med permanganat på grundvandsprøver fra Høfde 42 for at undersøge nedbrydningseffekten.

Da der ikke er nogen viden om persulfats effekt på organofosforpesticiderne, og da udgifterne til metoden vurderes at være høje, anbefales det i første omgang, at der ikke udføres laboratorieforsøg med persulfat.

På grund af den begrænsede økonomiramme for laboratorieforsøgene udføres forsøgene i første omgang på grundvandsprøver (uden fri fase). Til projektering af et pilotprojekt vil det dog også være nødvendigt at udføre forsøg med sedimentprøver og evt. fri fase.

4 Laboratorieforsøg

• 4.1 Formål
• 4.2 Prøvemateriale
• 4.3 Metoder
• 4.4 Beskrivelse af forsøg
  • 4.4.1 Forsøg med ozon
  • 4.4.2 Forsøg med Fenton
  • 4.4.3 Forsøg med Permanganat
• 4.5 Forsøgsresultater
  • 4.5.1 Analyse af vandprøven
  • 4.5.2 Ozon
  • 4.5.3 Fenton
  • 4.5.4 Permanganat
• 4.6 Vurdering
• 4.7 Anbefaling til pilotprojekt

På baggrund af anbefalingerne i afsnit 3 er der udført indledende laboratorieforsøg med tre forskellige oxidationsmetoder:

• Jernkatalyseret oxidation med brintperoxid (Fentons reagens)
• Oxidation med ozon og ozon + brintperoxid (AOP)
• Oxidation med permanganat

Forsøg med Fentons reagens er udført af ISOTEC i USA, forsøgene med ozon/brintperoxid er udført af ”Laboratory of Applied Process Technologies - (APT) i USA, mens forsøgene med permanganat er udført af COWI i COWIs laboratorium i Lyngby.

Resultaterne af forsøgene er beskrevet detaljeret i bilag 3.1 (Fentons og AOP) og 3.2 (permanganat). I dette afsnit gives en kort sammenfatning af de væsentligste resultater og konklusioner.

4.1 Formål

Formålet med laboratorieforsøgene er at undersøge effektiviteten af tre kemiske oxidationsmetoder til nedbrydning af forureningskomponenterne i en vandprøve udtaget i Cheminovas depot på Høfde 42. Laboratorieforsøgene er tilrettelagt med henblik på at udpege egnede kemiske oxidationsmetoder til en evt. fuldskala in situ oprensning af forureningen på Høfde 42.

4.2 Prøvemateriale

Der er udtaget vandprøve til laboratorieforsøg fra boring V04B i Høfde 42. Placering fremgår af bilag 1.2. Boringen er udvalgt af NIRAS på baggrund af:

• placeret centralt i tidligere depot- og nedsivningsområde
• repræsentative opløste niveauer af Chemicals Of Concern (COC)
• filtreringsforsøg har vist, at forureningsniveau i vand fra boring mht. COC i henholdsvis filtrerede og ufiltrerede vandprøver en nogenlunde ens
• ingen forekomst af fri fase

Vandprøver blev udtaget fredag den 21. april 2006 af NIRAS. Prøverne blev efter forpumpning udtaget ufiltrerede i 6 x 10-liters plastdunke, som efterfølgende hurtigt blev nedfrosset i fryser på lokaliteten. Der blev påfyldt ca. 7 l i hver plastdunk. Prøverne i de 6 plastdunke vurderes at være ensartede, idet de blev udtaget inden for ca. 15 min. COWI hentede de frosne prøver mandag den 24. april 2006. De frosne dunke blev anbragt i køletaske med køleelementer og sendt expres med luftfragt direkte til de 2 laboratorier i USA, hvor prøverne er modtaget den 26.04.2006 stadig nedfrosne. De to sidste prøver er i frosset tilstand sendt til COWIs laboratorium i Lyngby, hvor de er modtaget den 27. april 2006 (stadig frosne).

Prøverne er uden sedimentindhold, da man ved laboratorieforsøgene alene ønsker at undersøge nedbrydningen af de forurenende stoffer i vandfasen med henblik på at udvælge den bedst egnede oxidationsmetode. Der er således ikke udført forsøg med sediment.

4.3 Metoder

Teorien for de afprøvede oxidationsmetoder er beskrevet i kapitel 3 samt i bilag 2. Her skal anføres nogle korte bemærkninger til de enkelte metoder.

Ozon:

Ozon er alene et kraftigt oxidationsmiddel, som delvist fungerer gennem dannelse af hydroxylradikaler som mellemprodukter. Ozon er ustabil i vand, og stabiliteten afhænger bl.a. af pH og vandets sammensætning. Halveringstiden af ozon i vand vil typisk være ca. 15 minutter.

I kombination med brintperoxid dannes der hurtigt flere hydroxylradikaler, hvilket normalt gør denne kombination betydelig mere effektiv.

Når et laboratorieforsøg er gennemført, vil der normalt være behov for at fjerne overskud af oxidationsmiddel med et passende kemikalie (reduktionsmiddel), før prøven sendes til analyse.

Fenton:

Ved en Fenton oxidation anvendes brintperoxid som oxidationsmiddel og et jern(II) salt som katalysator. Herved dannes hydroxylradikaler, hvilket betyder, at man kan forvente nogenlunde samme effektivitet, som opnås med ozon + brintperoxid.

Processen forløber normalt bedst ved pH = 3-5, hvor udfældning af jern(III) hydroxid er mindre end ved neutral pH. Udfældningen kan begrænses ved at anvende en kompleksdanner sammen med jernsaltet. Overskud af oxidationsmiddel skal normalt fjernes efter forsøget, før prøver udtages til analyse.

Permanganat:

Permanganat er et meget stabilt oxidationsmiddel, som virker ved alle pH-værdier, men forbruget er mindst ved lav pH (< 3,5), hvor permanganat bliver reduceret til Mangan(II)-ioner. Da start-pH i vandprøven fra Høfde 42 er 3,12, vil der i starten fortrinsvis dannes mangan(II), men da pH stiger undervejs, kan der også forventes dannet en del brunsten (MnO₂).

Permanganat er et svagere oxidationsmiddel end ozon og hydroxylradikaler. Da permanganat kan nedbrydes af luft og lys, vil forsøgene foregå i lukkede flasker placeret i mørke. Efter forsøget må overskud af permanganat normalt fjernes før analyse for at få stoppet oxidationsprocessen.

4.4 Beskrivelse af forsøg

4.4.1 Forsøg med ozon

Der er gennemført fem forsøg med ozondosering alene samt med ozondosering og brintperoxid i blanding (AOP). I alle tilfælde foregår forsøget med 2,05 liter prøve, hvortil der løbende doseres en afmålt ozonmængde i form af en blanding af ilt og ozon indeholdende ca. 5 vægtprocent ozon. Udsugningsgassen fra processen indeholder ilt og ureageret ozon. Når ozonindholdet i gassen bliver for højt, stoppes forsøget. Ved forsøget tilstræbes følgende doseringer til de 2,05 liter-prøve:

Ozon alene: 190 mg/l

AOP(I): Ozon = 500 mg/l, H₂O₂ = ca. 300 mg/l

AOP(II): Ozon = 1000 mg/l, H₂O₂ = ca. 650 mg/l

Der måles nitrat, phosphat, sulfat og farve før og efter forsøget. Der er endvidere foretaget analyser på Cheminovas laboratorium for 22 stoffer omfattende pesticider og deres nedbrydningsprodukter samt visse urenheder fra råvarerne til parathionfremstilling.

Efter forsøget fjernes overskud af iltningsmiddel med metanol. Dog er der anvendt thiosulfat i prøverne til Cheminova- og microtoxanalyser. Efter første prøveserie glemte man at tilsætte metanol/thiosulfat, og man har derfor gentaget to af forsøgene fra første serie med den rigtige konservering. Efter kemikalietilsætning er prøverne opbevaret i køleskab eller fryser afhængig af de analyser, der skal foretages på dem.

4.4.2 Forsøg med Fenton

Der er gennemført fem forsøg med Fenton oxidation i 2 forsøgsserier samt et kontrolforsøg. Til alle forsøgene er anvendt 640 ml prøve i en glasreaktor med låg, hvortil der doseres de ønskede mængder af kemikalier. I første forsøgsserie er der doseret brintperoxid og modificeret Fentons reagens (katalysator) i tre omgange. I anden forsøgsserie er der doseret kemikalier i to omgange. Reaktionstiden har varieret fra 24 til 48 timer.

I alle forsøg har man fulgt indhold af brintperoxid og jern med henblik på at styre koncentration og dosering. Der blev analyseret samme parametre som ved forsøg med ozon.

Overskud af oxidationsmidler er fjernet i prøverne til analyse. Efter kemikalietilsætning er prøverne opbevaret i køleskab eller fryser afhængig af de analyser, der skal foretages på dem.

4.4.3 Forsøg med Permanganat

Forsøgene blev gennemført i 130 ml glasflasker med låg. En passende prøvemængde (120-130 ml) blev tilsat en given mængde kaliumpermanganatopløsning for at opnå den ønskede startkoncentration af permanganat.

Der blev indledningsvis gennemført forsøg med tre forskellige permanganatkoncentrationer, hvor forbruget af permanganat blev målt efter et døgn med henblik på at fastlægge den endelige dosering af permanganat ved nedbrydelighedsforsøgene.

Ved nedbrydelighedsforsøgene blev oxidationen undersøgt som funktion af tiden, idet prøverne blev analyseret efter 0, 2, 5, 24, 48 og 72 timer. Reaktionsflaskerne blev opbevaret i mørke i lukkede glasflasker (uden luft) ved 22-24 ^oC. Der blev undersøgt 2 parallelle prøver for hver reaktionstid.

Efter endt reaktion blev restindhold af permanganat målt fotometrisk og pH, redoxpotentiale, nitrat og farve blev ligeledes målt. Phosphat kunne ikke måles, da det meste phosphat blev udfældet sammen med mangan og brunsten.

50 ml prøve blev konserveret med natriumbisulfit for at fjerne overskud af permanganat, hvorpå prøven blev dekanteret og/eller filtreret samt nedfrosset og senere sendt til Cheminova for analyse af 22 pesticider og nedbrydningsprodukter.

4.5 Forsøgsresultater

4.5.1 Analyse af vandprøven

Vandprøven fra Høfde 42 er på Cheminovas laboratorium analyseret for 22 specifikke organiske forbindelser - herunder 5 pesticider, 11 nedbrydningsprodukter (incl. oxoner) samt 6 råvareurenheder og deres nedbrydningsprodukter. Der er endvidere analyseret for en række standard spildevandsparametre for at få et godt vurderingsgrundlag af prøven. Standardparametrene fremgår af tabel 4.1 og Cheminovas analyser, fremgår af tabel 4.2

Tabel 4.1: Generelle analyseresultater fra startprøven fra Høfde 42.

Analytechs analyser er baseret på danske standardmetoder. COWIs analyser er baseret på transportabelt måleudstyr og hurtigmetoder til fotometer. De amerikanske analyser er standard spildevandsanalyser.

Som det fremgår af tabellen, har grundvandsprøven en høj ledningsevne på grund af et stort saltindhold fra havvandet. COD er relativt høj, fordi vi bevidst har udtaget en prøve i en boring med høj forureningskoncentration. COD er ca. 4 gange højere end TOC, hvilket forekommer ret sandsynligt. pH er som ventet meget lav. Der er nogenlunde overensstemmelse mellem de forskellige målinger og analyser- dog har man i USA kun fundet den halve phosphatkoncentration af, hvad der er fundet i Danmark.

Tabel 4.2: Beregnet indhold i mg/l af kulstof, brint, kvælstof, ilt, phosphor, svovl og chlor ud fra Cheminovas analyse og det procentvise indhold af grundstoffer. Den procentvise fordeling på de 7 grundstoffer er beregnet for summen af de 22 analyserede komponenter. Resultatet af TOC, organisk-N og organisk-P er anført til sammenligning. Værdier med < er medtaget lig med detektionsgrænsen.

Det skal nævnes, at den aktuelle prøve som ønsket har et højt indhold af forurenende stoffer, men den er ikke karakteristisk for depotet som helhed. Depotet indeholder af pesticider overvejende parathion og amino-parathion og kun mindre mængder methyl-parathion og malation. I den aktuelle prøve er malathion og methyl-parathion dominerende. Det betyder dog ikke noget for de konklusioner, der kan drages af de udførte forsøg med henblik på en fuldskala løsning, da alle parathioner er til stede i prøven.

Det konstateres ud fra analyserne, at de 22 stoffer fra Cheminovas analyse repræsenterer 60-70% af de organiske forbindelser, der findes i vandet. Der er med andre ord en række uidentificerede organiske stoffer ud over de nedbrydningsprodukter, som er medtaget i tabel 4.2.

4.5.2 Ozon

Detaljer fra ozonforsøgene fremgår af rapporten fra Applied Process Technology. Hovedresultaterne fra disse forsøg fremgår af tabel 4.3.

Tabel 4.3: Resultat fra samtlige ozonforsøg. Koncentrationer angivet i mg/l. Forsøg 1 er udført 09.05.06 er uden kemisk fjernelse af oxidationsmiddel efter forsøget, mens prøverne fra forsøg 2 udført 15.05.06 er tilsat kemikalier (NA = ikke analyseret).

Klik her for at se Tabel 4.3

Forsøg med ozon alene (190 mg/l) resulterer i nedbrydning af alle pesticider, men der akkumuleres giftige nedbrydningsprodukter i form af methyl-paraoxon og malaoxon. Der sker kun en begrænset mineralisering målt på tilvækst i nitrat, sulfat og ortophosphat. Der forbruges kun ca. 30% af den ozonmængde, der skal til for at modsvare et COD-indhold på 170 mg/l. Der er tilsyneladende ikke den helt store forskel på, om forsøgene er stoppet med kemikalier eller ej, og derfor indgår alle forsøgsresultater i vurderingerne.

Forsøg med kombineret tilsætning af ozon og brintperoxid (AOP) giver fuldstændig nedbrydning af pesticider, og der påvises ingen nedbrydningsprodukter. Tilvækst i nitrat, sulfat og ortophosphat indicerer en fuldstændig mineralisering. Toxiciteten af den behandlede prøve er dog kun reduceret med en faktor 3, men det kan skyldes den meget lave pH-værdi i det rensede vand.

4.5.3 Fenton

Detaljer fra Fenton forsøgene fremgår af rapporten fra ISOTEC. Hovedresultaterne fra disse forsøg fremgår af tabel 4.4.

I alle forsøg nedbrydes pesticiderne fuldstændigt. Der konstateres en ophobning og stigning af nedbrydningsproduktet MP2-syre. Jo mere jernkatalysator der tilsættes, jo mere MP2-syre findes i det rensede vand. ISOTEC vurderer, at det kan skyldes, at jernkatalysatoren har et stort phosphatindhold, der reagerer med mellemprodukter og danner MP2-syre. Derfor bør man prøve en anden katalysator uden phosphat - evt. et rent jernsalt - for at eliminere dette problem.

Fenton oxidationen medfører en meget betydelig reduktion af toxiciteten (40-80 gange). Den store reduktion ved Fenton sammenlignet med ozon kan måske skyldes, at prøverne efter Fenton oxidation er neutrale, mens prøverne efter ozon behandlingen er meget sure (pH = ca. 2,5). Ved Fenton-forsøgene var det planlagt, at slut-pH skulle ligge mellem 3 og 6, men pH i de oxiderede prøver lå i stedet for i intervallet 3,62 - 8,30 - jo større dosering af Fenton-kemikalier, jo højere pH-værdi. Der er ikke nogen helt oplagt forklaring på denne pH-stigning, men en hypotese er, at der dannes basiske mellemprodukter ved oxidationen.

Tabel 4.4: Resultat fra samtlige Fenton forsøg. Koncentrationer angivet i mg/l. Forsøg udført som modificeret Fenton med tilsætning af jernkatalysator, der medfører pH-stigning (NA = ikke analyseret). Enhed er mg/l.

Klik her for at se Tabel 4.4

4.5.4 Permanganat

Forsøgene er gennemført på COWI´s laboratorier og alle forsøgsresultater og detaljer fremgår af bilag 3.2. Hovedresultaterne fra disse forsøg fremgår af tabel 4.5.

Tabel 4.5: Cheminova-analyser (mg/l) på prøver fra permanganat oxidationsforsøg. De anførte resultater er gennemsnit af to parallelle prøver. Resultaterne er korrigeret for fortynding med permanganat og konserveringsmiddel, der sammenlagt har givet ca. 10 % fortynding.

Ved forsøgene nedbrydes alle pesticider relativ hurtigt, men der akkumuleres nogle nedbrydningsprodukter i vandet under oxidationen først og fremmest E-OOOPO, paraoxon, methyl-paraoxon og malaoxon. Specielt oxoner er meget giftige, hvilket er en stor ulempe ved metoden. Det kan dog ikke udelukkes, at længere reaktionstid vil kunne føre til nedbrydning af disse forbindelser.

Der dannes ikke nitrat ved oxidationen, og det var ikke muligt at måle, om der blev dannet ortophosphat. Forbruget af permanganat var knap 600 mg/l efter 3 døgn med en svag stigende tendens ved længere oxidationstid. Der er fra start tilsat ca. 2600 mg/l permanganat - altså mere end 4 gange forbruget på 3 døgn. Forbruget af permanganat svarer til ca. halvdelen af det iltforbrug, som måles ved COD, hvilket viser, at der ikke er sket nogen komplet oxidation og mineralisering af de organiske stoffer.

Ved oxidationsprocessen stiger pH i reaktionsflasken fra ca. 3,14 til mellem 4,0 og 4,4 afhængig af reaktionstiden. Denne pH-stigning er helt som ventet ifølge reaktionsligningen.

4.6 Vurdering

I tabel 4.6 er foretaget en sammenligning og vurdering af de tre oxidationsmetoder til nedbrydning af pesticider og andre organiske forbindelse i depotet på Høfde 42.

Tabel 4.6: Sammenligning af metoder. ND = under detektionsgrænsen, P = produceret, T = transient reaktionsprodukt, +++ = hurtig nedbrydning, ++ = hovedparten nedbrudt, + = delvis nedbrudt

Alle metoder giver en hurtig nedbrydning af pesticiderne, men ved oxidation med ozon (alene) og permanganat opstår der giftige oxoner som nedbrydningsprodukt. Det gør umiddelbart disse processer mindre interessante til en fuldskalaløsning, med mindre oxonerne kan nedbrydes ved en forlænget reaktionstid.

Oxidation med ozon og brintperoxid (AOP) vurderes som den mest effektive ved de gennemførte forsøg. Denne metode nedbryder alle de analyserede stoffer i tabel 4.4 ved en ozon koncentration på 1000 mg/l, mens et par af nedbrydningsprodukterne (malaoxon og E-OOOPO) kan påvises i meget lave koncentrationer ved en dosering på 500 mg/l. Meget tyder på, at der opnås en fuldstændig mineralisering af de organiske stoffer ved denne metode.

Fenton oxidation viser også meget gode resultater med en enkelt undtagelse. Ved processen genereres MP2-syre, hvilket kan skyldes stor dosering af phosphat i den tilsatte jernkatalysator. Selv om MP2-syre ikke er meget giftig, er det dog ønskeligt, at så meget organisk stof som muligt nedbrydes. Såfremt dette problem kan løses, forekommer Fenton at være en meget anvendelig oxidationsmetode ved en fuldskalaløsning, da den rent praktisk er forholdsvis simpel at etablere og gennemføre.

4.7 Anbefaling til pilotprojekt

Rent teknisk ville permanganat være det mest velegnede som oxidationsmiddel på grund af stoffets stabilitet og langtidsvirkning. Når først permanganat er injiceret ned i depotet, vil det kunne virke i flere måneder efterhånden som stoffet trænger frem gennem jordlagene til de forurenede områder. Her kan man frygte, at AOP-løsningen vil have problemer på grund af ozons korte levetid i vand (halveringstiden er ca. 15 minutter). Tilsvarende vil Fenton-løsningen måske lide under, at den jernkatalyserede proces stopper, når der ikke er mere forurening tilbage på et givet sted. Det vil kræve ny dosering af jernkatalysator i forbindelse med, at brintperoxid trænger frem til nye forureningsområder.

Ud fra en kemisk vurdering suppleret med den praktiske løsning og økonomien for en fuldskalaløsning vil vi umiddelbart anbefale Fenton oxidation til pilotforsøg undersøgelser. Det skyldes, at Fenton processen er forholdsvis enkel at håndtere, og den har en passende stor aktionsradius fra injektionsstedet. Samtidig synes processen at være den mest effektive til nedbrydning af pesticiderne, hvis vi lige ser bort fra dannelsen af MP2-syre, som sandsynligvis kan undgås ved at ændre på de tilsatte Fenton-kemikalier.

Dog bør der først gennemføres nogle supplerende laboratorieforsøg omfattende:

• Fenton oxidation med anden Fenton katalysator samt rent jernsalt som katalysator
• Forsøg med permanganat oxidation med forlænget reaktionstid (10, 20 og 30 dage) for at konstatere, om oxoner og E-OOOPO kan nedbrydes fuldstændigt
• Orienterende forsøg med brintperoxid alene

Hvis ingen af disse tre undersøgelser kan give bedre resultater, kan det overvejes at anvende AOP til pilotforsøg i stedet, selv om denne metode er mere kompliceret at gennemføre i praksis og mindre effektiv til nedbrydning af pesticider end Fenton. Det må dog bero på en samlet vurdering af fordele og ulemper ved brug af de to metoder på lokaliteten. Såfremt flere af forsøgene falder positive ud, er det en ny situation, og planen for pilotforsøg må revurderes, idet de økonomiske og praktiske aspekter ved en fuldskalaløsning skal medtages sammen med de opnåede resultater.

5 Vurdering af anvendelse af kemisk oxidation i fuldskala

• 5.1 Formål
• 5.2 Forslag til afværgestrategi
  • 5.2.1 Indsatsområde
  • 5.2.2 Oxidationsmiddel
• 5.3 Konceptuel design
  • 5.3.1 Injektionsmetoder
  • 5.3.2 Oxidationsmiddel
• 5.4 Vurdering af metodens begrænsninger
• 5.5 Krav til spunsvæg og topmembran
• 5.6 Krav til grundvandssænkning, energiforsyning mm.
• 5.7 Miljøpåvirkninger
• 5.8 Vejrforhold og metodens anvendelighed
• 5.9 Omkostninger
• 5.10 Vurdering af nødvendig oprensningstid

Dette afsnit giver en konceptuel beskrivelse af fuldskalaoprensning med Fentons reagens. Der tages udgangspunkt i den eksisterende viden om forureningsforholdene på lokaliteten samt i den nye viden, som er fremkommet i denne rapport.

5.1 Formål

Formålet er at give en konceptuel beskrivelse af anvendelse af kemisk oxidation som afværgemetode med Fentons reagens, herunder at give et groft overslag over omkostninger.

5.2 Forslag til afværgestrategi

5.2.1 Indsatsområde

Indsatsområde er det øvre magasin inden for spunsvæggen. Dette område omfatter langt hovedparten af forureningsmassen, herunder den fri fase. På baggrund af forureningssituationen, som den kendes nu, er følgende væsentlige forudsætninger gjort i forbindelse med udarbejdelse af fuldskalaoprensning:

• Oprensningsareal er 19.000 m² svarende til det indspunsede område
• Efter etablering af spunsvæg og geomembram er tykkelsen af den umættede zone ca. 3-4 m
• Tykkelsen af den mættede zone er 3-4 m, svarende til det øvre sekundære magasin ned til det tynde lerlag, der adskiller det øvre fra det nedre magasin
• Volumen af indsatsområde i den mættede zone er ca. 76.000 m³
• Areal af område med fri fase er ca. 6.200 m². Den gennemsnitlige tykkelse af fri fase er ca. 0,05 m (0,01 – 0,25 m). Volumen af sand med fri fase er ca. 310 m³
• Den totale masse af forurening i indsatsområdet er ca. 245.000 kg, heraf udgør den sedimentbundne forurening ca. 237.000 kg, den fri fase ca. 8.100 kg og forureningen opløst i grundvandet mindre end 1.000 kg

5.2.2 Oxidationsmiddel

Der anvendes Fentons reagens, men det er endnu ikke fastlagt om der bruges modificeret Fentons uden pH regulering eller traditionel Fentons med lav pH. I begge produkter anvendes brintperoxid som en 50% opløsning.

5.3 Konceptuel design

Området opdeles i 2 delområder:

• Område 1 er et område med fri fase, svarende til ca. 6.200 m². I dette område forventes det, at mindst 85 % af forureningsmassen findes, svarende til 200.000 kg
• Område 2 er et område uden fri fase, svarende til ca. 12.800 m². I dette område findes ca. 45.000 kg forureningsmasse

5.3.1 Injektionsmetoder

Til injektion af Fentons reagens kan der anvendes "direct push boringer" (eksempelvis med Geoprobe) eller permanente boringer (traditionelle filterboringer) - eller en kombination af begge boringstyper.

Hvis der anvendes "direct push" metode forudsættes der 620 injektionsboringer i området med fri fase (3,5 m mellem hvert punkt svarende til 1,75 m influensradius). I området uden fri fase forudsættes 650 injektionsboringer (5 m mellem hver boring svarede til 2,5 m influensradius). I hver boring injiceres i 2 niveauer (top og bund) med 2 m filtre. Hvis der anvendes stationære (traditionelle filterboringer) kan afstanden mellem boringer være større. I område med fri fase kan boringerne eksempelvis placeres i ca. 5 m afstand (radius på 2,5 m). Injektionsboringer kan udføres jf. principper i figur 5.1. Her er dog anvendt stationære filterboringer.

Der påregnes doseret i 15 injektionsboringer pr. dag, når der doseres. Nogle af disse boringer vil være permanente boringer, men hvor mange permanente boringer, der etableres, vil først blive fastlagt under detailplanlægningen af fuldskalaprojektet.

Figur 5.1 Principskitse af injektion med fentons reagens med traditionelle boringer (fra Geocleanse.com)

5.3.2 Oxidationsmiddel

Ud fra ISOTEC’s erfaringer kan der antages et forbrug af iltningsmiddel pr kg forureningsmasse på ca. 1:20 for den sedimentbundne forurening og 1:25 for fri fase. Det betyder, at der for hvert kg forureningsmasse skal anvendes mellem 20-25 kg 50 % H₂O₂.

I område 1 med fri fase antages et forbrug på ca. 4.500.000 kg 50% H₂O₂ og i område 2 skal der bruges ca. 900.000 kg. Det vil sige et samlet forbrug på ca. 5.400 tons 50% H₂O₂ (ca. 4.500 m³ 50 % H₂O₂). Dette svarer til 2.700 tons 100 % H₂O₂.

Ud fra de gennemførte forsøg og kendskab til den kemiske sammensætning af pesticiderne kan det beregnes, at 1 kg pesticid forbruger ca. 1,5 kg ilt til fuldstændig oxidation. Det svarer til et forbrug på 3 kg 100 % H₂O₂ pr. kg pesticid. På den baggrund kan det beregnes, at ISOTEC´s forslag til dosering er på 4,4 gange det teoretiske forbrug, hvilket er rimeligt i betragtning af, at der også vil være et forbrug i sedimentet (NOD) samt et spild af H₂O₂ ved dekomponering.

Det vurderes, at der skal gennemføres 4-6 injektionsrunder i hvert af de to områder med ca. 6-8 ugers mellemrum. En injektionsrunde vil tage ca. 82 arbejdsdage i området med fri fase og ca. 108 arbejdsdage i området uden fri fase. Antages en koncentration på 5 % brintperoxid i den injicerede væske, skal der samlet tilsættes ca. 45.000 m³ af denne væske til jorden over de 12 injektionsrunder i de to områder.

5.4 Vurdering af metodens begrænsninger

Det vurderes, at der er gode muligheder for oprensning af forurening i det øvre magasin, både den opløste, sedimentbundne og den fri fase. Der kan dog være områder, som kan være vanskeligere at oprense, bl.a. områder som er mere finkornet eller områder som ikke rammes pga. kanalstrømning. Det vurderes også muligt at anvende metoden i det nedre magasin. Derimod vurderes det vanskeligere at oprense forurening i det indskudte lerlag på grund af de lavpermeable aflejringer.

5.5 Krav til spunsvæg og topmembran

Der er ingen særskilte krav til den etablerede spunsvæg. Mht. den etablerede plastmembran forudsættes det, at denne kan gennembores ved etablering af injektions- og moniteringsboringer. Ved injektion af Fentons reagens dannes der varme og gasser. I den forbindelse skal det vurderes, om der skal ske en ventilation af den umættede zone for at hindre at der sker spredning af gasser i den umættede zone eller optrængning af gasser til udeluften.

5.6 Krav til grundvandssænkning, energiforsyning mm.

Der er ingen krav til yderligere grundvandssænkning udover de allerede planlagte dræn, der udføres i forbindelse med spunsvæggen.

5.7 Miljøpåvirkninger

Pilotforsøget vurderes ikke at give nogle væsentlige negative effekter på det omgivende miljø. Forureningskomponenterne omdannes til kuldioxid, vand og diverse naturlige uorganiske ioner, dvs. helt ugiftige komponenter. Endvidere vil Fenton-katalysatoren tilføre jern og uorganiske saltrester. Jern vil efterhånden udfældes som ferrihydroxid eller andre tungtopløselige jernforbindelser og aflejres i depotet, hvilket dog ikke anses for noget problem. Der kan ske en vis afgasning af flygtige stoffer på grund af opvarmning af grundvandszonen. Disse dampe kan evt. opsamles gennem dræn med aktivt kulfilter. Der kan ligeledes ske en vis mobilisering af metaller, men dette vurderes ikke at være noget problem, da der ikke vil ske nogen spredning pga. spunsvæggen.

5.8 Vejrforhold og metodens anvendelighed

Vejrforhold vurderes ikke at have væsentlig betydning for arbejdets gennemførelse. Det kan dog i vinterperioden med frost og sne være vanskeligt at udføre injektion.

5.9 Omkostninger

Tabel 5.1 giver et groft overslag over udgifter til fuldskalaoprensning. Der er regnet med 6 injektionsrunder i både område med fri fase og område uden fri fase.

Tabel 5.1 groft overslag over udgifter til fuldskalaoprensning med Fentons reagens

* Enhedsprisen inkluderer brintperoxid, katalysator, arbejdskraft og doseringsudstyr

Det ses, at udgifterne på det foreliggende grundlag er estimeret til ca. 67 mill. kr. Overslaget er naturligvis forbundet med en del usikkerhed, og de faktiske udgifter kan gå begge veje. Antal af injektionsrunder kan fx godt være mindre end 6 runder som forudsat.

Vi har forudsat, at processen kan gennemføres med jern(II)sulfat som katalysator. Hvis yderligere forsøg viser, at der skal anvendes en modificeret jernforbindelse kan denne post stige med 5-10 mio. kr. Meget tyder dog på, at man kan nøjes med jernsulfat, da pH i depotet er meget lav.

Det skal bemærkes, at injektionsboringer foretages af det mandskab, der allerede er med i budgettet, og man bruger den borerig, som er med i udstyrsbudgettet.

Vi har i første omgang forudsat, at der skal gennemføres 6 injektionsrunder på henholdsvis 82 dage (i området med fri fase) og 108 dage i det område, hvor der ikke er fri fase. Det praktiske arbejde kan vise, at omfanget bliver mindre, hvilket først og fremmest vil spare arbejdskraft og udstyrsleje, men måske også kemikalier. Omfanget kan eventuelt også blive større, men det er mindre sandsynligt, da man i overslaget har kalkuleret med et relativ stor injektion.

Der kan eventuelt også spares på udstyrsposten, hvis man anvender en dansk entreprenør med dansk udstyr.

Der kan eventuelt også spares på supervisor indsatsen, hvis den amerikanske superviser kan styre to injektionsrunder ad gangen. Han vil i så fald i et vist omfang kunne erstattes af dansk arbejdskraft til billigere takst.

5.10 Vurdering af nødvendig oprensningstid

Med op til 6 injektionsrunder i hvert område med 2 måneders mellemrum mellem hver injektion vil selve injektionen alene vare 2-3 år, og det forudsætter endda, at der injiceres i begge områder samtidig eller med en vist overlapning. Planlægning, udbud og etablering af injektionsboringer vil ligeledes tage ca. ½ år. Den efterfølgende monitering og rapportering vil tage ca. ½ år. Det samlede tidsforløb vil da være 3 - 5 år.

6 Forslag til pilotforsøg

• 6.1 Formål
• 6.2 Forundersøgelser
• 6.3 Testområde
• 6.4 Projektbeskrivelse
• 6.5 Drift og kontrol
• 6.6 Monitering
• 6.7 Sikkerhed og sundhed
• 6.8 Miljøpåvirkninger
• 6.9 Økonomi
• 6.10 Tidsplan for drift og monitering

6.1 Formål

Det overordnede formål med pilotforsøg med Fentons reagens er at vurdere, om metoden er en cost-effektiv metode til oprensning af forureningen på Høfde 42. Pilotprojektet har følgende delformål:

• Kvantificere effekten og udbredelse af behandling af organofosforpesticider og beslægtede stoffer fra Høfde 42 på feltskalaniveau
• Bestemme influensradius af injektionen
• Bestemme det optimale injektionstryk
• Undersøge stabiliteten af brintperoxid i grundvandszonen
• Bestemme antallet af injektioner for at nå oprensningsmålet

6.2 Forundersøgelser

De tidligere begrænsede laboratorieforsøg på grundvandsprøver viste, at Fentons reagens er en lovende metode på lokaliteten, men forsøgene gav ikke alle nødvendige data til at designe et effektivt pilotforsøg. Det foreslås derfor, at der udføres supplerende laboratorieforsøg til predesign af pilotforsøget. Formålet med de supplerende forsøg er:

• Bestemme jordens oxidationsbehov (NOD). Dette er en nøgleparameter for at vurdere hvor meget iltningsmiddel, der skal tilsættes grundvandszonen. De tidligere test blev kun udført på grundvandsprøver. Testene skal udføres på jordprøver fra lokaliteten
• Forsøg med traditionel Fentons reagens. De tidligere forsøg blev udført med modificeret Fentons reagens ved neutral pH ved tilsætning af kompleksdanner (chelating agent) for at bibeholde Fe(II)-katalysatoren opløst. Traditionel Fentons reagens udføres ved lav pH uden tilsætning af kompleksdanner. Da grundvandet allerede har lavt pH på lokaliteten, kan traditionel Fenton oxidation muligvis anvendes uden brug af pH-regulering. Traditionel Fenton er relevant, da det er en billigere metode end den modificerede metode
• Forsøg med at reducere dannelsen af MP2-syre. Dannelsen af MP2-syre kan evt. skyldes de tilsatte reagenser, som blev tilsat ved laboratorieforsøgene (fx fosfat). De supplerende laboratorieforsøg skal klarlægge om der kan tilsættes andre katalysatorer, som hindrer dannelse af MP2 syre
• Optimering af dosering af Fentons reagens. De indledende forsøg var ikke målrettet mod at finde den optimale dosering for de lokalitetsspecifikke forhold. Det foreslås derfor at udføre supplerende forsøg for at bestemme den mest cost-effektive dosering af Fentons reagens
• Evt. behandling af DNAPL. Da en væsentlig mængde af forureningsmassen findes på ikke opløst form, foreslås supplerende test for at undersøge, om Fentons reagens kan behandle DNAPL. Disse forsøg er ikke medtaget i det økonomiske overslag.

Det forudsættes, at analysearbejdet af pesticiderne udføres af Cheminova.

6.3 Testområde

Der foreslås et testområde på ca. 100 m². Testområdet placeres i område med relativ kraftig jord- og grundvandsforurening og med mindre mængder af fri fase (i depot/nedsivnings-område eller indre randzone).

6.4 Projektbeskrivelse

Der udføres 4 injektionsboringer som placeres med ca. 5 m’s afstand, jf. figur 6.1. Hver injektionsboring indrettes med stålfiltre i top og bund af øvre magasin med ca. 2 m filterlængde. Alternativt anvendes ”direct push” med specialudstyr, som findes tilgængeligt i Danmark. Der udføres 3-5 injektioner i hver med 3-4 dages varighed og med 6-8 ugers mellemrum. I det økonomiske overslag er der forudsat 4 injektioner. Fentons reagens vil blive tilsat i en koncentration på 12-17 %. Ca. 1,9 m³ af Fentons reagens vil blive injiceret i hver boring i hver injektionsrunde. Det totale volumen af Fentons reagens vil være 30,4 m³ over 4 injektionsrunder.

Figur 6.1 Konceptuel design af pilotforsøg

Valg af Fenton katalysator (traditionel eller modificeret) vil blive bestemt på baggrund af de supplerende laboratorieforsøg.

Injektionsboringer udbygges efter principper på figur 5.1, dvs. med stålfiltre og afpropning med cementstabiliseret bentonit eller beton.

Fentons reagens vil blive tilsat under lavt til moderat tryk (0-4 bar og helst under 2,5 bar) for at få fordelt injektionsvæsken så ensartet som muligt i grundvandszonen. Dette giver erfaringsvis en god fordeling. Der er dog en vis risiko for kanalstrømning (preferential flow), men dette er svært at forudsige. Der forudsættes et injektionsflow på 10-15 l/min. Højere injektionsflow og tryk kan vise sig at være nødvendigt og vil blive vurderet under pilotforsøget.

I nogle af injektionsboringerne tilsættes konservative tracere (fx lithium eller fluorid) i injektionsvæsken for at bestemme influensradius.

Monitering vil blive udført med en kombination af nye og eksisterende boringer i varierende afstande fra injektionsboringerne.

6.5 Drift og kontrol

Det foreslåede pilotforsøg kræver kun lidt drift. Behandlingen vil ske gennem 3-5 injektioner, som hver varer 2-3 dage. Mellem de enkelte injektioner vil der ikke være nogen drift eller kontrol. Gennem injektionsperioden skal der gennemføres kontrol af injektionen (koncentration og flow) for at sikre den optimale oprensning. Der skal også ske kontrol af tryk i injektions- og moniteringsboringer. Det skal ligeledes kontrolleres, at der ikke sker nogen uhensigtsmæssig afdampning af giftige stoffer. Det bør endvidere foretages en løbende temperaturmåling (med dataopsamling) i et par boringer for at få en idé om varmeudviklingen ved processen.

6.6 Monitering

Der udtages grundvandsprøver før og efter de enkelte behandlinger. Det foreslås, at der udtages prøver hver måned efter den første behandling. Vandprøver analyseres for pesticidkomponenter (herunder nedbrydningsprodukter – Cheminova pakken), uorganiske oxidationsprodukter (nitrat, sulfat og fosfat) samt COD og TOC.

Koncentrationen af brintperoxid i grundvandet måles for at vurdere stabiliteten og udbredelsen heraf.

Der vil også blive målt for tracer, som tilsættes i injektionsvæsken i udvalgte boringer. Der udtages vandprøver fra udvalgte boringer for at vurdere spredning af tracer og influensradius af injektionen.

Det foreslås også at udtage jordprøver før den første injektion og efter den sidste injektionsrunde for at kvantificere oprensningseffekten af den sedimentbundne forurening og fri fase. Jordprøver analyseres for Cheminova-pakken.

6.7 Sikkerhed og sundhed

De væsentligste forhold vedrørende sikkerhed og sundhed er:

• kontakt med giftige stoffer ved prøvetagning og entreprenørarbejde. Den primære risiko vil relatere sig til forureningen af sedimentet med de stærkt humantoksikoloiske fosforinsekticider (organofosfater); parathion, methylparathion og malathion
• kontakt med injektionsvæske (primært brintperoxid) ved transport, opbevaring, og håndtering ved injektion
• risiko for optrængning af den injicerede væske til overfladen
• risiko for ukontrolleret spredning/lækage af pesticider til overfladen (grundvand, dampe)

Optrængning af gasser og væsker til overfladen er flere gange sket med Fentons reagens ved dårligt designede projekter. Dette kan især ske ved injektionsboringer eller ved for højt injektionstryk. På Høfde 42 vil risikoen for optrængning af Fentons reagens og flygtige stoffer være meget lille da der er etableret en plastmembran (lossepladsmembran). Det skal dog vurderes, om der skal ske udluftning af injektionsboringerne eller ventilation under membranen for at forhindre for stort overtryk.

I arbejdsplanen vil der blive udarbejdet retningslinier for injektionen, herunder at holde tryk under 2,5 bar. Tryk vil blive målt ved hver injektion, og det vil være muligt at udlufte boringer (til speciel opsamlingsbeholder med aktivt kul eller direkte til atmosfæren), hvis trykket bliver for stort.

Som et led i pilotforsøget vil der blive lavet en udførlig sikkerheds- og sundhedsplan for arbejdet, herunder håndtering af arbejder med forureningen og injektionsvæsken.

Fosforinsekticiderne er kendetegnet ved en stærk ram lugt af rådne æg eller hvidløg. Lugtgener vil udgøre en arbejdsmiljømæssig gene i forbindelse med projektet, og generne kan i sig selv medføre symptomer med utilpashed. Den kraftige lugt fra fosforinsekticiderne har dog den afledte positive effekt, at ingen beskæftigede på pladsen vil være i tvivl om, hvornår der er risiko for at være eksponeret for en sundhedsskadelig påvirkning fra de pågældende stoffer.

Fosforinsekticiderne optages meget let i kroppen ved indånding, ved kontakt med øjne og hudoverflader og ved direkte indtagelse til lige gennem mave-/tarmkanalen.

Det vurderes på forhånd at beskæftigede indenfor depotområdet i vid udstrækning skal benytte:

• Kemikaliresistente heldragter
• Kemikaliresistente handsker og støvler med tæt overlapning til heldragten
• Heldækkende øjen- og åndedrætsværn

Det forventes, at der ved injektionen vil være 3-4 beskæftigede personer i depotområdet. Arbejdspladsen kan således ikke betegnes som mandskabstung, hvilket helt klart vil lette og styrke sikkerhedsarbejdet og -koordineringen. Det vurderes derfor, at risikoen for svigt i sikkerhedsarbej-det er relativ lille.

6.8 Miljøpåvirkninger

Pilotforsøget vurderes ikke at give nogle væsentlige negative effekter på det omgivende miljø. Der kan ske en vis afgasning af flygtige stoffer på grund af opvarmning af grundvandszonen. Disse dampe kan evt. opsamles gennem dræn med aktivt kulfilter. Der kan ligeledes ske en vis mobilisering af metaller, men dette vurderes ikke at være noget problem, da der ikke vil ske nogen spredning pga. spunsvægen.

6.9 Økonomi

De økonomiske overslag for pilotprojektet er noget usikre. Udgifterne vil afhænge af, om projektet kan kombineres med de andre pilotforsøg mht. borearbejde, prøvetagning mm. Det anbefales, at GeoSyntec og ISOTEC indgår med specialistviden og bistand, herunder design af pilotforsøg samt bistand ved injektion. Udgifterne til hertil kendes dog ikke, da det vil afhænge af omfanget af deres arbejde. Vi forestiller os, at ISOTEC skal deltage og instruere ved opstart af forsøget. Herefter kan et dansk firme foretage de følgende injektioner, hvor der måske kan benyttes personale, som også deltager i de øvrige pilotforsøg. Hvis ISOTEC skal til Danmark mere end den ene gang, vil det fordyre projektet betydeligt. Det vurderes på nuværende tidspunkt, at omkostningerne vil være i størrelsesordenen 1,1 - 1,75 mio. kr. excl. moms. I tabel 6.1 er anført de væsentligste poster ved et pilotforsøg.

Der er ikke medtaget udgifter til forundersøgelser med Fentons reagens. Det vurderes, at disse kan udføres for ca. 60.000 - 120.000 kr excl. moms afhængig af hvilke forsøg, der skal udføres.

Tabel 6.1 Overslag over udgifter til pilotforsøg (2007 priser)

1) det forudsættes, at Cheminova-pakke analyseres af Cheminova uden omkostninger

Vi forestiller os ikke, at pilotanlægget skal sendes i udbud, da det er en relativ lille opgave, der kan udføres mest smidig og billigst ved, at vi finder en kvalificeret entreprenør samt foretager de nødvendige indkøb af udstyr. Såfremt Miljøstyrelsen ønsker, at COWI skal udarbejde tilbudsmateriale, skal denne post med i budgettet ud over de øvrige poster i tabel 6.1

6.10 Tidsplan for drift og monitering

Tabel 6.2 viser forslag til tidsplan for pilotforsøg. Det samlede tidsforbrug er ca. 15. måneder.

Tabel 6.2 Tidsplan for pilotprojekt

Referenceliste

/1/: A DNAPL Hotspot of organophosphoros pesticides. Høfde 42 Harboøre Tange. Ringkøbing County, October 2004.

/2/: Beregning af forureningsmasse. Høfde 42, Harboøre tange. NIRAS og Ringkøbing Amt, 22. november 2005.

/3/: Supplerende forureningsundersøgelse, Høfde 42 – Harboøre Tange, Ringkøbing Amt og NIRAS, 14. juni 2005.

/4/: Estimering af udsivning til Vesterhavet. HØFDE 42. Høfde 42 – Harboøre Tange, Ringkøbing Amt og NIRAS, 22. nov. 2005.

/5/: Høfde 42, Status over forureningssituationen ved høfde 42 på Harboøre Tange. Ringkøbing Amt, 2003.

/6/: Evaluering af nul-valent jern til oprensning af høfde 42 depotet. Eksamensprojekt af Annnika Fjordbøge, DTU. 1. november 2005.

/7/ Ringkjøbing Amt, januar 2001: Notat om undersøgelse af forurenings-situationen ved høfte 42 og "cheminovahullet" på Harboøre Tange 2001

/8/ Pugliese, P., Molto, J. C., Damiani, P., Marin, R., Cossignani, L., and Manes, J. (2004). Gas chromatographic evaluation of pesticide residue contents in nectarines after non-toxic washing treatments. Journal of Chromatography A. 1050, 185-191

/9/ Interstate Technology and Regulatory Council. 2005. Technical and Regulatory Guidance for In Situ Chemical Oxidation of Contaminated Soil and Groundwater. http://www.itrcweb.org/Documents/ISCO-2.pdf

Bilag 1 Lokaliteten

Bilag 1.1 Jordforurening

Bilag 1.2 Grundvandsforurening

Bilag 1.3 Fri fase udbredelse

Bilag 1.4 Områdeinddeling

Bilag 2 Litteraturstudie

Literature Review

In Situ Chemical Oxidation of Organophosphorous Pesticides in Groundwater

Prepared by:

10015 Old Columbia Road, Suite A-200
Columbia, Maryland 21046

and:

Department of Soil & Water
The Connecticut Agricultural Experiment Station
P.O. Box 1106
New Haven, CT 06504

GeoSyntec Project Number MR0487
GeoSyntec Document Number MD06272

April 2000

Table of Contents

1 Introduction

• 1.1 Site Conditions
• 1.2 Outline of this Document

2 Key Properties of Organophosphorus Pesticides

3 Overview of Oxidants for In Situ Treatment

• 3.1 Chemistry of Oxidants
  • 3.1.1 Fenton’s Reagent
  • 3.1.2 Permanganate
  • 3.1.3 Persulfate
  • 3.1.4 Ozone/Ozone with Peroxide
• 3.2 Advantages and Disadvantages of Each Oxidant
  • 3.2.1 Fenton’s Reagent
  • 3.2.2 Permanganate
  • 3.2.3 Persulfate
  • 3.2.4 Ozone/Ozone with Peroxide
• 3.3 General Considerations for In Situ Treatment
  • 3.3.1 Oxidant Properties
  • 3.3.2 Geology
  • 3.3.3 Plume versus Source Areas
  • 3.3.4 Application/Installation Methods
• 3.4 Full Scale Design, Operations and Maintenance, Safety and Cost Considerations
  • 3.4.1 Data Requirements for Design
  • 3.4.2 O&M Requirements
  • 3.4.3 Safety
  • 3.4.4 Cost

4 Fenton’s Reagent Reactions with Organophosphorus Pesticides

• 4.1 Reaction Chemistry with Organophosphorus Pesticides and other Contaminants at the Cheminova Site
• 4.2 Reaction Chemistry with Other Organophosphorus Pesticides
• 4.3 Summary of Implications for Treatability Test

5 Permanganate Reactions with Organophosphorus Pesticides

• 5.1 Reaction Chemistry with Organophosphorus Pesticides and other Contaminants at the Cheminova Site
• 5.2 Summary of Implications for Treatability Test

6 Persulfate Reactions with Organophosphorus Pesticides

• 6.1 Reaction Chemistry with Organophosphorus Pesticides and other Contaminants at the Cheminova Site
• 6.2 Summary of Implications for Treatability Test

7 Ozone and Ozone+Hydrogen Peroxide Reactions with Organophosphorus Pesticides

• 7.1 Reaction Chemistry with Cheminova Site Organophosphorus Pesticides and other Contaminants
• 7.2 Summary of Implications for Treatability Test

8 Case Studies of ISCO Field Applications

• 8.1 Case Study #1: Fenton’s Reagent Demonstration, Savannah River Site Aiken, South Carolina
• 8.2 Case Study #2: Ozone with Peroxide, Active Retail Gas Station, Riverside, California

9 Summary and Recommendations

10 References

Tables

Table 1: Summary of organophosphorus pesticide occurrence at the Høfde 42 Site

Table 2: Summary of key properties of organophosphorus pesticides

Table 3: Redox potentials of common chemical oxidants

Table 4: Oxidant properties to be considered for in situ treatment with ISCO (modified from ITRC, 2005)

Table 5: Summary of Cheminova contaminants degraded by the Fenton reaction

Table 6: Ozone reaction with Cheminova contaminants where products have been identified

Table 7: Summary of chemical oxidation case studies

Figures

Figure 1: Typical In Situ Chemical Oxidation Application

Figure 2: Reactions of Ozone in Water in the Presence of Reactive Solutes

1 Introduction

This report presents a literature review of in situ chemical oxidation (ISCO) technologies that are being considered for remediation of organophosphorous pesticides in groundwater at the Høfde 42 Harboøre Tange (Cheminova) Site in Ringkjøbing County, Denmark (the Site). The Site is located on the western coast of Denmark, on a beach adjacent to the North Sea. GeoSyntec and Dr. Joseph Pignatello of the Connecticut Agricultural Experiment Station and Yale University have prepared this report for COWI A/S as the first phase in a project to evaluate ISCO technologies for remediation of an organophosphorus pesticide contamination in groundwater at the Site. The findings of this literature review will be used to design laboratory treatability tests to further evaluate the effectiveness of selected ISCO technologies for application at the Site.

State-of-the-science literature regarding the use of ISCO techniques for the treatment of organophosphorus pesticides in groundwater has been reviewed. Particular emphasis has been given to chemical oxidation technologies that are applied for remediation in situ, including conventional Fenton’s reagent, modified Fenton’s reagent, ozone, ozone and peroxide, permanganate, persulfate, and heat-activated persulfate. Of these technologies, the focus of this document has been on those technologies that have already been shown, through laboratory research, to effectively oxidize phosphorothionate pesticides.

1.1 Site Conditions

This literature review has focused on parathion, methyl parathion, malathion, and amino parathion, which are primary contaminants of concern (CoCs) in the source area at the Site (Ringkjøbing County, 2004). These chemicals are present as an immiscible, dense nonaqueous phase liquid (DNAPL) in the upper sand aquifer at the Site. A groundwater plume of these pesticides and their degradation products extends from the DNAPL source area towards the North Sea. The dissolved phase chemicals have also been transported into the lower sand aquifer at the Site. The conditions in this plume are highly acidic (typical pH ranges from 2.4 to 4.3). A summary table of the measured concentrations of these compounds in the DNAPL source area, as presented in an overview of the Site conditions by Ringkjøbing County, Department of Environment and Infrastructure (2004), is provided in Table 1.

It should be noted that in addition to the organophosphorus pesticides listed in Table 1, lesser concentrations of alkylated polysulfides, triethyl phosphate and mercury are also present in the DNAPL at the Site. The scope of the literature review did not include evaluating the effects of chemical oxidation upon these chemicals; however, these effects should be evaluated prior to initiating any oxidation field program at the Site. For example, mercury is a redox-sensitive metal that has the potential to be affected by an ISCO program. Geochemical modeling and/or laboratory treatability testing could be used to assess the impact of an ISCO program on the solubility and speciation of this metal.

1.2 Outline of this Document

The remainder of this report outlines the results of the literature review including:

• Key properties of organophosphorus pesticides (Section 2.0);
• An overview of the ISCO technologies reviewed in this document, including their general chemistry, considerations for in situ treatment, advantages and disadvantages of each technology, and design, operation and maintenance, safety, and cost considerations for field implementation (Section 3.0);
• Reaction chemistry of Fenton’s reagent with organophosphorus pesticides and the implications for field implementation at the Site (Section 4.0);
• Reaction chemistry of permanganate with organophosphorus pesticides and the implications for field implementation at the Site (Section 5.0);
• Reaction chemistry of persulfate with organophosphorus pesticides and the implications for field implementation at the Site (Section 6.0);
• Reaction chemistry of ozone and ozone + peroxide with organophosphorus pesticides and the implications for field implementation at the Site (Section 7.0);
• A description of selected case studies of field applications of the candidate oxidants identified from the literature review (Section 8.0);
• A summary of the results of the literature review and recommendations for bench scale and field scale testing (Section 9.0); and,
• References (Section 10.0).

2 Key Properties of Organophosphorus Pesticides

The key chemical properties of organophosphorus pesticides reviewed in this memorandum are summarized in Table 2 (Chemfinder.com, 2004; Fjordbøge, 2005; Schwarzenbach et al., 2003; Montgomery, 2000; Budavari et al., 1996). Key properties of these pesticides that impact their partitioning and transport in soil and groundwater include the following:

• They are generally liquid at room temperature, with a density in the range of 1.2-1.3 grams per milliliter (g/mL), making them a DNAPL when released as a pure phase in groundwater;
• They have low aqueous solubility and vapor pressure;
• They partition readily into organic phases. As a result, these compounds adsorb strongly to natural organic matter in soils;
• They degrade to varying extents via hydrolysis, yielding water soluble compounds. However, at sites where the groundwater is relatively acidic (i.e., pH<5) such as the Høfde 42 site, this hydrolysis reaction may be very slow to non-existent (Montgomery, 2000); and
• They biodegrade to varying extents via both anaerobic and aerobic pathways; however, many of the reaction products are toxic themselves (Montgomery, 2000) and, biodegradation can be significantly inhibited at the low pH levels that are typical of the Høfde 42 site.

3 Overview of Oxidants for In Situ Treatment

ISCO is an emerging technology for the treatment of hazardous waste. ISCO refers to a group of specific technologies that each use differing combinations of oxidants and delivery techniques. ISCO has been shown to destroy or degrade an extensive variety of hazardous wastes in groundwater and soil, including fuel hydrocarbons, chlorinated solvents (e.g., perchloroethene [PCE] and trichloroethene [TCE]), fuel oxygenates (e.g., methyl-tert-butyl-ether [MTBE]), and polycyclic aromatic hydrocarbons (PAHs). To date, the technology has not been as widely applied with pesticides; however there are a number of bench scale studies that suggest that ISCO may be appropriate for field application for certain classifications of pesticides as well. A schematic of a typical ISCO field application is shown in Figure 1.

Various oxidants have been used in laboratory and field applications to aggressively destroy organic chemicals, including Fenton’s reagent, permanganate, persulfate, ozone, and ozone combined with peroxide. These oxidants react to varying degrees with organic contaminants (i.e., breaking molecular bonds of and capturing electrons from the contaminant) and convert them into degradation products. Depending on the parent compound, the final reaction products may be innocuous compounds commonly found in nature such as carbon dioxide (CO₂), water and inorganic chloride.

The following subsections provide an overview of the general chemistry of each of the oxidants, advantages and disadvantages for each oxidant, considerations for in situ treatment, and the design, operations and maintenance, safety and cost issues for ISCO.

3.1 Chemistry of Oxidants

Because not all ISCO treatments are applicable for all contaminants, site contaminants and conditions must be understood in order to choose the appropriate oxidant and delivery method. The treatment effectiveness of chemical oxidants currently in use varies based on several factors, including the redox potential (E°) of the oxidant, and the reactive specificity of the oxidant toward a given type of contaminant. Permanganate (E° – 1.70), for example, has been shown to be primarily effective for oxidizing chlorinated ethenes, but not chloroethanes or fuel hydrocarbons. In contrast, Fenton’s reagent (E° – 2.76) is known to oxidize fuel hydrocarbons, PAHs, pesticides, polychlorinated biphenyls (PCBs), and most types of chlorinated solvents. The redox potentials of the primary oxidants currently in use are summarized in the Table 3.

In selecting an appropriate ISCO technology, an understanding of the geochemical conditions at a given site is essential since the applied reagents could be consumed by natural organic matter or dissolved iron rather than the contaminants, resulting in poorer than expected treatment. Groundwater geochemistry may also need to be adjusted to more optimal conditions prior to treatment (e.g., lowering of pH during application of conventional Fenton’s reagent at sites where groundwater pH is near neutral).

Specific details of the chemistry of each oxidant are described further in the following subsections.

3.1.1 Fenton’s Reagent

The Fenton reaction has been studied extensively in regard to waste treatment applications (Pignatello et al., 2006). Hydrogen peroxide is an effective oxidizing agent; however, to achieve contaminant reduction in a reasonable time, iron or iron salts are used as a catalyst (the combination is referred to as Fenton’s reagent). It generates hydroxyl radicals through a series of reactions with hydrogen peroxide catalyzed by iron ions, which undergo a redox cycle between the +II and +III oxidation states. The hydroxyl radicals (OH^•) serve as powerful, effective and nonspecific oxidizing agents. The mechanism is complex but can be summarized by the following steps:

Fe(II) + H₂O₂ → Fe(III) + OH^- + HO^• (1)

Fe(III) + H₂O₂ → Fe(II) + HO₂^• + H⁺ (2)

HO^• + R-H → H₂O + R^• (3)

where R-H is an organic compound. The optimum pH is slightly less than 3, and rates normally drop precipitously above pH 4 (Pignatello et al., 2006).

There are several variants of the Fenton reaction that mainly contribute to the regeneration of soluble Fe(II). One of the most important is the photo-assisted Fenton, or photo-Fenton reaction, shown below, which is initiated when the solution is irradiated with ultraviolet (UV) or UV/visible light.

Fe(III)-L + hv → Fe(II) + L^• (L = OH^- or organic ligand) (4)

Although photolysis is not possible for in situ treatment, studies that involve photo-assistance are included in Section 4 for their relevance to the dark reaction. Photo-assistance generally increases the rate. Moreover, it often alters the product distribution, since some of the products undergo further oxidation because they form photo-labile complexes with Fe(III).

Low pH conditions are often impractical to maintain under field conditions due to the enormous buffering capacity associated with most native soils. In another variation of the Fenton reaction, it has been demonstrated that free radical generation and contaminant oxidation can be promoted at neutral pH using a modified Fenton’s process that uses hydrogen peroxide, Fe(II), and chelating agents (e.g., citric acid or nitrilotriacetic acid) to keep iron in solution without the need for acid pH conditions (Watts et al., 1999).

Fenton’s reagent produces a strongly exothermic reaction with a very short half-life. As a result, the persistence of Fenton’s reagent in the subsurface is relatively short, and the ROI of injected Fenton’s reagents is characteristically low.

3.1.2 Permanganate

Permanganate is an oxidizing agent that has an affinity for oxidizing organic compounds, particularly those containing electron-rich carbon-carbon (C=C) double bonds, aldehyde groups or hydroxyl groups. The reaction between permanganate and chlorinated ethenes involves an electrophilic attack on the ethene’s C=C double bonds and the formation of a cyclic hypomanganate ester. Rapid hydrolysis of the cyclic ester results in the production of CO₂. There are two permanganate salts, potassium permanganate (KMnO₄) and sodium permanganate (NaMnO₄). The half reaction for MnO₄^- for pH in the range of 3.5 to 12 is:

MnO₄^- + 2H₂O + 3e^- → MnO₂(s) + 4OH^- (5)

This reaction indicates that oxidation by MnO₄^- at neutral pH is accompanied by the production of manganese dioxide (MnO₂) solid and release of hydroxide (OH^-). However, under acidic conditions (for pH less than 3.5) the dominant half reaction for MnO₄^- is:

MnO₄^- + 8H⁺ + 5e^- → Mn²⁺ + 4H₂O (6)

Thus, under acidic conditions, hydrogen is consumed to produce water and Mn (II) is yielded.

In comparison to Fenton’s reagent, permanganate is a weaker oxidant, but it has a longer reactive half-life. Consequently, it is easier to control delivery and distribution of permanganate in the subsurface, and permanganate is more amenable to recirculatory designs for source area remediation.

3.1.3 Persulfate

Persulfate (a.k.a., peroxodisulfate; S₂O₈^2-) has received attention recently as a potential oxidant for ISCO treatment of organic contaminants. Persulfates are common oxidants used in plating, organic chemical synthesis, polymerization, and metal surface cleaning. The sodium salt (Na₂S₂O₈) is highly soluble in water (730 grams per liter (g/L) at 25 degrees Celsius (°C) and 860 g/L at 50°C) and can easily form a concentrated solution for subsurface delivery. Persulfate salts dissociate in water to persulfate anions which, although strong oxidants, are kinetically slow in destroying many organic contaminants. For ISCO applications, potassium persulfate has a low solubility, and the injection of ammonium persulfate may lead to the generation of ammonia, which is a regulated CoC in groundwater. Therefore, the most common salt used is sodium persulfate. The persulfate anion is a more powerful oxidant than hydrogen peroxide. Decomposition reactions vary with persulfate concentration, pH, and oxygen, and hydrogen peroxide or peroxymonosulfate can be produced. Under dilute acid conditions, hydrolysis of the persulfate anion yields bisulfate anions and hydrogen peroxide.

At ambient temperatures (15 to 20°C), the persulfate ion can act as an oxidant:

S₂O₈^-2 + 2e^- → 2SO₄^-2 E° = 2.01 V (7)

Thermal homolysis (Eq. 8) of persulfate or its reaction with a suitable reductant, such as ferrous ion (Eq. 9-10), leads to the sulfate radical, which dramatically increases the oxidative strength of persulfate. Manganese or copper can also be used as reductants for formation of the sulfate radical.

heat + S₂O₈^2- → 2 SO₄^{- •} (8)

Fe²⁺ + S₂O₈²^• → Fe³⁺ + SO₄^{- •} + SO₄²^- (9)

Fe²⁺ + SO₄^{- •} → Fe³⁺ + SO₄²^{- •} (10)

SO₄^{- •} + H₂O → SO₄²^- + OH^• (11)

3.1.4 Ozone/Ozone with Peroxide

Ozone (O₃) can be used alone or in combination with peroxide to form one of the strongest oxidants available for ISCO. Ozone can oxidize organic contaminants either by direct oxidization by ozone or generation of free radical intermediates. The hydroxyl radicals are nonselective oxidizers that rapidly attack organic contaminants and break down their carbon-to-carbon bonds. Ozone can oxidize compounds such as aromatics and chlorinated alkenes, although oxidation by hydroxyl radicals is faster than oxidation by ozone itself.

Most of the literature on the use of ozone alone or in combination with hydrogen peroxide has been in water rather than in soil treatment. Ozone is reactive with certain functional groups, but in water ozone also produces hydroxyl radicals that often account for most of the reactivity (Figure 2).

Hydrogen peroxide is frequently added to ozonated water to increase the ozone decomposition rate. By accelerating the ozone decomposition rate, the hydroxyl radical concentration is elevated, which increases the overall oxidation rate. The balance of the reactions generated is shifted to increase the contribution of indirect oxidation by the hydroxyl radicals over the direct ozone oxidation, which may be preferable for some target compounds.

3.2 Advantages and Disadvantages of Each Oxidant

The advantages and disadvantages of the various chemical oxidation methods are as follows:

3.2.1 Fenton’s Reagent

Advantages:

The oxidant materials are inexpensive and readily available.

A wide range of chemicals can be treated including chlorinated solvents, fuel hydrocarbons, coal tar, PCBs, and PAHs.

Disadvantages:

The technology is limited by the interference of subsurface impurities and carbonate since bicarbonate and organic matter will create competing reactions that hinder performance.

An extremely exothermic reaction occurs, which can create safety and handling issues. Ground heaving and surface damage is possible for poorly designed applications.

For conventional Fenton’s applications, addition of concentrated acid is required, and groundwater pH post-treatment can be quite low (<5). In this case, treatment with Fenton’s reagent can effectively sterilize the soil and limit secondary treatment choices.

As a result of the short reaction half-life, the ROI of the reagent can be relatively small, thereby requiring a higher density of injection wells relative to other ISCO options.

Associated heat and bubbling can cause significant volatilization of volatile target contaminants, thereby necessitating supplemental use of soil vapor extraction to capture vapors in some cases.

3.2.2 Permanganate

Advantages:

It is typically more stable and safer to handle than Fenton’s reagent, does not require pH adjustment and produces less heat and insoluble gas in the treatment zone.

The relatively long reaction half-life (lower reactivity) of permanganate allows for flushing of treatments throughout subsurface, improved delivery of oxidant (e.g., relative to Fenton’s reagent), and greater ability to oxidize contaminants diffusing from the aquifer matrix.

Disadvantages:

Permanganate treats a narrower range of contaminants than the other oxidants. Although it can treat chlorinated ethenes, permanganate is not effective at treating chlorinated ethanes and may have limited effectiveness against benzene, toluene, ethylbenzene, or xylenes (BTEX).

Permanganate can be expensive.

Permeability reductions can occur near DNAPL source zones due to the formation of MnO₂ precipitates (e.g., MacKinnon and Thomson, 2002; Dai and Reitsma, 2002; Lee et al., 2003) and/or rapid production of CO₂(g) (Dai and Reitsma, 2002), resulting in less effective treatment over time.

Recirculation systems are prone to fouling with MnO₂ precipitates.

Strongly oxidizing conditions are created that can persist post-treatment, which may impact the effectiveness or choice of polishing technology (if required).

Dissolved metals mobilization may occur in some aquifers, depending on the mineral content of the geological material present.

Manganese precipitated as MnO₂ may mobilize as dissolved manganese if the groundwater geochemistry becomes reducing upon termination of the treatment.

3.2.3 Persulfate

Advantages:

Unlike permanganate, persulfate does not result in the accumulation of manganese, a constituent which could become a CoC over the long-term if MnO₂ dissolves.

Sulfate is a primary end-product of ISCO with persulfate, and sulfate can serve as an electron acceptor to facilitate subsequent degradation of any co-occurring fuel hydrocarbons residual.

A wide range of chemicals including chlorinated solvents, fuel hydrocarbons, and PAHs can be treated.

Disadvantages:

Persulfate typically requires the use of heat or ferrous iron catalyst. Ferrous ions require low pH or chelating agents to remain in solution. It may be necessary to lower the pH as with peroxide systems to achieve this environment.

Fe(II) does not appear to effectively activate persulfate with chlorinated ethanes (1,1,1-trichloroethane [1,1,1-TCA], etc.) and methanes (chloroform, etc.). However, recent work with persulfate under alkaline conditions demonstrates effectiveness against these contaminants.

The catalytic effect of the iron appears to decay with time and distance from injection. This decrease could be the result of either poor transport of the dissolved Fe(II) in a soil environment or the depletion of the iron as it activates the persulfate. Chelated iron effectively increases the iron solubility and longevity of Fe(II) in the groundwater.

Low pH conditions may be generated by persulfate decomposition, which can cause dissolved metal concentrations to increase in the groundwater. Natural soil buffering capacity can help alleviate this phenomenon.

Persulfate may degrade soft metals such as copper or brass. Well construction and injection materials should be compatible with long-term persulfate exposure. Appropriate materials include stainless steel, high-density polyethylene, and polyvinyl chloride (PVC).

As with all oxidants, metals can be mobilized within the treatment zone due to a change in oxidation states and/or pH.

3.2.4 Ozone/Ozone with Peroxide

Advantages:

The gaseous nature of ozone allows for ease of delivery through the vadose zone compared with the liquid oxidants (Looney and Falta, 2000).

Disadvantages:

The short half-life of ozone substantially limits its ability to migrate through the soil, thus this oxidant is generally considered useful only for small scale or vadose zone applications.

Even with the addition of peroxide, there is still a short reaction half-life and the ROI of the reagent can be relatively small, thereby requiring a relatively high density of injection wells.

An ozone generation system requires a large capital investment.

Ozone can create an indoor air quality issue.

Ozone is highly reactive with aquifer solids and groundwater constituents, yielding a high oxidant demand.

As with all oxidants, metals can be mobilized within the treatment zone due to a change in oxidation states and/or pH.

As with Fenton’s reagent, ozonation can be expected to decrease the indigenous microorganisms within the treated site; however, microbial populations are expected to rebound at a rate that depends inversely on the duration of ozonation (Jung et al, 2005).

3.3 General Considerations for In Situ Treatment

The considerations to be made when evaluating ISCO as a remedy for a site include the properties of the oxidant, the hydrogeology, and the type and distribution of contaminants in the subsurface. Based upon these data, an evaluation of the most appropriate oxidant and application method can be made. These considerations are described in the following subsections. The specific reactions between the oxidants and the CoCs for the Site are further described in Sections 4-7.

3.3.1 Oxidant Properties

Table 4 presents key oxidant properties to be considered for ISCO. The chemical and physical properties of chemical oxidants vary widely. As such, the design considerations for ISCO remedies differ significantly between the various types of chemical oxidants. For example, ISCO using Fenton’s reagent is an unstable, exothermic treatment process, while treatment using permanganate is a stable process that does not result in heat generation. The half-life for Fenton’s reactions is on the order of seconds to minutes; while the half-life for permanganate and persulfate reactions is on the order of days to months. Similarly, ozone is highly reactive and unstable, with a short reaction half-life. ISCO remedies that use Fenton’s reagent in DNAPL source areas typically involve 3 to 5 batch injections of reagent over a 4 to 12 month period. In contrast, ISCO remedies that use permanganate or persulfate may involve batch injection or continuous recirculation of reagents throughout the treatment zone.

Applications of Fenton’s reagent in the field must recognize the potential for heat and gas generation, ground heaving (e.g., proximal to asphalt), and corrosion of materials in the subsurface. Conventional Fenton’s reagent remedies require low-pH conditions, which can be engineered via the addition of acids. Alternatively, neutral-pH Fenton’s formulations can be used that employ chelating agents that increase the solubility of Fe(II) and Fe(III) and prevent precipitation of iron oxides at circumneutral pH.

Most of the ISCO reagents can result in the mobilization of metals from aquifer solids, depending on the composition of the aquifer matrix. A primary end-product of ISCO using permanganate is the precipitation of MnO₂, which may accumulate and occupy a significant fraction of the aquifer pore space. ISCO using persulfate generates sulfate, an innocuous end product that can be subsequently used to stimulate biodegradation of any remaining organics. Similarly, Fenton’s reagent generates iron oxides (rust), which may subsequently stimulate anaerobic ferrogenic biodegradation of residual contamination.

3.3.2 Geology

ISCO has been applied in both unconsolidated and consolidated media. To effectively degrade contaminants, the oxidant must come into contact with the contaminant molecules. As with all remediation technologies that require the delivery of an amendment to the treatment zone, the more heterogeneous the media the more difficult it is to effectively distribute the amendments. Subsurface heterogeneities, preferential flowpaths, or low soil permeability can result in uneven flushing of the oxidant through the subsurface, resulting in untreated contaminants. To properly design and implement ISCO the treatment site must be adequately characterized including determining the nature and mass of contaminants (including the sorbed, dissolved and/or non-aqueous phases), having an understanding to the geology (including migration pathways for the contaminants), and an understanding of the hydrogeology of a given site.

3.3.3 Plume versus Source Areas

ISCO can be applied over a range of contaminant concentrations from source area to plume concentrations. Chemical oxidants treat contaminants in the dissolved phase; however, as degradation of the aqueous phase contaminants occur, enhanced desorption of the sorbed contaminants and dissolution of nonaqueous phase liquid (NAPL) contaminants can occur. The remedial design must account for the dissolved-phase, non-aqueous phase and sorbed mass for effective site remediation. For some ISCO technologies (e.g., Fenton’s reagent), source area applications are the most suitable use of the technology as a result of the short reaction half-life. For other ISCO technologies such as permanganate, both source and plume application are possible because of the better reactivity longevity of the oxidant. For certain plume applications, treatment using ISCO may be prohibitively expensive due to the size of the plume, or lower cost of other treatment alternatives (e.g., bioremediation). Chemical oxidation accelerates the remediation of NAPL source areas through treatment of dissolved phase contamination near the NAPL/water interface. The destruction of the dissolved-phase contaminants enhances the dissolution gradient at the NAPL/water interface, increasing the overall mass transfer of the contaminant from the NAPL to the dissolved phase, and thus depleting the NAPL at a faster rate. The maximum enhancement of NAPL removal is primarily based upon the total effective surface area over which oxidation can occur and the reaction rate. Other factors that influence NAPL removal include effective delivery of the oxidant to the contaminated media, consumption of the oxidant by other organic material in the aquifer, and the contaminant solubility.

3.3.4 Application/Installation Methods

In general, ISCO amendments are introduced to the subsurface through a number of injection wells or temporary/direct-push injection points. To achieve adequate contact between the oxidant and the contamination, an adequate fraction of the pore-volume of the target area must be filled or flushed with the oxidant. However, care must be taken not to displace the contamination with excessive amounts of oxidant injection. Once laboratory treatability testing has been conducted to ensure that ISCO is an appropriate remediation technology for a site, pilot studies can be conducted to provide the necessary information for full-scale design including the appropriate injection well/point spacing and appropriate injection flow rates for liquid or gas amendment delivery.

Oxidant concentrations need to be high enough to meet the natural oxidant demand of the aquifer, as well as the demand of any contaminant encountered during flushing. However, excessive concentrations of oxidant are not desirable due to potential impacts on secondary groundwater quality (e.g., color, pH, dissolved metals) and higher costs.

3.4 Full Scale Design, Operations and Maintenance, Safety and Cost Considerations

3.4.1 Data Requirements for Design

Location of all underground utilities;

Vertical and horizontal delineation of contaminant distributions in zone requiring treatment;

Vertical characterization of geology within treatment zone, to understand where high-flow layers exist; design needs to consider location of preferential flow paths. For Fenton’s applications, existence of preferential flow paths (e.g., in fractured bedrock) can significantly impact fate and transport of injected oxidants and off-gas;

Characterization of aquifer geochemistry, including any reduced metal species that are susceptible to oxidation due to chemical oxidants;

Hydraulic conductivity of zone requiring treatment (may require slug testing and/or aquifer pump testing);

Estimate of groundwater flow rates and travel times under operating conditions;

Soil oxidant demand test (can be screened in the field, using test kits, and/or fully studied in the laboratory);

In some cases, laboratory bench test to confirm treatability, determine site-specific treatment rates, and potential for mobilization of oxidizable metals (e.g., manganese and chromium) from aquifer solids; and

In some cases, pilot testing to confirm ROI, fate of injected oxidants, mobilization of metals, and overall treatment performance. Results from pilot testing are used to support design of a full-scale system.

3.4.2 O&M Requirements

For batch injection approaches, operation and maintenance (O&M) requirements are minimal because remedy implementation is rapid and of short duration;

For continuous injection or recirculation systems (e.g., permanganate recirculation in a NAPL source area), O&M can include replacement of in-line filters, maintenance of pumps, redevelopment of injection and extraction wells, maintenance of oxidant supply tanks, monitoring of water level in extraction and injection wells, etc.;

Performance monitoring (e.g., quarterly) to assess treatment performance, fate of oxidants, and mobilization of dissolved metals; and

For batch injection approaches, three to five injections might be required (as determined by performance monitoring) for treatment objectives to be achieved.

3.4.3 Safety

During the application of ISCO there are a number of health and safety considerations which are unique to these technologies, including the following (ITRC, 2005):

Safe storage and handling of the oxidants is essential.

If permanganate and persulfate are purchased as a solid powder, the powder must be controlled to control the potential for inhalation as it is harmful to the respiratory system.

Ozone requires careful monitoring as it is toxic to breathe and increases the flammability of many materials.

Ozone generation can require using high-voltage equipment.

Fenton’s reagent and persulfate can require the injection of multiple reagents for mixture in the subsurface. Proper injection equipment must be used to minimize the potential for above ground reactions, which can be exothermic or explosive.

If underground utilities exist there is a potential for preferential migration of oxidants and/or contaminants in the subsurface.

3.4.4 Cost

The major cost items that should be included in a cost estimate for ISCO are pre-treatment activities (laboratory and/or pilot study work), fixed cost items including injection point and sampling well installation; and, variable cost items including site supervision, chemicals, sampling and analysis for process control and O&M. For oxidants that are added in a batch approach, the major cost items are generally the installation of the injection points and the chemical reagents. For oxidants added in a recirculation approach the major cost items are generally the process equipment required for recirculation and the chemical reagents. Development of cost estimates for field work is generally developed based upon the results of laboratory testing which is used to confirm the reaction rates of CoCs and the oxidant demand of the site matrix. Costs for example Fenton’s applications are provided by U.S. Department of Energy ([USDOE]1999a), Yin and Allen (1999), and ITRC (2005).

4 Fenton’s Reagent Reactions with Organophosphorus Pesticides

4.1 Reaction Chemistry with Organophosphorus Pesticides and other Contaminants at the Cheminova Site

Among the hundreds of papers on the Fenton reaction there are relatively few dealing specifically with the Cheminova CoCs, and only one discusses treatment in soil media. Among the Cheminova CoCs, rate constants for the elementary reaction with hydroxyl radical (Eq. 3) are known only for 4-nitrophenol (k_OH = 3.8 × 10⁹ per mol per second [M^-1s^-1]), its conjugate base 4-nitrophenoxide ion (k_OH = 7.6 × 10⁹ M^-1s^-1), and dimethyl phosphate (k_OH = 1.2 × 10⁸ M^-1s^-1). (Radiation Chemistry Data Center of the Notre Dame Radiation Laboratory, http://allen.rad.nd.edu/). The former is a likely byproduct of methyl parathion and parathion oxidation, while the latter is a likely byproduct of methyl parathion oxidation.

Employing the dark Fenton reaction, Doong and Chang (1998) observed only 20% loss of malathion after 24 hours. However, in the photo-Fenton reaction, malathion was 94% reacted in 30 minutes (min) giving quantitative yield of sulfate, 35% yield of phosphate, and no loss of total organic carbon (TOC) (Huston and Pignatello, 1999). Organic products included formate, oxalate and acetate. In another photo-Fenton study (Doong and Chang, 1998), 1 gram per liter (g/L) Fe⁰ (as iron powder) gave comparable results as 50 micromoles (mM) Fe²⁺.

Methyl parathion was rapidly degraded by the photo-Fenton reaction (Pignatello and Sun, 1995). Under initial conditions listed in Table 5, methyl parathion reacted completely in 5 min, giving stoichiometric yields of sulfate and nitrate within that time. Phosphate was evolved stoichiometrically within 30 min. The transient organic intermediates identified included 4-nitrophenol, methyl paraoxon, dimethyl phosphate and oxalate. 4-Nitrophenol disappeared within 5 min. Under the same conditions, a commercial standard of uniformly-¹⁴C-labelled 4-nitrophenol evolved nitrate and ¹⁴CO₂, both quantitatively. A standard of dimethyl phosphate was degraded with a half-life of ~12 min. Importantly, maximum methyl paraoxon yield from methyl parathion was <<1%. Degradation of a commercially available analog of methyl paraoxon, ethyl paraoxon, degraded with a half-life of ~2 min, giving 4-nitrophenol as the major organic product.

Methyl parathion was treated in soil slurries under Fenton conditions using a chelated form of Fe(III) (Pignatello and Day, 1996). Since inorganic ferric ion is insoluble above pH ~3 (well below the normal pH of the soil, ~ 6), the chelating agent served to enhance the soluble concentration of Fe(III) and, thus, increase its availability in the Fenton reaction (Pignatello and Baehr, 1994; Nam et al., 2001). Two chelating agents were investigated: nitrilotriacetate (NTA) and N-(2-hydroxyethyl)iminodiacetate (HEIDA). Their performance was comparable.

Soil pre-equilibrated with methyl parathion was slurried with water (1:1 mass ratio) containing the Fenton reagents. The loss of methyl parathion, which leveled off after about 10 hours, reached as high as 88%, depending on reagent concentrations and temperature. The optimum concentration of the Fe-chelate was 0.01 moles per kilogram (mol/kg). Using this concentration of chelating agent, at 21°C, up to 80% loss of methyl parathion could be achieved with 6 mol/kg H₂O₂ added in two batches, while at the slightly elevated temperature of 35°C 88% loss could be achieved with 1 mol/kg H₂O₂, corresponding to a peroxide-to-methyl parathion molar ratio of about 130. This ratio is about an order of magnitude greater than that required to completely mineralize methyl parathion in water. The higher oxidant demand in soil is due to catalytic decomposition of H₂O₂ by soil components and scavenging of hydroxyl radicals by natural organic matter. A partial characterization of products indicated stoichiometric yields of nitrate and sulfate, extensive degradation of the ring (4-nitrophenol was detected in less than 5% yield), but only partial degradation of the organophosphorus group (dimethyl phosphate was determined in 22-36% yield). The presence of methyl paraoxon was not determined. The pH of the soil declined from 6 to 4.1 during the treatment.

4.2 Reaction Chemistry with Other Organophosphorus Pesticides

Dimethoate degradation by the photo-Fenton reaction was studied by Nikolaki et al. (2005). They observed extensive loss of TOC from solution and generation of sulfate (60% yield), phosphate (50%), and ammonium (40%) after ~3 h of UV irradiation. In addition, they detected transient levels of dimethyl phosphate, N-methylacetamide, and formic acid.

In the dark Fenton reaction (50 mM Fe²⁺, 0.6 mM H₂O₂, pH 7.2) Doong and Chang (1998) observed moderate losses after 24 hours of methamidophos (39%), diazinon (29%), phorate (28%) and EPN (12%). Much faster rates and greater losses of these organophosphorus pesticides were obtained when the samples were illuminated with a medium-pressure UV lamp. The order in reactivity was phorate > methamidophos > EPN > diazinon > malathion.

p-Nitrophenol, a common byproduct of the parathions, reacts rapidly with Fenton’s reagent, and has often been used to model reaction kinetics (Khan et al., 2005; Chirchi and Ghorbel, 2002; Goi and Trapido, 2002; Kiwi et al., 1994; Lipczynska-Kochany, 1992; Kavitha and Palanivelu, 2005). Degradation rapidly proceeds to ring-opened products that depend on whether the solution is illuminated (Kavitha and Palanivelu, 2005; Pignatello and Sun, 1995). In the dark, the major products are nitrite, nitrate, oxalate, acetate and CO₂. Only about 30% of the theoretical amount of CO₂ is evolved. Under UV or solar illumination, oxalate and acetate are mineralized via photolysis of their Fe(III) complexes to provide the remaining CO₂. In both dark and illuminated reactions, nitrite is oxidized and the final product is nitrate in stoichiometric yield.

4.3 Summary of Implications for Treatability Test

Successful Fenton treatment of soils on a laboratory scale has been demonstrated for chlorinated solvents, PAHs, PCBs, pesticides, explosives, fuels and fuel components (reviewed in Pignatello et al., 2006).

Commercial-scale systems based on Fenton technologies have been explored for the treatment of groundwater and soils by in situ and ex situ approaches (USEPA 1998; USDOE, 1999a). The general flow configuration for in situ treatment consists of a mixing head that combines catalyst and hydrogen peroxide solutions from separate reservoirs in the injection well. Positive displacement of reaction solutions into the aquifer are maintained by externally applied compressed air or the back pressure from CO₂ and O₂ generated from the oxidation reactions. In-place soil mixing has been used to increase contact between oxidants and contaminants, but this approach is generally limited to applications in shallow aquifer systems. Field trials conducted at sites with chlorinated solvent DNAPL contamination have shown some success at reducing groundwater contaminant concentrations (USDOE, 1999a). The observed ‘rebound’ of contaminant concentrations is likely attributable to poor contact between oxidant and contaminant zones, as well as desorption from the aquifer matrix.

Fenton technologies applied to the cleanup of natural solids face significant obstacles: interference by soil components, the pH limitation typical of Fenton reactions, difficulties in effective dispersal of reagents, and potential alteration of the soil environment.

The amount of hydrogen peroxide needed to transform, and especially mineralize, a given concentration of contaminant in soil is often far greater (factor of 10-100) than in aqueous systems due to i) the presence of natural organic matter; ii) nonproductive catalyzed decomposition of H₂O₂ to O₂ and H₂O, or iii) the presence of inorganic reductants in soil that consume H₂O₂ (Pignatello et al., 2006).

Lowering the pH to below 4 keeps Fe(III) soluble and reduces nonproductive decomposition of H₂O₂ (Baciocchi et al., 2003). Acidification of soil is difficult due to the high buffering capacity of soil, and is potentially polluting itself. The pH of the substrata at the Site (2.4 - 4.3) may be favorable for the Fenton reaction and obviate the need for acidification. If not, the addition of ferric ion chelating agents may be considered.

Oxidant addition can initially reduce aquifer microbial populations to nil (Miller et al., 1996), but populations can rebound (Ferguson et al., 2004; Chapelle et al., 2005). The formation of large amounts of iron oxyhydroxide precipitate (an end product of the Fenton reaction) may induce shifts in microbial communities to this solid as a terminal electron acceptor from others (e.g., SO₄²^-) (Chapelle et al. 2005).

Since Fe occurs naturally, one could hope to avoid having to add it. However, addition of peroxide alone is usually insufficient to achieve degradation before it decomposes nonproductively (Pignatello and Baehr, 1994; Miller et al., 1996). While iron minerals (goethite, magnetite, hematite) can catalyze the Fenton reaction, they are much less reactive than soluble iron, especially when the pH is not lowered.

Application of Fenton’s reagent in the field for in situ decontamination of aquifers may also suffer from problems related to effective dispersal of reagents. Iron(III) sorbs strongly to mineral surfaces and organic matter depending on pH and may not travel very far from the point of injection. Aquifer plugging from precipitated iron oxyhydroxides has been an issue in some field trials. Another problem is the rapid decomposition of peroxide near the point of injection which can result in gas eruption on the surface. Addition of phosphate stabilizers is thought to help in this regard (Kakarla and Watts, 1997), but after a while phosphate may be depleted by adsorption.

Fenton’s chemistry is the only oxidant commonly used for ISCO that has been well demonstrated to oxidize the types of organophosphorus pesticides found at the Høfde 42 site. It is recommended that treatability tests be conducted with Site groundwater to evaluate the treatment performance of conventional and/or modified Fenton’s reagent. The reactivity and degradation products from the organophosphorus pesticides should be quantified to evaluate the applicability of Fenton’s reagent for field pilot testing at the Site.

5 Permanganate Reactions with Organophosphorus Pesticides

5.1 Reaction Chemistry with Organophosphorus Pesticides and other Contaminants at the Cheminova Site

Permanganate oxidizes some organic compounds and is itself reduced to MnO₂ which is environmentally benign. The products of the reaction include hydrogen ions or hydroxide ions, depending on the target compound. Permanganate is much more specific in its reactivity with organic compounds than hydroxyl radical-generating reagents. In organic synthesis permanganate is widely used to hydroxylate or cleave alkenes, convert alcohols to ketones or carboxylic acids, oxidize aldehydes to carboxylic acids, and oxidize amines to nitro compounds. A few other less common oxidations are also known. In situ treatment applications of permanganate for soil or groundwater contaminants have largely been restricted to chlorinated alkenes, which are effectively mineralized by the reagent. Nevertheless, oxidation of compounds not typically thought of as susceptible to permanganate has been observed. Brown et al. (2003) spiked soil (1.3% organic matter) with a mixture of six PAHs and observed 8 to 72% degradation depending on the PAH. Gates-Anderson et al. (2001) obtained 99% loss each of naphthalene, phenanthrene and pyrene spiked to low organic carbon soils. MTBE is oxidized slowly (Damm et al., 2002). Even RDX (hexahydro-1,3,5-trinitro-1,3,5-triazine), an explosive in which the carbon atoms are on average in a highly oxidized state, was 87% mineralized by permanganate in slurries of aquifer solids (Adam et al., 2004).

Very few published studies exist of permanganate treatment of the Site contaminants. Pugliese et al (2004) investigated the ability of permanganate solutions to remove pesticide residues on nectarines. While removal of malathion and methyl parathion from the fruit by the reagent (25 milligram per liter [mg/L] KMnO₄) was no more effective than plain water, they showed in separate experiments that 0.1 molar (M) KMnO₄ converted malathion to malaoxon and methyl parathion to methyl paraoxon in 5 min, although the yields were not reported. The oxons are regarded as toxic. Nitrophenols are reported to be oxidized by permanganate (Radhakrishnamurti and Sahu, 1976). No data is available on the other Site contaminants.

5.2 Summary of Implications for Treatability Test

Well-to-well recirculation techniques have been successful in remediation of chlorinated alkene DNAPL contamination by permanganate (Lowe et al, 2002). However, in many field applications, increased injection pressures and reduced circulation rates occur as a result of pore clogging by particulate MnO₂ and the effects of CO₂ evolution (Siegrist et al., 2002; Li and Schwartz, 2000; Conrad et al., 2002; MacKinnon and Thomson, 2002). For example, in flow tank experiments, Li and Schwartz (2000) found that the DNAPL (TCE) mass removal rate decreased dramatically as treatment proceeded due to increasing flow divergence around zones of higher DNAPL saturation as MnO₂ precipitates and CO₂ gas reduced permeability in these zones. This can cause a large amount of permanganate to leave the contaminated zone during flushing (Lee et al., 2003), and thus the potential for pollution of the aquifer by permanganate. In other cases the effects of pore plugging were not significant (Struse et al., 2002; Crimi and Siegrist, 2003). For example, in aquifer solids taken from a DNAPL-contaminated Launch Complex at Cape Canaveral Air Station in Florida, Crimi and Siegrist (2003) showed there was potential for long-term immobilization of a portion of introduced manganese and no induced loss in subsurface permeability due to deposition of manganese oxides particles. Permanganate treatment did, however, cause elevated manganese, chromium, and nickel concentrations in site ground water within the treated region.

One advantage of permanganate is that it is more persistent than peroxide or ozone in the subsurface. Permanganate is stable in aqueous solution. In soil, it can be expected to be consumed by reaction with organic matter and metal ions in low oxidation states; this would have to be determined on soil specific to the site. In soil column and batch studies, Mumford et al. (2005) showed that permanganate reacts with aquifer materials by fast (>7 g of MnO₄ per kg per day (g/kg/day)) and slow (~0.005 g/kg/day) rates, but it was unclear whether the slower stage was due to intrinsically slower reactions or diffusion rate-limitations. Only a fraction (10-40%) of the organic carbon was mineralized over a 14-week period. Unlike peroxide or ozone, permanganate is not expected to be catalytically decomposed by soil components. Another advantage of permanganate oxidation over the Fenton reaction is that it generally does not require pH adjustment. Permanganate compared favorably with Fenton’s reagent for treatment of a mixture of volatile organic compounds (TCE, PCE and 1,1,1-TCA) or PAHs (naphthalene, phenanthrene, pyrene) in soil slurries (Gates-Anderson et al., 2001); the Fenton reaction required acidification to pH 3.

Various phase-transfer catalysts may be employed to facilitate reactions with permanganate. Co-solvents such as acetic acid, acetone, or tert-butyl alcohol (TBA) can increase the solubility of certain organic compounds in water where the oxidation takes place (House, 1972; Zhai et al., 2006). Kang et al. (2004) describe the use of paraffin wax-encapsulated KMnO₄ to treat chlorinated DNAPLs; the particles preferentially accumulated at the DNAPL interface where permanganate was released as the wax dissolved in the DNAPL. The surfactant sodium dodecylsulfate increased the rate of permanganate oxidation of TCE by several fold, even at concentrations below its critical micelle concentration (Li, 2004). Cationic phase-transfer catalysts—tetraethylammonium bromide, tetrabutylammonium bromide, and pentyltriphenylphosphonium bromide—also enhanced the rate of TCE destruction in two-phase mixtures of the DNAPL and water, although the effect was more modest (Seol and Schwartz, 2000). It is essential to recognize, however, that co-solvents may interfere with mineralization (Zhai et al., 2006), or may substantially increase the oxidant demand if the co-solvent can be oxidized by the oxidant. In addition, co-solvents typically need to be added at a minimum concentration of 10% of the solution in order to increase the solubility of target contaminants (Schwarzenbach et al., 1993).

Permanganate potentially can attack the thiophosphate group to form the corresponding oxon; the amino group of amino parathions to form the corresponding nitro compound; or the aromatic ring of the parathions and their byproducts, leading to hydroxylation of the ring and possibly ring cleavage. The products resulting from initial attack at these positions may react further with permanganate to give innocuous products. The limited treatability results available (Pugliese et al., 2004) suggested that organophosphorus pesticides are oxidized by permanganate, however the final products from this oxidation reaction are not known. Further laboratory work is necessary to further evaluate this oxidant’s capability with organophosphorus pesticides. Thus, a laboratory treatability test of permanganate is recommended for the Høfde 42 project.

6 Persulfate Reactions with Organophosphorus Pesticides

6.1 Reaction Chemistry with Organophosphorus Pesticides and other Contaminants at the Cheminova Site

Persulfate has only recently been widely evaluated for in situ treatment of groundwater contaminants. As a result, no data is available on Site contaminants or related compounds. A discussion of the reactivity of persulfate towards other organic contaminants follows. The thermally-induced reaction is slow at room temperature but still perceptible for many compounds (Huang et al., 2002; Huang et al., 2005). For example, Huang et al. (2002) observed half-lives of MTBE ranging from 14.8 hours (h) at 20°C to 0.25 h at 50°C at 31 mM initial persulfate concentration, affording an Arrhenius activation energy of 103 kilojoules per mol (kJ/mol). MTBE gave products typical of hydroxyl radical advanced oxidation processes (AOPs). However, MTBE reacted dramatically slower (roughly, factor of seven) in groundwater than in phosphate buffer solution. This was attributed to radical scavenging by carbonate ion.

Huang et al. (2005) monitored the rates of degradation of 59 volatile organic compounds (VOCs) present as a mixture in persulfate solution at 1 or 5 g/L. At the higher concentration, degradation was complete or nearly so for 37 out of 59 compounds. Activation energies (measured between 20 and 40°C) for select compounds ranged from 41 kJ/mole (vinyl chloride) to 92.9 kJ/mole (o-xylene). The most reactive compounds were the alkenes and the substituted benzenes. The least reactive were the halogenated alkanes.

Liang et al. (2003) observed rapid oxidation of 60 mg/L TCE or 1,1,1- TCA above 40°C in water using a 10:1 persulfate:contaminant molar ratio. Under comparable conditions in soil slurries (1:5 soil-water mass ratio), the half lives were several times greater. Organic matter appeared to act in a dual role as competing substance (i.e., radical scavenger) and reaction promoter by providing a source of ferrous ion. Further study showed that supplemental Fe(II) accelerates degradation of TCE (Liang et al., 2004a) but that Fe(II) participates as a stoichiometric reagent (i.e., is not regenerated from the (III) state) and leads to unproductive decomposition of persulfate presumably by scavenging SO₄^{- •} as shown in Eq. 9.

Addition of sodium thiosulfate following Fe(II) addition gave opposing effects of regenerating Fe(II) from Fe(III) on the one hand and scavenging SO₄^{- •} and OH^• on the other. Addition of citrate ion improved the ability of Fe(II) to accelerate degradation (Liang et al., 2004b). Optimum results were obtained at a persulfate-citrate-Fe²⁺-TCE molar ratio of 20:2:10:1. Citrate may influence the reaction through its ability to chelate Fe(II) or Fe(III), but exactly how it participates is unclear.

6.2 Summary of Implications for Treatability Test

Persulfate readily attacks aromatic compounds and, therefore, is likely to react with parathions and p-nitrophenol. Byproducts are expected to be similar to those obtained in hydroxyl radical AOPs. Few data exist, however, to allow prediction of the reactivity of persulfate towards Sulfotepp, malathion, triethyl phosphate and diethyl phosphate. The rate constant between hydroxyl and dimethyl phosphate (k_OH = 1.2 × 10⁸ M^-1s^-1; Buxton et al., 1988) is relatively low. In any case, reactions at in situ temperature are expected to be slow. The use of radio frequency or electrical resistance heating may be considered for activation of this oxidant. However, subsurface heating to achieve ISCO using persulfate would make a field program very costly. Given the lack of available laboratory evidence for treatment of organophosphorus pesticides with this oxidant, and the likely high costs associated with a heat activated persulfate field program, laboratory testing of persulfate is not recommended for the Høfde 42 project at this time.

7 Ozone and Ozone+Hydrogen Peroxide Reactions with Organophosphorus Pesticides

7.1 Reaction Chemistry with Cheminova Site Organophosphorus Pesticides and other Contaminants

The only data available on ozonation among the Site contaminants pertains to parathion, methyl parathion, and 4-nitrophenol. Information regarding treatment conditions, rates, and end products for ozone studies identified in this literature review are summarized in Table 6.

Ozonation of parathion dissolved in 95% ethanol resulted in substantial conversion to paraoxon accompanied by formation of sulfate (Gunther et al., 1970). Paraoxon was not destroyed under the same conditions. The reaction in ethanol may be relevant to Høfde 42 because the solvation properties of ethanol may mimic those of the DNAPL existing at the Site. In water, Ku et al. (1998) obtained stoichiometric yields of nitrate and sulfate after reacting parathion with ozone for 1 h. Paraoxon was not monitored. Phosphate and carbonate (from CO₂) evolved more slowly indicating the production of organophosphorus esters and ring breakdown products as transient intermediates. The rate of parathion loss was dependent on ozone concentration and independent of pH in the range 3-9. Laplanche et al. (1984) observed ozone-mediated decomposition of parathion to phosphate and p-nitrophenol. Meijers et al. (1995) report that, at the ozone dosage required to just reach the disinfection level of drinking water, methyl parathion was only partially degraded. Spencer et al. (1980) demonstrated ozone transformation of parathion to paraoxon on soil dusts or clay particles at 30% relative humidity. The yields of paraoxon varied with ozone concentration (30 ppbv or 300 ppbv) and were lower in the absence of UV light. The reactivity of paraoxon separately was not examined.

p-Nitrophenol, a common byproduct of the parathions, reacts quite rapidly with ozone in the dark and (especially) with UV light, and is often used to model ozone reaction kinetics (Beltrán et al., 1992; Shi et al., 2005; Beltrán et al., 2005; Gimeno et al., 2005; Goi et al., 2004; Yu and Yu, 2000; Yu and Yu, 2001; Ku et al., 1998; Barberis and Howarth, 1991). Among the products detected (Shi et al., 2005; Goi et al., 2004; Yu and Yu, 2001) are: catechol, 4-nitrocatechol, o-benzoquinone, p-benzoquinone, hydroquinone, phenol, fumaric acid, maleic acid, oxalic acid, formic acid, nitrate and oxidative coupling products.

Meijers et al. (1995) studied the ozonation of several organophosphorus pesticides in drinking water at ozone dosages required to just reach the disinfection level. The extent of degradation ranged from 96% to 28% and followed the order: dimethoate > diazinon ~ methyl parathion > chlorfenvinphos >tetrachlorvinphos. The extent of degradation increased with pH (7.2-8.3), temperature (5-20°C) and O₃/DOC ratio (0.53-0.95). The products were not identified. Meijers et al. (1995) observed that the addition of hydrogen peroxide, prior to ozonation, increased the formation of hydroxyl radicals, and improved the extent of pesticide treatment. Out of 23 pesticides tested, 21 including methyl parathion were effectively degraded by AOP at an ozone dosage of 3.0 mg/L (O₃/DOC = 1.4 g/g) and H₂O₂/O₃ ratio of 0.5 to 2.5 g/g. The authors concluded that relative to treatment by ozone alone, persistent pesticides can be degraded more effectively by dosing with hydrogen peroxide followed by ozonation.

7.2 Summary of Implications for Treatability Test

Ethyl and methyl paraoxon may be produced in high yield during ozonation reactions. Since these compounds are highly toxic, their concentrations must be monitored carefully in treatability studies, and conditions recommended for field application must be chosen to minimize or eliminate their formation.

Ozone can be injected as a gas or as a solution in water (solubility, 1 g/L at 0°C). For contaminants in the saturated zone, ozone can be injected either as a gas (sparging) or in aqueous solution. For contaminants in the unsaturated zone the preferred method is gas injection. Simulations indicate that ozone can be effective in the unsaturated zone provided that efficient circulation of ozone is achieved (Shin et al., 2004). These simulations were based on a contaminant (TCE) concentration of only 0.0015 mol/kg, far below what is necessary to form a residual NAPL phase. The rate of ozone delivery in the unsaturated zone is dictated by the relative gas-to-liquid film transfer rate (including chemical reaction) divided by the gas convection rate (Sung and Huang, 2002).

Little information on ozonation in the field is available in the peer-reviewed literature. Cases in which ozonation has been used for subsurface treatment (often, coupled with soil vapor extraction or air stripping) are summarized briefly in USEPA, 1998. Among these cases is one involving semi-volatile contaminants where — like the Site contaminants ¾ gas stripping is not likely to contribute significantly to contaminant removal compared to oxidation. At this former wood treating site in Sonoma, California (USEPA, 1998), contamination by pentachlorophenol and creosote extended from shallow soils down to the water table. After one month of continuous ozonation, pentachlorophenol and PAH concentrations in the solids were reduced by 38 to 99.5%. However, it was not specified where the samples were collected with respect to the water table.

Most reports in the primary literature are based on batch or soil column experiments, sometimes combined with mathematical modeling. When ozone is pumped through an artificially contaminated soil column the degree of degradation increases with increasing ozonation time, decreasing soil particle size, decreasing contaminant concentration, decreasing moisture content, and decreasing scavenger concentration (e.g., soil organic matter or bicarbonate ion) (Zhang et al, 2005; Masten and Davies, 1997).

Ozone must be delivered quickly to the contaminant plume because of its inherent instability. Ozone self-decomposes in both air and water, with a half-life of a few days and a few tens of minutes, respectively, at 20°C. Surfaces catalyze ozone decomposition. For example, the decomposition of ozone in a dry uncontaminated silica sand column of low organic matter content (0.03%) was first order and ozone had a half-life of 1.1 h (Yu et al., 2005). The rate of decomposition decreased with increasing moisture content. Shin et al. (2004) partitioned the rate of ozone decomposition in reactors with soil but without contaminant into self-decomposition, surface-catalyzed decomposition, and consumption by soil organic matter. The latter two processes are expected to dominate under field conditions. Organic matter, in contrast to surfaces, becomes less and less effective with increasing exposure to ozone. It was calculated that between 25 and 40% of soil organic matter is available for consumption by ozone. With increasing water content the rates of both surface-catalyzed decomposition and consumption by organic matter are reduced (Choi et al., 2002; Jung et al., 2004).

The question of whether or not hydrogen peroxide should be combined with ozone is relevant. The production of hydroxyl radical from ozone is not dependent on addition of hydrogen peroxide, since hydroperoxyl radical (HO₂^•) and its conjugate anion, superoxide ion (O₂^-•), may be generated by reaction of organoradical intermediates with oxygen. Compounds that convert OH^• to HO₂^•/O₂^-• ¾ not all do ¾ act as promoters of the chain (see Figure 2). Hydrogen peroxide can be a chain carrier though its conjugate anion (HO₂^- ; pK_a (H₂O₂) = 11.6). However, under the acidic conditions of the Site, the concentration of HO₂^- will be exceedingly low. It therefore seems unlikely that hydrogen peroxide would improve remediation at the Site. Nevertheless, hydrogen peroxide is known to enhance reactivity of ozone under certain conditions, and the combination of ozone and hydrogen peroxide has been shown to achieve effective treatment of certain CoCs, including methyl parathion and other persistent pesticides (e.g., see USEPA, 2004; and Meijers et al. 1995).

Further laboratory work is necessary to evaluate the capability of ozone and ozone + peroxide to treat organophosphorus pesticides. Thus, a laboratory treatability test of ozone and/or ozone + peroxide is recommended for the Høfde 42 project.

8 Case Studies of ISCO Field Applications

Two case studies have been selected to illustrate the efficacy of Fenton’s reagent and ozone with peroxide for in situ treatment. The first case study is an outline of the pilot-scale application of Fenton’s reagent at the Savannah River Site in Aiken, South Carolina (USDOE, 1999a). The second case study is a summary of an ozone and peroxide pilot and full scale tests conducted at an active retail gas station in Riverside, California (Brackin et al., 2005), which would be similar in approach to an ozone application for source zone remediation in groundwater. Table 7 summarizes the site conditions and remedial activities at both of these sites. Sections 8.1 to 8.2 below include summaries of the treatment operation and lessons learned from each application.

The following references available on the World Wide Web provide descriptions of additional ISCO case studies that demonstrate the efficacy of this technology for DNAPL remediation:

Environmental Security Technology Certification Program. 1999. In Situ Chemical Oxidation, Technology Status. http://www.estcp.org/documents/techdocs/ISO_Report.pdf

Interstate Technology and Regulatory Council. 2005. Technical and Regulatory Guidance for In Situ Chemical Oxidation of Contaminated Soil and Groundwater. http://www.itrcweb.org/Documents/ISCO-2.pdf

U.S. Department of Energy. 1999a. Fenton’s Reagent. DOE/EM-0484. Office of Environmental Management, Office of Science and Technology. http://apps.em.doe.gov/OST/pubs/itsrs/itsr2161.pdf

U.S. Department of Energy. 1999b. In Situ Chemical Oxidation Using Potassium Permanganate. DOE/EM-0496. Office of Environmental Management, Office of Science and Technology. http://apps.em.doe.gov/OST/pubs/itsrs/itsr167.pdf

U.S. Environmental Protection Agency. 1998. Field Applications of In Situ Remediation Technologies: Chemical Oxidation. Office of Solid Waste and Emergency Response. EPA 542-R-98-008. http://www.clu-in.org/download/remed/chemox.pdf

Yin, Y, and H.E. Allen. 1999. In Situ Chemical Treatment. Ground-Water Remediation Technologies Analysis Center. Technology Evaluation Report No. TE-99-01. http://www.groundwatercentral.info/org/pdf/E_inchem.pdf.

8.1 Case Study #1: Fenton’s Reagent Demonstration, Savannah River Site Aiken, South Carolina

A demonstration of ISCO using Fenton’s reagent was conducted at the Savannah River Site in Aiken, South Carolina. The site geology included permeable sands with low fines alternating with clayey sand and clay units. The area targeted for the demonstration was approximately 50 feet (ft) by 50 ft. The area was adjacent to a known source of DNAPL with an estimated TCE mass of 600 pounds (lbs) within the treated area. The demonstration was conducted over a 6 day period.

Fenton’s reagent, comprised of a catalyst solution of 100 parts per million (ppm) ferrous sulfate, pH-adjusted with concentrated sulfuric acid, was introduced to the subsurface using 4 injection points which used a patented mixing and injection process. Injections were conducted in batch mode with one batch injected per day. Following 6 days of injection, the site was characterized to determine treatment efficiency.

A destruction efficiency of 94% was achieved, based upon results of soil sampling. In addition, groundwater concentrations were substantially decreased from 119.49 mg/L PCE and 21.31 mg/L TCE before treatment to 0.65 mg/L PCE and 0.07 mg/L TCE. Corresponding increases in chloride concentration confirmed that the DNAPL removal was the result of oxidation. Some metals mobilization was observed, however the concentrations remained below levels of concern.

An evaluation of the costs to implement the Fenton’s reagent technology was completed based upon the results of the demonstration, and compared to the cost per pound of DNAPL treated by pump-and-treat. The break even point for costs for Fenton’s reagent treatment versus pump-and-treat was sensitive to the depth of contamination and total mass. For the Savannah River Site, the break even point ranged from 6500 to 9500 pounds of DNAPL as depth of contamination increased from 60 to 155 ft. For sites with less than 4000 lbs DNAPL the unit cost is >$100/lb DNAPL and for sites with approximately 1000 lbs of DNAPL unit costs increase to greater than $700/lb of DNAPL.

Lessons learned during the technology application included the following (text in italics has been excerpted from USDOE [1999a]):

Design Issues:

"The efficiency of the process increases at higher contaminant concentrations and decreases as target treatment levels become more stringent.

Higher H₂O₂ concentrations provide faster reaction times, significantly greater removal of DNAPL type contaminants, but less efficient H₂O₂ use.

Highly alkaline soils may require mineral acid addition to bring the pH into the optimal range.

Organic carbon content may impact treatment because the hydroxyl radical is relatively nonselective. However, no significant effect was observed with contaminant levels of 500-2000 ppm with total organic carbon of 0.1 to 1.3 (Watts et al. 1994).

For in situ groundwater treatment, the number and pattern of injectors and monitoring wells must be designed to ensure maximum coverage of the treatment zone. Because the cost is related to depth (cost per well was approximately $70/ft) and amount of DNAPL, the number and spacing of the wells becomes critical. The heterogeneity of the subsurface at the site will also control the number and spacing of wells required.

Duration of operation is not a linear function of volume of DNAPL. Factors affecting the duration of the treatment include: permeability, heterogeneity, and geochemistry of the aquifer."

Implementation Considerations:

"When implementing ISCO using Fenton’s Reagent, general operation considerations include:

pH of the system must be between 3 and 6, for traditional Fenton’s reagent.

The rate of the reaction increases with increasing temperature (although the efficiency declines above 40 to 50°C).

For most applications the valence of the iron salts used doesn’t matter (+2 versus +3) nor does it matter whether a chloride or sulfate salt of the iron is used, although chlorine salts may generate high rates of chloride during application.

Due to oxidation of the subsurface, metals that are mobile under these conditions may be released at some sites. This should be considered during the technology selection process.

Implementation of this technology does not require permanent infrastructure, such as a permanent power source (temporary power is required), permanent water and chemical tanks, etc. Temporary power is required for operation of the system. This is much less expensive for the short duration of operation, typically less than 1 month and in many instances 1 to 2 weeks. Also required is a constant supply of water for process, as well as emergency, purposes. For remote sites where a distribution line with potable water is not available tanks for water storage are appropriate. During the demonstration, approximately 1000 gallons of water per day were used for a 6-day period.

The end products of in situ oxidation are very appealing. No waste is generated from the treatment process, and no material is brought to the surface.

At complex sites in situ oxidation using Fenton’s Reagent should be considered in tandem with other technologies. For example, if in situ bioremediation is considered as a polishing step, the pH should be held above 4.0 during the treatment operations."

Note that the above reflect the conclusions and opinions stated in USDOE (1999a), do not entirely reflect the opinions of the authors of this literature review report.

8.2 Case Study #2: Ozone with Peroxide, Active Retail Gas Station, Riverside, California

A pilot test and subsequent full scale implementation of AOP was performed at an active retail gas station in Riverside, California. The site geology included silts and sandy silts to approximately 15 feet below ground surface (bgs), fine-and coarse-grained sands to about 35 feet bgs, and weathered bedrock between 35 and 40 feet bgs at some locations. The water table was generally observed at 33 feet bgs. The primary CoCs at the site included total petroleum hydrocarbon ([TPH] 5,000 – 35,000 micrograms per liter (µg/L)), MTBE (5,000 to 200,000 µg/L), TBA (1,000 to 20,000 µg/L), and with BTEX (500 to 5,000 µg/L).

A pilot test was conducted using two sparge wells placed in the center of a source area for the injection of ozone, oxygen, hydrogen peroxide, and air into the contaminated groundwater. This combination of reagents was selected to provide the AOP process, to distribute the oxidant in the subsurface, and to stimulate aerobic bioremediation. Subsequent to the pilot test, a full-scale remediation program was conducted that used six nested sparge points. The sparging devices were installed to a total depth of 40 feet bgs, about 8-10 feet below the top of the groundwater table. Four wells were used as monitoring points during the pilot test and five wells during the full scale program to observe ROI and water quality changes.

Groundwater samples were collected to evaluate changes in water quality. Samples were taken twice a week and analyzed for pH, temperature, conductivity, oxidation-reduction potential (ORP), and dissolved oxygen (DO), as well as the CoCs at the site. During the pilot and full scale tests ozone was injected into each sparging well at a 4% concentration at a rate of 0.25 lbs/well/day. During the pilot trial, hydrogen peroxide at a concentration of 7.75% was injected into each well at a rate of 1.2 lbs/well/day. For the full-scale operation, a 10% solution was injected at a rate of 0.5 lbs/well/day. Compressed air was used to assist in moving the oxidants away from the injection well.

The pilot test was operated for a period of 22 days. Results from the pilot test confirmed a substantial reduction in target contaminants at the monitoring points, although reduction in TBA concentrations was lower than that of the other target chemicals. The pilot test also confirmed an active ROI from the oxidant injection of 12 ft. Based upon these results, the full scale program was implemented.

The full scale program operated for a period of 3 months. The oxidant injection wells were placed around the perimeter of the gas station and downgradient of the source area. After three months of operation, concentrations in all monitoring and extraction wells onsite were below laboratory detection limits for TPH, BTEX, and MTBE. Also, the concentration of TBA was reduced by 90 to 99.95 % in the monitoring wells, suggesting additional treatment required to fully remove this chemical. Other important monitoring results included a slight pH increase (0.25 units) over the operating period, a slight increase in temperature of two degrees, an increase in DO, and an increase in ORP.

For this site, a high amount of chemical oxygen demand (COD) was present, which was compensated for by using a relatively high oxidant dose rate to achieve target chemical destruction. The ROI achieved during full scale operation was at least 20 feet.

9 Summary and Recommendations

Fenton’s reagent, permanganate, persulfate, ozone and ozone with peroxide are oxidants that are commonly used for treatment of organic contaminants. Certain ISCO technologies have been demonstrated to be a successful approach for remediation of NAPL source areas at numerous sites including the two case studies reported in this document. It should be recognized, however, that ISCO only treats dissolved-phase contamination. Therefore, the rate of treatment of NAPL source areas by ISCO is limited by mass transfer (dissolution) of NAPL constituents from the nonaqueous to the aqueous phase.

To date, research reported in the literature has reported oxidation of organophosphorus pesticides with Fenton’s reagent, permanganate, and ozone. Specific conclusions from this literature review for each of the oxidants reviewed includes:

• Fenton’s reagent is known to oxidize some organophosphorus pesticides (e.g., methyl parathion, malathion), and should be tested with Site groundwater to confirm treatability under Site conditions;
• Ozone has been demonstrated to oxidize some organophosphorus pesticides (e.g., parathion), and should be tested with Site groundwater to confirm treatability under Site conditions. The low pH conditions at the Site are favorable toward maximizing the reactive half-life of ozone in groundwater;
• The effectiveness of AOP (ozone + hydrogen peroxide) for treating methyl parathion has been demonstrated, and the combination of ozone + hydrogen peroxide is reported to yield more hydroxyl radicals than ozone alone. As such, ozone + hydrogen peroxide should be tested to confirm treatability of other Site organophosphorus pesticides under Site conditions;
• The literature is inconclusive regarding the treatability of organophosphorus pesticides by permanganate. Nevertheless, permanganate remains an attractive option for ISCO, given its success with other applications in Denmark and its relative ease of deployment; and
• No literature could be identified regarding treatment of organophosphorus pesticides by persulfate. In any case, the requirement for heat activation makes persulfate a relatively less attractive option for ISCO at the Site.

Based upon these recommendations, the bench treatability tests will be performed using Fenton’s reagent and ozone (or ozone + peroxide) treatments. Testing with permanganate should also be performed. The purpose of these tests will be to confirm the efficiency and extent of treatment of the principal contaminants at the Høfde 42 site by these chemical oxidants. The results from these treatability tests will be used to evaluate which oxidants would be appropriate for pilot scale testing at the Site.

10 References

References listed below include all the references that are cited in preceding sections, as well as references for additional publications that are relevant to ISCO technologies and treatment of organophosphorus pesticides by various chemical oxidants. All the references cited below were reviewed as part of the preparation of this report.

Adam, M. L., Comfort, S. D., Morley, M. C., and Snow, D. D. (2004). Remediating RDX-contaminated ground water with permanganate: Laboratory investigations for the Pantex perched aquifer. Journal of Environmental Quality 33, 2165-2173.

Baciocchi, R., Boni, M. R., and D'Aprile, L. (2003). Hydrogen peroxide lifetime as an indicator of the efficiency of 3-chlorophenol Fenton's and Fenton-like oxidation in soils. Journal of Hazardous Materials B96, 305-329.

Barberis, K., and Howarth, C. R. (1991). Reactivity studies in the ozonolysis of pollutants using a perforated spinning disk reactor to enhance mass-transfer. Ozone-Science & Engineering 13, 501-519.

Beltran, F. J., Gomezserrano, V., and Duran, A. (1992). Degradation kinetics of para-nitrophenol ozonation in Water. Water Research 26, 9-17.

Beltran, F. J., Rivas, F. J., and Gimeno, O. (2005). Comparison between photocatalytic ozonation and other oxidation processes for the removal of phenols from water. Journal of Chemical Technology and Biotechnology 80, 973-984.

Benitez, F. J., Beltranheredia, J., and Gonzalez, T. (1991). Absorption kinetics of ozone in aqueous-solutions of malathion. Ozone-Science & Engineering 13, 487-499.

Benitez, F. J., Beltranheredia, J., and Gonzalez, T. (1992). Ozonation of pesticides - variables effect on degradation - stoichiometric ratio. Anales De Quimica 88, 548-555.

Brackin, J., Hellmann, G., Gustafson, D., 2005. Aggressive remediation of adsorbed and dissolved gasoline, MTBE, and TBA contaminants using an innovative in-situ advanced chemical oxidation system. Proceedings of the 2005 NGWA Conference on MTBE and Perchlorate: Assessment, Remediation, and Public Policy.

Brown, G. S., Barton, L. L., and Thomson, B. M. (2003). Permanganate oxidation of sorbed polycyclic aromatic hydrocarbons. Waste Management 23, 737-740.

Budavari, S. et. al., 1996, The Merck Index, 12th Ed., Merck Research Laboratories.

Buxton, G. V., Greenstock, C. L., Helman, W. P., and Ross, A. B. (1988). Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals (^.OH/^.O-) in aqueous solution. J. Phys. Chem. Ref. Data 17, 513-886.

Chapelle, F. H., Bradley, P. M., and Casey, C. C. (2005). Behavior of chlorinated ethene plume following source-area treatment with Fenton's reagent. Groundwater Monitoring and Remediation 25, 131-141.

Chemfinder.com, 2004. Chemical Database. Downloaded from: http://chemfinder.cambridgesoft.com/, 21 March 2006.

Chirchi, L., and Ghorbel, A. (2002). Use of various Fe-modified montmorillonite samples for 4-nitrophenol degradation by H₂O₂. Applied Clay Science 21, 271-276.

Conrad, S. H., Glass, R. J., and Peplinski, W. J. (2002). Bench-scale visualization of DNAPL remediation processes in analog heterogeneous aquifers: surfactant floods and in situ oxidation using permanganate. Journal of Contaminant Hydrology 58, 13-49.

Crimi, M. L., and Siegrist, R. L. (2003). Geochemical effects on metals following permanganate oxidation of DNAPLs. Ground Water 41, 458-469.

Damm, J. H., Hardacre, C., Kalin, R. M., and Walsh, K. P. (2002). Kinetics of the oxidation of methyl tert-butyl ether (MTBE) by potassium permanganate. Water Research 36, 3638-3646.

Dai, Q., and Reitsma, S. 2002. Two-dimensional experimental studies of permanganate flushing of pooled DNAPL. Third International Conference on Remediation of Chlorinated and Recalcitrant Compounds; Monterey, CA; USA; 20-23 May 2003.

Doong, R. A., and Chang, W. H. (1998a). Photodegradation of parathion in aqueous titanium dioxide and zero valent iron solutions in the presence of hydrogen peroxide. Journal of Photochemistry and Photobiology a-Chemistry 116, 221-228.

Doong, R.A., and Chang, W.H. (1998b). Photoassisted iron compound catalytic degradation of organophosphorous pesticides with hydrogen peroxide. Chemosphere 37, 2563-2572.

Environmental Security Technology Certification Program. (1999). In Situ Chemical Oxidation, Technology Status. http://www.estcp.org/documents/techdocs/ISO_Report.pdf

Ferguson, S. H., Woinarski, A. Z., Snape, I., Morris, C. E., and Revill, A. T. (2004). A field trial of in situ chemical oxidation to remediate long-term diesel contaminated Antarctic soil. Cold Regions Science and Technology 40, 47-60.

Fjordbøge, A. (2005). Evaluering af nul-valent jern til oprensning af Høfde 42 Depotet. Institut for Miljø & Ressourcer, Danmarks Tekniske Universitet. November.

Gates-Anderson, D. D., Siegrist, R. L., and Cline, S. R. (2001). Comparison of potassium permanganate and hydrogen peroxide as chemical oxidants for organically contaminated soils. Journal of Environmental Engineering-ASCE 127, 337-347.

Gimeno, O., Carbajo, M., Beltran, F. J., and Rivas, F. J. (2005). Phenol and substituted phenols AOPs remediation. Journal of Hazardous Materials 119, 99-108.

Goi, A., and Trapido, M. (2002). Hydrogen peroxide photolysis, Fenton reagent and photo-Fenton for the degradation of nitrophenols: a comparative study. Chemosphere 46, 913-922.

Goi, A., and Trapido, M. (2004). Combined methods for remediation of nitrophenol-contaminated soil. Fresenius Environmental Bulletin 13, 1191-1196.

Goi, A., and Trapido, M. (2004). Degradation of polycyclic aromatic hydrocarbons in soil: The Fenton reagent versus ozonation. Environmental Technology 25, 155-164.

Goi, A., Trapido, M., and Tuhkanen, T. (2004). A study of toxicity, biodegradability, and some by-products of ozonised nitrophenols. Advances in Environmental Research 8, 303-311.

Gunther, F. A., Ott, D. E., and Ittig, M. (1970). Oxidation of parathion to paraoxon. II. By use of ozone. Bulletin of Environmental Contamination and Toxicology 5, 87-94.

Haapea, P., and Tuhkanen, T. (2005). Aged chlorophenol contaminated soil's integrated treatment by ozonation, soil washing and biological methods. Environmental Technology 26, 811-819.

Hafiz, A. A. (2005). Metallosurfactants of Cu(II) and Fe(III) complexes as catalysts for the destruction of paraoxon. Journal of Surfactants and Detergents 8, 359-363.

House, H. O. (1972). Modern Synthetic Reactions. 2nd/Ed. W.A. Benjamin, Inc., Menlo Park, CA.

Hsu, M. I. Y., and Masten, S. J. (1997). The kinetics of the reaction of ozone with phenanthrene in unsaturated soils. Environmental Engineering Science 14, 207-218.

Huang, K. C., Couttenye, R. A., and Hoag, G. E. (2002). Kinetics of heat-assisted persulfate oxidation of methyl tert-butyl ether (MTBE). Chemosphere 49, 413-420.

Huang, K. C., Zhao, Z. Q., Hoag, G. E., Dahmani, A., and Block, P. A. (2005). Degradation of volatile organic compounds with thermally activated persulfate oxidation. Chemosphere 61, 551-560.

Huston, P. L., and Pignatello, J. J. (1999). Degradation of selected pesticide active ingredients and commercial formulations in water by the photo-assisted Fenton reaction. Water Research 33, 1238-1246.

Interstate Technology and Regulatory Council. (2005). Technical and Regulatory Guidance for In Situ Chemical Oxidation of Contaminated Soil and Groundwater. http://www.itrcweb.org/Documents/ISCO-2.pdf.

Jung, H., Ahn, Y., Choi, H., and Kim, I. S. (2005). Effects of in-situ ozonation on indigenous microorganisms in diesel contaminated soil: Survival and regrowth. Chemosphere 61, 923-932.

Jung, H., Kim, J., and Choi, H. (2004). Reaction kinetics of ozone in variably saturated porous media. Journal of Environmental Engineering-ASCE 130, 432-441.

Kakarla, P. K. C. and Watts, R.J. (1997). Depth of Fenton-like oxidation in remediation of surface soil. Journal of Environmental Engineering-ASCE 123, 11-17, 1997.

Kang, N., Hua, I., and Rao, P. S. C. (2004). Production and characterization of encapsulated potassium permanganate for sustained release as an in situ oxidant. Industrial & Engineering Chemistry Research 43, 5187-5193.

Kavitha, V., and Palanivelu, K. (2005). Degradation of nitrophenols by Fenton and photo-Fenton processes. Journal of Photochemistry and Photobiology A: Chemistry 170, 83-95.

Khan, E., Babcock, R. W., Hsu, T. M., and Lin, H. (2005). Mineralization and biodegradability enhancement of low level p-nitrophenol in water using Fenton's reagent. Journal of Environmental Engineering-ASCE 131, 327-331.

Kiwi, J., Pulgarin, C., and Peringer, P. (1994). Effect of Fenton and photo-Fenton reactions on the degradation and biodegradability of 2-nitrophenols and 4-nitrophenols in water-treatment. Applied Catalysis B-Environmental 3, 335-350.

Ku, Y., Chang, J. L., and Shen, Y. S. (1998). Decomposition of ethylparathion in aqueous solution by ozonation. Water Air and Soil Pollution 103, 341-355.

Laplanche, A, Martin, G, Tonnard, F. (1984). Ozonation Schemes of Organo Phosphorus Pesticides Application in Drinking Water Treatment
Ozone: Science and Engineering 6, 207-219.

Lee, E. S., Seol, Y., Fang, Y. C., and Schwartz, F. W. (2003). Destruction efficiencies and dynamics of reaction fronts associated with the permanganate oxidation of trichloroethylene. Environmental Science & Technology 37, 2540-2546.

Li, X.D., and Schwartz F.W. (2000). Efficiency problems related to permanganate oxidation schemes. Second International Conference on Remediation of Chlorinated and Recalcitrant Compounds; Monterey, CA; USA; 22-25 May. pp. 41-48.

Li, Z. H. (2004). Surfactant-enhanced oxidation of trichloroethylene by permanganate - proof concept. Chemosphere 54, 419-423.

Liang, C. J., Bruell, C. J., Marley, M. C., and Sperry, K. L. (2003). Thermally activated persulfate oxidation of trichloroethylene (TCE) and 1,1,1-trichloroethane (TCA) in aqueous systems and soil slurries. Soil & Sediment Contamination 12, 207-228.

Liang, C. J., Bruell, C. J., Marley, M. C., and Sperry, K. L. (2004a). Persulfate oxidation for in situ remediation of TCE. I. Activated by ferrous ion with and without a persulfate-thiosulfate redox couple. Chemosphere 55, 1213-1223.

Liang, C. J., Bruell, C. J., Marley, M. C., and Sperry, K. L. (2004b). Persulfate oxidation for in situ remediation of TCE. II. Activated by chelated ferrous ion. Chemosphere 55, 1225-1233.

Lim, H. N., Choi, H., Hwang, T. M., and Kang, J. W. (2002). Characterization of ozone decomposition in a soil slurry: kinetics and mechanism. Water Research 36, 219-229.

Lipczynska-Kochany, E. (1992). Degradation of nitrobenzene and nitrophenols by means of advanced oxidation processes in a homogeneous phase: Photolysis in the presence of hydrogen peroxide versus the Fenton reaction. Chemosphere 24, 1369-1380.

Looney, B.B, and Falta, R.W. (2000). Vadose Zone Science and Technology Solutions. Battelle Press, Columbus OH.

Lowe, K. S., Gardner, F. G., and Siegrist, R. L. (2002). Field evaluation of in situ chemical oxidation through vertical well-to-well recirculation of NaMnO4. Ground Water Monitoring and Remediation 22, 106-115.

MacKinnon, L. K., and Thomson, N. R. (2002). Laboratory-scale in situ chemical oxidation of a perchloroethylene pool using permanganate. Journal of Contaminant Hydrology 56, 49-74.

Masten, S. J., and Davies, S. H. R. (1997). Efficacy of in-situ ozonation for the remediation of PAH contaminated soils. Journal of Contaminant Hydrology 28, 327-335.

Meijers, R. T., Oderwald-Muller, E. J., Nuhn, P., and Kruithof, J. C. (1995). Degradation of pesticides by ozonation and advanced oxidation. Ozone-Science & Engineering 17, 673-686.

Montgomery, (2000). Groundwater Chemicals: Desk Reference, 3^rd Edition. Lewis Publishers, 2000.

Miller, C. M., Valentine, R. L., Roehl, M. E., and Alvarez, P. J. J. (1996). Chemical and microbiological assessment of pendimethalin-contaminated soil after treatment with Fenton's reagent. Water Research 30, 2579-2586.

Moss, R. A., and Ragunathan, K. G. (1998). Remarkable acceleration of dimethyl phosphate hydrolysis by ceric cations. Chemical Communications, 1871-1872.

Mumford, K. G., Thomson, N. R., and Allen-King, R. M. (2005). Bench-scale investigation of permanganate natural oxidant demand kinetics. Environmental Science & Technology 39, 2835-2840.

Nam, K., and Kukor, J. J. (2000). Combined ozonation and biodegradation for remediation of mixtures of polycyclic aromatic hydrocarbons in soil. Biodegradation 11, 1-9.

Nam, K., Rodriguez, W., and Kukor, J. (2001). Enhanced degradation of polycyclic aromatic hydrocarbons by biodegradation combined with a modified Fenton reaction. Chemosphere 45, 11-20.

Nikolaki, M. D., Oreopoulou, A. G., and Philippopoulos, C. J. (2005). Photo-Fenton assisted reaction of dimethoate in aqueous solutions. Journal of Environmental Science and Health, Part B, 233 - 246.

Pignatello, J. J., and Baehr, K. (1994). Ferric complexes as catalysts for "Fenton" degradation of 2,4-D and metolachlor in soil. Journal of Environmental Quality 23, 365-370.

Pignatello, J. J., and Day, M. (1996). Mineralization of methyl parathion insecticide in soil by hydrogen peroxide activated with iron(III)-NTA or -HEIDA complexes. Haz. Waste. Haz. Mater. 13, 237-244.

Pignatello, J. J., Oliveros, E., and MacKay, A. (2006). Advanced oxidation processes for organic contaminant destruction based on the Fenton reaction and related chemistry. Critical Reviews in Environmental Science and Technology 36, 1-84.

Pignatello, J. J., and Sun, Y. (1995). Complete oxidation of metolachlor and methyl parathion in water by the photoassisted Fenton reaction. Water Research 29, 1837-1844.

Pugliese, P., Molto, J. C., Damiani, P., Marin, R., Cossignani, L., and Manes, J. (2004). Gas chromatographic evaluation of pesticide residue contents in nectarines after non-toxic washing treatments. Journal of Chromatography A. 1050, 185-191.

Radhakrishnamurti, P. S., and Sahu, S. N. (1976). Kinetics and mechanism of oxidation of nitrophenols and nitroanilines by potassium-permanganate. Journal of the Indian Chemical Society 53, 1154-1155.

Schwarzenbach, R.P., Gschwend, P.M., and Imboden D.M. (1993). Environmental Organic Chemistry. John Wiley & Sons, Inc. New York.

Seol, Y., and Schwartz, F. W. (2000). Phase-transfer catalysis applied to the oxidation of nonaqueous phase trichloroethylene by potassium permanganate. Journal of Contaminant Hydrology 44, 185-201.

Shi, H. X., Xu, X. W., Xu, X. H., Wang, D. H., and Wang, Q. D. (2005). Mechanistic study of ozonation of p-nitrophenol in aqueous solution. Journal of Environmental Sciences-China 17, 926-929.

Shin, W. T., Garanzuay, X., Yiacoumi, S., Tsouris, C., Gu, B. H., and Mahinthakumar, G. K. (2004). Kinetics of soil ozonation: an experimental and numerical investigation. Journal of Contaminant Hydrology 72, 227-243.

Siegrist, R. L., Urynowicz, M. A., Crimi, M. L., and Lowe, K. S. (2002). Genesis and effects of particles produced during in situ chemical oxidation using permanganate. Journal of Environmental Engineering-ASCE 128, 1068-1079.

Spear, R. C., Lee, Y. S., Leffingwell, J. T., and Jenkins, D. (1978). Conversion of parathion to paraoxon in foliar residues - effects of dust level and ozone concentration. Journal of Agricultural and Food Chemistry 26, 434-436.

Spencer, W. F., Adams, J. D., Shoup, T. D., and Spear, R. C. (1980). Conversion of parathion to paraoxon on soil dusts and clay-minerals as affected by ozone and UV Light. Journal of Agricultural and Food Chemistry 28, 366-371.

Staehelin, J., and Hoigne, J. (1985). Decomposition of ozone in water in the presence of organic solutes acting as promoters and inhibitors of radical chain reactions. Environmental Science & Technology 19, 1206-1213.

Struse, A. M., Siegrist, R. L., Dawson, H. E., and Urynowicz, M. A. (2002). Diffusive transport of permanganate during in situ oxidation. Journal of Environmental Engineering-ASCE 128, 327-334.

Sung, M. H., and Huang, C. P. (2002). In situ removal of 2-chlorophenol from unsaturated soils by ozonation. Environmental Science & Technology 36, 2911-2918.

U.S. Department of Energy. (1999a). Fenton’s Reagent. DOE/EM-0484. Office of Environmental Management, Office of Science and Technology. http://apps.em.doe.gov/OST/pubs/itsrs/itsr2161.pdf

U.S. Department of Energy. (1999b). In Situ Chemical Oxidation Using Potassium Permanganate. DOE/EM-0496. Office of Environmental Management, Office of Science and Technology. http://apps.em.doe.gov/OST/pubs/itsrs/itsr167.pdf

U.S. Environmental Protection Agency. (1998). Field Applications of In Situ Remediation Technologies: Chemical Oxidation. Office of Solid Waste and Emergency Response. EPA 542-R-98-008. http://www.clu-in.org/download/remed/chemox.pdf

U.S. Environmental Protection Agency. (2004). AOP treats dioxane-contaminated groundwater. Technology News and Trends, Issue 11. March.

Trapido, M., and Kallas, J. (2000). Advanced oxidation processes for the degradation and detoxification of 4-nitrophenol. Environmental Technology 21, 799-808.

Wang, Y. X., Liang, Z., Yuan, X. M., and Xu, Y. S. (2005). Preparation of cellular iron using wastes and its application in dyeing wastewater treatment. Journal of Porous Materials 12, 225-232.

Watts, R. J., Bottenberg, B. C., Hess, T. F., Jensen, M. D., & Teel, A. L. (1999). Role of reductants in the enhanced desorption and transformation of chloroaliphatic compounds by modified Fenton’s reactions. Environmental Science & Technology 33, 3432-3437.

Watts et al. (1994). On Site Treatment of Contaminated Soils Using Hydrogen Peroxide. Project Report T9234-06, Washington State Transport Center, Washington State University.

Yin, Y, and H.E. Allen. (1999). In Situ Chemical Treatment. Ground-Water Remediation Technologies Analysis Center. Technology Evaluation Report No. TE-99-01. http://www.groundwatercentral.info/org/pdf/E_inchem.pdf.

Yu, C. P., and Yu, Y. H. (2000). Identifying useful real-time control parameters in ozonation process. Water Science and Technology 42, 435-440.

Yu, C. P., and Yu, Y. H. (2001). Mechanisms of the reaction of ozone with p-nitrophenol. Ozone-Science & Engineering 23, 303-312.

Yu, D. Y., Bae, W., Kang, N. G., Banks, M. K., and Choi, C. H. (2005). Characterization of gaseous ozone decomposition in soil. Soil & Sediment Contamination 14, 231-247.

Zhai, X. H., Hua, I., Rao, P. S. C., and Lee, L. S. (2006). Cosolvent-enhanced chemical oxidation of perchloroethylene by potassium permanganate. Journal of Contaminant Hydrology 82, 61-74.

Zhang, H., Ji, L., Wu, F., and Tan, J. (2005). In situ ozonation of anthracene in unsaturated porous media. Journal of Hazardous Materials 120, 143-148.

Tables

TABLE 1 SUMMARY OF ORGANOPHOSPHORUS PESTICIDE PRESENCE AT THE HØFDE 42 SITE

(Modified from Ringkjøbing County, [2004])

TABLE 2 SUMMARY OF KEY PROPERTIES OF ORGANOPHOSPHORUS PESTICIDES

Click here to see Table 2

TABLE 3. REDOX POTENTIALS OF COMMON CHEMICAL OXIDANTS

TABLE 4. OXIDANT PROPERTIES TO BE CONSIDERED FOR IN SITU TREATMENT WITH ISCO (modified from ITRC, 2005)

TABLE 5. SUMMARY OF CHEMINOVA CONTAMINANTS DEGRADED BY THE FENTON REACTION

TABLE 6. OZONE REACTION WITH CHEMINOVA CONTAMINANTS WHERE PRODUCTS HAVE BEEN IDENTIFIED

TABLE 7. SUMMARY OF CHEMICAL OXIDATION CASE STUDIES

Figures

Figure 1. Typical in-situ chemical oxidation application

Figure 2. Reactions of ozone in water in the presence of reactive solutes (Staehelin and Hoigné, 1985). M represents a reactive solute, such as an organic compound.

Bilag 3 Laboratorieforsøg

Bilag 3.1 Fentons reagens og ozon/brintperoxid

Prepared for:

COWI
Odensevej 95
DK-5260 Odense S
Denmark

Laboratory Treatability Tests of
Chemical Oxidation Technologies for
Groundwater Remediation

Høfte 42 Harboøre Tange,
Ringkjøbing County, Denmark

Prepared by:

10015 Old Columbia Road, Suite A-200
Columbia, Maryland 21046 USA
GeoSyntec Project Number MR0487

July 1, 2006

Executive Summary

A series of laboratory treatability tests were conducted by GeoSyntec Consultants (GeoSyntec) and its subcontractors to evaluate alternative in situ chemical oxidation (ISCO) technologies for groundwater remediation at the Høfde 42 Harboøre Tange site in Ringkjøbing County, Denmark (Site). The primary contaminants at the Site are the organophosphorus pesticides parathion, methyl parathion, malathion, 2-methyl-4-chlorophenoxyacetic acid (MCPA) and ethyl-sulfoteb. These chemicals are present both as dissolved phase in groundwater and as an immiscible, dense nonaqueous phase liquid (DNAPL) in the upper sand aquifer at the Site. In addition to these pesticides, a number of related organic contaminants exist in the Site soil and groundwater, including (i) constituents related to the pesticide production (E-OOOPS, M-OOSPS, EEM-OOSPS, MME-OOSPS), and (ii) pesticide and pesticide manufacturing contaminant degradation products (4-Cl-cresol, para-nitrophenol [PNF], amino-parathion, amino-methyl parathion, MP-1, EP-1, MP-2-Syre, EP-2-Syre, paraoxon, malaoxon, methyl-paraoxon, E-OOOPO, EEM-OOSPO). Mercury is also a significant contaminant at the Site; however, the tests reported herein did not evaluate treatment of mercury.

The treatability tests evaluated two ISCO technologies: ozone and Fenton’s reagent. Using groundwater collected from the Site, each test evaluated treatment efficiency and extent of treatment over a range of oxidant dosages. Treatment efficiency and extent was measured in terms of destruction of the primary pesticides, pesticide manufacturing contaminants, and pesticide degradation products, as well as accumulation of inorganic oxidation products (e.g., sulfate, nitrogen oxides, and phosphate) and reduction in toxicity (i.e., via Microtox ® bioassay).

The ozone tests were performed by adding oxidant to groundwater in sealed batch reactors. Ozone was used as the sole oxidant in one treatment at a concentration of 190 mg/L. In the remaining treatments, ozone and peroxide were used in combination at concentrations of 500 mg/L ozone + 250 mg/L peroxide, and 1000 mg/L ozone + 500 mg/L peroxide. Peroxide was added in these other treatments to enhance the ozone reactivity, which was limited in the laboratory tests by the acidic conditions of the Site groundwater and the relatively short duration over which the study was completed. Ozone at each dose tested achieved complete (100%) destruction of the pesticides (parathion, methyl parathion, MCPA, and malathion) and the pesticide reaction products (MP-1, EP-1, MP-2-Syre, EP-2-Syre, 4-Cl-cresol, PNF). Pesticide manufacturing contaminants (E-OOOPS, M-OOSPS, EEM-OOSPS, MME-OOSPS) and the oxidation product EEM-OOSPO were also reduced by 100%. E-OOOPO was produced at the lower and middle ozone dosages, but was not observed in the high ozone dose treatment suggesting that this oxidation product may be produced transiently during a field application of the technology. Oxons (paraoxon, malaoxon, methyl-paraoxon) were produced with addition of 190 mg/L ozone, but were not detected in the highest oxidant treatment. This data suggests that these oxidation products may be produced transiently in a field application of the technology. Treatment at the highest oxidant dose achieved a 3-fold reduction in groundwater toxicity. The low pH of the Site groundwater, both before and post-treatment, may have limited the ability of ozone/peroxide to achieve greater reductions in toxicity as measured by Microtox® assay.

The Fenton’s reagent tests were also performed by adding oxidant (and catalyst) to groundwater in sealed batch reactors over a range of oxidant and catalyst concentrations. Fenton’s reagent at each dose tested achieved complete (100%) destruction of the pesticides (parathion, methyl parathion, MCPA, and malathion) and the pesticide manufacturing contaminants (E-OOOPS, M-OOSPS, EEM-OOSPS, MME-OOSPS) and the oxidation products EEM-OOSPO and E-OOOPO, and paraoxon were reduced by 100% (malaoxon and methyl-paraoxon were not detected in any sample). Pesticide reaction products (MP-1, EP-1, EP-2-Syre, 4-Cl-cresol, PNF) were reduced by 100% at all Fenton’s reagent doses. MP-1 was produced in the lowest dosage treatments, but was not observed in the higher dosage treatments suggesting that this oxidation product may be produced transiently during a field application of the technology. MP-2-Syre concentrations decreased for the low dose batch reactors but increased in all of the high dose batch reactors. The low dose data indicates that this compound can be destroyed by Fenton’s chemistry, however the high dose data suggests that MP-2-Syre also forms as a reaction product from hydrolysis of other organic compounds (e.g., methyl parathion) in the Site groundwater. Treatment with a high Fenton’s dose achieved a 77-fold reduction in groundwater toxicity, as measured by Microtox® bioassay.

The results of these bench tests indicate that Fenton’s reagent and ozone are both potentially highly effective technologies for destruction of the Site contaminants, and associated manufacturing contaminants and degradation products. The extent and rate of treatment by Fenton’s reagent and ozone was impressive in both cases. One of the few differences observed between the results of the two tests was that the toxicity reduction achieved was significantly greater in the Fenton’s treatability test. One or both of these technologies might be successful at the field scale, although selection of one or both of these technologies for field implementation should only proceed after a careful evaluation of the site-specific conditions. In addition, it is recommended that additional bench testing be performed as a part of any pre-design testing. Such bench testing could include optimization of reagent dose (to avoid wasting reagent), evaluation of soil oxidant demand, treatment of sorbed-phase pesticides, treatment (enhanced dissolution) of DNAPL, evaluation of mercury transformation consequent to treatment, and adjustment of reagents employed to avoid/minimize formation of MP-2-Syre.

Table of Contents

1 Introduction

• 1.1 Background
• 1.2 Overview and Scope of Work

2 Objective

3 Groundwater Collection

4 Ozone Study

• 4.1 Setup and Sampling Details
• 4.2 Results

5 Fenton’s Study

• 5.1 Setup and Sampling Details
• 5.2 Results

6 Key Findings and Conclusions

• 6.1 Chemical Oxidation using Ozone
• 6.2 Chemical Oxidation using Fenton’s Reagent

7 References

List of Tables

Table 1: Summary of Oxidation Treatability Study Controls and Treatments

Table 2: Summary of Oxidation Treatability Study Analytical Results from External Laboratories

Table 3: Summary of Treatability Study Results by Treatment and Constituent

Table 4: Elemental Component Mass Balance Calculations

List of Figures

Figure 1: Ozone - May 9 2006 Trial External Laboratory Results

Figure 2: Ozone and Peroxide - May 9 2006 Trial External Laboratory Results

Figure 3: Ozone - May 15 2006 Trial Laboratory Results

Figure 4: Ozone and Peroxide - May 15 2006 Trial External Laboratory Results

Figure 5: Fenton's Reagent - Low Dose External Laboratory Results

Figure 6: Fenton's Reagent - High Dose Laboratory Results

List of Appendices

Appendix A: ISOTEC Report

Appendix B: APT Report

List of Abbrevations and Acronyms

1 Introduction

1.1 Background

A series of laboratory treatability tests were conducted by GeoSyntec Consultants (GeoSyntec) and its subcontractors to evaluate alternative treatment technologies for groundwater remediation at the Høfde 42 Harboøre Tange site in Ringkjøbing County, Denmark (Site). The Site is located on the western coast of Denmark, on a beach adjacent to the North Sea. The primary contaminants of concern in the source area at the Site are the organophosphorus pesticides, parathion, methyl parathion, malathion, 2-methyl-4-chlorophenoxyacetic acid (MCPA) and ethyl-sulfoteb (Ringkjøbing County, 2004). These chemicals are present both as dissolved phase in groundwater and as an immiscible, dense nonaqueous phase liquid (DNAPL) in the upper sand aquifer at the Site. In addition to the pesticides, a number of other organic compounds exist in the Site soil and groundwater, including:

• contaminants related to the pesticide production (E-OOOPS, M-OOSPS, EEM-OOSPS, MME-OOSPS); and
• biotic and/or abiotic reaction products of the pesticides and the pesticide contaminants (4-Cl-cresol, para-nitrophenol [PNF], amino-parathion, methyl-amino parathion, MP-1, EP-1, MP-2-Syre, EP-2-Syre, paraoxon, malaoxon, methyl-paraoxon, E-OOOPO, EEM-OOSPO).

A groundwater plume that contains these pesticides, their contaminants and degradation products extends from the DNAPL source area towards the North Sea. The dissolved phase chemicals have also been transported into the lower sand aquifer at the Site. The conditions in this plume are highly acidic (typical pH ranges from 2.4 to 4.3). Mercury is also present in significant concentrations in the DNAPL at the Site; however, treatment of mercury was not evaluated in the tests reported herein.

GeoSyntec and its subcontractors reviewed state-of-the-science literature regarding the use of in situ chemical oxidation (ISCO) techniques for treatment of organophosphorous pesticides in groundwater (GeoSyntec and the Connecticut Agricultural Experiment Station, 2006). The literature review identified Fenton’s reagent, ozone, and permanganate as ISCO technologies that are potentially effective for treatment of the contaminants at the Site. Accordingly, screening-level laboratory treatability tests were performed to evaluate these technologies using Site groundwater to further evaluate the feasibility of using ISCO at the Site. The treatability test approach and analytical methods were performed in general accordance with a work plan memorandum submitted to COWI on 30 March 2006. The treatability tests evaluating ozone and Fenton’s reagent were performed at the laboratories of Applied Process Technology, Inc. (APT of Pleasant Hill, California, USA) and In Situ Oxidative Technology, Inc. (ISOTEC of West Windsor, New Jersey, USA), respectively. Treatability tests to evaluate the feasibility of using permanganate were conducted and reported separately by COWI.

1.2 Overview and Scope of Work

Screening-level bench treatability tests were performed to determine the effectiveness of various chemical oxidant options for in situ treatment for the unique mixture of contaminants at the Site. All the bench tests used aquifer groundwater samples collected from the Site. The purpose of these tests was to confirm the efficiency and extent of treatment of the principal contaminants at the Høfde 42 site by these chemical oxidants. Data from these tests can be used to aid in the selection of an effective remedial technology for field-scale application at the Site. Parameter measurements obtained in the bench tests (e.g., contaminant reactivity, oxidant dosage) can be used to support the design of the field-scale remedy. Analysis of organic chemicals present at the Site, including the pesticides, their contaminants and reaction products was provided by Cheminova A/S. In addition, anion measurements (e.g., NO₂^-, NO₃^-, PO₄^3-, and SO₄^2-) were performed to evaluate whether complete oxidation of the target chemicals had been achieved, and changes in toxicity were measured (via Microtox® analysis) in samples from select treatments.

The remainder of this report is divided into seven sections. Section 2 presents the objectives of the study. Collection of Site materials for the treatability work is presented in Section 3. The approach, methods and results of the ozone study is presented in Section 4. Section 5 presents the approach, methods and results of the Fenton’s reagent study. Section 6 presents the key findings and conclusions from the treatability studies. Report references are provided in Section 7.

2 Objective

The purpose of ozone and Fenton’s reagent treatability tests was to confirm the efficiency and extent of treatment of the principal contaminants at the Site by these chemical oxidants over multiple oxidant dosages. The findings of the literature review, together with the results of the treatability tests, were used to identify feasible and effective ISCO technologies for potential implementation at the Site. The bench tests involved evaluation using groundwater samples collected from the Site. The focus of the treatability tests was on the organophosphorus pesticides and associated degradation products and manufacturing contaminants listed in Section 1.1.

3 Groundwater Collection

Groundwater samples were collected by COWI on 24 April 2006 from a well representative of the source area conditions at the Site. Groundwater was collected directly into four-3 gallon containers, minimizing headspace to the extent possible. The groundwater samples were express-shipped to APT’s laboratory in California and ISOTEC’s laboratory in New Jersey. For the purposes of sample preservation, groundwater samples were frozen prior to shipment, packed in ice in coolers during shipment, and immediately transferred to a freezer upon receipt at the laboratories. The groundwater was completely thawed one day prior to initiating the treatability studies.

4 Ozone Study

A treatability test was performed to evaluate the reactivity of the target chemicals with ozone and the rate of ozone consumption by the Site groundwater. The treatability test involved a series of batch tests with Site groundwater at a range of ozone dosage concentrations. While the treatability studies were designed to evaluate ISCO using ozone alone, peroxide was added in some treatments to enhance the ozone reactivity (i.e., increase the concentration of hydroxyl radicals), which was limited in the laboratory tests by the acidic conditions of the Site groundwater and the relatively short duration over which the study was completed.

The setup and sampling details are summarized in Section 4.1 and the results are described in Section 4.2. APT’s report for the ozone treatability study is included in Appendix A and includes a more detailed description of the study setup and analysis details. Table 1 summarizes the treatments and experimental controls.

4.1 Setup and Sampling Details

Batch tests were performed by adding oxidant to groundwater to a sealed batch reactor. Ozone was used as the sole oxidant in one treatment at an approximate concentration of 190 milligrams per liter (mg/L; labeled SP-190). In the remaining treatments, ozone and peroxide were used in combination at approximate concentrations of 500 mg/L ozone + 300 mg/L peroxide, and 1000 mg/L ozone + 650 mg/L peroxide (labeled SP-500, and SP-1000 respectively). In the combined ozone and peroxide treatments, hydrogen peroxide and ozone were added in sequence to Site groundwater to promote oxidation of the target chemicals. The treatment period in the reactor ranged from one to five hours, with an increase in treatment period with oxidant dosage. A control was constructed to assess the pre-treatment concentrations of the analytes in the Site groundwater (labeled SP-0 or source water). After completion of the tests conducted on 9 and 10 May 2006, it was discovered that the test performance samples were not quenched to remove any excess oxidant. For this reason, a second set of tests (including a second SP-0, SP-190, and SP-1000) were conducted on 15 May 2006 with ozone quenched. Samples for anion analysis were quenched using a methanol solution, while samples for pesticide and toxicity analysis were quenched with thiosulfate.

A list of the analyses conducted and the number of analyses for each treatment is included in Table 1 and is further detailed in Appendix A. The details of sample collection are also presented in Appendix A, including the holding times, sample volumes, preservation, analytical methods, and laboratories selected for each analysis. Throughout this document analysis conducted at the treatability study subcontractor laboratories (APT and ISOTEC) are described as “internal” analyses, while analyses conducted at analytical laboratories are described as “external” analyses. The analyses conducted at each sample event included high performance liquid chromatography (HPLC) and gas chromatography (GC) analysis conducted by Cheminova to identify organic compounds present in the Site groundwater that were either: 1) pesticides; 2) chemical contaminants from the manufacture of parathion and their oxidation products; 3) abiotic and/or biotic in situ reaction products of 1) and 2) [denoted as “pesticide reaction products” in this document]; and 4) oxons. A detailed listing of the constituents included in the HPLC and GC analyses is included in Table 1. In addition to these organic analyses, samples for anions (sulfate, phosphate, nitrate and nitrite), color, oxidant, turbidity, alkalinity, oxidation-reduction potential (ORP) and pH were collected for each treatment and control. All samples were analyzed by published analytical methods similar to those of the United States Environmental Protection Agency, as appropriate.

Toxicity testing was conducted for the source water (SP-0) and from the SP-1000 treatment to evaluate changes in groundwater toxicity consequent to treatment. The toxicity of the sample contents was evaluated via a Microtox® bioassay testing system that measures the light output from freeze-dried luminescent bacteria (Photobacterium phosphoreum). The light-producing mechanism in these bacteria is tied to the metabolic processes of the cell. Specifically, when the bacteria are killed or the bacteria's light-producing mechanism is changed or damaged by toxic substances, resulting in a reduction in luminescence (light output). These changes in light output were measured to calculate an IC50 for each sample, which is defined as the initial solution concentration, and is a calculated toxicity value representing the sample concentration expressed in % estimated to cause a 50% response by the exposed test organisms.

The change in oxidant and target chemical concentrations was used to assess the oxidant demand of the groundwater. This oxidant demand in combination with the target chemical concentration data collected was used to evaluate the feasibility of this technology for potential application in the field.

4.2 Results

The laboratory analytical results are summarized in Table 2, including anions, organic compounds, toxicity and color from the external laboratories. The concentration of detected chemicals in the Site groundwater prior to treatment and after treatment at each of the three treatment conditions are presented in Figures 1 and 2 for the trial completed on 9 April 2006, and in Figures 3 and 4 for the trial completed on 15 April 2006. The results from these two trials were very similar, with small differences in concentrations attributed to the use of the ozone quench. Because the performance samples in the first trial were not quenched, discussion of the results presented here focuses primarily on the second trial. Results for each of the four categories of organic compounds, as well as the anion and color results, are presented on separate data plots in each of these figures. Results for internal analyses analyzed including oxidant concentrations, alkalinity, ORP and pH are presented in Appendix A. Analytical reports from the external laboratories are included in Appendix C.

The results presented in Figures 1 through 4 clearly indicate that virtually all organic compounds were reduced by the addition of ozone or ozone and peroxide with concentrations decreasing with higher oxidant concentrations. The relative change in concentration of the organic compounds by the oxidant addition is presented in Table 3, with the following key results:

Pesticides, (parathion, methyl parathion, MCPA, and malathion), were reduced by 100%. Ethyl sulfoteb was not detected in any samples.

Pesticide manufacturing contaminants (E-OOOPS, M-OOSPS, EEM-OOSPS, MME-OOSPS) and the oxidation product EEM-OOSPO were reduced by 100%. E-OOOPO was produced in SP-190 and SP-500 treatments, but was not observed in the SP-1000 treatment suggesting that this oxidation product may be produced transiently during a field application of the technology.

Pesticide reaction products (MP-1, EP-1, MP-2-Syre, EP-2-Syre, 4-Cl-cresol, PNF) were reduced by 100%. Methyl-amino-parathion and amino-parathion were not detected in any samples.

Oxons (paraoxon, malaoxon, methyl-paraoxon) were produced with addition of ozone alone (SP-190), but were not detected in the highest oxidant treatment (SP-1000). This data suggests that (i) oxons may be produced transiently in a field application of the technology, particularly if ozone alone is used; and (ii) the combination of ozone and hydrogen peroxide may achieve more complete treatment in the field.

Treatment achieved a 3-fold reduction in groundwater toxicity, as measured by Microtox® bioassay.

Nitrate, phosphate and sulfate, inorganic reaction products from the oxidation of the organic compounds, increased with higher dosage of oxidant. Nitrite was not detected in any samples.

Color was reduced in most oxidant treatment conditions. At the highest treatement a slight increase in color was observed.

Table 4 presents a calculation of the change in presence of nitrogen, phosphorus, and sulfur from the organic and inorganic (i.e. anions) components in the samples collected, as well as the contribution of these elements to the samples from any reagents used in the study. This mass balance calculation allows for an evaluation of whether there were substantial unmeasured reaction products in the treatment samples collected. It should be noted that the observed change in sulfate concentrations was relatively small as compared to the total initial concentration, and was likely within the experimental error of the test. It can be seen that at the highest treatment concentration (1000 mg/L ozone) that the decrease in organic nitrogen and phosphorus was quite similar as the increase in inorganic nitrogen and phosphorus, suggesting that the oxidation of the organic compounds was complete. Results for sulfur were similar to that of nitrogen and phosphorus, although the measurements were within the experimental error of the test. However at the lower treatment concentrations (i.e. 190 mg/L ozone) the total concentration of these elements in the measured organic and inorganic components is different, suggesting some incomplete oxidation products were present which were not quantified by the analyses available for the study. These mass balance calculations suggest that some transient reaction products from ozone treatment of the organic compounds that were not part of the analytical suite may have been formed at the lower ozone dose. However, the apparent complete oxidation of the organic compounds in the highest treatment suggests that these unknown reaction products will be oxidized with sufficient reagent dose and/or longer contact-time.

The toxicity of the sample contents was evaluated via a Microtox® bioassay. The toxicity data results presented in Table 2 is the calculated sample concentration expressed in % that causes a 50% response by the exposed test organisms. Because the “control” sample is more toxic compared to the “treated” sample, a greater percent concentration of “treated” sample is needed to generate 50% response than the “control” sample for a fixed exposure time. For 5 minutes of exposure time, a “control” sample concentration of 0.44% resulted in a 50% of the exposed test organisms being killed versus the treated sample concentration of 1.38% to kill 50% of the exposed test organisms. The acidity of the Site groundwater (i.e. pH ~3 both before and after treatment) likely played a strong role in the relatively high toxicity of the Site groundwater.

The oxidant demand of the groundwater collected from the Site was relatively low, and the oxidation of the Site contaminants was not limited by presence of non-target organic compounds or alkalinity.

5 Fenton’s Study

ISCO treatability tests with modified (neutral-pH) Fenton’s reagent were performed using groundwater samples collected from the Site. These studies were used to provide information about the reactivity of the organic chemicals and any potential interactions that may influence the design of an oxidant injection system. The following tests were performed:

• Batch Tests with Groundwater, Low Dose: Site groundwater was mixed with one and three dosages of a low concentration of Fenton’s reagent; and
• Batch Tests with Groundwater, High Dose: Site groundwater was mixed with one, two, and three dosages of a high concentration of Fenton’s reagent.

The low and high doses were used to evaluate any differences in oxidant performance as a result of the treatment concentration.

The following sections summarize the setup and sampling details for the groundwater batch tests (Section 5.1) and results from the tests (Section 5.2). ISOTEC’s report for the Fenton’s reagent treatability study, which is included in Appendix B, provides a more detailed description of the study setup and analysis details. Table 1 summarizes the treatments and experimental controls. The analytical methodologies employed for the treatability study are summarized in the subcontractors report in Appendix B.

5.1 Setup and Sampling Details

Batch tests were performed by adding Fenton’s reagent to the Site groundwater in sealed batch reactors. A number of treatment and controls were prepared. The quantity of reagents added to each treatment, including hydrogen peroxide and catalyst, are summarized in Table 2 of ISOTEC’s report (Appendix B). The reagent solution added to achieve Fenton’s chemistry had a pH between 5 and 6; no other chemicals were added to adjust the pH of the Site groundwater. The low dose treatment reactors received one, and three dosages (treatments D and E) of hydrogen peroxide and catalyst to achieve equivalent concentrations of 0.9% and 2.8% of hydrogen peroxide and 1.8 millimoles per liter (mM) and 3.7 mM of catalyst, respectively. The high dose treatment reactors received one, two and three dosages (treatments A, B and C) of hydrogen peroxide and catalyst to achieve equivalent concentrations of 2.7%, 5.3%, and 7.9% of hydrogen peroxide and 5.3 mM, 10.6 mM, 15.9 mM of catalyst, respectively. The multiple dosage approach was used during the test to increase treatment efficiency, minimize gas formation and the resulting pressure buildup. Distilled water was used to compensate the difference of reagent volumes applied between reactors. Controls included a baseline sample (F-BGC) collected to assess the pre-treatment concentrations of the analytes in the Site groundwater and a control reactor (F-BGCont) which received an equivalent volume of distilled water instead of reagent. A time gap of approximately 24-48 hours was maintained between dosages, with the exact timing based upon oxidant measurements. All reactors (control and treatment) were left undisturbed for a minimum of 24 hours or until all the peroxide was consumed before analytical sample collection.

A list of the analyses conducted and the number of analyses for each treatment is included in Table 1 and is further detailed in Appendix B. The details of sample collection are also presented in Appendix B, including the holding times, sample volumes, preservation, analytical methods, and laboratories selected for each analysis. The analyses conducted at each sample event were similar to that for the ozone study and included HPLC and GC analysis conducted by Cheminova to identify organic compounds (listed in Table 1) present in the Site groundwater. In addition to these organic analyses, samples for anions (sulfate, phosphate, nitrate and nitrite), color, peroxide, ferrous iron, ORP and pH were collected for each treatment and control. All samples were analyzed by published analytical methods similar to those of the United States Environmental Protection Agency, as appropriate. Analytical lists and methods were the same as described for the ozone study. The change in target chemical concentration over time was used to assess the extent and rate of chemical degradation.

For the baseline sample and the highest Fenton’s reagent dose (treatment C), toxicity testing was conducted to evaluate changes in groundwater toxicity consequent to treatment. The toxicity was calculated as the sample concentration expressed in % that causes a 50% response by the exposed test organisms. For a fixed exposure time, because the “control” sample is more toxic compared to the “treated” sample, a greater percent concentration of “treated” sample is needed to generate 50% response than the “control” sample. For 5 minutes of exposure time, a “control” sample concentration of 0.55% resulted in a 50% of the exposed test organisms being killed versus the treated sample concentration of 42.35% to kill 50% of the exposed test organisms.

5.2 Results

The results of analysis of the target compounds at the Site are summarized in Table 2, including anions, organic compounds, toxicity and color from the external laboratories. The concentration of detected chemicals in the Site groundwater prior to treatment and after treatment at each of the three treatment conditions are presented in Figures 5 and 6 for the low and high dose treatment conditions. Results for each of the four categories of organic compounds, as well as the anion and color results, are presented on separate data plots in each of these figures. Results for analyses analyzed by ISOTEC including peroxide and ferrous iron concentrations, ORP and pH are presented in Appendix B. Analytical reports from the external laboratories are included in Appendix C.

The results presented in Figures 5 and 6 clearly indicate that the presence of virtually all organic compounds were reduced by the addition of Fenton’s reagent and that the extent of treatment for certain constituents increased with increasing oxidant dose. The relative change in concentration of the organic compounds by the oxidant addition is presented in Table 3, with the following key results:

Pesticides, (parathion, methyl parathion, MCPA, and malathion), were reduced by 100% in all the treatments, including the lowest peroxide dose. Ethyl sulfoteb was not detected in any samples.

Pesticide manufacturing contaminants (E-OOOPS, M-OOSPS, EEM-OOSPS, MME-OOSPS) and the oxidation products EEM-OOSPO and E-OOOPO were reduced by 100% in all the treatments, including the lowest peroxide dose.

Pesticide reaction products (MP-1, EP-1, EP-2-Syre, 4-Cl-cresol, PNF) were reduced by 100%. Methyl-amino-parathion and amino-parathion were not detected in any samples. MP-1 was produced in the lowest dosage treatments (treatment A and D), but was not observed in the higher dosage treatments (treatments B, E and F) suggesting that this oxidation product may be produced transiently during a field application of the technology. MP-2-Syre concentrations decreased for the low dose batch reactors (treatments D and E; Figure 5) but increased in all of the high dose batch reactors (treatments A, B, and C; Figure 6). The low dose data indicates that this compound can be destroyed by Fenton’s chemistry, however the high dose data suggests that MP-2-Syre also forms as a reaction product from hydrolysis of other organic compounds (e.g., methyl parathion) in the Site groundwater. Methyl parathion is known to undergo enzymatic hydrolysis to form MP-2-Syre and PNF (Cho et al. 2002), and therefore methyl parathion may also be susceptible to abiotic hydrolysis in the presence of certain catalysts.

Paraoxon was reduced by 100% in all treatments. Malaoxon and methyl-paraoxon were not detected in any samples.

Treatment at the highest Fenton’s reagent dose achieved a 77-fold reduction in groundwater toxicity, as measured by Microtox® bioassay.

Nitrate, nitrite, phosphate and sulfate, which were present in the reagents used in the study and were also inorganic reaction products from the oxidation of the organic compounds, increased with concentration of oxidant.

Color as measured by the external laboratory increased in many of the oxidant treatment conditions. Precipitates generated during the oxidation treatment likely interfered with the measurement of color, making this data difficult to interpret. Color measurements by ISOTEC (Figure 2, Appendix A) using filtered samples showed a substantial decrease from initial conditions for most treatment conditions.

Table 3 presents a calculation of the change in presence of nitrogen, phosphorus, and sulfur from the organic and inorganic (i.e. anions) components in the samples collected, as well as the contribution of these elements to the samples from the reagents used in the study. The high concentration of nitrogen, phosphorus and sulfur in the reagents used in the study make it difficult to use the mass balance calculations to evaluate whether any unmeasured organic species were present in the samples.

The pH in the treatment samples increased substantially as a result of the addition of Fenton’s reagent (Figure 3 of ISOTEC’s report; Appendix C). The pH in the control sample was 2.91, while the post-treatment sample pH ranged between 3.6 and 8.3, increasing with concentration of oxidant.

The toxicity of the sample contents was evaluated via a Microtox® bioassay and the results are presented in Table 2. The results of toxicity testing indicated that the toxicity of the groundwater decreased substantially - 77 fold - as a result of Fenton’s reagent addition (at its highest dose). The substantial decrease in toxicity was likely the effect of both the treatment of the organic compounds and the increase in pH from highly acidic conditions to slightly basic conditions.

6 Key Findings and Conclusions

Laboratory treatability tests were conducted to evaluate the performance of various in situ remediation technologies for treatment of groundwater at the Cheminova site in Ringkjøbing County, Denmark. The following technologies and treatments were tested in batch reactors using aquifer groundwater collected from the Site:

• Chemical oxidation using Fenton’s reagent; and
• Chemical oxidation using Ozone;

The results of the treatability tests are summarized in Table 3, which presents general treatment results for individual contaminants. Primary findings of the tests are also summarized in the sections below.

6.1 Chemical Oxidation using Ozone

Treatment of Site groundwater using ozone+peroxide at a dose of 1000 mg/L ozone achieved complete destruction of all pesticides, contaminants from the pesticide manufacturing process, pesticide reaction products, and oxons. E-OOOPO and malaoxon were generated in lower concentration treatments, suggesting that these compounds may be observed transiently during a field application of the technology. Ozone treatment resulted in a 3-fold reduction in toxicity of Site groundwater. This toxicity reduction observed in the ozone test was more than 20 times lower than that observed for the Fenton’s study, which may be linked to pH of the post-treatment samples. pH in the ozone-treated samples remained acidic, whereas the Fenton’s reagent treatments resulted in a slightly basic pH.

Based upon comparison of the results observed in the ozone and ozone and peroxide treatments, it is expected that either of these reagents could be effective at field treatment of the target compounds. Although oxons were produced at the lower ozone dose (190 mg/L), the short duration of the test did not allow a determination of whether these products would persist in the presence of ozone over durations more relevant to field scale application. While the low pH of the Site groundwater limited the reactivity of the ozone when used solely in the batch reactor for the short duration study, application of the technology in the field would be over a longer treatment period, and it is not expected that acidity would preclude use of ozone alone as an ISCO approach. Nevertheless, a higher ozone dose, possibly combined with hydrogen peroxide, might be required to achieve complete treatment without production of oxons.

6.2 Chemical Oxidation using Fenton’s Reagent

Fenton’s reagent, at a peroxide dose of 7.9 %, completely degraded all pesticides, contaminants from the pesticide manufacturing process, oxons detected in the site groundwater and most pesticide reaction products. MP-1 was generated in lower concentration treatments, suggesting that these compounds may be observed transiently during a field application of the technology. MP-2-Syre was the only compound that was generated and not subsequently oxidized by higher treatment in the Fenton’s reagent test. This compound was not generated in the ozone study, which achieves oxidation of organic compounds through similar reaction pathways (i.e., production and reaction of hydroxyl radicals). This suggests that the specific reagents used in the Fenton’s study (such as phosphate which is used as a stabilizer of peroxide) may have prevented complete oxidation of MP-2-Syre. This compound is likely a hydrolysis product from methyl parathion, following a similar reaction pathway as that for production of EP-2-Syre from parathion as described by Atkor Innovation (2004) in their review of natural attenuation processes at the Site. As shown by Cho et al. 2002, MP-2-Syre is a known product of the hydrolysis of methyl parathion. Results from Fenton’s study suggest that this reaction may be catalyzed by peroxide and/or metal ions. The molar concentration of MP-2-Syre in the highest dose treatment (treatment C) was several times higher than that of methyl-parathion in the initial sample, suggesting that other pesticides and related compounds may have also served as parent compounds for the production of MP-2-Syre. It is possible that with alternative reagents and/or avoiding the use of phosphate in a field application that this compound would not be generated. Fenton’s reagent at the highest dose tested achieved a 77-fold reduction in toxicity of Site groundwater. These results suggest that MP-2-Syre does not possess significant toxicity as measured by the Microtox® bioassay.

7 References

ATKOR Innovation, 2004. Høfde 42 Depotet: Litteratureundersøgelse og modelbeskrivelse af naturlig nedbrydning af parathion. 24 May 2004.

Cho, C. M-H., A. Mulchandani, and W. Chen, 2002. Bacterial cell surface display of organophosphorus hydrolase for selective screening of improved hydrolysis of organophosphate nerve agents. Appl. Environ. Microbiol. 68(4):2026-2030.

GeoSyntec Consultants and Department of Soil & Water, Connecticut Agricultural Experiment Station, 2006. Literature Review: In Situ Chemical Oxidation of Organophosphorus Pesticides in Groundwater.

GeoSyntec Consultants, 2006. Work Plan for Chemical Oxidation Treatability Tests Høfde 42 Harboøre Tange, Ringkjøbing County, Denmark. March, 2006.

Ringkjøbing County, Department of Environment and Infrastructure, 2004. A DNAPL hotspot of organophosphorous pesticides, Høfde 42 Harboøre Tange, County of Ringkjøbing, Denmark. October, 2004.

LAboratory Treatability Study Report

Cheminova Site
Ringkjobing County, Denmark

June 26, 2006

Prepared For

GeoSyntec Consultants
130 Research Lane, Suite 2
Guelph, Ontario
Canada N1G 5G3

Prepared By

In-Situ Oxidative Technologies, Inc.
51 Everett Drive, Suite A-10
West Windsor, New Jersey 08550

ISOTEC Project No. 801000

Table of Contents

1 Executive Summary

2 Study Objectives

3 Sample Collection and Preparation

4 Laboratory Treatability Study

• 4.1 Reagent Selection
• 4.2 Establishing Experimental Controls
• 4.3 Experimental Setup
• 4.4 Sample Analysis

5 Treatability Study Results

• 5.1 GW-test Results
  • 5.1.1 Pesticide and Degradation Product Results
  • 5.1.2 Toxicity Results
  • 5.1.3 Anion Results
  • 5.1.4 Color Results
  • 5.1.5 pH

6 Summary and Conclusions

1 Executive Summary

In-Situ Oxidative Technologies, Inc. (ISOTEC^SM) was retained by GeoSyntec Consultants (GeoSyntec) to conduct an in-situ chemical oxidation (ISCO) bench-scale laboratory treatability study (study) on soil and groundwater samples collected from the Cheminova site located in Ringkjobing County, Denmark. Targeted contaminants of concern (COCs) for the study are organophosphorous pesticides primarily consisting of insecticide parathion (EP3). The purpose of the study was to determine the potential effectiveness of ISOTEC’s modified Fenton’s process to treat COCs in dissolved phase in site groundwater. The modified Fenton’s reagent promotes contaminant destruction via oxidizing and reducing free radicals including hydroxyl radicals, superoxide radicals and hydroperoxide anions.

The treatability study consisted of a groundwater test (GW-test) to evaluate the COC treatment effectiveness. The test evaluated five different reagent dosages ranging from low to high treatment conditions in an effort to determine the optimal treatment dosage for a potential field application. Analytical samples were submitted to multiple laboratories including Columbia Analytical Services (CAS), Cheminova lab (Cheminova) and Enviro-Test Labs (ETL) for analyses of a variety of parameters (see Section 3.1).

For the reagent dosages evaluated, treatability study results indicated the following:

• Pesticides including MCPA, methyl-parathion, parathion and malathion were reduced by greater than 99% to non-detect (ND) concentrations for the lowest treatment evaluated.
• Parathion degradation products including MP-1, EP-2 Syre, EP-1, PNF, amino-methyl-parathion, and amino-parathion were reduced by greater than 99% to ND concentrations for the lowest treatment evaluated. The only exception was MP-2 Syre (diethylthiophosphoric acid or DETP), which increased in concentration with increasing reagent dosage. It is hypothesized that the presence of phosphate in the ISOTEC reagent mix to promote stabilization of hydrogen peroxide may have led to formation of MP-2 Syre via catalytic hydrolysis or other mechanisms. Therefore, it is suggested that this assumption be tested in a follow-up test using the ISOTEC reagent with no phosphate addition, and if confirmed, the field application of modified Fenton’s reagent should exclude phosphate.
• Groundwater toxicity levels were significantly reduced following modified Fenton’s treatment;
• Individual anion concentrations including nitrate-nitrogen, nitrite-nitrogen and sulfate, and orthophosphate-phosphorous all increased with each treatment possibly due to their formation from pesticide mineralization or their presence in the ISOTEC reagent mix.
• Groundwater color analysis performed by CAS showed inconclusive results possibly due to presence of suspended solids in certain treated samples, which interfered with the analysis. ISOTEC visual observations during the test and also quantitative spectrophotometric absorbance measurements of color absorbance indicated that the sample color intensity (as measured by absorbance) decreased with increasing treatments.

2 Study Objectives

The objectives of the study were as follows:

• Evaluate the COC treatment effectiveness of the modified Fenton’s reagent on site groundwater samples; and
• Select the most effective reagent dosage (i.e. volume) for a potential field scale application at the site.

3 Sample Collection and Preparation

GeoSyntec personnel provided a groundwater sample for the treatability study. The groundwater sample (designated as Hofde 42) was collected from the Cheminova site on April 24, 2006 and shipped to ISOTEC. The samples were kept frozen during shipment and at the ISOTEC facility until commencement the test.

Prior to initiating the treatability study experiment, the sample was thoroughly thawed, and portions of the groundwater were sub-sampled and submitted for various chemical analyses (Table 1 below) to determine the sample initial characteristics. The “initial” sample was designated as F-BGC.

Table 1. List of Analytical Parameters

4 Laboratory Treatability Study

The treatability study experiment, hereafter referred to as GW-test, was performed on the groundwater sample. It consisted of the following four steps:

1. Reagent selection,
2. Establishing experimental control,
3. Experimental setup, and
4. Sample analysis.

4.1 Reagent Selection

ISOTEC’s modified Fenton’s reagent contains a proprietary catalyst and an oxidant. The oxidant used in the reagent was H₂O₂ and the catalyst was ISOTEC’s patented Catalyst 4260 (Cat-4260). Cat-4260 is a circum-neutral pH (e.g. 5-8) organometallic complex with high mobility within the subsurface. Based on historical contaminant levels noted at the site and previous experience with treatment of the compounds of concern, ISOTEC selected this catalyst for the experiments. The stoichiometric molar ratio of Cat-4260 to measured site contaminants was determined and then used to prepare the Cat-4260 containing reagents.

4.2 Establishing Experimental Controls

An experimental “control” sample (identified as F-BGC Control) was set up during the experiment to document the following:

• reduction or changes in concentrations of the target constituents due to sample dilution by reagent volumes injected, and
• reduction in concentrations of the target constituents due to volatilization caused by room temperature test conditions.

The “control” sample was set up exactly the same way, remained at, and was subject to the same conditions as the associated “treatment” reactors. However, the “control” reactor was injected with distilled water instead of reagent (see Section 4.3 below). The volume of distilled water injected was identical to the volumes of reagent injected into the “treatment” reactors.

4.3 Experimental Setup

The GW-test was performed in six (6) identical 1-liter glass reactors sealed with screw-up caps fitted with Teflon-liners. One of the six reactors served as “control” reactor (identified as F-BGC Control, see Section 4.2 above) while the remaining five as “treatment” reactors (identified as F-GT-A through E). Exactly 640 ml of groundwater was introduced into each reactor leaving enough headspace for injection of reagent into reactors and receive a variety of reagent dosages to represent different treatment conditions.

For each “treatment” reactor, a predetermined amount of modified Fenton’s reagent was injected as small incremental dosages. ”Treatment” reactors F-GT-A through C received one, two or three reagent dosages and reactors F-GT-D and F-GT-E received one and three reagent doses at a lower reagent level as shown in Table 2 below . In terms of the volume of groundwater being treated, the treatment reactors F-GT-A through E received an equivalent hydrogen peroxide concentration of 2.7%, 5.3%, 7.9%, 0.9% and 2.8%, and catalyst concentration of 5.3 millimoles per liter (mM), 10.6 mM, 15.9 mM, 1.8 mM and 3.7 mM, respectively (Table 2).

Table 2. Treatment Dosage Summary

The multiple dosage approach (incremental approach) was used during the test to increase treatment efficiency, minimize gas formation and the resulting pressure buildup. Distilled water was used to compensate the difference of reagent volumes applied between reactors. The “control” reactor received an equivalent volume of distilled water instead of reagent. A time gap of approximately 24-48 hours was maintained between dosages and the next dosage was provided only after ensuring the peroxide concentration from the previous injection dropped to less than 100 ppm in a given reactor. All reactors (control and treatment) were left undisturbed until the majority of peroxide was consumed before analytical sample collection. Analytical samples were collected by individually decanting water from each of the “control” and “treatment” reactor into corresponding laboratory pre-cleaned and certified containers and submitted for corresponding analyses as summarized in Table 3.

4.4 Sample Analysis

The analytical services were provided by three laboratories – Cheminova Lab of Harboare, Denmark, CAS of Rochester, NY, USA and ETL of Winnipeg, Manitoba, Canada. The following table (Table 3) summarizes the details of the chemical analyses of the treatability study. In addition, ISOTEC also measured H₂O₂ and Fe levels using Hach Test Kits during the experiment. This data was collected to ensure that the majority of injected reagents were consumed in each reactor prior to the next reagent application.

Table 3. Sample Analyses Summary

Note:
1. Anions include nitrate, nitrate-nitrogen, nitrite, nitrite-nitrogen and sulfate.
2. Orthophosphate was reported as phosphorous.

5 Treatability Study Results

Treatability study results are presented in Table 4 and discussed below. Pesticide treatment effectiveness is evaluated by comparison of “treated” sample data with the associated “control” sample data. As discussed in Section 4.2, “control” sample underwent the same conditions as all “treated” samples but received zero dosage of reagent. Therefore, the differences in contaminant concentrations between “treated” samples and the associated “control” sample best represent the treatment effectiveness. For discussion purpose, all non-detect (ND) values (or U values) are assumed to be equal to zero in the contaminant reduction calculation.

5.1 GW-test Results

Results for the GW-test are presented in Table 4. Pesticide and degradation product results are discussed in Section 5.1.1, toxicity results in Section 5.1.2, anion results in Section 5.1.3 and color results in Section 5.1.4.

5.1.1 Pesticide and Degradation Product Results

Pesticide and associated degradation product results are presented in Table 4. Data indicate that the modified Fenton’s reagent has reduced the concentration of all detected pesticides including MCPA, methyl-parathion, parathion and malathion by >99% to ND levels with the lowest treatment evaluated (i.e. 1 dose @0.9% oxidant). Total pesticide levels decreased from a “control” concentration of 14.26 mg/l to ND levels.

Among the pesticide degradation products including MP-1, EP-2 Syre, EP-1, PNF, amino-methyl-parathion, amino-parathion, malaoxon, methyl-paraoxon and paraoxon, each individual compound which showed a detectable “control” concentration was eventually treated to ND levels except for MP-2-Syre (dimethylthiophosphoric acid), which showed increases with each increasing treatment. Please note that the pesticide degradation product MP-1 was detected at low concentration after the lowest treatment but was treated to ND levels following subsequent doses indicating that this compound may appear transiently during field implementation but is eventually mineralized.

The following plot (Figure 1) shows a graphic representation of the results discussed above.

Figure 1 Pesticide and degradation product results as a function of dose

Based on the above plot, it is clear that all pesticides and degradation products except MP-2-Syre were treated to ND for the lowest dose evaluated. For the remainder of the doses, MP-2-Syre concentration more or less followed a linear increasing trend with increasing reagent dose. It may be noted that some MP-2 Syre was treated at the lowest treatment dose (from 4.25 mg/l to 2.55 mg/l), which indicates that MP-2 Syre can be treated by modified Fenton’s treatment. If MP-2-Syre were formed from parathion or other detected pesticide degradation, its concentration should not have increased after the lowest dose since all detected pesticides were completely mineralized with the lowest dose (>99% treatment) and no more parent pesticide mass is available to generate MP-2-Syre as a daughter product. Hence, it is hypothesized that the presence of phosphate in the ISOTEC reagent mix to promote stabilization of hydrogen peroxide may have led to formation of MP-2 Syre via catalytic hydrolysis of transient intermediate compounds from pesticide mineralization or other mechanisms. We believe that addition of phosphate-free reagents will mitigate the potential for MP-2-Syre formation. Even though some MP-2-Syre formation may occur via catalytic hydrolysis of pesticides, this compound should eventually be completely mineralized. Therefore, it is recommended that the modified Fenton’s reagent used for field application be prepared without phosphate. In summary, we believe that MP-2-Syre is amenable to oxidation and will degrade with subsequent treatments.

5.1.2 Toxicity Results

The toxicity of the sample contents was evaluated via a Microtox^® bioassay testing system that measures the light output from freeze dried luminescent bacteria (Photobacterium phosphoreum). The bacteria’s light-producing mechanism is tied to the metabolic processes of the cell. When the bacteria are killed or bacteria's light-producing mechanism is changed or damaged by toxic substances, a reduction in light output results. These changes in light output are measured to calculate an IC50 for each sample, which is defined as the initial concentration, and is a calculated toxicity value representing the sample concentration expressed in % estimated to cause a 50% response by the exposed test organisms.

The toxicity data presented in Table 4 is the calculated sample concentration expressed in % that causes a 50% response by the exposed test organisms. For a fixed exposure time, because the “control” sample is more toxic compared to the “treated” sample, a greater percent concentration of “treated” sample is needed to generate 50% response than the “control” sample. For example, for 5 minutes of exposure time, an “initial” sample concentration of 0.55% is needed versus a “treated” sample concentration of 42.35% to kill 50% of the exposed test organisms. Overall, the results indicate a significant reduction in sample toxicity from “initial/background” to “treated” sample.

5.1.3 Anion Results

Table 3 presents anion results. Individual anion concentrations including nitrate-nitrogen, nitrite-nitrogen and sulfate, and orthophosphate-phosphorous all increased with each treatment possibly due to their formation from pesticide mineralization or their presence in the ISOTEC reagent mix. Overall, the concentrations increased with increasing reagent dosage.

5.1.4 Color Results

Table 3 presents results from the color test. Data indicates fluctuations in color concentrations as a result of the application of modified Fenton’s reagent. We believe the suspended solids present in the samples submitted to CAS may have interfered with color measurements via USEPA Method 110.2.

ISOTEC conducted color measurements on filtered samples via a spectrophotometer prior to shipping the samples to CAS. The color absorbance was measured at a wave length of 470 nano meters (nm). Color intensity is directly proportional to the absorbance (i.e. greater the absorbance greater is the color intensity). The color absorbance data collected by ISOTEC is plotted in Figure 2.

Figure 2 Color absorbance as a function of dose

Results indicate an overall decrease in the sample color absorbance following modified Fenton’s treatment, suggesting decreasing color intensity with increasing treatment.

5.1.5 pH

The final pH value measurements indicate that the treatment occurred in the pH range 3.62-8.30 with the “control” sample present at a pH of 2.91. The following plot shows the pH change as a function of increasing reagent dose. It is clear from the above plot that increasing reagent dose increased the sample pH. The ISOTEC reagents were prepared at a pH of 5-6 with no additional pH adjustment of the sample performed. Under normal buffering associated with addition of pH 5-6 reagents, the final pH was expected to be in the range of 3.0-6.0. Since the final pH of treatment samples B, C, D and E was in the range 7.88-8.30, it is hypothesized that the pH may have increased as a result of alkaline byproduct formation.

Figure 3. pH value as a function of dose

6 Summary and Conclusions

The treatability study consisted of a pesticide GW-test. Results indicate that the modified Fenton’s reagent was effective towards complete destruction of all pesticides and degradation products with the exception of MP-2-Syre. It is hypothesized that MP-2-Syre may have formed due to phosphate presence within the ISOTEC reagent mix via catalytic hydrolysis of transient intermediate compounds from pesticide mineralization or other mechanisms. Therefore, it is suggested that this assumption be tested in a follow-up test using the ISOTEC reagent with no phosphate addition, and if confirmed, field application of modified Fenton’s reagent should exclude phosphate.

Groundwater toxicity levels were significantly reduced following modified Fenton’s treatment. Individual anion concentrations all increased with each treatment following the expected trend. Color absorbance of the samples decreased following the modified Fenton’s treatment indicating decreased color intensity.

The data suggests that a pilot study should be conducted at the site to gather additional data on the effectiveness of this remedial alternative on a large-scale basis under field conditions. A pilot application would also serve as an initial step towards complete site remediation. Upon request, the pilot study design and approach including the reagent volumes to be applied will be provided in a pilot study work plan issued under separate cover.

Tables

Applied Project Number P-1835

HiPOx™ Technology Lab Testing:
Cheminova

Report Authors: Keel Robinson and Reid H. Bowman

Report Date: June 20, 2006

Prepared for

Leah MacKinnon
GeoSyntec Consultants
130 Research Lane, Suite 2
Guelph, Ontario N1G 5G3

APPLIED PROCESS TECHNOLOGY, INC.
3333 Vincent Road, Suite 222, Pleasant Hill, CA 94523
Phone: (925) 977-1811; Fax: (925) 977-1818
www.aptwater.com

Table of Contents

7 Background Information

• 7.1 HiPOx Technology
• 7.2 Objective of Evaluation
• 7.3 Process Water Information

8 Test Equipment and Procedures

• 8.1 Test Equipment Description
• 8.2 Test Procedures

9 Discussion

• 9.1 Ozone Only Results
• 9.2 HiPOx Ozone and Hydrogen Peroxide Results

10 Conclusion

7 Background Information

7.1 HiPOx Technology

The HiPOx™ process developed by Applied Process Technology, Inc. (Applied) is an Advanced Oxidation Process (AOP) that uses ozone (O₃) and hydrogen peroxide (H₂O₂) to destroy organic compounds. Ozone dissociates as well as reacts with hydrogen peroxide to produce an intermediate, hydroxyl radical (•OH). Hydroxyl radicals are the second most powerful oxidizing agent found in nature. These hydroxyl radicals react very rapidly to oxidize organic contaminants to non-hazardous compounds, carbon dioxide, and water.^[1] The oxidation of the organic contaminants does not increase the temperature or pressure of the treated water because of the low mg/L or sub-mg/L concentration of contaminants.

Advanced oxidation chemistry using ozone and hydrogen peroxide to create hydroxyl radicals is base-catalyzed and is very well known. The overall balanced reaction for ozone and hydrogen peroxide to yield hydroxyl radical is shown in Equation 1.^[2]

2O3 + H2O2 → 2•OH + 3O2 (Equation 1)

7.2 Objective of Evaluation

The objective of this laboratory trial was to determine the efficiency of the HiPOx technology utilizing ozone and hydrogen peroxide to reduce the concentrations of pesticides and to reduce the toxicity in the groundwater. The final application will be an in-situ process and not an ex-situ process, however, the goal was to confirm that an Advanced Oxidation Process (AOP) could treat the contaminants before conducting in-situ AOP testing.

7.3 Process Water Information

The groundwater sample for the bench test was collected by GeoSyntec Consultants from the Cheminova Site. The water sample was received on May 8, 2006, and was labeled P-1835. The sample was frozen when received. Once received, the sample was logged in and stored in a freezer to maintain the sample frozen until tests were performed.

8 Test Equipment and Procedures

8.1 Test Equipment Description

The HiPOx reactor is constructed of PVC. All ozone piping is constructed of either 316 Stainless Steel or Teflon. The hydrogen peroxide piping is constructed of either 316 Stainless Steel or polyethylene. The ozone generator utilized was an ASTeX Model 8200. The ozone is injected under pressure into the water to be treated. Mixing of the ozone into the water is accomplished with an in-line static mixer. The vent from the system is passed through an ozone destruct unit manufactured by Pacific Ozone.

8.2 Test Procedures

Test Conditions

The first HiPOx bench test was conducted on May 9-10, 2006. A second HiPOx bench test was conducted on May 15, 2006. The water sample was thawed prior to the first test and immediately tested. The remaining water was stored frozen until the second test. The frozen water was thawed immediately prior to the second test on May 15, 2006. The pH, alkalinity and turbidity were measured on P-1835 prior to the bench testing.

The test procedures are described below. Sample designations reflect sample and treatment levels. For example, SP-190-1835 represents water sample P-1835 treated with 190 mg/L of ozone. There were two types of oxidation methods evaluated in the bench tests, ozone only and Advanced Oxidation (Ozone and Hydrogen Peroxide). The dosing levels in the ozone only tests were limited to the solubility of ozone in the water sample.

Ozone Only Test Procedure:

The test procedure was as follows: A sample (2.05 liters) was treated with ozone (5,530 mls, 5.19% by wt in oxygen, 185.9 mg/L). Samples of the treated water (SP-190-1835) were taken and analyzed for nitrate, nitrite, sulfate, orthophosphate, color, pesticide and Microtox. The amounts of ozone applied during each run are summarized in Table 1 (Run #1) and Table 2 (Run #1).

Advance Oxidation (Ozone and Hydrogen Peroxide) Test Procedure:

The test procedures were as follows. Sample designations reflect sample and treatment levels. For example, SP-500-1835 represents water sample P-1835 treated with 500 mg/L of ozone. The test procedure was as follows: A sample (2.05 liters) was treated with hydrogen peroxide (12.0 mls, 5.0% by wt, 292.2 mg/L). This mixture was then treated with ozone (15,060 mls, 5.04% by wt in oxygen, 491.7 mg/L). Samples of the treated water (SP-500-1835) was taken and analyzed for nitrate, nitrite, sulfate, orthophosphate, color, pesticide and Microtox. The amounts of ozone and hydrogen peroxide applied during each run are summarized in Table 1 (Run #2 & 3) and Table 2 (Run #2).

Table 1. Summary of Experimental Conditions for P-1835 May 9-10, 2006 Lab Test⁵

1 Residual at the end of the Run in the water
2 2.40 wt % ozone was detected in the vent
3 0.39 wt % ozone was detected in the vent
4 0.14 wt % ozone was detected in the vent
5 The analytical samples taken were not quenched

Table 2. Summary of Experimental Conditions for P-1835 May 15, 2006 Lab Test⁴

1 Residual at the end of the Run in the Water
2 2.61 wt % ozone was detected in the vent
3 0.51 wt % ozone was detected in the vent
4 The analytical samples were quenched. The nitrate, nitrite, sulfate, and orthophosphate samples were quenched with methanol, while the pesticide and Microtox samples were quenched with sodium thiosulfate.

Analytical Testing

Samples for nitrate, nitrite, sulfate, orthophosphate from the HiPOx test were sent for analysis to STL San Francisco, 1220 Quarry Lane, Pleasanton, Ca 94566, Phone (925) 484-1919. Samples for Microtox were sent to Envirotest Laboratories, 745 Logan Ave, Winnipeg, MB R3E 3L5 Canada Phone: (204) 945-3705. Samples for pesticides were sent to Cheminova, Attention Bo Breinbjerg Thyborønvej, 78 DK-7673 Harboøre, Denmark Phone: +45 96 90 96 90. The samples collected on May 9 and 10 were not quenched. On May 15, the nitrate, nitrite, sulfate, and orthophosphate samples were quenched with methanol, while the pesticide and Microtox samples were quenched with sodium thiosulfate.

Applied Analytical Methods:

The turbidity meter used was an Orbeco-Hellige Model 965-10 Serial # 2222. The pH was measured with an Oakton Model Ph Tester 3⁺. Alkalinity was measured using a Hach Model 5-EP test kit. Hydrogen peroxide residuals were measured with Quantofix Peroxid 25 test strips. Ozone residual was measured with a Hach Model OZ-2 test kit. The nitrate concentrations were measured using a HACH Model# 820, for the nitrate and Nitraver 5 test reagent.

9 Discussion

9.1 Ozone Only Results

The results and conditions for the treatment of P-1835 with ozone are summarized in Tables 3 (Run #1) and Table 4 (Run #1). The water sample, P-1835, was clear and had a strong chemical odor when the sample was received. The addition of ozone was discontinued when ~ 190 mg/L of ozone was added because the ozone concentration detected in the vent off of the lab reactor exceeded 2.5 weight % in oxygen. The ozone concentration feeding the lab reactor was 5.0 weight % in oxygen. Therefore 50% of the ozone being supplied to the system was being vented. This is attributed to the limited solubility of ozone in water and the relatively slow rate of ozone reacting with contaminants. However, the ozone only treatment of P-1835 resulted in a majority of the pesticides being oxidized with the exception of trace amounts of MP-2-Syre (0.40 mg/L) and 4-chlorocresol (0.47 mg/L). This represents a significant reduction in the concentration of pesticide in P-1835. Three compounds concentration increased after ozone only treatment EOOOPO, Methyl–paraoxon and Paraoxon. These three compounds are oxidation by-products of one or more of the parent pesticides present in P-1835. Also, a significant amount of the pesticides appear to have been mineralized to their elements. The mineralization is based on the amount of nitrate, orthophosphate and sulfate produced after treatment with ozone.

The amounts of nitrogen, phosphorus and sulfur derived from the pesticides are shown in Table 5 (May 10, 2006 SP-190-1835, Ozone Only) and Table 6 (May 15, 2006, SP190-1835, Ozone Only). For example, the calculate amount of nitrogen derived from the pesticides is 1.528 mg/L. The nitrogen is converted to nitrate when oxidized. The nitrogen (1.528 mg/L) in the pesticides is equal to 6.77 mg/L as nitrate. Nitrogen has a molecular weight of 14 and nitrate has a molecular weight of 62. Therefore, the nitrogen concentration is multiplied by 4.43 (62/14) to calculate the amount of nitrogen as nitrate. P-1835 had an initial nitrate concentration of 15 mg/L.^[3] If all the nitrogen in the pesticides had been converted to nitrate, the total nitrate would be 21.8 mg/L. The observed nitrate in the SP-190-1835 on May 10, 2006 was 18 mg/L. This indicates that approximately half of the nitrogen in the pesticides was mineralized.

The phosphorus and sulfur derived from the pesticides are converted to orthophosphate and sulfate, respectively, when oxidized. The corrected phosphorus as phosphate and sulfur as sulfate derived from the pesticides is shown in Table 5 and 6 as “Corrected”. P-1835 had an initial orthophosphate and sulfate concentrations of 4.7 and 460 mg/L, respectively. If all the phosphorus and sulfur in the pesticides had been converted to orthophosphate and sulfate, the total phosphate and sulfate concentration would be 31.3 and 495.5 mg/L, respectively. The observed orthophosphate and sulfate in the Ozone Only test on May 10, 2006 were 13 and 560 mg/L, respectively. The orthophosphate concentration observed indicates that approximately half of the phosphorus in the pesticides was mineralized. The sulfate concentration observed are greater than expected based on the pesticides as the only source of sulfur. There may be other sulfur containing compounds in P-1835 that contribute to the high sulfate concentrations observed. The same trends observed for nitrate, orthophosphate and sulfate were observed in the May 15, 2006 SP-190-1835 (Table 6).

The samples taken on May 10 were not quenched and the samples taken on May 15 were quenched. The concentration of pesticides in both SP-190-1835 taken on May 10 and May 15 were very similar. This indicates that the residual ozone in the samples had very little impact of the pesticides concentration. The nitrate, orthophosphate and sulfate concentrations observed were different. The anion sample SP-190-1835 sample taken on May 10, which was not quenched, had lower nitrate concentrations but higher phosphate and sulfate concentration than the anion SP-190-1835 sample that was quenched on May 15. The anions results for the SP-1000-1835 quenched or not quenched followed the same pattern as the SP-190-1835. The significance in the difference nitrate, orthophosphate and sulfate concentrations observed in the quenched and the sample not quenched is unclear.

9.2 HiPOx Ozone and Hydrogen Peroxide Results

The results and conditions for the treatment of P-1835 with ozone and hydrogen peroxide are summarized in Tables 3 (Run #2 and 3) and Table 4 (Run #2). In Table 3, the results for two ozone doses of 491.7 and 949.0 mg/L are presented. The only pesticide detected in SP-500-1835 was MP-2-Syre (0.32 mg/L). This represents a very significant reduction in pesticide concentration in P-1835. Two oxidation by-products were observed in SP-500-1835, EOOOPO (0.7 mg/L) and Malaoxon (0.2 mg/L). In SP-1000-1835, all pesticides were below their detection limits and no by-products were observed. Although hydrogen peroxide was present to react with the ozone to yield the hydroxyl radical, ozone was detected in the vent. The concentration of ozone detected in the vent was significantly less than in the ozone only tests. This is pointed out because the concentrations of ozone shown in Tables 3 and 4 are the applied ozone doses and the concentration of ozone consumed maybe less than the applied ozone doses.

Using the same analysis of nitrogen, phosphorus and sulfur used to determine the degree of mineralization of the pesticides in the ozone only test, it appears nearly complete mineralization occurred when ozone and hydrogen peroxide were used. The mineralization is based on the amount of nitrate, orthophosphate and sulfate found in P-1835 and the amount of these inorganics after treatment with ozone and hydrogen peroxide. The amounts of nitrogen, phosphorus and sulfur derived from the pesticides are shown in Table 7 (May 10, 2006, SP-1000-1835) and Table 8 (May 15, 2006, SP-1000-1835). P-1835 had an initial nitrate concentration of 15 mg/L (Table 7). If all the nitrogen in the pesticides had been converted to nitrate, the total nitrate would be 21.8 mg/L. The observed nitrate in SP-1000-1835 May 10, 2006 was 21 mg/L. This indicates that greater than 96% of the nitrogen in the pesticides was mineralized.

If all the phosphorus and sulfur in the pesticides had been converted to orthophosphate and sulfate, the total phosphate and sulfate concentration would be 31.3 and 495.5 mg/L, respectively. The observed orthophosphate and sulfate in the SP-1000-1835 on May 10, 2006 were 40 and 470 mg/L, respectively, (Table 7). The orthophosphate concentration observed is higher than can be attributed to phosphorus in the pesticides. There may be other phosphorus containing compounds in P-1835 that contribute to the high orthophosphate concentrations. The sulfate concentration observed was slightly less than expected based on the sulfur in the pesticides.

In the May 15, 2006 SP-1000-1835 (Table 8), the nitrate concentrations in SP-1000-1835 were slightly greater, 24.0 mg/L (observed) vs. 20.7 mg/L (estimated) than can be attributed to the nitrogen in the pesticides. The observed concentrations of orthophosphate and sulfate are consistent with 100 % mineralization of the phosphorus and sulfate in the pesticide.

There were only three toxicology test run, SP-0-1835 on both May 10 and 15 and SP-1000-1835 on May 15, 2006. Applied is not qualified to address the toxicological aspects of the testing.

Table 3 Summary of Results and Conditions for May 10 Test.

1 The value presented is the average of two analysis
ìg/L - micrograms per litre
mg/L - milligrams per litre
v/v (%) - volume per volume percentage
NA - Not Analyzed
NM-Interference with the low pH No measurement could be conducted

Table 4 Summary of Results and Conditions for May 15 Test.

1 The value presented is the average of two analysis
ìg/L - micrograms per litre
mg/L - milligrams per litre
v/v (%) - volume per volume percentage
NA - Not Analyzed
NM-Interference with the low pH No Measurement could be conducted

Table 5. Inorganic Material Balance for Ozone Only May 10, 2006 Test

1 Nitrogen was correct to Nitrate, Phosphorus was corrected to orthophoshate and sulfur was corrected to sulfate
2 The initial concentration for nitrate, orthophoshate and sulfate
3 Sum of the intial concentration plus the inorganic produced

Table 6. Inorganic Material Balance for Ozone Only May 15, 2006 Test

1 Nitrogen was correct to Nitrate, Phosphorus was corrected to orthophoshate and sulfur was corrected to sulfate
2 The initial concentration for nitrate, orthophoshate and sulfate
3 Sum of the intial concentration plus the inorganic produced

Table 7. Inorganic Material Balance for SP-1000-1835 May 10, 2006 Test

1 Nitrogen was correct to Nitrate, Phosphorus was corrected to orthophoshate and sulfur was corrected to sulfate
2 The initial concentration for nitrate, orthophoshate and sulfate
3 Sum of the intial concentration plus the inorganic produced

Table 8. Inorganic Material Balance for SP-1000-1835 May 15, 2006 Test

1 Nitrogen was correct to Nitrate, Phosphorus was corrected to orthophoshate and sulfur was corrected to sulfate
2 The initial concentration for nitrate, orthophoshate and sulfate
3 Sum of the intial concentration plus the inorganic produced

10 Conclusion

HiPOx technology, which uses ozone and hydrogen peroxide, was very effective in the oxidation of the pesticides in P-1835. The pesticide concentrations were below detection limits for all pesticides when an ozone dose of ~950 mg/L of ozone was applied with ~ 650 mg/L of hydrogen peroxide. The pesticides appear to have been mineralized based on the material balance of nitrogen, phosphorus and sulfate.

When testing with ozone only, the ozone dose could not exceed 190 mg/L because of the limited solubility of ozone in water and the relatively slow reaction rate of the ozone with the pesticides. However, the pesticide concentrations in P-1835 were significantly reduced but complete mineralization was not observed with ozone only and oxidation by-products of the parent pesticides were observed.

This series of tests were attempting to address the effectiveness of ozone and ozone/ hydrogen peroxide in an in-situ application. The limitation of ozone only experiments in this test was loss of ozone because of limited solubility of ozone in water. In an in-situ application of ozone, the limited solubility of ozone may not be significant. In an in-situ application the ozone may be applied at a slower rate providing time for the ozone to react with the pesticide contaminants. Also, in an in-situ application the ozone may react with material in the soil to generate hydroxyl radicals.

When requested by GeoSyntec, Applied will provide an engineering cost estimate for a PulseOx in-situ treatment system at the Cheminova site. The PulseOx system offered by Applied has the capacity to inject ozone only and ozone and hydrogen peroxide into the subsurface.

[1] W. Glaze and J. Kang, J.Amer. Water Works Assoc., 80, 51, (1988).

[2] J. Staehelin and J. Holgné, Environ. Sci Technol., 16, 676 (1982)

[3] The nitrate values reported in this section are the nitrate values obtained from STL and not the nitrate values observed with the Applied Hach test.

Bilag 3.2 Permanganatforsøg

Laboratorieforsøg med kaliumpermanganat.

Udført af COWI A/S

Indhold

Sammenfatning og konklusioner

1 Indledning

2 Oxidation med permanganat

3 Test procedure

• 3.1 Metodebeskrivelse
• 3.2 Analyse- og målemetoder

4 Resultater

• 4.1 Analyse af startprøven fra høfde 42
• 4.2 Orienterende forsøg med permanganat forbrug
• 4.3 Redoxpotentiale som funktion af pH
• 4.4 Oxidationsforsøg
• 4.5 Nøjagtige forsøg med permanganat forbrug
• 4.6 Farvefjernelse med aktiv kul

5 Konklusion

Bilag 1: Orienterende forsøg med permanganatforbrug

Bilag 2: Oxidationsforsøg

Bilag 3: Nøjagtige forsøg med permanganatforbrug

Bilag 4: Komplet resultatskema med Cheminova-analyser

Sammenfatning og konklusioner

I forbindelse med forsøgene udviklede vi et regneprogram til beregning af elementsammensætningen af vandprøven ud fra de 22 organiske forbindelser, som blev analyseret på Cheminovas laboratorium.

Analyserne af startprøven viste, at man ved Cheminovas analyse af 22 organiske forbindelser formentlig havde identificeret ca. 2/3 af det samlede indhold af organiske stoffer. Der er således en række andre organiske forbindelser i vandet, som ikke er identificeret og kvantificeret

Cheminovas analyser kunne redegøre for 58,4 % af det organiske kulstof, 38,8 % af det organiske kvælstof og 65,5 % af det organiske phosphor i det oprindelige vandprøve fra høfde 42. Den oprindelige prøve var meget sur med en pH-værdi på 3,5 - velegnet til en permanganat oxidation, hvor der overvejende vil dannes mangano ioner.

Ved oxidation med permanganat forbrugte vandprøven 640 mg/l MnO4 over 3 døgn. Det svarer ca. til 50 % af det iltforbrug, som er bestemt ved COD på prøven, idet COD = 170 mg/l. Det viser, at en oxidation med permanganat ikke er en komplet af alle organiske forbindelse til vand og kuldioxid.

Oxidation med permanganat viste en overraskende hurtig nedbrydning af pesticider, nedbrydningsprodukter og råvarer urenheder. De fleste stoffer forsvandt inden for to timer, men enkelte stoffer krævede dog op til 72 timer, før de var helt væk fra reaktionsflasken.

Desværre blev der ikke i alle tilfælde opnået en fuldstændig nedbrydning af de organiske stoffer. Parathion, methyl-parathion og malathion blev således i et vist omfang (20-35 %) oxideret til de tilsvarende oxoner, der er mindst lige så giftige som parathionerne. Også E-OOOPS blev oxideret til den tilsvarende iltholdige forbindelse: E-OOOPO. På grund af oxon dannelse og manglende nedbrydning heraf synes metoden ikke umiddelbart velegnet som fuld-skala oprensningsmetode ved høfde 42, selv om det rent teknisk og praktisk nok er det letteste oxidationsmiddel at anvende.

Da permanganat er kendt for sin langtidsvirkning, kan det naturligvis ikke helt udelukkes, at oxonerne kan nedbrydes ved en betydelig længere varende iltningsproces end de 72 timer, der er anvendt ved COWIs undersøgelse. Det skal dog nævnes, at udviklingen i oxonkoncentration de første 72 timer af forsøget ikke tyder på, at koncentrationen falder med tiden - snarere tværtimod.

1 Indledning

I COWIs oprindelige oplæg til forsøg med kemisk oxidation var oxidation med permanganat nævnt som en mulighed, der måske skulle undersøges, men i det indsendte tilbud var disse forsøg ikke medtaget i budgettet.

På styregruppemøde i Miljøstyrelsen den 7. april 2006 var der enighed om, at permanganat burde undersøges, da det på grund af sin langtidsvirkning ville være meget velegnet at anvende i fuld skala, såfremt stoffet var i stand til at oxidere parathioner og nedbrydningsprodukter heraf.

COWI udarbejdede et forslag til forsøgsplan, og der blev bevilget ekstra midler af Miljøstyrelsen til at gennemføre disse ekstra forsøg på COWIs laboratorium i Lyngby. Denne undersøgelse og resultaterne herfra er nærmere beskrevet i denne rapport.

2 Oxidation med permanganat

Natrium og kalium salte af permanganat er velkendte oxidationsmidler, som har vundet en del udbredelse inden for jord- og grundvandsrensning. Permanganat udmærker sig ved at være stabil, og virkningen af stoffet kan derfor ofte holde sig i flere måneder i forbindelse med in-situ rensning, hvor det injiceres i grundvandet.

Litteraturen angiver, at permanganat er meget velegnet til oxidation af PCE, TCE, DCE, VC, BTEX, PAH og phenoler. Permanganat kan også - om end med større besvær - oxidere benzen og pesticider. Permanganat er derimod ikke særlig effektiv over for TCA, tetrachlorkulstof, CHCl₃ og PCB. Permanganat er generelt god til oxidation af kulstof-kulstof dobbeltbindinger, men mindre effektiv til oxidation af aromatiske forbindelser.

Ved pH < 3,5 omdannes permanganat til Mn⁺², hvilket betyder, at iltningstrinnet går fra +7 til +2 efter ligningen:

MnO₄^- + 8H⁺ + 5e^- → Mn⁺² + 4H₂O

Ved pH mellem 3,5 og 12 omdannes permanganat til MnO₂ (brunsten), hvilket betyder, at iltningstrinnet går fra +7 til +4 efter ligningen:

MnO₄^- + 2H₂O + 3e^- → MnO₂ + 4OH^-

Den dannede brunsten vil dog efterhånden kunne omdannes til Mn⁺² ved fortsat reaktion med de reducerende stoffer, som er til stede, og det giver på længere sigt øget udbytte af det tilsatte permanganat.

Ved pH > 12 omdannes permanganat til manganat, hvilket betyder, at mangans iltningstrinnet går fra +7 til +6 efter ligningen:

MnO₄^- + e^- → MnO₄^-2

Som det fremgår af disse reaktioner, kan permanganat i sur væske (pH < 3,5) oxydere 5 gange så meget stof, som i basisk væske (pH > 12) og 67 % mere end ved pH = 3,5-12.

Standard oxidationspotentialet for permanganat er 1,7 V sammenlignet med brintperoxid på 1,8 V, ozon på 2,1 V og chlor på 1,4 V. Permanganat er således et svagere oxidationsmiddel end brintperoxid og ozon, men lidt stærkere end chlor.

Da natriumpermanganat er mere opløseligt i vand end kaliumpermanganat, vil man ofte foretrække at dosere en opløsning af natriumpermanganat, men prisen spiller naturligvis også en rolle.

3 Test procedure

Hovedformålet med denne laboratorietest er at finde forbruget af permanganat, samt at undersøge i hvor høj grad permanganat kan nedbryde parathioner og nedbrydningsprodukter heraf på en vandprøve fra høfde 42.

3.1 Metodebeskrivelse

Vandprøverne anbringes i forseglede flasker (reaktorer)med ca. 130 ml væskeindhold, idet der tilsættes en passende mængde permanganat. Flaskerne henstilles i mørke ved stuetemperatur (22-24 ^oC) i 3 døgn, og der udtages prøver til analyser efter 2, 5, 24, 48 og 72 timer. Ved analysen kontrolleres pH og redoxpotentialet samt indhold og forbrug af permanganat. Endvidere analyseres løbende for indhold af phosphat, ammoniak-N og nitrat-N samt farve.

Efter den ønskede reaktionstid stoppes oxidationsprocessen ved at fjerne overskud af permanganat i flasken ved tilsætning af natriumbisulfit. Der tilsættes 1 ml opløsning med 400 g/l af natriumbisulfit. Ved doseringen forsvinder den violette farve af permanganat straks, hvilket samtidig er beviset på, at permanganat er destrueret og oxidationsprocessen dermed stoppet.

Doseringen af permanganat er fastlagt ud fra en indledende test med dosering af tre forskellige koncentrationer af permanganat. Flaskerne opbevares i mørke ved stuetemperatur (22-24 ^oC) i 1-3 døgn.

Redox og pH blev målt direkte på prøven, mens permanganat først blev målt efter dekantering eller filtrering på et 0,45 µ membranfilter.

Efter dosering af natriumbisulfit blev der straks målt ammoniak-N, nitrat-N, phosphat og farve.

Prøve (50-60 ml) til analyse for parathioner og nedbrydningsprodukter blev straks nedfrosset i 100 ml PE-flasker, og de blev samlet sendt til Cheminova til analyse efter hele forsøgsserien.

3.2 Analyse- og målemetoder

I COWIs laboratorium er anvendt følgende analyse- og målemetoder:

Det skal bemærkes, at permanganat ved de første forsøg blev målt efter metode 1), men målingen var meget ustabil, hvilket gav stor måleusikkerhed. Senere fandt vi ud af, at målingen var noget mere stabil, når man anvendte metode 2), hvilket var tilfælde ved de sidste forsøg, hvor det nøjagtige forbrug af permanganat blev bestemt over 7 døgn.

Udgangsprøven fra høfde 42 er endvidere analyseret hos Analytech Miljølaboratorium for følgende parametre:

Cheminovas analyser omfatter den såkaldte "Cheminova pakke" (17 stoffer), hvortil kommer oxoner samt E-OOOPO og EEM-OOSPO (i alt 22 stoffer). Cheminova har anvendt følgende analysemetoder:

Pesticider, oxoner og produkturenheder analyseret ved GC efter ekstraktion

Nedbrydningsprodukter analyseret direkte ved HPLC.

De 22 stoffer ved Cheminovas analyser fremgår af tabel 4.2

4 Resultater

4.1 Analyse af startprøven fra høfde 42

Startprøven er dels analyseret på COWIs laboratorium og dels på Analytech Miljølaboratorium for udvalgte parametre samt på Cheminovas laboratorium. COWIs analyser er foretaget med simpelt måleudstyr og fotometre med det formål løbende at følge udviklingen i forsøgene.

Tabel 4.1: Generelle analyseresultater fra startprøven fra Høfde 42.

Tabel 4.2: Analyse af startprøven fra høfde 42 for pesticider, nedbrydningsprodukter, råvarer urenheder og oxidationsprodukter. Prøven er analyseret af Cheminovas laboratorium ved HPLC og GC.

Tabel 4.3: Beregnet indhold i % af kulstof, brint, kvælstof, ilt, phosphor, svovl og chlor ud fra den kemiske bruttoformel for de enkelte komponenter, der er analyseret af Cheminovas laboratorium.

Tabel 4.4: Beregnet indhold i mg/l af kulstof, brint, kvælstof, ilt, phosphor, svovl og chlor ud fra Cheminovas analyse og det procentvise indhold af grundstoffer. Den procentvise fordeling på de 7 grundstoffer er beregnet for summen af de 22 analyserede komponenter. Resultatet af TOC, organisk-N og organisk-P er anført til sammenligning.

I tabel 4.3 er den procentvise sammensætning opdelt på grundstofferne (kulstof, brint, kvælstof, ilt, phosphor, svovl og chlor) beregnet for de 22 Cheminova-stoffer. I tabel 4.4 er den procentvise sammensætning beregnet ud fra Cheminovas analyse af startprøven, og resultaterne for kulstof (TOC), kvælstof (organisk-N) og phosphor (organisk-P) er sammenlignet.

Diskussion

Vandet har meget lav pH-værdi samt en høj ledningsevne som følge af infiltreret havvand. Vandet har en svag gul farve og lugter kraftigt af svovlforbindelser. Bemærk, at COWIs analysemetoder tilsyneladende giver lidt for høje værdier for ammoniak og nitrat, mens der er fin overensstemmelse for phosphat.

Ud fra tabel 4.4 giver de 22 komponenter i Cheminovas analysepakke følgende procentvise fordeling på grundstoffer

Det bemærkes, at man ved Cheminovas analyse for 22 stoffer kun har fundet stoffer svarende til 58% af det fundne TOC. Tilsvarende har man fundet 39% af organisk-N og 66% af organisk-P. Der findes med andre ord stadig en 30-40% organiske stoffer, som vi ikke kender navnene på.

Går vi ud fra TOC = 45,2 mg/l og et gennemsnitligt kulstofindhold i de organiske forbindelser på 35% svarer det til i alt 132 mg/l organisk stof (med samme grundstoffordeling). Lægges organisk-P til grund for beregningen fås 118 mg/l organisk stof. Regnes ud fra organisk kvælstof fås et noget højere tal, som der dog er en del usikkerhed på. Med et forholdsvis stort iltindhold i det organiske stof stemmer det meget godt overens, at indholdet af organiske stoffer på 120-130 mg/l svarer til en COD-koncentration på 170 mg/l.

4.2 Orienterende forsøg med permanganat forbrug

Der blev i første omgang foretaget 3 forsøg med forskellige doseringer af permanganat for at finde frem til en passende dosering ved de endelige forsøg. Tre flasker (á ca. 130 ml) blev fyldt med prøve og tilsat varierende mængder (1, 4 og 10 ml) stamopløsning med kaliumpermanganat (MnO4 = 34,2 g/l). Stamopløsningen blev fremstillet ved udvejning af teknisk kaliumpermanganat, som efterfølgende blev analyseret såvel fotometrisk som ved titrering med en kendt opløsning med jern(II) salt. Det komplette laboratorieskema fra forsøgene findes i bilag 1.

Her skal hovedresultaterne præsenteres.

Tabel 4.5: Koncentration i mg/l af MnO₄ i reaktionsflasken ved dosering af henholdsvis 1, 4 og 10 ml permanganat stamopløsning. Startkoncentrationen af permanganat er beregnet ud fra fortyndingsgraden af MnO4-stamopløsningen.

Fig 4.1: Koncentration i mg/l af MnO4 i reaktionsflasken ved dosering af henholdsvis 1, 4 og 10 ml permanganat stamopløsning

Resultaterne i tabel 4.1 tager ikke højde for, at der er lidt forskellig prøvemængde i de tre reaktionsflasker som følge af, at der er doseret forskellige rumfang permanganat. Udregnes forbruget af MnO4 pr. liter startprøve, er resultatet for de tre forsøg efter 72 timer:

Det skal bemærkes, at C-1 var farveløs efter 24 timer svarende til, at alt permanganat var blevet forbrugt.

Diskussion

Koncentration af permanganat i reaktionsflasken kan naturligvis ikke stige med tiden, som resultaterne viser. Når det ser sådan ud i tabel 4.2 og fig. 4.1 skyldes det den lidt unøjagtige kolorimetriske målemetode, som blev anvendt til bestemmelse af permanganat ved de første forsøg

Forsøg C-1 viser tydeligt, at alt permanganat blev brugt ved den aktuelle dosering, idet prøven var affarvet efter 24 timer. Det faktiske forbrug, vil derfor være større end den dosering af permanganat, som blev anvendt i dette forsøg det vil sige større end 264 mg MnO4 pr. liter prøve. Noget tyder på et faktisk forbrug mellem 300 og 350 mg pr. liter prøve. I de følgende forsøg vil blive anvendt en dosering på 10 ml stamopløsning til en reaktionsflaske med i alt 130 ml indhold for at sikre et rigeligt overskud af permanganat.

4.3 Redoxpotentiale som funktion af pH

Redoxpotentialet for en oxidationsproces er afhængig af de i processen indgående stoffer - herunder reduktions- og oxidationsmidler samt pH. Ved oxidation med permanganat, hvor der både kan dannes Mn⁺² (mangano) og MnO₂ (brunsten,) kan det være vanskeligt at beregne redoxpotentialets afhængighed af pH. Vi har derfor valgt at bestemme sammenhængen mellem redoxpotentiale og pH for det aktuelle spildevand eksperimentelt, idet vi går ud fra, at begge oxidationsprocesser vil finde sted, da startprøvens pH ligger lige omkring 3,5.

I en opløsning med ca. 120 ml prøve og 10 ml stamopløsning af permanganat opnås en permanganat koncentration på 2630 mg/l. I denne opløsning er pH-værdien gradvis blevet hævet ved tilsætning af saltsyre og sammenhørende værdier af pH og redoxpotentiale er målt. Sammenhængen fremgår af fig. 4.2. Redoxpotentialet falder 72,2 mV ved en pH-stigning på 1 enhed.

Fig 4.2: Redoxpotentialet som funktion af pH i vandprøve fra høfde 42 tilsat 2630 mg/l permanganat. Der opnås som ventet en lineær sammenhæng: Redox = -72,2 x pH + 1286 Den tilnærmede linie er indtegnet på grafen. Som det fremgår, falder pH med 72,2 mV ved en pH-stigning på 1.

4.4 Oxidationsforsøg

Ved de endelige oxidationsforsøg var det planen at følge pH, redox og permanganat koncentrationen samt eventuelt phosphat, ammoniak og nitrat. De fuldstændige resultater fremgår af bilag 2. I fig 4.2 er udviklingen i pH, redox og permanganat afbildet.

Fig 4.3: Forløb af pH, redoxpotential og permanganat koncentration som funktion af reaktionstiden i timer.

Det viste sig, at der var store måletekniske problemer med at bestemme ammoniak og phosphat i de reagerede prøver efter fjernelse af overskydende permanganat.

Ved ammoniakmålingen skal der tilsættes to reagenser - et reagens med konditioneringskemikalier og et andet reagens med farvestof. Det viste sig, at der allerede kom farve i prøven, når konditioneringskemikalierne blev sat til. Derfor måtte dette farvebidrag modregnes i det endelige resultat. Da man samtidig skulle fortynde prøven 20-50 gange blev resultatet alt for højt og usikkert.

Ved phosphatmålingen blev der målt noget mindre koncentration end på udgangsprøven. Det skyldes formentlig, at phosphat fældes som manganophosphat eller sammen med brunsten. Det var ikke umiddelbart muligt at bringe bundfaldet i opløsning, så den korrekte koncentration kunne bestemmes.

Nitrat kunne laves med samme nøjagtighed som på startprøven, men der blev ikke påvist nogen stigning i nitratindhold. Det var helt som ventet, at permanganat ikke er et stærkt nok oxidationsmiddel til at ilte kvælstof til nitrat.

Da måleresultater for ammoniak og phosphat ikke er pålidelige og anvendelige til at følge nedbrydningsforløbet, er de ikke medtaget i resultatskemaet, bilag 2.

Analyser af pesticider og deres nedbrydningsprodukter fremgår af tabel 4.6. Der er udført 2 parallelle forsøg for reaktionstiderne: 2, 5, 24, 48 og 72 timer, og det er gennemsnittet af de to forsøg, som findes i tabel 4.6. I bilag 4 er alle resultater præsenteret. I tabel 4.6 er endvidere medtaget resultatet af den oprindelige prøve, samt en kontrolprøve, som har henstået i 72 timer uden tilsætning af permanganat.

Tabel 4.6: Cheminova-analyser (mg/l) på prøver fra oxidationsforsøg. De anførte resultater er gen-nemsnit af to parallelle prøver. Resultaterne er korrigeret for fortynding med permanganat og kon-serveringsmiddel, der sammenlagt har givet ca. 10% fortynding. Alle resultater fremgår af bilag 4.

Resultaterne er afbildet grafisk i fig 4.4 til 4.9.

Fig 4.4: Nedbrydning af organiske stoffer som funktion af oxidationstiden.

Fig 4.5: Nedbrydning af pesticider som funktion af oxidationstiden.

Fig 4.6: Nedbrydning af nedbrydningsprodukter som funktion af oxidationstiden.

Fig 4.7: Nedbrydning af råvarer urenheder og oxidationsprodukter heraf som funktion af oxidationstiden.

Fig 4.8: Nedbrydning o dannelse af oxoner som funktion af oxidationstiden.

Fig 4.9: Nedbrydning af sum af organiske stoffer som funktion af oxidationstiden.

Diskussion

Forbruget af permanganat ligger på 300-550 mg/l med et relativt stort forbrug i starten. Det er målt nogle uforklarlige svingninger undervejs - formentlig på grund af den unøjagtige målemetode.

Der er en løbende stigning i pH-værdien under forsøget, idet pH starter ved 3,4 og slutter ved 4,4 efter 72 timer. Denne pH-stigning kan skyldes generering af hydroxylioner ved oxidation af permanganat til brunsten, men det kan også skyldes dannelsen af mangano ioner, da denne proces forbruger brintioner i vandet. pH-stigningen er under alle omstændigheder en naturlig følge af oxidationsprocessen.

Redoxpotentialet falder tilsyneladende lidt under forsøget, hvilket er en kombination af, at permanganat forbruges og pH falder. Vi har undersøgt (afsnit 4.3), at en stigning på 1 pH-enhed medfører et fald i redoxpotential på 72 mV.

Alle pesticider oxideres med permanganat og forsvinder fuldstædigt. Bort set fra MCPA sker nedbrydningen inden for de første to timer. MCPA forsvinder først helt efter 3 døgn, men er allerede halveret efter 2 timer.

Pesticiderne bliver dog ikke oxideret fuldstændig til vand og kuldioxid. Det konstateres nemlig, at der ved oxidationen opstår betydelige mængder oxoner. Paraxon er nedbrydningsprodukt af parathion, methyl-paraoxon er nedbrydningsprodukt af methyl-parathion og malaoxon er nedbrydningsprodukt af malathion. 20 til 35 % af de de parathioner omdannes til de respektive oxoner, og koncentrationen af oxoner holder sig næsten konstant - med tendens til en svag stigning - under hele oxidationsforløbet på 72 timer. Det kan dog ikke helt udelukkes, at oxoner vil kunne nedbrydes ved en meget længere oxidationstid, men det er ikke blevet undersøgt.

Også nedbrydningsprodukterne - hvoraf MP2-syre, EP-2 syre og PNF er de dominerende - nedbrydes stor set fuldstændigt ved oxidationen inden for de første to timer. Der findes dog en rest på 0,3 mg/l af PNF selv efter 72 timers oxidation.

Urenheder fra de anvendte råvarer oxideres også fuldstændigt inden for de første to timer, men der opstår nogle oxidationsprodukter heraf, hvor et dobbelt bundet svovlatom erstattes med et iltatom. EEM-OOOSPS bliver til EEM-OOOSPO, der dog efterhånden også nedbrydes. E-OOSPS bliver til E-OOSPO, hvor det ser ud til at denne omdannelse næsten er fuldstændig efter 72 timer. Man kunne dog forestille sig, at den oxiderede form også forsvandt ligesom EEM-OOSPO, hvis oxidationstiden var noget længere end 72 timer.

Sammenligner vi Cheminovas analyser af den oprindelige prøve med en prøve, der har henstået i en lukket flaske i mørke ved 24 ^oC uden kemikalietilsætning, så er omdannelse i denne reference prøve meget begrænset. Der måles en samlet reduktion af de organiske stoffer fra 76,2 til 73,3 mg/l svarende til et reduktion på 2,9 %.

Ser vi på det samlede restindhold af de analyserede organiske stoffer hos Cheminova, så skyldes det hovedsagelig oxoner og E-OOSPO. Da oxonerne er lige så giftige som parathion, står vi således med et relativt stort indhold af giftige oxoner, hvilket betyder, at permanganat ikke vil være en tilfredsstillende rensemetode.

4.5 Nøjagtige forsøg med permanganat forbrug

Da de første forsøg gav en meget unøjagtig bestemmelse af permanganat forbruget, blev der gennemført endnu et forsøg, hvor forbruget af permanganat blev målt på en mere nøjagtig måde (metode 2).

Der blev foretaget to parallelle forsøg med ca. 120 ml vandprøve og 10 ml permanganat stamopløsning samt en enkelt referenceprøve med 120 ml DI-vand tilsat 10 ml permanganat. Flaskerne blev på den måde fyldt helt op med væske, og der blev sat låg på. Flaskerne blev opbevaret i mørke ved 22-24 ^oC i 7 døgn. De fuldstændige resultater fremgår af bilag 3. Resultaterne fremgår af tabel 4.7, og i fig. 4.10 er resultaterne vist grafisk.

Tabel 4.7: Forbrug i mg/l af MnO₄ som funktion af tiden

ved tilsætning af 10 ml permanganat stamopløsning til

120 ml vandprøve fra høfde 42. 7 døgn ved 22-24 ^oC.

Fig 4.10: Forbrug af permanganat i vandprøve fra Høfde 42 som funktion af tiden. Slutkoncentrationen er korrigeret for forbrug i referenceprøven (nul-prøven).

Det totale forbrug på 640 mg/l permanganat svarer til 108 mg iltforbrug pr. liter, hvis vi forudsætter, at alt permanganat omdannes til Mn⁺². Tilsvarende er iltforbruget kun 65 mg/l, hvis vi forudsætter, at alt permanganat omdannes til brunsten. Da begge processer i virkeligheden forløber, vil iltforbruget formentlig være et eller andet sted mellem 65 og 108 mg/l svarende til ca. halvdelen af vandprøvens COD-koncentration på 170 mg ilt pr. liter. Det kan umiddelbart fortolkes således, at permanganat har oxideret halvdelen af de stoffer, som bliver oxideret ved en COD-bestemmelse, men man kan ikke umiddelbart vide, hvilke stoffer der er blevet nedbrudt og i hvilket omfang.

Diskussion

Den mere nøjagtige permanganat bestemmelse viser, at permanganatforbruget er ca. 585 mg/l (ukorrigeret) over 3 døgn og 640 mg/l (korrigeret) over 7 døgn. Det er betydelig mere, end der blev fundet ved det orienterende forsøg, hvor der blev anvendt en mere unøjagtig metode til måling af permanganat. De to parallelle forsøg følger stort set hinanden.

Permanganatforbruget svarer til en oxidation på ca. 50% af de stoffer, som nedbrydes ved en COD-bestemmelse. Ved en nærmere gennemgang af Cheminovas analyser kan man se, hvilket stoffer der er blevet omdannet eller nedbrudt.

Resultaterne stemmer udmærket overens med, at Cheminovas analyser har vist, at man ved den kemiske oxidation i et vist omfang kun får dannet første trin i nedbrydningen. Vi har således set, at parathioner delvisomdannes til oxoner, hvor blot et dobbeltbundet svovlatom erstattes med et iltatom. Det samme gælder også stoffet E-OOOPS, der omdannes til E-OOOPO. Vi må regne med, at der er flere af disse delvist oxiderede stoffer, som ikke er identificeret og kvantificeret ved Cheminovas analyse.

4.6 Farvefjernelse med aktiv kul

Vandprøven udtaget ved høfde 42 er svag gullig. I styregruppen har man diskuteret, at den farve formentlig skyldes ikke-polære forbindelser, som bør kunne fjernes med aktiv kul. Derfor blev der som et led i forsøget gennemført et orienterende forsøg med rensning af vandprøven med aktiv kul.

200 ml vandprøve blev behandlet med 5 g aktiv kul (AQ 40 fra Chemviron Carbon) i bægerglas under omrøring i 30 min. Herefter blev en vandprøve filtreret gennem et 0,45 µ membranfilter og farven blev målt på Hanna C100 fotometer i PCU (Platinum Cobalt Units). Det ubehandlede spildevand målte efter filtrering i membranfilter (0,45 µ) til 45 PCU. Efter aktiv kulrensning måltes 5 PCU svarende til fjernelse af 90 % af spildevandets farve. Formodningen holdt således stik, og farven kunne stort set fjernes med aktiv kul.

5 Konklusion

Analyserne af startprøven viste, at man ved Cheminovas analyse af 22 organiske forbindelser formentlig havde identificeret ca. 2/3 af det samlede indhold af organiske stoffer. Cheminovas analyser kunne redegøre for 58,4 % af det organiske kulstof, 38,8 % af det organiske kvælstof og 65,5 % af det organiske phosphor i det oprindelige vandprøve fra høfde 42. Den oprindelige prøve var meget sur med en pH-værdi på 3,5 - velegnet til en permanganat oxidation, hvor der overvejende vil dannes mangano ioner.

Ved oxidation med permanganat forbrugte vandprøven 640 mg/l MnO4 over 3 døgn. Det svarer ca. til 50 % af det iltforbrug, som er bestemt ved COD på prøven, idet COD = 170 mg/l. Det viser, at en oxidation med permanganat ikke er en komplet af alle organiske forbindelse til vand og kuldioxid.

Oxidation med permanganat viste en overraskende hurtig nedbrydning af pesticider, nedbrydningsprodukter og råvarer urenheder. De fleste stoffer blev nedbrudt inden for to timer, men enkelte stoffer krævede dog op til 72 timer for komplet nedbrydning.

Desværre blev der ikke i alle tilfælde opnået en fuldstændig nedbrydning af de organiske stoffer. Parathion, methyl-parathion og malathion blev således i et vist omfang (20-35 %) oxideret til de tilsvarende oxoner, der er mindst lige så giftige som parathionerne. Derfor synes metoden ikke umiddelbart velegnet som fuld-skala oprensningsmetode ved høfde 42.

Da permanganat er kendt for sin langtidsvirkning, kan det naturligvis ikke helt udelukkes, at oxonerne kan nedbrydes ved en betydelig længere varende iltningsproces end de 72 timer, der er anvendt ved COWIs undersøgelse. Det skal dog nævnes, at udviklingen i oxonkoncentration de første 72 timer af forsøget ikke tyder på, at koncentrationen falder med tiden - snarere tværtimod.

Ved forsøget blev der ikke genereret nitrat som følge af oxidation af organiske kvælstofforbindelser. Det var ikke muligt at måle, om der blev genereret phosphat på grund af måletekniske problemer, idet phosphat formentlig blev udfældet af mangan eller brunsten og forsvandt fra opløsningen. Vi forsøgte ikke at måle, om der blev genereret sulfat, fordi baggrundskoncentrationen af sulfat i prøven var meget høj, hvilket gjorde det vanskeligt at måle små ændringer præcist.

Bilag 1: Orienterende forsøg vedr. forbrug af permanganat

Bilag 2: Forsøgsresultater, 0-5 timer.

Bilag 2: Forsøgsresultater, 24-72 timer.

Bilag 3: Forsøg med nøjagtigt forbrug af permanganat

Bilag 4: Komplet resultatskema med Cheminova-analyser fra oxidationsforsøgene. Minimumsværdierne er excl. de værdier, som er mindre end.

Klik her for at se Bilag 4

| Til Top | | Forside |

Version 1.0 Oktober 2007 • © Miljøstyrelsen.