Corrosion behaviour of zinc and aluminium in simulated ......FI9700043 STUK-YTO-TR 123 FEBRUARY 1997 Corrosion behaviour of zinc and aluminium in simulated nuclear accident environments

FI9700043STUK-YTO-TR 123FEBRUARY 1997

Corrosion behaviour of zincand aluminium in simulatednuclear accident environments

J. Piippo, T. Laitinen, P. SirkiaVTT Manufacturing Technology

In the Finnish Centre for Radiation and Nuclear Safetythe study was supervised byTimo Karjunen

This study was conducted by order ofthe Finnish Centre for Radiation and Nuclear Safety

The conclusions presented in the report are those of the authorsand do not represent the official position of the Finnish Centrefor Radiation and Nuclear Safety.

FINNISH CENTRE FOR RADIATION AND NUCLEAR SAFETYP.O.BOX 14, FIN-00881 HELSINKI, FINLANDTel. +358-9-759881Fax +358-9-75988382

ISBN 951-712-177-6ISSN 0785-9325

Oy Edita AbHelsinki 1997

FINNISH CENTRE FOR RADIATIONSTUK-YTO-TR 123 AND NUCLEAR SAFETY

PIIPPO, Juha, LAITINEN, Timo, SIRKIA, Pekka (VTT Manufacturing Technology). Corrosionbehaviour of zinc and aluminium in simulated nuclear accident environments. STUK-YTO-TR 123.Helsinki 1997. 25 pp.+ Appendices 5 pp.

ISBN 951-712-177-6ISSN 0785-9325

Keywords: zinc, aluminium, corrosion, hydrogen, accidents, particulates, sludge

ABSTRACT

Zinc and aluminium are used as anodic coatings and isolation materials in nuclear power plants. Athigh temperature environment they are supposed to oxidise and produce large amounts of hydrogenwhich is a danger for the power plant safety. The solid corrosion products may, together with insu-lation debris created as a result of a pipe break, also clog the pump suction strainers.

The solubility of zinc and aluminium and the stability of the corrosion products were estimatedusing thermodynamical calculations. The corrosion rates of zinc and aluminium were determined insimulated large pipe break and in simulated severe accident cases. An in situ on line measurementtechnique, which is based on the resistance measurement of sample wires, was used.

In the large pipe break case the corrosion rates of zinc and aluminium were determined at pH 8 andpH 10 in deaerated and in aerated solutions. Tests were also performed in aerated 0.1 M boratebuffer solution at pH 9.2. Temperature range was 130°C.. .50°C. The corrosion of zinc appears to berelatively fast in neutral or mildly alkaline aerated water, while both high pH and deaeration tend toreduce the corrosion rates of zinc. The aeration and pH elevation decrease the corrosion rate ofaluminium. Borate content increased especially the corrosion rate of aluminium.

The simulation of the severe accident case took place in the pH range 3-11 in chloride containingsolutions at 50°C temperature. The corrosion rate of aluminium was lower than that of zinc, exceptfor the solution with pH 11, in which the corrosion rate of aluminium was practically identical tothat of zinc. Both metals corroded more rapidly in the presence of chlorides in acidic and alkalicconditions than in the absence of chlorides at neutral environment. The behaviour of zinc and alu-minium was also monitored in high temperature water at 170°C and steam at 300°C.

The experimental results showed that the corrosion rate of zinc decreases with increasing pH, whichis in agreement with the decreasing solubility of zinc with increasing pH observed in thermodyna-mical calculations. The solubility of zinc decreased as the temperature was increased. According tothe calculations ZnO is the stable form of the corrosion products of zinc, which was in agreementwith the experimental results.

FINNISH CENTRE FOR RADIATIONAND NUCLEAR SAFETY STUK-YTO-TR 123

PIIPPO, Juha, LAITINEN, Timo, SIRKIÄ, Pekka (VTT Valmistustekniikka). Sinkin ja alumiininkorroosio ydinvoimalaitosonnettomuutta simuloivissa olosuhteissa. STUK-YTO-TR 123.Helsinki 1997. 25 s.+ liitteet 5 s.

ISBN 951-712-177-6ISSN 0785-9325

Avainsanat: sinkki, alumiini, korroosio, vety, onnettomuudet

TIIVISTELMÄ

Sinkkiä ja alumiinia käytetään pinnoite- ja eristemateriaaleina ydinvoimalaitoksissa. Mahdollisenonnettomuuden yhteydessä niiden oletetaan hapettuvan korkeassa lämpötilassa ja tuottavan suuriamääriä vetyä, mikä voi vaarantaa ydinvoimalaitoksen turvallisuuden. Hapettumisen yhteydessä syn-tyvät kiinteät korroosiotuotteet sekä putken rikkoutumisen yhteydessä vapautuneet eristeriekaleetvoivat myös tukkia kiertovesipumppujen imusiivilöitä.

Sinkin ja alumiin liukoisuutta ja niiden korroosiotuotteiden stabiilisuutta arvioitiin termodynaamis-ten laskelmien avulla. Sinkin ja alumiinin korroosionopeudet määritettiin ydinvoimalaitosonnetto-muuksia simuloivissa ympäristöissä. Mittauksissa käytettiin in situ menetelmää, joka perustuu koe-kappaleina käytettyjen näytelankojen sähkövastuksen mittaukseen.

Ydinvoimalaitoksen höyryputkirikkoa kuvaavaa onnettomuutta simuloitiin hapettomissa ja hapelli-sissa liuoksissa pH arvoilla 8 ja 10. Lisäksi kokeita suoritettiin boraattipuskuriliuoksessa, jonka pHoli9.2. Sinkin ja alumiinin korroosionopeudet määritettiin lämpötiloissa 130°C...50°C. Sinkin kor-roosio on suhteellisen nopeaa neutraaleissa ja lievästi emäksisissä ympäristöissä. pH:n nosto jahapen poisto liuoksessa pienentivät sinkin korroosionopeutta. Hapen poisto liuoksesta ja pH:n nos-to lisäsivät alumiinin korroosionopeutta. Alumiini liukeni nopeasti boraattipuskuriliuoksessa.

Vakavaa onnettomuutta simuloitiin pH arvoilla 3-11 50°C lämpötilassa siten, että happamissa jaemäksissä olosuhteissa liuos sisälsi klorideja. Alumiinin korroosionopeus oli pienempi kuin sinkin,paitsi pH:ssa 11, jossa niiden korroosionopeudet olivat samat. Molemmat metallit syöpyivät nopeam-min klorideja sisältävissä olosuhteissa happamissa ja emäksisissä ympäristöissä kuin kloriditto-massa neutraalissa ympäristössä. Sinkin ja alumiinin korroosionopeudet määritettiin myös 170°Clämpöisessä vedessä ja 300°C lämpöisessä höyryssä.

Tasapainolaskelmien mukaan pH:n nosto pienentää sinkin liukoisuutta, minkä myös kokeet vahvis-tivat. Laskennallisten arvioiden mukaan sinkin liukoisuus pienenee lämpötilan kohotessa. Kokeetja termodynaamiset laskelmat osoittivat, että sinkkioksidi (ZnO) on stabiili kiinteä korroosiotuote.

FINNISH CENTRE FOR RADIATIONSTUK-YTO-TR 123 AND NUCLEAR SAFETY

CONTENTS

ABSTRACT Page

TIIVISTELMA 4

NOMENCLATURE 6

1 INTRODUCTION 7

2 EQUILIBRIUM SOLUBILITY OF ALUMINIUM AND ZINC 92.1 Background 9

3 TEST MATERIALS, EQUIPMENT AND EXPERIMENTAL PROCEDURES 12

4 TEST CONDITIONS AND RESULTS 154.1 Large pipe break case 15

4.1.1 Analysis of corrosion products 164.1.2 Corrosion of zinc 164.1.3 Corrosion of aluminium 18

4.2 Severe accident case 184.2.1 High temperature tests 204.2.2 Effect of chloride 20

5 DISCUSSION 225.1 Experimental observations 22

5.1.1 Large piper break case 225.1.2 Severe accident case 23

5.2 Experimental results vs. thermodynamical calculations 23

6 CONCLUSIONS 24

REFERENCES 25

APPENDIX 1 26APPENDIX 2 29APPENDIX 3 30


NOMENCLATURE

/ = wire lengthm = massr = wire radiusR = electrical resistancet = timepr = specific resistance of a materialpd = density of a material

STUK-YTO-TR 123FINNISH CENTRE FOR RADIATION

AND NUCLEAR SAFETY

1 INTRODUCTION

The containment buildings both in pressurisedwater reactors (PWR's) and boiling water reac-tors (BWR's) house large quantities of zinc andaluminium. Zinc can be found in zinc-basedpaints and galvanised surfaces in steel liners,cable trays, conduits, walkways, gratings, insu-lation covers, and various supports, while alu-minium is used in fans, blades, hubs, and valves,for example. The zinc and aluminium massespresent in the containment may vary greatly de-pending on the plant design, but values as highas 10-20 tons of zinc and 1-10 tons of alumini-um have been given for certain plants.

During normal operation, when containment at-mosphere is dry and the temperatures are mod-erate, corrosion of zinc and aluminium is veryslow. However, in accident situations the condi-tions can be very different. For example, in thecase of a large pipe break nearly all surfaces inthe containment may be wetted with hot water.Provided that the surface area prone to corrosionis large, considerable quantities of both gaseousand solid corrosion products can be formed. Therelease of a gaseous product, namely hydrogen,may increase the containment pressure and causefires. Slow pressurisation can be important espe-cially in the inerted containments, where verylittle oxygen is available for recombination. Thesolid corrosion products in turn may affect boththe functioning of the systems for cooling waterrecirculation and cleaning, and the behaviour ofthe fission products dissolved in the cooling wa-ter.

Corrosion of zinc and aluminium was studiedquite intensively in the 1970's and the early1980's [1-8] The studies were limited to the as-sessment of the hydrogen production potential

in a case of a large pipe break, which formed thedesign basis for the reactor containments. Basedon these studies, it has been concluded that inaccidents which do not result in extensive fueloverheating, hydrogen is predominantly pro-duced by corrosion of zinc-based paints, galva-nised materials and aluminium, together withwater radiolysis.

A major shortcoming in these studies is that lit-tle or no attention was paid to the formation ofsolid corrosion products. Consequently, they didnot provide any data that could be used whenassessing for example the pump suction strainerperformance during a large pipe break. This is-sue was reopened after the Barseback incident in1992, in which inadvertent relief valve blowdown directly to the upper drywell of the con-tainment damaged pipe insulation and causedclogging of the core and containment spray suc-tion strainers. Since then a lot of attention hasbeen paid to estimate the nature and amount ofdebris present in the containment water poolsduring an accident. While detailed analysis hasbeen made concerning the corrosion productspresent in the pools during normal operation, nosuch assessment has been possible for the corro-sion products produced during an accident.

Since also severe accidents involving considera-ble core damage have been brought under inves-tigation, the data base concerning corrosion ofzinc and aluminium has become insufficient evenfor the estimation of hydrogen production poten-tial alone. During severe accidents the contain-ment may be pressurised allowing water temper-ature to be higher than in the design basis acci-dents. If also the containment cooling is lost, thecontainment materials may contact superheated


steam flowing directly from the primary circuit. This report describes a set of experiments thatThis high temperature range was not covered in were performed in order to obtain quantitativethe early experiments. data on corrosion rates and solid products of zinc

and aluminium in conditions typical for both de-Corrosion during a severe accident may be en- sign basis and severe accident conditions. Thehanced also due to the release of chloride, which experiments were performed by VTT Manufac-is contained in the cable insulation material dur- turing Technology on the contract of the Finnishing normal operation. Yet the effects of chloride Centre for Radiation and Nuclear Safety (STUK).on the corrosion of zinc and aluminium have notbeen studied.


AND NUCLEAR SAFETY

2 EQUILIBRIUM SOLUBILITY OFALUMINIUM AND ZINC

2.1 Background

Both soluble species and solid corrosion prod-ucts can be formed when aluminium and zinc areoxidised in an aqueous environment. The natureof the corrosion products depends on the prevail-ing conditions. If the physical and chemical con-ditions (e.g. pH) vary in the vessel, release ofsoluble species to the aqueous environment at onelocation may lead to precipitation of solid prod-ucts in the other parts of the vessel. The startingpoint for the understanding of the dissolution andprecipitation behaviour of aluminium and zincand their relevant oxidation products is based onthermodynamic equilibrium calculations.

Niyogi et al. [5] reviewed solubility product datafor compounds assumed to control the solubilityof aluminium and zinc in the temperature range20.. .30 °C. They used the concept of aluminiumhydroxides to refer to such compounds asA12O3-3H2O (A1(OH)3) and A12O3-H2O(AIO(OH)). Aluminium hydroxide precipitateswere, under certain conditions, suggested to un-dergo a recrystallization process through inter-mediate hydration states to more stable and lesssoluble forms. For instance, aluminium hydrox-ide might in acidic conditions initially precipi-tate as A1(OH)3, and then slowly convert toA12O3«3H2O. On the other hand, in alkaline con-ditions aluminium hydroxide might initially pre-cipitate as AIO(OH), and then convert toA12O3»H2O. The solubility data of Al presentedin ref. [5] were based on the assumption thatA1(OH)3 and AIO(OH) determine the solubilityof Al. Zinc was also assumed to precipitate aszinc hydroxide [5], the solubility product ofwhich determines the concentration of total sol-uble Zn in the aqueous solution. In the presenceof aqueous boron, precipitation of Zn as

Zn(BO2)»H2O may also be of significance. Niy-ogi et al. [5] presented no results concerning sol-ubility at higher temperatures.

Fineschi et al. [6] discussed the behaviour of alu-minium in terms of different regions: at low pHvalues dissolution as Al3+, at near neutral pH val-ues passivation due to an aluminium oxide film,and at high pH dissolution as A1O2\ The oxidewas reported to consist of A12O3»H2O at roomtemperature and of A12O3»3H2O at temperaturesabove 75 °C.

It is worth noting that the thermodynamic prop-erties of surface oxides may differ from those ofbulk oxides (as commented also by Fineschi etal. [6]). The thermodynamic data available forsolid oxidation products have most often beendetermined for bulk oxides, which results in someuncertainty in any equilibrium calculations forsurface films. Also, it has to be taken into ac-count that the surface films can only seldom beconsidered to be in an equilibrium state.

2.2 Thermodynamicalcalculations

In this work the equilibrium solubility of alumin-ium and zinc was estimated in different pH val-ues and at different temperatures using Chem-sage 3.0.1 software (GTT Technologies, Germa-ny). The thermodynamic data for solid com-pounds were obtained from the HSC database(Outokumpu Research Oy, Finland) and for theaqueous species from GEM Aqua database (GEMSystems, Finland.).

The total concentration of dissolved Zn (presentas Zn2+(aq), HZnOy(aq) and Zn(OH)2(aq)) result-ing in the precipitation of ZnO or Zn(OH)2 in


three different conditions (25 °C and 3.5 bar, 80 °Cand 3.5 bar, 165 °C and 7 bar) in the pH range 5 -11 are given in Figure 1. ZnO limits the solubil-ity of zinc in the whole pH range. The solubilityof zinc decreases with increasing pH and withincreasing temperature. The behaviour at 25 °Cin Figure 1 is closely similar to the curve pre-sented by Niyogi et al. [5].

The total concentration of dissolved Al (presentas Al3+(aq), A1O2(aq) and Al(OH)4(aq)) result-ing in the precipitation of A12O3 or A1(OH)3 in

three different conditions (25 °C and 3.5 bar, 80 °Cand 3.5 bar, 165 °C and 7 bar) in the pH range 5 -11 are given in Figure 2. A12O3 limits the solubil-ity of aluminium in the whole pH range. At roomtemperature the solubility of aluminium shows aminimum at pH 5.8, while at 80 °C and at 165 °Cthe solubility increases with pH over the wholepH range 5-11. The solubility increases with tem-perature (except for pH < 5). The curve for 25 °Cin Figure 1 is qualitatively similar to the curvepresented by Niyogi et al. [5].

2

0

21 "8j-ioM-12

-14

-16

i i

i i

i

• upper curve: 25 °

middle curve: 80

lower curve: 165

4 6

V

C, 3.5 bar°C, 3.5 bar

°C, 7 bar

8

pH

10

Zn

12

Figure 1. The total solubility of zinc (as Zn2*(aq), HZnO2 (aq) and Zn(OH)2(aq)) in aqueous solutionsin different conditions.

10


AND NUCLEAR SAFETY

'woX,

ion/

mol

trat

(con

cent

2

0

-2

-4

-6

-8

10

- upper curve:

- middle curve

" lower curve:

-

-

. < - • • • • • • '

i i i i

4

165 °C,

:80°C,

25°C,3

i i i

6

7 bar

3.5 bar

.5 bar

1 1 1 I 1

8

pH

Al

*


3 TEST MATERIALS, EQUIPMENT ANDEXPERIMENTAL PROCEDURES

Zinc and aluminium wires, aluminium plates andhot-dip galvanised steel plates were used as testmaterials. The chemical composition of the zincand aluminium wires supplied by Goodfellow Ltdis presented in Table I. The aluminium plates andthe hot-dip galvanised steel plates were takenfrom TVO NPP pipe installations and foot grat-ings, respectively.

Tests were performed in a titanium cladded auto-clave equipped with necessary input and outputpipelines and a temperature control. Tempera-tures varied from 300 °C to 50 °C depending onthe test. Volume of the autoclave was 7 dm3.

Water purified in a Milli-RO 15 water system(Millipore) was used as test solution. Its pH andchloride concentration was adjusted using lithi-um hydroxide (LiOH) and hydrochloric acid(HC1). In one test the pH of the solution was ad-justed using 0.1 M Na2B4O7 solution. All thechemicals were of pro analysis grade. The solu-tions were pumped from a reservoir via a filterinto the autoclave and further cooled in a heatexchanger before returning to the reservoir. Thevolume flow of the solution was regulated to 10ml/min by a pump in all the tests. The total vol-ume of the solution was 20 dm3 in each test.

The environment inside the autoclave was pos-sible to adjust oxidising or non-oxidising by bub-bling air or nitrogen through the solution, respec-tively. Both gases were decompressed to the de-sired pressure before supplying into the auto-clave. The exact conditions of individual tests aredescribed in the next chapter.

The diameter of the zinc and aluminium wireswas 1 mm. The wires were coiled around Teflonbars in order to eliminate the short circuiting and

Table I. Chemical compositions of zinc and alumin-ium wires. All impurity contents are shown in ppm.

Itemnumber:

Cadmium:

Calcium:

Chromium:

Copper:

Iron:

Lead:

Magnesium:

Manganese:

Silicon:

Silver:

Zinc

ZN005150/1

7

STUK-YTO-TR 123

FINNISH CENTRE FOR RADIATIONAND NUCLEAR SAFETY

A solution sample was taken from autoclave forpH measurement and for cation concentrationanalysis, if necessary. Cation concentration anal-ysis was performed using atom absorption spec-trometry (AAS). Oxidation products from theautoclave bottom were collected for chemicalanalysis and grain size characterisation. Thechemical analysis of the corrosion products wasperformed using X-ray diffraction (XRD) and X-ray fluorescence (XRF) analysis.

The corrosion rates of zinc and aluminium wireswere determined also by measuring on-line theelectrical resistance of the corroding wires dur-ing the tests. Corrosion reduces the diameter ofthe metallic part of a wire as a function of time,which increases its electrical resistance. Themeasurement was performed by feeding a con-stant current of 100 mA through the wires usinga Yokokawa 7651 power source. The voltage dropover the wire was measured using a Keithley 182sensitive digital voltmeter at intervals of five orten minutes. The data from the digital voltmeterwas saved to a microcomputer. The electrical re-sistance was calculated from the voltage drop andthe current using the Ohm's law.

Knowing the electrical resistance (/?), its timederivative (dR/dt), the specific resistance of amaterial (pr) and the wire length (/), it is possibleto determine the wire radius decrease rate dr/dtas a function of time

—\P A. = -600000

The radius change rate is converted to weight lossrate because it is more informative. Conversionwas performed by multiplying corrosion rate bythe density of the material

dm\gdt I m dt

(2)

where pd is the density of material. Note that thetests were not reproduced. Thus no statistical es-timation of the measurement accuracy could bemade.

Figure 3. Installation of wires and aluminium plate.13


The proportion of the oxide remaining on the wire convert to ZnO and A12O3, respectively. The hy-surfaces was estimated by calculating the mass pothetical weight gain due to the oxide growthof the corroded metal using the corrosion rate was calculated. The difference of the real weightdata based on the resistance measurement meth- gain and the calculated weight gain was com-od. All the zinc and aluminium was supposed to pared to the mass of the oxide.

14


AND NUCLEAR SAFETY

4 TEST CONDITIONS AND RESULTS

The performed tests can be divided into two groups according to the type of theaccident they simulate. The first group (denoted with A) describes the materialbehaviour during a large pipe breakage accident (LOCA) and the second group(denoted with B) during a severe accident.

4.1 Large pipe break case

As a large steam or water pipe breaks during aLOCA accident in a reactor containment, tem-perature and pressure are controlled using con-tainment spray systems. At the beginning of theaccident the temperature is high but it decreasesrapidly.

A summary of the test conditions is given in Ta-ble II. At the beginning of tests the temperaturewas 130°C, and it was lowered stepwise to 50°C.

The testing periods took approximately one dayat 130°C and at 110°C, and two days at the othertemperatures. pH values 8 and 10 were used inorder to simulate the effect of LiOH that is add-ed into spray water in BWR's for the iodine chem-istry control. The effect of redox potential on thecorrosion rates was studied by performing thetests in oxygen-free water and in water saturatedwith air. Borate buffer solution (pH 9.2) was usedin one test in order to simulate the accident tak-ing place in a VVER-440 environment.

Table II. Test conditions in large pipe break experiments.

Test Nr.

A1

A2

A3

A4

A5

T/°C

1301109070501301109070501301109070501301109070507090110130130110907050

pH/beginning

8

8

10

10

9.2

pH/end

6.5

7.1

9.4

9.4

9.3

Environment

H2O + LiOH

H2O + LiOH

H2O + LiOH

H2O + LiOH

H2O + Na2B4O7

Time

6h36 h2d2d2d1 d1 d3d2d2d1 d1 d2d2d3d1 d1 d2d2d2d1 d1 d1 d1 d1 d1 d2d2d2d

Note

N2-pressurisation500 dm7h

air-pressurisation500 dmVh

N2-pressurisation

air-pressurisation500 dm7h

air-pressurisation

15


Table III. The concentrations of dissolved Zn, Fe and Al in the process solutions after the tests.

Test Nr.

A1A2A3A4A5

Environment

pH8

pH10

borate, pH 9

Gas

N,air

Nsairair

Cation concentration

Zn / mg/l21.53.42.71.8

64.5

Fe / [ig/\30656

690

Al / ug/l—

Table IV. Description of the cormsion products of zinc formed on zinc wires and on galvanised steel plates together with the weight loss data of the wires andplates in the large pipe break experiments.

Visualinspection

Change ofweight (withoxide layer) [%]

Change ofweight(Washed) [%]

Change ofweight (withoutoxide layer) [%]

Proportion ofZnO remainingon surface [%]

RunA1

Zinc wire

Dark colouredoxide layer onthe wire.

0.14

-0.06

90

galvanisedsteel plates

Slightcorrosionvisible on thesamples.

0.09

RunA2

Zinc wire

Light browncolouredpowderyoxide layer onthe wire.

-0.18

-3.26

-4.60

80


Plentiful ofpowderycorrosionproducts onthe samples.Coloured lightbrown

0.24

Run A3

Zinc wire

Wire coveredby darkcolouredoxide layer

-O.02

-0.04

-0.43

80


Somepowderycorrosionproductsvisible on thesamples.

0.13

Run A4

Zinc wire

Powderyoxide layer onthe wire, notformedregularly

-0.26

-0.37

-1.08

70


Powderycorrosionproducts onthe samples.Coloured lightbrown

0.18

RunA5

Zinc wire

Wire coveredby dark oxidelayer

0.88

0.43

-1.28

90


Oxide layernot visible.Some darkspots.Surfacecoarse.

0.11

00

H

H

9JO

2

> 5JO >

H O


4.1.3 Corrosion of aluminium

Figure 5 shows the corrosion rate of aluminiumbased on the measurement of the electric resist-ance, while the numerical values are given inAppendix 2.

The proportion of the oxide remaining on the alu-minium wire surface, the results based on themeasurement of the weight loss of aluminiumplates together with a description of the oxida-tion products on aluminium wires are given inTable V. At pH 8 and 10 major part of the oxideremains on the surface. The pH elevation increas-es the proportion of oxides remaining on surfaceby about 20 %-units. In the borate containingsolution only about 20 % of the oxide stays onthe aluminium surface.

4.2 Severe accident case

If the spraying systems do not work during aLOCA the temperature in the containment re-mains high for a long time. The high tempera-ture affects the corrosion rates of structural ma-terials in the containment and it can also increasethe decomposition rate of the cable insulationmaterials. The decomposition may cause the re-lease of gases containing chloride that affects thecorrosion of structural materials in the contain-ment.

A summary of the test conditions is given in Ta-ble VI. Experiments were performed using zincand aluminium wires as test materials.

10

E?

1 r

0.1 r

0.01 r

0.001

—•— pH 8, nitrogen—C— pH 8, air—*— pH 10, nitrogen-£— pHIO, air—O— borate, air

40 50 60 70 80 90 100 110 120 130 140

Temperature / °C

Figure 5. Corrosion rates (converted to g/m2h)for aluminium as a function of temperature in different pHand gas environments determined by measuring the resistance increase of the corroding wires.

18

Table V. Description of the corrosion products of aluminium wires and plates together with the weight loss data of the wires and plates in the large pipe breakexperiments.

Visualinspection

Change ofweight (withoxide layer) [%]

Change ofweight (washed)

[%]Change ofweight (withoutoxide layer) [%]

Proportion ofAI2O3remaining onsurface [%]

RunA1

Aluminiumwire

Wire coveredby thin oxidelayer.

0.59

0.10

60

Aluminiumplate

Plate coveredby thin oxidelayer.

0.50

0.28

Run A3

Aluminiumwire

Wire coveredby thin oxidelayer.

0.98

0.73

0.60

60

Aluminiumplate

Plate coveredby thin oxidelayer.

1.69

1.34

0.34

RunA5

Aluminiumwire

Very thinoxide layer onthe wire

2.22

2.18

2.11

80

Aluminiumplate

Brightnessfaded, thinoxide layer

0.45

0.43

0.41

Aluminiumwire

Very thinoxide layer onthe wire

1.63

1.57

1.53

80

Aluminiumplate

Light oxidelayer withwhitecorrosionspots onlower side,upper sidedarkened

0.66

0.43

0.38

Aluminiumwire

Light oxidelayer of thewire. Surfacewas corroded

-11.51

-13.31

-14.67

20

Aluminiumplate

Surface wascorroded

-10.71

-12.19

-13.38

C/3Ha

9JO

to

2zXnwz

at/3 O

H O«J Z


Table VI. Test conditions in experiments where a severe accident was simulated. The pH values aremeasured at the beginning of experiments.

Test Nr.

Run B1

Run B2

Run B3

RunB4

Run B5

Run B6

Run B7

T/°C300

170

50

50

50

50

50

PH—

-6.5

-7

5.07

3.0

9.0

11.0

Environment

Superheated steam

H2O + [OJ-200 ppb

H2O + [O2]~8 ppm at start

H2O + 1E-5 mol/l HCI

H2O + 1E-3 mol/l HCI

H2O + 5E-4 mol/l HCI +5.6E-4 mol/l LiOHH2O + 5E-4 mol/l HCI +7E-3 mol/l LiOH

Time

1d

4d

4d

4d

3d

5d

5d

Observe

water flow 0.2 dm7min

N2-overpressure

N2-overpressure

N2-overpressure

N2-overpressure

4.2.1 High temperature tests

These measurements simulated the early timesequence in an accident as the hot steam (300 °C)is discharging from a broken cooling water pipe(Run B1) and the later period as the water hascooled to 170°C(RunB2).

Pure water deaerated with nitrogen was used astest solution. The water was pumped on to theautoclave bottom where it was evaporated andsuperheated at the test temperature of 300 °C. Inthe experiments performed at 170°C the dissolvedoxygen content of the input water was control-led to be 200 ppm± 20 ppm. The results are sum-marised in Table VII.

The average grain size of the oxide product de-posit formed at 170 °C was ca. 30 u,m.

4.2.2 Effect of chloride

The effects of chloride concentration and pH werestudied by making tests at uniform temperatureof 50 °C. The pH was varied from 3 to 11 and thechloride concentration was varied from 1E-5mol/l to 1E-3 mol/l. The results based on themeasurement of the resistance of the zinc andaluminium wires are presented in Figure 6, whilethe corresponding numerical values are given inAppendix 3.

Table VII. Description of the corrosion products together with the weight loss data for zinc and alumin-ium in the severe accident experiments.

Visualinspection

Change ofweight [%]

Weight lossrate [g/(m2h)]

Run B1 (300 C)Zinc wire

The lowestpart coveredby a strongoxide. Rest ofwire coveredby light oxidelayer.

0.099

3.00

Aluminiumwire

No visualchangesobserved.

-0.001

0.05

Run B2 (170 °C)Zinc wire

Abundantcorrosionproducts inautoclave.Wire brittle,covered withblack deposit.Wire extendedapproximately5 m.

-51.35

11.27

Aluminiumwire

Thick looselayer of oxide.Wire extendedapproximately1 m.

^7 .50

7.15

Galvanisediron wire

Generalcorrosion,grey coloured.

20


AND NUCLEAR SAFETY

Table VIII. The relative weight loss data of zinc and aluminium wires.

Run Nr.

B5B4B3B6B7

pH3

5

7

9

11

[HCI] / mol/l

1.0-10"3

1.0-10"5

—

5.0-10"4

5.0-10^

[LiOH] / mol/l

———

5.6-10^7.0-10^

Zn / m-mjmo 1 %-0.37-0.18-0.13-0.33-0.23

Al / m-m^m,, / %0.570.840.931.840.35

The corrosion rate of aluminium is lower thanthat of zinc, except for the solution with pH 11,in which the corrosion rate of aluminium is prac-tically identical to that of zinc. Both metals seemto corrode more rapidly in the presence of chlo-rides in neutral and acidic conditions than in theabsence of chlorides at neutral pH.

During the tests the dimensions of the wires didnot change. At pH 3 zinc was covered by a darkoxide layer and the aluminium by a thin lightcoloured layer. At pH 5 both the materials hadthin oxide layers. In neutral solution zinc had alight coloured brittle oxide layer and aluminiuma very thin oxide layer. At pH 9 the zinc wire was

covered by a light powdery layer. Corrosion prod-ucts were observed on the bottom of the auto-clave. No signs of oxide formation could be de-tected on aluminium. At pH 11 aluminium wascovered by a thin oxide layer whereas no chang-es on the zinc wire surface could be observed.

The results of the relative corrosion rate deter-mination based on weight loss measurements ofzinc and aluminium wires are presented in TableVIII. Results show that the corrosion rate of alu-minium wires could not be determined by meansof the weight loss method because their weightincreased due to the oxide formation in all thetests

s:

£

-dm

/dt

0.14

0.12

0.10

0.08

0.06

0.04

0.02

0.00

-

-

-

-

-

o-—

)

\

c

4

\ /

/

6 8

pH

/

/

m

—O-

10

I

- Zn- Al

12

Figure 6. The corrosion rates (converted to g/m2 h) of zinc (closed symbols) and aluminium (open sym-bols) based on the measurement of the resistance of the wires in different environments. The compositionsof the test solutions are presented in Table VI.

21


5 DISCUSSION

5.1 Experimental observations

The corrosion rates were determined usingweighing and a technique based on the measure-ment of electric resistance of wires. The resist-ance measurement was found to be more suita-ble in this study because it is an in situ techniqueand because the corrosion products on the spec-imens did not affect the results.

5.1.1 Large piper break case

The measurements show that the corrosion ratesof zinc and aluminium may vary greatly depend-ing on the water temperature and chemistry. Cor-rosion of zinc appears to be relatively fast in neu-tral or mildly alkaline aerated water, while highpH and deaeration both tend to reduce the corro-sion rates. The use of borated alkaline water in-duces rapid corrosion at high temperatures, whileat temperature range 50°C.. .90°C corrosion rates

are moderate. Contrary to expectations, zinc cor-rosion did not always slow down when tempera-ture decreased. In some cases a change in pHduring the tests may have contributed to this un-expected behaviour. However, the same was ev-idenced also in the test with borated water, inwhich pH remained stable. This behaviour mustbe connected with the decreasing solubility ofZnO with decreasing temperature.

A lot of data on aluminium and zinc corrosion isavailable but only few of these are similar to theenvironments of this test series. The results ob-tained for zinc in borate containing solution canbe compared to the data presented in [2,3] shownin Figure 7. The measurements were performedin aerated 0.1 M borate solution at pH 9 [2] andin solutions containing 2000 ppm and 4000 ppmB at pH 9 [3]. Estimates by the same authors arealso presented.

le+1

le+0 r

•aS le-2 r

le-3 r

le-4

this work, 4000 ppm B, pH 9[2], 2500 ppm B, pH 9[3], 2000 ppm B, pH 10[3], 4000 ppm B, pH 10[2], estimation[3], estimation

20 40 60 80 100 120

Temperature / °C

140 160

Figure 7. A comparison of estimated corrosion rate data [2, 3} for zinc and of zinc measured in 0.1 Mborate buffer solution at pH 9 and of galvanised steel measured by [2, 3].

22


AND NUCLEAR SAFETY

The corrosion rates are in good accordance attemperatures below 100°C, but at higher temper-atures the corrosion rates obtained in this workare higher. The estimated corrosion rates [2] aresomewhat higher at temperature region from70°C to 90°C than the experimentally obtainedcorrosion rates in this work.

5.1.2 Severe accident case

Both zinc and aluminium are more susceptibleto corrosion in the presence of chlorides in neu-tral and acidic conditions than in the absence ofchlorides at neutral pH. The corrosion rate of alu-minium is lower than that of zinc, except for thesolution with pH 11, in which the corrosion rateof aluminium is practically identical to that ofzinc.

Loyola and Womelsduff [3] have measured thehydrogen formation rate on galvanised steel indemineralised water at 48°C temperature. Theyobserved a corrosion rate of 0.088 g/m2h that isin good accordance to 0.079 g/m2h measured forzinc in this study at pH 5 in the presence of chlo-rides. On the other hand, their result 39.4 g/m2hat the temperature of 168°C is clearly higher than11.3 g/m2h measured for zinc at 170°C in thisstudy.

Frid et al. had determined the corrosion rates oftechnically pure aluminium (AA 1050) in deaer-ated water at pH 5 at temperatures 50°C, 100°Cand 150°C [4]. The corrosion rate was independ-ent of temperature and exposure time. The cor-rosion rate of 14.4 mg/m2h is lower than the 30mg/m2h observed in this study. Same authors re-peated the same tests at pH 9 and obtained a cor-rosion rate of 12.6 mg/m2h. As presented in Fig-ure 7 and in Appendix 3 a somewhat higher cor-rosion rate of 51 mg/m2h was obtained in this

study. The higher corrosion rates obtained in thisstudy are probably due to the chlorides.

5.2 Experimental results vs.thermodynamicalcalculations

The results of the analysis of the corrosion prod-ucts, and the experimentally observed corrosionbehaviour of zinc and aluminium agree in sever-al aspects with the results of the thermodynamiccalculations.

Equilibrium calculations indicated that the solidcorrosion product should be in the form of ZnO,which was also detected in XRD analysis. Onthe other hand, the concentration of dissolved zincdecreases with pH in water containing no borate,which also agrees with the theoretical calcula-tions reported in Chapter 1. Unfortunately, thecontent of dissolved aluminium and the amountof solid corrosion products of aluminium weretoo low to be detected with the methods used.Accordingly, no comparison between the theo-retical calculations and the experimental resultscould be made.

The corrosion rate of zinc (determined by meansof the measurement of the electric resistance ofa zinc wire) decreased with increasing pH, whichis in agreement with the decreasing solubility ofzinc with increasing pH. On the other hand, theexperimental dependence of the corrosion rateof aluminium on pH was not so straightforward.The main trend seemed to be that aluminiumcorrodes more rapidly at a lower pH, which isnot in agreement with the solubility calculationsof aluminium. Therefore, it can be suggested thatthe corrosion of aluminium proceeds as a solidstate process in which the role of solubility is notvery significant.

23


6 CONCLUSIONS

The corrosion rates of zinc and aluminium weredetermined in simulated large pipe break and insevere accident cases. An in situ on line meas-urement technique that was based on the resist-ance measurement of the sample wire was used.The technique was well applicable to this workbecause of its in situ measurement possibility andof its high accuracy compared to the conventionalweight loss measurement method.

In the large pipe break case the corrosion ratesof zinc and aluminium were studied at pH 8 andpH 10 in deaerated and in aerated solutions andin 0.1 M borate buffer solution at temperatures130°C...50°C. The corrosion of zinc appears tobe relatively fast in neutral or mildly alkalineaerated water, while high pH and deaeration bothtend to reduce the corrosion rates. The aerationand pH elevation decrease the corrosion rate ofaluminium.

The simulation of the severe accident case tookplace in the pH range from 3 to 11 in chloridecontaining solutions. The corrosion rate of alu-minium was lower than that of zinc, except forthe solution with pH 11, in which the corrosionrate of aluminium was practically identical to thatof zinc. Both metals corroded more rapidly inthe presence of chlorides in acidic and alkalicconditions than in the absence of chlorides atneutral environment.

The solubility of zinc and aluminium and the sta-bility of the corrosion products were estimatedusing thermodynamical calculations. The exper-imental results and the thermodynamical calcu-lations were in fair agreement.

24


AND NUCLEAR SAFETY

REFERENCES

[1] Griess JC, Bacarella AL. The corrosion ofmaterials in reactor containment spraysolutions. Nuclear Technology 10, 1971:546-53.

[2] van Rooyen D. Hydrogen release rates fromcorrosion of zinc and aluminium. BNL-NUREG-24532 Informal report. May 1973:1-37.

[3] Loyola VM, Womelsduff JE. The relativeimportance of temperature, pH and boric acidconcentration on rates of H2 production fromgalvanised steel corrosion. NUREG/CR2812. November, 1983.

[4] FridW, Karlberg G, Sundwall SB. Hydrogengeneration from aluminium corrosion inreactor containment spray solutions.Proceedings of the Second InternationalConference on the Impact of Hydrogen onWater Reactor Safety. Albuquerque, NewMexico, October 3-7, 1982. NUREG/CP-0038. pp. 440-50.

[5] Niyogi KK, Lunt RR, Mackenzie JS.Corrosion of aluminium and zinc in contain-ment following a LOCA and potential forprecipitation of corrosion products in thesump. Proceedings of the Second Inter-national Conference on the Impact ofHydrogen on Water Reactor Safety.Albuquerque, New Mexico, October 3-7,1982. NUREG/CP-0038. pp. 401-423.

[6] Fineschi F, Lanza S and Lombardi G.Hydrogen generation in LOCA conditions:the contribution of aluminium corrosion.Contribution to the Hydrogen Sub-Group,Working Group 2, Comitato NazionaleEnergia Nucleare (Contract AC-3), EECBruxelles, December 1980: 1-58.

[7] Zittel HE. Post-accident hydrogen generationfrom protective coatings in powergenerations. Nuclear technology 19, 1973:143-6.

[8] Lopata JR. Control of containment H2 levelsevolved from zinc primers during a LOCA.Power Engineering / November 1974: 48-51.

25


APPENDIX 1The grain size distributions of the individual tests of the large pipe break case.

6 7 8 10 20 30 40

Particle Diameter (um)

60 60 70 100 200

Figure Al-1. The cumulative particle size distribution of the corrosion products obtained in testAl.

100-

90-

80-

70-

I 17 8 10 20 30 40 50 60 70


100 200

Figure Al-2. The cumulative particle size distribution of the corrosion products obtained in test A2.

26


AND NUCLEAR SAFETY

APPENDIX 1

100-

90-

80-

70-

60-

50-

40-

30-

—T 1 120 30 40


I 1 1—'—'—I50 60 70 100 200


9«

i0

1aQ•s

UU

100-

90-

80-

70-

60-

50-

40-

30-

20-

10-

n •

/

/

/

/

/

/

'

u i i i i i • i

4 5 6 7 8 10

/

20 30 40 50 60 70Particle Diameter (um)

100 200


27


APPENDIX 1

10 201—30

~1 1 1 I—'—"—i40 50 60 70 100 200



28


AND NUCLEAR SAFETY

APPENDIX 2a) Corrosion rates (g/m2h) and wire radius decrease rates (fxm/h) measured for zinc in a large pipe

break case at different temperatures in different aqueous environments determined by measuringthe resistance increase of the corroding wires.

t / °c

130110907050

t / ° C

130110907050

N,0.3170.8530.0880.1470.117

N20.04440.11950.01240.02060.0164

Weight loss ipH8

air4.9571.5570.1630.2520.747

ate of zinc / -dm/dt [g/(m2h)]

N,0.2960.1310.0230.0090.011

Wire radius decrease rate opH8

air0.69430.22090.02290.03530.1047

N20.04150.01840.00320.00120.0015

pH10air

0.3690.3820.0580.0190.017

borateair

1.264.450.040.030.05

F zinc / -dr/dt [um/h]pH10

air0.05170.05350.00810.00270.0024

borateair

0.17610.62310.00560.00420.007

b) Corrosion rates (g/m2h) and wire radius decrease rates (jxm/h)measured for aluminium in a largepipe break case at different temperatures in different aqueous environments determined by meas-uring the resistance increase of the corroding wires.

t / °c

130110907050

t / ° C

130110907050

N,0.4120.2070.1160.0870.080

N20.15240.07670.04290.03210.0298

Weight loss rate of zinc / -dm/dt [g/(m!h)]pH8

air0.1970.6990.0480.0370.004

N,0.7010.2830.0310.0430.012

Wire radius decrease rate oipH8

air0.07310.33280.01780.01370.0014

N,0.25980.10490.01140.01580.0045

pH10air

0.0720.1500.0120.0060.006

borateair

3.061.231.450.600.15

F zinc / -dr/dt [nm/h]pH10

air0.02670.05550.00460.00220.0022

borateair

1.13350.45390.5370.2220.0556

29


APPENDIX 3Corrosion rates (g/m2h) measured for zinc and aluminium in a severe accident case at 50 °C inaqueous environments (different pH and Cl concentrations) determined by measuring the resis-tance increase of the corroding wires.

PH357911

[HCI] / mol/l1.0-103

1.0-105

—5.0-10"5.0-10"

[LiOH] / mol/l———

5.6-10"7.0-103

Zn/-dm/dt /g /m 2h0.1420.1080.0790.1200.110

Al / -dm/dt /g /m a h0.0300.0300.0160.0510.126

30

Documents

Corrosion behaviour of zinc and aluminium in simulated ......FI9700043 STUK-YTO-TR 123 FEBRUARY 1997 Corrosion behaviour of zinc and aluminium in simulated nuclear accident environments