hydrogen effects in materials - hysafe · hydrogen effects in materials brian somerday sandia...

Hydrogen Effects in Materials

Brian SomerdaySandia National Laboratories

Livermore, CA, USA

3rd European Summer School on Hydrogen SafetyUniversity of Ulster, Belfast, UK

July 28, 2008

Outline

•Introduction–H2 interactions with metal components– Safety concern: hydrogen embrittlement (HE)

•Methods to assess material performance

•Variables affecting HE– Material– Environmental– Mechanical

Outline

•Considerations for materials selection

•Component design–Design approach–Material property measurement– Example design analysis

•Recommendations

H2 containment components

On-board fuel tanks Manifold components

H2 containment components

PipelinesTransport and stationary tanks

H2 interactions with metal components

•H2 gas

1) H2 gas2) H2 adsorption3) H2 dissociation to H

4) H absorption5) H diffusion

Two principal safety concerns

• H2 permeation through wall of structure– Effectively results in H2 leak

• Hydrogen embrittlement (HE) of structural metal– Crack propagation through wall of structure

– HE mechanisms involve H in solution• Temperatures < 200 oC• Other H-related degradation mechanisms

operate >200 oC

H2H2

H2

H2

H2

H2 H2

stress

stress

defect

component surface

H2H2

H2H2

H2

HH

H HH2 H2 H2H2

H absorbsinto surface

H2

H2

H2H

H

H HH2 H2 H2H2

H concentratesat defect

H

HHHH

H2H2

H2H2 H2H2 H2 H2H2 HH H

H H HH

H

H causes embrittlementand crack propagation

HE enables crack propagation

Hydrogen embrittlement mechanisms

Hydrides

Hydrogen-EnhancedLocalized Plasticity(H.E.L.P.)

Hydrogen-EnhancedDecohesion

H

H

HH

HH

σ

σ σ

σ

H

σ

σ

τ1 τ2

HHHHH

HHH

HH

HH

H

H

H

τ1 > τ2

σ

σ

HHH

H

HH

HH

HH

H HH

σ

σ

void

particle

hydrides

C. Beachem, H. Birnbaum, I. Robertson, P. Sofronis

A. Troiano, R. Oriani

D. Westlake, H. Birnbaum

HE due to hydrogen-enhanced decohesion

“Intergranular” HE can be devastating

Barthélémy, 1st ESSHS, 2006

Material performance: test methods

strength of materials:σu, σy, εf, RAS-N

fracture mechanics:KIH, KTHda/dN vs ΔK

dl/dt > 0

H2H2

H2

H2H2

H2

H

HHH

HH

HH

dl/dt > 0

l

HHH

H H

H H

H

HH

dδ/dt ≥ 0

H

H

H

HH H

H

H

H

H

HH H

H

δ HHHH2 H2H2H2H2

H2

H2H2

dδ/dt ≥ 0


HE manifested by reduced fracture toughness

H2 gas pressure (kpsi)0 5 10 15 20 25

Stre

ss in

tens

ity fa

ctor

, K (

ksi-i

n1/2)

0

50

100

150

200

250

300

H2 gas pressure (MPa)0 25 50 75 100 125 150

Stre

ss in

tens

ity fa

ctor

, K (

MP

a-m

1/2 )

0

50

100

150

200

250

300KJIcKTH

X100 ferritic steelWOL specimensC-L orientation25 oC

HHH

dδ/dt ≥ 0

H2H2H2H2

H2

H2

H2H2

HHH

dδ/dt ≥ 0

H2H2H2H2

H2

H2

H2H2


HE manifested by increased fatigue crack growth rate

Stress intensity factor range, ΔK (ksi⋅in1/2)1 10 100

Fatig

ue c

rack

gro

wth

rate

, da/

dN (

in/c

ycle

)

10-8

10-7

10-6

10-5

10-4

10-3

10-2X42 ferritic steelfrequency = 1 Hz

1000 psi H2, R=0.11000 psi N2, R=0.1

HHH

dδ/dt > 0

H2H2H2H2

H2

H2

H2H2

HHH

dδ/dt > 0

H2H2H2H2

H2

H2

H2H2

Cialone and Holbrook, Met Trans A, 1985

•Fuel tanks– Aluminum or polymer lined composite

– H2 pressure < 70 MPa

•Manifold components– Austenitic stainless steel

– H2 pressure < 138 MPa– Low material strength– No welds

Material performance: service experience

Material performance: service experience

•Storage tanks– Low-alloy ferriticsteel

– H2 pressure < 42 MPa– Limited material strength

– No welds

•Pipelines– C-Mn ferritic steel– H2 pressure < 14 MPa– Low material strength– No pressure cycling

H2 compatibility guidelines for metals

•Materials favored for H2 service– austenitic stainless steels, aluminum alloys, low-alloy ferritic steels, C-Mn ferritic steels, copper alloys

•Materials commonly avoided in H2 service–high-alloy ferritic steels, nickel alloys, titanium alloys

HE susceptibility can be sensitive to many variables

Variables affecting HE

•All metals can be susceptible to HE depending on–Material variables– Environmental variables–Mechanical variables

•Trends demonstrated from laboratory test methods–Emphasize fracture mechanics data– Emphasize ferritic steels, austenitic stainless steels, aluminum

Material variable: strength (hardness)

Yield strength (MPa)600 700 800 900 1000 1100

KTH

(M

Pa-

m1/

2 )

0

20

40

60

80

100

120

Yield strength (ksi)90 105 120 135 150

K TH (

ksi-i

n1/2)

0

20

40

60

80

100Pressure Vessel SteelsH2 gas pressure = 41 MPa4130 steel4145 steel4147 steel

Loginow and Phelps, Corrosion, 1975

•Strength promotes HE in all materials•Technologically important trend

HHH

dδ/dt ≥ 0

H2H2H2H2

H2

H2

H2H2

HHH

dδ/dt ≥ 0

H2H2H2H2

H2

H2

H2H2

Material variable: alloy composition

Alloy composition affects intergranularHE in ferritic steels

[Mn + 0.5Si + S + P] (wt%)0.0 0.2 0.4 0.6 0.8 1.0

KTH

(M

Pa-

m1/

2 )

0

20

40

60

80

100Ni-Cr-Mo ferritic steelσYS = 1450 MPa110 kPa H2 gas296 K

B7Mn=0.007Si=0.002P=0.003S=0.003Mn=0.02

Si=0.01P=0.014S=0.003

Mn=0.09Si=0.01P=0.012S=0.005

Mn=0.02Si=0.27P=0.0036S=0.005

Mn=0.23Si=0.01P=0.009S=0.005

B2Mn=0.68Si=0.08P=0.009S=0.016

Mn=0.72Si=0.01P=0.008S=0.005

B6Mn=0.72Si=0.32P=0.003S=0.005

Mn=0.75Si=0.20P=0.006S=0.004

Bandyopadhyay et al., Met Trans A, 1983

HHH

dδ/dt ≥ 0

H2H2H2H2

H2

H2

H2H2

HHH

dδ/dt ≥ 0

H2H2H2H2

H2

H2

H2H2

Material variable: alloy composition

Nickel affects deformation and/or martensite formation in stainless steels

San Marchi et al., IJHE, 2008

H-precharged

non-charged 316 stainless steels

HH

HH

HH

dl/dt > 0

Material variable: welding

•Weld microstructure can promote HE•Hardness, residual stress also affect HE

δγ

Stainless steel weld Weld tested after H2 exposure

δγ

base metal weld

Environmental variable: H2 pressure

HE more severe at higher H2 pressure for all materials

H2 gas pressure (MPa)0 30 60 90 120 150

K TH (

MP

a-m

1/2 )

0

30

60

90

120

150

180

H2 gas pressure (ksi)0 5 10 15 20

KTH

(ks

i-in1

/2)

0

20

40

60

80

100

120

140

160Pressure Vessel Steels

4130 steel (σYS=634 MPa)4145 steel (σYS=669 MPa)4147 steel (σYS=724 MPa)

Loginow and Phelps, Corrosion, 1975

HHH

dδ/dt ≥ 0

H2H2H2H2

H2

H2

H2H2

HHH

dδ/dt ≥ 0

H2H2H2H2

H2

H2

H2H2

HE can be maximum at ambient temperature

HH

HH

HH

dl/dt > 0

Caskey, Hydrogen Compatibility Handbook for Stainless Steels, 1983

Environmental variable: temperature

Gas impurities can inhibit or intensify HE in ferritic steels

Environmental variable: H2 purity(d

a/dN

) H2+

addi

tive /

(da/

dN) H

2

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

O20.10%

2.25 Cr-1 Mo ferritic steelσYS = 430 MPa1.1 MPa H2 gasΔK = 24 MPa√mfrequency = 5 HzR = 0.1293 K

CO0.99%

SO21.10%

H2O0.03%

CH40.98%

CO21.01%

CH3SH1.04%

H2S0.10%

Fukuyama et al., Pressure Vessel Technology, 1992

HHH

dδ/dt > 0

H2H2H2H2

H2

H2

H2H2

HHH

dδ/dt > 0

H2H2H2H2

H2

H2

H2H2

Water vapor promotes HE in aluminum alloys

Environmental variable: H2 purity

Speidel, Hydrogen Embrittlementand Stress Corrosion Cracking, 1984

HHH

dδ/dt ≥ 0

H2H2H2H2

H2

H2

H2H2

HHH

dδ/dt ≥ 0

H2H2H2H2

H2

H2

H2H2

Mechanical variable: loading rate

HE more severe at lower loading rates in all materials

Loading rate, dK/dt (MPa-m1/2/min)0.1 1 10 100

KIH

(M

Pa-

m1/

2 )

0

20

40

60

80

1004340 ferritic steelσYS = 1235 MPa550 kPa H2 gas297 K

HHH

dδ/dt ≥ 0

H2H2H2H2

H2

H2

H2H2

HHH

dδ/dt ≥ 0

H2H2H2H2

H2

H2

H2H2

Clark and Landes, ASTM STP 610, 1976

HE more severe at lower load cycle frequency in all materials

Mechanical variable: load cycle frequency

ΔK (MPa-m1/2)0 15 30 45 60

da/d

N (μ m

/cyc

le)

0.1

1

10

100SA 105 ferritic steel103 MPa H2 gasload ratio = 0.1

1.0 Hz35 MPa He gas

1.0 Hz

0.1 Hz

0.01 Hz

0.001 Hz

HHH

dδ/dt > 0

H2H2H2H2

H2

H2

H2H2

HHH

dδ/dt > 0

H2H2H2H2

H2

H2

H2H2

Walter and Chandler, Effect of Hydrogen on Behavior of Materials, 1976

Considerations for materials selection

1) HE data or service experience should not be extrapolated


KTH

(M

Pa-

m1/

2 )

0

30

60

90

120

150

180


KTH

(ks

i-in1

/2)

0

20

40

60

80

100

120

140



Yield strength (MPa)600 700 800 900 1000 1100

KTH

(M

Pa-

m1/

2 )

0

20

40

60

80

100

120

Yield strength (ksi)90 105 120 135 150

KTH

(ks

i-in1

/2)

0

20

40

60

80

100Pressure Vessel SteelsH2 gas pressure = 41 MPa4130 steel4145 steel4147 steel

•H2 compatibility must be established for specific conditions


2) HE can be severe in alloys considered compatible with H2

HH

HH

HH

dl/dt > 0

316 stainless steels

•Intersections of variables important


3) Materials selection may balance H2compatibility with other constraints

highcryogenic to lowintermediate

to highintermediatealuminum

low to intermediate

ambient to highlow to highlow

C-Mn/low-alloy steels

highcryogenic to highlow to highhighaustenitic SS

HE resistanceTemperature rangeSpecific strengthCostMaterial


4) Materials must be qualified for H2 service based on– Material property measurements under

relevant conditions• Material variables• Environmental variables• Mechanical variables

– Design approach for component

Design based on fracture mechanics

Local stress drives crack extension

σlocalH2H2

H2

H2

H2

H2 H2

stress (σ)

defect

component surface

H2H2

H2H2

H2H2 H2H2 H2 H2H2

σlocal = σc

stress (σ)

)(2

θπ

σ ijij frK

=

•Local stress described by stress-intensity factor, K

•Crack extension governed by critical K values

Design based on fracture mechanics

Component design: sustained-load cracking

H2 induces time-dependent crack propagation under static loading

pa

pa

H2H2H2

N2N2

N2

time

a

p

crack in H2

crack in N2

Sustained-load cracking proceeds when K > KTH

paH2H2

K = p * f(a/t,Ri,Ro)

RiRo

t

time

a

K

KTH

measured in laboratory

Component design: sustained-load cracking

Crack growth under cyclic loading accelerated by H2

Δpa

Δpa

H2H2H2

N2N2

N2

a

Number of pressure cycles, N (time = N/f)

p

crack in H2

crack in N2

Δp

Component design: fatigue crack growth

Fatigue crack growth rates (da/dN) are a function of ΔK

ΔpaH2H2

ΔK = Δp * f(a/t,Ri,Ro)

RiRo

t

da/dN

ΔK

measured in laboratory H2 gas

measured in laboratory N2 gas

da/dN = C[ΔK]m

Component design: fatigue crack growth

Component design analysis

•Objective: calculate number of pressure cycles, Nc, to grow crack to critical length, ac

Calculation requires material property measurements: KTH and da/dN vs ΔK

a t

ac

Number of pressure cycles, N Nc

critical crack depth forsustained-load cracking cycles tocritical crack depth

K = p * f(a/t,Ri,Ro)

ΔpaH2H2

RiRo

t

a

Measurement: fatigue crack growth

HHH

dδ/dt > 0

H2H2H2H2

H2

H2

H2H2

HHH

dδ/dt > 0

H2H2H2H2

H2

H2

H2H2

Measurement gives crack length vsnumber of cycles in design analysis

Δa = C[Δp * f(a/t,Ro,Ri)]m

Number of pressure cycles, N

a

ΔK (MPa-m1/2)0 15 30 45 60

da/d

N (μ m

/cyc

le)

0.1

1

10

100SA 105 ferritic steel103 MPa H2 gasload ratio = 0.1

1.0 Hz35 MPa He gas

1.0 Hz

0.1 Hz

0.01 Hz

0.001 Hz

Measurement: fatigue crack growth

da/dN = C[ΔK]m

Measurement: cracking threshold

Load

(P)

Time in H2

Po ∝ Ko

Loading bolt

Load cell

13 mm

PTH ∝ KTH

Wedge Opening Load (WOL) specimen

Measurement gives critical crack depth (ac) in design analysis

ac/t = g(KTH,p,Ro,Ri)ac

Number of pressure cycles, N Nc

a


KTH

(M

Pa-

m1/

2 )

0

30

60

90

120

150

180


K TH (

ksi-i

n1/2)

0

20

40

60

80

100

120

140



Measurement: cracking threshold

Example design analysis for H2 pipeline

•Parameters for H2 pipeline(ASME, Hydrogen Standardization Interim Report, 2005)

– X42 ferritic steel– Inner radius, Ri = 15 cm– Wall thickness, t, from 0.5 to 1.3 cm

•Maximum pressure, p = 10 MPa

•Existing defect with depth aoand length parallel to pipe axis

p

RiRo

aot

• Case 1: t=0.8 cm (σh=65% SMYS) and ao/t=0.10



p

10 MPa

N

1 MPa

cycle 1ΔK1=f(ao,Δp,Ro,Ri,t)Δa=(2.51x10-12)ΔK16.56a1=ao+Δa

cycle 2ΔK2=f(a1,Δp,Ro,Ri,t)Δa=(2.51x10-12)ΔK26.56a2=a1+Δa

ao a1 a2


p

RiRo

aot

σh

Material/component performance can be quantified with design analysis

number of pressure cycles0 50000 100000 150000 200000 250000

crac

k de

pth

/ wal

l thi

ckne

ss, a

/t

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

hydrogent=1.3 cmσh=43% SMYS

nitrogent=1.3 cmσh=43% SMYS

X42 Pipelinepressure cycle = 1 - 10 MPainner diameter = 30 cm

X

X

critical crack depthcalculated from KIc

critical crack depthcalculated from KTH


Fatig

ue c

rack

gro

wth

rate

, da/

dN (

in/c

ycle

)

10-8

10-7

10-6

10-5

10-4

10-3


da/dN=(3.55x10-11)ΔK3.83

da/dN=(2.51x10-12)ΔK6.56

1000 psi H2, R=0.11000 psi N2, R=0.1


Performance depends on both material properties and component design

number of pressure cycles0 5000 10000 15000 20000 25000 30000 35000 40000 45000

crac

k de

pth

/ wal

l thi

ckne

ss, a

/t

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

case 1t=0.8 cmσh=65% SMYS



X42 PipelineH2 gaspressure cycle = 1 - 10 MPainner diameter = 30 cm

X

X X

critical crack depthscalculated from KTH


Fatig

ue c

rack

gro

wth

rate

, da/

dN (

in/c

ycle

)

10-8

10-7

10-6

10-5

10-4

10-3


da/dN=(2.51x10-12)ΔK6.56

1000 psi H2, R=0.1

case 1 initial ΔK

case 2 initial ΔK

case 3 initial ΔK


1) Austenitic stainless steels, aluminum alloys best candidates for H2 service– Based on service experience and testing– Do not extrapolate performance or test data

• Define temperature, alloy composition for austenitic stainless steels

• Define water vapor content of gas for aluminum alloys

– Design analysis can quantify safety margins

Recommendations

2) Ferritic steels are susceptible to HE under wide range of conditions– Service experience may be adequate but

design analysis required in many cases– Measure properties under specific conditions

• Material: strength, alloy composition, fabrication (e.g., welding)

• Environment: H2 pressure, H2 composition, temperature

Recommendations

3) Nickel alloys, titanium alloys are generally not recommended for H2 service– Design analysis required– May require other measures such as H2

permeation barrier

Recommendations

hydrogen effects in materials - hysafe · hydrogen effects in materials brian somerday sandia...

Documents