SECA SOFC Seal MeetingJuly 8-9, 2003

Sandia Brazing Research and Modeling Capabilities*

F. Michael HoskingMaterials and Process Sciences Center

Sandia National LaboratoriesAlbuquerque, NM 87185-0889

* Sandia is a multiprogram laboratory operated by Sandia Corporation,a Lockheed Martin Company, for the United States Department of Energy’s

National Nuclear Security Administration under Contract DE-AC04-94AL85000.

Principal SNL/NM Technical Brazing ContactsMetals: Mike Hosking & John Stephens, Org. 1833 & Chuck Walker, Org. 14171

Ceramic: Ron Loehman, Jill Glass & Sandy Monroe, Org. 1843Microanalysis: Joe Michael, Tom Headley, Paul Kotula & Paul Hlava, Org. 1822

Modeling: Steve Burchett, Frank Dempsey, Rick Givler & Jerry Wellman, Org. 9100

SECA SOFC Seal MeetingJuly 8-9, 2003

Brazing is widely used to join metals and ceramics to each other for a variety of high reliability components

New active brazing alloys simplify hermetic joining issues

Basis: Oxide-forming additives to filler metal (e.g., V, Ti, Zr, Hf) promote direct wetting and adhesion to the ceramic materials (Al2O3, Si3N4, cermet, …)

Payoff: Eliminate complex Mo-Mn metallize / Ni-plate & related processing for conventional alloys

Issue: Complex reactions between active braze elements and ceramic/metal base materials need to be understood to control and optimize process

• headers• connectors• feedthroughs• thermal batteries• high voltage tubes• electromechanical devices• storage containers

TEM ImageTEM Image

TiO, Ti2O & Cu3Ti3O reaction products

SECA SOFC Seal MeetingJuly 8-9, 2003

Braze flow visualization and dissolution reactions are important components in understanding the braze process

x distanceto

LabVIEW + IMAQ Vision Analysis

∆x = 0.19 mm

≈ 10 D (1.3 mm)

Laminar Flow

approx. 5 Hz

x distanceto

LabVIEW + IMAQ Vision Analysis

x distanceto

LabVIEW + IMAQ Vision Analysis

∆x = 0.19 mm

≈ 10 D (1.3 mm)

Laminar Flow

approx. 5 Hz

Recent Accomplishments• Conference paper / presentations (IBSC’03;

CSM Materials Science Seminar; C6 ASCI Solidification Working Group)

• S&T of Welding & Joining journal paper• In-Situ Visualization of Braze Flow patent

1095ºC

1145ºC1120ºC

(at% Ni)

1095ºC

1145ºC1120ºC

(at% Ni)

66019010181140 (H2Frn)

321020181145

1.519514121120

0.5175631095

ResidentTime Above

1085ºC

JointThickness

(µm)

EMPA Ni Content(wt. %)

Terminal Solidus

(wt. % Ni)

Specimen Temperature

(ºC)

66019010181140 (H2Frn)

321020181145

1.519514121120

0.5175631095

ResidentTime Above

1085ºC

JointThickness

(µm)

EMPA Ni Content(wt. %)

Terminal Solidus

(wt. % Ni)

Specimen Temperature

(ºC)

Note: At 1145ºC, Cliquidus ~ 10 wt. % Ni

0

20

40

60

80

100

0 1000 2000 3000 4000 5000

Ni

Cu

Con

cent

ratio

n (w

t. %

)

distance (µm)

Cu-Ni Traces Through Center of Cu Braze1120C Peak Work Temperature

1120ºC

EMPA Traces

1120ºC

EMPA Traces

Cu-Ni binary braze flow & dissolution experiments

Braze Capillary Flow Visualization

SECA SOFC Seal MeetingJuly 8-9, 2003

Active brazing technology is bridging a wide range of needs because of the fundamental knowledge being developed

0

20

40

60

80

100

120

Frac

ture

Stre

ngth

(M

Pa)

94% 94% 94%99.8% 99.8% 99.8%

Au - 16.0Ni -1.75V - 0.75Mo

Au - 18.0Ni - 0.8Ti

Au -17.5Ni - 2.5Ti

Alumina ceramic / braze interface reactions vary significantly with active element

Alumina

Au-Ni-Mo-2V

reaction layer

10 nm

XX--ray diffractionray diffraction

(Fe, Ni,Co)3Ti

… no TiAlumina

Kovar (Fe-29Ni-17Co)

Ag-Cu-Ti

Competing reactions between dissimilar base materials can introduce unwanted results

V activation

Ti activation… compatibility

issuesZr activation

Alumina

Au-Ni-Mo-2V

reaction layer

10 nm

XX--ray diffractionray diffraction

(Fe, Ni,Co)3Ti

… no TiAlumina

Kovar (Fe-29Ni-17Co)

Ag-Cu-Ti

Competing reactions between dissimilar base materials can introduce unwanted results

V activation

Ti activation… compatibility

issuesZr activation

Fracture strengths for 94 & 99.8% alumina brazed with Ti & V-bearing Au-Ni filler metals exhibit dependency on glassy phase and more stable TiOx reaction product

Ag-2Zr (wt. %) ABA

reaction layer

Alumina

ZrO2

15 µm

SECA SOFC Seal MeetingJuly 8-9, 2003

A multidisciplinary approach to characterizing the different brazing reactions and properties is necessary for success

0

20

40

60

80

100

120

140Average Tensile Strength (MPa)

BrazingTemperature

& Time

Determining effects of reaction products & microstructures on properties (hermeticity & strength) as a function of active element (thermodynamics / kinetics), base material (adhesion / dissolution), brazing conditions (temperature, time, atmosphere, heating cycle& gradients, orientation), surface preparation (ceramic air-firing).

Contact Angle vs. [SiO2]

0

10

20

30

40

50

60

70

80

0 20 40 60 80 100

[SiO2]

Wet

ting

Ang

le

2.5%Hf3.4%Hf6.8%Hf

1100°C for 30 min in Ar

Contact Angle vs. [SiO2]

0

10

20

30

40

50

60

70

80

0 20 40 60 80 100

[SiO2]

Wet

ting

Ang

le

2.5%Hf3.4%Hf6.8%Hf

Contact Angle vs. [SiO2]

1100°C for 30 min in Ar

0

10

20

30

40

50

60

70

80

0 20 40 60 80 100

[SiO2]

Wet

ting

Ang

le

2.5%Hf3.4%Hf6.8%Hf

1100°C for 30 min in Ar

200 µm

96% Al2O3

Ag-2.5 wt.% Hf

200 µm96% Al2O3

200 µm

96% Al2O3

200 µm

96% Al2O3

Ag-2.5 wt.% Hf

200 µm96% Al2O3

Ag-2.5 wt.% Hf

200 µm96% Al2O3

200 µm

96% Al2O3

Ag-2.5 wt.% Hf

200 µm96% Al2O3

200 µm

96% Al2O3

200 µm

96% Al2O3200 µm96% Al2O3

Ag-2.5 wt.% Hf

200 µm96% Al2O3

Ag-6.8 wt. % Hf

ASTM F19Test Specimen

ASTM F19Test Specimen

Al-500 alumina brazed with Ag-3.4Hf & 0.25 mm Kovar™

SECA SOFC Seal MeetingJuly 8-9, 2003

Ability to correlate braze interface reactions with desired properties will yield more producible & reliable joints

0

21

42

63

84

105

0 20 40 60 80 100

AlSiFeCoNiAgHf

Co

nce

ntr

atio

n (

wt.

%)

Distance (µm)

Job #PBR02014Disk 2file 95

P. F. Hlava020207

Trace From Kovar Through Ag - 3.4Hf Braze Alloy into Alumina (Trace 3)Sample Brazed 990°C/10 min., dry H

2

BSE image shows continuous HfO2reaction layer at ceramic interface Kovar AL-500

0

21

42

63

84

105

0 20 40 60 80 100

AlSiFeCoNiAgHf

Co

nce

ntr

atio

n (

wt.

%)

Distance (µm)

Job #PBR02014Disk 2file 95

P. F. Hlava020207

Trace From Kovar Through Ag - 3.4Hf Braze Alloy into Alumina (Trace 3)Sample Brazed 990°C/10 min., dry H

2

BSE image shows continuous HfO2reaction layer at ceramic interface Kovar AL-500Kovar AL-500

Single crystal sapphire & polycrystalline alumina brazed with Ag-2Zr

94% alumina brazed with Ag-2Zr is hermetic, fails in the bulk ceramic and has 130 MPa tensile strength

Ag-3.4Hf ceramic braze joints also demonstrate similar excellent properties (hermetic & 125 MPa)

Recent Accomplishments• Implementation of active brazing in production• Conference paper / presentations (ASM Materials

Solutions 2002; AWS IBSC’03; ACerS 2003)• Metallurgical Transactions A paper (in publication)• Coating System for Direct Ceramic Brazing patents

SECA SOFC Seal MeetingJuly 8-9, 2003

Diffusion bonding offers an alternative joining processCeramic-metal and metal-metal joints are possible

a) Hard ceramic and comparatively soft metal surfaces come in contact (T, t, P)

b) Metal surface yields under high local stressesc) Deformation continues mainly in the metal,

leading to interface diffusion & void shrinkaged) Metallurgical bond is formed

TWI, Ltd, 2000.

SECA SOFC Seal MeetingJuly 8-9, 2003

FEA Brazing Model Development & Validation

ThermalResponses Stress & Failure

Predictions

Fluid Flow Responses

Design &Production

FixturedAssembly

+

Engineering Sciences Center Simulations

SECA SOFC Seal MeetingJuly 8-9, 2003

FEA Thermal Modeling of Furnace Brazing

Work Rack & Parts

• Nonlinear, 3-D transient thermal finite element code

• Ability to mesh very fine details (large node and surface radiation enclosure)

Typical Model Inputs• materials density• thermal conductivity• specific heat• emissivity• thermal boundary

conditions• convective heating by hydrogen

gas is assumed negligible• heating driven by radiation from

furnace elements & conductionfrom Mo shelves to work piece Simulations processed on massively

parallel teraflop compute server

SECA SOFC Seal MeetingJuly 8-9, 2003

Example: Two-Step, Cu + 50Au-50CuBraze Assembly Process Characterization

Oven - 1/2 section

FEA Thermal Model

Shelves/fixtures

backup ringweight

housingweight

fixture alignment

backupring

Au-Cu2nd Braze

Cu1st Braze

base

large flange

small flange

pins

housing frame

backupring

Braze Fixture

ceramic ring

SECA SOFC Seal MeetingJuly 8-9, 2003

Process Space Simulations

braze

Oven control TemperaturesTop ---Sides ---Btm ---

equalize

Data acquired:• Peak brazing temperature• Time braze is melted• Temperature uniformity• Determine optimization

parameters in process spaceversus experimental “point solutions”

Trends at Peak Temperature:• Top and bottom shelves are

hottest (trim heater settings)• Middle shelf most uniform• Center unit of five on each shelf

is coolest (~ 5°C lower)• 10 min. hold at set temp. gives

uniform heating, but increases time above melting

• Brazing temperature is uniform within individual units

4x24x5 7x2

7x5

Process Space:• Thermal boundary conditions:

– 1st ramp rate (10, 40°C/min)– 1st hold time (5, 25 min)– 1st hold temp (930, 960°C)– 2nd ramp rate (10, 40°C/min)– peak temp (1000, 1030°C)– time @ peak temp (0, 10 min)– cooling rate (10, 40°C/min)

• Number of shelves (4, 7)• Number of fixtures/shelf (2, 5)• Shelf type (solid, perforated)• Fixture material (Mo, Kovar®)• Emissivity (low, high)

Two-level fractional factorial design for 12 factors in 16 unique trials

SECA SOFC Seal MeetingJuly 8-9, 2003

Example: DoEx Screening Sensitivities forThermal Characterization of Brazing Process

h l it O h ld H th ld P th ld 1 2 h l T k t6

8

1 0

1 2

1 4

Factor sensitivities to the brazing conditions on the central shelf

47

2

525

5

960

930

0

40

10

1040

No

1030

Yes

1000

40

10

main effect factor significance

# of

She

lves

1stho

ld te

mp.

1stho

ld ti

me

Tim

e @

pea

k

1stra

mp

rate

2ndra

mp

rate

Hol

e in

shel

ves

Peak

tem

p.

Coo

ling

rate

# of

Uni

ts

time

abov

e m

eltin

g, m

in.

10