A

LSAdeHa©

K

1

icjobcwuftitratuf

0h

Available online at www.sciencedirect.com

ScienceDirect

Journal of the European Ceramic Society 34 (2014) 765–772

Glasses for laser joining of zirconia ceramics

Floriana-Dana Börner a,∗, Wolfgang Lippmann a, Antonio Hurtado a, Bastian Schön b

a TU Dresden, Institute of Power Engineering, Chair of Hydrogen and Nuclear Energy, George-Bähr-Str. 3b, 01062 Dresden, Germanyb Schott AG, Site Landshut, Christoph-Dorner-Strasse 29, 84028 Landshut, Germany

Received 7 March 2013; received in revised form 5 September 2013; accepted 25 September 2013Available online 19 October 2013

bstract

aser joining of ZrO2 ceramics using glasses and glass ceramics as sealing components requires optimized systems. The ternary systemsiO2–BaO–B2O3 and BaO–SrO–SiO2 were selected as a basis for development of suitable glass compositions for the laser joining process.dditives such as CaO, TiO2, Al2O3, and MgO were used to control the crystallization processes and hence the thermal expansion coefficientsuring glass synthesis. The glass viscosity, the strength of the ceramic-glass-ceramic joint, and the joint tightness are other important glass prop-

rties which were optimized for the laser process. For glass G018-345, this yielded strengths of up to 225 MPa (Weibull modulus of m = 8.6) ande leak rates of up to 4.3 × 10−5 mbar l s−1. Because of the varying viscosities obtained, the optimized glass systems could be used selectively in

temperature range of 700–900 ◦C. 2013 Elsevier Ltd. All rights reserved.

on; Z

hh

swwttotbsatGawm

eywords: Laser joining; Glass sealant; Mechanical strength; Thermal expansi

. Introduction

Current trends in materials development are showing anncreasing demand for functionalized components and an asso-iated need for joining materials of different compositions. Theoining technology that leads to the best joint depends mainlyn the material parameters and the compatibility of the parts toe joined.1–3 In this paper, the joining of yttria-stabilized zir-onia (YSZ) using laser radiation is described. A diode laserith combined beams of wavelengths 808 nm and 940 nm wassed as the laser source. In comparison with conventionally usedurnace technology, laser technology offers the decisive advan-ages of extremely short process times and pinpointed energynput. Development and optimization of laser-compatible andhermally stable glasses are examined in detail. The ZrO2 mate-ial used in the experiments was selected due to its popularity as

design and functional material and its high development poten-ial, as well as its selective oxygen ion permeability.1,4 YSZ is

sed as a solid electrolyte in the production of SOFCs and as aunctional layer in oxygen sensors.1 The production of zirconia

∗ Corresponding author. Tel.: +49 0351 463 34904; fax: +49 0351 463 39611.E-mail address: floriana.boerner@tu-dresden.de (F.-D. Börner).

tpeBaci

955-2219/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.jeurceramsoc.2013.09.019

rO2 ceramics

ousings for sensor and display applications using laser joiningas also been investigated.5

Extensive work was performed in the recent past to developpecial sealing-glasses for joining zirconia ceramics. Meltingas usually performed in a furnace, yielding borosilicate glassesith a B2O3:SiO2 ratio of approx. 0.55. BaO was added to adjust

he thermal expansion coefficients, La2O3 was added to alterhe viscosity, crystallization was suppressed through additionf Al2O3.6,7 For other SOFC glass applications using the sys-em RO–Al2O3–SiO2 (R = Ba, Ca), the BaO/CaO ratio variedetween 6 and 8.8 Other work dealt with glass systems in whichpecific crystallized components formed during joining. The rel-tively long process times of the furnace-based technology ledo a near-equilibrium defined glass microstructure in the seam.lass ceramics are preferential to conventional glasses for SOFC

pplications due to the higher joint strengths that can be achievedith them.9,10 Thus, for example, addition of MgO to a com-ercially available glass from Schott AG significantly increased

he strength of the joint due to the formation of crystallinehases such as BaAl2Si2O8, SiO2, and MgSiO3.9 The sameffect was achieved through specific conditioning of selectedaO–Al O –SiO glass compositions through addition of CaO

2 3 2nd B2O3.10 The main component that crystallized for thisomposition was BaSiO3. Similar observations were made innvestigations on the same base system modified with MgO and

7 opean Ceramic Society 34 (2014) 765–772

BBahZcf

srpsw

--

-

-

-

glajlraccm

piiobtosr7hd

2

2

Cs

sspiijaa

totbat3bopic

beiatS

As

66 F.-D. Börner et al. / Journal of the Eur

2O3.11 In this case, the crystalline components Ba2Si3O4 anda5Si8O21 as well as small amounts of crystalline BaMgSiO4nd Ba2MgSi2O7 were formed. The compositions describedere yielded joints exhibiting high strengths in combination withrO2 as well as chemical stability and high density in the spe-ific application conditions typically found in high-temperatureuel cells.

In general, the ceramic microstructure adjacent to the jointeam is not changed during joining. Unlike welding, solderingequires introduction into the joint region of a glass that is com-atible with both components to be joined.3,12–14 Within thecope of the work presented here, the developed sealing-glassesere optimized to achieve the following:

Good compatibility with the two components to be joined; Adequately high absorption of laser radiation by the sealing-

glass; Suitable flow properties of the molten glass in the joint gap

and very good wetting of the ceramic material; Specific thermal expansion coefficient of the glass adapted to

that of the ceramic material; Joining temperature during the laser process at least 100 K

above intended application temperature.

In laser-based joining, the energy required for melting thelass and heating the joint edge regions is introduced via theaser radiation in the joint area. Interaction between the lasernd the ceramic material hence determines the quality of theoint. Nonoxide ceramics are practically nontransparent to theaser wavelengths of the diode laser. This means that the laseradiation is absorbed on the surface of the ceramic componentnd energy transport into the body of the ceramic occurs via heatonduction. The high thermal conductivity of nonoxide ceramicsounteracts the thermally induced stresses, thereby preventingicro- and macrocrack formation.12

In contrast, oxide ceramics such as ZrO2 are partially trans-arent for the specific wavelengths of diode lasers. This resultsn absorption of the laser radiation as well as the thermal energynto the volume of the ceramic body and is the reason whyxide ceramics, despite having low thermal conductivities, cane heated using laser radiation in short times without allowablehermal stress being exceeded. However, the partial transparencyf oxide ceramics is strongly dependent on temperature. Mea-urements of optical properties as a function of temperatureevealed a linear increase in laser absorption up to approx.50 ◦C followed by an exponential increase in absorption atigher temperatures.14 This fact must be taken into accounturing use of the laser technology.

. Materials and methods

.1. Materials

A commercial ZrO2 ceramic material (ZN101B) fromeramTec GmbH was used for joining. The material was iso-

tatically pressed and sintered at 1500 ◦C.

v

sc

Fig. 1. Base system BaO–SrO–SiO2 (Source: Sci. Glass).

Three glasses were developed by Schott AG for use asealing-glasses in a laser joining process. The thermal expan-ion coefficients of the glasses were adapted as closely asossible to that of the zirconia ceramic material without increas-ng the joining temperature to >1100 ◦C. All of the glassesnvestigated showed crystalline phases in the glass matrix afteroining, enabling the thermal expansion coefficients to be suit-bly adapted and adequate thermal stability of the joint to bechieved.

The first glass developed (G018-339) had a composition inhe ternary system SiO2–BaO–B2O3. Earlier work performedn the base system SiO2–BaO–B2O3 showed crystallization ofhe phases sanbornite (BaSi2O5) and cristobalite (SiO2). Cristo-alite must be avoided due to its anisotropic expansion. Thenisotropic thermal expansion of cristobalite is caused by a phaseransformation in the temperature range between 200 ◦C and00 ◦C. This phase transformation affects the thermal expansionehavior strongly and is undesirable due to a possible formationf cracks. Thus, CaO was added to the glass system to sup-ress the formation of cristobalite in the glass. TiO2 was alsontroduced into the glass composition as a nucleation aid to favorrystallization of the desired phase, sanbornite.

The ternary system BaO–SrO–SiO2 was selected as anotherase system. This system includes, among other features, autectic point at which the crystalline phase SrBa2Si3O9 precip-tates (Fig. 1). This phase has a relatively low melting point atpprox. 1100 ◦C as well as high SrO and BaO contents, sugges-ive of a high thermal expansion coefficient. The SrO, BaO, andiO2 contents were selected to match the eutectic composition.

The glass G018-340 was synthesized through addition ofl2O3, MgO, CaO, and B2O3. These additive components were

elected to reduce the glass formation temperature as well as theery strong crystallization tendency.

The third glass, G018-345, was yielded from application-pecific adaptation of its predecessor. The composition washanged minimally (MgO concentration increased and Al2O3

F.-D. Börner et al. / Journal of the European Ceramic Society 34 (2014) 765–772 767

Table 1Properties of the glasses developed by Schott AG for the laser joining process.

Glass properties G018-339 G018-340 G018-345

System BaO–SiO2–B2O3 BaO–SrO–SiO2–B2O3 BaO–SrO–SiO2–B2O3

α20–300 [K−1] 9.7 10.1 10.3Tg [◦C] 628 602 598Tcrystallization [◦C] 767 836 980Hemisphere temp. [◦C] 997 852 828Flow temp. [◦C] 1012 918 876L ◦

atpc

Mtwwt

pAtgtti

2

aj

raat

ttto

rc(

nns

fis

ogfit

t

tchr

rttt

wtsw

3

characteristics such as wetting of the ceramic surface with theinvestigated molten glasses, the quality of the ceramic-glass jointsurface, and the formation of stress cracks.

aser joining temp. [ C] (for bar–cylinder joint) 1120

ddition omitted) to achieve a further reduction in the meltingemperature and to lower the viscosity of the glass. The mainhysical properties of the glasses developed by Schott AG areompiled in Table 1.

Glasses were synthesized using high-purity raw materials.ixtures of defined composition were melted in Pt crucibles at

emperatures between 1500 ◦C and 1600 ◦C; the molten glassas then quenched using water-cooled metal rollers. The shardsere transferred to porcelain drum mills, where they were milled

o powders with a d50 grain size of approx. 10 �m.Ethanol was added to the glass powder to form a viscous sus-

ension, which was applied as a thin layer to the joint surface.fter evaporation of the volatile component the glass concen-

ration was 0.25–0.3 mg/mm2. For obtaining the adhesion of thelass powder to the ceramic material required for the process,he applied glass was thermally treated at a temperature betweenhe hemisphere temperature and the flow temperature for eachnvestigated glass prior to the laser process.

.2. Laser process

A diode laser (Rofin Sinar GmbH) with wavelengths 808 nmnd 940 nm and a maximum power of 3.1 kW was used for laseroining of the ceramic parts.

Two different types of tests were performed: joining of twood-shaped ceramic bars and joining of a rectangular bar inside

cylindrical ceramic part. The laser beam was guided through focusing element with a focal length f = 305 mm for achievinghe bar–bar joint.

The circular irradiated area had a diameter of 6.4 mm. Duringhe laser process the joint was centered in the laser beam andhe irradiated ceramic material was heated to the glass joiningemperature. A water-cooled absorber ensured that no damageccurred beneath the joined specimen.

During the joining process the bar-shaped samples wereotated to achieve uniform sample heating. The rotation deviceonsisted of two jaws driven synchronously by an electric motorFig. 2). The optimum rotational speed was 6.7 rad/s.

Rotation of the ceramic during the joining process was notecessary for producing the bar–cylinder joint, eliminating theeed for thermal pretreatment of the glasses in an intermediate

tep.

The laser beam was guided over the scanner after it passed thexed optical element of focal length f = 405 mm. The selectedcan path was circular with a diameter of 4 mm. The scan path

1035 960

n the cylinder surface was a circle of diameter 15 mm (shown inreen in Fig. 3, left). The optimum working distance between thexed optical element and the cylinder surface was determined

o be 370 mm.During laser radiation the laser focal point was moved along

he path shown in red in Fig. 3 at a speed of 500 mm/s.The options for temperature measurement were limited due

o the technological characteristics of the laser process. Theonventional variation with a thermocouple could not be usedere because the thermocouple would strongly absorb the laseradiation and lead to artifacts in the measured signal.

A suitable variation proved to be IR thermography, whichequires knowledge of the emission coefficients for the materialo be measured. An emission coefficient ε of 0.81 was found forhe ZrO2 material. The measurement method used to determinehis is described in detail in the literature.14

A type VarioCam hr IR thermal camera from Infratec GmbHas used. An infrared block filter protects the sensor units in the

hermal camera. The distance between the filter and the sampleurface was 80 cm, and the surface temperature recording rateas 2 Hz.

. Characterization methods

Light microscopic methods were used to test seam quality

Fig. 2. Test setup for laser joining of bar-shaped specimens.

768 F.-D. Börner et al. / Journal of the European Ceramic Society 34 (2014) 765–772

(left:

gmy(twE

rpH3Aotgtbi

nfii

sEpait

ZT2pt

σ

w[

wwdbp

mXamlced

4

4

According to the datasheet provided by the manufacturer,CeramTec AG, the ZrO2 ceramic used had a thermal expansion

Fig. 3. Path of travel of the laser beam on the irradiated specimen

The microstructures and compositions of the crystallizedlasses were investigated by means of scanning electronicroscopy (SEM) and energy-dispersive X-ray (EDX) anal-

sis. Measurements were made using a microscope from Zeiss1540VP with INCA-Energy/Oxford Instruments EDX detec-or) with acceleration voltages of 10 kV and 20 kV. Elementsith atomic numbers of 6 and higher could be detected withDX analysis; boron could not be detected.

The glass viscosity was measured using a high-temperatureotational viscometer (especially built by Schott AG). The flowroperties were investigated using a hot stage microscope (EMI,esse Instruments). With this technique, a cube (edge length:

mm) hand-pressed from the glass powder was heated on anl2O3 plate in air at a heating rate of 10 K/min until fusingccurred. The sealing temperature was usually chosen to be inhe same range as the so-called hemisphere temperature of thelass or deviated from it (typically) by ±20 K. The hemisphereemperature is the temperature at which the original sampleody melts to form a hemisphere. The hemisphere temperatures associated with a viscosity log η of approx. 4.6.

The sintering temperature was also measured with this tech-ique. The sintering temperature is the temperature where therst shrinkage, due to sintering of the glass powder, of the cube

s detected.For strength testing of the ceramic-glass-ceramic joints, the

tandardized four-point-bending-test in accordance with DINN 843-1 was used.15 The special arrangement of the loadingins (Fig. 4) yielded information on the strength of the jointlone. The bars, which were joined at their faces, were positionedn a way ensuring that the joining zone was centered betweenhe loading pins.

Testing was done using a material testing machine fromwick & Roell with control of the applied force (Fmax = 5 kN).he load was selected to ensure that fracture occurred approx.0 s after the test was started. The spacing between the loadingins was 20 mm and 40 mm between the support pins. The frac-ure stress was calculated according to the following equation:

= 16IA

πd30

F (1)

schematic diagram; right: image taken during the laser process).

here IA, span [mm]; d0, diameter [mm]; F, forceN].

For comparison of the measurement parameters, 20 samplesere tested and a Weibull plot of the measured flexural strengthsas obtained for each test series for all glasses. The Weibullistribution function is frequently used for describing failure ofrittle materials and characterizing the failure by means of thearameters m (Weibull modulus) and σ0 (mean strength).4

Information on the tightness of the joints was obtained byeans of a helium leak test with a mass spectrometer PhoenixL 300. Helium leak systems are based on detection of helium

s a test gas. This technique is one of the most sensitive testethods for measuring joint tightness and can detect very low

eak rates. In the integral leak test, specimens are placed in a testhamber, evacuated, and then pressurized with helium. Heliumscaping due to leaks is suctioned by a high-vacuum pump andetected.

. Results and discussion

.1. Thermal properties

Fig. 4. Sample position during strength testing.

F.-D. Börner et al. / Journal of the European Ceramic Society 34 (2014) 765–772 769

Table 2Properties of the crystallized specimens.

Crystallized glasses G018-339 G018-340 G018-345

Sintering temperature [◦C] 951 931 939Softening temperature [◦C] 981 941 945Spherical temperature [◦C] 992 946 979Hemispherical temperature [◦C] 1009 954 998F

cd9tmecat

G9ib

pugt

ct

4

ecGiojto

500 600 700 800 900 1000 11002

4

6

8

10

12

14

lg / d

Pa

*s

temperature / °C

G018-339

G018-340

G018-345

F

a

1jdcgtca1mwpp

jjpcgh

low temperature [◦C] 1010 970 1012

oefficient α20–1000 ◦C of 10.4 × 10−6 K−1. For the glasseseveloped by Schott AG, thermal expansion coefficients of.7–10.3 × 10−6 K−1 were measured. Glass G018-345 had thehermal expansion coefficient closest to that of the ceramic

aterial. Even with the maximum difference in thermalxpansion �α of 0.70 × 10−6 K−1 (glass G018-339), no stressracks were found during cooling. Hence, it was concluded thatll three glasses could be used to produce low-stress joints inhe laser joining process.

Melting temperatures of <1100 ◦C were achieved. Glasses018-340 and G018-345 already flowed at temperatures of18 ◦C and 876 ◦C, respectively. The interval between sinter-ng temperature and flow temperature was extremely short foroth glasses, presenting an advantage in terms of process time.

To evaluate the maximum operation temperature, the flowroperties of crystallized specimens (Table 2) were investigatedsing the hot stage microscope. For these studies, the developedlasses were annealed for 48 h at their crystallization tempera-ure.

Due to the crystallization occurring in the glasses, the appli-ation temperature can be specified to be approx. 900 ◦C sincehey show their first sintering above 900 ◦C.

.2. Laser joining experiments

When irradiated by the laser, the glasses showed differ-nt flow behaviors. Glass G018-339 yielded a highly viscousrystallized system at the joining temperature, whereas glass018-345 represented a system with very good flow at the max-

mum process temperature and good filling or tight sealing evenf narrow gaps (Fig. 5). The observations made during the laser

oining process were in agreement with the measured viscosi-ies. Glass G018-345 exhibited a much lower viscosity than thatf glass G018-339 at the same temperature.

ttZ

Fig. 5. ZrO2 ceramics joined using glasses G018-339 (

ig. 6. Change in viscosity with increasing temperature for developed glasses.

The measured viscosities are shown as a function of temper-ture in Fig. 6.

For producing the bar–bar joint, a constant laser power of70 W could be used for all of the glasses investigated. Whereasoining temperatures of about 1200 ◦C were required for the pro-uction of stable joints with glass G018-339, the laser processould be stopped at temperatures of 1040 ◦C and 980 ◦C forlasses G018-340 and G018-345, respectively. Through reduc-ion of the melting temperature of glass G018-345, processontrol for obtaining a reproducible seam quality was easier tochieve. In this case, the working point was found to lie approx.50 ◦C below the couple-in temperature of the ZrO2 ceramicaterial to be joined. The couple-in temperature is the point athich the temperature increases exponentially at a constant laserower. Control, monitoring, and reproducibility of the joiningrocess are difficult at temperatures approaching this point.

For the bar–cylinder geometry, the laser power required foroining was 330 W, which was held constant during the entireoining process for all glasses. Higher powers could shorten therocess considerably but would lead to crack formation in theylindrical part. The joining trials with glass G018-339 yieldedood results, which, however, were associated with a relativelyigh glass viscosity. Increasing the joining temperature to reduce

he viscosity caused the temperature of the ceramic materialo rise above 1100 ◦C, nearly the couple-in temperature of therO2 ceramic. Being the case with the bar–bar geometries, with

left), G018-340 (middle), and G018-345 (right).

770 F.-D. Börner et al. / Journal of the European

tpagtthMlG

ta

ugtwsj

4

Ggw1caSIg

ygwquantities of Mg and Ca. The needle-like crystals showed onlytraces of Ca and no Mg. The gray platelet-like crystals contained

Fig. 7. Investigation of bar–cylinder joint by computer tomography.

his geometry, precise control and hence reproducibility of therocess was only possible to a limited extent in this temper-ture range. Much better joining results were achieved withlasses G018-340 and G018-345 due to the lower viscosities andhe lower melting temperatures of these glasses compared withhose of glass G018-339. With the procedure described above,omogeneous and adequately strong joints could be produced.acroscopic cracks were not found in the joined part after the

aser process. The joining temperatures were 1035 ◦C for glass018-340 and 960 ◦C for glass G018-345.Computer tomographic investigations on the joints indicated

hat the molten glass flowed easily into the gap between the barnd the cylindrical part to form a tightly sealed joint (Fig. 7).

ad

Fig. 8. Changes in surface temperatures of joined parts subjec

Fig. 9. Micrographs of glass G018-340: 500× (le

Ceramic Society 34 (2014) 765–772

The heating of the glass-ceramic joint is illustrated in Fig. 8sing the example of glass G018-340. It can be seen that with theiven test setup, the energy from the laser beam was localized inhe joining region. Symmetric heating of the sample and slightarming of the support surface ensured that no uncontrolled

tress due to large temperature differences could arise in theoined parts.

.3. XRD and EDX analysis of the joined area

EDX investigations on the microstructure of Schott glass018-340 indicated the presence of idiomorphic crystals elon-ated in part up to 20 �m. In some cut sections, the crystalsere hollow. Round pores (gas bubbles) with diameters of up to0 �m were also found (Fig. 9). The glass matrix was shown toonsist primarily of O, Si, Sr, and Ba, but it also included smallmounts of Mg, Al, and Ca. The crystals mainly contained O,i, Sr, and Ba. Mg, Al, and Ca were not detected in the crystals.n addition, there was slightly less Sr in the crystals than in thelass matrix.

The same investigations performed on glass G018-345ielded two different types of crystals (Fig. 10). Apart fromray platelets, bright needles were observed. The glass matrixas found to contain mainly O, Si, Sr, and Ba as well as small

small amount of Ca and more Mg than the surrounding matrixid.

ted to laser radiation (example: ZrO2 + Lot G018-340).

ft) and 2000× (right) (Image: Bosch AG).

F.-D. Börner et al. / Journal of the European Ceramic Society 34 (2014) 765–772 771

4

Notcwco

d

fi

tG(

g

Table 3Calculated Weibull modulus m and characteristic strength σ0 for glasses opti-mized for joining ZrO2.

Joint σF [MPa] WeibullModulus m

σ0 [MPa]

ZrO2–G018-339–ZrO2 227.00 ± 15.00 3.58 251.00ZrO2–G018-340–ZrO2 190.00 ± 8.00 6.58 204.00ZrO2–G018-345–ZrO2 212.56 ± 21.00 8.35 225.15

Table 4Measured He leak rates for laser-joined ceramic materials using glasses G018-340 and G018-345.

Material system Joint Leak rate [mbar l s−1]

ZrO2 ZrO2–G018-340–ZrO2 2.9 × 10−4

Z

ooiwts

3Ns

5

icpa

Fig. 10. Micrograph of glass G018-345 (Image: Bosch AG).

.4. Mechanical properties

The failure origin was in the joint region for all specimens.o fractures in the ceramic materials adjacent to the joints werebserved. Damage to the ZrO2 ceramics during laser irradia-ion could be excluded. For this study, bar-shaped specimensonsisting of monolithic ZrO2-ceramic samples were irradiatedith the laser power required for the investigated joining pro-

esses. Testing the strength of irradiated samples showed a meanf 706 MPa, the standard deviation was 32 MPa.

As an example, Fig. 11 shows the graph of strength used foretermining the Weibull modulus for glass G018-345.

The slopes of the lines ln(ln(1/1−F)) = f(ln(σB)) yielded theollowing values for the Weibull modulus m for the glassesnvestigated (Table 3):

Although the means of the fracture stress was nearly identical,he probability of joint fracture was considerably higher for glass018-339 at low stresses. This resulted from the different slopes

m values) of the distributions.The highest slope was obtained with joints produced with

lass G018-345. The high value indicates good reproducibility

Fig. 11. Weibull plot of strength for the ZrO2–G018-345–ZrO2 joint.

oTwmjccsia

awgwoctwfp

rO2 ZrO2–G018-345–ZrO2 4.3 × 10−5

f the joining process with this glass system. A high reliabilityf the laser joining process for achieving a bonded joint in thenvestigated ZrO2 material system can also be assumed. Theork performed repeatedly showed that especially for extreme

hin seams (<20 mm) between the joined ceramic materials, hightrengths were achieved.

The measured leak rates for glasses G018-340 and G018-45 are listed in Table 4 and indicate adequate leak tightness.o measurements were made for glass G018-339 because this

ystem is relatively unsuitable for geometries with thin gaps.

. Conclusions

In the present work, it was shown that the laser processs a suitable method for joining of the investigated zirconiaeramic material using barium silicate glasses as sealing com-onents. The good mechanical properties of the joints could bechieved with a high reproducibility for the selected glasses outf the ternary systems BaO–SiO2–B2O3 and BaO–SrO–SiO2.he properties viscosity and wettability of the sealing glassesere found to be very important, in particular for the develop-ent of solder systems for short time processes such as the laser

oining technology. Even small differences in viscosity broughthanges in the laser process control as well as the possible appli-ation area. Thus, the developed solder G018-339 was moreuitable for filling large gap volumes due to a higher viscos-ty, while the solders G018-340 and G018-345 were also morepplicable for very narrow joint gaps.

Although the joining by means of a laser source needs only short time, it could be shown that also crystallized glassesith defined structure may be used as solder components. In thelasses G018-339, G018-340, and G018-345, crystalline phasesere formed during the laser joining, supporting the adaptationf the thermal expansion coefficients as well as the later appli-ation temperatures of the joints. It is also conceivable to usehe process of laser joining for high temperature applications,

hich only have been realized by a thermal process furnace so

ar. Because of the glass properties achieved, application tem-eratures of between 700 ◦C and 900 ◦C are expected for the

7 opean

dos

A

oGf(

SgC

R

1

1

1

1

72 F.-D. Börner et al. / Journal of the Eur

eveloped glass solders. Their applications would be in the rangef the SOFC and the production of exhaust gas sensors as well asensors for the temperature measurement in exhaust gas systems.

cknowledgments

The results presented here were obtained within the scopef the “CeraJoin” project. The authors would like to thank theerman Federal Ministry of Education and Research (BMBF)

or providing financial support over the course of the projectreference number: 03X0507F).

Dr. Köhler from R. Bosch AG in Stuttgart, Germany, and Dr.chunck from CeramTec GmbH in Plochingen, Germany, areratefully acknowledged for the SEM and EDX work and theT investigations, respectively.

eferences

1. Kollenberg W. Technische Keramik Grundlagen – Werkstoffe – Verfahren-stechnik, vol. 134. Essen: Vulkan-Verlag; 2004. p. 515.

2. Verein Keramische Industrie e.V. Brevier Technische Keramik; 1999. p.131–42.

3. Messler Jr RW. Joining of engineered ceramics. Adv Mater Process2004;162(11):43–5.

4. Michalowski L. Neue keramische Werkstoffe. Elektrische Funktionskeramik.Leipzig, Stuttgart: Dt Verlag für Grundstoffindustrie; 1994. p. 108–12.

1

1

Ceramic Society 34 (2014) 765–772

5. Kasch S, Köhler G, Müller H, Wächter S. Prozess- und funktionsangepasstesLaserstrahllöten mit Glaslot für Sensor- und Displayanwendungen. DVS2007;243:201–4.

6. Ghosh S, Kundu P, Das Scharma A, Basu RN, Maiti HS. Microstructureand property evaluation of barium aluminosilicate glass-ceramic sealant foranode-supported solid fuel cell. J Eur Ceram Soc 2008;28:69–76.

7. Sohn SB, Choi SY, Kim GH, Song HS, Kim GD. Suitable glass-ceramic sealant for planar solid-oxide fuel cells. J Am Ceram Soc2004;87(2):254–60.

8. Saswati Ghosh P, Das Scharma A, Kundu P, Basu RN. Glass-ceramicsealants for planar IT-SOFC: a bilayered approach for joining electrolyteand metallic interconnect. J Electrochem Soc 2008;155(5):B473–8.

9. Eichler K, Solow G, Otschik P, Schaffrath WBAS. (BaO Al2O3 SiO2)– glasses for high temperature applications. J Eur Ceram Soc 1999;19:1101–4.

0. Meinhardt KD, Kim DS, Chou YS, Weil KS. Synthesis and properties of abarium aluminosilicate solid fuel cell glass-ceramic sealant. J Power Sources2008;182:188–96.

1. Gödeke D, Dahlmann U. Study on the crystallization behavior and thermalstability of glass-ceramics used as solid oxide fuel cell-sealing materials. JPower Sources 2011;196:9046–50.

2. Lippmann W, Herrmann M, Hille C, Hurtado A, Reinecke AM. Laser joiningof ceramics. CFI Ceram Forum Int 2008;85(13):60–4.

3. Heilmann F. Dreidimensionales Laserhybridlöten zum dauerhaften Fügengleicher und ungleicher Oxidkeramiken. Dissertation, Dresden; 2009.

4. Börner FD, Lippmann W, Hurtado A. Laser-joined Al2O3 and ZrO2 ceram-ics for high-temperature applications. J Nuclear Mater 2010;405:1–8.

5. DIN EN 843-1. European Standard/Monolithic Ceramics. Mechanical prop-erties at room temperature. Part 1: determination of flexural strength; 1995.