Electrophoretic and Electrolytic Deposition of Ceramic Particles on ...

2 downloads 0 Views 10MB Size Report
Consultant on Analytical Investigation. G. Starck ...... consulting with experts of Wright-Patterson Laboratories. Tables 1 and 2 ...... erent csi.-ates. fti.uther on.
AEOSR.R3. 93

AD-A265 141

0

44

I

Electrophoretic and Electrolytic Deposition of Ceramic Particles on Porous Substrates

L. Gal-Or, S. Haber, S. Liubovich •

5, --

Israel Institute of Metals

1

--

Technion Research and Development Foundation

74

Final Scientific Report 3

March 1, 1988 - November 30, 1992

DTIC ELECTE

SBD

MAY2 7.1993

Prepared for

EOARD - England, London AFOSR-Bolling, Washington, D.C.,20332, U.S.A.

93-11951 • II lN Iii 11vr1111 fOP P1111 distri but ton Unlalsited.

93 -5 26 072

Electrophoretic and Electrolytic Deposition of Ceramic Particles on Porous Substrates

L. Gal-Or, S. Haber, S. Liubovich

Israel Institute of Metals Technion Research and Development Foundation

Final Scientific Report March 1, 1988 - November 30, 1992 AFOSR Grant 89-0474

Prepared for EOARD - England, London AFOSR - Boiling, Washington, D.C.,20332, U.S.A.

Approved for ni?.I e, relcaso; distributiouj uu1imited.

Electrophoretic and Electrolytic Deposition of Ceramic Particles on Porous Substrates

L. Gal-Or, S. Haber, S. Liubovich

Israel Institute of Metals Technion Research and Development Foundation

Final Scientific Report March 1, 1988 - November 30, 1992 AFOSR Grant 89-0474

Prepared for EOARD - England, London AFOSR - Bolling, Washington, D.C.,20332, U.S.A.

Approvedt fory• di~tribuLt ia

i ! release; Luilimi ted.

I

REPORT DOCUMENTATION PAGE Puohicleooffff

cufCrd* o th-r fM, cct e,

;amffTfffrO4"0 tfO

l

o

In.

:'

DUOnfMe t o

~Ce'a,

'-rmftior , ntlmiittO to &-eraq# Ifour der 4,00Pse ,r-uo-C frf _rn o #'aCcOr'•Oelmfq arto

f•,flCW'

fie cOIlt.•C•flf Of fflfV

Il

ýenO co-me"ts rl90fOe g ftr0 Ms

cotiemoh a' 'nformaincf,c -v~oing suggeIfron 1cf feoucif'q On'sturcentA I Wasnmlt~qorimrcf MtDO~J a l ft O*,MC4.S..xe .ZC4.AlrrnqioC. VA '2C2243CZ and torfa 00,(eolmor qe~relt'M 6.09boel.O'Def-off kiftCI

I. AGENCY USE ONLY (Ledve blank)

2. REPORT DATE

ForrmlApproved OM NO 0704.0188 *tffl% fdf

,I'O

data'Ct *11&'

t'fffa" q Of 11, g O'n ,

0"t. ,

,,

21S 'tffw ,0~'iCtCIJ07Q04118) WCV-floln 0C ;owlC fo 10

lIketO~

oreoa

andC~tf ia on,

13. REPORT TYPE AND DATES COVERED Finai

March 1, 1988 - Sept. 30, 1992

4. TITLE AND SUBTITLE

Electrophoretic and Electrolytic Deposition of Ceramic Particles on Porous Substrates

Report

S. FUNDING NUMBERS

AFOSR89-0474

6. AUTHOR(S)

L. Gal-Or, S. Liubovich, S. 7. PERFORMINt,

1Haber

ORGANIZATION NAME(S) AND ADORESS(ES)

..

8. PERFORMING ORGANIZATION REPORT NUMBER

Technion R&D Foundation Technion City, Haifa, 32000, Israel 9. SPONSORING; MONITORING AGENCY NAME(S) AND ADDRESS(ES)

10. SPONSORING/ MONITORING

- EOARD 223/231 Old Marylebone Rd. London NW1 STH UK - AFOSR/NE, Lt.Col. L. Burggraf, Boiling AFB DC 20332-6448

AGENCY REPORT NUMBER

11. SUPPLEMENTARY NOTES

12a. DISTRIBUTION /AVAILABILITY STATEMENT

12b. DISTRIBUTION CODE

Unlimited Approved for publ 10 release distribution unlimitsd, '13. ABSTRACT (Maximum 200 word•)

Electtopnoretic deposicion of ceramic particles and their penetration into tne pores •f graphite and a 2D C-C composite were demonstrated and studied tneoreticaiiv andc experiimental v for colloidal SiO,, SiC, SiN 4 , AI 2 TiOs and HfTIO,. The effect of deposition parameters (field intensity, particle concentration, ratio of dielectric constant/viscositv of fluid) on the amount of induced material was experimentallv studied. Penetration depth of particles as function of above parameters was analyzed theoreticaily. Penetration is enhanced by large Peclet numbers and is therefore higher in aqueous suspensions, while coatina morphology is better in propanoi suspensions. OiF-imal conditions for coverage and penetration in a two-stage process *;ere tperelore aetermined. Subsequent deposition of two layers ý.as demonstrated and studitd for Si 3 N4 on AI 2 TiO5 and SiC on a glass ceramic. Single (Si 3 XN, AI 2 TiO5 , HfTiO,) and multilaver ,Si,%,N on Ai 2 TiO,, and Si:N4 on fused SiO0) coatings were deposited, sintered in a N2 atmosmnere and their phase composition, and Inorphology studied. Oxidation resistance of coated specirvens was studied at 83U0C in continuous and intermittent exposures. Grapnite exnaust nozzles were coated and tested in a9 air breatning enaine simulation. Very efficient protection was achieved. Electrolvtic deposition of ZrO2, A1lO3 and Al.O-ZrO2 was achieved from aqueous solutions of ZrO(NO3 ), and AII(O 3 )3 . The deposition kinetics and mrcrostructure of the deposits were studied. The initial amorphous deposits transformed into crystalline pnases with nonocrsztaiiLre dimensions after heat treatment.

'14. SUBJECT TERMS

...

Electrophoresis,

Ceramics,

Impregnation,

Coating,

Electrolytic Deposition 17. SECURITY CLASSIFICATION OF REPORT

Unclassified

NSN 7540-01-280-5500

Ia. SECURITY CLASSIFICATION OF THIS PAGE

Unclassified

1.. 15.

NUMBER OF PAGES

16. PRICE CODE 19.

SECURITY CLASSIFICATION OF ABSTRACT

20. LIMITATION OF ABSTRACT

Unclassified

Unlimited Saar'.ao ;orm, 298 (Rev P-

o

C

-0

-4

m~2

~+0

4E-4

0

020

1!t

0

R

i

a

0C

C)

C)~C CO~Z4 -i

Enn -4-4t-to

CV

.

02q

0

C~4 10

0)0 C44

C*4i

m

Go co~

00

40

10

ed

to

w +ot(

V= Lo =0ocG

00

oj

CV)

C

4)

at

.~C4

2

C4

0

'

0

s L

-4

~

&MC4C

M4

n

z

(n)

U

m

Wo4

'4

a

a

i

0

0

-W

.S!

4

Ur1Sri 4

I

c

Iz

0

0

t

W

0

+0

:EI

-

-4.

*

An

4

-

-

cc 0

0

=6

z zL00

""

0

1:W

I-

25

uC.

I

I

4

V4

i

J

0

'w

E

~

o

w~ *-

EZ

H >+

0

o U,.;

w u U24;-

w

00

0

E.~

w

in

o

r

el

-

0

m

-4

ul

4

07

.O

C4

U-

z

x

..

0

co

0

%-

cn

RW~

.4 V, .Lz

to

wZ-

4-T4

W~

C14

I4

I

0j

w

Iq A.

A.

IV

4..

a.

93~

<
2-

.-'

C)

11X

-

4r

-J

o

-

X

X

-

=

I

X

p

-

C,;

C.)

-

-

z

,-'

-

x

3

C)

.-

'-

x

-: -

a

-

0

Lfl

r'*)

0

tfl

0

C'J

N

LC

0

Lfl

-

-4

C.,'

.c

1

00 0

C)

IV')

CN

ssol

1421L

00

7ýA

4r1

-4

I-

C-)

r

4

X4

ccf

___

Q1n

Mir//

I

L

LCV)0

Nrn

-

C-

W

l)a

Cj

0

to ".7

-0

('10

Isllq-

00

~~Y0

________________________

0 CA (00

0 ?A..

AL2a

0

0

ýo

" -AA

i'z-

64

4.

5ummary and Conclusions The work presented in this report concentrated on the final steps of the

development

of electrophoretic

carbon-carbon composites.

protective

coatings

for

graphite

and

The emphasis was therefore on the sintering of

these coatings and the testing of their protective properties. The deposition conditions were determined in the previous stages of this project. The coating systems studied were: Al 2TiO 5 , HfTiO 4, Si3N 4 , SisN4 on AI 2TiO 5 , Si 3N 4 on fused SiO 2 and Si3N4 on BN. The Al 2TiO 5 and HiTiO 4 coatings were selected due to their compatibility with graphite and C-C in terms of thermal expansion.

However, it was expected that due to the

anisotropicity of the thermal expansion coefficient microcracking will take place and therefore additional overcoating will be needed for oxidation protection. The HfTi0

4

disintegrated after sintering and no viable undercoating

could be obtained.

Al 2TiO 5 was impregnated and uniformly deposited.

Sintering at 1650"C in N2 resulted in adherent coatings which however decomposed partially into A12 0 3 and TiO 2 as shown by the X-ray diffraction pattern. Si 3N4 coatings were obtained both in single and multiple deposition stages. Multiple deposition was done primarily in order to change the holder position to minimize defects in the coatings. After sintering at 1650'C in N 2 uniform, adherent coatings were obtained with exte,,tive anchoring in the pores and voids of the substrate. Interaction with the substrate took place as indicated by diffraction peaks of SiC and Si. Si3N 4 coatings with underlayers of A12TiO 5 were obtained.

Uniform

layers of Si 3N 4 with penetration into the Al 2TiO5 underlayer formed in the "green" state.

Afte: sintering cross-sections showed dense layers of Si 3N4

65

anchored in the substrate with a discontinuous layer of a Ti containing component below the Si3N 4 and an even distribution of the element Al in the entire coating. Very extensive anchoring of the Si3N 4 coating is seen both in graphite and C-C specimens.

Overall coating thicknesses were up to 88A.

Adherent coatings of Si 3 N4 on fused SiO 2 were obtained with anchoring in the substrate. Horizontal and vertical microcracks are seen in the coatings after sintering. Impregnation of BN into the pores prior to Si3N4 coating was attempted. Since both B and N are not visible in the EDS-SEM no evidence of penetration was seen except for weight gain after impregnation. Shaped objects, exhaust nozzles and casting crucibles, were coated with Si 3N 4 + A12TiOs and Si 3N 4 + Si0 2. Uniform internal and external coatings were obtained using shaped electrodes. Oxidation experiments were carried out by recording the weight loss during continuous and repeated exposures of 830" C.

Best results with

graphite specimens were obtained with the Si3N4 + A12TiO 5 system where in some cases nil weight loss was recorded after 7 repeated 1-hr exposures, a regime that exposed the specimens to repeated thermal shock. No beneficial effect of an underlayer of fused SiO 2 or BN was found in laboratory specimens when compared with Si3N4 only. However, very good results were obtained with an underlayer of Si0

2

on exhaust graphite nozzles under

specific testing conditions. On C-C specimens low weight losses were obtained with SiSN 4 and Si 3N4 on A1 2TiO 5 in continuous exposure experiments. In most repeated exposure tests high weight losses were recorded. However, in some experiments Si 3N4 coatings gave low weight loss. The irreproducibility of the results in both

66

materials and in particular in the case of C-C may stem from a major defect in the coatir5 formed on the location of the holder during deposition. Although repeated depositions were made this problem may not have been solved completely and should be elaborated in further experiments.

REPORT DOCUMENTATION PAGE ¶set-main 4:00v, cO.tbon ot imiar'maita P.J~d.Cta~r~$q h,*m. 1n.ntarnn the 43ta nnl,•t a.Carl, cOm ttflglfga r r,.toucfing :thil d luqqesItOMS ma" Or,,to. .~ctudclnq n'CCtl 3Z02.and to u- 0.': H•,lwaV. SWtte 1204. Aretq0tn. VA •

~'o

v

1. AGENCY USE ONLY (Leave blank)

0MB No

~

.7O4•.t88

~

-.:~oa~~t .e.' ng IOI 0e 1 hour >e ,ý,ooee. inaveiNth~Ue cMtom . :. vm, Other #o o.rC.g O M "tt M.te rConaml(It"% lt Of intorrm ton. $endd tNirtioIcuate Selrmes.O,'fOEtiw tar :ntornsucin Oieeto ns A.:ot4i.oorn li l 03a Warenl-.U.. . Aý ddOn 0uaqet. 8*0U W-. Orft *rOec(OJOA4S4t t and o"Ic '-j'-t4-4c. :.

eV*

:t.

3. REPORT TYPE AND DATES COVERED

2. REPORT DATE I August 30, 1991

Annual

report

1 July 90-30 June 91 5. FUNDING NUMBERS

4. TITLE AND SUBTITLE

Electrophoretic and electrolytic deposition of ceramic particles on porous substrates

AFOSR-89-0474

6. AUTHOR(S)

L.

Gal-Or,

S.

Haber, S.

Liubovich 8. PERFORMING ORGANIZATION

7. PERFORMING ORGANIZATION NAME(S) AND AODRESS(ES)

REPORT NUMBER

Technion Research and Development Foundation Technion City, Haifa 32000,

Israel

10. SPONSORINGIMONITORING

9. SPONSORING/ MONITORING AGENCY NAME(S) AND ADDRESS(ES)

= EOARD 223/231 Old Marylebone Rd.,

AFOSR/NE, L:.Col. L. Burggraf, DC 20332-6448

London NWI STH UK

AGENCY REPORT NUMBER

Bolling AFB

11. SUPPLEMENTARY NOTES

1 20. 0t5TIEuT!CN ::DE

STATEMENT

12a. OISTRISUTION/AVA)LABWLITY

Unlimited.

13. ABSTRACT (Maximum 200worts)

The electrophoretic deposition of low expansion ceramic materials (AlTiO5 and HfTi04) on porous graphite and their penetration into the pores was studied. Penetration is

1

enhanced by large Pe numbers and is

therefore higher in aqueous

suspensions while coating morphology is improved in ipropanol suspensions. Optimal conditions for coverage and penetration in a two-stage process were determined. The feasibility of subsequent electrophoretic deposition of two different layers of materials with zeta potentials of the same polarity was demonstrated for Si 3 N4 on Al 2 TiO5 (cathodic deposits) and SiC on glass ceramic (anonic deposits). Layer thicknesses in

14.

these experiments were in

the range of 100-300 ;j .

Electrolytic Deposition, Zirconia, Alumina 17.

In

parallel studies on electrolytic deposition from ionic aqueous solutionsthe deposition of A12 03 and co-deposition of A120 3 +ZrO2 from solutions of AI(NO 3 ) 2 It was shown that the atomic ratio and ZrO(N0 3 ) 2 were demonstrated and studied. of A!/Zr in the deposit can be varied by varying this ratio in the solution. The as-denosited coatinas are amorphous and became crystalline after heat treatment. ¶5, NUMBER OF PAGES SUBJECT TERMS Electrophoresis, Ceramics, Penetration,

SECURITY CLASSIFICATION OF REPORT

Uncla3sified

!8.

SECURITY CLASSIFICATION OF THIS PAGE

Unclassified

19.

16. PRICE CODE SECURITY CLASSIFICATION OF ABSTRACT

20- LIMITATION OF ABSTRACT

Unlimited

Unclassified

C "SS, 0 ,narard ,.40

298 'Rev L11S

'-89)

2

ELECTROPHORETIC AND ELECTROLYTIC DEPOSITION OF CERAMIC PARTICLES ON POROUS SUBSTRATES L.

S.

Haber,

S.

Gal-Or,

Liubovich

Summary of Annual Report

1 July 1990 - 30 June 1991

INTRODUCTION The present

phase of the research project concentrated on the

following aspects: -

Surface

deposition

ceramic materials -

such

as

2

TiO5

ceramic and SiC. Electrolytic deposition

of

co-deposition of -

penetration

graphite. Multilayer electrophoretic ceramic materials:

-

and

Al

two

Al

2

TiO5

a

S 3 N4 single

oxides

low-expansion

and HfTi0 4 on porous

deposition and

of

of

two

as

well

oxide

(Al 2 0 3 +ZrO 2 )

ionic soltions. Theoretical analysis of particle penetration

The interest in

the depositAon of

impregnation

appropriate ceramic at elevated

of

materials

temperatures.

as glass

(A1

2

from

03)

and

aqueous

into pores

low-expansion ceramics stems

from the intention to apply electrophoretic coverage and

different

deposition for the

carbon-carbon

composites with

so as to prevent A1

2

TiOs

its oxidation

and

HfTi0

4

are

characterized by low thermal expansion coefficients

similar to

that of

as coating

carbon-carbon

materials may

provide

.

the

Therefore,

their

use

neccessary compatibility with the

3

the

substrate from

of

view

of

expansion.

thermal

known that the expansion of these materials is

is

it

However,

point

anisotropic and

therefore

has been realized for some time that

It

place upon sintering.

microcracking of the coating takes

the complicated problem of carbon-carbon protection will have to be solved by the application of several different coatings. It

was thus considered of great importance

electrophoretic Si3N4 on

to study sequential

deposition of another ceramic material

such as

the Al 2 TiOs as a remedy of the detrimental effect of

microcracking. The above mentioned studies were performed on the basis of the conclusions from theoretical

the

was

now

deposition extended

experimental

Thus

studied and

then

the

The

effect

demonstrated.

the

morphology

of ZrO2 , studied in

to

oxide-Al 2 03.

kinetics,

basic

studies of the deposition of colloidal

The electrolytic year,

previous

its

codeposition

deposition

of

and

of

of

deposition

composition

of

S10 2 .

the previous with

another

only was first

A1 2 0 3

feasibility

and

co-deposition parameters the

was

on the

coatings

was

studied. The penetration

of particles

it

pores

which

analyzed theoretically

(an

consists

of

into a porous substrate assuming are

closcd

open

at the far end was

interconnected

porous

structure was considered previously).

EXPERIMENTAL The electrophoretic out on

porous

and

electrolytic

graphite substrates

porosity average

pore

size

60 it).

depoqition was carried

(UCAR

grade 45,

Initial

48% volume

experiments with

C-C were performed on a 2-D composite supplied by the USAF. AI 2 TiO 5 , prepared BaCO3

addition

HfTiO4 was

at

and

acquired

average size of 70 u.

the Israel an

average

from

Cerae

Ceramics Institute,

had a 5%

particle size of 1.4 u. Corp.

The

and had originally an

4

Si 3 N 4 particles

SiC>40 u.

were

the

in

range of

0.5-1.0

u and that

of

Suspensions of the ceramic particles were prepared

in isopropanol,

water

and

mixtures

Particle concentrations were 30,

of

the

two

100 and 250 g/l.

solvents.

the electric

field was in the range of 5-300 V/cm and deposition t.mes were 30 sec-6

0

min.

The amount

of

impregnated

material was determined by weight

change after removal of the external deposit. The multilayer substrate,

deposition was carried out by immersion of the

coated

with

the

first

the second material while still

layer in

the suspension of

wet.

The composition and morphology of the multilayer deposits were studied on crossections obtained by cutting coated specimens that were

infiltrated

with

an

epoxy resin and cured for 10

hrs. These studies were performed with a JEOL 840 SEM equipped with an EDS unit. The electrolytic

deposition was carried out from solutions of

0.02-0.2M AI(NO,),

with and without 0.01-0.05M ZrO(NO 3 ) 2 .

The deposition was done at constant current with c.d.'s range of 15-60 mA/cm2

the

and deposition times of 20 sec - 40 min.

Some deposits were heat-treated at 1200"C for 1 hr. and microanalysis

in

were

studied in

Morphology

the JEOL 840 SEM and phase

content with an x-ray diffractometer Phillips Model PW-1730.

SUMMARY OF RESULTS In contrast studies it the cathode

to was both

positive charge point since

the

SiO

2

particles

deposited

in

previous

found that both Al 2 TiOs and HfTiO 4 .deposit on from

water and ipropanol

suspensions.

The

on the particles indicates a high isoelectric

the pH of these suspensions was determined as

8.1

in

water

SiO2

and

6.2

deposition,

the porous However,

in

it

propanol.

As in the previous study of

was found that

substrate

is

penetration of AITiO 5

higher

in

water than in

the addition of x20% ipropanol

maximal penetration

probably

due

to

ipropanol.

to the water lead to a better

wetting of the

graphite as demonstrated in simple wetting experiments. 6% of

substrate weight of Al

at 100

V/cm,

however,

is

due to

30

g/l

2 TiO5

after

lower

About

were induced into the pores

60 min.

The surface deposition,

both higher and morphologically

a

into

better

fraction of penetrated material

in

propanol

and lower H,

formation on the cathode. It

was

thus

penetration

concluded

a

two

impregnation from relatively

low

ipropanol min).

at

It

that

stage an

duration

was

process

aqueous

shown

that

coverage

would

be

75-100 V/cm)

(200 V/cm)

deposition

optimal

and

employed:

solvent with 20% ipropanol

(60 min,

higher field

intensity and

for

and coating in

and shorter duration

impregnation time.

The

for

(x2

increases with field

morphology

of

surface

deposits and impregnated crossections were studied. The feasibility

of

subsequent

deposition

of

two different

layers of

materials with zeta potentials of the same polarity

was shown

for Si

glass ceramic out

by

3

transferring

first

in

However, layer

these experiments

The co-deposition

Al(NOj)

3

the

from

is

carried

the

specimens with an Epoxy

analysis

and

first of the

observed. were in

of

two

of

the

The layer

resin and

Si and Ti x-ray imaqes

the SEM and the existence of

solutions

÷ZrO(NO3 ),

deposits.

of

penetration

was

from aqueous

substrate

and SiC on

second to continue the deposition

Elemental

were performed shown.

coated

The deposition

Crossections of the two layers were prepared

impregnation curing.

(cathodic deposits)

deposits). the

the

second material.

its

on AITiQ5

(anodic

suspension into

after

N4

Si

3 N.

the two layers

and SiC into the

thicknesses obtained

in

the range of 150-300 u.

oxides containing

was demonstrated

by an electrolytic salts

cf

process both

and studied for Al

Prior to that the deposition of AIO

3

cations 2

O3'ZrO,

was studied.

6

The effect of field intensity on the penetration in the optimal solvent is seen in Fig. 4. The penetration increases with field intensity in the range tested (further increase in field intensity caused excessive beating of the suspension and was therefore not studied). A deccelerating effect of deposition time on penetration is seen in Fig. 5. Figs. 6 & 7 describe the morphology of the coating and penetration. The surface deposit of AI 2 TiOs is seen in Fig. 6. The coverage is uniform on a macroscale with a high degree of roughness reflecting the porous surface of the substrate. It was found that a denser and more uniform deposit is obtained in ipropanol at high field intensities and short deposition times. A crossection of a penetrated specimen obtained at optimal penetration conditions compared with a reference untreated specimen is seen in Fig. 7.

b) Multilayer Deposition

A schematic description of the multilayer deposition is given in Fig. 8. A deposit of particles A is formed in suspension A, the coated specimen is removed from the suspension and while still wet introduced in suspension B. A deposit of particles B is formed. A more thorough study of this two-stage deposition process should be done in the future. So far, two systems were deposited: Si 3 N4 on top of Al.TiO, and SiC on a glass ceramic deposit. In each system the zeta potential of the different materials had the same sign, Si 2 N4 and A12 TiOs deposited both on the cathode, while SiC and the glass ceramic migrated to the anode. When couples of materials with opposite zeta potential were studied, the first deposit migrated away from the substrate when its polarity was changed so as to deposit the next layer. Yet conditions can be envisaged where such two-layer deposition can still be performed and these will be studied in a future work. Crossections of the two layer deposits are seen in Figs. 9-12. Fig. 9 shows an optical micrograph of the layer of AlaTiO and Si2NA as well as an SEM x-ray image of Si and Ti at a higher magnification. The overall layer thickness is '350 um. The element Si is seen in both layers with a

much higher existence in

Its in the outer layer (the Si 3 N4). density the inner layer (the A12TiO5 ) is explained by

into the initial porous Si3 N, some penetration of The element Ti, however, exists layer. ("green") Al 2 TiOQ An interesting the inner layer. as expected only in the "pegging" of the inner layer in the pores feature is The thickness of the SixN4 is &200 um of the substrate. Composition graphs of the and that of Al 2 TiO5 t150 um. 11 showing again the the two layers are seen in Fig. penetration of Si into the underlying Al 2 TiOs layer. An inner Two anodic deposits are shown in Figs. 11, 12. glass ceramic layer (a300 u) with an elemental composition Fig. 11a and an x-ray image characterized by the seen in An outer layer of SiC (f150 v) with traces of element K. The x-ray image (Fig. 12b) shows Fe (from the anode). again the penetration of SiC (characterized by the element The Si) into the underlying layer of the glass-ceramic. amount of Si 3 N4 deposited on AI 2 TiOs as function of deposition time was determined with the intention to study layer on the of the underlying AI 2 TiO, the effect It is seen that the (see Fig. 13). deposition of Si3N4 existence of the underlying layer does not affect the deposition kinetics probably due to the high porosity of "green the in still is which layer the first state"

CONCLUSIONS A low expansion ceramic material-aluminium titanate (A12TiOT) deposited and penetrated into porous was electrophoretically The Al 2 TiOs which is positively charged deposits graphite. on a cathodic substrate at a pH=8.1 indicating a relatively amount of material that The point. high isoelectric in an aqueous solution of 20% maximal penetrated was an outcome of the balance This composition is ipropanol. between the high E/7 ratio of water and the improved wetting of the substrate in the presence of ipropanol. Penetration increases with electric field intensity in the For optimal coverage range studied and with deposition time. and penetration a two stage process should be applied: a 20% ipropanol solution in water.(at 100 impregnation in 6% of substrate weight of AI 2 TiOs is about g/l V/cm, 30 penetrated after 60 min.).

8

-

Surface coating in ipropanol at a high field strength (200 V/cm) and short duration ý2 min.).

At optimal penetration conditions the whole crossection of a 7 mm thick substrate of porous graphite was penetrated. Multilayer deposition of different ceramic materials is possible by subsequent deposition from different suspensions of materials with zeta potential of the same polarity. This was demonstrated by forming an inner layer of AI 2 TiOs and an outer layer of Si 3 N. (cathodic deposits) as well as an inner layer of a glass -eramic and an outer layer of SiC (anodic deposits). The individual layer thicknesses were in the range of 150-300 u.

In the range of thickness discussed, the existence of the underlying "green" layer did not affect the kinetics of deposition of the external layer.

ACKNOWLEDGEMENT

Research sponsored Research,

Air

Force

the

by

Air

Force

Systems Command,

Office U.S.A.,

of Scientific under Grant No.

89-0474. The U.S.

authorized to reproduce and distribute Governmental _ purposes not withstanding any

Goevernment is

reprints for

copyright notation thereon.

9

LIST OF FIGURES

Fig.

1:

Ceramic thermal expansion characteristics 14).

Fig.

2:

Amounts of Al 2 TiO5 deposit.d and porous graphite as function concentration in water.

Fig.

3:

Model for a two-stage deposition process.

Fig.

4:

Amount of AlTiO5 penetrated as function of field intensity (20% ipropanol in H2 0, 60 min, 30 g/l).

Fig.

5:

Amount of Al 2 TiO5 penetrated as function of deposition time (20% ipropanol, 50 V/cm, 30 g/l).

Fig.

6:

Surface deposit on graphite obtained in at 200 V/cm, 2 min, 30 g/l.

ipropanol

Fig.

7:

Crossections of treated (20% ipropanol, 60 min) and untreated specimens.

100 V/cm,

Fig.

8:

Schematic description of multilayer formation.

Pig.

9:

Crossection of a double layer deposit of Si 3 N4 (xl0) (100 g/l, ipropanol, 50 V/cm, 2 min) and AI 2 TiOs (30 g/l, ipropanol, 200 V/cm, 90 sec) and x-ray images (xl00).

penetrated in of ipropanol

Fig. 10:

Elemental composition of the two layers a) Si 3 N4; b) Al 2 TiOn

Fig. 11:

Crossection of a double layer deposit of SiC on glass ceramic (xl0). SiC (100 g/l, ipropanol, 50 V/cm, 15 sec), glass ceramic (250 g/l, ipropanol, 200 V/cm, 10 sec).

Fig. 12:

Elemental composition of the two layers b) glass ceramic (i100).

Pig. 13:

Amount of SiN, d..positep function of deposition time.

nn

A)-TiO,

a)SiC

170 0) as

10

BIBLIOGRAPHY

1.

L. Gal-Or, S. Liubovich, S. Haber, "Deep Electrophoretic Penetration and Deposition of Ceramic Particles Inside Porous Substrates - Experimental", J. Elc'trochem. Soc., in press.

2.

L. Gal-Or, S. Liubovich, S. Haber, "Deep Electrophoretic Penetration and Deposition of Ceramic Particles Inside Porous Sunstrates An Analytical Model", J. Electrochem. Soc., in press.

3.

D.W.

4.

I.

5.

K.L.

6.

J.R. Strife, (1988).

7.

R.W. Lynch, (1972).

8.

G.A.

McKee,

Jawed,

Carbon,

24,

D.C. Nagle,

Luthra, Carbon,

Parks,

J.E.

B.

Chem.

737(1986).

Mat.

Res.

Bull.,

21,1391(1986).

26,217(1988).

Sheenan,

Morosin,

Rev.,

J.

Ceramic

Am.

65,177(1985).

Bulletin,

67,

369

Ceram. Soc.,

55,

409

TEMPERATURE (*0

Soo

200

2000

1500

1000 STABIUZED HfO2

1.5

Alz03

T602 STABIUZEO0T0 rz r

1.6

Y203 1-4

"'B2 12

I

!2 1.0

T

0.8

06

12SO

0.4

0.2

rig.

80

C.D C.)

0 00

L0Lo

_

CVto0

CCto



I

00

TWO

STAGE

_

PROCESS:

_

I

,

I

I)OPTIMAL Pe>>

CONDITIONS FOR IMPREGNATION: Pe = (U )-r U' =

•H0= 81 b.,H2 0 = Ilcp

•prop=

18.3

I•prop= 2 cp

20% propanol in H2 0 II)OPTIMAL CONDITIONS FOR COATING: 100%

propanol

. less penetration of particles

Fig.

3

{N I-

-\

E

---

o0( C)

E oi

I

9 O!IL•IV

%

oo• 00

0

00

,\

to

LO

LOO

i o

-,-

--

-

LOO!JCi

%

9OLL

toV %

If

0C

Fig 6

Fig 7

A

C

-.

,~4b

AC

A

A

.K

Substrate Coated with A

Fig 8

A

Substrate Coated with A and B

Si 3N4

on

,AI2Ti05

X-ray

Si

X-ray Ti

Fig 9

:Ps

,

.-

0.Q'.

R

,

b

~7

if

5.

>

iK

'.

,I

•ii

iX-RAY: jL ive:

-

1100: P r se t:

I.,

EMi :•SI34

Fig 10

o'

R e ma i n i n:

Os'

Fig I

L::,-R t,-: :Live-:.

I0 - ,Fre k).t.:. 1,cr',f-.

loft-

R'~na ir n :

i.i

I I

SI..

X-ray of Si Iit

: •.

i

1 00

-..

FS= 4-K tIEM1 :'SIC:

,L

1.~t.

.•:

z F'" -,. r.:

11 .2

V

,h

31 5=

" LC"

-"

110CI

-C•

-F

'ItJ

It

X-ray of K

':"

. .

,.:|,

'A u

1 12

Fig 12

>;

z

EE

ES' >()

CE,

Cl

'nVon

CDJ

Ul)

m~

0 cn

1

0

04J

04

0

LO

SINGLE AND MULTILAYER ELECTROPHORETIC DEPOSITION AND PENETRATION OF LOW-EXPANSION CERAMICS ON POROUS SUBSTRATES

L.

Gal-Or,

S.

Liubovich

Institute of Metals - Technion, of Technology

Israel Institute

ABSTRACT

Previous electrophoretic deposition and penetration studies were further extended to the deposition of a low expansion ceramic material

-

A1 2 Ti0 5

multilayer deposition.

AI 2 TiOs into However,

the

maximum

propanol in

on

graphite surface.

porous

graphite

and to

It

was found that the penetration of

pores is

higher in water than in propanol.

penetration is

water

efficient and

-

probably The

due

surface

obtained in

a 20% mixture of

to enhanced wetting of the deposition

morphologically better in

is

both

propanol.

Thus,

more for

optimal average and penetration a two stage process should be employed. The feasibility

of

subsequent

deposition

of two different

layers of materials with zeta potentials of the same polarity was shown for deposition of SiaN4 on AI 2 TiOs (cathodic deposits) and deposits). hundreds of

of

SiC

on

Crossecttions

a of

glass

ceramic

layer

(anodic

the two layers in the order of

microns were prepared and elemental analysis and

x-ray images are shown.

INTRODUCTION Basic principles and parameters of electrophoretic deposition of ceramic particles in the pores of a conductive porous substrate, were described in two pievious publications (1,2]. These reports summarized studies of deposition of colloidal Si0 2 from aqueous and nonaqueous suspensions on porous graphite. It was shown, experimentally 111, that penetration is enhanced by a high ratio of dielectric constant to viscosity of the solvent as well as by increased particle concentration and electric field. At optimal conditions the whole crossection of the porous graphite substrate was penetrated and a skeleton of SiOa with the same shape and dimensions as the original graphite specimen was obtained after burning off of the graphite. A theoretical analysis [2] of the penetration enables to predict the penetration depth as function of two non-dimensional parameters, the Peclet number (Pe) and the deposition rate (A). Large Pe and low X enhance the penetration. In addition to the above mentioned results the feasibility of deposition of a variety of other oxide and non oxide ceramic materials (fused SiO2 , SiC, SiN) was demonstrated. The present report is an extension of concetrating on the following two aspects: Surface deposition ceramic material -

and

penetration

of

these a

studies

low expansion

- A1 2 TiOs - aluminium titanate on porous

graphite. Multilayer electrophoretic deposition which includes deposition of two consecutive layers of different composition: AI2 TiO 5 and Si 3 N4 as well as a glass ceramic and SiC.

The interest in the low expansion ceramic and the multilayer deposits stems from the intention to apply electrophoretic deposition for the coverage and impregnation of carbon-carbon composites with ceramic materials so as to prevent oxidation at elevated temperatures. Carbon-carbon (C-C) composites are of great interest for aerospace applications since they exhibit excellent high-temperature mechanical stability and low weight. However, they are characterized by a residual porosity and low oxidation resistance in air at temperatures above 500"C [3-5]. Another feature of these composites is their low thermal expansion coefficient (1-3 x 10-/*C). The ceramic protective coating should therefore be characterized

3

by a low expansion coefficient so as to prevent its cracking during use. (Al 2 TiOS) has one of the lowest thermal expansion coefficients as seen in Fig. 1 (61. Therefore, it is a plausible candidate for the application mentioned above. However, the expansion of this material is anisotropic and therefore microcracking will occur when applied as a coating (7]. The detrimental effect of the microcracking on the oxidation protection can be remedied by another coating deposited on the Al 2 TiOs such as Si 3 N4 . A1 2 0 3 -TiO 2

In addition to the practical interest described above the study of sequential electrophoretic deposition of different ceramic materials is of interest for understanding the process and widening its scope. Thus another system (glass ceramic and SiC) was also deposited.

EXPERIMENTAL PROCEDURE The deposition and impregnation were performed on porous graphite - UCAR grade 45 with a 48% volume porosity and an average pore size of 60u. (The porous graphite serves as a model material for the C-C composite). The specimens were 20x20x7 mm. rhe ceramic materials deposited were: -

A1 2 TiO5 prepared at the Israel Ceramics Institute with 5% BaCO3 added for sintering enhancement. After attrition milling the average particle size was 1.42 u.

-

Si 3 N4 -0.5 V. SiC - 600 mesh ( N02-

+ 20H-

Dissociation of the salts: - AI(NO)

A A>

-

3

- ZrO(NO,) 2

3)

40H-

+ 4e ->

+ 2H 2 0

02

- NO"

2)

expected to be the following:

->

+ 3NO+

ZrO2

-

+ 2N03-

ZrO2° + H2 0

Interaction of hydroxyls with the cations: A1

3

-

+

Zr(OH)

2

30H-

->

2

20H-

'

+

AI(OH)3

->

Zr(OH) 4

--

Zr(OH)

22

3

4)

Dehydration of the hydroxides: 2AI(OH)

A

3 -

Zr(OH)4

2

+ 2H 2 0

Zr0 2

->

The hydroxides

+ 3H2 0

03

form

thus

electrochemical generation

of

in

two

OH-

and

between the cations and the hydroxyls. description of the co-deposition is

successive a

chemical

steps: reaction

A schematic

given in

Fig.

1.

The yield of the deposit depends on the faradeic efficiency of OH- formation (affected by reduction reactions that do not result in

OH-

the fraction substrate.

formation such as H- + e -)

of

hydroxide

The

latter

formed

parameter

diffusion coefficients of OHZr(OH) 2 2 ÷ on deposit will A13 - and

H2 and

at sites removed from the depends

on

the

relative

on one hand and that of A13 - and

other.

The ratio of A12 0 3 and ZrOa in the depend on the relative diffusion coefficients of

Zr(OH)22-

generated at of the

the

H H + H ->

the

cations

which

compete

for the OH-

ions

cathode,

and the solubility constants (K.,) two hydroxides formed. A quantitative analysis of the

kinetics of deposit formation and its take into account diffusion phenomena hydroxide formation.

all these factors assuming that are the rate determing steps in The

the the

present paper describes the effect

of electrochemical morphology and

composition will have to

deposition composition of

parameters the

A1 2 03

on

the

and Al2 03

rate, +

ZrO2

deposits. EXPERIMENTAL Deposits were obtained on graphite specimens (UCAR45) 2 x 10 x 20 mm. They were polished with a 1000 grit SiC abrasive paper, then rinsed with ethanol in an ultrasonic bath, washed with distilled water and dried in 0.02M - 0.2M Al(NO0).

air.

Electrolytes were:

with and without

0.01M - 0.05M ZrO(NO3 ) 2

4

Deposition was

carried out at constant current with a Horizon

Model SR-365 d.c. 14 x

28

mm

were

measured with experimental the

AVO

Cell

meters.

set-up is

voltages A

seen in

of

Fig.

2.

curves were

stirrer

obtained

Deposits were an accuracy 1200"C for

with

dried in of

1

0.05 hr in

during a

air

currents were

drawing

of

the

Current densities were

and deposition times were

20 sec - 40 min.

magnetic

and

schematic

the range of 15-60 mA/cm

range

with a

used.

Two counter anodes made of Pt

2

varied in in

power supply.

Mild stirring

was applied

deposition.

Polarization

PAR

Model

270

potentiostat.

and their weight determined within

mg.

Some deposits were heat treated at

argon.

Morphology and microanalysis were

studied with a JEOL SEM Model JSM 840.

X-ray diffraction was

performed with a diffractometer Model PW-1730 operated at 40 KV and

40 mA using Cu k-alpha radiation and scanning speed of

1.2 deg/min.

RESULTS AND DISCUSSION Cathodic polarization solution of Al(NO)) first

and

-1000

region is

mv SCE in

A limiting

current

of

setting in

at

same

the

similarity in

by the

seen (down to -800

place

Fig.

3a,b.

mV SCE in

a In

the

the second solution respectively). mA/cm2

is

seen

potential

of

-2000

u10

the two curves is

reaction taking current being

an AI(N0 3 ) 3 solution and in

and ZrO(NO 3 ) 2 are given in

3

both curves a Tafel

curves in

in

in

both curves

mV

SCE.

The

a result of the same cathodic

both

systems

with

the

limiting

determined by the hydroxide deposit rather than

concentration

of

the

NO3-

which

is

higher in

the

combined solution. Fig.

4

presents the variation in

time for two current densities.

cell voltage with deposition

At the higher c.d.

(25 mA/cm2 )

the voltage increases steeply for the first 3.5 min as a result of the build-up of the hydroxide deposit and then levels off. weight of

Yet,

as

seen

in

the deposit continues

Fig.

5 for the same c.d.,

the

to rise slowly after 3.5 min.

5

The cell

voltage

local breakdowns

stability in

deposition

significant

increase

2

mg/cm /10sec) change in

form on

then

breakdown.

The

lower c.d.

(20

voltage takes

.

A

its

As

different,

the

the

of the A1 2 0,

stage

due

to

the

a monotonic

as-deposited, in

on the deposit

increase in

is

seen in

deposit

Fig.

(1G min). Figs.

consisting

6a.

A

seen in

very Fig.

6a of

dense

6b.

Two

microcracks that formed during

non-uniform contraction of the wet

bubbles attributed to H2 formation as a result of

H, reduction.

In

preliminary studies

aim of controlling

the density

of

drying rate is An attempt

cracks

to thick

current.

Pulses

overcome

the

formation of defects and yet to

deposit

was

made

of

voltage

reached when

was shown that

can be reduced significantly when the

by

the

use

2 min duration at 25 cA/cm'

with intermissions of 5 sec (see Fig. increase in

(to be published later)

the drying rate it

lowered.

obtain a

obtained.

by

entire time interval

seen

ions

voltage/time curve at the

close-packed

is

depends

created

the individual clusters is

coating and with the

build-up

sites

is

for

The

the nonconductive deposit

further

kinds of defects are also seen: the drying

conductive and the OH-

mA/cm2)

relatively

structure of

20 secs (:0.3

explained by the fact that at the

of

individual clusters

The slope of the

the first

behaviour place

without further

slows down to about 1/10 of that.

reduction

The morphology

continue

for

surface.

substrate

of reduction which

cell Voltage.

high

rate is

entire

availability of

& b

to

whole susbstrate is

the

covers the

the

very

deposits

beginning the

process in

is

and

explained by the formation of

the film creating sites

enable the

deposition curve

is

up

applying

to

a

7).

were applied

A relatively smooth

much higher voltage than that

a continuous current

Accordingly,

of a pulsed

(25V

vs.

08V)

was

the deposit was thicker and smoother.

This effect can be explained by the possibility of creation of new nucleation

sites

for deposit formation during the current

intermission. X-ray diffraction deposits fired

patterns at

1200"C

of

dry as-deposited coatings and for

1

hr,

according

to

the

6

recommendation of (dried in

(5),

are seen in

Fig. 8.

The as-deposited

air)

coating does not show any diffraction pattern except that of the graphite. Crystallization of

the

deposit

takes

place during the heat

treatment with the formation of a aA1 2 03 . The crystallite size, as calculated with the Scherrer equation (6) from the line broadening effects exist

of

the

due

to

110 the

reflection, substrate,

assuming no strain

was found to be about

1oam. Deposition from solutions ZrO(NO 3 ) 2 was carried out 14"C. The dependence seen in Fig. 9. A place in

spite

deposition of

of

containing at two c.d.

both AI(NO 3 ) 3 and (30 and 60 mA/cm2 at

of cell voltage on deposition time is more gradual rise in the voltage takes the

A1 2 0 3

higher only

c.d.s when compared with the

(see

Fig,.

4).

A

possible

explanation is the difference in deposit properties as a result of the presence of Zr(OH) 4 . From our previous work (3) it can be inferred that the voltage reached prior to breakdown onset is much higher for Zr(OH) 4 , at the same c.d. and so is the deposit weight per unit area. The AI(OH)a forms probably a denser Zr(OH)

4

film is

The phase

at lower thicknesses and thus is

modified when

codeposited. structure

of

a combined deposit was studied after

forming at 1200"C for 1 hr. The x-ray diffraction spectrum shows the presence -of two phases: tetragonal ZrO2 and %A12 0 3 (Fig. 10). The were studied in

morphology of the deposit and microanalysis the SEM by x-ray imaging of Al and Zr and

local elemental

semiquantitative

the x-ray

analyses.

Fig. 11-12 shows

image of Al and Zr in a combined deposit along with

the spectrum

and

semiquantitative

results

of the Al and Zr

content. The relative solution is

of the A1 3 - and ZrO 2 - ions in the to affect the deposit composition. The

concentration expected

latter was therefore studied (in terms of AI/Zr atomic ratios as function of the same ratio in the solution). Fig. 12 shows

7

this

dependence.

concentrations solution, can be of the

It in

is

the

seen

deposit

hence the Al(OH)3

is

that

is

two

ions

diffusion rate

A1

and,

hydroxide formation.

higher

preferentially

explained by the differences 3

much

the

in

Al

than in

the

deposited.

This

diffusion coefficients

and ZrO2 with the Al- 3 therefore,

relative

having a higher

being more available for the

LIST OF FIGURES

Fig.

1:

Schematic description of the co-deposition of ZrO2 +AI 2 0

3

.

Fig.

2:

Experimental set-up for deposition.

Fig.

3:

Cathodic polarization curves in a.

0.05M ZrO(NOa)

2

b.0.O5M Al(NO3 )3+0.05M ZrO(N0 25"C, 3 mV/sec, vs. SCE. Fig.

4:

Cell voltage

as

function

densities (0.05M AI(NO3 )3, Fig.

5:

6:

),

of

time for two current

10 min,

25"C).

Deposit weight as function of time (0.05M AI(N0

Fig.

3

3

)j,

25 mA/cmr,

25"C).

A1 2 0 3 deposit from 0.05M AI(NO3 )3, 25 mA/cm2,10 min.

at

a) X75

b) X750 Fig. 7:

Voltage/time curves for deposition of A1 2 0 3 with pulsed current: 0.05M AI(NO)) 3 , 25 mA/cm 2 , t..=2 min toff

Fig.

8:

- 5 sec,

25'C.

X-ray diffraction spectra from: a) as-deposited A1 2 0 3 b) after firing at 1200'C for I hr.

Fig. 9:

Cell voltage

as

function

of

time for two current

densities for deposit formed from 0.05M Al(NO3 ) 3 +0.05M ZrO(NO 3 ) 2 at 14"C, 60 and 30 mA/cm 2 .

t=5 min c.d.

9

Fig.

10:

X-ray diffraction pattern of co-deposit.

Fig. II:

X-ray image and local analysis of a deposit formed from a 0.05M AI(NO3 ) 3 +O.OSM AI(NO0) 2 solution.

Fig. 12:

Atomic ratio of AI/Zr in deposit as functior of this ratio in solution.

10

ACKNOWT.FDGEMENT

by

Research sponsored Research,

Air

the

Force

Air

under Grant No.

U.S.A.,

Systems-Command,

Force

of Scientific

Office

89-0474. The U.S.

Goevernment

authorized to reproduce and distribute

is

withstanding

not

purposes

Governmental

reprints for

any

copyright notation thereon.

BIBLIOGRAPHY

1.

J.A.

2.

J.A.

Am.

Switzer, Switzer.

Ceram. R.J.

Soc.

Bull.

66,

MRS

Phillips,

1521(1987). Proc.

Symp.

121,

111(1988).

3.

L.

I.

Gal-Or,

Soc.

In

R.

Silberman,

Chaim,

Jour.

Electrochem.

press. Silberman,

4.

R.

Chaim,

I.

5.

A.

Ayral,

J.C.

6.

B.D.

Cullity,

Droguet, "Elements

Addison-Wesley Pub.

Co.,

L.

ibid.

Gal-Or,

Jour. of

Res.

Mat. X-ray

Reading,

4,

970,

1984.

Diffraction',

Mass.,

1978.

p.

254

NO 3 - Electrolyte e Substrate

NO 2 -

OH"

Zr(OH) 2++ = Zr(OH) 4 SJ4

(Cathode)

---- o-e

e

OH"

e

H20

H2 e.. 01-1F

Fig I

AI+3

AI(OH) 3

POWER SUPPLY - -(.'*)

-

23 45-

Power supply Ammeter Voltmeter Solution Holder

67 -

Graphite specimen Electrodes - Platinum

89-

Stirrer Plate

5

(-)

"

-

"

••4

-

-

-

-_

Fig 2: Experimental Set-up For Electrolytic Deposition

-

|

a 5 ' " . ..

.......

,J 10

I

,i , S,"'1

S ii " i...1

"'

*

y I SJI""l ...

,

..

z

-2400

b

Zii: 0 10

I%,a

0

'

1.0 C0 t( UA/CM^2)

. Is '

*I

..

I

"

'4 10

10

..

K,

0

iE

5-40

, ,,.... a

-. 0

10

Fig 3

10

I

, , , ,,,,I

1

2 10..

... , , , ...5! I0

, , , , .. S.I, '4 10

,. . 10

(Njj

ci

L'1

30VI -10A

00

Ln

CI00 (Z3/AlN)i

H

dci

a

b Fig.

6

Ln

jcLLJ

-9l

L

I

I

30V~lA i

CIO

IDO -j

A jb&) P)~ 7

(11i 047t LO I~7b

Co

1D LL)

ci I. cx r':

Lr") Cl% cr)

'O

a

Ia

c

Kz>e Go

LL.

Ein

U,

z Lii

AL

14

(A)

ilO

t

(vzLT0) 1V.J

(L~~

(ZOZ) I z
100). Cases in which X = 0 would obviously result in the deepest penetration. It rmght be achieved by introducing a repelling force between the particles and the pore walls, stronger than van der Waals forces. The asymptotic value of the cumulative probability for large values of z is always unity, since after a long time the probability that a particle has been deposited anywhere over the pore walls is unity (the particle cannot disappear). Figure 7 summarizes the foregoing results. We define the penetration depth as the zlb value for which P, - 0.9. In other words, 10% of the particles would pebetrate deeper than the value shown in the figure for a given set of parameters, the Damk6hler number X,and the Peclet number P,. Obviously, the 0.9 numerical value chosen for P, is quite arbitrary and a lower or a higher percentage could have been chosen. The basic interpretation of the results, however. .,ould remain unaltered. It is obvious from Fig. 7 that the penetration depth for a given value of X grows almost linearly with the Peclet number (which is not at all surprusing). The slope of the curves depends strongly on X,;the smaller the 4, value the deeper the particle penetration. Conclusions The following general conclusions can be drawn from the foregoing analysis. Deep penetration and coating is feasible by electrophoretic processes. Enhanced penetration is obtained for large electrophoretic Peclet numbers and small Damk6hler numbers Electroosmosis (directly dependent on the electroosmotic permeability -f and thus on the sign of the wall zeta potential) increases the Peclet number and thus penetration only if the electroosmotic

-"

4

2

6

8

10 1

Pe Fog 7. P"s deo (deflee Wthe do•irm "Ave ef 11b fo whid the is 90% piabildhy tht the pwmete ho dwmowd ow a Y le" time) v. t tro Pelet au. (t Dankair namber is use as a Parameter). and electrophoretic velocities coincide in direction. Le. 4w and 4, possess the same signs. The diffusion coefficient of the colloidal particles has a dual effect. Decreasing diffusion coefficient causes increasing Peclet number which has a favorable effect on penetration depth. Decreasing diffusion coefficient causes increasing deposition rate which has an uniavorable effect on penetration depth. The process can be controlled by a judicious selection of part. dce size and concentration, solution properties, electric field intensity and its time protocol Quantitative comparison with experimental results that we presented in Part 11 is not yet possible since only global results of total mass deposition inside the pores was measured and not how mass is distributed inside the porous slab (which is the main result we presented in this paper). Further experimental work is required to analyze this very important feature of deep electrophoretic deposition. which will determine its advantage over other coating processes. On the other hand. the total of mass deposited L side the porous slab depends on the rauo between the flux of particles depositing over the exterior surface of the slab and flux of particles penetrating it. Thus. further theoretical work is required to determine this ratio dependence on the process parameters before comparison can be made with the experiments conducted in Part IM Notwithstanding, several recommrnendations based on the theoretical results were made and found in qualitative agreement with experiment, e.g. the electroosmotic permeability depends Linearly on the ratio rv• which depends solely on solvent properties. Enhancement of penetration is predicted to occur if this ratio is high. Several liquids were examined (water, isopropanol, ethanol and pentanol) and water nossessing the highest ratio has shown the best results (Fig. 9 in Part ID. A rather limited quantitative comparison with Fig. 9 in Part II can be made. We calcu. late the ratio of SiO2 percentile of particles deposited inside porous graphite and migrating in a certain solvent to SiO%percentile of particles migrating in a given solvent, say, water. solvent

y(theory)

y(expenment)

0.052 0.112 0.283

0.078 0.29 0.42

Pentanol Propanol Ethanol where

%SiOsolventl " %SiO~water) The foregoing table reveals that the trend predicted in this paper is defirutely correct and there is a fixed bias which can be attributed to inaccuracies in measurements (difficulty in removing the deposit over the extehor sir-

J. Electrochem. Soc., Vol. 139, No. 4. April 1992 C The Electrocrhemical Society. Inc. face of the graphite slab), variations in zeta potentials and mathematical model simplifications. Acknowledgment Research sponsored by the Air Force Office of Scientific Research Air Force Systems Command U.S.A.F. under Grant No. AFSOR 89-0474. The U.S. Government is Authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright notation thereon. Manuscript submitted Feb. 15, 1991; revised manuscript received Oct. 17, 1991. APPENDIX The flow through porous media generated by an electric potential gradient is discussed in (28) and is briefly recapitulated. The governing field equations for the flow through a long circular tube induced by an electric potential gradient are based on Stokes' equations for low Reynolds number flows. iLe. V = Vp, V-v=0 [A-1] where v and p are the velocity and pressure fields and -n stands for the viscosity of the fluid. The boundary conditions which the velocity field satisfies over the cylinder walls are (17) f4. a6 , i,, = 0 at r = b

4"

-

[A-2]

where f is the dielectric constant of the fluid, ;,, is the zeta potential at the cylinder wall, d stands for the electric potential, b is the radius of the tube and (r, z) are polar coordinates. the z coordinate coinciding with the cylinder axis. It should be noted that the no-slip condition can no longer be applied and in its place we use condition [A-21. The solution for v is easily derived for a fully developed flow, namely, avpaz = 0. It is a simple superposition of the effects of the pressure gradient and the electric potential gradient. Integrating over the tube cross section. the mean interstitial velocity is obtained b, ap V,=

- -

-

8-q az

q. -

a'

--

am

[A-3]

Integrating over the entire cross section of the porous medium the superficial velocity is v=

81

az

where ov stands for the medium porosity. Since a homogeneous porous medium contains pores in all possible directions, a simular procedure can be applied for pores oriented along the x and t, directions. Expressions similar to [A-4] can be obtained for v and ý,, and fur, ther generalized to derive the following expression for v in vectorial notation [A-51

where

r=,

= FPj

such

as

the

silica: 2SiO-

b) adsorption mode,

+ H2 0

(negatively charged particles)

such as the adsorption of H- from water or

from weak acids: n AI(OH)

3

+HNO 3->[AI(OH)

3

1-,xH'+xNO-3

(positively charged particles). The surface

charge

opposite charge layer is

once formed is

derived

from

the

then balanced by ions of an solution and thus a double

formed.

Usually the

polarity

of

experimentally because spontaneously when

mixed

the the with

particles particles the

has to be determined acquire

solvent.

the

Further,

charge this

charge may be reversed upon addition of ionic compounds. According to constitutes of

the a

model

proposed

by

Stern

the double layer

rigid part with a linear potential gradient,

4

and a

diffuse part with a non-linear potential gradient termed "Zeta pqtential".

CHARGED PARTICLE ++ +

+

+

IONIC CLOUD

+

+ + +

+ +

+

+

+

0

+

+

++

POT ENTIAL

DISTANCE Fig. 1:

The double layer and potential gradient surrounding a charged particle.

The Zeta

potential

can

be

determined

by

two main methods:

microscopic determination of the velocity of individual particles using the Smoluchovski equation or in the Burton cell where a pure filtrate of solvent is poured on the suspension, a field is

applied

and

the

movement

of the sharp boundary is

monitored [2]. When an electric field is

applied to a charged particle it

will

5

move to

the

electrode with the opposite charge.

However,

the

diffuse double movement. The to which

it

layer around the particle interferes with its diffuse charge tends to move with the particle is attached but on the other hand it is influenced

by the electric field which pulls it The particle

apparently wins.

It

in the opposite direction.

moves through the liquid with

a diffuse egg-shaped double layer surrounding it (although it is not actually carry'ing the oppositely charged ionic atmosphere but rebuilding it ions in the the movement produces a

is in

rather

leaving

front as it

part

of

it

behind

moves along).

The tendency of the diffuse d.l. to move in the direction opposite to of the particle has an effect on the velocity. It "drag"

which

slows down its movement

(see Fig.

[3]:

DRIVING FORCE

RELAXATION FORCE

RETARDING

SFORCES

(ELECTRIC)

VISCOUS FORCE

L64-SHAPAe IONIJC CLOUD

Fig:.

and

Egg-shaped ionic cloud and forces acting on a moving particle.

2)

6

The mobility of the particle is derived by equating electric force with the frictional resistance and relaxation force. Thus the mobility U is given by: )=

EEý/4nX[l+f(kr)I

the the

where:

U

-

mobility

E

-

field

E

-

dielectric

-

zeta potential of particles

-

viscosity of fluid

K

-

1/double layer thickness

r

-

radius of particle

strength constant of fluid

f(kr) varies

between

0 and 1 for small and large values of kr

respectively. f(kr) represents the relaxation phenomenon, The Smoluchowski equation is

ii=(E-E-

Values of

ý

u

k [Al(OH)

3

1. XH'+XN0

3

placed in water an electric charge develops on it from the transfer of ions between the surface and the liquid balanced by ions of an The surface charge layer is phase. opposite charge in a diffuse layer on the solution side of the The two layers constitute the electric double interface.

When silica

layer.

is

The

potential

just

potential outside

constitute the charge of sign of the surface charge.

is the

a

measure

layer

of

the surface;

of

the

adsorbed the sign of

electrical that

ions •

is

the

10

can be measured also by passing a solution under pressure P through a packed bed of particles and measuring E across the bed.

=4w

-

rIX E/EP

viscosity.

X - specific cond. E - dielectric constant Most silicas (:Si-OH)

have

a

certain

population

of

silanol

groups

which can dissociate:

ESiOH+OH- -)>

i(SiO')+HzO

Thus a negative charge forms on the surface. Certain ions

such

as A13- can be adsorbed and change the sign of the 4 potential. On top of them citrate ions can be adsorbed to change again to-(91. Consecutive transformation of charge is following events [101: Silica (neg.

charged)

+ A13 - ->

possible such as in the

positive charge citrate ions ->

+

neg.

charge

Various degrees of hydration are possible. Thus silica particles prepared by precipitation in water or long ageing in it have surfaces entirely covered with ESiOH groups. If heated at >400"C the following happens: £Si

2ESi-OH

->

=

O+H2 0(siloxane groups)

O/ 2 Si At 1200"C only siloxane remains and hydration is

slow.

11.

on the

solvent.

Therefore,

the

coated

electrode

must

be

determined experimentally for each material since the particles acquire the charge spontaneously when mixed with the solvent. Further this charge may be reversed upon addition of ions. 2.2

ELECTROPHORESIS IN POROUS STRUCTURES

previous sections the assumption was made that the particle is suspended in an infinite fluid. When electrophoresis in porous structures is considered the presence of rigid boundaries will affect both the electric field and the In the

velocity of the particles. According to

J.

Anderson

(11,12,131

the

pore wall produces

three effects on the particle velocity: -

-

-

The applied electric field exerts a force on the d.l. of the An electrosmotic the wall is charged ( I .' pore wall if flow of the fluid is thus proc ted which either augments or of the particle velocity electrophoretic opposes the . depending on the polarity of ,. vs. The pore wall distorts the electric flux (current lines) around the particle thereby intensifying the local electric field so that the particle velocity is enhanced. The pore wall creates additional viscous stresses in the

fluid which retard the particle velocity. Anderson analyzed the effect of the presence of a pore when the located in the centerline of a long pore. Particle particle is Two geometries are considered: a neglected. interaction is cylindrical one and a slit

X=a/R

x=a/B (B

a

- particle radius,

(R - pore radius)

- half width).

The basic equations which apply to this system are:

nVIV-Vp=O

(conservation of charge) (Stokes eq., velocity dominated by

7°V=0

viscous stre-ses). (conservation of mass)

17-E=O

12

Solving these .quations one obtains: n UL=[(-1.2899X+l•8963X5 -l.0278X0 +O(Xs)x• for a cylinderical configuration, U,=[l-0.

and:

2 6 7 7X3+0. 3 3 8 3X-0,0402Xe÷O(Xa)$E(,pDZw)Ev 47M

for a slit

configuration

It

be

should

noted

that the pore size effect enters first

at

Therefore X has a weak effect. This is because of a fast decay of electrical and velocity disturbances from a particle A,3

moving in an unbounded field. The disturbance to the electric field is Ep=.L (

) 3[1_I -

V =_1a

3[

1

given by:

-3Ec

T7 3 UO

u

(ýV - w) Ea

Both fall off with X3 . The pore

wall

intensifies

the

electric

field

m..ving

the

particle faster but the effect on the hydrodynamics leads to a larger retardation, thus overall the velocity is reduced as the pore size decreases. In a

closed

system

the

average flow of solution through the porous medium is zero and a macroscopic pressure gradient develops to oppose the electroosmotic flow produced by the electric field. The cross-section available for the particle 1 is ( -X)l for a cylinder and (1-X) for a slit. cylinder: U=-•' E.[ 1(

(2 X_72)-_(i-r)X,289X3]

"p• slit: U

•--A p 47M E.[I÷(_-.w) (),-

2 IpI -_...

"0.p2677,'] I

13

Here the. electrophoretic effect comes in at 3 and is weak while the pressure flow has a contribution dependent on and has therefore a stronger contribution. It

is

to be stressed that in

the above analysis the pore walls were considered to be non-conductive. 2.3 It

ELECTROCHEMICAL DEPOSITION OF CERAMIC FILMS has

recently

[141

been described in literature that it is possible to synthesize ceramic films from water soluble ceramic precursors by an electrochemical method. The method is based on cathodic reduction of water or another anion as N0 3 -, to generate OH-.

A

secondary

appropriate metal formation of the Switzer (141 has catholyte was contained 1.0M

1.0M

ion

reaction

present

between

the

OH- and an

in

the solution can result in hydroxide and oxide of that metal. Thus, synthesized CeOm in a system where the cerous

ammonium

nitrate

and the anolyte

NaNO 3 .

The cathodic reactions which took place on a platinum cathode at c.d. of 50mA/cm2 were: 2H 2 O+2e ->

H2+20H-E,--0.83V

N03-+H2 0+2e ->

N0 2 -+20H-

Eo = 0.01V

The mechanism of the secondary reaction is may be: Ce 4 *÷4OH-

->

not clear yet but it

Ce(OH) 4

Ce(OH) 4 -2H 2 0 ->

CeO 2

The powder deposited on the cathode was identified by X-ray diffraction to be ceric oxide with a cubic lattice constant of 0.5409nm and a crystallite size of 7 .0nm. After ultrasonic dispersion the material was found to have a narrow particle size distribution with an average particle size of l.8U. The

14

morphology of

the

ceramic

film

depend

will

mainly

on the

Thus highly electric-conductivity of the oxide formed. conductive materials can form thick dense films or powders. It

is

possible

This is

which a

substrates.

By

by an anodic reaction.

0.5M

hydrolysis

to

an hydroxide or

has deposited thallic oxide films from a

1141

acetate in

thallous by

1.OM NaOH on silicon

reactions oxides of Ni,

similar

deposited

Mn were

undergoes

it

Switzer

solution of

oxides

synthesize

on the oxidation of a metal ion to an oxidation

based

state at oxide.

to

Cu,

Co,Fe and

Tend and Warren 1151 while Sakai et al.

f16] have deposited a mixed Pb-Ti oxide film. The methods

described

above

differ

from the electrophoretic

deposition of colloidal ceramic particles. that they

can

become

powerful

However,

we believe

for coating flat and

methods

complex shapes as well as for impregnation of porous structures with ceramic materials. 2.4

ELECTROCHEMICAL REACTIONS IN POROUS ELECTRODES

in Electrochemical reactions studied in two systems mainly: in fuel

electrodes

cell

electrodes

porous

have

been

and in electrochemical reactors used

for processes such as metal ions removal from waste water. Two groups

of

diffusion porous kinetics in researchers One of

such

exist:

electrodes

forced

flow

through

and

Theoretical analysis of reaction have been performed by numerous

electrodes. systems

(17,18,19,20].

the models used in

the analysis of diffusion electrodes

is the "straight pore" model

(see Fig.

5).

15

Z=O' TI=

Z=i n=TJ

0

r2"

a b b_

PLdz

Za

Fig.

This model

The straight pore model for diffusion electrodes and the current distribution.

assumes

that the electrode consists of a number of identical, straight, nonintersecting, cylindrical pores running

through the entire length of a conductive matrix. The basic

assumptions

filled with

in

this

model

are that the pores are

electrolyte, no potential drop exists in the solid

electrode so that potential gradients are in the solution only, the concentration of reactants is high so that no mass transfer limits exist. The potential distribution in such a system is:

TI

4RT z - qO = -F

16

The current,

up to point Z-I., and the total pore current

-

is:

I=tanf.-3 (XC)

1 k= os-'-h Yo

2io.Z2 F K=.......

---

2RT

22

where io exchange current S-

overpotentiai.

The current distribution will depend mainly on conductivity of the solution. For i2