kECURITY CLASSIFICATION OF THIS PAGE q ,. p F r _Form Approved

REPORT DOCUMENTATION PAGE OMB No. 0704-0188

lb. RESTRICTIVE MARKINGSA D-A 218 946 __ _ _ _ _ _ __ _ _ _ _ _

3. DISTRIBUTION /AVAILABILITY OF REPORT

4. PERFORMING ORGANIZATION REPORT NUMBER(S) S. MONITORING ORGANIZATION REPORT NUMBER(S)

6a. NAME OF PERFORMING ORGANIZATION 6b. OFFICE SYMBOL 7a. NAME OF MONITORING ORGANZATI -

(if applicable) WTNE E ____ __6c. ADDRESS (City, State, and ZIP Code) 7b. ADDRESS (City, State, and ZIP Code ,-'o ,

250, Route de 1'Empereur <"' \"92508 RUEIL-MALMAISONFRANCE IK

Sa. NAME OF FUNDING/SPONSORING Bb. OFFICE SYMBOL 9. PROCUREMENT INSTRUMENT IDENTIFKATION NUMBERORGANIZATION (If applicable)

/ DAJA 45-89-M-0234

8r- ADDRESS (City, State, and ZIP Code) 10. SOURCE OF FUNDING NUMBERS

PROGRAM PROJECT TASK WORK UNITELEMENT NO. INO. NO. ACCESSION NO.

11. TITLE (include Security Classification)

IMPROVEMENT STUDY OF DISILICIDE THERMOELECTRIC MATERIALS

12.PERSONALAUTHOR(S) Stanislas and Hubert SCHERRER

13a. TYPE OF REPORT 13b. TIME COVERED 14. DATE OF REPORT (Year, Month, Day) 15. PAGE COUNTFROM Sept89ToJan9O 1990 January 12 14

16. SUPPLEMENTARY NOTATION

17. COSATI CODES 18. SUBJECT TERMS (Continue on reverse if necessary and identify by block number)

FIELD GROUP SUB-GROUP

19. ABSTRACT (Continue on reverse if necessary and identify by block number)

------ In many applications at high temperature, the semiconducting compound- Fe Si2 has been already proposed for power generation. The goal of this

part of the study is to obtain pure semiconducting ( - Fe Si2 phase as a

e- 4, material of reference for thermoelectric properties.To start with, we ,elaborated the material, b Ihe melting of elemental Ironand Silicon pow&es at 1250 C and the transition of the metallic (Cc) tosemiconducting Q) phase is obtained by annealing. This classicalpreparation of'o - Fe Si2 was time consuming and because of the phasediagram results in the production of inhomogeneous alloys.

Continued on next page20. DISTRIBUTION/AVAILABILITY OF ABSTRACT 21. ABSTRACT SECURITY CLASSIFICATION

3 UNCLASSIFIED/UNLIMITED [=_ SAME AS RPT. 0 DTIC USERS72a. NArOE OF RIFSPONSIBLE INDIVIDUAL 22b. TELEPHONE (Include Area Code) 22c. OFFICE SYMBOL

DD Form 1473, JUN 86 Previous editions are obsolete. SECURITY CLASSIFICATION OF THIS PAGE

2

We then considered mechanical alloying, as an alternate method ofproducting P3 - Fe Si2. This method for elaborating the compound from lowertemperature must avoid the difficulties of sublimation and evaporation ofthe components. It also allows the optimization of P - Fe Si2.

We- pe the results of the two elaborating methodsby xlray diffractionmeasurements. These results on3 - Fe Si2obtained by mechanical alloying,

-as far -w4-aw,- were original. This process now appears to be a reliablemethod to introduce amounts of dopant in the material. It must now bedeveloped. ,

Acce35lo1 For14TIS GTU &

iDTIC T:

I Una n ou: dSu7ztjlficatto

By

.t,.p..) l ty CodeS

! !Availand/orDi_,c I p cialS-i ,

.. QL,

AD

IMPROVEMENT STUDY OF IRON DISILICIDE

THERMOELECTRIC MATERIALS,

Final Technical report

by

Stanislas and Hubert SCHERRER

January 1990

United States Army

EUROPEAN RESEARCH OFFICE OF THE U.S. ARMY

London England

CONTRACT NUMBER: DAJA'-89-M-0234

T.N.E.E.

Approved for Public Release ; distribution unlimited

January 1990

IMPROVEMENT STUDY OF IRON DISILICIDE

THERMOELECTRIC MATERIALS

This study concerns the first Dart of the program entitled

"Imprnovement ,study o ion d~siticide t~ermoeizctAic matetiaLs".

Tn many applications at high temperature, materials for to-

wer 7eneration must have a larae figure of merit Z, a stabilit-i of sur-

faces with respect to oxidation, sublimation and evaporation, a low

cost for elaboration. The semiconductLng compound B - Fe Si 2 meets these

requirements and has been already proposed for power generations. iron

disilicide is doped with cobalt to obtain the N-type material. The ?-

tvpe is obtained either by manganese (used by KOrA'SU commercially) or

by aluminiun (U. BIMF-1OLZ - Universityi of KARLSRUE). The manganese gi-

yes a Z of around K.5.0 - while aluminium is supposed to giveZ 1.5.10O K (CO ° C).

The goal of this tart of the study is to obtain inure semicon-

ducting 3 - Fe Siz nhase from elemental 325 mesh iron and silicon oowder

under the conditions of phase diagam (fig. 1). Thermelectric oroperties

of this material must be a reference for later studies. At once, we ela-

borate the material by the same classical method as KOMATSU or U. 3B1UICLZ

the synthesis of elemental iron and silicon powder is achieved by "elt .

at 12500 C and the transition of the r-,r. 71ic (c) to semiconducting (3)

phase is obtained by annealing. Then, w,. ind it is very difficult to main-

tain the stoichiometric composition during the a - .phase transformation.

Consequently to this method it is irmpossible to control adjustement of the

doping material. Using the competence and the experiments of our laborato-

ry, we elaborate with success 3 - Fe Si 2 by mechanical alloying. This ne-

thod elaborating the compound from low temperature must avoid the diffcul-

Lies of sublimation, evaporation of the components and allows the optim7isa-

tion of B - Fe Si2.

S1 (wtf0//)50 54 58

1200-Fe Si +cxFe SI7 2 a FeSI 2 (xFeS12+Sl

4-

ai 1000 ~ ei

E 986

S 980 - (FeS 2 +

-FeSI + P3FeS12960 -pFeS12 95

PFe SI 2 SI940 -

65 70 75

SI if 0/0)

Figure I1: Phase diagram of the binary system Fe-Si

1. CLASSICAL SYNVTHESIS OF PURE IRON DisILICIDE (~-Fe Si z).

Acccr-fing -coa classical cDrocess -4- can. be o3btained by arrnea-

~ £75 f -m ce a -etal7li-c Dnase. Th are 7:'aterials iro=n (F

p935~,silicor -'39,995 %)were -urified i a recuc2ing a--.^s-± ere_. Then hecor~zunds -n 3 t-oo0 ero oooiio er ele

cv ntucoion -ea'-n~ in a qua-tz amp~oule unmder 'e-K a m'csoh-ere. --n 7this-;ay we obtained hea - metal'lic t-hase of Te S i2 and hen.e~s ~ v

me at - 3 hase tasona~n

Aft er colin -o room t_-e-ora~e --he ingot w..as zrount

ill. Te Dowder :*- as filled; into a cylindrical, :ox of 1C n- i fiameter

and a ram Dressur-e (7,- >T7) w.,as apied4,. Tesancles wer-e annealed- --~K .mder ,,acuum :for several hours. The subsecuent s7teir rocess

*zas Derforr'et for 3hours at VL5 K under a D e 2 72xure-Th

-_a~s _t4on to 7th1e sercnu -- ohase w..as achieved by a 7 :cr se-

vieral days at 1075 K under hihvacuum. :n t he t-able bDelow -.-e show he

measuremernt : o h-' e -cr ) ower at, room 7emera7=2e.

N* sample 1 (1) (2) (3) I(4)(5

16/ .7,3 2u6,5 ')04, cq' -5c

Wle note,- -:hat all samples were always n-tyo-es. Th!-e time

-or the phase trans_-tion from t-he ret!-alli-c h_4_h t-emrera--cre t o s sen-m-

-onducting phase is found -to be ver sensitive t-o t-he 5stoicn--om-etr4 con-_

tositon of the ma erial. Thus -her-celectric Dower chanzed in t-he same

manner.

- LL -

Some measurermenis on electrical resistivity and thermnal conduc-

tivity wereoerformed. Sampies had a very high resistivity about 50 m C .M.

T':his value was not surorising because the samples were undoped and -ne

thermal conductivit was found as a classical value X = 5.95 w.m-.K

X-ray diffraction e~asurements on these samnles showed c:-at

- Fe Si 2 line vanished but we noted the .-nortant Dresence of Fe 2i

lines (fig. 2).

Intensity

t= 1-FeSi1

2= FeSi

2 2

t20

60 70 10 go 10 1l 1223 10

I I

1 I\ I

Figure 2 •X-Ray diffraction pattern of P3-FeSi 2 Sample [classical synthesis]

From the ?hase diagram, it seemed that it w.as very difficult

to keep the stoichiometric composition during the a -8 3?hse transform.a-

tion. Consequently to this method it is impossible to control adjustement

of the doping material.I I

11. - PREPARATION OF PURE 5 IRON DISILicIDE ( Fe Si.2) By MECHANICALALLOYING. -

the classical c rec7aration, of F'e --i 2 , s;' aw. -as --

~cs~~ant ±Lfcut ecause of th.e ochase :.iagram Fe :D or -=4

cles are 7oresent and therefore -result_ n -!'-e cdUCt4cn 0. Lmorc-enecu

alloys. --- add ition, --he high vaour- oressure of a'i'ur (o xace

at -eevac-ed ternoeratures makes the corntrolled addici :_on of -:hi4s dn

ifc Li the melt. Th,,ese -4 ' ::=4culci es -ich cast aPlovs '-ed us -o cn

sicer 7:echanical allloyng as an. alt-ernate cehdc rcC~c 2 on'er-

7;celeccrjc 7;aterials.

. echanical lovn offers oh osblt fr C* or_

~c~ca± nzer~er~it/and 7 :-*-sze and of eva1',a- 4 _ d~crs:m

-ore C amenble -o --he melcin- orocess. t- is a -7echnicue n whic-hcr-

calineintrn~calicor ele.nencal --owdters a_ _ oyed 3hoe ec-e-.

of collision evern-s inside a hih enerT/ ball nil n a 7acrcscoc::i sca-

le -- is orocess i-s accormelished throug h a neoeat-ed fractu7re and col. we,-

2,L~ :-rocess. t- i's desirable --o ore7are ther-noelectric 7anerial-s _inh lw

-herral conducni-_vi-z' so t-haz a terneerature gradient -can easily- b-e 7azirna-

ned = a d-evi-lce, 7rovid ed t-he el1ectrical conduc-7_vinv-_ an eeL-ec< o~:

cien can te <eo7~ Suitably lar -e . 7Vi dence suggests that nechanical elv

,rcud o.roduce a 2'an er~al .. i-ch exterelv small < 1 u mn) 7arni _cle sze w.~nichn

shoul result in a S4iifican-! reduction in t-he-rnal ccnducniv_* '. --o achie-

vre alloyving we used a vibrati.g slhakezr and o)1an~e-!arv =_'es :)f~ :inI,-ecual success.

Ill. - MECHANICAL ALLOYING W='T A VIBRATORY SHAKER TYPE OF MILL.-

The Fe --i 2 samules wereo-reoared from elemental 325 -nesh silicon

and iron -Dow.der. 7ne amunts of': each comronent wNere w.,eighed in anargon-

i,-'-lled 7love box --o obtain a nominal Ccmosition of Fe Si2 and Dlaced z

a nunsszen carbide vial along w-ith a 3 --ram 1C =_ diameter "C bDall and'

seale(d inder argon. Tn t-he initial atterrts t-o achieve allovinz t-hese vials

wqere- loaded into a vibratoryj shaker tyce of m-lu ig )

Fe +~ Sipowder

Shaking mouvement

:F:gure j 7,"-raor~r shaker -7-e 7,-77,!

:(-?,av 1uifract-lon :rteasuremervts aen of- '-e 7cwcer aroer - ccur

0: -Drccessina showed tht he Si and I-on *-- e s hadna ar, 2:c_

allovLn7 had not occured. Ln he Iur'e ~cli_-ches a andcorendd'

o cer resuectivelv t-aken free in th!-e vial arrd a geanet on --he -.,a1I's

th'-e vial. No change could be oDbse-red as o :,r lich-e C fr' 3 'curs or,, a-rc-

cessinz. MIany other at-tem~ts -.zere zeade uo :o 111 h-Ours.

:n -the furS _-, t) nd cffr l, 15 and 3C hoCurs r'es-,ec-_-vely

some changes appeared, enlarging ol: '(-'av -- -es end aearance of7 3 Fe :2

X-ray 'Lie.

it was felt that in' t-his -pe -7f 1 nil,jSUffrZi*ent energy w.,as

disioaedto cause allovying to occur.

7-

Intensity xiO 2.5 . 0

4.05

3.20 Si

2.45

1.80 Si . --

1.25 Si " Fe0.o Fe Si Si Si -

eFe

0.45

0.20

0.05

20.0 40,0 60.0 80.0 100.0 120.0 140.0

Intensity (b)5.00 (b)

4.05

3.20

2.45

1.80

1.25

0.80

0.45 olh1

0,20

0.05

20.0 40.0 60.0 80.0 100.0 120.0 140.0

xto 2[ntensity 5.00

4.05 (c)

3.20

2.45

1.80

1.25

0.80

0.45

0.20

0.05

.. 0 . 0 .' 0 2 . 2020.0 40.0 60.0 80.0 100.0 120,.0 140.0

I I

intensity -.A -

1. ;2 ,(a)

).2

).50

1). --! 8 10

').28 a

22

10.) 40.) ;0.0 30.) '00.1 20.)

Intensity ,I%

1. ;2(b)

). 38

0.72

0.50 "

1-32

0.02

20.) 40. ) 60.)? ?0. 1.'0.) 120.)

Intensity to 212.)0 1

(c)1.29

0.39

0.32 4N r

0.18f0.08

0.02

20. 40.0 60.) 30.0 100.0 120.0

Figure 5 X-Ray diffraction pattem of samples after ball milling

IV. - MECHANICAL ALLOY.G WITH A PLANETARY TYPE OF MILL.-

':!-e -v,,als filledi under --he same ccnditi--OnE as before w..erelcacet aDlaecary :7'i -. qich rapidly rcoat as e ach v anc L S

Ln- ooooCS4-7c:. 7:o -. e -Dr:Lmry royta-17oon axis of the Dla-_for- 6 ).ore emo-s -ere -rade -o ocormise the tocmoonents 7ass, numrter

bals ~c--oe of -rocessing. For all -the followinz a--o-S, th,-e

mrun-ts ofach ::cmocnent were an~~e (arc.l -v~ -----

1'ebx -o obtain a norminal CC= ositio of Ta i 2 n:lodia

7-ing-sten carbid-e T-l 1 wihseven, -~am '1C om ia.e-_er "3,C balls

and sealed urmder arloo-n. Thie .latforo r-otates at the r-ate of 3'L0ror and,

Ia vials at 380 r,7,.X-rays ifrconosre~: a-nof he ow-

.er after 2 hours3 :f -.,,,sn~ ere shc,,- in- fiozre a

:.e ser-z -- e -axis-e:tce -

- r Ta i2 (cuadratic) lines Iand 7

- 3 Fe Si 2 (or-ohorombic) 'lne 2

- --e :zi (cubic) : lin'e

- --he iine S 731owC thle s~itnos:,ceserc30.

After '-hours of Drocessina we noted in t-he fizzre T

- lnes I and -- decrease in a sinfcaietnner

- lie 2 (2 S Ti) increases oonsiocerahiv

_ ie 4 shc-is --he li'saooearance of' 7e Ti

- the d ecreasing of line 5 confirms -the zransforo'auion,

of F e Eiz in* rela _on w.,ith a -. Li 2 and Fe Ti;

^..e think t-hat t-his is a god a to obta-i 3 Ta >2'

Opposite

Rotative platform

WC Vials

Balls

Fig 6: Planetary type of mill

.Intensity

'I IS ". "(a)

MI I

' II

a -I= 1 3

-.FeSi = 24;

i ". rFeri = 4

Intensity

(b)

I Il l I l

If

5, 7-, Be N, S 100 116 12 130 140 150

(a)13 -ei afe 22;5rb)

I-

IL

t:IIi 1 11

NJ I I

LI A.11 I

P. 20II 14

,.., ~ ,.v, 5' 4 70 so 9o iw) ito 1229 i 144) 15

Fig 7:• X-Ray diffraction pattern of mechanical alloyed samples(a) - after 2 hours(b) -after 4 hours

-he figue 3 shows an exper1ment of D. S. C. (Difrni al37Can 4 -~ Z c lriLmetry) with -o-der aken after '-our of -DrocessL-g.

-,e note -ha-7 -he . 5 nfi~ t-akes -D7ace at a lower te.=era-

-tur-e Q-4CC-5CO0 C) than --he tamoeratu"re_ :Drdicted ty t-he o)hase darm

_hS 43 robablv d-ue t:o -the e.-stence of4: many energeti-_c defects in-he,owder -whihn aeneled quickly. These derfects :avourise cer-ta_-.;.y

t7he -Di-ase -rarisfor7-at-ons.

:n the fig',re 9a are reor_-esentat.ed the X-rays lin'es of t-he

-:cwder after one '-our Of ridinz. Th-Ien t he -owder 4S annealed (SCComui~oe hour and the X-ray;s soect ra are represented in- 3itr b.

CONCLUSION.-

The results obtai'ned from a olane-zar-7 -all were found t-o *-e

suitable t-o elaborate the semiconduct!ig 3 Fe Hi2 ohase. The raDlC'-i

t'he orocess (about 2 h-ours of g;rindiJng and annealing) w.,as not-ed.

Results about D. S. C. shocwed that t-he a - 3 ohase tasto

was a-. a 'Lower te.Rmerate due to an anrealinz of very, energ-etic defect s

int-=,oduced _tUring mcchanicolall7oyn'g.

These results on 3 Fe i2 obtained by mechanical alloying,

as far -as we know, were original Then thi4s prccess seems a reliable me-

lahd -to introduce amrunts of Iooant in the material and tist be cdevelooed.

Heat Flow (mW) Mass: 89,300 mg Rate: 10,1 'C,'mn

7I

213

i /

/

Temperature ('C

Figure 8: a transition observed by Differential Scanning Calorimetry

- 14. -

4A

Intensity

F t(a)tcc. . FeSi

I, [I S

I'A

I| I II CI l pl I I

Ik II .l I I

I tesity I IIt

.-. I. t !ii?S I I It I/

I(b)

FeSi FeSi FeSi

. . .. -, , -- + " - -

50 Go 70 soto 's

Intensity

()a (b)

II

FeiI'

I

54

iue9 -a ifato atr fmcaia loeape

(a n orofbl iln

(biaepu oehu faneln 50C