citation doc url bulletin ... · pdf filefor turbine blades are at times coated with the anti...

10
Instructions for use Title On the Atmospheric Oxidation of Pure and Siliconized Ta Metal Sheets at 1000℃ Author(s) Nishida, Keizo; Hachinohe, Mitsuo Citation 北海道大學工學部研究報告 = Bulletin of the Faculty of Engineering, Hokkaido University, 79: 139-147 Issue Date 1976-03-19 Doc URL http://hdl.handle.net/2115/41343 Type bulletin (article) Additional Information There are other files related to this item in HUSCAP. Check the above URL. File Information 79_139-148.pdf Hokkaido University Collection of Scholarly and Academic Papers : HUSCAP

Upload: dinhnga

Post on 24-Mar-2018

214 views

Category:

Documents


2 download

TRANSCRIPT

Instructions for use

Title On the Atmospheric Oxidation of Pure and Siliconized Ta Metal Sheets at 1000℃

Author(s) Nishida, Keizo; Hachinohe, Mitsuo

Citation 北海道大學工學部研究報告 = Bulletin of the Faculty of Engineering, Hokkaido University, 79: 139-147

Issue Date 1976-03-19

Doc URL http://hdl.handle.net/2115/41343

Type bulletin (article)

Additional Information There are other files related to this item in HUSCAP. Check the above URL.

File Information 79_139-148.pdf

Hokkaido University Collection of Scholarly and Academic Papers : HUSCAP

;iヒ幸毎・道プく学コニ学音i‘石序5豊酒粕テ

㌶ッ79.場・ (1昭孝1…5!脅の

Bulletin of the Faculty of Engineerins;,

   Hokkaido University, No. 79 (1976)

On the Atmespheric Oxidation ef Pure and Silicenized

               Ta Metal Sheets at 10000C

Keizo NISHIDA and Mitsuo HACHINOHE       (Received September 20, 1975)

Abstract

   With a pure silicon vapor technique Ta sheets were siliconized in an evacuated quartz

ampoule up to 49 hr at 10000C, and then these siliconized sheets as well as untreated

ones were oxidized at 10eOOC for 15 min in an atmospheric air in order to determine the

resistance of siliconized Ta metal against oxidatioR. Severe oxidation of pure Ta sheets

almost prevented by using this method under the above-metioned conditions. Next, the

mechic nism of this protection process of siliconizing was discussed.

1. lntroduction

   Investigati,on and development of new metallic materials aganist the high-

temperature oxidation attack are in stroRgly demand, since the environment in

which metallic materials ai-e used has recently become increasingly detrimental.

   There are yarious methods to improve the anti-corrosioR property of refractory

rnetals. The practice of surface coating with vapored metals has dyawn special

attention as an effective method. For example lnconel alloys now widely used

for turbine blades are at times coated with the anti-corrosion metal by a vapor-

diffusion ’method, resulting in a satisfactory modification of the oxidation re-

slstance.

   From thl,s point of view, the present investigatin was carried out to examine

in what manner a siliconized Ta sheet shows better resistance against the oxidation

attack than that of an untreated Ta sheet, because pure Ta metal fails easily in

atmospheric environrnent above 10000C in spite of its excellent mechanical proper-

ties at elevated temperatures aRd furthermore, the vapor pressure of Si is relatively

high and easily produced.

   So far the oxidation of pure Ta metal has been investigated by Peterson etal.i), Gulbransen and Andrew2}, Bakish3), M[ichael‘〉, and Albrecht et al“”}. Recently,

there is a detailed report by Dooley et al.6’7) on the oxidized structure of single

and polycrystalline Ta metals. General}y, the oxidation of polycrystalline Ta is

found to obey a logayithmic corrosion rate law below 5000C and a parabolic rate

law in a range of 500 to 10000C, and it is oxidized linearly above 10000C.

   The effect of alloying elements for Ta metal against oxidation was investigated

comprehensiveiy by Kloop et al.S), who showed good improvment of anti-oxida-

tion properties in the order of Ni, Mo, Cr, and Si in atmospheric oxidation for

lhr at 12000C. From those results, if Si could penetrate into Ta metal through

its surface, the treated metal would show good resistance in air-oxidation at elevated

temperatures. Consequently, siliconizing of Ta sheets in Si vapor will be probably

Metals Research lnstitute, Faculty of Engineering.

140 Keizo NlsMDA and Mitsuo HACHINOHE 2

accompiished relatively easily and an excellent resistance of the metal aginst the

oxidation attack may wel! be obtained.

2. Experirnetal proeedures

   Test pieces for the preseRt experiment were prepared from sheets (O.3 mm

thick) made by Van steel Co., U. S. A. These were cut down to a size of about

10×10mm and annealed at正000QC for l hr iなaquartz ampoule evacuated of air

up to 4×10-6 mmHg and then their surfaces were polished up to #1000 emery

paper fol}owed by washing and degreasing.

   Such test pieces were capsuled once more in a quartz tube at an order of

10-6mmHg with pure Si powder (T200 mesh) prepared by The Shin-etsu Chemical

lndustry Co. Ltd. and diffusion-annealed in an electric furnace as shown in Fig. 1

for various lengths of time at 10000C.

   Tabie 1 shows the details of this siliconizing experiment, listing the purities

and vapor pressures of both metals. The siliconized metal sheets were thenoxidized in the apparatus shown in Fig. 2, iR which a test piece was hung in a

vertical quartz tube through which air was allowed to fiow up at a rate of

leO cc/min. This fiow rate of air was determined after comparing its effect on

the oxidation rate of pure Ta sheets. Moreover, the variation of oxidation rate

was checked with or without desiccating the fiowing air and the effect of such

treatmeRt oR the corrosion was found to be much less.

   These oxidized test pieces were observed microscopically and diffractometrically

and also the sectioned parts were checked by means of an elecron probe microa-

nalyzer (EPMA).

「孫、。,.i (一200mesh)

1 ....g.一9.pmsu/ rl一一IO’6mmHg

Fig. 1

「一一葡’.,,・..黒6ムt&【ロ「パ)sh包et o   o

1 eoxioxo.3mm)L一

Table 1

Schematic presentation of siliconlzing

ampoule and its furnace

Test piece and si}iconizing condition

Time(hour)

1000。C …   4   9   16  25  36  49

SILICON

X9.999%Si

 TANTALUMX9.9%Ta min,n,03%Femαx.n.03%Cmqx.

VAPQR PRESSURE  TOOO。C              ,

10-5mmHg 10-21mm”9,

MEしTING  TEMPERATURE

1410 。C 2996 。C

I

E

t

1,

1

一Air outlat

  Sjljconjt

gt

SPecimen

  一 Air inlet

Ftow meter

Air pur叩

o

Fig. 2 Schematic presentation of

     oxidation apparatus

3 On the Atmospheric Oxidation of Pure and Siliconized Ta Metal Sheets at 1000eC 141

                                3. Results

    3.1 Oxidation of pure Ta sheet

   On account of the shape of a test piece oxidation over a long period failed

owing to its descent. Hence the oxidation test was carried out within 15min at

10000C. The results obtained for the change in weight with the square root of

corrosion time was shown in Fig. 3 (a) in which the oxidation seems to obey a

parabolic rate law in general, showing only a slight deviation from that law.

This behavior is more clearly indicated by a log-log expression .shown in Fig. 3

(b),which shows that the relation of weight gain(」アのwith time(の;AW=ktn

holds with the value of n==O.64, because it is mainly due to the abnormal

oxidation at the corners of the test piece as will be stated later. From the

experimental results the metal shows a weight increase of about 23.5 mg/cm2 in

15皿in at lOOO。C and the rate constant k in this case is abut 4.0(mg/cm’2 mini?2).

   The surface of oxidized sheet was pale grey and was firmly adhered to the

substrate, but the oxide scales showed some cracks from which white oxide was

 30

A?5

毛,。

It T,

g,,

.tug

ls

 o  oCG)

   一

@,I   一

 ■

I

1

Li

:hkN

Fig. 4

 ra)

 ’わノ

   ?a

Fig. 5

            50

           ρ「.0

            52。

           も            E           vlO            ≡…           ’07            団5

           三           .93            ゆ           ≧2

             T2.

@ 3  4  S  1  2 3 5710 Time(mln)    (b)     丁imeCmin)

Fig.3Weight gain vs. time

/に’、  -

ずF

 /’o.

n=0、5

./ 5づ

1

O署idqtiqn o野Tq 5hoet αt IOOOoC for15 min@      F.【【5cα1。;3・1♂ζpn

  Pur噂.Tα

 T(亀05

Crロ¢k

20ド

一   一

20 30

     Distapce(V)

EPMA profile from oxidized Ta sheet

30 40 50 60 70’ eO 90 ILIO 11.) 12J 130

     20 rdeg.) iv・ith CrKot

X-ray diffraction patterns from oxidized

Ta (Ta20s)

(a); Surface, (b): White oxide powder

竃2爵1:…爵…由自営癬. 衛}.群嚇…灘麟;幽幽;幽幽…幽幽囎.

ζ…嚇i.鷺…農…嬢.響;.㈱《轟…{趨1;

職欝繋騨犠1Ch6毛6丁1 E一.ternww.1’tihd”S66ti6fitt’1”

views of oxidized Tasheet

Photo. 2 Sectional view of oxide

with markers on its sur-

face.

left : Ta, right : plated Ni

142 Keizo NISH工DA and Mitsuo HAcHINOHE 4

seen extruding. The stripes seen on the oxidized surface may be scratches from

the polishing abrasion. At tlte corners of the test piece abnormal behayior was

seen. Namely, the above-stated white oxide crystals extruded and appeared curled

up as seen in Photo. 1. Detailed inspections showed that the corners had noscales as pointed out previ,ously by Dooly et a17).

   The microstructure of the sectioned part of oxidized meta1s was shown in

Photo. 2. The oxide appears dark grey, and the metal/oxide interface was not

fiat, showing an irregular oxidation. This fact would be attributabl,e to the

preferential oxidation at the grain boundary of the metal substrate. An EPMAresult showed the distribution of Ta in the section as seen i.n Fig. 4 for example,

from which the scales seemed to be Ta,O, (identified later) equaliy thyoughout

the oxide scale. Such scales are very hard and are adhered to the substrate as

stated above, so that the removal of it from the metal substrate was very difficult.

   The removal was not done to check the oxide crystal by rneans of an x-ray

diffractometer.

   The x-ray diffraction of the oxidized surface made for various oxidation times

all indicate the same crystal patterns, and that from th.e white oxide crystal

powder shows also the same crystal, although it contained some other diffyaction

patterns which were probably due to the dfference between the orientations ofthe same crystal (see Fig. 5 (a) and (b)). ln order to investigate the oxidation

mechanism of Ta metal, alumina particles (O.5 pt¢) were placed on the surface of

the test piece as markers. After the oxidation the almina maker was still found

on the oxide scale・ of Ta sheet, as seen in Photo. 2. Consequently, the oxidation

of Ta metal at 10000C was concluded to be performed by only an inward diffusion

of oxygen atoms through the oxide (Ta,O,,). According to Kubaschewski et al.9)

Ta20s was classified into a metal-excess type (n-type semiconductor), but as a result

of consideration of these atomic sizes and observation of the marker i fi this

experiment this oxide may be rather regarded as an anion-deficit n-type at 10000C,

so that movable species seems to be oxyen atoms alone rather than Ta atoms.

    3.2 Siliconizing of pure Ta sheet

   The surface appearance of siliconized Ta sheets is dark grey with a disap-

pearance of the metallic luster observed before the treatma.ent, but theye was no

difference among the test pieces for various treating intei”vals.

   The results from the x-ray diffraction analysis for each test piece are shown

in Fig. 6, in which x-ray patterns on its surface for a short time (4hr) are almest

the same (TaSi,) as those for a long time (49hr) except the former shows other

peaks. These peaks were clearly identified as Ta,Si,, so that the compound formed

for a short time may be regarded to be i.n a transient phase. Therefore, the

surface seems to consist mainly of the same compound in spite of the various

treating intervals.

   The behavior on w’hich pure Ta metal was siliconized was shown in Fig. 7.

This figure shows that the weight gain (mg/cm2) of Ta sheets obeys a parabolic

rate law, with a two-step siliconizing.

    Next, Photo. 3 shows representative sectional features of these test pieces.

The formation of alloy正ayers was very slow and£he structure of them was not

clear for a short time (4 and 9 hr), but after 16 hr the layer was th,ickeRed and

a metal/alloy interface was fiat and also the adhesion between metals and alloys

seemed to be good. For comparison the hardness of the surface was examiRed

and the metal subtsrate was about Hv=115, while alloy iayer showed Hv=一1032

5 On the AtmQspheric Oxidation of Pure and Siliconized Ta Metal Sheets at 10000C 143

Fig・. 6

鮮5「吊’.嘩.ve  ・:.   ・ ㌦双. di=・1    筆      ・・.   ;‘夢;   ・±    .蹴.ニ                                            ご  と                             ド

欝灘鍛 紬醜ゴ(a) 4hr

        .髪.

       t

4こ,  .5㌧1  6J   70   30   9J   PO  ノノ1フ  1?0

   2θrde.9/ w~舗 」一r’κ感

X-ray diffraction patterns from

siliconized Ta(a): 4hr, ((b): 49hr. x: TasSi3,

0thers: TaSi2

   ’t ” t/”

(b) 9hr

 重;     1・;

130

(c) 16hr

一100

c51-e

も1’4

1’a3

0for2’i

     0 1234S67             Timethr)

Fig. 7 Siliconizing of Ta sheet at 10000C

25

   (d)25hr.     (e)36hr                           3’

  Photo.3 Sectional views of siliconized Ta sheets      缶20          at 10000C;upper part Ni plated, lower      台          part:Ta;intermediate layer:silicide      一                                    ド                 あ                                              暑 15

                                              ぢ

↑                   鴇 th-                              210 ℃                                     着 に                                                                                    ロ                                                                     エニ

 £                                  ト

 も

 き ε ε ・ミ                                               0   1   ?   3   →   E

                                                      Weight Gαin(m9’⊂111乞)

                                            Fig.9Weight gain vs. thickness of

               D’stance‘ノs,                    alloy layer of siliconized Ta

    Fig.8EPMA patterns from siliconized Ta        sheet

±50.From the above results the alloy layer was found to be very hard. so that

this coating may be very useful for the other fields.such as antiabrasive parts.

   The distribution of Si as well as Ta in the treated section was examined by

an EPMA me.thod with results as shown in Fig.8. In this figure another compound

in case of the low concehtration of Si for short-time treatments was present(4 and

ghr). This wes identi且ed as Ta5Si3. After a g hr.treatment, higher silicide w.as

clearly seen in the figure, and the compound was determined to be TaSi2 corre-

sponding to the x-ray diffraction results. Moreover, from the detailed examlnatlon

of this pro丘le the inside of the alloy layer seemed to have a sl.ight solubility of

Si in Ta metal(in this figure it could not be indicated clearly). For example,

the 16hr-treatment showed a penetration depth of about lOμ. Of course, the

phase diagram of a Ta-Si system1。)indicates that Ta has a solubility of about

O,Q2 wt%of S.i, so that the above stated obs.ervation is highly probable. Further-

mqre, the TaSi, compound has no solubility in the present phase diagraln, but in

a diffusion treatment such as the present investigation a calculation from the

且gure of the results for 36 hr-treatment showed to have a solubility range of

about 1.4 wt%Si.

   As the thickness of an alloy layer for different treating intervals also indicated

the same relation against the square root of time as Fig.7, the thickness of an

4伽r@I

9hr ’ hr 25拘r@I

  1

R6hr

9 F勧ro丁b

一凸  一 「 一

9 -  一

一一  層 一 ●.

puqrtz c隊stαis“しdr一「αy

e.U,。・1・、3,1げ・p暁

ri

jAP cry5tαl

rtK的一r剛euUs・d・・1♂cρη

而 而 掬 知 而

  跨’0

s’

@曜   .

s’晒.s「.’

、4

3’’

s’,  .

一 ” ., 一 7  一

144 Keizo NISHIDA and Mitsuo HACHINOHE 6

alloy layer was found to have a linear relation with silicollized amount(mg/cm2)

as shown in Fig.9. But there is a certain veriety of Iayer depths at different

positiohs even in the same test piece.

    3.3 0xida1:io皿of siliconized Ta sheets

   In order to compare the anti-oxidation property of siliconized Ta sheets with

that of pure Ta ones, Ta sheets siliconized for vafious intervals were oxidized

for 15 min at lOOO。C under the same condition as that of the latter. Fig.10

shows the results obtained. This indicates that test pieces treated for a short

time were severely oxidized, but an amount of oxidation was markedly decreased

in those siliconized for g hr. For example, the 4 hr-treated test piece was almost

covered with the white powdered oxide covering the surface, but those treated for

along time showed orlly a slight oxidation and a grain boundary attack wasnoted. These oxidized surfaces were also checked by means of all x-ray dif-

fractometer as seen in Fig.11, in which it was revealed that the test piece treated

fQr 4 hr had the same patterns of.simple Ta2050xide and some of those of the

original TaSi2 alloy. The surface of a 49 hr-treated sheet is considered almost

the same as the original one.

    Here it is noteworthy that the existence of Si oxide was not recognized and

it is considered that there is no Si oxide on the test piece. However, the Si

oxide might be not detectable by means of an x-ray diffractometer because of its

amorphous structure even if it was contained in the. oxide scale. For the purpose

          7

                                            4ゐ「.s

          6

                                            s hrs        ロ        筆5

        芝                                            49んr∫        ロ        s4

        .E        O                                                 2θ(deg、ノ wノ’わCrKOt

        ロ        看3         Fig.11 X.,ay.diff,ac・i。n p。tt。m・f・。m

        塁2                               0xidation of siliconized Ta

                                                    o:Oxide(Ta205),                                                                 x=TaSi2

          1

o

oπ’ゴσ”α70∫5〃σσπ‘2団階

@   副000・Clbr 15層.’丹

1

1

1 2 3  4 5 6 7

         Siliconizing time(hr)

Fig. 10 Oxidation of siliconized Ta at

      10000C

(q)4hr 切16hr (⊂)25hr (dl 36 hr (e)4.9hr

.   一   . .    一 .    一   - 一   ■ .    {   . 一    . PureTq

1Tq205 1 工⊆s2

一   . 一    .

Oxide[qye.「芦lllci働

「   l@      x由e       ’

Sill,ide  、

I  l垂сニ.Si贋

唖・  15陛聖

ll IC kI       「

P1C「q⊂k1 Fu115c.α【e:

R.葺。5cpm

@106叩mfo「丁αL匡r

?C窒ri K.凶 1・、1 ピ「 ρ●

  「

@「@’j

    20H

黶@ .  一  ・ 曹   一 一   ・

.]v」l l,_製   」_.

  1.s

@ トー

@且.⊥.

o o o o D D o 諾

o 0 o x x o.o x x

∫ o x o 罵 o x耳 篤 x 裏κ x x x

2G 30 4050   60  70   δ0  90  100  1ア0  120  β0

8壬

s

5診2

sE

騨…雛繍議灘町民

.穣

勢.f

n「獣t.:

…瞬1灘裁!撮魏蕪鎌上騰…

                   Distance(H)

Fig. 12 EPMA patterns from oxidized Ta sheet after

       siliconizing for various times

騰識撫灘灘藻轍繍書懸蕪繋犠.

:轟醐

  ..i蝋瞬.:騰§晒達.1醸級.蜜.....嚢醗麟..... .

Photo. 4 Sectional views of oxidized Ta

       sheets aiter siliconizing for

       varlous t1エneS

7 On the Atmospheric Oxldation of Pure and Silieonized Ta Metal Sheets at 10000C 145

          alloy layer

oxidation. This is probably attributable to the migration

of Si toward the alloy layer during the oxidation process.

    3.4 Kinetics of the oxidation of siliconized Ta

        sheets

   In order to investigate the oxidation behavior of

siliconized Ta sheets, 49 hr-treated test pieces were used

at 10000C up to 70 min. The results obtained are

shown in Fig. 13. From the observation it is suggested

that siliconized Ta sheets were oxidized obeying a

parabolic rate law with a much lesser corrosion rate(rate constant k==O.181) than that of the pure Ta sheet

(k =一 4.e).

of the detection of the existence of such an oxide, the newly prepared pure Si

powder was oxidized for more than 1 hr under the same condition as before, but

still such oxide could not be detected. Hence it is probable that some oxide filrn

may exsist in spite of the lack of a reliable detecti.on method. This shall also

be discussed later in the paragraph where checking with an EPiMA method will

be described.

   Photos. 4 (a) to (e) indicate some of the sectioned microstructures of oxidized

samples. According to Photo. 4 (a) the test piece after 4 hr-treatment has the

same metal/oxide interface as pure Ta sheets corresponding to the EPtMA result

(Fig. 12) and the alloy layer is already destroyed (refer Fig, 12 (a)), ln this case

it is clear tha.t the oxide layer reaches the Ta substrate.

   Moreover, there is a characteristic distribution of constituents, i. e., Si seems

to enrich the oxide, Accordingly, it may be surmized that there may be some Si

oxide in the scales, in spite of the fact that the compound is unknoun. Another

special feature which is notable is that there is an even more enriched zene ofSi in the alloy layer than the concentration before t2r一一一一一….一一一一一一…,一一一……一一,一一一.,.....

一. IQI

芝ao,sE

E’6 a6

a芸an(届’a

 O,2

r”

了es.

SIIIc

atlO

  k=0.翠8

   1t..一..rrttt. Gttt-ttTt

   l

   l

一一…一…

G一一一一

   l

   I

......k...

   ∈ ノ

Fig. 13

           」2  4  6  e lo   Time(min)

 Oxidation ef sili-

 conized Ta sheet at loooc

4. Consideration of results

   From the above stated results it is suggested that the oxidation behavior is

closely related to the siliconizing (mg/cm2) as seen in Fig. 14. lt is quite clear

shat the siliconizing amount of about 1 mg/cm2 shows a strong suppressing effect

against oxidation. Moreover, another semi-log expression of the above relation

is shown in Fig. 15. There are two stages in the behavior, namely the first

stage of the reaction corresponds to the oxidation in which the weight gain

decreases largely with the increase in an amount of siliconizing. The latterstage of the curve involves further the oxidation of remaining Ta silicide (TaSi,)

in the substrate and the weight gain by oxidation decreases in proportion to an

amount of siliconizing in this representation, so that these two stages must be

discussed separately.

   Of course, the former stage of oxidation corresponds to that of Ta一一rich silicide

(Ta,Si,) as stated before. Thus, the oxidation of this alloy layer will be relatively

rapid and under the present experimental conditions the alloy layer is oxidized

up to the thickness formed after 9 hr-treatment. Nevertheless, the thin oxide

layer formed strongly retarded further oxidation.

   Here, considering the latter stage of the oxidation, it would be, at first,

necessary to clarify the relation between the oxidation of this alloy layer and the

146 Keizo NlsHIDA and Mitsuo HACHINOHE 8

25

g這

ヨ。1

.二

6 s1

黶D

            [.…マ..             一

@           口

        Silicenizlng(mgicm’)

Fig. 14 Siliconizing vs. oxidation

Toe

70

50

3e

20

 10:s

s.萎

.9

6 1.O

 O.7

 0.s

o,]

o.i U

Fig. 15

 1   2   3   4   5   6   Siticaml:ing Cmg!cma)

Oxidation vs siliconizingcurvefor Ta sheet at 10000C

diffusion behavior of Si atoms in this layer. The observation that in the latter

stage the oxidation is not too extensively retarded even with a relatively thick

alloy layer will be expiained by another mechanism. lt is generally assumed that

Si atoms migrate into Ta metal and then the released Ta atoms at the interface

will combine with oxygen atoms because oxygen atoms readily migrate throughthe oxide. lf Si atoms oniy migrate towards Ta in the alloy layer and Ta atoms

only combine with oxygen, then Si atoms will be forced to migrate into the Ta

metal leaving oniy Ta oxide behind, so that the TaSi, layer will not change its

thickness even after oxidation for a long time.

   we then examined whether the above stated assumption of the oxid. ation of

alloy layers was valid by using 49 hr-treated test pieces. However, th’e change

in the thickness of alloy layer could not be clearly recognized before and after

the oxidation for about l hr because of the small variation of each alloy layer

formed under the same conditions and the irregularity of the alloy thickness at

various positions of the same test piece. Accordingly, it is suggested that the

flux of Si atoms will decrease with the thicker alloy layer on account of the

decrease・in the concentration gradient, so that Ta atoms remaining on the surfaces

per unit time will be increasingly decreased by thicker layers, resulting in the

decrease of the oxidation of the alloy.

   Since it is recognized from the EPMA penetration curve in the sectioned ailoy

layer that there are some amount of Si even in the oxide layer and that Si is

much enriched in the alloy near the alloy/oxide interface, the above assumption

will be partly valid.

5. Conclusion

   Ta sheets were siliconized in the pure Si vapor in an evacuated quartz ampoule

up to 49 hr at 10000C, and then oxidized in flowing air in order to compare with

the results of oxidation of pure Ta sheets for 15 min at IOOOOC. The results

obtained were as follows:

   (1) Siliconizing of Ta sheets obeys a parabolic rate law and shows a two-stage

behavior, the first is formation of a lower silicide layer (Ta,Si,) with a slower

9 On the Atm◎spheric Oxidation o正Pure and Siliconized Ta M、etal Sheets at 1000。C 147

rate, and next after about 9 hr the second stage corresponding to the formation

of a higher silicide layer (TaSi,) with a higher growth rate occurs.

   (2) Oxidation of pure Ta sheets showed an anomalous behavior in which the

corners of a test piece were strongly oxidized and curled back in spite of the

fact that the center of the sheets were of a compact type. However,both oxides

show the same compound (Ta,O..) with different growth orientations at each po-

sitions. By using alumina markers this chemical process was determined to

proceed with the inward diffusion of oxygen atoms through Ta oxide.

   (3) Oxidation of siliconized Ta sheet showed a greater resistance with the

increase in the amount of siliconizing and the behavior of oxidation was divided

into two stages corresponding to the formation of two kinds of silicides. The

former stage showed a large decrease of corrosion because of the protection of

oxide against further oxidation and the latter one did not show too large a

resistance and could be controlled with the inward diffusion of Si atoms through

the silicide layer after the formation of an initial oxide layer. ln the present

experiment the oxidation of Ta sheets showed a retardaition down to about one

hundredth on account of the siliconized layer formed in 49 hr at 10000C.

Acknowledgement

   The authors wish to express their hearty thanks to the Shinetsu Chemical

Industry Co. Ltd. for supplying high purity Si metal used in the experiment.

1)

2)

3)

4)

5)

6)

7)

8)

9)

10)

11)

12)

                             References

Peterson R. X. W. H. Fassell, Jr. and M. E. Wadsworth: Trans. AIME, 20日目(1954), 1038.

Gulbransen E. A. and K. F. Andrew:Trans. AIME, 188 (1950), 586.

Bakish R.: J. Electrochem. Soc. 105 (1958), 71.

Michael A. B. : “The Oxidation of Co-base and Ta-base Alloys”, Reactiye Metals, Vol. 2,

   Metallurgy Soc. Conf.,工959, Interscience Publisher Inc., New York.

Arbrecht W. A. W. D. Klopp, B. G. Koeland, and R. 1. Jaffee: “Reaction of Pure Ta

   with Air, N2 and 02” AIME Symposium on Preparation, Purification and Properties

   of Refractory Metals, Chicago, Nov. (1959).

Dooly R. B. and J. Stringer: Corrosion Science, 10 (1970), 265.

Dooly R. B. and J. Stringer: X Less-Common Metals, 24 (1971), 139.

Klopp W. D. D. J. Maykuth, and R. 1. Jaffee: Trans. ASM,1 53 (1961), 137.

Kubaschewski O. and B. E. Hopkins: “Oxidation of Metals and Alloys”, Butterworths,

   London (1962), 2nd ed., 24.

H’ansen M. and K. Anderko: “Constitution of Binary Alloys” McGrow-Hill, New York   (1958), 1194.

Kubaschewski O. and E. LL. Evans : “ Metallurgical Thermochemistry ”, An lnternational

   Series of Monographs edited by G. V. Raynor, Pergamon Press, LoBdon (1956), 2nd

   ed., 272; 274.

Powers R. W. and M. V. Doyle: L Appl. Phys. 30 (1959), 514, and Trans. AIME, 209

   (1957), 1285.