General Disclaimer. One or more of the Following Statements may affect this Document

General Disclaimer

One or more of the Following Statements may affect this Document

This document has been reproduced from the best copy furnished by the organizational source. It is being released in the interest of making available as much information as possible.

This document may contain data, which exceeds the sheet parameters. It was furnished in this condition by the organizational source and is the best copy available.

This document may contain tone-on-tone or color graphs, charts and/or pictures, which have been reproduced in black and white.

This document is paginated as submitted by the original source.

Portions of this document are not fully legible due to the historical nature of some of the material. However, it is the best reproduction available from the original submission.

Produced by the NASA Center for Aerospace Information (CASI)

t

G ^'

r

NASA TECHNICAL '240PANDUM _{NAV, TM-75129}

EFFECTS OF Fe 203 _{ADDITION ON THE NITRIDATION OF SILICON POWDER}

Y. Hasegawa, Y. Inomata, K. Ki,jima, and T. Matsuyama

(NASA-TN-75129) EFFECTS OF Fe203 ADDITION N78-11196 ON TEE NITRILATIO:1 OF SILICON POWDER

(National Aeronautics anal Space

Administration) 16 p HC A02/NF A01 CSCL 11D G3/24 Unclas

Translation of "Shiricon Funmat t,su no Chikka Katei ni Oyobosu Fe 2 0 3 Tenka no Eikyo," Yogyo-_{Kyokai-Shi, (Journal of the}

Ceramic Society of Japan; Vol. 85 (2), No. 975, 1977, pp. 83_-90.

NATIONAL AERONAUTICS AND SPACE ADMINISTRATION

i

t ( _{STANDARD TITLE PAGE}

I. Report No, 2. Government Accession No. J. Recipient's Catalog No. S 7M

-7124

4. Title and Subtitle S. Report Date

EFFECTS OF Fe 2 0 3 ADDITION ON THE August, 1977

a, Performing Organization cad. NITRIDATION OF SILICON POWDER

7, Authors) 0. Performing Organisation Report No.

Y. Hasegawa, Y. Inomata, K. Ki,jima

lo, work unit No,

and T. Matsuyama, National Institute

for Research in Inorganic Materials

11. Contract at Grant No.

NASw-2790

4, Performing Organization Name and Address

10, Type of Report and Period Covered

Leo Kanner Associates

Redwood Cii:y, California 94063

Translation 12. Sponsoring Agency Nume and Address

National Aeronautics and Space

Administration, Washington, D.C. 20546

14, Sponsoring Agency Code

15. Supplementary Notes

Translation of "Shiricon Funmatsu no Chikka Ketei ni

Oyobosu Fe203 Tenka no Eikyo," Yogyo-Kyokai-Shi, (Journal

of the Ceramic Society of Japan); Vol. 85 (2), No. 97 8 , 1977, pp.

83-90.

Id. Abstract

The reaction of silicon powder and nitrogen was studied III

the range of 1300-1400°C. When an addition of Fe 0 was

more than 0.8wte, the reaction was linear and com^Oed to samples with no Fe 2 0 , the reaction velocity increased 5-?0

times. The reaction were mediated by the process of

peel-ing and crackpeel-ing in a thin layer of Si

2Nformed on the

silicon particles or on the surface of tAe Fe-Si melts. As

the addition of Fe 0 3 increased, the rection activation en

ergy for highly puK samples decreased. Fe 0 which

exceeded the Si N 4 solubility limits was fiK ly converted

to d-Fe.

17. Key Words (Selected by Author(s)) 18. Distribution Statement

Unclassified-Unlimited

19. Security Clauif. (of this report) 20. Security Classil, (of this pope) 21. No, al Pages 22. Price

Unclassified Unclassified 1 f i [ 1 NASA•NO

I

l

4

^

I

j

EFFECTS OF Fe 2 0 3 ADDITION ON THE NITRIDATION OF SILICON POWDER

X. Hasegawa, X. Inomata, and K. Kijima

National Institute for Research in Inorganic Materials, Ibaraki--ken T. Matsuyama

R and D Division of Noritake Co., Ltd. Nagoya, 451

The reaction of silicon powder (purify: 99%, average particle

size 15 lam) and nitrogen (0 2 <0.5 ppm, dew point<-60°C) was studied

in the range of 1300-1400°C. When the addition of Fe 2 0 3 to the

silicon was more than 0.8 wt%, the progress of the reaction was linear and compared to samples with no addition the reaction veloc-ity increased 5-10 times.

The reactions were mediated by the process of peeling and cracking in the thin layer of S1 2 N 4 formed either on the silicon

particles or on the surface of the Fe -Si melts or by the process

of dissolution eduction from the Fe-Si solution. As the addition

of Fe 2 0 3 to the silicon increased, the reaction activation energy

for highly pure samples decreased from about 160 kcal/mol to about

133 kcal/mol. Fe 2 0 3 which exceeded the Si 3 N 4 solubility limits

was finally converted to a-Fe. The solubility limit of Fe in

S1 3 N 4 was in the range of 0.24-0.49 at% at 1400°C

1. Background /46

There has been much research carried out concerning the re-reaction of silicon powder and nitrogen. It is known that certain impurities inside the silicon powder used for nitridation acceler-ate the nitridation reaction. Among these impurities, iron

compounds are very effective [1, 2]. This kind of research is

being conducted continuously, but there has been no research

dis-cussing the velocity of the reaction 13, 41.

* Numbers in the margin indicate pagination in the foreign text.

^

►

I

^

E

Factors other than impurities in the silicon have been re-ported as affecting speed of nitridation, for example minute

quantities of oxygen in the nitrogen atmosphere [5-10]. For this experiment we established a comparatively pure nitrogen atmosphere

(0 2 <0.5 ppm, dew point<-60 0 C), and studied the effects of Fe 2 0 3 ad-dition, chiefly in respect to velocity [11-14]. The results have been integrated and reported below.

Many additives have been reported effective in the accelera-tion of nitridaaccelera-tion. The reason we have chosen iron, which is easy to combine by a silicon pulverization process, is that it is the most universal of powders for industrial use, and an oxide, be-cause minute amounts of powder are easy to procure as reagents.

2. Experiments

2.1 Materials

The silicon powder used in the experiments was 99.99% pure and was from the Taka,jundo Chemical Co., Ltd. Lump silicon was pulverized with a stainless steel ball mill. The powder from which impurities had been removed by acid treatment had average particle size of about 15 pm, and the chief impurity was iron.

0.007 wt% was detected by X-ray fluorescence analysis. The pow-der was also analyzed for aluminum and calcium, but none was de-tected. The powder was not analyzed for oxygen content. The particle distribution was measured by the Omnicon Pattern Analysis System of the Shimazu Manufacturing Co., Ltd. The results are shown in table 1.

The nitrogen used for the reactions was obtained from a gun of highly pare liquid oxygen by evaporation (0 2 <0.5 ppm, dew point<-60 1 C). T,-,e Fe 2 0 3 used as additive was of special quality with particle size below 0.5 Um.

2

2.2 Nitridation

One gram of the silicon powder and Fe 2 0 3 was precision

weighed and placed in an agate mortar. The mixture was moistened

with 1 or 2 drops of distilled water and after mi-ing with care so

as not to pulverize further, was shaped into a disc using a 15 mm

steel mold. After drying a day and a night, the disc was reweighed

and the experiments carried out. The packing rate of the powder

was about 60%, and the thickness of the sample was about 4 mm.

TABLE 1. SIZE DISTRIBUTION OF After drying, the sample

SILICON POWDER USED IN THE

EX-was placed in an alumina PERIMENTS

(Al 2 0 3 >99%) crucible and

lowered into a vertical tubular Diameter of Number of overaita Percent

particle oaa> particle (wt";) _{furnace, the core of which is}

Total

0.5.

84ue 100.9

5779 10019 an alumina tube. A

thermo-.0 z.9

2 2041 99.97

4929 99.50 balance was a

pplied

continuous-4.0

1.0

200 90.7

u72 97,0 1 Y • After the atmos heric ni-P

5.0

0.0

1015 91.5

811 92.0 trogen had been displaced, the

7.0 8.0

010 00.4

442 87.0

tem

0.9 19

288

70.9 80

.7

zBB

1501C/h until the determined

14

]I

102 08'1

127 59.1

temp

19, is

57 1910

40 41.0 while injecting 500 cc of

ni-20

zz

40 21.5

20 4815

tro

en

4' 44

_^._ '2' 2.._ temperature was reached, it

was maintained,and the increase in weight and the temperature

was recorded. The experimental temperatures were near 1300, 1350,

and 1400°C. The amount of additive was chosen after consulting

Suzuki's results [2], and were by wt% of the unreacted mixture 0,

0.8, 1.6, 3.2 and 6.2.

/47

Ii

3. Results

3.1 Rection Velocity

,

i

The results concerning

! Y

01 temperature and reaction

veloc-°of ity are in tables 2-1 through

oa l i i ! 2-5 for each F'e 2 0 3 addition.

The items concerning the re~ action are not from the begin-. r

F4r, I. C,drul„ic,I rci,,,i„, 1JOW1. 1 a Illd T ning because the data are from

III N1, cn for Ii,o „wc„11 _{jil,w&r In after the temperature inside}

TIIIIIC I, I:I,d I ,pll,l YIIIIWR ll,, uld of

1110 rv°ntr,l,• _{the furnace had become}

regu-lar. The reaction velocity A (um/h) is from the increave in weight measured by the thermobalance, which is the change in the speed of decrease in the size of silicon particles. This was done as previously reported [12], but this time, dead time was not

calculated. AT/At were cllculated separetely for 511 reaction

range units for the value of A. The relationship formula neces-sary for this conversion is 12

X

= E

n=l

x: silicon after reaction, the weight as opposed to the amount initial silicon

m: number of particle layers

T: decrease in size of silicon particles, the thickness of silicon after reaction

Rn: the average size of particles n n: the weight of particles n

The relationship between x and T according to this formula are shown in figure 1.

The value of A in table 2, i.e. the calculation of OT/At, was calculated by contrasting the amount of silicon in the sample

4

TABLE 2-1. CALCULATED REACTION

HATE,A m/h. ADDED Fe 2 0 3 : 01;

Reaction Range or JT Js A

tent Into rend Olt

(C) ( o) Gnn) (10 (mill) 1350 1357 10.15 0.114 6.82 0017 1100 1408 1015 0.114 0.24 0,415 1108 10-20 0.125 0.16 0.781 1408 20-25 0.130 0.20 0.623 1408 25.30 01117 2.00 0.071 1408 30-35 01159 6.50 0.029

` Table 2 . 2. Culculuted renctlon r010. A pin/h. Added Fe,e.l o.ewt%

Reaction Range of JT Jr A

lem0eralure reaction (C) (56) Gnn) '(h) Ganth) 1500 130D 79•,6 51111 2.11 9.917 1310 15-20 0.125 1.14 0.053 1310 2025 0.130 1100 0.085 1310 25'30 0.147 1.78 0.083 1310 3036 0.169 1.21 U.131 1350 1309 1046 01114 0.57 01170 1301 15.20 0.12;1 1.08 0.110 1302 20-25 01136 0.40 0.317 1363 20-30 01147 0.37 013]7 1303 30'35 0.109 0.38 0.418 1363 35 40 0.173 0.39 0A 1303 4045 0.187 044 0.310 1363 45.60 0.207 0.62 0.331 1363 50-55 0.230 1.02 0.225 1100 1400 10 is 0.114 0.31 0.368 1401 1020 0.123 0.37 0.338 1401 2025 0.135 0.20 0.623 1401 20'30 0.147 0.17 0.865 1401 3035 0.169 0.11 1.45 1401 35.40 0.173 0.12 1.44 1401 40.45 0.181 0.12 1.50 1401 45.50 0.207 0.14 1.48 (sample weight-amount of Fe 203) with the detected increase in weight. Strictly speaking, since the reaction system was an open one, there were identi-cal reactions with silicon

and the alumina crucible and between silicon and the oxygen in the silicon, the oxygen in the atmosphere, and the oxygen in the Fe 2 0 3 [4, 9, 15, 161 with the result that the

in-crease in weight did not always correspond to the reaction rate. However, as will be explained later, the scattering was very slight and has been ignored in the calculations of table 2.

3.2 The Condition of the

Sil con Nitride Formed

.w.

3401 55-65 0730 0.10 slot

The mineral composition of

3/01 56.80 0.780 0.17 1.61 p

1401 65-75 01787 0.17 1.69

1401 e7-70 0.319 0.21 1.33 the nitrided samples as

deter-

1101 _- 75-75 01357 0.37 0.997

mined by the X-ray diffraction

method were investigated

1171,

and the results arb shown in table 3. X-ray fluorescence analysis indicated that the iron additive did not scatter outside the sys-tem, but remained inside the nitrided sample. In samples with more than 3.2 wt% Fe 2 0 3 added, all of it remained as metallic

iron. The addition of Fe 2 0 3 in large quantities caused the gener-ationof Si 2 N2 0. The amount of 0 phase material in the nitride, in general, was greater at high temperatures.

5

r6

a

TABLE 2-3. CALCULATED REACTION

RATE. A m/h. ADDED Fe 2 0 3 : 1.6wt%

Reaction Itanxa of 17' 41 A

let" ^^^ore 1eo,^lon

l7 m) (h) O"t1t)

IJ00 iJ01 1015 D•II{ 1.0! 0.058

1307 15-20 0.125 1.00 0.069 1307 20 23 0.130 1.18 0.115 1301 25 00 0.147 1.22 0.120 1307 30-35 0.159 1.32 0.110 1307 3340 0.173 1.40 0.124 1301 4045 0.187 1.60 0.117 1301 46 so 0501 1.96 0.106 1307 50.55 0.230 3.04 0.076 13D7 55-60 0126 0.55 01037 1350 1363 1520 0.123 0.74 Or loo 1305 20125 0.136 0.68 0.200 1305 2530 01147 0,44 0.331 1305 30-35 0.160 0.35 0.454 t365 3540 0.173 0131 0.408 1365 40,45 0.187 0.32 0.681 1305 4560 0.207 D.40 0.618 1301 60.65 01210 0.40 0.6T6 1365 65.50 0.256 0.45 0.569 1365 60-65 01287 0160 0.513 1365 61 70 0.319 0.71 0.449 1365 70-TO 0.3337 0.90 0.397 1400 1196 1520 0.125 0.20 0.025 1390 20.20 0.130 0.20 0.080 1400 26-30 0.147 0.24 0.013 1400 30.35 0.159 0.16 0.994 1100 35-40 0173 014 124

Parts of samples with dif-ferent amounts of Fe 2 0 3 additive nitrid.d at 1400°C were lightly pulverized in an agate mortar and the resultant powde observed by electronic microscope. The results are shown in figure 2.

In all cases, minute

rectangu-lar crystals were ob-.,, •ed.

There was no special relation between the differences in the amounts of Fe 2 0 3 added and the configuration of the Si 2 N 4 formed.

4. Observations

4.1 Reaction Velocity

wu.

1400 40-45 0.157 0106 2

.34

Figure 3 is an Arrbenius

lot of table 2.

In table 2

1404 80^66 0.230 0108 2.85 p f

1401 65.00 0.286 0. 2.30

1407 OO.fiS 0.28T 0.12

12

2.09

h erbolic reactions were noted

1407 66.70

0.319

oa5

in systems with no Fe 0 added

_

0.

y

2 3

but linear reactions were noted when there were additions ex-cept got the initial period of the reaction. This tendency can be seen in figure 3. When different amounts of Fe 2 0 3 are compared in figure 3, the addition of Fe 2 0 3 to the silicon p owder acceler-ated nitridation, but addition of Fe 2 0 3 above 1.6 wt% was not al-ways effective. Except for low temperatures, the speed of nitri-dation accompanied the increase of additive but there was a

tendency for saturation. The nitridation speed at 1400°C with an additive of Fe 2 0 3 of more than 1.6 wt% was 5-10 times higher than when there was no additive. This effectiveness at acceleration was remarkable at low temperatures.

6

1 6 A rewlh Hem #4 bt Free %1 4► 'wl* •.0 as Y IT 71a 11.• 14 1 14 T IT 1 •4 1 2. 1. Ml.a 11 't t•2 4 2' Y4 v t • 94 3 3 • '2 t It 4 ^/4 11 v7 T 1 • vY 1 1U PKI 0 • 0 a D4 ••• 0 *a-•

UNIGI YAL PAGE la

OF

Pont QE;A_LJ7T 'rob iv s. mn.tal ,vntrlatmulon ..41, elated frunl X-ray data and liar w,

It" erh, ,, ..0. %d a..i It,. v '1 ./i.A rr•ultm (or tbr • t- it FVO), rr1,6-,1 kemUr.n h•.,;r. n... '^ '»Y ddbmrh^h _

b wl»r.dt,rr _{{ . ,} t I nr ^t ^I,N,/1 f wl't f w 1 0 1 Nttt a ! 1110 It 4 U a 11(to I^ m 1.)6o I. v 1490 It N 3.6 1104 l) Y 1350 12 2 4 r 14(w) I{ 4 M 3 0••• 13(m) 12 P 116x1 12 2 1 1 1400 1/ 1 V dete• ted

F 2••0 I1tt4 It Y 1 0 dirt- red

13511 1 1 1 U.o detn led

1490 11, 7 n detecaed

• l mtentY of .r t) yr to totel St•N. Inntted by IN um. bon,

•• Gwtd by the It."imtr,n .rl St,N,O

••• ,t fr wn• dvtr 1• d to all 9anpler ft.trA-d —th 3 2 »nJ o 2 w1;. I r,t 1, addtit ar

loo^i;r •Q

"/• ^

_f

Fiit^'

?

_{'"h e,Oj 6 ?}

_F

_:

F'Ir. ft. liirrtnm qurrugt.tphF nl pu"dered maniple nitrified at ah„ut WNW, Nlinrralu«t.Yl

data of thus manwive art• mhuwn In Table :1.

9 SO 50 60 61 62 63 64 (1/!) to' Fig. 3(1) too so I 20

1

101

100 -t ----T r -Fi,;, J 6 'Ia So + 1 i ° I I 10 S OS i° I of 8 005 o r_ SA S9 60 61 62 61 64 I n to loo so 20 GO 03 a 02 01 005 too ^Y^C2 3 ? °ti --^ SO }-10 03 0 ^ ea 02 01 ^-- -005

—L

Fig. 3. Arrhenlua plot of calculnivil unn rat". A /ah

in Tahle 2 1-2 5.

Tahlr 4 innti,rtt rncrk% ralr ulntrd frnrn I tK 3.

Addy 1 I . ,r P. Activnhnn enrray kcal molt u l.e Ins 3 2 ISU c.2 133

• 1S6' r and Iss"' k—1 mnl w,re rep-ted.

Table 1 4. Calculated reacton rate, A !nn/h Ad-lyd Fell),: ;1.1 w1% Nea,II,," N.n0. of jr Ir A W.11^raturo l.arllnn (i ) l A! (on) 11%) (ANA) 1100 1113 110 1/ U. 111 7 24 01001 1316 16 w o 173 O N 0 161 1116 to 23 0 116 1 A U 111 1716 15 3a U 117 1 16 0 129 1116 30 Is W150 1 M P is0 1316 16 60 0 171 1 16 0 129 1414 40 65 0.161 1.62 U 131 1116 65 60 0 24 1 64 0. 126 1116 6u 66 0 130 2 14 n 1"7 1316 IS 0o 0 236 1 ,V 0 126 1316 60 6s 0 217 2 60 U Its 1116 66 7o 0319 6 10 o. 4M 1360 1 I Is t0 0 lb 0 N 0 772 1'67 20 as 0.1% 0 6 1) 0 I97 1354 76 30 0 111 0 61 0 761 1160 1,0 30 0 0. 0 so 0.176 1362 36 60 0.173 0 66 0.376 1166 60 61 0 161 0.39 0.6H 1166 65 60 0 J*r7 0.61 0 1106 Ito$ 60 66 0 23 0 62 U 566 11,165 66 60 012'16 0 62 0 610 1365 60 65 0.167 0 66 0 016 :!a6 N 10 0.119 O.N ONO -- keaclv,,n Nnn11a n1 IT 1► .1 lemperalun rem, twn 7 91) 0.n11 h) h - 141.5 70 76 0.317 x.72 0.666 1600 1146 35 0.136 0.16 0.706 1347 2'1 30 0.147 0.16 0.419 t 349 3„ 39 0.100 0 is 0.473 11,1 35 131 0.173 0..16 I.(AN 11 1 61, 65 0.1471 a 16 1.17 1401 45 5o 0.201 0.16 1.74 1401 5113 0.M 0.114 1.211 Uot As 60 0.306 0.16 1.62 1601 60 66 0.267 0 16 1 710 1601 66 70 0 $19 0 16 1 77 1601 70 71 0,367 0 IN I a

Table 4 show: the reaction activation energy found from figure 3. The average valuee between 1300-140000 are

re-corded in the table. The

val-ues could not be recorded when there was no additive, but in the range where the addition of Fe 2 0 3 was small, the values of highly pure reactions which were 156 kcal/mol [11], and

1 1")8 kcal/cool [12], agree very

closely. As the Fe 2 0 3 additive increased the values tended to become lower. This was obvious in reaction with an additive of

6.2 wt%, and it is possible

that when the additive is large, the rate of reaction is deter-mined by a different mechanism.

4.2 The State of the Iron During Nitridation and r,,,^

.acid

---When Fe 2 0 3 coexists with a sufficient quantity of sili-con, as in the present experiment, it is estimated that reactions such as the following take place during nitridation.

Tmble 1 / Maculated reaction rate, A !'rn'11. Added 1-4,041 6.2 wt%

k.. tl.,n kma. .d

it

6en11M/alYle f1.IIUn

11 1 ` ^^ '„wl hi !;an hi 111111 111. Is 70 o.lib lIn 0-102 lilt vi 26 01126 1.10 0,IL4 13:1 16 30 0.161 1.26 0.119 1321 11, 26 0.169 0. W 9.16e 1,21 36 N o 1T1 0.66 0.2'11 1121 60 J6 01111 0.76 01216 1221 15 S0 0-207 0.99 0.23! 1121 Su AS n 210 1.00 0.:10 1211 55 60 0. no 1.10 0.213 1121 60 66 0.261 1 .2 0 tab 1121 06-70 0.310 1 H 0.2.16 1121 10 16 : 357 :.m 0179 1360 1.160 16-20 o lib 0.52 0.160 1151 :0 66 a 126 0.51 U 252 1256 25 3D 0 161 0.06 0.210 1260 70 36 0.169 0.62 11.256 136) 3;160 U.171 060 0.2x9 1360 60 66 0.161 0.117 0.160 1460 65 60 0.2117 0.62 0.193 an0 S+. 56 0 230 0 61 11, 661 !]6U S6 60 0.266 0.61 0.6:6 136, 6066 0.261 0.66 0.662 1160 65-10 0.319 0.19 01661 1360 To-711 0.351 0.63 0.567 1600 1395 25-20 0.167 0.22 0.669 1391 30 A U.16V _­10 1.69 13,09 35-60 0.173 0.16 0.961 1199 60 63 0.167 0.16 117 1600 15 61) so?0 0.10 1 29 1611 W 55 0.230 0.16 1 26 14 A 55 6o 01256 0.17 1.51 1196 60 66 0.297 0.17 1. i9 1606 t5 70 0 319 0.19 1.77 11011 T'1 76 _0.367 0.11 I.Y6

The similar reactions 51 3 11 4 -Fe 2 0 3 and Si-Si 3N4-Al203 are also thought possible [15,

163. In any case, when the

re-action system is open as in this experiment, Si0 is transferred out of the reaction system dur-ing the reaction period because it is volatile [15,163, and iso-lated iron or silicide such as

FeSi., is generated. It is

be-lieved that when the temperature is above 1206°C, the common melt-ing point of FeSi,,-Si [1$3, a si-licon solution of iron is gener-ated. If the reaction proceedr;

sufficiently, the solution limit of Si 3 N 4 ie exceeded, and the iron additive, as shown in table 3, exists as metallic Iron which retains its equilib-rium with Si 3 N 4 at less than one atmosphere of nitrogen [193.

r•

For example, in table three, for the sample which was nitrided at 1300 0 C with. a Fe 2 0 3 additive of 3.2 wtb, a lOp residue of sili-con was noted in the sample after nitridation. In spite of this, the generation of metallic iron was noted. This can probably be explained by lack of uniformity of reactions in the samples. As-cording to tagle 3, when more than 3.2 /;t/O' of Fe 2 o 3 (corresponding

to 0.49 at/'O Fe in the Si 3N4 when nitridation was completed) was

added, metallic iron was generated. Consequently, we can say that in the present experiment, the solubility of iron into Si 3 14, at dear 1400 0 C was in the range of 0.24-0.49 at/00. This value is

10

11 i 3 I I ^

almost in agreement with that of another experiment [193 in which the solubility of iron synthesized in Si 3N4with the same starting materials was in the range of 0.4-0.6 at%.

The amount of disassociated silicon noted by X-ray diffraction and recorde in table 3 is, in general, smaller than that calculated

from the results from the thermobalance. This is possibly due to _.w.

the fugacity of S10 mentioned earlier. When Fe 2 0 3 was added at 6.2 wt%, there was more than 100% nitridation of the sample at near 1400°C. This is thought to have some connection with the fact that Si 2 N 2 0 was noted in the sample.

4.3 Structure and Shape of the Silicon Nitride Formed

The Si '

4 enerated by nitridation of silicon in these

experi-ments was, as shown by table 3, largely u phase. As the tempera-ture became higher and the Fe 2 0 3 was increased, there was a slight tendency for a phase to increase, but not markedly.

The relation of temperature to the amount of a phase Is

pro-ba!'ly due to the fact that

_a

tem-peratures [17]. It is believed that first a phase is generated, and this by going through a process of subllmation,recrystalliza-tion is converted into a phase. The increase of a phase which accompanies an increase in Fe 2 0 3 addition is probably dur to the acceleration of the recrystallization in the solution 113, 171J. In Figure 2 can be seen some idiomorphic clear minute crystals thought to have been formed by crystal frowth.

4.4 Reaction Mechanism

It is believed that the reactions of table 2 in wLkoh Fe 203 was added proceeded almost linearly. This fact must be satisfied by the reaction mechanism. One corroboration is the process of

peeling and cracking of the thin layer of action or the silicon or on the surface o iron [12]. However, in the case of Fe 203 ticipates in the reaction; thus the S13N4 be attributed, for example, to the V.L.S. the dissolution eduction process recorded

S1 3 N 4 formed by the re-C the solution containing

additive, the melt par-thin layer can possibly mechanism [4,20], or to as S.L.S. [21, 221•

The average activation energy 130-160 kcal/mol (table 4) be-tween 1300-1400°C is comparatively near the diffusion barrier, i$6 kcal/mol [14], of nitrogen in Si 3 N 1 , synthesized under almost the same conditions using identical silicon powder. It was noted that for additions of above 1.6 wt% of Fe 2 0 3 (0.57 atjli in silicon as an Fe additive, 0.24 at% in completely nitride silicon) the amount of increase in reaction r— locity hardly changes, and the velocity of nitridation was five to tenfold. This fact is in ac-cord with separate results [19] where the solubility limits of

iron in S1 3 N 4 near 1400°C and in silicon at 1400°C were 0.4-0.6 at%. It also agrees closely with the fact that the autodiffusion vate of hycrogen in Si,N 4 raises one unit when the addition of iron is above the solubility limits [14].

The significance of the above is that the reaction pr,aceeds according to the diffusion rate. The progress of the reaction is linear because, as mentioned before, there is an intervention of peelong and cracking of the Si 3N4skin formed on the silicon or on the surface of the solution containing iron or an intervention of a dissolution eduction process. Hyperbolic reactions were noted in highly pure powder with no iron additive. The reason for this is not clear, but compared with previous research [12], particle size was coarse and it is possible that minute quantities of oxygen in the atmosphere participated in the reactions.

r

peeling and cracking of the thin layer of action or the silicon or on the surface o iron [12]. however, in the case of Fe203 ticipates in the reaction; thus the S13N4 be attributed, for example, to the V.L.S. the dissolutioi, eduction procesF recorded

S] 3N4 formed by the re-f the solution containing

additive, the melt par-thin layer can possibly mechanism [4,20], or to

as S.L.S. [21, 22]. .w

The average activation energy 130-160 kcal/mol (table 4) be-tween 1300-1400°C is connaratively near the diffusion barrier, 186 kcal/mol [14], of nitrogen in Si 3 N j# synthesized under almost the same conditions using identical silicon powder. It was noted that for additions of above 1.6 wt% of Fe 2 0 3 (0.57 at°0 in silicon as an Fe additive, 0.24 atN in completely nitride silicon) the amount of increase in reaction — locity hardly changes, and the velocity of nitridation was five to tenfold. This fact is in ac-cord with separate results [19] where the solubility limits of

iron in Si 3 N 4 near 140C°C and in silicon at 1400°C were 0.4-0.6 at%.

It also agrees cl r ^sely with the fact that the autodiffusion rate of hycrogen in '`,N4 raises one unit when the addition of iron is above the solubility limits [14].

The significance of the above is that the reaction pr-needs according; to the diffusion, rate. The progress of the reaction is linear because, as mentioned before, there is an intervention of peelong and cracking of the Si 3 N 4 skin formed on the silicon or on the surface of the soluticn containing iron or an intervention jof a dissolution eduction process. Hyperbolic reactions were

rioted in highly pure powder with no iron additive. The reason for this is not clear, but compared with previous research [12], particle size was coarse and it is possible that minute quantities of oxygen in the atmosphere participated in the reactions.

5. Summary

The effects cf the addition of Fe 2 0 3 on the reaction of sili-con (purity: 99.99%, average particle size; 15 um) and nitrogen

(0 2 <0.5 ppm, dew point<-60°C) were studied at temperatures in the

vicinity of 1300-1400°C. When there was no Fe 2 0 3 added, the

re-actions proceeded hyperbolically, because the particles were coarse.

When 0.8 wt% of Fe 2 0 3 was added to the mixture, linear reactions

were noted. In the range above 0 .8 wt%, the reaction velocity was

raised five to ten times. The Fe 2 0 3 added in excess of the

solu-bility limits of Si 3 N 4 remained as metallic iron after nitridation. /52 The solubility limits of iron in Si 3 N 4 are believed to be in the

range of 0. 24-0.49 at% at near 1400°C. A tendency was seen for

the reaction activation energy to decrease as the amount of iron additive was increased. In cases when the addition of iron was

small, the value was about 160 kcal/mol, and when the addition was

6.2 wt%, the value was 133 kcal/mol.

It is believed that the reactions were determined by the dif-fusion rate of nitrogen in the Si 3 N 4 . With the addition of Fe 203, the reactions bEcame linear. This was thought to be due to the intervention of peeling and cracking of the Si 3 N 4 skin formed on the silicon or on the surface of the solution containing iron, or that it is mediated by a dissolution edu.tien process.

^3

I

,

REFERENCES

1. Popper, P. and S. Ruddlesden, Trans. Brit. Ceram. Soc. 60,

603-26 (1961).

2. Suzuki, H., Bulletin of Tokyo Institute of Technology 54,

16347 (13 3

3. Gribkov, V., V.A. Silaev, B.V. Schchetanov, E.L. Umant• sev and

A.S. Isaikin, Sov. Phys.-Crystallogy.'16, 852-54 (1972).

4. Messier,D.R,. and P. Wong, J. Am. Ceram. Soc. 58, 480-85 (1973).

5. Naruse, W., M. No,jiri and M. Tada, Nihon-kinzokugakkai-Shi 35,

731-38 (1971).

6. Mangels, J.H., J. Am. Ceram. Soc. 58, 354-55 (1975)•

7. Blegen, K., "Special Ceramins No. 6," edited h,; P. Popper. The British Ceramin Research Association, Stoke-on Trent, 1975, pp . 223-244•

8. Wild, S., P. Grieveson and K.H. Jack, "Special Ceramins No. 5," ibid., 1972 , pp. 223-244.

9. Sin-Shong-Lin, J. Am. Ceram. Soc. 58, 271-273 (1975)•

10. Mitomo, M., J Am. Ceram. Soc. 58, 527 (1975).

11. Hxettinger, K.J., High Temp.--h Pressures 1, 221-230 (1969).

12. Inomata, Y., and Y. Uemura, YoNyo-kyokai-Shi 83, 244-248 (1975)•

13• Inomata, Y., ibid. 83, 497-500 (1975)•

14. Kijima, K. et ate., to be published.

15. Messier, D.R. and G.E. Gazza, J. Am. Ceram. Soc. 58, 538-40 (1975).

16. Tnomata, Y., K. Yukino and T. Wada, Yo&Eo okai-Shi 84, 254 (1976).

17. Inomata, Y., YoEyo-Kyokai•-Shi 82, 522-526 (1974)•

18. Hansen, M. Constitution of Binary Alloys, McGraw-Hill, Inc., 1958, PP . 711-717.

19. Yamamura, H., K. Ki,jjz a, S. Shirasaki, Y. Inori to and H. Suzuki, J. Mat Soc. 11 3 1754-1?55 (1976).

20. Wagner, R.S. and Ellis, W.C., Trans. AIMS. 233, 1053-1064 (1965).

21. Kaiser, W., and C.D. Thurmond, J. APpl. Phys. 30,, 427-431 (1959)•

22. Inomata; Y., and T. Yamane, J. Crystal Growth 21, 317-318 (1974)•

14

RE ERENCES

1. Popper, P. and S. Ruddlesden, Trans. Brit. Ceram. coc. 60,

603-26 (1961). —

Suzuki, IJ., Bulletin of Tok_{2 o Institute of '7echnolog54,}

163-77 (1., , .

3. Gribkov, V., V.A. Silaev, B.V. Schchetanov, F.L. Umantsev and

A.S. Isaikin, rjoy. Phys. -Crystal lo,! -,y. 16 , 852-54 (1 172.) .

4. Me3sler, D.R. and P. Wong, J. _{Am. Ceram. Soc.} 58, 480_{-85 (1973) .}

5. Naruse, W., M. No,jiri and M. Tada, Ilihon-kin,okugakkai-Shi 35,

731-38 (1971).

—

6. Mangels, J.H., J. Am. Ceram. Soc. 58, 354-55 (1975).

7. Blegen, K., "Special Cendmiris :do. 6," edited ',%; P. Popper. The British

Ceramin Research Association, Stoke-on Trent, 1975, pp. 223-244.

8. Wild, S., P. Grieveson and F.H. Jack, "Special Cerardn) No. 5," ibid.,

1972, pp.

9. Sin-Shone -Lin, J. Am. Ceram. Soc. 58, 271-273 (19,75). 10. Mitomo, M., J Am. C(-ram. Soc. 58, 527 (1975).

11. ibettin M. K.J., Iii T`erT).-limb Presstwes 1, 221-230 (1969).

12. II-romata, Y., and Y. Uemura, YYoW2= rokai-Shi 83, 244-248 (1975) .

13. Tnomata, Y., ibid. 83, 497-500 (1975).

14. Ki,j irm, K. et a'.1., to be published.

15. Messier, D.R. and G.E. Gazza, J. Am. Ceram. Soc. 58 538-40 (1975).

16. Tnomata, Y . , F. Yukino and T. Wada , Yei y c--1,7yokai-Shj 84, 2.54 (1976) .

17. Tnornata, Y. , YX 82, 522-526 (1974).

18. IJansen, M. Constitution of Binary Alloys, McGraw-I?--I ll, Inc., 1958, PP. 711-717.

19. Yamwawa, H. , K. Ki j ima, S. Siiirasaki, Y. IrlonLctt? and I1. Suzuki, J. Mat

Soc. 11, 1754-1''55 (19(6).

20. Wagner, R.S. and Ellis, W.C., Trans. AI^E. 233, 1053-1064 (1965).

21. Kaiser, W., and G.D. Thurmond, J. Aj-)P 1. Phys.

30, 427-431 (1959).

22. Inomat,a, Y.; and T. Yariane, J. Cr sta1 GroArth 21, 317-318 (1974).

14