Nucleation of helium precipitates in nickel observed by HDS

560

Journal of Nuclear Materials 122 & 123 (1984) 560.564 North-Holland, Amsterdam

NUCLEATION

OF HELIUM PRECIPITATES

IN NICKEL OBSERVED BY HDS

A. van Veen, J.H. Evans*, L.M. Caspers and J.Th.M de Hosson** Applied Physics Department, Delft University of Technology/IRI Mekelweg 15, 2629 JB Delft, The Netherlands * Metallurgy Division AERE Harwell, Oxon UK **Materials Science Center, University of Groningen, Nijenborgh 18, 9747 AG Groningen, The Netherlands

Thermal Helium Desorption Spectrometry (HDS) has been used to study the room temperature nucleation of helium precipitates at point defects inNi(llO), notably HeV defects at depth - 20 nm below the crystal surface. Helium is injected into the crystal by 50 eV He ion-irradiation which causes no atomic displacements. It has been observed that He,V defects with occupation from n = 2 He to n = 4 He bind helium equally strongly,but weaker than for HeV. For n35 He the binding increases rapidly. The observed behaviour is attributed to helium induced trapmutation and agrees qualitatively with results of atomistic calculations in nickel for this case. Helium precipitation at near surface trapping sites is held responsible for the observed increase of helium release ternperatures with helium dose when an undamaged crystal is irradiated. Preliminary TEt! observations of Ni specimens irradiated with 50 times higher helium doses than the maximum dose used in the HDS experiments indicated planar clustering of the helium. 1. INTRODUCTION

emission

The generation

of helium in metals exposed

the intense neutron irradiation tors has a considerable ment of radiation binds strongly

other defects,e.g. the metal matrix

Helium

vacancy com-

and

stabilizes

nuclei. Though helium

on the develop-

in these metals.

to vacancies

plexes and therefore

of fusion reac-

influence

damage

to

the early void

initially binds less to

foreign atoms substituted in I,2 , clusters of self-intersti-

of prismatic

conversion

loops, platelet bubble

and helium cracking

dislocations4'5'6.

processes

Helium desorption

at

measure-

ments provided

the information

submicroscopic

stages of precipitation.

on the early, For the

bee metals MO and W a large number of data has been collected

on multiple

helium trapping at

defects which enable; Fxtrapolation visible

precipitates

to TEM

’ . THDSexperiments

metals are rather scarce.For

onfcc

Ni(lOO) Kornelsen

tials and most of the dislocations, thebinding

and Edward B report on small helium vacancy com-

energy

gradually with the amount of

plexes but no attempt has been made so far to

by these defects

study precipitate

increases

helium collected

leading to

Low energy helium

injected at room tempera-

ture into metal specimens malisation

migrates

in the metal and therefore

controlled

after ther-

freely towards defects sites deeper

the nucleation

can be used to follow

and growth of precipitates

way. For molybdenum

in a

application

of

this method has led to the discovery with TEM of the platelet morphology

of precipitates

ted at small point defects3 yielded

information

growth by

0022-3 115/84/$03.00 0 Elsevier Science Publishers (North-Holland Physics Publishing Division)

tained on the growth of helium precipitates previously

prepared HeV defects

in Ni(ll0) by

helium filling up to 50 He per precipitate. section 2 experimental

In

details are described

followed by results and discussion Conclusions

at

in section 3.

and final remarks are given in sec-

tion 4.

nuclea-

and further has

on platelet

growth.

In this paper we report on the results ob-

some form of helium precipitates.

B.V.

2. EXPERIMENTAL The helium desorption

spectrometer

and the

561

A. van Veen et al. f Nucleation of helium precipitates

experimental

procedure

used in this work is

r

similar

to that used in previous precipitation studies of helium in molybdenum 697 . ANi(ll0)

s, ’

Sl

I

H

G

crystal of purity better than 99.995% has been sputter cleaned and heated may times to 15OOK (ramp heating 40 K/s) in UHV. Inherent to the method, thecrystal time a desorption HeV defects 1 keV He-ions

is heated spectrum

to 14DOK every

is recorded.

are produced

by irradiation

(dose 1.5~10~2 cme2) followed

ramp annealing

A desorption

trum taken after this treat~nt

spec-

shows a single

peak due to HeV dissociation

(HeV+HetV)

corresponding

l.lx101THecm-2

with the release of

and thus indicates

number of HeV defects. tation experiments repeated

by

to 740K (40 K/s) and cooling

down to room temperature.

desorption

with

the above treatment was

but with intermediate

irradiation

an identical

In the helium precipi-

50 eV He-ion

at varying dose before the crystal

finally was degassed.

3.1. Helium

induced "trap-mutation"

Fig.1 shows helium desorption HeV defects

spectra for

decorated

by 50 eV He-ion bombard-2 ment with doses varying from 0 to 1Ol5 He cm . The bottom spectrum HeV dissociation. spectra reveal

shows the H-peak asigned Further helium desorption

the occurrence

of new peaks deno-

ted M and K.TheS-peaksareobservedalso undamaged

crystal

to

is irradiated

when an

with low energy

16

12

9

4

CRYSTAL

3. RESULTS AND DISCUSSION

TEMPERATURE

t=‘K)

FIGURE 1 Helium desorption spectra for a Ni(ll0) crystal containing HeV defects which has been irradiated with varying doses of 50 eV He-ions. The average number ofextraheliumtrappedperinitial HeV defectisindicated. Peaksand series ofpeaksare indicated by S (surface related) and G,H,M,K (see table 1). ments by He(Ethresh

- 100eV)8.They

ascribed

these peaks to helium release from surface ted sites located within

the dynamical

rela-

range of

helium. Therefore, only the G peak and the high

the He ions at some lattice units of depth be-

temperaturepeaks

low the surface.

multiple average

should be attributed

helium filled vacancies

to the

(HenV). The

resulting

number of helium cn> per defect is ob-

tained by integration

of the helium content

H,M and K peaks and dividing

inG,

this number by the

initial number of HeV defects. With regard to the S-peaks Edwards

Kornelsen

below the threshold

irradiation

and

evidence

from the nresent work is reported

in

section 3.4. The evolution with dose of the H and G peak is shown in fig.2.

On helium filling the H

peak remains constant whereas

found for Ni(100) similar

ture peaks for He-ion

Further exoerimental

the G peak grows

and reaches a maximum helium population.

There-

low tempera-

after the H and G peak both decrease while

at energies

total number of helium trapped at the defects

energy for atomic displace-

increases

apnroximately

linearly.

the

562

A. van Veen et al. /Nucleation of helium precipitates

TABLE 1. Helium release data for He,V complexes



n

I !

1 H 900 2 G 720 I$ ;; II

10

5

; M 750

2.0 1.0

; 9 10K 950 II

0.5

I”

10'3

10'4 HELIUM

ION

II 0 II

10'5

DOSE

50 1400

(cmT2)

FIGURE 2 Peak population of H- and G-peak, and the total number of trapped helium vs the helium ion dose (ion energy 50 eV). Calculated curves (H,G) are given for nmax=4 and nmax=5 (see also the text). Filling theory'

applied

deeper

(- 20 nm average depth is calculated

(- 0.5 nm) of the 50 eV He-ions, ber of defects

2.82'I 1.81 1.70 1.95 2.20 1.62

emission. HenV2+He thermal >2.5 vacancy trapping. He,V,*He 'V

1.02 1.16 1.38 1.41

1.25 1.37

Ni which have trapped

i helium

as follows:

E"yHe = helium migration H-peak population

(see calculated

temperature are observed

is observed

= j Ni d

curves in fig.2). spectra no peaks associated

defects are observed

than the G-peak.

at lower

Instead M-peaks

between G and H and around the H

number of HeV, and p is a filling constant dependent on the trapping size of the defects

and

depth of the helium. The ave-

rage number of trapped helium per defect ci>= x. (note that = + 1). The experimental ling curves in fig.2 are matched when it is assumed

fil-

reasonably well

that He,V defects with n=2 all helium,

in excess of one,

to the G peak as follows =

from He,V with n<4 (nmax=4He).

release temperature

"trap mutation".

nments and estimates

we call

Peak assig-

of the dissociation

ener-

gies involved are given in table 1. The results 10 of atomistic calculations by Wilson et al and De Hosson et all:

also given in table 1, predict

quite well the trends found for He,V dissociation from n=l to n=4 equal binding of the 2

In particular the nearly th nd helium is through 4

in agreement with the above peak assignment

for the 7th through

the 5th helium which only

beyond the

the onset of this rapid increase of the helium

binding

f i Ni i=l

and that the H -peak gets contributions

As in molybdenum

the G-peak. The calculations

3

G-peak population

already at temperatures

H-peak temperature.

where P is the helium dose, No is the initial

to n=4 contribute

energy

peak. Upon average filling of
Ni = No (l/i!) (uP)'exp(-VP)

the implantation

2.05 1.63 2.09 2.04

*first order release; nential factor -IDI s-1 is assumed;E ?.I 9pH'e'+BxYP =E 9 e+EM*He

with He,V (n#)

gives the num-

He,V

2.631° I0 1.44 1.35 1.51 1.35 1.76 1.40

In the desorption

) than the implantation depth

HenV2

HeV -+He 2.34*) He2V 'He 1.82 He3V 'He ' 'He ' He$P;He He4V 21.8

to the present case,

where the helium traps (HeV) are situated much

with MARLOWE

release mechanism He,V

4 /:; /

/

peak/ TinK

periment. emission

However

is not observed

in the ex-

it might be possible

of a mutation

of

predict lower th helium than for 10

produced

that

interstitial

(MPI)"

occurs before helium is released,

e.g.

+ He V n2

discrepancy

+ I

discussed

He,v2 -f HeJ2

f He

from EB=2_25eVfor

He is concerned.Asimilar

found for trapmutation

10

In fig.3

filling of the HeV defects.

n=7 toEB=0.3eVfor

release temperature the He,V defects

amounts

increaseof

binding after

to - 2 eV and explains

the

He release tem-


(n>lO)

results are shown of further helium

n=10 which makes the above process likely to

experimental

in Mo is

in refs.7 and 12.

occur at - 9 or 10 He.Thehefium MPI emission

the release

3.2. Growth to larger precipitates

~FI-b~~d~ng energies quoted by WSlson et al decrease

A discrepancyexistswhere

ofthe5ththrough8th

via defect reactions He,V

perature.

Xt appears that the

of the helium attached

increases

to

rapidly withdose;

He per defect the temperature

at

has in-

creased to 1400K (0.8 T,). It should be remarked that trapping of thermally cies during heating

Thus

the precipitates

to equilibrium

(EL c $)=2.5

is calculated

per defect at 9SOK. perature

eV more

to be trapped this tem-

beyond

will firstly convert

bubbles before dissociation

bubble diffusion

takes place. Therefore

parison can be made with calculated ding energies

vacan-

plays a role in the helium

release process. Assuming than one vacancy

generated

or

no com-

helium bin-

of helium precipitates

of this

size (n>lg),

3.3. Surface related precipitation The earlier assignment

of S-peaks

to release

from surface related helium traps was given support by experiments

which

involved low ener-

gy (SO eV) Ar-ion

irradiation of the crystal. -2 A dose of 1014 Ar-ions cm was sufficient for complete

removal of the S-peaks, whereas

doses as high as lOI

of the HeV defects was observed. dynamical

at

cm-2 only partly removal

interaction

Apparently

the

of the argon projectiles

or nickel recoil atoms with the shallow

implan-

ted helium caused its release. The deeper situa-

, 4

A 8

0 12

FIGURE 3 Hel"ium desorption spectra fof5Ni(LJ0) irradiated with varying dose (0-l.SSxlO cm" ) 50 eV Heions. The two spectra plotted for every dose number corresnond with a) a crvstal containinq previously prepared HeV defects and b) without these defects. The average number of helium trapped per HeV defect fs indicated.

ted HeV defects were affectedonlybythe tion of irradiation which migrated

produced

to deeper

layers and caused re-

lease via the recombination Though the evolution

reaction HeV+

vacancies.

the evaluation

growth at bulk defects

dicate how precipitates

I +He.

with dose of the surfa-

ce peaks in fig.3 complicates precipitate

interac-

self-interstitials

of

it might in-

grow in the absence of

It is observed

that trapping of

564

A. lian Veen et at. / Nurleation ~~h~l~urn precipitutes

helium in shallow

traps leads by increased

ling to helium apparently

fil-

in deeper traps (fig.

3 top spectrum). interstitial

A scheme for precjpitation of 10 helium given by Wilson eta1 might

be applicable

for this case. These authors quo-

500 K probably will prevent the growth of these precipitates

sothata

deeper defects

The near-surface precipitation

constant

zone 'damaged' bytheabove

may have some similarity with the

te a binding energy of - 1 eV for He5 precipita-

helium rich surfaces

te and a

formed on nickel surfaces.

MPI binding energy of .6I eV which

make it likely that MPI emission temperature

occurs at room

or slightly beyond. The He5V defect

thus created binds helium stronger

Helium desorption

spectra taken beyond the critical dose for blistering I3 show low temperature

The authors TEM observation

for valuable of Ni(llD) specimensthatwere

uptodosesof1017He/cm2,

50 eVHe,

vealed a variety of images with sizesofthe der of 10 nm. Insomecasestheimages milar appearance

as the platelets

However,

reor-

showed sifound in molyb-

also groupsofsmall

bubbles

desorption

(-450K)

The number of helium atoms trapped per defect cannot be extra polatedwell results,

becausethehelium

measured

only to-50

more the continued

fromthepresent

filling
He/trap

(fig.2). Further-

growth of surface relatedpre-

cipitates might have causedareductionin flowofheljumtothe

HDS

interiorofthe

the

crystal.

are indebted to K.T. Westerduin

assistance

1. E.V. Kornelsen and A.A. van Gorkum, J.Nucl.Mat.92 (1980) 79. 2. G.J. van der Kolk, A.S. Hydra, A. van Veen and L.M. Caspers, Nucl.Instr.Meth.209/210 (1983) 1047.

J.H. Evans, A. van Veen and L.M. Caspers, Nature 291 (1981) 310.

4. J.H. Evans, A. van Veen and L.M. Caspers, Scripta. Met. 15 (1981) 323. 5. J.H. Evans, A. van Veen, J.Th.M. de Hosson, R. Bullough and J.R. Willis, this volume. 6. A. van Veen, L.M. Caspers and J.H. Evans, J.Nucl.Mat. 103/104 (1981) 1181. 7. A. van Veen, L.M. Caspers,

4. CONCLUSIONS The HDS results obtained tation at mono-vacancies The occurence

for helium precipi-

in nickel reveal:

oflltrap mutation" of He,V at

n _ 5 He. High thermal stability

of relatively

helium vacancy clusters agreement

mistic calculations.

small

(20-50 He).

with results of atoA discrepancy

exist for

the binding of the 5ththrough 8thhelium

atom.

In addition, evidence is found for considerable growth of precipitates

near the Ni(ll0) surface,

Future work will be aimed to circumvent problem of near-surface Helium injection

the

helium precipitation.

at elevated

temperatures

e.g.

in the THDS exneriments,

REFERENCES

3.

lying on {lllj planes could be identified.

Qualitative

have

ACKNOWLEDGEMENT

3.4. TEM RESULTS

denum3'6.

created when blisters

similar to the SI peak in this work.

than the

He5 defect.

irradiated

flow of heliumto

is guaranteed.

J.H. Evans, W.Th.M. Buters and Rad.Effects

in print.

8. E.V. Kornelsen and D.E. Edwards jr., in: Appli~atjons of ion beams to metals, eds.S.T. Picreaux et al (Plenum, New York, 1974)p521. 9. A. A. van Gorkum and E.V. Kornelsen, tion Effects 42 (1979) 93.

Radia-

10.W.D. Wilson, C.L. Bisson and M.I. Baskes, Phys.Rev. B24 (1981) 5616. ll.J.Th.M. de Hosson, L.M. Caspers and A. van Veen, Radiation Effects, in print. 12.L.M. Caspers, A. van Veen and T.J. Bullough, Radiation Effects, in print. 13.3. Ehrenberg, B.M.U. Scherzer and R. Behrisch, Radiation Effects, in print.