Tải bản đầy đủ - 0 (trang)
2 Recent Mössbauer Studies on Fe in Si

2 Recent Mössbauer Studies on Fe in Si

Tải bản đầy đủ - 0trang

8 Nuclear Methods to Study Defects and Impurities in Si Materials



393



An “interstitial” iron mapping was considered to be possible so far in B-doped Si

using Photo-Luminescence (PL) [31] and W-PCD [32]. The analysis was based on

a Zoth and Bergholz model [33], taking into account the measurements on the carrier

life times before and after breaking Fe-B pairs. This mapping technique can only

be applied for electrically active Fe impurities, but not for inactive Fe impurities.

Accordingly, a direct observation of all Fe impurities including interstitial Fe and

Fe-B pairs appears to finally be necessary on an atomistic scale, in order to achieve

not only the picture for Fe impurities in Si materials, but also Si wafer and device

fully controlled processes.

Mössbauer spectroscopy appears to be ideal to characterize Fe impurities in Si

crystal [7–14, 34–36]. Recently, a series of the experimental investigations [36–60]

has been performed using Mössbauer spectroscopy on Fe impurities in p-type and

n-type Si materials such as single crystal Si, multi-crystalline (mc) Si wafers, and

even mc-Si solar cells. These results lead to the hypothesis that we need a new

picture for Fe impurities in Si materials: the Fe impurities in the Si matrix exist

not only on the interstitial sites with Feint 0/C1 and Feint C -B pair, but also on a

higher charge state of Feint 2C associated with defects as well as on the substitutional

sites with Fesub 0 and Fesub . Furthermore, these components are found to transform

to each other in the Mössbauer spectra by changing experimental conditions such

as temperature, external voltage, electron or light irradiation and external stress as

well as by changing the Fermi level, carriers and their concentrations, and the device

structures in the vicinity of 57 Fe probes.

In the following sections, first of all, recent results from Mössbauer spectroscopy

will be discussed, summarizing the isomer shifts for different Fe components.

Secondly will be presented: Mössbauer absorber spectra after annealing at 1273 K,

measurements at high temperatures from 1273 K down to room temperature,

Mössbauer absorption spectra after 100 keV 57 Fe implantation, and finally spectra

under external stress. Furthermore, in order to discuss isolated Fe in Si materials, online emission spectra after GeV 57 Mn implantation in Si wafers will also be shown.

Based on this complete series of experimental results a consistent interpretation will

be presented for the components appearing under different experimental conditions.



8.2.2 Isomer Shift to Distinguish Fe Impurities in Si

Mössbauer spectroscopy resolves the hyperfine interaction between electrons and

a nuclear probe, i.e. 57 Fe in the present case, observing directly the shifts and/or

the splitting of the nuclear levels through recoil-free ”-ray resonant absorption or

emission, which is called Mössbauer effect and is described in Sect. 8.1.2 of this

chapter. The typical experimental set-up is shown in Fig. 8.3. Since the magnitudes

of the shifts and the splittings are typically of the order of 10 7 –10 9 eV, we can

use a Doppler energy shift of the ”-rays to scan the whole spectrum, moving the

absorber with respect to the source with moderate velocities of the order of mm/s.

Therefore, the spectrum is usually presented as a function of the Doppler velocity



394



G. Langouche and Y. Yoshida



in mm/s for the case of 57 Fe spectrum. The Si sample containing 57 Fe impurities

is either a source in accelerator facilities, or, more conventionally, an absorber in

laboratory set-ups.

We can distinguish three major kinds of hyperfine interactions. The quadrupole

interaction and the magnetic interaction give rise to a hyperfine splitting, as

described in Sect. 8.1.2 of this chapter, which can be studied by other hyperfine

interactions techniques as well. Mössbauer spectroscopy is unique in that it is

sensitive also to the electric monopole interaction, which is the Coulomb interaction

between the protons of the nucleus and the electrons (dominantly s-electrons)

penetrating into the nucleus. This interaction causes a shift of the Mössbauer

spectrum, which is called the “isomer shift • [mm/s]” (see also Sect. 8.1.2 in this

chapter). Isomer shift values depends on the charge state, high or low spin state,

lattice site and bonding properties such as covalency and electronegativity. The

isomer shift is influenced by defects in the surrounding Si matrix also.



8.2.3 Fe Components Observed in Mössbauer Spectra

Table 8.1 summarizes the isomer shifts for different Fe components observed at

room temperature in Si materials. The same colouring for each component will be

used throughout the rest of the chapter, when Mössbauer spectra are presented with

the different Fe components. The isomer shift, which is proportional to the electron

density at the nucleus, corresponds to the energy shift of the nuclear level for a

Mössbauer component on a Doppler velocity scale of mm/s against the spectral

centre of ’-Fe at room temperature. The data have been obtained in different

Mössbauer experiments on 57 Fe impurities, which were differently introduced into

Si samples, and subsequently measured at different temperatures: (1) deposition

at room temperature, and diffusion and measurements at high temperatures up

to 1273 K [37–42], (2) highly energetic implantation of the mother isotope of

57

Mn/57 Fe into Si [36, 43–47] and (3) deposition and diffusion, and measurements at

room temperature [48–52]. All the results can be analysed by the new model based



Table 8.1 Summary of Mössbauer components observed so far in Si



8 Nuclear Methods to Study Defects and Impurities in Si Materials



395



on the interstitial Fe atoms with neutral, C1, and C2 states (defect associated),

in addition to substitutional Fe with neutral, possibly C1 and 1 states. We also

identify Fe-B pairs in a highly B-doped sample, which is very close to the isomer

shift value for interstitial Fe with neutral charge state [51]. In Table 8.2, we summarize dominant Mössbauer experiments using completely different methodologies

in different groups. One should notice that the different experiments use different

nuclear decays, yielding different excess carrier concentration at the 57m Fe nucleus,

and consequently producing different charge states of 57m Fe probes in Si materials.

Moreover, the assignments of the Fe components in Table 8.1 were originally

based on a theoretical calculation for neutral interstitial Feint and substitutional Fesub

[61]. There are, however, still rather large discrepancies between theoretical values

and our experimental values, which are presented in Table 8.1. Nowadays, one could

use much better theoretical tools [62], but the isomer shifts deduced for the charged

interstitial and substitutional Fe atoms do not agree with the experimental values.

Our assignments for the charged Fe states are, therefore, proposed by taking into

account the experimental conditions at the 57 Fe nuclear probes in Si materials [46,

47, 51, 52] referring to the theoretical predictions for the Fe deep levels in Si [63,

64]. It should be mentioned that none of the spectra contained any Fe-silicides

or Fe-oxides whose isomer shifts are well known. A peculiarity of Mössbauer

spectroscopy is that it provides information on all Fe components present in the Si

matrix, even if they are not electrically active, while all other evaluation techniques

are only sensitive to electrically active components.



8.2.4 Unique Features of Mössbauer Spectroscopy

To control defects, light elements and metallic impurities in Si crystals, one intends

to use nowadays an evaluation technique, which provides atomistic information.

However, there is still no well-established technique accessible for single vacancy

and interstitial in Si crystals, while dislocations and extended defects which might

be decorated with metallic impurities such as Fe atoms can be directly observed, for

instance, by High-Resolution Electron Microscope (HREM). Chapter 7 “Electron

Beam Induced Current (EBIC) and Cathode Luminescence (CL)” provides such

images of electrically active dislocations and extended defects. In addition, Chapter

3 describes “Deep Level Transient Spectroscopy (DLTS) and Laplace DLTS”,

yielding the deep levels of impurities in Si devices. On the other hand, light

elements such as oxygen and nitrogen can be observed via Fourier Transform

Infra-Red (FTIR) spectroscopy, combining with simulation techniques for the

atomic vibrations of such light elements (Chap. 4). In the following sections, we

demonstrate the results of Mössbauer spectroscopy, which appears to be ideal to

study the basic nature of Fe impurities in Si materials. The experiments are always

challenging because of the detection limit, and therefore, the spectroscopy can be

applied mainly for Si materials which are intentionally contaminated by the stable



6. 60 keV

57

Mn implanted at

RT-800 K

7

500 MeV

57

Mn implanted

at RT-1000 K



5.



4.



3.



2.



1.



Diffused or implanted

element and

temperature

100 keV 57 Fe

implanted at

RT&1073 K

57

Fe diffused and

measured at

RT-1273 K

10 MeV 57m Fe

implanted at

300–850 K

57

Co diffused at

873–1473 K

for 5–30 min

30 keV

57

Co implanted

at 50–300 K



Total number

of 57 Fe to obtain 57 Fe depth profile

one spectrum

and concentration

2 1015 /cm2

R D 87 nm

32 nm

2 1020 /cm3

1 1016 /cm2

D D 1–500 m

0.1–200 m

1017–21 /cm3

5 1010 /cm2

R D 10 m

5 m

1 1016 /cm3

5 1013 /cm2

D D 30 nm

10 nm

4 1015 /cm3

2.5 1012 - 1

R D 30 nm

1013 /cm2

12 nm

5 1019 - 2

1020 /cm3

11

2

5 10 /cm

R D 57 nm

22 nm

2 1017 /cm3

2 1011 /cm2 R D 100 m

50 m

2 1013 /cm3

57

weeks

Co source /3.7GBq

100 ns 104–6 ”/cm2 s



days

EC-decay of 57 Co to

100 ns 57m Fe



t 2 min

“-decay of

£meas 100 ns 57 MnC to 57m Fe

108 57 MnC /cm2 s

t 2 min

“-decay of 57 Mn25C to

£meas 100 ns 57m Fe 107

57

Mn25C /cm2 s



£meas



t



t 2 ns

Coulomb excitation

£meas 100 ns of 57 Fe

104 57m Fe/cm2 s

t 30 min

EC-decay of 57 Co to

£meas 100 ns 57m Fe



£meas



t



Time after

diffusion or

implantation

Radiation fluence

57

t days

Co source/3.7 GBq

£meas 100 ns 5 106 ”/cm2 s



Table 8.2 Summary of experimental conditions and observed spectral components



0



0 (wafer)

1/0

(pn-junc.)



C1 (wafer)

0/C1/C2

(pn-junc.)



0



C1



1/0/C1?



C1



0



0



0/C1*



C1



0/C1/C2



[43–47]



[14]



[35]

*re-analysis

by [60]

[12]



[13]



[37–42]



Charge states Charge states

of interstitial of substitutional

57m

57m

Fe

Fe

Reference

0

0

[53]



396

G. Langouche and Y. Yoshida



8 Nuclear Methods to Study Defects and Impurities in Si Materials



397



isotope 57 Fe and as well as by mother isotopes of 57 Fe, i.e., either 57 Co or 57 Mn

nuclear probes.

In the following we sum up the unique features of 57 Fe Mössbauer spectroscopy

when it is applied for the study on Fe impurities in Si crystals.

1. The most important is that the 14.4 keV ”-rays are “recoil-free” absorbed or

emitted without disturbing the electronic states nor the atomic vibrations of both

Fe and Si atoms, i.e., Mössbauer effect. This is in contrast to other evaluation

techniques such as PL, CL and EBIC where strong laser or electron beam are

used to create the electro-active Fe states, and subsequently to measure emitted

photons and induced current, respectively. The ”-ray intensity of 105–8 photons/s

is typically used in Mössbauer experiments, while PL, CL and EBIC techniques

use many orders of magnitude higher intensities for excitation, i.e., 1018 –1020

photons and electrons/s, respectively.

2. Mössbauer spectroscopy provides information corresponding to all dominant Fe

components existing in Si matrix, even if they are not electrically active, while all

other evaluation techniques are only sensitive to electrically active components.

Furthermore, different lattice sites and charge states of solid solution of Fe in Si

can be distinguished in Mössbauer spectroscopy and have Mössbauer parameters

which are different from Fe silicides and oxides.

3. The concentration of 57 Fe is proportional to the resonance area. The detection

limit is different for source and absorber experiments, and close to 1011 and 1015

57

Fe atoms/cm3 , respectively.

4. In Mössbauer spectroscopy, there are different measuring time-scales involved

in the different experimental methods, depending on how 57 Fe probes were

introduced into the Si matrix and subsequently how the measurements are

triggered by the nuclear decay from the excited state of the 14.4 keV 57m Fe

nuclear level. Accordingly the atomistic information of Fe impurities is obtained

through the hyperfine interactions during the nuclear life time, i.e., 98 ns of this

level.

5. Not only the 57 Fe concentration, but also the carrier and their concentrations at

around the 57 Fe probes are rather different in the following experiments which

have been performed since more than 50 years: (1) 57 Fe implantation [6, 53],

(2) 57 Fe diffusion at SIST [6, 37–42, 48–53], (3) 57m Fe Coulomb-excitation and

recoil-implantation at HMI [13], (4) 57 Co diffusion [35], (5) 57 Co implantation

at Leuven [7, 12], (6) 57 Mn on-line separation and low-energy implantation at

CERN [14] and (7) 57 Mn on-line separation and high-energy implantation at

RIKEN [36, 43–47].

6. After 57 Co (T1/2 D 270d) [7, 12] and 57 Mn (T1/2 D 1.45 m) [14, 44–46] implantation into Si wafers the spectra consisted mainly of Fesub 0 and Feint C , as presented

in Table 8.2.



398



G. Langouche and Y. Yoshida



8.2.5 Mössbauer Spectra of 57 Fe Deposited on n-type Si Wafers

We first start looking at the Mössbauer absorber experiment on n-type Si wafers

with a thickness of 540 m containing 6 1018 Sb cm 3 and also 1 1012 P cm 3 ,

respectively [50]. The wafers were cut into square plates of 20 20 mm2 . After

removing a surface oxide layer by a 20 % HF solution, 57 Fe stable isotopes were

deposited onto the Si wafers with a thickness of 3.3 nm, and finally, the wafers

were annealed at 1273 K for 1 week. We used a set-up enabling us to measure

an in-situ Mössbauer spectrum under light illumination at room temperature. The

Si sample was fixed in an acryl-holder with an angle of 45ı against the 14.4 keV

”-ray direction. Accordingly, the sample was kept to be electrically isolated from

the ground during UV illumination. The UV light source was a xenon lamp, and a

mirror module was used to select an UV energy range between 3.5 and 5.5 eV. We

measured each Mössbauer spectrum for 1 week with a 1.85GBq 57 Co-in-Rh source,

yielding a ”-intensity of the order of 108 s 1 on the sample, which was ten orders of

magnitude lower than the light photon intensity, i.e., 1018 s 1 . This means that the

influence on the carrier production in the sample from the 14.4 keV and other higher

energetic ”-rays emitted from the Mössbauer source are considered to be negligible.

Mössbauer spectra of 57 Fe in n-type Si (6 1018 Sb/cm3 ) are shown in Fig. 8.18a.

The uppermost spectrum was measured in the dark before UV illumination, the

middle spectrum under UV illumination, and the lowest spectrum in the dark after

UV illumination. The spectrum before illumination consists of two singlets. The

spectrum under UV irradiation, however, changes from the spectrum in the dark: a

broad singlet, C, appears, while the areal fractions of A and B decrease. Finally,

the spectrum after illumination returns back to the initial spectrum in the dark,

consisting only of A and B again. In Fig. 8.18b, furthermore, Mössbauer spectra

of 57 Fe in n-type Si (1012 P/cm3 ) are presented: in the dark (before), under UV

illumination, and in the dark (after), respectively. Both the spectra before and under

UV illumination consist of two singlets A, B. In the case of the spectrum after

illumination, on the other hand, the spectrum consists of the components, A, B, and

C and D in addition. In the case of the Sb-highly doped Si, the new component C

appears only under illumination, but in the case of low P-doped Si the spectrum does

not change, but the linewidth of singlet B gets slightly narrower. After illumination,

however, the spectrum changes considerably: two singlets, C and D, are emerging

at the expense of the area fractions of the singlets, A and B.

Before we interpret the components A, B, C and D in Fig. 8.18a, b, we should

mention that we did not include potential Fe-silicides in the analysis of our spectra.

This can be concluded by investigating a Mössbauer spectrum of poly-crystalline

“-FeSi2 which was measured at room temperature [59]: The “-FeSi2 spectrum

in Fig. 8.19 appears to be a broad doublet with resonances at around 0.2 and

C0.3 mm/s. The former exists close to the component A in Fig. 8.18, while the latter

between A and C. A fitting analysis of our spectra reveals that there is certainly no

strong component of “-FeSi2 present in our spectra, as one would expect from the

reported Fe solubility [29].



8 Nuclear Methods to Study Defects and Impurities in Si Materials



Fig. 8.18 Mössbauer spectra of 57 Fe in (a) n-type Si (6

(1 1012 P/cm3 ) (From Ref. [50])



Fig. 8.19 Mössbauer spectra of “FeSi2 at RT (From Ref. [58])



399



1018 Sb/cm3 ), and (b) n-type Si



400



G. Langouche and Y. Yoshida



We have measured the spectra of 57 Fe doped n-type Si samples before, under

and after UV illumination with a photon intensity of 1018 s 1 . The singlets A and

B shown in Fig. 8.18 are the dominant components, which can be assigned to

substitutional Fesub 0 and interstitial Feint 0 , respectively, on the basis of theoretical

calculations of the isomer shifts [61, 62] in addition to our previous experimental

work described in this Chapter. In n-type Si “interstitial Fe” is considered to form

a “donor” level at 0.40 eV from the valence band, and therefore, the charge state

of interstitial Fe is expected to be neutral in both samples. On the other hand, the

components C and D are observed as small satellites which must be produced only

by the carrier injections through UV illumination and subsequently by the carrier

trapping processes at 57 Fe probes. Accordingly, the singlet C appears to correspond

to interstitial Feint C , which must yield a lower electron density than that of Feint 0 ,

i.e., more to the right on the Doppler velocity scale, while the singlet D is tentatively

assigned to substitutional Fesub , related to an accepter state of substitutional Fe,

the existence of which was predicted in a first principle calculation [63]. The isomer

shift of the most left hand side resonance corresponds to the highest electron density

in comparison with those of the other singlets, which appears to be consistent with

the accepter state of Fesub . The same components as A, B, C, and D were observed

immediately after highly energetic implantation of 57 Mn/57 Fe, where the formations

of Fe clusters and silicides can be absolutely excluded because of the experimental

conditions, as will be explained in the Sects. 8.2.11 and 8.2.12.

The heat treatment at 1273 K would provide a Gaussian distribution of 57 Fe

substitutional atoms from the surface down to about 1 m, and in addition, a rather

homogenous distribution of 57 Fe interstitials at least down to 10 m in the Si wafer,

as was measured using SIMS [39]. This was confirmed by etching the surface

layer and by the subsequent measurement of a spectrum, in which the substitutional

component nearly disappeared, while the interstitial one remained. Accordingly, the

substitutional Fe atoms are supposed to be illuminated directly from UV light. The

interstitial Fe atoms are, however, expected to be distributed homogeneously, which

means that only a small part of the interstitial Fe atoms are directly illuminated. The

UV illumination may cause a transformation either from the substitutional Fe atom

to the interstitial Fe atom or from the neutral substitutional Fe and the interstitial Fe

to their charged states. Under illumination: in the low P-doped sample, the excess

carriers react with the charged states of the substitutional and the interstitial Fe

atoms effectively, so that only the neutral states can be seen, while in the highly Sbdoped sample, the transformation from the charged state to the neutral state of the

interstitial Fe atoms seems to be partly suppressed due to the majority carriers and

the Sb dopants. This results in the singlet of Feint C . After illumination: in the case

of low P-doped sample with a very low electron concentration, the excess carriers

are supposed to diffuse and subsequently to react with neutral substitutional and

interstitial Fe atoms, leading to the Feint C and Fesub components which emerge

in the dark after UV illumination, on one hand. In the case of highly Sb-doped

sample, on the other hand, the excess carriers are compensated and therefore, the

same spectrum is obtained as before illumination.



8 Nuclear Methods to Study Defects and Impurities in Si Materials



401



8.2.6 Low Energy 57 Fe Implantations into Si

As the second example, we discuss the Mössbauer absorber experiment after

implantation of 57 Fe stable isotopes into Si wafers [53]. The ion implantation

technique enables us to introduce a well-defined concentration profile of atoms

such as boron and phosphorus dopants into semiconductors such as Si crystals. This

technique is, therefore, widely used for electronic device processing, although the

implantation produces also cascade defects consisting of vacancies and interstitials

in the matrix due to energetic collisions of implanted atoms with host atoms during

the slowing-down process. We apply this technique to study Fe impurities in Si

matrix. We introduce 57 Fe probes into the Si matrix, and subsequently measure the

transmission Mössbauer spectra, as shown in Fig. 8.20 (left).

57

FeC ions were implanted into p-type CZ-Si (1019 B/cm3 ) and FZ-Si (1015 B/cm3 )

wafers at room temperature and also at 1073 K with an energy of 100 keV [53]. The

total dose was 2 1015 57 Fe /cm2 .Subsequently, Mössbauer absorption spectra were

measured for 1 week at room temperature using a 1.85-GBq-57Co-in-Rh source. The

spectra are shown in Fig. 8.20 (left): the two upper spectra correspond to 57 Fe in CZSi (1019 B/cm3 ), and the two lower to 57 Fe in FZ-Si (1015 B/cm3 ), respectively. The



2㽢1015 57Fe/cm2

RT



(2)



RT



1073K











-0.1



0.0



0.1



0.2



0.3



0.4



Collision Events / Å-1



1073K



Fe distribution /arbitrary unit



(1)











Depth / μm



Fig. 8.20 (Left) Mössbauer spectra of 57 Fe in (1) CZ- and (2) FZ-Si measured at room temperature

(From Ref. [53]). The implantations of 57 FeC ions with an energy of 100 keV were performed up

to a total dose of 2 101557 Fe/cm2 at room temperature and also at 1073 K, respectively (right).

The depth profile of 57 Fe and defects distributions are simulated for room temperature without

diffusion effect by SRIM [65]



402



G. Langouche and Y. Yoshida



depth profile of 57 Fe impurities and the numbers of collision events are shown in

Fig. 8.20 (right) as a function of depth, which were estimated by a SRIM simulation

[65] for the case of 100 keV-57 FeC -implantation at room temperature. The range

and the straggling are 87 and 32 nm, respectively. As discussed in Sect. 8.1.2. earlier

Mössbauer experiments [10] showed that 70 kV Fe implantation in Si at high fluence

(1016 cm 2 ) resulted in the formation of an amorphous region at the end of slowingdown process due to overlapping cascade damages. The present Mössbauer spectra

were recorded with a somewhat higher implantation energy and lower fluence and

the spectrum shape is similar for the RT implantation and for the implantations at

1073 K where amorphization should not be present. We therefore feel justified in

analysing our spectra with two sharp absorption lines, suggesting that 57 Fe probes

are stopping at lattice positions not in an amorphous region, but at well-defined

lattice sites, as will be discussed in the following.

All the spectra mainly consist of substitutional Fesub 0 and interstitial Feint 0

components at the left (green) and the right hand sides (yellow), respectively. The

area fraction of substitutional Fesub 0 is slightly higher after the implantation at

1073 K than after that at room temperature, suggesting that interstitial Fe atoms

encounter vacancies produced by the implantation, forming substitutional Fe during

implantation at 1073 K, because of the higher diffusivities of Fe interstitial and

vacancy. The spectra indicate that the charge state of interstitial Feint 0 corresponds to

the neutral state, although one would expect the charged state Feint C from the dopant

concentrations. 100 keV-FeC ions will lose their energy mainly through elastic

collisions during the slowing down processes. Accordingly, the following reactions

are considered to occur possibly: Feint C C V ! Fesub 0 ; Feint C C e ! Feint 0 . The

vacancies V are considered to be distributed over the Fe profile, as is shown in

Fig. 8.20 (right).

We have shown the absorber Mössbauer spectra, which were obtained either

by the annealing at 1273 K for 1 week in Sect. 8.2.5, or by implantation at room

temperature and also at 1073 K in the present section. The spectra of n-type

Si wafers consist of interstitial Feint 0 and substitutional Fesub 0 as the dominant

components, and additionally Feint C and Fesub after trapping excess carriers

injected by UV-illumination during the measurements. In p-type Si wafers, on the

other hand, only the components of interstitial Feint 0 and substitutional Fesub 0 are

observed, which is contradictory to what one would expect by taking into account

the position of both the Fermi levels and the deep level for Fe interstitial, i.e.,

0.38 eV above the Si valence band [29]. One should notice that all spectra clearly

contain the substitutional component Fesub 0 and this is completely different from the

model which has been accepted for Fe impurities in Si materials in the Si community

[29, 30]. In order to clarify the nature of interstitial Feint 0 , Feint C and Feint 2C , and

substitutional Fesub 0 and Fesub , we investigated further 57 Fe deposited Si samples

at high temperature, where 57 Fe probes are thought to be distributed over the whole

sample depth.



8 Nuclear Methods to Study Defects and Impurities in Si Materials



403



8.2.7 High Temperature Absorber Experiments

The spectra of n-type FZ-Si with 1.9 1019 As/cm3 [42, 53] are shown in Fig. 8.21

(left). The spectra were measured (1) at room temperature immediately after the

deposition of 57 Fe with a thickness of 1.4 nm, (2) at 1273 K for about 10 days, and

finally (3) at room temperature again after the measurement at 1273 K. The spectrum

(1) consists of substitutional Fesub 0 and interstitial Feint 0 and Feint C , respectively.

The spectrum (2) at 1273 K can be fitted with three singlets, Fesub 0 , Feint 0 and

Feint 2C , the third component which is temporally assigned will be discussed later.

The isomer shift of the Feint 2C component corresponds to C1.2 mm/s, which could

be due to an interstitial Fe in the vicinity of a defect. After the high temperature

measurements, the spectrum (3) at room temperature can be fitted with Fesub 0 and

Feint 0 .

Furthermore, the spectra shown in Fig. 8.21 (right) of 57 Fe in p-type FZ-Si with

3.6 1015 B/cm3 were measured in the same way as in Fig. 8.21 (left). The spectrum

(1) consists of the same components of Fesub 0 , Feint 0 and Feint C , as can be seen in

Fig. 8.21 (left), but additionally another component appears on the left hand site of

Fesub 0 , which appears to be due to Fesub related to an accepter level predicted by

theoretical work [63]. The spectrum (2) at 1273 K can be fitted with three singlets,



19



-3



15



n-type Si (As:1.9㼤10 cm )

100.1



100.1



100.0



100.0



99.9



99.9



99.8



100.0

99.9



1273K



100.1

100.0



Normalized counts



Normalized counts



100.1



99.7



99.8



RT



99.7



99.8



-3



p-type Si (B:3.6㼤10 cm )



RT



100.1

100.0

99.9

99.8



1273K



100.1

100.0



99.9



99.9



99.8



RT



99.7

-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0



Velocity / mms-1



99.8



RT



-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0



Velocity / mms-1



Fig. 8.21 (a) Mössbauer spectra of 57 Fe in n-type Si (As D 1.9 1019 cm 3 ), and (b) Mössbauer

spectra of 57 Fe in p-type Si (B D 3.6 1015 cm 3 ), measured (1) at room temperature after 57 Fe

deposition, (2) at 1273 K, and (3) at room temperature after high temperature measurement



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

2 Recent Mössbauer Studies on Fe in Si

Tải bản đầy đủ ngay(0 tr)

×