Temperature dependencies of hydrogen-induced blistering of thin film multilayers

TL;DR: In this article, the influence of sample temperature on the development of hydrogen-induced blisters in Mo/Si thin-film multilayers was investigated and it was shown that the areal number density of blisters decreases with increasing exposure temperature, whereas individual blister size increases with exposure temperatures up to ∼200°C but decreases thereafter.

Abstract: We report on the influence of sample temperature on the development of hydrogen-induced blisters in Mo/Si thin-film multilayers. In general, the areal number density of blisters decreases with increasing exposure temperature, whereas individual blister size increases with exposure temperatures up to ∼200 °C but decreases thereafter. Comparison as a function of sample temperature is made between exposures to a flux containing both hydrogen ions and neutrals and one containing only neutrals. In the case of the neutral-only flux, blistering is observed for exposure temperatures ≥90 °C. The inclusion of ions promotes blister formation at <90 °C, while retarding their growth at higher temperatures. In general, ion-induced effects become less evident with increasing exposure temperature. At 200 °C, the main effect discernable is reduced blister size as compared with the equivalent neutral-only exposure. The temperature during exposure is a much stronger determinant of the blistering outcome than either pre- or post-annealing of the sample. The trends observed for neutral-only exposures are attributed to competing effects of defect density thermal equilibration and H-atom induced modification of the Si layers. Energetic ions modify the blistering via (temperature dependent) enhancement of H-mobility and re-crystallization of amorphous Si.

Summary

1. Introduction

• Hydrogen-induced blistering of materials is a complex phenomenon1-3 that can have a strong dependence on the exposure conditions, with sample temperature being one of the critical parameters.
• In the case of exposure to a neutral-only hydrogen flux, blistering was the result of layer delamination occurring exclusively at the outermost Mo-on-Si interface.
• This induces an overall compaction of the ML structure as the silicide layers produced are more dense than those of their original elemental constituents.
• In addition, the relative importance of the sample temperature during hydrogen irradiation as compared with either pre- or post-annealing of the sample at higher temperatures than during the actual exposure is studied.

2. Experimental.

• The Mo/Si ML samples investigated in this work are similar to those studied previously23-26.
• The hydrogen source was a capillary-type thermal cracker (Oxford Applied Research TC50).
• They were stabilized at the desired target temperature prior to commencing exposure.
• Similar exposures were performed with the addition of a biased deflector plate near to the sample and parallel to the incident particle direction.
• For samples exhibiting non-uniform damage, the analysis was performed in the region exhibiting the most damaged.

3.2. Effects of pre- and post-exposure annealing

• The effects on blistering of pre- and post-annealing of samples to elevated temperature was investigated.
• Figure 6 shows the comparison of individual blister volume versus area derived from these samples.
• In the case of post-annealing, the change is the result of preferential growth of the larger component (compare Fig. 5 (a)&(d)).
• The effect is most pronounced for the “thermal neutral” blisters; the number of “ion-induced” blisters formed is relatively unaffected by pre-annealing.

4. Discussion

• The authors begin by giving a synopsis of their previous observations and interpretations of blister formation in Mo/Si MLs, before addressing how those processes are influenced by temperature-dependent effects.
• The temperature dependencies the authors report can be accounted for on the basis of defect densities established as a function of Texposure and/or Tanneal and exposure-induced modifications to the layer structure.
• As noted above, defects introduced by energetic ions modify the uptake and mobility of hydrogen by the ML.
• While blisters clearly identifiable as “ion-induced” are not evident at Texposure>150 °C ion-induced pre-blister voids may still act to sequester hydrogen, thereby retarding H2-filled blister growth.
• Thus, defect equilibration contributes to enhanced hydrogen mobility in a-Si but the effect is counteracted with increasing Tanneal by interfacial layer growth and densification.

5. Conclusion.

• Evaluation of the effects of sample temperature on hydrogen-induced blister formation in Mo/Si multilayer (ML) samples illustrates that the temperature during exposure DOI: 10.1063/1.4875484 is the dominating factor in the development of blisters.
• Effects related to thermally-induced changes to the ML structure and post-exposure redistribution of incorporated hydrogen are of lesser importance.
• The general trends observed for increasing exposure temperature were of an increase in blister size and a decrease in areal number density up to 200 °C, followed by a decrease in blister size there-after.
• In the case of the neutral-only exposures, these trends are attributed to the competing effects on a-Si layers of defect density thermal equilibration (enhancing Huptake and mobility) and H-atom induced re-crystallization (reducing H diffusion and enhancing H2 out-diffusion).
• The inclusion of ions in the irradiating flux enhances blister formation at lower temperatures by introducing defects and strained bond that promote Huptake and diffusion.

Figures

Figure 3: Plot of the radii of the large and small components of “double blister” features as a function of exposure temperature based on the averaged result of measurement from ten individual features. Inset: Ratio of the two radii.

Figure 7: Plots of (a) the blister count, (b) the total volume occupied, and (c) the average blister volume after exposure to the ion+neutral flux as a function of pre-/post-annealing temperature. Negative (positive) annealing temperatures signify pre- (post-)annealing. The data points are derived from analysis of 50x50 µm2 AFM scans.

Figure 6: Volume versus area data of individual blisters derived from 50×50 µm2 AFM scans of samples after 5 hours of exposure to the ion+neutral flux at Texposure=100 °C. (a) No pre/post-annealing; (b) 200 °C pre-anneal; (c) 300 °C pre-anneal; and (d) 250 °C post-anneal. “Ion-induced” and “thermal neutral” blisters are represented by filled and open circles, respectively. The reference curve from figure 2 is again reproduced.

Figure 1: Optical microscopy (x100) images of samples after exposure at fixed temperatures for 5 hours to the flux from the capillary cracker at a 1 sccm H2 flow. The left-hand-side segment of each image represents an exposure to unmodified flux (ion+neutral); the righthand-side segments represent exposure with deflection of charged species (neutral-only). The

Figure 5: Shaded image representation of 20×20 µm2 AFM scans from samples after 5 hours of exposure to the ion+neutral flux at 100 °C with: (a) no pre-/post-annealing; (b) 200 °C pre-annealing; (c) 300 °C pre-annealing; and (d) 250 °C post-annealing.

Figure 4: Plots of (a) blister count, (b) total occupied volume, and (c) average blister volume as a function of sample temperature during exposures to the ion+neutral and neutral-only fluxes. Data points are derived from the same AFM scans used to produce Figure 2. All values have been normalized to represent a 50x50 µm2 equivalent area.

Figure 2: Plots of individual blister volume versus corresponding area covered derived from AFM scans of samples after exposure at fixed temperature. The panel labels indicate Texposure and the scan area used for the analysis. Data from samples exposed to the ion+neutral flux are represented by circles (filled and open indicate “ion-induced” and “thermal neutral” blisters, respectively). Data from samples exposed to the neutral-only flux are represented by crosses (“thermal neutral” blisters only). The solid line is a reference curve to aid comparison

Figure 8: (a) Schematic representation of ‘thermal neutral’ process of hydrogen uptake and blister formation. (b) Schematic representation of ‘ion-induced’ H uptake and hydrogen transport to deeper layers.

Full text

A.S. Kuznetsov, M. A. Gleeson and F. Bijkerk (2014). "Temperature dependencies of hydrogen-induced blistering of thin
film multilayers." Journal of Applied Physics 115(17): 173510.
DOI: 10.1063/1.4875484
Temperature dependencies of hydrogen-induced blistering of thin film multilayers
A.S. Kuznetsov
1
, M.A. Gleeson
1,1
and F.Bijkerk
1,2
1
DIFFER – Dutch Institute for Fundamental Energy Research, Postbus 1207, 3430 BE Nieuwegein, The
Netherlands.
2
MESA
+
Institute for Nanotechnology, University of Twente, Postbus 217, 7500 AE Enschede, The
Netherlands.
Abstract
We report on the influence of sample temperature on the development of hydrogen-induced
blisters in Mo/Si thin-film multilayers. In general, the areal number density of blisters
decreases with increasing exposure temperature, whereas individual blister size increases with
exposure temperatures up to ~200 °C but decreases thereafter. Comparison as a function of
sample temperature is made between exposures to a flux containing both hydrogen ions and
neutrals and one containing only neutrals. In the case of the neutral-only flux, blistering is
observed for exposure temperatures ≥90 °C. The inclusion of ions promotes blister
formation at <90 °C, while retarding their growth at higher temperatures. In general, ion-
induced effects become less evident with increasing exposure temperature. At 200 °C the
main effect discernable is reduced blister size as compared with the equivalent neutral-only
exposure. The temperature during exposure is a much stronger determinant of the blistering
outcome than either pre- or post-annealing of the sample. The trends observed for neutral-
only exposures are attributed to competing effects of defect density thermal equilibration
and H-atom induced modification of the Si layers. Energetic ions modify the blistering via
(temperature dependent) enhancement of H-mobility and re-crystallization of amorphous Si.
1
Corresponding author: M.A.Gleeson@differ.nl
1

A.S. Kuznetsov, M. A. Gleeson and F. Bijkerk (2014). "Temperature dependencies of hydrogen-induced blistering of thin
film multilayers." Journal of Applied Physics 115(17): 173510.
DOI: 10.1063/1.4875484
1. Introduction
Hydrogen-induced blistering of materials is a complex phenomenon
1-3
that can have
a strong dependence on the exposure conditions, with sample temperature being one of the
critical parameters. The temperature dependence of blister formation is an open issue in
fusion research where hydrogen retention
4-6
and hydrogen-induced damage to plasma-facing
materials
7-9
are being investigated. In relation to thin films, blister and crater formation
occurs during annealing of hydrogenated Si-Ge multilayer (ML) structures
10
. Heating can
induce structural modifications such as relaxation and crystalline phase transformation,
promote diffusion and nucleation, and in the case of mixed-material systems induce
intermixing and compound formation. All such changes influence the behaviour of
hydrogen within the system.
The selection of appropriate annealing/exposure temperature is important for the
process optimization in thin layer transfer techniques both for standard ion implantation and
strained-layer driven layer detachment
11-21
. Localization of either hydrogen or stabilized
vacancy complexes (depending on the specifics of the strain states) at interfaces in layered
materials is a commonly observed phenomenon. It is proposed as a means of producing
better quality thin silicon layers for silicon-on-insulator related applications
18-20
. Layer quality
is improved because the use of buried interfaces to control detachment allows for significant
reduction, or even elimination, of the energetic ion fluence required to induce delamination.
In this work we report on the influence of sample temperature on the development
of blisters in Mo/Si ML samples. These samples are composed of alternating layers of
nanometer thick amorphous Si (a-Si) and polycrystalline Mo with mixed Mo-Si interfacial
regions
22
. They are susceptible to two distinct hydrogen-induced blistering processes
23-26
,
which are attributed to the formation of H
2
-filled blisters and to the formation and clustering
2

A.S. Kuznetsov, M. A. Gleeson and F. Bijkerk (2014). "Temperature dependencies of hydrogen-induced blistering of thin
film multilayers." Journal of Applied Physics 115(17): 173510.
DOI: 10.1063/1.4875484
of hydrogen-vacancy complexes producing void blister
24
. The former occurs under the
influence of thermal H-atom irradiation; the latter process is observed when energetic (100’s
of eV) ions are present in the irradiating flux. These processes are not entirely independent.
In particular, the inclusion of energetic ions in the irradiating flux modifies the development
and growth of H
2
-filled blisters
24
.
Hydrogen-induced blistering in such MLs is observed to be localized near the Mo-
on-Si interfaces
23
. This is attributed the strained states and defects that are introduced near
these interfaces by the transition between materials and the local atomic structures
24
. In the
case of exposure to a neutral-only hydrogen flux, blistering was the result of layer
delamination occurring exclusively at the outermost Mo-on-Si interface. When ions were
also present in the irradiating flux delamination was observed at the outermost two Mo-on-
Si interfaces. Apart from direct penetration to the second and deeper bilayers, energetic ions
were proposed to induce transport of hydrogen from gas-filled bubbles near the outermost
Mo-on-Si interface across the mixed-phase interfacial region into the underlying Si layer.
A preliminary study on the temperature dependence of blister formation in Mo/Si
MLs illustrated that the substrate temperature had a significant impact on the formation of
blisters
26
. However, the complexity of the system left many open questions since the MLs
themselves are subject to structural deterioration as a result of annealing
27-29
. This is primarily
the result of increased intermixing at the interfaces resulting in the growth of molybdenum
silicide layers at the expense of the individual Mo and Si layers. This induces an overall
compaction of the ML structure as the silicide layers produced are more dense than those of
their original elemental constituents.
This manuscript aims to provide a detailed evaluation of the influence of sample
temperature on the formation of blisters in these structures. Atomic force microscopy
3

A.S. Kuznetsov, M. A. Gleeson and F. Bijkerk (2014). "Temperature dependencies of hydrogen-induced blistering of thin
film multilayers." Journal of Applied Physics 115(17): 173510.
DOI: 10.1063/1.4875484
(AFM) measurements are used as the basis for evaluating the extent and type of blistering.
Comparison is made between the temperature dependencies in the case of exposure to an
exclusively neutral flux and to a flux containing energetic ions. In addition, the relative
importance of the sample temperature during hydrogen irradiation as compared with either
pre- or post-annealing of the sample at higher temperatures than during the actual exposure
is studied. Substrate temperature during the exposure is demonstrated to have the biggest
impact on the outcome.
2. Experimental.
The Mo/Si ML samples investigated in this work are similar to those studied
previously
23-26
. They were deposited on a super-polished Si wafer by magnetron sputtering
with additional ion polishing of the deposited Si layers. The thicknesses of individual layers
in the samples were ~3 nm for Mo and ~4 nm for Si. The samples were transferred through
air both prior to and post hydrogen exposure and no pre-treatment step was applied before
either exposure or analysis.
The hydrogen source was a capillary-type thermal cracker (Oxford Applied Research
TC50). It was operated at a power of 55 W with a 1 SCCM H
2
flow. In addition to the
neutral atomic and molecular hydrogen flux, the source also produced an ion current on the
order of ~75 nA. The majority of ions produced have energies >800 eV, due to the positive
bias (+1000 V) used to e-beam heat the capillary
24
.
Samples were exposed for 5 hours to the unmodified hydrogen flux from the
thermal capillary cracker at various set-point temperatures ranging from 50 °C to 275 °C,
under otherwise identical conditions. They were stabilized at the desired target temperature
prior to commencing exposure. Heating was radiative, via a back-mounted filament and the
4

A.S. Kuznetsov, M. A. Gleeson and F. Bijkerk (2014). "Temperature dependencies of hydrogen-induced blistering of thin
film multilayers." Journal of Applied Physics 115(17): 173510.
DOI: 10.1063/1.4875484
temperature was monitored and controlled on the basis of a K-type thermocouple that was
spot-welded to one of the sample mounting clamps. Similar exposures were performed with
the addition of a biased deflector plate near to the sample and parallel to the incident particle
direction. A bias of -1100 V was used to deflect charged species from the flux prior to
interaction with the surface. When the bias was applied the drain current measured on the
sample during exposure dropped to <1 nA. Exposures are characterized as “neutral-only” or
“ion+neutral” depending on whether the biased plate was present or absent, respectively.
Damage to exposed samples was typically evident to the naked eye as a
discoloration/dulling of the original mirror-like surface. Regions covered by clamps during
the exposure always appeared undamaged. Exposed samples were characterized by optical
microscopy and atomic force microscopy (AFM). For samples exhibiting non-uniform
damage, the analysis was performed in the region exhibiting the most damaged. In the
current work this corresponds to the on-axis position of the cracker capillary. In cases where
no damage was visible or where the damage appeared uniform, the analysis position was
selected with reference to the non-uniform samples.
3. Results.
3.1. Temperature dependencies of “ion+neutral” and “neutral-only” exposures.
As reported in
24
, samples exposed to the ion+neutral flux at temperatures (T
exposure
)
of 85-100 °C typically exhibited a distinctive concentric damage pattern on the surface. This
is primarily due to the presence of ions; when a neutral-only flux is used the damage across
the exposed surface appears uniform. After the current ion+neutral exposures, the samples
exposed at temperatures between 50 °C and 100 °C again exhibited the well-defined central
spot surrounded by regions of more diffuse discoloration. Those exposed at 150 °C and 200
5

Frequently asked questions

Q1. What contributions have the authors mentioned in the paper "Temperature dependencies of hydrogen-induced blistering of thin film multilayers" ?

The authors report on the influence of sample temperature on the development of hydrogen-induced blisters in Mo/Si thin-film multilayers. Energetic ions modify the blistering via ( temperature dependent ) enhancement of H-mobility and re-crystallization of amorphous Si. 1 Corresponding author: M. A. Gleeson @ differ.