2019-06-07 Educational seminar at EP-3 University of Wuerzburg
I will present particular experiments and related results with FTIR spectrometer, so one may consider these experiments complimentary for you research.
2019-06-07 Characterization and research of semiconductors with an FTIR spectrometer
1. Characterization and research
of semiconductors
with an FTIR spectrometer
Particular experimental designs and related research possibilities
Leonid Bovkun
Friday EP-3 seminar, 07/06/2019
3. FT-IR spectrometer
is based on Michelson interferometer
fixed mirror
source
signal
HeNe laser 632.8 nm
3
controlled mirror
beamsplitter
4. FT-IR spectrometer
is based on Michelson interferometer
fixed mirror
source
signal
HeNe laser 632.8 nm
4
Multiplex Advantage
FT-IR required
5. maintenance-free FT-IR analyzer
28 cm43 cm
37 cm
https://new.abb.com/products/measurement-products/analytical/ft-ir-and-ft-nir-analyzers/laboratory-spectrometers/mb3000
sample
5
source
sample
beamsplitter
detector
• Room temperature DLATGS detector
• Spectral range 485 to 8,500 cm-1
• Resolution better than 0.7 cm-1
6. high-end research instruments
https://www.bruker.com/products/infrared-near-infrared-and-raman-spectroscopy/ft-ir-research-spectrometers/vertex-series/vertex-
8080v/applications.html
Pharma
Determination of the absolute configuration of molecules (VCD)
Characterization of stability and volatile content of medical drug products by thermal analysis (TGA-FTIR)
Differentiation of polymorphs of active pharmaceutical ingredients in the far infrared region
Polymers and Chemistry
Identification of inorganic fillers in polymer composites in the far infrared region
Dynamic and rheo-optical studies of polymers
Determination of volatile compounds and characterization of decomposition processes by thermal analysis (TGA-FTIR)
Reaction monitoring and reaction control (MIR fiber probe)
Identification of inorganic minerals and pigments
Surface Analysis
Detection and characterization of thin and monolayers
Surface analysis combined with polarization modulation (PM-IRRAS)
Material Science
Characterization of optical and highly reflective materials (windows, mirrors)
Investigation of dark materials and depth profiling by Photo-Acoustic Spectroscopy (PAS)
Characterization of the emittance behavior of materials
Research & Development
Continuous and Step Scan technology for amplitude/phase modulation spectroscopy
Rapid, interleaved and Step Scan technology for experiments with high temporal resolution (Step Scan / Rapid Scan / Interleaved TRS)
Characterization of periodically ordered microscopic materials, known as metamaterials
High resolution spectroscopy for gas analysis with resolutions better than 0.06 cm-1
Instrumentation for vacuum FTIR beamline installations
Stopped-flow methods for enzyme catalysis experiments
External adaptation of ultrahigh vacuum measurement chambers
FTIR spectroelectrochemistry for the in-situ investigation of electrode surfaces and electrolytes
Semiconductors
Determination of oxygen and carbon contents in silicon wafers
Low temperature transmittance and photoluminescence (PL) measurements of shallow impurities for quality control
6
8. Motivation
What for do we need an FT-IR spectrometer?
• Photoluminescence (characterization)
• Transmission in magnetic field (science)
To explore narrowgap systems
8
9. PL principle and physics
https://en.wikipedia.org/wiki/Photoluminescence
Evaluation of optical quality and properties:
• bandgap and higher bands
• exciton binding energies
• impurity binding energies,
• deep-level defect presence,
• recombination mechanisms
• surface structure and excited states
9
o radiative
o non-radiative
10. PL principle and physics
A.-S. Gadallah and M. M. El-Nahass, Advances in Condensed Matter Physics 2013, 11, 234546 (2013).
2.38 eV
3.26 eV
3.11 eV (donor-acceptor)
1.70 eV
10
exciton
3.26 eV
example study of ZnO films with He-Cd laser:
• bandgap
• exciton binding energies
• deep-level defect presence
Very-very far from FIR
11. ICL 3-6 microns InAs/GaInAs
https://nanoplus.com/en/news/icl-3000-6000-nm/
R. Weih, L. Nähle, S. Höfling, J. Koeth, and M. Kamp, Appl. Phys. Lett. 105, 071111 (2014).
R. Weih, “Interband Cascade Lasers for Gas Sensing in the Mid Infrared Spectral Region” PhD Thesis 2018
11
Introduction
The main research objectives of ‘Technische Physik’
are related to the fabrication and characterization
of semiconductor nanostructures. The group
works on the development of nanostructure
patterning technologies for optoelectronic
applications as well as for basic physics studies of
low dimensional photonic and electronic systems.
A complete chain of processing equipment, ranging
from molecular beam epitaxy over several
lithographic techniques to etching and deposition
systems is used to fabricate devices and
nanostructures for spectroscopy and transport
experiments.
In addition to the cleanroom facilities, the group
operates a wide range of characterization tools
(optical spectroscopy, transport measurements,
laser characterization, high frequency
measurement...), which are used to investigate the
performance of devices and fundamental aspects of
low-dimensional structures.
12. Room-temperature setup
R. Weih, “Interband Cascade Lasers for Gas Sensing in the Mid Infrared Spectral Region” PhD Thesis 2018
sample = source
BS KBr / CaF2
N2-cooled
InSb, HgCdTe
photo-excitation
N2 purge
FTIR
12
λ = 532 nm
13. Optimal growth parameters
Robert Weih, “Interband Cascade Lasers for Gas Sensing in the Mid Infrared Spectral Region” PhD Thesis 2018 13
Variation of substrate temperature
450 °C – high intensity, small FWHM
Variation of As flow
defects
formationfrom
literature
As-Sb exchange reaction at interfaces
14. Challenges of the FIR range
• undesired absorption of medium
• intensive background radiation
• weak transitions’ intensity
14B. A. Bernevig, T. L. Hughes, and S. C. Zhang, Science 314, 1757 (2006)
M. Orlita et al., Nat Phys 10, 233 (2014).
CdHgTe
gapless
structures:
bulk with xc
QW with dc
pump it
next
slide
15. Modulated PL spectroscopy
J. Shao, W. Lu, X. Lü, F. Yue, Z. Li, S. Guo, and J. Chu, Review of Scientific Instruments 77, 063104 (2006).
J. Shao, F. Yue, X. Lü, W. Lu, W. Huang, Z. Li, S. Guo, and J. Chu, Appl Phys Lett 89, 182121 (2006).
15
sample
beamsplitter
laser
excitation
detector
vacuum
Signal = PL + thermal + He-Ne
• undesired absorption of medium
• intensive background radiation
• weak transitions’ intensity
16. Modulated PL spectroscopy
J. Shao, W. Lu, X. Lü, F. Yue, Z. Li, S. Guo, and J. Chu, Review of Scientific Instruments 77, 063104 (2006).
J. Shao, F. Yue, X. Lü, W. Lu, W. Huang, Z. Li, S. Guo, and J. Chu, Appl Phys Lett 89, 182121 (2006).
16
sample
beamsplitter
laser
excitation
detector
lock-in
electronic
controller
PC
reference from chopper
vacuum,
step-scan
Signal = PL
17. Experiment design for PL in HgTe QW
S. V. Morozov, V. V. Rumyantsev, et. al., Appl Phys Lett 104, 072102 (2014). 17
18. Bandgap determination
S. V. Morozov, V. V. Rumyantsev, et. al., Appl Phys Lett 104, 072102 (2014). 18
4
3
2
1
0.17 xc
xc
19. Temperature effects
S. V. Morozov, V. V. Rumyantsev, et. al., Appl Phys Lett 104, 072102 (2014). 19
1. Eg increase with T
2. shallow impurity/defect
cm-1
50 (1)
108 (2)
220 (3)
218 (4)
radiative limit 1.8
20. PL-setup with time resolution
S. V. Morozov, V. V. Rumyantsev, et. al., Appl Phys Lett 104, 072102 (2014). 20
laserτ = 400 ns
21. PL experiment conclusions
S. V. Morozov, V. V. Rumyantsev, et. al., Appl Phys Lett 104, 072102 (2014).
Fast measurements in controlled environment
• Allow quick selection of the best sample in array
• Sensitive to defects and impurities – allow adjustments of
growth
• Allow to determine bandgap
Optional time resolved spectroscopy
• allow to determine the carriers lifetime
• and study recombination mechanisms
21
26. B = 0
26
Landau levels in a single 2D parabolic band
𝐸 𝒑 =
𝑝 𝑥
2 + 𝑝 𝑦
2
2𝑚
𝐸 𝑛 = ℏ𝜔𝑐(𝑛 +
1
2
)
B ≠ 0
n = 3
n = 2
n = 1
n = 0
electric dipolar
selection rules
27.
CR & interband transitions
Band structure Landau levels
conduction band
valence band
Energy(meV)
α
nm-1 B (T)
27
28.
CR & interband transitions
Band structure Landau levels
conduction band
valence band
Energy(meV)
nm-1
α
β
α-
γ
B (T) 28
36. 36
Sensitivity down to 0.1 % of the relative change with the magnetic field
https://emfl.eu/find-experiment/infrared-spectroscopy/
DLATGS
37. Magneto-optical spectroscopy
Contactless but sensitive in the broad range
Powerful tool to study band structure
Complimentary features:
• optical gating
• polarization-sensitive
• temperature effects
Experiment
Band
structure
Theory
38. 38
Three-layer heterostructures InAs/GaSb/InAs
Liu, C., et al., Quantum Spin Hall Effect in Inverted Type-II Semiconductors. Physical Review Letters, 2008. 100(23): p. 236601.
Krishtopenko, S.S. and F. Teppe, Quantum spin Hall insulator with a large bandgap, Dirac fermions, and bilayer graphene analog. Science
Advances, 2018. 4(4).
39. 39
Conical and parabolic subbands in low-fields
Krishtopenko, S.S., et al., Cyclotron resonance of dirac fermions in InAs/GaSb/InAs quantum wells. Semiconductors, 2017. 51(1): p. 38-42.
E2
E1 «conical»
40. 40
Conical and parabolic subbands in low-fields
Krishtopenko, S.S., et al., Cyclotron resonance of dirac fermions in InAs/GaSb/InAs quantum wells. Semiconductors, 2017. 51(1): p. 38-42.
E2
E1 «conical»
optical gating
optical gating
41. 41
Experimental results
Ruffenach, S., et al., Magnetoabsorption of Dirac Fermions
in InAs/GaSb/InAs “Three-Layer” Gapless Quantum Wells.
JETP Letters, 2017. 106(11): p. 727-732.
43. 43
Studies of the valence band
normal band
structure
Gapless
Dirac cone
inverted band
structure
Mikhailov, N.N., et al., Int. J. of Nanotechnology, 2006. 3(1): p. 120.
Dvoretsky, S., et al. J. of Electronic Materials, 2010. 39(7): p. 918-923.
Samples
•
•
•
•
•
4.6 nm
5.0 nm
5.5 nm
6.0 nm
8.0 nm
x=0
y=0.7
y=0.7
44. α– and β– is characteristic for p-type only
44
opaque regions
dQW = 6 nm (near gapless )
α−
𝐻 = 𝐻8∙8
53. A slide to impress the public
53
QC0541-THz2 @ 5,8·1011 cm-2
54. Magneto-optical spectroscopy
Powerful tool to study band structure
- characteristic B-field dependence and effects of non-
parabolicity
- determination and correction of the model parameters or
growth parameters
- access to matrix elements of transitions
Contactless but sensitive in the broad range of energies, beyond
gating possibilities. Sensitive to levels below pinning energy in the
valence band
Complimentary features:
• optical gating – speculative but yet useful
• polarization-sensitive – rich data
• temperature effects – valid for many systems
55. Transmission in magnetic fields
Band structure determination and controlled modification
Photoluminescence
Interband transitions between Landau levels that are forbidden within the
framework of the widely used axial dispersion law model are observed. It is due
to anisotropy of chemical bonds at heterointerfaces and the absence of an
inversion center in the crystal lattice.
Talk conclusions
55
FT-IR is a versatile technique allowing broadrange spectroscopy and vast applications
for semiconductors characterisation and research
Notes de l'éditeur
Good afternoon dear colleagues, I would like to thank everyone for coming and showing interest towards my talk.
The title one can read on the slide is representing my main project at EP-3. Today I will present you particular experiments and related results, so you may start think wider and consider further usage of spectrometer for you research.
Here’s the outline from my talk.
At first, I will introduce the device called FT-IR: it’s principle and applications followed by a motivation statement.
Then I will show two different experimental approaches for photoluminescence and transmission, from experimental setup toward selected results related to narrow-gap samples
An FTIR is typically based on The Michelson Interferometer Experimental Setup; an example is shown in Figure 1.
The interferometer consists of a beam splitter, a fixed mirror, and a mirror that translates back and forth, very precisely.
The beam splitter is made of a special material that transmits half of the radiation striking it and reflects the other half.
Radiation from the source strikes the beam splitter and separates into two beams.
One beam is transmitted through the beam splitter to the fixed mirror and the second is reflected off the beam splitter to the moving mirror.
The fixed and moving mirrors reflect the radiation back to the beamsplitter.
Again, half of this reflected radiation is transmitted and half is reflected at the beam splitter,
resulting in one beam passing to the detector and the second back to the source.
A Fourier Transform InfraRed (FT-IR) Spectrometer is an instrument which acquires broadband Near InfraRed (NIR) to Far InfraRed (FIR) spectra. Unlike a dispersive instrument, i.e. a grating monochromator or spectrograph, FTIR spectrometers collect all wavelengths simultaneously. This feature is called the Multiplex Advantage.
FT-IR Spectrometers are often simply referred to as FTIRs. But for the purists, an FT-IR is a method of obtaining infrared spectra by first collecting an interferogram of a sample signal using an interferometer, and then performing a Fourier Transform (FT) on the interferogram to obtain the spectrum. An FT-IR spectrometer collects and digitizes the interferogram, performs the FT function, and displays the spectrum.
On this slide is presented as manufacturer states – first maintenance free FTIR analyser
One may analyze solids, liquids, pastes, gels and intractable materials addressing the most common sampling needs. Spectroscopic performance (typical at 25°C with DTGS detector)
Spectral range 485 to 8,500 cm-1
Resolution better than 0.7 cm-1
Apodized resolution adjustable 1 to 64 cm-1, in increments of 2
Maximum signal-to-noise ratio (root-mean-square, 60s, 4 cm-1, at peak response): 50,000: 1
Signal sampling: 24-bit ADC
Short-term stability: < 0.09 %
Temperature stability: < 1 % per C˚
Frequency repeatability (@ 1918 cm-1): < 0.001 cm-1
Frequency accuracy (@ 1918 cm-1): < 0.06 cm-1
On this slide I’ve copied applications of high-end research spectrometer from Bruker.
Do not try to read, I will read it for you!
Basically, for pharmacy and chemistry, one can study everything that adsorbs at specific energies in the given region of infrared radiation.
While for semiconductors we mainly interested in electronic excitations.
Anyway, I put this slide just to clarify particular things - FTIR has a lot of applications in different fields,
so whatever I will present you today is a tiny piece of a usage.
Anyway, if you are not yet impressed here goes another slide, showing spectrometer itself and optional accessories one can purchase to extend it’s applicability.
In our case it’s important
to have vacuum inside the spectrometer
to have external optical inputs and outputs out of the spectrometer box
for example, having external output allows usage of He-cooled bolometer, standing nearby the spectrometer, because it is too big
Photoluminescence (PL) spectroscopy is a traditional method for semiconductors characterization. It provides rich data on the optical quality and properties of the material, like bandgap, exciton and impurity binding energies, deep-level defect presence, etc.
Band gap determination
Band gap is the energy difference between states in the conduction and valence bands, of the radiative transition in semiconductors. The spectral distribution of PL from a semiconductor can be analyzed to nondestructively determine the electronic band gap. This provides a means to quantify the elemental composition of compound semiconductor and is a vitally important material parameter influencing solar cell device efficiency.
Impurity levels and defect detection
Radiative transitions in semiconductors involve localized defect levels. The photoluminescence energy associated with these levels can be used to identify specific defects, and the amount of photoluminescence can be used to determine their concentration. The PL spectrum at low sample temperatures often reveals spectral peaks associated with impurities contained within the host material. Fourier transform photoluminescence microspectroscopy, which is of high sensitivity, provides the potential to identify extremely low concentrations of intentional and unintentional impurities that can strongly affect material quality and device performance.
Recombination mechanisms
The return to equilibrium, known as “recombination”, can involve both radiative and nonradiative processes. The quantity of PL emitted from a material is directly related to the relative amount of radiative and nonradiative recombination rates. Nonradiative rates are typically associated with impurities and the amount of photoluminescence and its dependence on the level of photo-excitation and temperature are directly related to the dominant recombination process. Thus, analysis of photoluminescence can qualitatively monitor changes in material quality as a function of growth and processing conditions and help understand the underlying physics of the recombination mechanism.
Surface structure and excited states
The widely used conventional methods such as XRD, IR and Raman spectroscopy, are very often not sensitive enough for supported oxide catalysts with low metal oxide concentrations. Photoluminescence, however, is very sensitive to surface effects or adsorbed species of semiconductor particles and thus can be used as a probe of electron-hole surface processes
This slide is very illustrative to my view, although the experiment is performed with grated spectrometer far from far-infrared. On the left picture the results for two samples with different preparation techniques are shown. Lets confirm that PL is at least sensitive to preparation technique, since we observe different response.
Both samples have peaks close to bandgap, but in first case its explained as free exciton radiative recombination, while in the second case – donor-acceptor pair is speculated in the paper.
Another bump is observed in the energy range of the oxygen antisite level – thus photoluminescence is sensitive for defects. The broad band centered at around 730 nm with a full width at half maximum (FWHM) of about 155.5 nm might be due to transition from conduction band to oxygen vacancy level.
Here I just show you how photoluminescence experiment can determine bandgap and access deep-level defects.
In the following slides I will show you how FT-IR was used in the group of our colleagues from Technische Physik. Their main activities are related to the fabrication and characterization of semiconductor nanostructures with a wide range of characterization tools. I would like to focus your attention towards interband cascade lasers in the range of 3 to 6 microns, where a typical PL sample is composed of 5 symmetric W QWs separated by 20 nm GaSb.
In the PL setup one use excitation with a 50 mW pump laser with a wavelength of 532 nm, radiation from the sample then send it as source signal to FTIR.
That is why it is important to have an option of external input in the spectrometer.
The charge carrier pairs generated due to the high excitation energy in the entire sample relax in the most energetically favorable state and can recombine there radiatively or non-radiatively. The spectrometer is equipped with a KBr and a CaF2 beam splitter and liquid nitrogen cooled detectors (InSb, HgCdTe).
Additionally, since the sample is located outside of spectrometer, one can make use of low temperature and other environmental effects in experiment.
Here’s two particular cases how PL helped in adjustment of growth parameters for such a devices, both taken from Robert Weih PhD thesis.
On the left picture he presented results for different substrate temperatures. The substrate temperature increases, a shift in emission towards shorter wavelengths can be seen. The highest intensity as well as the lowest half width (29.1 meV) could be determined for the grown at 450 ◦C sample.
On the right picture yet another useful experiment, this time with different Arsenic pressure in the growth chamber. Starting from the value reported in the literature, they try to use lower pressures to compensate lower substrate temperature and lower growth rate in their system. Slight decrease might be beneficial in terms of optical output but inevitably shifts the line towards shorter wavelength. Further decrease of the Arsenic flow leads to formation of visible defects reflecting in broadening of luminescence line. Since the lattice constant remains the same corresponding changes were attributed to modifications at the interfaces.
*** thesis google translate below
Variation of substrate temperature For identification of the optimum substrate temperature, a PL sample series was prepared in which the substrate temperature at the end of the buffer layer for the growth of the W-QWs was lowered to 440, 450 and 460 ◦C. Figure 5.8 shows the results of the PL measurement and the HRXRD measurements at a temperature of 20 ◦C. The resulting parameters are summarized in Table 5.1. As expected, as the substrate temperature increases, a shift in emission towards shorter wavelengths can be seen. The highest intensity as well as the lowest half width (29.1 meV) could be determined for the grown at 450 ◦C sample.
Variation of As flow
Starting from an As flow of (1.6 ± 0.2) × 10-6 torr, a series of samples was prepared under variation of the As flow. In the literature, an As flow, which allows stable InAs growth, is considered optimal [86]. Due to the lower growth rate used in comparison and the comparatively low substrate temperature, a variation towards lower As fluxes was made. In addition to the control sample, samples with As fluxes of (1.2 ± 0.2) × 10-6 torr and (0.8 ± 0.2) × 10-6 torr were produced in the InAs layers. The Sb flux during the growth of the Ga0.665In0.335Sb layer was (1.8 ± 0.2) × 10-6 torr due to the high growth rate. The results of the PL and HRXRD measurements are shown in Figure 5.9. It can be seen that as a result of the decrease in the As flux to (1.2 ± 0.2) × 10-6 torr, the PL intensity could be increased by more than 120% and the half width could be reduced to 24.4 meV. If the As flow is reduced further to a value of (0.8 ± 0.2) × 10-6 torr, the intensity decreases again and the half width increases. This is due to an increased formation of growth defects, which were already visible under the optical microscope. Obviously, the As flow used at a substrate temperature of 450 ◦C is no longer sufficient to enable high quality layer growth. Furthermore, as As flow decreases, there is a shift to higher transition energies, which can be explained by HRXRD measurements. In Figure 5.9b it can be seen that the superlattice reflections shift as a result of the change in the As flow. The mean strain of the structure increases with decreasing As flow. Since the superlattice period does not change, this can only be due to a change in the interfaces and thus the strength of the As for Sb exchange reaction. A higher As flux results in an increased formation of GaAs - like bonds at the interface. Furthermore, As - atoms remaining in the chamber are built into the Ga0.665In0.335Sb layer, which leads to a narrowing of the bandgap and a reduction of the strain, thus explaining the reduction of the transition energy with increasing As flow. Table 5.2 summarizes the results of the study. In conclusion, the As flux in the growth of InAs films in W - QW should be chosen to be so small that it allows stable InAs growth. By variation of the Sb flow during the growth of the Ga0.665In0.335Sb film, no significant change of the luminescence yield could be observed, as long as the Sb flux does not fall below a value of (1.5 ± 0.2) × 10-6 tor
So far so good, photoluminescence doing great job for visible range and near-infrared. What is about far-infrared range?
As many researchers are aimed at gapless or narrow bandgap systems, which examples are shown on the slide, one have to mention corresponding challenges
First and easiest – absorption of the air and water, just one have to establish a stable vacuum condition. I will highlight it here, since we’re speaking of seconds and minutes of integration time, the stability is valued higher than vacuum quality, typical pressure in Bruker spectrometer is <0.02 mbar
PL measurements in such narrow gap semiconductors are complicated
by (i) intensive room temperature background radiation;
(ii) a low probability of interband radiative transitions in the far infrared range.
Here I present the straightforward approach to measure PL spectroscopy, we excite carriers with laser and collect radiation from the sample. It goes through beamsplitter to form the interferogram at the detector.
The overall signal is superposition of small photoluminescence, huge thermal background coming from setup itself. For specific regions influence of internal He-Ne laser used for mirror positioning is also noticeable.
Here we introduce additional modulation, implemented in a combination of mechanical chopper and dual-channel LIA. Now, in principle, one can pick up only PL signal modulated at reference frequency. The highest sensitivity is achieved in step-scan mode, where mirror stays in the series of fixed position and signal accumulated after some delay to decouple effect of moving mirror.
In the following slides I would like to present you the setup that I‘d like to start with here in EP-3 reported in the work of my former colleagues from Nizhny Novgorod. The sample is located in cryostat and illuminated by the laser to generate electron-hole pairs in the surrounding QW barriers of CdTe. The transmission geometry of the experiment is available since GaAs substrate is transparent in the most of the IR-range.
Now, experiment. Here’s for samples with different compositions denoted as x and further parameters in the table. Top figure is for photoluminescence measurement at 18 K, where the typical peak corresponds to the bandgap value. As you may notice, going from 4 to 1, bandgap is decreasing as it should be since we approaching critical composition of 17% of cadmium in the ternary solution.
Lower figure corresponds to the photoconductivity measurements, where the signal is preamplified from the samples itself, in other words – no additional detector needed. Once you start to populate the valence band – conductivity signal dramatically increases, you can see the abrupt edges, corresponding to bandgap as well. The sharp eye may notice a slight shift which is explained by the temperature difference.
The PL spectra were also investigated in a wide temperature range from 18K to 200K at a fixed pump power. Fig. 4 exemplifies the PL spectra for sample #120613 measured at different temperatures. In contrast to most semiconductors,
in Hg1xCdxTe solid solutions with x<0.48 the bandgap increases with temperature.2 This results in the PL spectra shift to the higher frequency range.
Dependence of the PL line width on temperature is worth mentioning. In the case of interband radiative recombination of free carriers, the full width at half maximum (FWHM) of a PL line is known to be 1.8 kBT. However, for
all samples the PL line FWHM at 18K reached 4–6 meV, thus significantly exceeding the expected value, which is 1.55 meV at 18 K. As temperature is increased the PL line FWHM in kBT units decreases for all samples, until at
T70K it becomes 2 and does not change significantly with further increase in temperature. Table II shows the data for the PL line FWHM measured at different temperatures for sample #120613.
A significant difference between the experimental and theoretical values for PL line width at 18K shows that the additional shallow impurity/defect PL is important at low temperatures. Interband PL line broadening associated with
the manifestation of impurity PL at low temperatures has been observed in Ref. 9. At T>70 K, PL line width expressed in kBT units approaches the theoretical limit and is determined mostly by interband transitions of free carriers.
In order to determine the main mechanisms of photoexcited carriers recombination both PL and PC relaxation at pulsed optical pumping were investigated. Fig. 5 represents 3D picture of time resolved PL signal with time resolution of 400 ns (sample was excited with Nd:YAG laser pulses, kexc¼1064 nm, 10 ns in duration, and 3mm beam diameter).
One can see two emission lines in the spectrum. The long wavelength line (1) with maximum at 1210 cm1 results from the radiative recombination in the layer with low Cd content (x¼0.223). The second line (2) at 4230 cm1 corresponds to the bandgap of the lower “barrier” layer (x¼0.48; see Fig. 1).
At low excitation power, the PC decay is purely exponential (see Fig. 6(a)) and carrier lifetimes increase with the bandgap (see Table I). The latter indicates that carrier recombination is non-radiative in the low excitation limit. Review of literature data2,10,11 suggests that the dominating mechanism of carrier recombination is Shockley-Read-Hall process. It is natural to assume that in case of low defect density this recombination channel will be saturated with the growth of non-equilibrium carrier density and the onset of other processes (radiative and Auger recombination) will take place. Indeed, as the excitation power increases photoresponse decay becomes non-exponential: a region of faster relaxation process appears just after the excitation pulse (Fig. 6(a))
Expensive setup 100+ k€
Vacuum spectrometer with TRS and additional components as sources, beamsplitters, mirrors and external detectors for FIR and MIR range
In a month me and Florian will participate in the event in Nizhny Novgorod, performing not only scientific exchange and discussion but also take over particular experience in photoluminescence experiments.
Going back to outline to remind the one who fall asleep during photoluminescence experiment – you can have a fresh new start and follow up the transmission experiment.
Here’s the basic transmission experiment is illustrated. One have to block the optical path between beamsplitter and detector with a sample to reveal specific absorption.
The frequencies where absorption lines occur, as well as their relative intensities, primarily depend on the electronic and molecular structure of the sample.
The frequencies will also depend on the interactions between molecules in the sample, the crystal structure in solids,
and on several environmental factors (e.g., temperature, pressure, electromagnetic field).
The lines will also have a width and shape that are primarily determined by the spectral density or the density of states of the system.
As we are interested in narrow gap semiconductor system with small effective masses, characteristic energy of optical response belongs to far-infrared region.
Additionally, HgTe QW structures are sufficiently transparent in the far-infrared region, so the study of transmission spectra and polarization experiments could provide valuable information about band structure.
For a two-dimensional parabolic band in magnetic fields we obtain set of Landau levels, separated by the value of energy hwc, as one can see from this equation.
Optical transitions with resonant frequency allowed only between landau levels with index difference of plus or minus one, where sign defines the polarization of adsorbed light.
In real structures (and especially in gapless semiconductors) we can have non-parabolic conduction band, which gives rise to non-equidistant Landau levels.
In this case, energy of intraband transitions might be different for a fixed magnetic field,
so it can be resolved in the experiment, providing information about band structure and Fermi energy
In real crystal we have conduction and valence band . Each band gives a fan of Landau levels.
Thus, we can observe interband transitions, as beta and alpha-minus.
In our theoretical model selection rules are the same for both types of transitions.
It is worth to mention, that in gapless and narrowband structures,
energies of CR and interband transitions might be of the same range, as it seen from the picture.
With increase of magnetic field not only the spacing between Landau level increasing, but also degeneracy for each of them, so less and less Landau levels are occupied.
Due to the occupation effect, we cannot observe transitions from empty levels, as well as transitions to fully occupied levels.
The intensity of the line is proportional to amount of carries absorbing the light.
To summarize, observation of magnetooptical transitions can provide information concerning effective mass and bandgap of a system.
Magnetooptical picture obtained for a broad range of magnetic field coupled with occupational effect allow to reconstruct Landau levels,
directly linked with a band structure.
Here I present two different views on Landau levels, from transport point of view and from optics point of view.
I would like to highlight that in transport experiment one may no difference within internal nature of Landau level and everything is grayscale
In optics, on the contrary, one have to keep in mind landau level index and dominant spin to rule out forbidden transitions of low probability.
Another advantage of magnetooptics is that one can “feel” and follow energy of the Landau level if it goes beyond pinning energy in the valence band, as it happens for b- transition,
Having a gated structure one may reconstruct landau levels sweeping the gate voltage and observing different plateaus.
In the left corner the example of such reconstruction is shown, reaching 100 meV. Further sweeping is limited with the value of the gate voltage
On the other side – results of magnetooptical experiment in the bulk sample reaching 340 meV, which still can be enhanced by chooseing another optical configuration of source, beamsplitter and filters.
Here I combine PL experiment rotated by ninety degree, and transmission experiments, where all the absorption features are actually plotted in the other figure. In this particular case for each subband there is one dominant transition which you can extrapolate towards zero-field and and extract energy values useful for band structure understanding. Additionally, one may notice that additional transition e( is all the way on the magnetooptics but hardly seen in PL response.
In order to perform magneto-optical experiments a system of coupled Fourier-spectrometer and magnet is used.
Radiation from source is passing through beamsplitter and travel inside the light-pipe towards sample. Transmitted light is detected with a bolometer, located right after the sample, and send to computer after amplification for further analysis.
Sample and reference is mounted on top of rotational holder, so it is possible to perform absolute measurements of transmission. The holder is located here and then one may see attached bolometer. The light from spectrometer travel in parallel beam and entering the probe, located inside the cryostat with superconducting magnet.
So here I present raw spectra for transmission experiment in the empty tract. The overall response is defined by optical configuration, in other words combination of beamsplitter, source and detector. In red and blue are the signal for two different sealed bolometers.
Lower panel shows the ratio between two consecutive measurements integrated for 512 scans, which is a few minutes of time. The stability of the system is of high importance in that experiment.
Sensitivity is reaching to 0.1 % of the relative change with the magnetic field.
Let me show you some illustrative examples of magneto-optics
I am not aiming to explain you in details underlying physics for all the cases, I will just try to present you the tool of landau level spectroscopy, and its particular applications.
The two-layer InAs / GaSb QWs are the second semiconductor system in which the state of a two-dimensional topological insulator was predicted theoretically [96] and then discovered experimentally [97].
The GaSb valence band edge is by 0.15 eV higher than the bottom of the conduction band in InAs, which makes it possible to control the band structure of the two-layer InAs / GaSb QW by selecting the layer widths.
Moreover, QW itself bounded by AlSb barriers (specific material with a large bandgap) to provide overall confinement.
The lack of the inversion symmetry in the growth direction in the "double-layer" InAs / GaSb quantum wells leads to the "crossing" of the E1 and H1 subbands at k ≠ 0 and, consequently, to the gap opening in the spectrum due to the hybridization of the subbands. Alas, it is stated that one cannot obtain Dirac cone at the Г-point.
Luckily, the latter is possible in three-layer system and supported by theoretical calculations for a narrow range of growth parameters. As you may notice, second layer of InAs manifested as appearance of the second conduction subband.
One subband is conical, another is parabolic. As grown sample have an electron sheet density higher then 10^12 and thus unable observation of distinction effect between two subbands.
Inset show the magnetooptical experiment with QCL source at 3 THz and sweeping of magnetic field up to 5.7 Tesla. On the right – calculation of effective mass in k dot p theory, where for high concentration it coincide but diverge at the bottom of conduction band, where conical subband lost its rest mass.
Using optical gating one may decrease carrier density taking advantage of persistent photoconductivity effect.
Cyclotron line undergo two modifications, the area became smaller since less carrier participates in absorption,
but more important is a prominent broading, revealing co-existence of two conduction subbands supported with calculations in k dot p theory.
In quantizing (up to 34 T) magnetic fields, the presence of "conical" subbands is manifested in the appearance of a new absorption line due to a transition from the lowest Landau level in conduction band.
Here‘s a small bonus for you – photo of magnetooptical experiment in the high magnetic field laboratory in Grenoble, France, where I spent half of my PhD.
The only difference from low-field setup – is high quality collimated beam travelling from spectrometer towards cryostat with resistive magnet.
The samples were grown in Novosibirsk by the group of N. Mikhailov and S. Dvoretsky with MBE setup. GaAs was used as substrate, following thick layer of CdTe to remove the lattice mismatch, in other words not to introduce any strain. The pure HgTe quantum well is located between barriers, and protective layer is covering the structure. The samples of different width were studied, covering the whole range of normal, near gapless and inverted band structure.
Message – one can define type of conductivity and roughly evaluate concentration in quantized regime.
At the left part of the slide presented experimental data obtained for the almost gapless sample.
All the spectra corresponds to relative transmittance, in other words, normalized by zero-field spectra of the sample,
opaque regions are masked with grey.
On the left side solid lines are Landau levels and arrows denote observed transitions. The Fermi level is shown with dotted line.
At 10 Tesla all Landau levels in the valence band are occupied and all Landau levels in conduction band are empty.
There are two interband transitions observed, one of them is beta transition well known in the literature for studies of n-type samples.
In lower magnetic field we additionally observe intraband transition in the valence band denoted as beta-minus.
I’d like to notice that these transitions are characteristic for p-type samples only and may be used as reference in future studies.
At this slide experimental data is depicted with different symbols, while calculation in axial model is show with lines.
Since all of samples under consideration have relatively small density of holes, high field spectra contain mentioned earlier interband transitions alpha-minus and beta. Extrapolation of transition energy toward zero magnetic field allow to estimate energy gap of the sample. With increase of QW width the energy gap closes and then opens up again for sample with the inverted band structure.
At the low magnetic fields highest Landau levels in the valence band are depopulated, allowing observation of intraband transitions beta-minus and delta-minus.
Mixing of states are revealed in the experiment – you cannot sense it in the transport
Another interesting experiment – temperature dependence of the band structure
For the sample with normal band ordering two interband transitions alpha and beta shifts up with the temperature.
In the inverted ordering, gap first closes at 90K and then reopens again as one can track from magneto-optical response.
We also perform measurement of magnetotransport in 4-terminal Van der Pauw configuration to determine type of conductivity and carrier density in the sample. It is possible to modify carrier density with illumination of LED, located nearby the sample. In order to perform polarization-sensitive spectroscopy we use combination of fixed and rotating linear polarizers to study effect of Faraday rotation.
At the bottom of the page presented the color plot of absorption, where absorption features are shown in yellow, green and blue in order of increase the effect.
Initially, the Fermi level of the sample are located in the valence band with magnetooptical response shown here.
Illumination of the sample with LED introduce additional electrons from surrounding barriers.
Thus, we were able to lift up the Fermi energy to the conduction band.
In this particular case we observe strong absorption line within conduction band.
From the slope of cyclotron resonance lines one can obtain values of effective mass with classical formula.
It is seen, that the valence band is occupied with heavier particles.
Let me show you some illustrative examples of magneto-optics
