close

Вход

Забыли?

вход по аккаунту

?

el%3A19970209

код для вставкиСкачать
pyramidal pit of an Si (111) surface [4], it is not preferable because
the following tip metal fabrication is difficult. The LT-GaAs layer
was grown on the Si substrate using molecular beam epitaxy. The
growth temperature was 300°C. After lpn of growth, the wafer
was annealed at 600°C for l0min. The growth rate was 1 W h .
The equivalent pressure ratio of As, to Ga was -10. The RHEED
pattern was spotty during growth and the annealing process. This
is reasonable because LT-GaAs on a Si substrate becomes close to
a poly-crystal due to lattice mismatching. After patterning the LTGaAs film becomes a triangular cantilever. The Pt film was sputter evaporated to make the probe tip electrode and lead strip. The
tip electrode is necessary to make electrical contact with a DUT.
Therefore, a hard metal which has a chemically stable surface is
preferable. The thickness of the Pt film is 300 nm and the final tip
radius is 700 nm. Finally, the Si substrate under the cantilever
region was removed by lapping and then followed by chemical
etching with a KOH solution. In this process, although a LTGaAs film without annealing is easily dissolved in KOH solution,
the annealed one survived perfectly.
The resulting special probe worked well. Mechanical strength
and flexibility of the cantilever is good enough to bend more than
60”. No tip abrasion was found by a SEM even after image observation of a 1 pn square area 100 times. The calculated force sensitivity of the probe is 0.3N/m.
the incident beam is absorbed by the PCSS to generate carriers
and the other 30% is reflected to a beam position detector. A normal SFM uses a CW laser beam to detect the probe position. The
SFOEM uses just one pulse beam for both the probe position control and PCSS switching. Since the periodic frequency of the optical pulse train is sufficiently high, the pulse beam operation causes
no disturbance to probe position control. A pulse shape obtained
with a 2.4ps FWHM is shown in Fig. 4. From this result the time
resolution of the SFOEM is better than 2ps.
Conclusion: We fabricated a LT-GaAs probe for the SFOEM to
measure the ultrafast signals of ultra-small devices. The probe
worked well and obtained a 2.4ps pulse which means a time resolution of better than 2ps.
0 IEE 1997
Electronics Letters Online No: 19970197
18 November 1996
K. Takeuchi and A. Mizuhara (Teratec Corporation, 2-9-32 Naka-cho,
Musashino-shi, Tokyo 180, Japan)
References
and
MIZUHARA, A.:
‘Scanning
tunneling
optoelectronic microscope with 2 ps time resolution’, Electron.
Lett., 1996, 32, (18), pp. 1709-1711
2 HOU, A.S , HO, F., and BLOOM, D.M.:‘Picosecond electrical sampling
using a scanning force microscope’, Electron. Lett., 1992, 28, (25),
pp. 2302-2303
1 TAKEUCHI, K.,
3 GUPTA, S.,
FRANKEL, M.Y.,
SMITH, F.w.,
VALDMANIS, J . A ,
WHITAKER, J.F.,
and CALAWA, A.R.: ‘Subpicosecond
carrier lifetime in GaAs grown by molecular beam epitaxy at low
MOUROU, CA.,
4
temperatures’, Appl. Phys. Lett., 1991, 59, (25), pp. 3276-3278
AKAMINE, s.,
CARVER, T.E.,
and QUATE, c.F.::
‘Microfabrication of cantilever styli for the atomic force
microscope’, J. Vac. Sei. Technol., 1990, AS, pp. 3386-3396
ALBRECHT, T.R.,
M
delay line
*
piezo scanner
Results of a two year radiometric
measurement programme in New Zealand
1435131
M.J. Rodda and A.G. Williamson
Fig. 3 Setup of pump-probe measurement using SFOEM
Measurement: Fig. 3 shows a system setup for the SFOEM measurement. It is a so called pump-probe method [I]. DUT are coplanar strips (CPS) fabricated on a LT-GaAs layer on a semiinsulating GaAs substrate. It is the same as that used in the previous STOEM work. We measured an electrical pulse generated by
illuminating the gap region between two strips with an optical
pulse train with 1OOfs duration, 80MHz periodic frequency, and
830nm wavelength.
5
time, p s
10
Annual and cumulated two year statistics of rainfall rate, sky
noise temperature and attenuation are presented, derived from
two years radiometric measurements at 11.6GHz in New Zealand.
Total attenuation statistics are well predicted by both the ITU-R
method and the PARC model when a rain height of - 2 . 3 h is
used. The ITU-Rs ‘ K zone rainfall rate profile accurately
represents the two year cumulative rain statistics.
Introduction: This Letter presents both single and two year cumulated statistics of radiometric measurements at 11.6GHz, collected
from April 1993 to March 1995. The objective of the programme
was to obtain, using co-located equipment, concurrent point rainfall rate and slant-path Ku-band attenuation statistics. Statistics
from the first year of measurements at the New Zealand site have
already been published [l]. Hereafter, the first year of measurements (April 1993 to March 1994) will be referred to as Year 1,
and the second year (April 1994 to March 1995) as Year 2.
.......... .............
,......
0
Indexing terms: Radiometry, Rain, Radiowave propagation
15
Site: The equipment was installed at the Tamaki campus of The
University of Auckland. The site has co-ordinates 174.86 E, 36.89
S and is -42m above sea-level. The antenna’s elevation angle was
42.3” and its azimuth angle was 0.5” west of true north. New Zealand is ascribed a ‘ K type rain climate zone by the ITU-R [2].
Auckland’s average annual rainfall is -1 198mm.
Fig. 4 Photo-excited electrical pulse measured by SFOEM
A probe beam separated from the pump beam is introduced
through a singlemode optical fibre and focused onto the PCSS on
the probe. The average beam power at the probe is 1mW. 70% of
326
Equipment: The equipment, provided by INTELSAT, was very
similar to that used in the joint African radiometric programme
[3]. A 1.8m diameter parabolic reflecting antenna was used in conjunction with a Dicke-switched radiometer operating at 11.6GHz.
ELECTRONICS LETTERS
13th February 1997
Vol. 33
No. 4
The system was calibrated at regular intervals by external application of hot and cold loads. Rainfall was measured using a colocated tipping bucket rain-gauge.
Fig. 2 shows the worst month rainfall rate profiles for each of
the two years. The more intense events tended to occur about the
equinoxes (MarcWApril and SeptemberIOctober) rather than the
traditional mid-winter months (July, August).
Measurements: Rainfall-rate measurements were based on the time
interval between gauge tips. Total attenuation was calculated
using radiometer measurements in conjunction with an assumed
physical medium temperature of 285K and an antenna integration
(‘H’) factor of 0.95. Over the two year period, the system uptime
(when both the radiometer and rain-gauge were operational) was
99.36%; the uptime for Year 1 and 2 being 99.94 and 98.78%,
respectively.
Results Cumulative statistics of rainfall rate, sky noise temperature and total attenuation have been calculated for both years and
also cumulated over the two-year period. Event statistics of rainfall rate and attenuation have been derived over similar periods.
120
16
m
12
U
sa
3
2 L
~
0
103
10-1
100
% tlme atten > ordlnote value
10-2
10 1
102
Fig. 3 Annual and worst month attenuation statistics
year
worst
(1)
2
month
(11) year 1 worst month
(111)
1
(iv) year 2
year
;loo
~~- -
annual statistics
annual statistics
E
E 80
2 60
-6
i
B
LO
20
0
lo3
IO-^
IC?
100
% tlme rolnfaU rate >ordinate
IO‘
value
102
m
Fig. I Comparison of measured rain statistics and ITU-R rain profiles
(i) 2 year measured
(io ITU ‘ K
(iii) ITU ‘H’
(iv) year 2 measured
(v) year 1 measured
(i) Rainfall rate measurements: Fig. 1 shows five rainfall-rate distributions. They are the measured cumulative distributions for each
individual year as well as the two year cumulated result, and the
standard ITU-R rain profiles for regions ‘H’ and ‘ K . Year 1 was
significantly drier than average and its statistics are well approximated by the ITU-R’s ‘H’ zone profile. Year 2 was slightly wetter
than average; rain rate statistics cumulated over both years are
accurately modelled by the ‘ K region ITU-R rain profile.
A comparison of the two year measured rain rate statistics with
predictions by the Rice-Holmberg model has been carried out.
Using an average annual accumulation (M) of 856” (the mean
of the two annual rainfalls) it was found that no single value of
the Rice-Holmberg model parameter, p, provided a good fit to the
data over the entire range of percentage times. p represents the
ratio of rainfall produced by convective rain processes to the total
annual rain accumulation. Given that many attenuation prediction
models require knowledge of the rain rate if it exceeded 0.01% of
the time (Go,),there is perhaps good reason to bias the choice of
p so it provides accurate predictions around this percentage time,
in which case p turns out to be in the range 0.12-0.14. In the past,
New Zealand has been associated with lower values of p; 0.05 to
0.1 [4]. Values in this range tend to yield accurate predictions for
percentage times <0.01%, but under-estimate the rain rate at
larger percentage times.
(ii) Attenuation measurementr The cumulative and worst month
attenuation statistics of each year are displayed in Fig. 3. Fig. 4
compares the two year cumulated result with predictions made
using the ITU-R attenuation model [5] and the PARC (Propagation analysis for ram and clear air) model.
43“lh was
Both prediction models require knowledge of hO,;
used for all predictions, this figure coming from the measured two
year cumulated statistics, although the ITU-R ‘ K zone rain
profile would also have promded an accurate estimate of Go,
(42“ih). Both prediction models involve an effective ram height
based on the height of the 0” isotherm The curves labelled ‘ITUR 3.94km’ and ‘PARC 4 k”were derived when the models used
their own internally generated rain height estimates based on the
earth station’s latitude. These predictions over-estmate the attenuation exceeded for 0 01% of the tme by 37.6 and 33.5%, respectively The generally accepted upper bound for year-to-year
variation in attenuation is 34%. Local meteorologists report that
the height of the freezing layer varies from 1.5 to 2km in wmter,
increasing to a maxnnum of 3 km in summer. Realistically an
average rain height would he close to 2 km. A rain height of
2.3km provides a best fit to the cumulated two year measurements
and a prediction by the [TU-R model using this rain height is
shown in Fig 4
16
m12
m
C
‘
5
8
3
‘2
5 L
0
10-3
IO?
1U’
100
101
% ttme atten =. ordinote value
102
Fig. 4 Compuriwn of measured attenuation statistics with ITU-R and
PARCmodels
~
two year measured
3 4 km
PARC4km
I T U - R 2 3 km
A ITU-R
W
0
It should be noted that both the ITU-R and PARC models only
predict rain attenuation. The radiometric measurements are of
total attenuation so 0.5dB (gaseous attenuation) was added to the
raw predicted results to yield the curves shown in Fig. 4.
10-3
104
10-1
100
% time ordinate exceeded
IO‘
(iii) Event analyses: In addition to the cumulative analyses presented in this Letter, event and inter-event analyses have been carried out. Attenuation event data for the two year period is
summarised in Fig. 5. The vertical (log of frequency) axis indicates
the number of times an attenuation threshold was exceeded for a
specified duration interval. All attenuation events exceeding the
102
161112/
Fig. 2 Worst month rain statistics
(i) year 1
(ii) year 2
ELECTRONICS LETTERS
13th February 1997
Vol. 33
No. 4
327
8dB threshold lasted <1Omin. A similar event analysis of rainfall
rate (not presented here) showed that all events at thresholds
2 4 0 " h lasted <lOmin.
1.26W CW iff fraction- limit^^ InGaAs
lifier at 780nm
S. O'Brien, R.S. Geels, W.E. Plan0 and R.J. Lang
Conclusions: The main results of a two year radiometric and rainfall measurement programme have been presented. The first year
was significantly drier than average; the second year was marginally wetter than average. The cumulated two year statistics of rainfall rate are accurately modelled by the ITU-R ' K zone rain
profile. The Rice-Holmberg rain model is less accurate but gives
good predictions of 0.01% time exceedances when a p value of
0.12 is used. The ITU-R and PARC attenuation prediction models
give very similar results; both attenuation models over predict the
cumulated two year statistics when estimates of the rain height
based on station latitude were used. A rain height of -2.3km produced accurate attenuation predictions.
Indexing terms: Semiconductor junction lasers, Semiconductor
optical ampliJevs
Flared amplifiers fabricated with InGaAsP alloys have produced
1.26W CW at a wavelength of 780nm in a single-lobed,
diffraction-limited far field pattem. Under quasi-CW conditions,
2.4W of diffraction-lited power has been obtained. This
represents the highest diffraction-limited power reported at
780nm and the highest diffraction-limited power reported in the
InGaAsP material system.
Discrete and monolithic master oscillators power ampllfiers
(MOFAs) utilising flared amplifiers have achieved the highest diffraction-limited output powers from diode lasers. Flared amplifiers have achieved between 1 and 3W CW while wide aperture
flared amplifiers have reached output powers of 5W CW [l 71.
Most previous work with flared amplifiers has involved the use of
the (In)GaAs/(&)GaAs material system operating at wavelengths
between 860 and 1020 nm. However, at shorter wavelengths the
fabrication of monolithic photonic integrated circuits such as
monolithic MOPAs (M-MOPAs) is complicated by the presence of
Al-containing cladding layers. For example, strong oxide layers
formed on exposed AlGaAs surfaces can prohibit the high quality
regrowth necessary for grating formation with M-MOPAs.
Recently, the Al-free InGaAsP/GaAs material system has been the
subject of considerable development, demonstrating high incoherent powers at 810 and 980nm [8, 91. In this Letter we describe the
fKst use of InGaAsP/GaAs materials in flared amplifiers and the
first demonstration of >1W CW diffraction-limited output power
in the 780nm wavelength regime from any material system. Flared
amplifiers are fabricated which emit at 780nm and operate in a
diffraction-limited far field pattern to 1.26W CW and to 2.4W
under quasi-CW conditions.
The InGaAsP materials used in this work were grown by metal
organic chemical vapour deposition (MOCVD). The epitaxial
structure, grown on a GaAs substrate, consisted of a standard separate confinement heterostructure with an InGaAsP quantum well
active region (h -780nm) surrounded by InGaAsP waveguide and
cladding layers. Broad area lasers 1.5"
long with l o o p apertures fabricated from the material operated to 4.SW CW and to
15W under short pulse (40011s at 1kHz) conditions. The 1.5"
long lasers exhibited slope efficiencies of 0.92WlA and a threshold
current density of 230Aicmz. Plots of threshold current density
against inverse cavity length indicated a transparency current density of 110A/cm2while a plot of inverse differential quantum efficiency against cavity length yielded an internal loss of 4cm-1. The
vertical divergence of the lasers was measured to be 27" full width
at half maximum (FWHM) and 52" full width at the l/e2 points.
~
1
intervd,s
*"""
m
Fig. 5 Two year attenuation event analysis, April 1993
~
Murch 1995
The results reported in this Letter were obtained in the perfonnance of Contract INTEL-1076 for the International Telecommunications Satellite Organization ('INTELSAT'). Any views expressed
in this Letter are those of the authors, and not necessarily those of
INTELSAT.
Acknowledgments: The authors would like to thank F. Haidara
(INTELSAT), J. Allnutt, D. McCarthy (formerly with INTELSAT) and C. Zaks (formerly with COMSAT Laboratories) for
their enthusiastic support of this work.
0 IEE 1997
Electronics Letters Online No: 19970209
I 1 December 1996
M.J. Rodda and A.G. Williamson (Department of Electrical and
Electronic Engineering, The University of Auckland, Private Bag 92019,
Auckland, New Zealand)
References
master oscillator
M.J., and WILLIAMSON, A.G.: 'Results of 11.6GHz
radiometric measurements in New Zealand', Electron. Lett., 1996,
32, (4), pp. 397-399
RODDA,
Recommendation ITU-R PN.837-1, Propagation in Non-Ionized
Media, 1994 PN Series Volume
ARBESSER-RASTBURG,B., ZAKS, c., MCCARTHY, D., ROGERS, D.v., and
ALLNUTT, J.E.: 'The joint African radiometric measurement
programme', Int. J. Sat. Commun., 1990, 8, pp. 141-149
RICE, P.L., and HOLMBERG, N.R.: 'Cumulative time statistics of
surface-point rainfall rates', IEEE Trans., 1973, COM-21, pp.
1131-1136
Recommendation ITU-R PN.618-3, Propagation in Non-Ionized
Media, 1994 PN Series Volume
328
[;%:4
InGaAsP
flared amplifier
Fig. 1 Schematic diagram offlured ampl$er setup
Master oscillator is (A1)GaAs singlemode laser operating at 780nm
Flared amplifier is fabricated with InGaAsP alloys
A schematic diagram of the experimental setup is shown as an
inset in Fig. 1. The singlemode output from a discrete (AIGaAs)
Fabry-Perot master oscillator operating at 780nm is injected into
the InGaAsP flared amplifier with a high NA 1:1 imaging optics.
The maximum output power from the master oscillator was
5SmW, as measured after the collimating lens. The width of the
emitting aperture of the flared amplifier is 1 3 0 ~ .
ELECTRONICS LETTERS
13th February 1997
Vol. 33
No. 4
Документ
Категория
Без категории
Просмотров
2
Размер файла
462 Кб
Теги
3a19970209
1/--страниц
Пожаловаться на содержимое документа