+44 (0)24 7671 8970
More publications     •     Advertise with us     •     Contact us
 
Technical Insight

InP looking good for future higher speeds (IPRM Conference)

Exotic compounds, new solutions to old problems and record-breaking device performance were just some of the things discussed at IPRM'01. Colombo Bolognesi reports.
The Thirteenth International Conference on Indium Phosphide and Related Materials (IPRM) was held May 1418 in Nara, Japan. A new attendance record was set with some 470 participants, including 210 from outside Japan. This year s IPRM conference program was rich with exploratory work on novel electronic/optoelectronic materials, including GaInNAsSb/GaNAs QWs, InMnAsSb/InSb heterostructures, GaInNP, TlInGaAs/InP double heterostructures and GaAs1xBix which features a temperature-insensitive band gap. Optical devices Arakawa et al. from the Furukawa Electric Company reported the use of CBr4 to perform in situ etching of InP in MOCVD reactors. Using SiN stripe masks at a temperature of 600C, it was found that InP is easily etched while AlGaInAs layers can be used as effective etch stop layers. For InP the etch rate was found to be proportional to the CBr4 flow in the chamber and etch rates varied with group III element. In etched the fastest, followed by Ga then Al. The etched surfaces were smooth enough for epitaxial regrowth with a reduced defect density and a lower residual oxygen concentration, as observed by SIMS measurements. Lee et al. from Osaka University reported temperature-stable TlInGaAs/ InP double heterostructure (DH) LEDs grown on (100) InP substrates with a Tl composition of 6%. The DH LEDs were operated up to 340C near 1.58 m and exhibited an electroluminescence temperature coefficient of 0.09 meV/K. The improved temperature stability could turn out to be an important asset for WDM laser diodes. A new type of electroabsorption (EA) modulator for high bit rate WDM applications was proposed and simulated in a presentation by Wakita et al. from Chubu University. The device consists of a transverse p-i-n structure using a strain-compensated AlGaInAs MQW stack grown on InP. This allows the field to be applied in a direction parallel to the stack layers. The structure is believed to allow greater incident optical power levels since the carriers generated flow parallel to the layers instead of escaping the MQW stack by crossing the barriers as in classical modulators. The predicted device performance includes a phenomenal 3 dB bandwidth of 250 GHz and a driving voltage of less than 2 V for a 20 dB extinction ratio. Takeuchi from NTT presented work on an ultrafast EA modulator integrated with a DFB laser in both lumped and traveling-wave electrode configurations. The traveling-wave EA-DFB exhibited a bandwidth of over 50 GHz as well as 40 Gbit/s eye diagrams. Ougazzaden et al. from Agere demonstrated tandem EA modulators suitable for 40 Gbit/s transmission over long distances using return-to-zero modulation. DC extinction ratios of 28 and 22 dB were measured at 3 V for 120 and 80 m long devices, respectively. The 3 dB bandwidth for the 80 m device was 40.5 GHz and a dynamic extinction ratio of 13 dB was achieved with a 2.5VPP drive at 40 GHz. Electronic devices In the field of HBTs, a group from Simon Fraser University reported InP/GaAsSb/ InP DHBTs with both ft and fmax values greater than 300 GHz. These operated at record current densities in excess of 500 kA/cm2. Mathew et al. from The University of California at Santa Barbara (UCSB) reported transferred substrate AlInAs/ GaInAs/InP DHBTs with ft = 165 GHz and fmax = 300 GHz. UCSB also reclaimed the fastest ECL static frequency divider record in a post-deadline paper. A couple of very impressive papers from l Institut d à&mille;lectronique et de Microlectronique du Nord (IEMN, France) revealed that better scaling can further improve the characteristics of sub-100 nm gate HEMTs built on metamorphic and/or InP substrates. The group showed results from a "standard structure" and an "optimized structure", both grown by MBE. The standard structure consisted of a 20 nm InGaAs channel (53% In), with a 5e12 cm2 Si delta-doping plane separated from the channel by a 5 nm InAlAs spacer layer. A 12 nm InAlAs (52% Al) top barrier was capped with a 10 nm n+ InGaAs layer. In the optimized structure, the top barrier was reduced to 8 nm and Al content raised to 65%. The Si delta-doping plane was increased to 6e12 cm2 with a 3 nm spacer. The In content of the channel was raised to 65%. Both structures gave well behaved HEMTs, with the standard structure giving an ft of 270 GHz and fmax of 260 GHz for a 60 nm gate finger. The optimized structure gave an ft of 240 GHz and an fmax of 470 GHz for a 70 nm gate finger (see ). Endoh et al. from Fujitsu reported using a similar layer structure (see ) to the IEMN group to fabricate the fastest transistor ever built. This HEMT had a short-circuit current ft of 396 GHz. The device featured a 25 nm gate and operated at VDS of 1 V with a 2 m sourcedrain separation. Are we running out of indium? The evening rump session theme was "InP-based materials and device research for the Internet era". This attempted to plot a course of development for InP research towards 2010. The rump panel speakers were Sawada (Sumitomo), Nelson (IQE), Oki (TRW), Kasukawa (Furukawa), Subbanna (IBM) and Wada (FESTA). Sawada launched the evening discussion by stating that his company faces market demands for higher quantities of InP substrates with larger diameters and improved flatness and cleanliness. The discussion soon migrated to the question of whether or not the world supply of indium metal can sustain the InP industry. The bad news was, it appears from the discussion, that the world may run out of indium within the next 20 years. The better news is that InP substrates consume only about 1% of indium worldwide, while the major user of indium metal is indium tin oxide (ITO). The supply of indium for InP technology may, in the long run, depend on the ability to recycle ITO. Thinking back on that discussion and recalling Marie Meyer s survey of gallium supply problems (see Compound Semiconductor Dec 2000/Jan 2001, p43) may tempt one to play devil s advocate and ask whether God really intended us to work with III-Vs at all after he gave us silicon heresy! Nelson (IQE) painted a bright financial picture for InP devices with an estimated $50 billion market by 2005. Nelson also suggested that metamorphic buffers on larger area, cheaper GaAs substrates might be a clever way to work around the indium supply problem. This comment prompted a query from Agilent s Tom Low who pointed out that some devices are temperature sensitive and benefit from the higher thermal conductivity of InP compared to GaAs. This further raises the question of how the thermal properties of metamorphics compare to InP substrates. Oki (TRW) provided an interesting point of view with his perception of the over-capacity in the cell phone and fiber-optics market, quoting the number of companies working on tunable VCSELs as 60. His expectations are that as the industry matures, many of these companies will disappear or be absorbed into larger units. Subbanna (IBM), who chose to make a career in the realm of silicon, rocked our fragile InP world with news of 120200 GHz SiGe HBTs built on 8 inch sturdy, cheap, and crystallographically near-perfect silicon substrates. In terms of reliability, Subanna pointed out that Si has no clear failure modes like those of III-Vs, and that defect diffusivity in Si is some 10 orders of magnitude lower than those found in GaAs at 150C. The full technical program of the Thirteenth IPRM can be found at www.knt-ec.com/IPRM01/program.html.
×
Search the news archive

To close this popup you can press escape or click the close icon.
×
Logo
×
Register - Step 1

You may choose to subscribe to the Compound Semiconductor Magazine, the Compound Semiconductor Newsletter, or both. You may also request additional information if required, before submitting your application.


Please subscribe me to:

 

You chose the industry type of "Other"

Please enter the industry that you work in:
Please enter the industry that you work in: