AFAIK in the telecom buisness nobody uses VCSEL laser diodes yet, although they have a potential for Dense Wavelength Division Multiplexing. What is the big disadvantage compared to FP, DFB or external laser diodes? Will there ever be a high volume production of any VCSELs?
VCSELs are not useful for optical sources to single mode std. fiber- limited to 2km, and some cases 10km reach. Telco networks enjoy economy of scale as you know with enormous SMF cables installed, so edge emitting lasers have been able to be produced for SONET/SDH, ATM, other links for distances in excess of 80km with source wavelegnths maching SMF low attenuation and 0-dispersion wavelengths (1310/1550nm). This allows manufacturing for higher bandwidth/data rate/longer reach applications. LD product that does not meet the highest specification can still be used for shorter reach links, as well as just relaxing production techniques for lower performance parts noting materials and processes are essentially the same (i.e. FP's for 2km SR links- almost same price as VCSEL these days). In this scenario, and understanding there are manufacturing similarities of FP and DFB LD's, is it now easier to understand? SMF fiber price reductions during the telecom/i-net bubble burst and a slightly proactive change in FP LD diode prices ~ 2000-2001 I think helped influence this too. VCSELs are extremely useful for coupling to multimode fiber and are improving every year. They did eat LED's lunch in communications. The SAN and LAN market is still using lots of MMF so it's not like VCSEL volumes are so poor. Optimizing VCSEL to the low dispersion wavelength of SMF (~1310nm) has been expected (feared) to put a dent in laser market share, but there are difficulties that stall this, mainly due to it's multimodal nature, and so VCSELa seam stuck in
850nm domain for mass production at present. I feel there is high volume VCSEL production now.
I'm sorry I have not mentioned I'm talking only about the 1.3/1.55um bands. AFAIK there are a couple of producers for monomode VCSEL especially for ~1.55um band, but these companies sell just for specialities in highly sophistcated measuremnt technology like NIR spectroscopy. They are very monomode,have high efficiency and lower output power compared with DFB. VCSEL for this range are relatively new on the market due to the technological problems with the different material (InP) used. Is this the only reason why they are not yet in the telecom market or are the other challenges left? Urs
I see. The Telecom market has been purely cost driven with equipment makers giving multi-million dollar systems away to carriers for testing (sometimes for over a year) before ever collecting a payment or even a solid purchase order. Therefore, these equipment builders (Alcatel, et. al.) have passed the pain on down the line to their component suppliers. These component guys can't finance their customer's future, so they just drop their prices. So I'm assuming that research is still needed to bring long wave VCSEL mass production to the levels of volume-cost efficiency that the LD fabs are at now. A real VCSEL expert should comment on exact technical challenges though as I'm out of touch w/VCSEL mfrng.