Alternatives

• Author: Clemm, Christan; Löw, Clara; Baron, Yifaat; Moch, Katja; Möller, Martin; Köhler, Andreas R; Gensch, Carl-Otto; Deubzer, Otmar
• Pages: 35-38

RoHS Annex II Dossier, final

Indium phosphide

35

9 ALTERNATIVES

9.1 Availability of substitutes / alternative technologies

Optoelectronics

Indium phosphide, and alloys of indium phosphide with related compounds such as InGaAsP or

InGaAlAs, are considered to be unparalleled for use within transmitters or receivers in fibre optic

communications systems at 1.3 µm or 1.55 µm wavelength ranges.92

According to SMART Photonics, InP is the only direct bandgap semiconductor with a bandgap that

can be tuned to emit at a wavelength range between 1,200 and 1,700 nm. This functionality is

considered to be needed for optical communications > 1km to 10,000km with no substitute available

for reliable lasers in this wavelength range.93

At some wavelengths, however, substitution of InP by other semiconductor materials appears to be

possible, but according to Coherent, this would be associated with reduced performance, higher

production cost, unknown reliability, etc. Gallium arsenide could be an alternative for InP in some

applications that are not sensitive to the emission wavelength of the laser.94

As pointed out by Infinera, early in the development of optoelectronic devices for fibre-optic

communications, gallium arsenide (GaAs)-based devices, which are ideally suited to 0.85 m

transmission, were tested at 1.3 or 1.5 m. T hese efforts, however, ultimately failed due to high-

defect density that is inherent to highly-strained or lattice-mismatched, indium-containing alloys

grown on GaAs substrates. Based on these results, indium phosphide was established as the

substitute for GaAs devices and apparently offered better technical performance. In the early history

of optoelectronic device development, suggestions to use devices based on II-VI semiconductors95

(such as CdZnSe) were also made but were abandoned in the 1990s due to high defect density and

poor mechanical stability inherent in these materials.96

Moreover, Infinera mentions that some commercial suppliers of optoelectronic components

operating at 1.5 m employ a silicon-based photonics technology. However, in Si photonics, the

active devices are fabricated from InP and placed on a Si substrate for integration with other optical

functions. Therefore, Infinera considers Si photonics to be a viable integration technology for InP-

based devices, but not to represent a substitution path for InP. The rationale for this conclusion is

that silicon by itself cannot be used for lasers or direct amplification. For Infinera, silicon appears to

be ideal only for simpler, single wavelength applications, and for co-packaging with active devices

in the “pluggable” market for client optics and metro transponders. Likewise, optoelectronic devices

emitting at other wavelengths (IR, visible, and UV) may be fabricated from other III-V materials and

may find other commercial applications. Infinera concludes however, that there is no alternative

available to InP for high-capacity, long-haul networks based on DWDM technology.97

Photodetection devices could provide the best opportunities for substitution. According to

Lumentum, for this application field, germanium would be a suitable substitute when integrated with

92 Op. cit. Lumentum (2018)

93 Smart Photonics (2018), Contribution of Smart Photonics submitted on 15.06.2018 during the stakeholder consultation

conducted from 20 April 2018 to 15 June 2018 by Oeko-Institut in the course of the study to support the review of the

list of restricted substances and to assess a new exemption request under RoHS 2 (Pack 15); find the link in the annex

94 Op. cit. Coherent (2018)

95 II-VI compound semiconductors are obtained by combining group II elements with group VI elements.

96 Op. cit. Infinera (2018)

97 Op. cit. Infinera (2018)