As already indicated, the photovoltaic market is absolutely dominated by silicon. Only a tiny fraction (0.4%) corresponds to non-silicon materials, such as CdTe, ClGS, GaAs, etc. (see Fig. 2). In the last few years, poly-silicon has been gaining share at the expense of monocrystalline silicon. Both approaches altogether, taking 84% of the market, can be

called wafer technologies. The 16% remaining is basically shared by amorphous silicon (both for indoor and outdoor applications), ribbon silicon and silicon heterojunctions.

It is quite clear that p-i-n amorphous silicon cells, which once reached nearly 30% of the PV market, have been loosing share not only in favour of wafer-based devices, but also, and very particularly in favour of silicon heterojunctions, who have conquered 6% of the total PV market in very short time. Very remarkable is the fact that this whole share corresponds to a single company: Sanyo (Japan). The group, who named their cells HIT® (Heterojunction with Intrinsic Thin layer) achieved already in 1994 20.0% efficiency in a

1- cm2 cell of this kind, and have reached 17.3% for 100-cm2 cells at industrial level and 15.2% for production modules4. These figures are unambiguously indicative of the enormous potential of SHJ technology.

In contrast to the well-defined and well-funded Japanese PV programme, the situation in Europe is characterised by a strong diversity in national programmes, some of which are individually strong, but with a high degree of fragmentation. On the other hand, a successful European RTD strategy imposes the collaboration of a large number of public and private institutions. In the 6th Framework Programme (FP6), positive actions have been launched to increase collaboration and co-ordination. A number of research groups and companies are working hard on SHJ cells in Europe. Some projects are now running under FP5 based on hybrid technologies and low-temperature formation of the junction as a main issue, such as MOPHET, METEOR and ADVOCATE7. Their individual results are promising and reveal an excellent scientific level. The need to address the fragmentation of European R&D in this field must however be recognised by creating a permanent structure ensuring the harmonisation of the whole R&D on SHJ cells. Here are just some non-exhaustive highlights of the ongoing activity of European groups involved in SHJ-cell research, with indication of some of their main lines of action (given in alphabetical order of organisations):

v CNR-IMIP (I): PECVD deposition of a-Si, pc-Si, SiC alloys; Er-doped silicon thin films; fully pc-Si thin films with very low H-content even at low temperatures; growth of pc-Si on plastic substrates; optimisation of c-Si/pc-Si heterojunction solar cells (p=15%)

ч/ CNR-IMM (I): SHJ cells using single and double microcrystalline / amorphous emitters (Л=14%).

* ECN(NL): Processing of inorganic thin-film cells: development of methods to grow silicon films by plasma sprayed silicon on ceramics. Advanced surface and bulk passivation and silicon nitride anti-reflection. Development of special ceramic substrates for film silicon applications and "adapted” solar cell concepts.

* Fh-ISE (D): c-Si thin-film solar cells (high-temperature approach) on low-cost silicon and ceramics. Methods: Si-APCVD (using SiHCl3 at ~1200°C), zone melting recrystallisation Reference cells: one-side contacting (Si thickness ~45pm) q=19.2%, direct epi (37pm) on p+-Cz: 17.6%,: Activity foreseen for the future: Transfer of results on reference cells to "realistic” substrates, scaling of the substrates / cells to 10x10cm2. Bringing c-Si thin-film to a state where pilot production can begin.

* HMI (D): Deposition of amorphous and thin-film crystalline-silicon solar cells by PECVD and ECR-CVD; defect spectroscopy by optical, electrical and electron-spin resonance methods; seed-layer approach for a poly-Si thin solar cells on glass:metal-induced crystallisation techniques for seed-layer formation, epitaxial growth a T<600°C by ion — assisted deposition techniques, low-temperature emitter technology. Solar cells using heterostructures such as a-Si:H / c-Si. Results: q=17.1% for a-Si:H(n)/c-Si (p) cells using FZ — cSi wafer. Hall-mobility measurements, in-situ lifetime measurement during plasma process.

* LPICM (F): study of plasma processes and growth of thin films through the use of in­house in situ diagnostic techniques such as UV-visible ellipsometry; Kelvin probe and Time Resolved Microwave Conductivity; stable single-junction p-i-n solar cells based on polymorphous silicon with efficiencies close to 10%; simple dry process to passivate silicon wafers with a quality comparable to that achieved by using standard HF cleaning procedures.

* TU-Delft (NL): microcrystalline, protocrystalline Si and a-SiGe:H by PECVD and Expanding Thermal (Plasma ETP-CVD CVD; studies of defect-state distribution in a — Si:H and pc-Si:H, PECVD; single — junction a-Si:H solar cells ninit=10.3% (no back reflector); tandem a-Si:H/a-SiGe:H solar cell r|init=8.7% (no back reflector); Rd(i)=0.1-

0. 2 nm/s; ETP CVD: single junction a-Si:H solar cell ninit=6.7% (no back reflector), Rd(i)=0.85 nm/s.

* Univ. Barcelona (E): HWCVD microcrystalline p-type emitters, HWCVD microcrystalline cells at low temperature (<200°C). Plans for next future: development of HIT structures with HWCVD-deposited emitters.

* Univ. Ljubljana (SL) and Rome (I) "La Sapienza”: modelling and simulation together with characterisation of thin-film semiconductor materials and electronic devices; simulator with electrical model for heterostructures and with optical model for multilayer structures with smooth or rough interfaces. Particularly remarkable are the activities done by Rome University in collaboration with ENEA by using chromium silicide as the activated layer in high-conductivity emitters in a-Si / c-Si HJ. In particular an efficiency of 17% was obtained on a CZ Si based device of 2.25cm 2 total area.

* Univ. Patras — Plasma Technology Laboratory (GR): Process development, Scale-up, plasma modelling, characterization and control for high rate deposition of amorphous and microcrystalline silicon.

The institutions of the authors of the present paper, who hold a mutual collaboration on

several projects, and are deeply involved in SHJ, do in turn develop the following activity:

v CIEMAT DER (E): applications of deposited silicon (amorphous, microcrystalline and hybrid silicon). PECVD of amorphous silicon at high growth rates. Dry etching and passivation of c-Si surfaces. Wide-bandgap, highly conductive emitters. Development of silicon p-i-n and HJ cells and position sensors.

v ENEA (I): CR Portici: amorphous silicon solar cells using a wide range of technological options (Si, SiN, SiC, SiGe, and a-Si multijunctions, Hot-Wire CVD and VHF-PECVD), laser-scribing full technological process for large-area modules, seed layer by standard LPCVD for epitaxial growth of thick polysilicon films, Solid-Phase Crystallisation (SPC), and Laser-Induced Crystallisation (LIC), development of processes suitable for industrial applications (screen printing, TCO), passivation and dry / wet conditioning of silicon surfaces for a-Si / c-Si HJ technology (n=17% 2,25 cm[1] [2] [3] on CZ Si based devices in collaboration with Rome Univ.), and in collaboration with CR ENEA Casaccia PV lab, Laser doping and screen printed contacts in cSi Technology.

CONCLUSIONS

Photovoltaic research is looking for breakthroughs which can make solar cells competitive as soon as possible. Silicon absolutely dominates the market and will continue to be in a preferential position for a long time. The evolution of wafer technology largely depends on our capacity to develop new cheaper and thinner silicon, whereas thin-film silicon research is focused on improving material quality and crystallinity. In this scenario, a novel, hybrid approach, that of silicon-heterojunction cells, has already demonstrated its capacity to introduce an important milestone in the search of new approaches for photovoltaics. This is possible thanks to the combination of the best of two worlds: the good carrier lifetimes of wafer and ribbon silicon, and the versatility, accuracy and good passivation properties of thin films. The best industrial PV modules so far are silicon-heterojunction devices. This technology has already conquered 6% of the global PV market in a very short time. An important effort should me made in Europe to support and co-ordinate the excellent work done by the European groups involved in this research. This is the only way to quickly fill the present gap with Japan.

ACKNOWLEDGEMENTS

The authors would like to thank for the collaboration, the information received or the discussions held with: J. Andreu & J. Bertomeu (Univ. Barcelona, E), G. Bruno &

M. Losurdo (CNR-IMIP, I), W. Fuhs, S. Gall & F. Wunsch (HMI, D), D. Mataras (Uni-Patras, GR), F. Palma & G. De Cesare (Rome University, I), S. Reber (FhISE, D), P. Roca i Cabarrocas (PICM, F), F. Zignani & C. Summonte (CnR-IMM, I), M. Topic (Univ. Ljubljana), M. Zeman (TU Delft, NL), W. J. Soppe (ECN, NL), and last but not least, J. J.Gandia (CIEMAT, E), M. Tucci & E. Bobeico (ENEA, I). The financial contribution to silicon — heterojunction research by the European Commission by way of the Mophet project is also acknowleded.