ECS Journal of Solid State Science and Technology, 1 (6) P263-P268 (2012) P265

Figure 4. SEM top view of the LC GeSi film with pre-patterned lines;pre-patterning is indicated by the parallel lines. The large grains ((2–50)× (2–3) μm2) are confined in between the pre-patterned lines (i.e. withinthe 3 μm wide space). Reprinted with permission from B. Rangarajan, I.Brunets, P. Oesterlin, A. Y. Kovalgin and J. Schmitz, MRS Proceedings,Vol. 1321 (2011).

the grains within a defined space (3 μm wide), which may lead tobetter-controlled characteristics of properly-placed devices.

The samples were further compared to as-deposited poly-Ge0.85Si0.15 films of similar thickness. As visible in Fig. 6, these filmsexhibit smaller-sized grains of approx. 200 nm. Moreover, atomicforce microscopy reveals a much higher surface roughness (RMS) ofthe as-deposited poly-GeSi film (12.6 nm) compared to the LC GeSifilm (2.5 nm).

Grain orientation and residual stress.— A standard X-ray diffrac-tion (XRD) θ-2θ scan was performed on laser-crystallized and as-deposited poly-GeSi film. Both films exhibit the (111), (220) and

Figure 5. SEM top view of LC GeSi film without pre-patterned lines. Thelarge grains are formed without any confinement leading to length axis rotationsas high as 26◦. Reprinted with permission from B. Rangarajan, I. Brunets, P.Oesterlin, A. Y. Kovalgin and J. Schmitz, MRS Proceedings, Vol. 1321 (2011).

Figure 6. SEM top view of as-deposited poly-GeSi film showing small grainswith a typical size of 200 nm. Reprinted with permission from B. Rangarajan,I. Brunets, P. Oesterlin, A. Y. Kovalgin and J. Schmitz, MRS Proceedings,Vol. 1321 (2011).

(311) 2θ peaks. The XRD scan of a-GeSi film confirmed its amor-phous nature with no observable 2θ peaks.

In order to gain further insight into these materials, grazingincidence angle XRD (GAXRD) scans were also performed usingincidence angle ω ranging from 2.5◦ to 4◦. By GAXRD, it is possibleto detect the crystal planes oriented at different angles α (Fig. 7) withrespect to the sample surface, which are otherwise not detectableusing a standard θ-2θ scan. The low angle further limits the X-raypenetration into the substrate, so any angular dependence can be usedto resolve depth information.

Fig. 8 shows the GAXRD results for LC GeSi film and as-depositedpoly-GeSi film (under the same data acquisition conditions). One cansee that: (i) shifted (311) 2θ peaks have emerged adjacent to theexpected (311) 2θ peak position; (ii) and when the incidence angle ωmoves from 2.5◦ to 4◦, the shift increases toward higher 2θ angles,for all the films. This shift represents a reduced lattice spacing d,indicating a tensile stress along the lateral direction (i.e. along thefilm surface), and is observed for both the LC and as-deposited GeSifilms. Similarly, shifting of the 2θ peaks with respect to ω indicates agradient in stress as a function of depth.

For different incidence angles ranging from 2.5◦ to 4◦, the calcu-lated α (defined in Fig. 7) for all the individual shifted (311) 2θ peaks is25◦, for all the films. This indicates that the diffraction peaks originatefrom the lattice planes all oriented at 25◦ to the surface. The existenceof such-oriented (311) planes is further confirmed through EBSD. TheEBSD pole figure (see Fig. 9) shows the angular distribution of (311)lattice planes in the sample rolling direction (RD) i.e. with respect tothe film surface. Two main orientations can be seen: large number of(311) planes occur at an angle of 25o to the surface (first maximum)while the second maximum appears at around 65o. From the statisticalanalysis (not shown) we conclude that more than 60% of the grains

Figure 7. Definition of various angles used in GAXRD.

P266 ECS Journal of Solid State Science and Technology, 1 (6) P263-P268 (2012)

1

100

10000 = 2.5°Shift

Existing peak

1

100

10000 = 3°

Co

un

ts Shift

Existing peak

1

100

10000 = 3.5°Shift

Existing peak

1

100

10000

52 54 56 58 60

= 4°Shift

Existing peak

= 2.5°ShiftExisting

peak

= 3°ShiftExistingpeak

= 3.5°Shift

Existingpeak

52 54 56 58 60

= 4°Existing

peak

Figure 8. Shifted (311) 2θ peaks are shown in detail for LC GeSi film and as-deposited poly-GeSi film indicating a residual tensile stress. The shift increasesfor higher ω angles.

Figure 9. The pole figure, obtained using EBSD for (311) planes along thelateral growth (scanning) direction. It shows angular distributions with respectto the film surface.

begin their growth as (001) planes and finish as (311) planes. Thisindicates a preferential evolutionary selection toward (311) planes.

For the LC GeSi films, the shifted (311) 2θ peak with the highestintensity counts (17733) appears at ω = 3◦, and for the as-depositedpoly-GeSi film at ω= 2.5◦ (with a count of 933). This indicates that thelattice planes with the corresponding d have the highest prominence.Considering these two prominent peaks, the strain for both types ofGeSi film is calculated as shown in Table I. The LC GeSi film shows

Figure 10. Raman Spectra of as-deposited poly-GeSi and LC GeSi with pre-patterned lines. The left hand shift of Ge-Ge peak corroborates the resultsobtained from the XRD.

a strain (�d/d0) of 3.4%; for the as-deposited film the strain reaches2.7%. The former further exhibits a 19-times-more-intensive shifted(311) 2θ peak than the latter; apparently, a much higher number ofstrained (311) planes exists in the LC film compared to as-depositedpoly-GeSi.

In order to further verify the existence of strain, Raman measure-ments were done using a 30-mW, 532 nm laser beam at room tem-perature. For a relaxed Ge0.85Si0.15 film, the first-order Ge-Ge peakshould appear around 298 cm−1.29 It is considerably offset for bothlayers, as shown in Fig. 10. The left-hand shift compared to 298 cm−1

indicates that both layers have tensile stress, and highest in the LCGeSi film. This is in line with the XRD results. It is important to notethat no Ge-Si and Si-Si Raman peaks were observed in the spectrumdue to the low concentration of these bonds30 (expected, given thehigh germanium fraction).

The residual tensile stress in the layer can be attributed to thedifferent thermal expansion coefficients of the Ge0.85Si0.15 film (5.2× 10−6 ◦C−1 31) and the SiO2 layer beneath it (0.5 × 10−6 ◦C−1,both values at 300 K). During the laser treatment, the GeSi film ismelted. As it cools down and solidifies, its shrinking is obstructed bythe underlying oxide layer.

Fig. 11 shows the concentration of Ge and Si atoms along thethickness, as obtained by X-ray Photoelectron Spectroscopy (XPS).Both a-GeSi film prior to crystallization and LC GeSi film - arecompared. The atomic concentrations are constant along the thick-ness of the a-GeSi film. Yet, the upper layer of the LC GeSi film isgermanium-enriched. This segregation of Ge to the surface occurs inorder to lower the surface energy of the crystallized film.32 This isfurther enhanced by multiple melting due to the large pulse overlap.The vertical gradient in the material composition leads to a gradient inthermal expansion coefficient, and can also contribute to the residualstress in the LC GeSi films.

Optical properties.— Spectroscopic Ellipsometry (SE) was used todetermine the refractive index n, extinction coefficient k and the opticalbandgap of the GeSi films. The optical functions are highly relevantfor GeSi application in optoelectronic devices such as photodiodes.The Tauc–Lorentz (TL) model33 was used to parameterize these films.

The TL model was initially developed for parameterizing amor-phous materials and later extended to polycrystalline materials.34,35

For the a-GeSi film, a single TL oscillator was used representing a

Table I. Strain calculation for LC GeSi film and as-deposited poly-GeSi film. All XRD spectra were obtained under the same acquisition conditionsto facilitate comparisons.

(311) 2θ (311) 2θ (311) 2θ

normal shifted shifted peak d0 d �d/d0

Material Type ω peak peak intensity counts (Å) (Å) (%)

As-deposited Poly Ge0.85Si0.15 2.5 54.17◦ 55.81◦ 933 1.647 1.602 2.7LC Ge0.85Si0.15 3.0 54.35◦ 56.41◦ 17733 1.642 1.586 3.4

ECS Journal of Solid State Science and Technology, 1 (6) P263-P268 (2012) P267

Figure 11. XPS depth profile detailing the composition of a-GeSi films beforeand after the laser crystallization process. Ge enrichment can be observed atthe top layer of the LC film. Part of it is reprinted with permission fromB. Rangarajan, I. Brunets, P. Oesterlin, A. Y. Kovalgin and J. Schmitz, MRSProceedings, Vol. 1321 (2011).

single inter-band transition. A TL oscillator is described by four pa-rameters: the Tauc bandgap Eg, the peak transition energy E0, thebroadening term Br and the transition-matrix-element related param-eter Amp. The obtained values of the four parameters are shown inTable II.

Two TL oscillators (with Eg coupled, i.e., using the same Eg) wereused to describe the LC GeSi film and the as-deposited poly-GeSifilm. Each oscillator represents the inter-band transitions occurring ata specific peak transition energy level (E0), as shown in Table II. Thevalues of E0 for both materials can be related to those of pure Ge(4.4 eV and 2.3 eV) and Si (4.4 eV and 3.4 eV).36

The n and k for the GeSi films with respect to photon energyare shown in Figs. 12a and 12b. The first E0-peak for the LC GeSiappears at 2.4 eV, lower than 2.6 eV of the as-deposited poly-GeSifilm. This indicates a higher Ge fraction in the LC film. Given the highsensitivity of SE to the uppermost layers for light absorbing films, thisobservation is consistent with the Ge-enrichment observation by XPS(see Fig. 11).

Relaxed mono-crystalline Ge0.85Si0.15 has a bandgap of 0.84 eV.37

Using the TL model fit, the as-deposited poly-GeSi indeed ex-hibits a 0.83 eV optical bandgap. The a-GeSi film shows a widerbandgap of 0.97 eV. The bandgap of the LC GeSi film is as small as0.35 eV. To note, the accuracy of extracting this particular value canbe low because it falls outside the spectral range detectable by ourSE. Nevertheless, we can expect narrowing of bandgap caused by asignificant tensile strain.38,39 In contradiction, as-deposited poly-GeSiexhibits an extracted bandgap very close to the value of unstrainedmonocrystalline material. But one have to bear in mind that the XRDresults show a 19-times lower intensity of the shifted (311) 2θ peakfor the as-deposited poly-GeSi layer compared to the LC GeSi layer(see Table I). This point to a lower number of strained (311) planesand, therefore indicates to a smaller impact on the bandgap.

Band gap narrowing can widen the application range of GeSi inoptoelectronics. For example, one can use LC GeSi films to realizeinfrared photodetectors for wavelengths between 1 and 5 μm. Hencefurther analysis is required (such as photoluminescence spectroscopy)to confirm the extracted bandgap value.

Figure 12. The (a) refractive index n and (b) extinction coefficient k of thea-GeSi film, as-deposited poly-GeSi film and LC GeSi film. The inset showsa zoomed-in view for k indicating the absorption near the edge.

Activation of ion-implanted dopants.— As-deposited poly-Ge0.85Si0.15 films were implanted with three dopants: BF2

+, As+ andP+. This was followed by dopant activation. The activation of dopantswas inferred from the sheet resistance value of the film obtained us-ing a four-point probe. (In case the sheet resistance did not changeupon anneal, the activation was considered unsuccessful.) By multi-plying the sheet resistance with the film thickness (measured usingSE), the resistivity was obtained. Resistivity values below 20 m�-cm, indicating an active dopant concentration above 1019 cm−3 (formonocrystalline Ge, see40), were considered desirable.

Furnace annealing (FA) and rapid thermal annealing (RTA) wereperformed with various combinations of annealing time and tempera-ture. FA resulted in successful activation of boron dopants but neitherAs nor P got activated. RTA led to comparable results, see Table III.The n-type dopants (As and P) have high diffusivity in pure Ge, andare reported to out-diffuse.41 As a result, given sufficient thermal bud-get, these impurities will move out of the Ge. A similar problem canbe expected in poly-Ge0.85Si0.15 films, which can explain the lack ofAs and P activation by FA and RTA.

On the other hand, green laser annealing (LA) activated all threedopants (see Table III). The laser energy density per pulse was chosenonly to activate the dopants without melting the film. This successfulactivation of n-type dopants could be attributed to the sub-μs-longlaser pulses, limiting the diffusion of the dopants.

Table II. Tauc-Lorentz oscillator parameters used for a-GeSi film, as-deposited poly-GeSi film and LC GeSi film.

Material No. of TL Osc. Eg(eV) E0(eV) Br(eV) Amp(eV)

a-Ge0.85Si0.15 TL1 0.96 3.0 3.0 178.5LC Ge0.85Si0.15 TL1 0.34 2.4 0.3 13.2

TL2 Coupled to TL1 Eg 3.9 1.3 64.9As-deposited Poly-Ge0.85Si0.15 TL1 0.83 2.6 0.9 32.4

TL2 Coupled to TL1 Eg 3.9 1.5 70.6

P268 ECS Journal of Solid State Science and Technology, 1 (6) P263-P268 (2012)

Table III. A comparison of different dopant activation methods used for dopant implanted poly-Ge0.85Si0.15 films including a single-step methodinvolving crystallization of dopant implanted a-Ge0.85Si0.15 films. The thickness of all the films was 100 nm. The resistivity for all films wasdetermined from four-point probe sheet resistance measurements.

Resistivity (m�-cm)

Method Material Process Parameters Boron As P

FA poly-Ge0.85Si0.15 600–800◦C; 15–30 minutes 6.7–7.8 >200a >200a

RTA poly-Ge0.85Si0.15 700–900◦C; 5–60 seconds 5.0–5.5 >200a >200a

RTA poly-Ge0.85Si0.15 500◦C; 5 s to 5 mins 5.0–5.5 >200a >200a

LA poly-Ge0.85Si0.15 95% overlapping; 0.4 Jcm−2; FWHM: 40 μm 10 6.9 31.2LC with implanted dopants a-Ge0.85Si0.15 99.75% overlapping; 0.6 Jcm−2; FWHM: 40 μm 90.6 2.7 17.6

a The same resistivity was measured before and after dopant implantation and activation.

BF2+, As+ and P+ were also implanted in a plain a-Ge0.85Si0.15

film and subsequently laser crystallized using the green laser. Thisinvolved melting and re-crystallization of the implanted a-Ge0.85Si0.15

films. This allowed significant reduction in the resistivity values ofAs- and P-doped films (Table III). In opposite, the resistivity of B-doped GeSi has increased. The single-step treatment was however notoptimized for sequential super-lateral growth; the grain size was inthe range of 200 nm, i.e., similar to that of as-deposited poly-GeSi.

Conclusions

In this work, we investigated crystallization of a-Ge0.85Si0.15 filmsand dopant activation in amorphous and polycrystalline Ge0.85Si0.15

using a green laser, for the purpose of semiconductor device fabrica-tion at a low temperature. Green laser crystallization of a-Ge0.85Si0.15

films with pre-patterned lines resulted in the formation of large grainswith a grain length in the range of 10–25 μm (30% of the surfacearea) expanding to 35 μm (5%) and further up to 50 μm (2%) with atotal of 64% longer than 2 μm. The grains were confined between thepre-patterned lines (i.e. within a width of 3 μm) thereby facilitatingthe fabrication of semiconductor devices in a subsequent fabricationprocess. This would likely enhance the electrical device character-istics. Crystallization without the pre-patterned lines led to randompositioning of grains with a large angle of deviation in the crystal’slength axis.

Laser crystallization of a-Ge0.85Si0.15 films led to a strain of 3.4%due to the presence of residual tensile stress. Band gap narrowing wasobserved in the laser crystallized film, which can widen the applicationrange of GeSi in infrared photodetectors.

Green laser annealing was able to successfully activate both p-type(boron) and n-type (As and P) dopants in poly-Ge0.85Si0.15 films. Onthe other hand, both furnace annealing and rapid thermal annealingfailed in activating the n-type dopants. The one-step laser treatmentof n-type implanted a-Ge0.85Si0.15 films resulted in their reduced re-sistivity compared to the two-step laser treatment.

Acknowledgments

The authors gratefully acknowledge the support of the Smart MixProgram of the Netherlands Ministry of Economic Affairs and theNetherlands Ministry of Education, Culture and Science. The authorsthank Tom Aarnink (LPCVD), Buket Kaleli (Raman Spectroscopy),Lan Anh Tran (AFM), Gerard Kip (XPS) and Mark Smithers (SEM),all from MESA+ Institute for Nanotechnology, University of Twente,for their support.

References

1. A. Pruijmboom, C. E. Timmering, J. M. L. van Rooij-Mulder, D. J. Gravesteijn,W. B. de Boer, W. J. Kersten, J. W. Slotboom, C. J. Vriezema, and R. de Kruif, Mi-croelectronic Eng., 19, 427 (1992).

2. S. Datta, J. Brask, G. Dewey, M. Doczy, B. Doyle, B. Jin, J. Kavalieros, M. Metz,A. Majumdar, M. Radosavljevic, and R. Chau, IEEE BCTM 10.2, 194-197 (2004).

3. J. Liu, D. Pan, S. Jongthammanurak, K. Wada, L. C. Kimerling, and J. Michel, Opt.Express, 15(2), 623 (2007).

4. A. Splett, T. Zinke, K. Petermann, E. Kasper, H. Kibbel, H.-J. Herzog, and H. Presting,IEEE Photonics Technology Lett., 6(1), 59 (1994).

5. A. Witvrouw, L. Haspeslagh, O. V. Pedreira, J. De Coster, I. De Wolf,H. A. C. Tilmans, T. Bearda, B. Schlatmann, M. van Bommel, M.-C. de Nooijer,P. H. C. Magnee, E. J. Lous, M. Hagting, J. Lauria, R. Vanneer, andB. van Drieenhuizen, J. Microelctromechanical systems, 19(1), 202 (2010).

6. Y. K. Choi, T. J. King, and C. Hu, IEEE Trans. Electron Devices, 49(3), 436 (2002).7. G. Hock, E. Kohn, C. Rosenblad, H. von Kanel, H.-J. Herzog, and U. Konig, Appl.

Phys. Lett., 76(26), 3920 (2000).8. T. J. King, J. P. McVittie, K. C. Saraswat, and J. R. Pfiester, IEEE Trans. Electron

Devices, 41(2), 228 (1994).9. S. A. Bhave, B. L. Bircumshaw, W. Z. Low, Y.-S. Kim, A. P. Pisano, T. J. King,

and R. T. Howe, Solid-State Sens., Actuator and Microsystems Workshop, p. 34–37(2002).

10. E. Kasper, Properties of Strained and Relaxed Silicon Germanium, p. 123–129,Inspec (1995).

11. J. G. Fossum and A. Ortiz-Conde, IEEE Trans. Electron Devices, 30(8), 933 (1983).12. A. K. Ghosh, C. Fishman, and T. Feng, J. Appl. Phys., 51(1), 446 (1980).13. G. K Celler, J. Cryst. Growth, 63(3), 429 (1983).14. Paul Ch. van der Wilt, B. D. van Dijk, G. J. Bertens, R. Ishihara, and

C. I. M. Beenakker, Appl. Phys. Lett., 79(12), 1819 (2001).15. C.-H. Kim, I.-H. Song, W.-J. Nam, and M.-K. Han, IEEE Electron Device Lett.,

23(6), 315 (2002).16. L. Mariucci, A. Pecora, R. Carluccio, and G. Fortunato, Thin Solid Films, 383(1–2),

39 (2001).17. T.-K. Chang, C.-W. Lin, C.-C. Tsai, J.-H. Lu, B.-T. Chen, and H.-C. Cheng, Elec-

trochem. Solid-State Lett., 8(1), G14 (2005).18. T. Chen, R. Ishihara, M. R. T. Mofrad, Sten. Vollebregt, J. van der Cingel,

M. van der Zwan, H. Schellevis, and K. Beenakker, IEDM, p. 21.1.1–21.1.4 (2010).19. J. Schmitz, Nucl. Instrum. Methods Phys Res Sect. A, 576(1), 142 (2007).20. H. J. Kim and J. S. Im, Appl. Phys. Lett., 68(11), 1513 (1996).21. S. Sedky, M. Gromova, T. Van der Donck, J.-P. Celis, and A. Witvrouw, Sens. Actu-

ators, A, 127(2), 316 (2006).22. T. Sameshima, H. Watakabe, H. Kanno, T. Sadoh, and M. Miyao, Thin Solid Films,

487(1–2), 67 (2005).23. C. Eisele, M. Berger, M. Nerding, H. P. Strunk, C. E. Nebel, and M. Stutzmann, Thin

Solid Films, 427(1–2), 176 (2003).24. J. Holleman, A. E. T. Kuiper, and J. F. Verweij, J. Electrochem. Soc., 140(6), 1717

(1993).25. J. S. Im, H. J. Kim, and M. O. Thompson, Appt. Phys. Lett., 63(14), 1969 (1993).26. M. Lee, S. Moon, M. Hatano, K. Suzuki, and C. P. Grigoropoulos, J. Appl. Phys.,

88(9), 4994 (2000).27. R. S. Sposili and J. S. Im, Appl. Phys. Lett., 69(19), 2864 (1996).28. K. Murakami, O. Eryu, K. Takita, and K. Masuda, Phys. Rev. Lett., 59(19), 2203

(1987)29. H. H. Burke and I. P. Herman, Phys. Rev. B, 48(20), 15016 (1993).30. F. Pezzoli, Lucio Martinelli, E. Grilli, M. Guzzi, S. Sanguinetti, M. Bollani,

H. D. Chrastina, G. Isella, H. von Kanel, E. Wintersberger, J. Stangl, and G. Bauer,Mater. Sci. Eng. B, 124–125, 127 (2005).

31. M. E. Levinshtein, S. L. Rumyantsev, and M. S. Shur, Properties of Advanced Semi-conductor Materials GaN, AlN, InN, BN, SiC, SiGe, p. 149–188, John Wiley & Sons,New York (2001).

32. P. C. Weakliem and E. A. Carter, Phys. Rev. B, 45(23), 13458 (1992).33. H. Fujiwara, Spectroscopic ellipsometry: principles and applications, p. 170–172,

John Wiley & Sons, New York (2007).34. Z. G. Hu, J. H. Ma, Z. M. Huang, Y. N. Wu, G. S. Wang, and J. H. Chu, Appl. Phys.

Lett., 83(18), 3686 (2003).35. Y. J. Cho, N. V. Nguyen, C. A. Richter, J. R. Ehrstein, B. H. Lee, and J. C. Lee, Appl.

Phys. Lett., 80(7), 1249 (2002).36. J. R. Chelikowsky and M. L. Cohen, Phys. Rev. B, 14(2), 556 (1976).37. R. Braunstein, A. R. Moore, and F. Herman, Phys. Rev., 109(3), 695 (1958).38. M. E. Kurdi, G. Fishman, S. Sauvage, and P. Boucaud, J. Appl. Phys., 107(1), 013710

(2010).39. J. Michel, J. Liu, and L. C. Kimerling, Nat. Photonics, 4, 527 (2010).40. D. B. Cuttris, Bell System Technical Journal, 40(2), 509 (1961).41. A. Axmann, M. Schulz, and C. R. Fritzsche, Appl. Phys., 12(2), 173 (1977).