Room-temperature electroluminescence in the mid-infrared (2-3 μm) from bulk chromium-doped ZnSe
3
Room-temperature electroluminescence in the mid-infrared „2–3 m… from bulk chromium-doped ZnSe Julien Jaeck, Riad Haidar, Emmanuel Rosencher, Marcel Caes, and Michel Tauvy ONERA Chemin de la Hunière, 91761 Palaiseau Cedex, France Stéphane Collin, Natalie Bardou, Jean Luc Pelouard, and Fabrice Pardo CNRS/LPN, Route de Nozay, 91460 Marcoussis, France Philippe Lemasson CNRS/LPSC, 1 Place Aristide Briand, 92195 Meudon, France Received May 30, 2006; revised August 25, 2006; accepted August 25, 2006; posted September 13, 2006 (Doc. ID 71336); published November 9, 2006 Electroluminescence associated with impact excitation or ionization of deep Cr 2+ impurity centers in bulk ZnSe is reported. A broad signal of mid-infrared luminescence between 2 and 3 m is observed once the biased bulk ZnSe device runs into a nonlinear conduction regime. Optical powers in the nanowatt range have been measured at room temperature. The different mechanisms involved in this intracenter infrared light emission are discussed. © 2006 Optical Society of America OCIS codes: 230.2090, 260.3800, 260.3060, 250.0250. Luminescence can be generated by impact excitation or ionization of impurity centers in semiconductors. Visible electroluminescence was widely studied in Mn 2+ -doped ZnS as early as 1936 1,2 (Destriau effect). More recently Scarmacio et al. 3 studied the electrolu- minescence at 3.5 m from impact ionization of Fe 2+ in InP. 3 This experiment was performed at very low temperature 5K, mainly because of multiphonon coupling. 4 Terahertz electroluminescence has also been demonstrated in phosphorous- and boron-doped silicon wafers at 5 K. 5,6 Since the pioneering paper by DeLoach et al. 7 in 1996, Cr 2+ -doped ZnSe has emerged as an alternative material for infrared (IR) emission in the 2–3 m range. Thanks to numerous advantages [low phonon energy 250 cm -1 , large optical cross section, good thermal conductivity 8 ], this material is particularly fit for optically pumped laser action. 9,10 Room- temperature electroluminescence of polycrystalline Cr 2+ : ZnSe was recently demonstrated in the ohmic regime. 11 In this Letter we demonstrate room- temperature mid-IR electroluminescence 2–3 m of a single crystalline Cr 2+ :ZnSe device running in a nonlinear conduction regime, where impact excita- tion and ionization are very likely to prevail. ZnSe : Cr single crystals (thickness t =546 m) were prepared by solid-phase recrystallization of ZnSe polycrystalline materials, 12 followed by a Cr 2+ doping up to 1.94 10 19 cm -3 . These concentrations were de- duced from peak absorption measurements at the wavelength of 1.8 m, assuming an absorbance cross section of =0.9 10 -18 cm 2 . 7 Metal electrodes (20 nm Cr and 200 nm Au layers) were evaporated on both sides of the samples. The backside electrode covers the whole surface and is used as the ground contact, while the topside elec- trodes are processed into 100 m 8 mm pad by a liftoff technique (see Fig. 1). The metal lines are 100 m apart and a few tens of micrometers away from the sample edges to avoid arc effects. The samples are mounted on ceramics on which electrical contacts are provided by needle probes. Large biases (up to 1800 V in cw operation) can be applied to the sample. The emitted light is side collected through an uncoated f /1.25 ZnSe lens. The exit edge of the sample is imaged onto a HgCdTe camera (spectral bandwidth 0.8–2.6 m). Figure 2 shows the log plot of the current versus voltage I – V curve. Clearly, an ohmic regime is ob- served at low biases (up to 800 V), while a nonlinear regime starts for higher applied voltages. The thresh- old corresponds to an applied field of 1.5 10 4 V/cm, which seems too small to be explained by a band-to- band avalanche mechanism over the whole volume of the ZnSe crystal. 13–15 However, it is highly probable that the electric field in the ZnSe volume is not ho- Fig. 1. Structure of the device. December 1, 2006 / Vol. 31, No. 23 / OPTICS LETTERS 3501 0146-9592/06/233501-3/$15.00 © 2006 Optical Society of America
Transcript of Room-temperature electroluminescence in the mid-infrared (2-3 μm) from bulk chromium-doped ZnSe
December 1, 2006 / Vol. 31, No. 23 / OPTICS LETTERS 3501
Room-temperature electroluminescence in themid-infrared „2–3 �m… from bulk
chromium-doped ZnSe
Julien Jaeck, Riad Haidar, Emmanuel Rosencher, Marcel Caes, and Michel TauvyONERA Chemin de la Hunière, 91761 Palaiseau Cedex, France
Stéphane Collin, Natalie Bardou, Jean Luc Pelouard, and Fabrice PardoCNRS/LPN, Route de Nozay, 91460 Marcoussis, France
Philippe LemassonCNRS/LPSC, 1 Place Aristide Briand, 92195 Meudon, France
Received May 30, 2006; revised August 25, 2006; accepted August 25, 2006;posted September 13, 2006 (Doc. ID 71336); published November 9, 2006
Electroluminescence associated with impact excitation or ionization of deep Cr2+ impurity centers in bulkZnSe is reported. A broad signal of mid-infrared luminescence between 2 and 3 �m is observed once thebiased bulk ZnSe device runs into a nonlinear conduction regime. Optical powers in the nanowatt rangehave been measured at room temperature. The different mechanisms involved in this intracenter infraredlight emission are discussed. © 2006 Optical Society of America
OCIS codes: 230.2090, 260.3800, 260.3060, 250.0250.
Luminescence can be generated by impact excitationor ionization of impurity centers in semiconductors.Visible electroluminescence was widely studied inMn2+-doped ZnS as early as 19361,2 (Destriau effect).More recently Scarmacio et al.3 studied the electrolu-minescence at 3.5 �m from impact ionization of Fe2+
in InP.3 This experiment was performed at very lowtemperature �5 K�, mainly because of multiphononcoupling.4 Terahertz electroluminescence has alsobeen demonstrated in phosphorous- and boron-dopedsilicon wafers at 5 K.5,6
Since the pioneering paper by DeLoach et al.7 in1996, Cr2+-doped ZnSe has emerged as an alternativematerial for infrared (IR) emission in the 2–3 �mrange. Thanks to numerous advantages [low phononenergy �250 cm−1�, large optical cross section, goodthermal conductivity8], this material is particularlyfit for optically pumped laser action.9,10 Room-temperature electroluminescence of polycrystallineCr2+:ZnSe was recently demonstrated in the ohmicregime.11 In this Letter we demonstrate room-temperature mid-IR electroluminescence �2–3 �m� ofa single crystalline Cr2+:ZnSe device running in anonlinear conduction regime, where impact excita-tion and ionization are very likely to prevail.
ZnSe:Cr single crystals (thickness t=546 �m) wereprepared by solid-phase recrystallization of ZnSepolycrystalline materials,12 followed by a Cr2+ dopingup to 1.94�1019 cm−3. These concentrations were de-duced from peak absorption measurements at thewavelength of 1.8 �m, assuming an absorbance crosssection of �=0.9�10−18 cm2.7
Metal electrodes (20 nm Cr and 200 nm Au layers)were evaporated on both sides of the samples. The
backside electrode covers the whole surface and is0146-9592/06/233501-3/$15.00 ©
used as the ground contact, while the topside elec-trodes are processed into 100 �m�8 mm pad by aliftoff technique (see Fig. 1). The metal lines are100 �m apart and a few tens of micrometers awayfrom the sample edges to avoid arc effects. Thesamples are mounted on ceramics on which electricalcontacts are provided by needle probes. Large biases(up to 1800 V in cw operation) can be applied to thesample. The emitted light is side collected through anuncoated f /1.25 ZnSe lens. The exit edge of thesample is imaged onto a HgCdTe camera (spectralbandwidth 0.8–2.6 �m).
Figure 2 shows the log plot of the current versusvoltage �I–V� curve. Clearly, an ohmic regime is ob-served at low biases (up to 800 V), while a nonlinearregime starts for higher applied voltages. The thresh-old corresponds to an applied field of 1.5�104 V/cm,which seems too small to be explained by a band-to-band avalanche mechanism over the whole volume ofthe ZnSe crystal.13–15 However, it is highly probablethat the electric field in the ZnSe volume is not ho-
Fig. 1. Structure of the device.
2006 Optical Society of America
3502 OPTICS LETTERS / Vol. 31, No. 23 / December 1, 2006
mogeneous, owing to residual free carriers, and thatthe electric field is far larger at the ZnSe surface thanits average value. Finally, the samples are destroyedfor applied fields above 2.8�104 V/cm.
Figure 3 shows the luminance versus current �L–I�versus the V–I curves. As could be expected, the elec-troluminescence is enhanced in the nonlinear regime.Figure 3 shows also a visualization of the emission atthe crystal facet, with three electrodes being biased.Two points deserve attention: first, the emission isvisible only under the biased electrodes, and it isspread over the entire thickness of the sample. Thisconfirms the bulk nature of the emission mechanism.Second, no filamentation is observed in the ZnSesample.
In order to determine the spectral content of theemitted light, a circular variable filter is then in-serted between the sample and the camera (Fig. 4):the emitted light spectrum is centered on 2.4 �m andrather broad (extrapolated FWHM of 0.8 �m). TheCr2+ absorption spectrum is shown in this same fig-ure. The Stokes shift is clear evidence of the nonther-mal origin of the signal. The emission is thus unam-biguously related to the electrical excitation of the5E→ 5T2 transition of the Cr2+ ion in the ZnSe host.One can also notice that the electroluminescence
Fig. 2. Log plot of the I–V curve.
Fig. 3. V–I versus L–I curves for the Cr2+:ZnSe device.biased. The crystal facet is indicated: (a) below the threshold
(IR emission).peak at 2.4 �m is somewhat shifted from the usual2 �m position (Jahn–Teller effect16). We believe thatthis is an indication of strong reabsorption in theZnSe:Cr sample.
Concerning the electroluminescence quantumyield, the typical detected signal corresponds to an IRemitted flux in the nanowatt range, for 1 mW of ap-plied electrical power, leading to an electrical-to-optical quantum efficiency in the 10−6 range.
Clearly, IR emission rises with the current density.As explained above, the origin of this conductionmechanism is still unclear: it might be due toFowler–Nordheim injection at the Au–ZnSe inter-face, trap-controlled space-charge-limited current, ora band-to-band avalanche mechanism.1,17 This latterhypothesis is, however, very unlikely, since no visiblelight emission corresponding to the ZnSe bandgapcould be observed.
Concerning the mid-IR emission, it is noteworthythat transport mechanisms in heavily compensatedsemiconductors are very complex17 and depend criti-cally on various factors such as (i) the electron and/orhole injection capabilities of electrical contact, (ii) theexact charge state of the traps (trap-filled limitedcurrent), and (iii) the balance mechanisms betweenelectron and hole capture cross sections for perma-nent steady state.18 All of these data are largely un-known in Cr2+-doped ZnSe devices. Nevertheless,electron mobility being far larger than hole mobilityin ZnSe (effective mass of 0.21 for electrons versus0.6 for holes19), the main ionization mechanism prob-ably stems from impact by hot electrons in the ZnSeconduction band originating from the high electricfields.
Different mechanisms may then explain the excita-tion of the 5E→ 5T2 transition by the flowing elec-trons: Fig. 5(a), direct excitation of intracenter Cr2+
transitions or, 5(b), impact ionization of the Cr2+ elec-tron toward the conduction band followed by an in-tracenter recombination process. Experiments are inprogress to investigate these different hypotheses.
In conclusion, we have demonstrated room-temperature mid-IR �2–3 �m� electroluminescence
ht, visualization of the IR emission. Three electrodes arefacet is dark, and (b) above the threshold it appears bright
Rigthe
Cr center in ZnSe is displayed for comparison.
E→ T2 optical transition.
December 1, 2006 / Vol. 31, No. 23 / OPTICS LETTERS 3503
from bulk Cr2+:ZnSe crystals unambiguously relatedto Cr2+ intracenter emission. Further experimentsare in progress to enhance the quantum yield. Thisresult paves the way to electrically pumped laser ac-tion by impact excitation in bulk semiconductors,particularly in the mid-IR.
The authors are deeply indebted to Eric Bringuier,Sergey Mirov, and Irina Sorokina for fruitful discus-sions. E. Rosencher is also with the Département dePhysique, Ecole Polytechnique, 91128 Palaiseau,France. J. Jaeck’s e-mail address [email protected].
References
1. E. Bringuier, Philos. Mag. B 75, 209 (1997).2. G. Destriau, J. Chim. Phys. Phys.-Chim. Biol. 33, 587
(1936).3. G. Scamarcio, F. Capasso, A. L. Hutchinson, T. Tanbun-
Ek, D. Sivco, and A. Y. Cho, Appl. Phys. Lett. 68, 1374(1996).
4. K. Seeger, Semiconductor Physics: an Introduction(Springer-Verlag, 1973).
5. T. N. Adam, R. T. Troeger, S. K. Ray, P.-C. Lv, and J.Kolodzey, Appl. Phys. Lett. 83, 1713 (2003).
6. P.-C. Lv, R. T. Troeger, T. N. Adam, S. Kim, J.Kolodzey, I. N. Yassievich, M. A. Odnoblyudov, and M.S. Kagan, Appl. Phys. Lett. 85, 22 (2004).
7. L. D. DeLoach, R. H. Page, G. D. Wilke, S. A. Payne,and W. F. Krupke, IEEE J. Quantum Electron. 32, 885(1996).
8. I. T. Sorokina, Opt. Mater. 26, 395 (2004).9. I. T. Sorokina, E. Sorokin, A. Di Lieto, M. Tonelli, R. H.
Page, and K. I. Schaffers, J. Opt. Soc. Am. B 18, 926(2001).
10. S. B. Mirov, V. V. Fedorov, K. Graham, I. S. Moskalev,V. V. Badikov, and V. Panyutin, Opt. Lett. 27, 909(2002).
11. L. Luke, V. V. Fedorov, I. Moskalev, A. Gallian, and S.B. Mirov, in Proc. SPIE 6100, 61000Y (2006).
12. P. Lemasson, J. O. Ndap, S. Fusil, A. Rivière, B.Qu’hen, A. Lusson, G. Neu, E. Tournié, G. Geoffroy, A.Zozime, and R. Triboulet, Mater. Lett. 36, 162 (1998).
13. G. A. Baraff, Phys. Rev. 128, 2507 (1962).14. T. D. Thompson and J. W. Allen, J. Phys. C 20, L499
(1987).15. H. Ishikura, T. Abe, N. Fukuda, H. Kasada, and K.
Ando, Appl. Phys. Lett. 76, 1069 (2000).16. R. H. Page, K. I. Schaffers, L. D. DeLoach, G. D. Wilke,
F. D. Patel, J. B. Tassano, S. A. Payne, W. F. Krupke, K.T. Chen, and A. Burger, IEEE J. Quantum Electron.33, 609 (1997).
17. M. A. Lampert and P. Mark, Current Injection in Solids(Academic, 1970).
18. A. W. Livingstone and U. W. Allen, J. Phys. C 3, 2468(1970).
19. J. Singh, Physics of Semiconductors and Their
Fig. 4. Spectral analysis of the mid-IR emitted light by acontinuously variable filter. The absorption spectrum of the
2+
Fig. 5. Proposed mechanisms for Cr2+ mid-IR emission.Impact excitation process: impact due to a hot electron inthe conduction band promotes an electron from the funda-mental to the excited state of Cr2+ followed by a intracenter5E→ 5T optical transition (a). (b) Impact ionization process:the impact promotes an electron from the fundamentalstate of Cr2+ to the ZnSe conduction band, followed by itscapture in the Cr2+ excited state and by an intracenter5 5
Heterostructures (McGraw-Hill, 1993).
