Keywords: Semiconductor lasers, Vertical cavity surface emitting lasers.

Transcription

1 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI.9/JQE..9, IEEE Journal of nm VCSELs with P-type δ-doping in the Active Layers for Improved High-Speed and High-Temperature Performance Kai-Lun Chi, Dan-Hua Hsieh, Jia-Liang Yen, in-nan Chen, Jason (Jyehong) Chen, Hao-Chung Kuo Fellow IEEE, Ying-Jay Yang, and Jin-Wei Shi *, Senior Member IEEE Abstract In this paper, we study the influence of p-type modulation doping on the dynamic/static performance of highspeed nm VCSELs with highly strained multiple quantum wells (MQWs). The studied device structure has a / asymmetric cavity design, which can let the internal transit time of injected carriers be as short as that of / cavity design and further improve its performance in terms of speed and output power for high single-mode (SM) operation. Our proposed VCSEL structure with p-type doping shows superior modulation speed with an output power comparable to that of the un-doped reference device under room temperature operation. Furthermore, when the operating temperature reaches, there is significant improvement in both the modulation speed and maximum power of the p-doped structures. According to our simulation, this can be attributed to the change in the quasi-fermi levels of the injected carriers after the addition of p-doping in the active layers, which minimizes the electron leakage under high-temperature operation. T Keywords: Semiconductor lasers, Vertical cavity surface emitting lasers. I. INTRODUCTION he rapid growth of global internet traffic has lent a strong impetus to improving the required data rate. The target of the next generation optical interconnect framework is a total data rate of up to Gbit/sec []. It is highly desirable to increase the data rate pre-channel and reduce the number of lanes so as to reduce both the footprint and the power consumption of the G optical transreceiver module. Recently, the Optical Internetworking Forum (OIF) has targeted a data rate per channel as high as Gbit/sec (Common Electrical Interface (CEI)-G) in the next generation veryshort reach (VSR) G interconnect framework []. Due to huge propagation loss and dispersion of the electrical Gbit/sec signal transmitted on the printed circuit board, the optics transreceiver module, which is usually composed of a high-speed nm vertical-cavity surface-emitting laser (VCSEL) and photodiodes [-], must be placed as close as possible to the electronic integrated circuits (ICs). Kai-Lun Chi, in-nan Chen, and Jin-Wei Shi are with the Department of Electrical Engineering, National Central University, Taoyuan, Taiwan ( * : jwshi@ee.ncu.edu.tw). Dan-Hua Hsieh, Jason (Jyehong) Chen and Hao- Chung Kuo are with the Department of Photonics, National Chiao-Tung University, Hsinchu, Taiwan. Jia-Liang Yen is with the Department of Information Technology, Takming University of Science and Technology, Taipei, Taiwan. Ying-Jay Yang is with the Department of Electrical Engineering, National Taiwan University, Taipei,, Taiwan. The heat generated from ICs during high-speed operation, will lead to an increase in the junction temperature in the VCSEL, which usually degrades its high-speed performance. This makes the pre-emphasis driving-circuit of the VCSEL on the transmitter side as an essential component to compensate for the high-temperature induced degradation in the electrical-tooptical (E-O) bandwidth when the operation speed (> Gbit/sec) is close to the maximum E-O bandwidth (~ GHz) attainable by the VCSEL under room-temperature () operation [7]. However, the addition of pre-emphasis ICs would definitely increase the cost as well as power consumption in the linking channel. Furthermore, the enhancement in linking performance of a channel with embedded pre-emphasis ICs is still sensitive to the ambient temperature and frequency response of whole channel [7]. It is thus vital to have a VCSEL light source with almost invariant high-speed performance from to operation to produce the desired data rate of Gbit/sec per lane [-]. One of the most promising solutions to further enhance the modulation speed and minimize the hightemperature induced bandwidth degradation is the incorporation of highly strained multiple-quantum wells (MQWs) in the active layers in high-speed 9- nm VCSELs [,]. State-of-the-art high-speed performance has been reported for the 9 nm VCESL under operation []. However, it is not feasible to use such a large indium (In) mole fraction (> %) for the nm VCSEL, as for the highly strained MQWs in the 9 nm VCSEL due to the limitations of well thickness. The incorporating of p-type doping in the active layers, is another possible way to further improve the differential gain (modulation speed) and minimize the highcurrent (junction temperature) induced gain compression [- ]. Experimental observations show that the improvement in speed performance due to p-doping is more pronounced in unstrained MQWs than in strained ones [] and the improvement in dynamic performance of high-speed 9 nm VCSEL is marginal [7]. In this paper, we study the influence of the p-type modulation doping techniques on the dynamic/static performance of a high-speed nm VCSEL structure with highly strained MQWs and a / asymmetric cavity design. With this novel cavity design we can achieve an internal transit time for injected carriers as short as for the / cavity design and improve the performance in terms of speed and output power for single-mode (SM) operation [9,] (c) IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See for more information.

2 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI.9/JQE..9, IEEE Journal of The rest of this paper is organized as follows. First, we discuss the superior performance of our asymmetric cavity design over the traditional symmetric cavity VCSEL. Then we further investigate the influence of p-type doping on the dynamic/static performance of VCSELs with asymmetric cavities. Our measurement results show that the devices with p-type doping in the barrier layers exhibit superior speed performance to that of the undoped references. The observed advantages of p-type doping in VCSELs become more apparently under operation. According to our simulation, such improvement can be attributed to changes in the quasifermi levels of the injected carriers inside their active regions after the added p-type doping. Compared with our previous study [], this study provides a more detailed comparison of the dynamic performance of the p-doped and un-doped references with different sizes of oxide-relief apertures. The results demonstrate that this is a promising way to further improve the static/dynamic performance in the next generation of high-speed nm VCSELs for achieving extremely high data rates under high ambient temperature operation. II. DEVICE STRUCTURE Figures and show conceptual cross-sectional and top views of the studied device, respectively. With additional Zn-diffusion apertures in the top p-type DBR layers, we can not only manipulate the number of optical transverse modes inside the VCSEL cavity but also reduce the differential resistance [,9,9]. In addition, the oxide layer for current confinement is removed in our oxide-relief structure by using selective wet chemical etching to reduce its parasitic capacitance [,9,9]. The diameters of the Zn-diffusion (W Z) and oxide-relief apertures (W o) of the measured devices are specified in the figures below. The fabricated device has a ~ m diameter active mesa, which is integrated with the slot line pads for onwafer high-speed measurement, as shown in Figure. For details of the fabrication process please refer to our previous work [9,9]. Pad BCB SiO d W Z P contact Zn diffusion Oxide relief Active region N contact µm (Al.9Ga.As) oxidation layers are included in the epi-layer design to reduce the parasitic capacitance [,]. A short hole transit time is desired for high-speed direct modulation of the VCSEL. One effective approach to reducing the internal carrier transit time is to downscale the cavity length up to /, however, the strong confinement in a / cavity VCSEL may lead to degradation in the maximum saturation output power. Here, we demonstrate an asymmetric / cavity design that releases the trade-off between speed and output power of the VCSEL. Figures and show the simulated optical intensity distributions inside structures with cavity lengths of / and /. Details of the cavity structure for each case are also shown in the enlarged image in this figure. As can be seen, both the / and reference / cavity VCSEL share the same active layer design ( pairs of In.Al.Ga.7As/Al.Ga.As (/Å ) MQWs) without doping. Both structures have the same Fabry-Perot (FP) dip at nm. It can be seen that in the / cavity design, the active MQWs are closer to the p-side, which can greatly shorten the hole transit time, decrease the spatial hole burning (SHB) effect, and enhance its speed performance. For details of the / cavity VCSEL epi-layer structure please refer to our previous work [9]. The aperture sizes and depths of each studied device are specified in the figure captions. Optical Intensity (a.u.) Thickness Doping (cm - ) N-Al. Ga. As nm E7 U-In. Al. Ga.7 As (well) nm P-Al. Ga. As P-Al. Ga. As (Barrier) P-Al. Ga. As ~ nm ~ nm ~ nm..... Distance from Substrate ( m) ~E7 (or ) P-Al. Ga. As nm E7 N-side λ/ P-side N-side P-side Distance from Substrate ( m) Figure. Refractive index profile and simulated optical field intensity inside the asymmetric λ and λ VCSEL cavities. The zoom-in image shows details of the corresponding epi-layer structures and the doping profiles in the cavity layers. λ/ Thickness Doping (cm - ) N-Al. Ga.7 As nm E7 U-In. Al. Ga.7 As U-Al. Ga.7 As nm nm P-Al. Ga.7 As nm E7 Refractive Index Undoped layer µm Figure. Conceptual cross-sectional views and top-view of the demonstrated VCSELs. The epi-layer structure, purchased from LandMark, is grown on a semi-insulating GaAs substrate, which is composed of three In.Al.Ga.7As/Al.Ga.As (/Å ) MQWs sandwiched between a -pair n-type and 9-pair p-type Al.9Ga.As/Al.Ga.As Distributed-Bragg-Reflector (DBR) layers with an Al.9Ga.As layer ( nm thickness) above the MQWs for oxidation. Two more additional shallow III. MEASUREMENT RESULTS AND DISCUSSIONS Here, we discuss the influence of the cavity design on the internal carrier transit time. Devices with sizes of W Z, W o, and d are chosen as having high SM performance [9]. This is because high SM performance VCSELs usually suffer from a significant SHB effect and the hole injection/transit time can induce a serious low-frequency roll-off which becomes the major bandwidth limiting factor [9]. Figure shows the measured L-I curves of the SM VCSELs with the proposed / (devices A and B) and / devices C and D) cavity designs. Figures and show the corresponding measured bias dependent output optical spectra of devices A and B, respectively. LandMark Optoelectronics Corporation, No., Nanke 9th Rd., Shanhua Dist., Tainan City 7, Taiwan (c) IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See for more information.

3 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI.9/JQE..9, IEEE Journal of Power (mw) / / Device A Device B Device C Device D Figure. Measured L-I curves of demonstrated SM VCSELs with λ (devices A and B) and λ (devices C and D) cavity lengths. Both devices have undoped MQWs. (W z/w o/d=7/9/. m). Relative Intensity (a.u.) Device A / ma ma ma 7 Wavelength (nm) Wavelength (nm) Figure. Measured bias dependent output optical spectra of demonstrated SM VCSELs with λ (device A) and λ (device C) cavity lengths. Both devices have undoped MQWs. (W z/w o/d=7/9/. m). We can clearly see that with the same threshold current of ~ ma, both devices demonstrate a high SM performance across the whole range of bias currents. Furthermore, the / cavity VCSEL has a higher maximum SM (saturation) power. The superior L-I performance of the / cavity VCSELs over that of the / device might be attributable not only to the different cavity lengths (optical confinement) as discussed above, but also differences in the reflectivity of the top p-type DBR mirrors. During material growth, the targeted reflectivity for the top mirrors in both structures is the same value, ~99.7%. However, there could be some variation in film thickness between runs during VCSEL wafer growth arising from the use of different chambers which would lead to variation in the reflectivity of the DBR mirrors. Figure shows the measured bias dependent electrical-to-optical (E-O) frequency response of these two device structures. These traces were measured by a lightwave component analyzer (LCA), which was composed of a network analyzer (Anritsu 797C) and a calibrated GHz photoreceiver module (New focus -S). The shortening of the drift-distance of the holes in both cavity structures minimizes the SHB effect induced low-frequency roll-off in the O-E responses of these two demonstrated SM VCSELs (to less than db) [9]. Furthermore, the / design offers a wider - db E-O bandwidth ( vs. GHz) under a higher saturation bias current ( vs. ma) than does the / reference. From the dynamic/static measurement results, we can conclude that the demonstrated asymmetric / cavity VCSEL structure fabricated using a Zn-diffusion approach really does benefit the speed and output power performance under high SM operation. - Device C / ma ma Response (db) Thus, we adopt this epi-layer structure and then further study the influence of p-type barrier doping on the dynamic/static performance. - ma GHz GHz ma GHz ma GHz - / Device A / - Device C Figure. Measured bias dependent E-O frequency responses of demonstrated SM VCSELs with λ and λ cavity lengths. Both devices have undoped MQWs (W z/w o/d=7/9/. m). Two kinds of epi-wafers, with exactly the same structure as discussed above and / asymmetric cavity design, were grown. The only difference was that p-type doping was added in the barrier layers of one of the MQWs (structure I). The barrier layers in the reference structure (structure II) were left un-doped. The effective barrier height of electron can be increased with the p-type barrier doping level, which can suppress hot electron leakage and improve high-temperature performance of device []. However, if the p-type doping level goes too high, it would tremendously increase the free-carrier absorption loss in active layers and degrade laser performance. We thus had to well control the barrier doping by monitoring the intensity of photoluminescence (PL) spectrum of active layers during test runs of material growth. The p-type deltadoping density ( nm in thickness) we eventually used in the barrier layer ( nm in thickness) of our device is slightly over 7 cm -. A detailed p-type doping profile for our active layer design appears in Figure. Both MQW designs (structures I and II) exhibited peaks in the photoluminescence wavelength at around ~ nm (structure I (p-doping):. nm; structure II (undoped):.9 nm). The measured FP dips for structures I and II grown on the VCESL wafers differed slightly from the targeted FP dip wavelength ( nm), being around. and nm, respectively. The deviations in the central wavelengths and FP dip were mainly due to variations in the film thickness or the composition of the alloy during material growth. Figures shows the measured light output power (L) and bias voltage (V) versus current (L-V-I) characteristics observed at roomtemperature () and for structures I and II. Figure 7 shows similar measurement results for VCSELs with a smaller oxide-relief aperture (- vs. -7 m). In each figure, the traces are measured for two typical devices. In conflict with the devices shown in Figure, here we just chose the sizes of W z/w o for multi-mode performance of VCSELs under test in order to pursue their maximum modulation speed [9]. We can clearly see that in a comparison of Figures and 7 that the p- type doping devices exhibit a very close L-I-V performance compared to that of the reference sample without p-type doping, under operation. In contrast, when the ambient temperature reaches, the devices with p-type doping, with two different oxide-relief apertures (~ and ~ m) both show the superior output power performance to the undoped device. - ma GHz 9GHz ma GHz V -997 (c) IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See for more information.

4 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI.9/JQE..9, IEEE Journal of Power (mw) % % O C Voltage (V) Figure. Measured L-I and I-V curves of demonstrated devices under (solid symbols) and (open symbols) operation with (black symbols) and without (red symbols) p-type modulation doping in the active MQWs. W z/w o/d=9// m. Power (mw) % % :~ m O C Voltage (V) Figure 7. Measured L-I and I-V curves of demonstrated devices under (solid symbols) and (open symbols) operation with (black symbols) and without (red symbols) p-type modulation doping in the active MQWs. W z/w o/d=9// m. Table I shows details of the threshold currents (I th), threshold current density (J th), and slope efficiency (dl/di) of these devices under and operation. It can be seen that, compared with undoped references, the influence of p-type doping on the threshold current (density) is not very significant. According to previous studies, p-type doping in active layers without strain usually leads to an increase in the threshold current []. This is because the profiles of the conduction/valence (Ec/Ev) bands are not symmetric and the downward shift in the quasi-fermi levels after the incorporation of p-doping increases the hole density faster than it decreases the electron density (P increases approximately exponentially while N decreases approximately linearly). Thus, the NP product actually increases, resulting in a higher transparency current density []. On the other hand, as discussed above, there is significant compressive strain in our active layer, which would let both the profiles of E c/e v bands and position of electron/hole quasi-fermi levels (/) become more symmetric, thereby diminishing the increased threshold current density expected after p-doping []. TABLE I STATIC CHARACTERISTICS OF VCSELS UNDER AND OPERATIONS O C O C :~ m :~7 μm I th () I th ( O C) J th () J th ( O C) dl/di() (at ma) dl/di( O C) (at ma).7 ma.7 ma ka/cm ka/cm. W/A. W/A Undoped.7 ma. ma ka/cm 7 ka/cm.7 W/A. W/A :~μm I th () I th ( O C) J th () J th ( O C) dl/di() (at ma) dl/di( O C) (at ma). ma. ma ka/cm 7 ka/cm. W/A. W/A Undoped. ma.7 ma 9 KA/cm ka/cm.7 W/A. W/A In addition, compared with the undoped reference sample, there is a significant improvement in slope efficiency (dl/di) after the incorporation of p-type doping in the active layers under the same bias currents ( or ma) and under high-temperature (T) operation. This phenomenon can be attributed to the fact that the electron quasi-fermi level in the p-doping device is closer to the conduction band minimum in the active region than in the undoped devices, which thus minimizes electron leakage and degradation in output power under high-t operation. Figure shows the simulated band diagram of a single quantum well (QW) in the active region as obtained using the commercial simulation software: Photonic Integrated Circuit Simulator in D (PICSD). In our simulation, a uniform p-type doping profile across the whole barrier layer is assumed due to the possible out-diffusion of the p-type dopant during material growth. The detail numbers of material parameters used for our simulations can be referred to the previous work []. Based on these reported values, we can further determine the effective mass of electron/hole in our In.Al.Ga.7As well layer, which play important roles in the simulation results. The used values of effective mass are given in the caption of Figure. Different p-type doping concentrations of 7 cm - and 7 cm - are chosen for the comparison with the case of undoped barrier. Figure to (c) and (d) to (f) show the simulated band diagrams at two different ambient temperatures as and o C, respectively. The black line indicates conduction band (E c) and first heavy hole band (E HH). The red line indicates the first light hole band and the blue dash line indicates the quasi-fermi levels of electron () and hole (). As can be seen, when the p-type doping level increases from un-doped to 7 cm -, the energy separation between E c and reduces from () to () mev under ( o C) operation. This indicates that the p-doping can minimize the electron leakage and improve the high-temperature performance of device. Furthermore, the higher p-type doping QW exhibits a more symmetric distribution of quasi-fermi levels of electron/hole, which means an enhanced differential gain and modulation speed of device [-]. More enhancement of differential gain can be expected by further increasing the doping level to over cm - in barrier layer. However, this would lead to a serious degradation in PL intensity of active layers as discussed. We eventually chose a moderate doping level (> 7 cm - ) in our barrier layers as discussed.. un-doped o C.9.7 Figure. Simulated band diagram of In.Al.Ga.7As/Al.Ga.As QW with different barrier p-doping levels at ( to (c)) and o C ((d) to (f)). (a,d) undoped. (b,e) 7 cm -, and (c,f) 7 cm -. Electron /hole effective mass:./. m Crosslight Software Inc., - Lougheed Hwy, Vancouver, BC, VM A, Canada (c) IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See for more information Ec-Efn= - mev Ehh-Efp=. mev (d) un-doped o C Conduction band.9.7 Conduction band Ec-Efn= - mev Ehh-Efp=. mev. 7 cm - Conduction band o C.. Ec-Efn= - mev Ehh-Efp=.77 mev. (e) 7 cm - o C.7 E fn Conduction band Ec-Efn= - mev Ehh-Efp= 7.7 mev. (c) 7 cm - Conduction band o C..... Ec-Efn= - mev Ehh-Efp= 7. mev.9.7 (f) 7 cm - Conduction band. o C Ec-Efn= - mev Ehh-Efp= mev.9.7

5 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI.9/JQE..9, IEEE Journal of Response (db) Response (db) Figures 9 and show the measured bias dependent electricalto-optical (E-O) frequency responses (under and operation) of devices with and without p-type doping, respectively. We can clearly see that even with the same oxiderelief aperture, the p-doping devices show significantly faster speed performance than the undoped ones, under both and operation ma 9mA ma 9GHz GHz GHz 9GHz Figure 9. Measured bias dependent electrical-to-optical (E-O) frequency responses of demonstrated devices with p-type modulation doping in the active MQWs under (solid symbols) and (open symbols) operation. W z/w o/d=9// m GHz GHz GHz ma ma 9mA GHz Figure. Measured bias dependent electrical-to-optical (E-O) frequency responses of demonstrated devices with undoped active MQWs under (solid symbols) and (open symbols) operation. W z/w o/d=9// m. The solid and open symbols in Figure represent the summarized measured E-O bandwidth results for several p- doped and undoped devices, with three different oxide-relief diameters (~, ~, and ~7 m), under room-temperature () and operation, respectively. As can be seen, there is some variation in the measured -db bandwidths with each aperture size, which can be attributed to variations in the process. These include uncertainty of the surface state during the p- and n-type ohmic contact metallization and uniformity during the wet oxidation processes. Nevertheless, there is a gradual improvement in the bandwidths of most devices with a downscaling of the size of the oxide-relief apertures. Furthermore, regardless of the aperture size, the p-type doped devices not only exhibit a faster modulation speed under operation but also show less degradation in the O-E bandwidth when the ambient temperature reaches (~ vs. ~%) than those of the undoped ones GHz GHz ma ma O C GHz - - O C - Undoped - Undoped % 9% O C O C % 7 Oxide Aperture ( m) - - GHz 7% 9GHz Undoped 9GHz % ma ma O C O C 9% 7 9 Oxide Aperture ( m) Figure. Measured -db E-O bandwidths of VCSELs with and without p-type modulation doping of the active layers for different oxide-relief apertures under (solid symbols) and (open symbols) operation. The W z/d of all the measured devices is fixed at 9/ m. Both dynamic and static measurement results show that p- doping of the strained active layers of nm VCESLs can effectively improve their modulation speed and output power over that of the undoped reference samples, and this improvement becomes more apparent when the ambient temperature reaches. The improvement in speed performance is mainly because the electron quasi-fermi level is closer to the conduction band minimum after p-type doping, as discussed in Figure. This would thus lead to a more symmetric distribution of /, an increase in the differential gain, and enhancement in modulation speed of the VCSELs [-]. IV. CONCLUSION In this work, we demonstrated an asymmetric / VCSEL cavity structure and compared it with a symmetric / cavity structure. Although such a novel design has a longer cavity length (/ vs. / it can sustain a short hole transit time and exhibit a faster modulation speed and higher saturation power for SM operation. We further investigated the influence of p-type doping in the active layer of this VCSEL structure on the dynamic and static performance of the device. Compared to an undoped reference, the p-doped device had the same threshold current but demonstrated superior speed and output power performance, which became more apparent under operation. According to our simulation, the improved performance of the p-doping device is due to the change in the electron quasi-fermi level in the active region. Overall, the p- type doping technique proposed here should play an important role in the next generation OI system, where high density package induced device-heating and bandwidth degradation is a major issue for > Gbit/sec operation of VCSELs. Acknowledgement: This work was sponsored by the Ministry of Science and Technology in Taiwan under grants MOST --E--- and MOST --E-9- -CC. References [] H. Isono, H. Sakamoto, Y. Miyaki, T. Tanaka, T. Takahara, B. Martin, DMT relative cost consideration, IEEE P.bs Gb/s Ethernet Task Force, Norfolk, VA, USA, May. Available: [] Presentation.pdf [] P. Moser, P. Wolf, G. Larisch, H. Li, J. A. Lott, and D. Bimberg, Energyefficient oxide-confined high-speed VCSELs for optical interconnects, Proc. SPIE, Vertical-Cavity Surface Emitting Lasers VIII, vol. 9, pp. 9, Feb.,. [] E. Haglund, P. Westbergh, J.S. Gustavsson, E.P. Haglund, A. Larsson, M. Geen, and A. Joe, GHz bandwidth nm VCSEL with sub- fj/bit energy dissipation at - Gbit/s, Electron. Lett., vol., no., pp. 9-9, July,. [] D. M. Kuchta, A. V. Rylyakov, F. E. Doany, C. L. Schow, J. E. Proesel, C. W. Baks, P. Westbergh, J.S. Gustavsson, and A. Larsson, A 7-Gb/s NRZ Modulated -nm VCSEL-Based Optical Link, IEEE Photon. Technol. Lett., vol. 7, pp.77-, March,. [] Kai-Lun Chi, Yi-uan Shi, in-nan Chen, Jason (Jyehong) Chen, Ying- Jay Yang, J.-R Kropp, N. Ledentsov Jr., M. Agustin, N.N. Ledentsov, G. Stepniak, J. P. Turkiewicz, and Jin-Wei Shi, Single-Mode nm -997 (c) IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See for more information.

6 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI.9/JQE..9, IEEE Journal of VCSELs for Gbit/sec On-Off Keying Transmission Over km Multi- Mode Fiber, IEEE Photon. Technol. Lett., vol., no., pp. 7-7, June,. [7] D. M. Kuchta, A. V. Rylyakov, C. L. Schow, J. E. Proesel, C. W. Baks, P. Westbergh, J.S. Gustavsson, and A. Larsson, A Gb/s NRZ Modulated nm VCSEL Transmitter Operating Error Free to 9 C, IEEE/OSA Journal of Lightwave Technology, vol., no., pp. -, Feb.,. [] H. Li, P. Wolf, P. Moser, G. Larisch, A. Mutig, J. A. Lott, and D. H. Bimberg, Impact of the Quantum Well Gain-to-Cavity Etalon Wavelength Offset on the High Temperature Performance of High Bit Rate 9-nm VCSELs, IEEE J. Quantum, Electron., vol., pp. -, Aug.,. [9] Kai-Lun Chi, Jia-Liang Yen, Jhih-Min Wun, Jia-Wei Jiang, I-Cheng Lu, Jason (Jyehong) Chen, Ying-Jay Yang, and Jin-Wei Shi, Strong Wavelength Detuning of nm Vertical-Cavity Surface-Emitting Lasers for High-Speed (> Gbit/sec) and Low-Energy Consumption Operation, IEEE J. of Sel. Topics in, vol., no., pp. 7, Nov.,/Dec.,. [] M. Liu, C. Y. Wang, M. Feng and N. Holonyak, Jr., nm Oxide- Confined VCSELs with Gb/s Error-Free Transmission Operating up to C, in Technical Digest of Conference on Lasers and Electro-Optics, paper SFL., San Jose, CA, USA, May. [] H. Hatakeyama, T. Anan, T. Akagawa, K. Fukatsu, N. Suzuki, K. Tokutome, and M. Tsuji, Highly Reliable High-Speed.- m-range VCSELs With InGaAs/GaAsP-MQWs, IEEE J. Quantum Electron., vol., pp. 9-97, June,. [] L. A. Coldren and S. W. Corzine, Diode Lasers and Photonic Integrated Circuits, chapter, John Wiley & Sons, New York, 99. [] K. Uomi, Modulation-doped multi-quantum well (MD-MQW) lasers. I. Theory, Jpn. J. Appl. Phys., vol. 9, pp. 7, Jan. 99. [] K. Uomi, T. Mishima, and N. Chinone, Modulation-doped multiquantum well (MD-MQW) lasers. II. Experiment, Jpn. J. Appl. Phys., vol. 9, pp. 9, Jan. 99. [] N. Hatori, A. Mizutani, N. Nishiyama, A. Matsutani, T. Sakaguchi, and F. Motomura, F. Koyama, and K. Iga, An Over -Gb/s Transmission Experiment Using a p-type Delta-Doped InGaAs GaAs Quantum-Well Vertical-Cavity Surface-Emitting Laser, IEEE Photon. Technol. Lett., vol., no., pp. 9-9, Feb., 99. [] A. Schönfelder, S. Weisser, I. Esquivias, J.D. Ralston, J. Rosenzweig, Theoretical investigation of gain enhancements in strained In.Ga.As/GaAs MQW lasers via p-doping, IEEE Photon. Technol. Lett., vol., pp.7-77, April, 99. [7] Y. Zheng, C.-H. Lin, A. V. Barve, and L. A. Coldren, P-type δ-doping of highly-strained VCSELs for Gbps operation, IEEE Photonic Society Meeting, Burlingame, CA, USA, Sep.,, pp. -. [] Kai-Lun Chi, in-nan Chen, Jia-Liang Yen, Wei Lin, Shi-Wei Chiu, Jason (Jyehong) Chen, Hao-Chung Kuo, Ying-Jay Yang, and Jin-Wei Shi, Strong Enhancements in Static/Dynamic Performances of High-Speed nm Vertical-Cavity Surface-Emitting Lasers with P-type δ-doping in Highly Strained Active Layers, Proc. OFC, Anaheim, CA, USA, March,, pp. TuD.. [9] Jin-Wei Shi, Zhi-Rui Wei, Kai-Lun Chi, Jia-Wei Jiang, Jhih-Min Wun, I- Cheng Lu, Jason (Jyehong) Chen, and Ying-Jay Yang, Single-Mode, High-Speed, and High-Power Vertical-Cavity Surface-Emitting Lasers at nm for Short to Medium Reach ( km) Optical Interconnects, IEEE/OSA Journal of Lightwave Technology, vol., pp. 7-, Dec.,. [] I. Vurgaftman, J. R. Meyer, and L. R. Ram-Mohan, Band parameters for III V compound semiconductors and their alloys, Journal of Applied Physics vol. 9, pp. -7, no., June,. Kai-Lun Chi was born in New Taipei City, Taiwan on Feb.,, 9. He is now working on the Ph.D. degree in the Department of Electrical Engineering at National Central University, Taoyuan, Taiwan. His current research interests include high-speed optoelectronic device measurement and high-speed VCSELs and LEDs for optical interconnect applications. Dan Hua Hsieh received his B. S. degree in optoelectronics from Yuan-Ze University, Taiwan in. He is currently pursuing Ph. D. degree in Department of Photonics, National Chiao Tung University, Hsinchu, Taiwan. His research focus on the design, fabrication, and characterization of high-speed nm vertical cavity surface emitting lasers (VCSELs) for short reach and high speed communication networks Jia-Liang Yen was born in Kaohsiung, Taiwan on September, 99. He received his M. S. degree and Ph.D. degree in Electrical Engineering from National Taiwan University in 99 and, respectively. He is currently an assistant professor at Takming University of Science and Technology (TMUST) in the Department of Information Technology, Taiwan. His research interests include cloud computing and artificial intelligence, especially the power control of cloud system. in-nan Chen received his Master degree in Electrical Engineering from National Central University, Taiwan in. His research focus on the fabrication and characterization of high-speed nm vertical cavity surface emitting lasers (VCSELs) for short reach and high speed communication networks He is currently working in LandMark company, Tainan, Taiwan. Jason (Jyehong) Chen received his BS and MS degree in Electrical Engineering from National Taiwan University, Taiwan, in 9 and 99 respectively and the Ph.D. degree in Electrical Engineering and Computer Science from University of Maryland Baltimore County, Maryland, USA, in 99. He joined JDSU in 99 as senior engineer and obtained U.S. patents in years. He joined the faculty of National Chiao-Tung University, Taiwan,, where he is currently a professor in the Institute of Electro- Optical engineering and department of photonics. Prof. Chen published more than papers on international journals and conferences. His research interests focus on hybrid access network, long reach passive optical network and optical interconnects. Hao-Chung Kuo received the B.S. degree in physics from National Taiwan University, Taipei, Taiwan, the M.S. degree in electrical and computer engineering from Rutgers University, New Brunswick, NJ, USA, in 99, and the Ph.D. degree from the University of Illinois at Urbana Champaign, Urbana, IL, USA, in 999. He has an extensive professional career both in research and industrial research institutions that includes: Research Assistant at Lucent Technologies, Bell Laboratories, from 99 to 99; and a Senior R&D Engineer in Fiber-Optics Division at Agilent Technologies from 999 to and at LuxNet Corporation from to. Since October, he has been with the National Chiao Tung University (NCTU) -997 (c) IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See for more information.

7 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI.9/JQE..9, IEEE Journal of as a Faculty Member of the Institute of Electro-Optical Engineering. He is currently the Associate Dean, Office of International Affair, NCTU. He has authored and coauthored international journal papers, two invited book chapter, six granted and pending patents. His current research interests include semiconductor lasers, VCSELs, blue and UV LED lasers, quantum-confined optoelectronic structures, optoelectronic materials, and solar cell. He is an Associate Editor of the IEEE/OSA JOURNAL OF LIGHTWAVE TECHNOLOGY and a Guest Editor of the IEEE JOURNAL OF SELECTED TOPICS IN QUANTUM ELECTRONICS issue on Solid-State Lighting in 9. He received the Ta-You Wu Young Scholar Award from the National Science Council Taiwan in 7 and the Young Photonics Researcher Award from OSA/SPIE Taipei chapter in 7. He was elected as an OSA Fellow and the SPIE Fellow in. Ying-Jay Yang was born in I-Lan, Taiwan, in 9. He received the B.S. degree in electrical engineering from National Taiwan University in 97, the M.S. degree and the Ph.D. degree in electrical engineering from North Carolina State University, in 9 and 97 respectively. During his Ph.D. work he invented the first quantum well Transverse Junction Stripe (TJS) lasers and also the first CW operation strained-layer TJS lasers. From 97 to 99 he was an engineer at Hewlett Packard, working on the development of. um InGaAsP LEDs for FDDI. From 99 to 99 he joined Lockheed Palo Alto Research Laboratory as a research scientist. He worked on the vertical-cavity surface emitting lasers (SELs), invented the first single transverse mode SELs and the first optoelectronic integration circuits (OEICs) with a SEL and a FET. Since February 99 he joined the Department of Electrical Engineer, National Taiwan University, where he is now an professor. His current research areas are semiconductor materials, and devices including lasers, modulators, quantum devices, and OEICs. Jin-Wei Shi (M SM ) was born in Kaohsiung, Taiwan on January, 97. In, he joined the Department of Electrical Engineering, National Central University, Taoyuan, Taiwan, where he is a professor from. In - and, he joined the ECE Dept. of UCSB as a Visiting Professor. His current research interests include ultrahigh speed/power photodetectors, electro-absorption modulator, THz photonic transmitter, and VCSELs. He has authored or coauthored more than book chapters, Journal papers, conference papers and hold patents. He was the recipient of Da-You Wu Memorial Award (c) IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See for more information.