Oxidation process was regular process for VCSEL fabrication. Most high speed VCSEL was also made by the same method [7-8]. But only oxide-confined structure could be not enough to make device approaching high speed [9].
Parasitic capacitance was crucial parameter for speed limitation. That was, the region between bonding pad and oxide-layer formed considerable capacitance.
There were many efforts to reduce the capacitance reported based on reduction of pad area [10], mesa implantation [11], and BCB process [12-13]. Mesa implantation was commercial semiconductor process that fabrication was easier than BCB process. For this, we chose oxide-only and oxide-implant VCSELs to compare with their high speed characteristics in first section. In second section, we compared with different oxide aperture size VCSELs. Oxide aperture diameter resolved transverses mode which affected modulation speed. We still found out if NFP and spectrum width was concerned with modulation limitation.
Static and high speed characteristics of these VCSELs were still investigated.
Equivalent circuit modeling was investigated in section 4-3. With VCSEL modeling, the equivalent components were extracted and observe the modulation limitation caused by structure.
4-1 Comparison oxide-confined and oxide-implant VCSELs
Oxide-confined VCSEL and oxide-implant VCSEL had same process structure, shown in Figure 4-1. Process structure was described in Chapter 3.
Both VCSEL had the same oxide aperture as 8 µm and mesa as 22 µm.
Oxide-implant VCSEL had an isolation region as implant region was kept away from mesa to prevent damage from destroying the active region and hence voided triggering reliability issues.
The static characteristics of both VCSEL were shown in Figure 4-2.
Threshold current of oxide-only and oxide-implant VCSEL was 1.4 and 2.3 mA respectively. It meant that implant process provided additional current confinement made oxide-implant VCSEL having lower threshold current. Series resistance of both was similar as 91 and 108 Ohm. Slope efficiency, defined as conversion from injection current to output power, was 0.484 and 0.337 W/A.
Power conversion efficiency, indicated in Figure 4-2, appeared that oxide-implant VCSEL had higher light output up to 18%. That was, VCSEL maintained less power dissipation avoided parasitic heating which affected output power.
The measured modulation response curves for oxide-only and oxide-implant VCSELs were shown in Figure 4-3(a) and 4-3(b), respectively. The maximum modulation bandwidth of both VCSEL was 2.3 and 8.6 GHz. The roll-off as the lower frequency part of modulation response curve of oxide-only VCSEL was serious caused by parasitic effect. The resonance frequency increased with the bias current as expected. The plots in Figure 4-4 shown that,
for both different process VCSEL, proportional value of resonance frequency versus (I-Ith)1/2 as D-factor was 2.75 and 2.49 GHz/mA1/2. Modulation current efficiency factor (MCEF), indicated in Figure 4-4, appeared modulation bandwidth as function of (I-Ith)1/2. Oxide-implant VCSEL had 3.06 GHz/mA1/2; furthermore, bandwidth of oxide-only VCSEL was limited by parasitic effect for high current injection. Compared with both modulation responses, of oxide-only VCSEL with large capacitance, which introduced by the thin oxide layer, resulted in parasitic effect. On the other hand, the additional implantation in oxide-implant VCSEL reduced the capacitance of device without increasing the resistance much.
For commercial high speed performance testing, oxide and oxide-Implant VCSEL were measured as eye diagram shown in Figure 4-5. Eye diagram of oxide-only VCSEL shown that jitter was about 30 ps and fall-time tail touched 10Gbps mask. That was, oxide-only VCSEL could not apply to 10Gbps communication. Oxide-implant VCSEL had small jitter below 20 ps. It still had rise-time/fall-time were 44 and 54 ps respectively. For that, oxide-implant VCSEL passed 10Gbps clarified.
4-2 Comparison oxide-implant VCSELs in different oxide aperture sizes
There were three oxide aperture sizes VCSEL investigated as their high speed performance. Oxide aperture was fabricated as 6, 7, and 8 µm. The static characteristics of these VCSEL were shown in Figure 4-6. Threshold current were 0.8, 0.8, and 1.4 mA respectively. Series resistance of VCSEL was 91, 149, and 174 Ohm. It demonstrated that small device had large resistance than large
oxide aperture ones. Slope efficiency was 0.347, 0.387, and 0.484 W/A.
Small signal modulation response of different oxide aperture size of VCSEL was shown in Figure 4-7 as 6, 7, and 8 µm respectively. At lower frequency range, modulation bandwidth of small oxide aperture VCSEL was roll-off and large aperture VCSEL was flat. The difference was attributed to the additional parasitic effect. In Figure 4-8, D-factor of three VCSEL was 2.75, 4.89, and 5.14 GHz/mA1/2. Modulation current efficiency factor (MCEF), indicated in Figure 4-8, was 3.06, 5.54, and 5.97 GHz/mA1/2 respectively. We observed that small oxide aperture VCSEL had better D-factor but worst MCEF at high current injection.
The limitation of modulation response was possible induced by parasitic effect and heating. For the three oxide aperture size VCSEL, we could find out the optimum oxide aperture size was 7µm for high speed operation.
4-3 Equivalent circuit design and modeling
We established an equivalent circuit, depended on VCSEL structure, for investigate limitation caused by parasitic effect. The opposite equivalent circuit components were defined as Figure 4-9. Oxide-only and oxide-implant VCSEL were compared their VCSEL structure using Agilent ADS software to simulating and obtained results as S-parameter, introduced in Chapter 2.
The simulated and measured data showed better fitting as Figure 4-10. The equivalent circuit components could be extracted with reflection coefficient measured by network analyzer. Reflection coefficient, S11, was also determined impedance of circuit. Extraction components value was listed as Table 4-1 and
4-2 for oxide-only and oxide-implant VCSEL respectively. For this, we found that parasitic capacitance was effectively reduced due to implantation from 2 pF down to 0.36 pF, shown in Figure 4-10, but hindered current flow and increased parasitic resistance simultaneously. Even though, the trade-off had to be mode to minimize the RC product affected modulation bandwidth. For this, oxide-implant had better modulation bandwidth than oxide-only VCSEL was demonstrated.
Figure 4-1 Oxide-only and oxide-implant VCSEL structure
Substrate n-DBR p-DBR
active layer
n-contact p-contact
Implant region Oxide layer
0 4 8 12 16 20 0
1 2 3 4
Current (mA)
Voltage (V)
0 1 2 3 4 5
Output power (mW)
0 4 8 12 16 20
0 2 4
Current (mA)
Voltage (V)
0 1 2 3 4
Output power (mW)
Figure 4-2 Typical LIV curve for oxide-only and oxide-implant VCSELs (b) Oxide-implant VCSEL
(a) Oxide-only VCSEL
Figure 4-3 Modulation responses for both oxide-only and oxide-implant VCSELs
(b) Oxide-implant VCSEL (a) Oxide-only VCSEL
0 2 4 6 8 10 12 14 16 18
Modulation response (dB)
Frequency (GHz)
Modulation response (dB)
Frequency (GHz)
Figure 4-4 Resonance and 3dB frequency as a function of (I-Ith)1/2 for oxide-only and oxide-implant VCSELs
0.6 1.2 1.8 2.4 3.0
2 3 4 5 6 7 8 9
Frequency (GHz)
(I-Ith)1/2 (mA1/2) fr oxide-implant
f3dB oxide-implant fr oxide f3dB oxide
Figure 4-5 Eye-diagram for oxide-only and oxide-implant VCSELs (a) Oxide-only VCSEL
Ox O xi id de e -o - on nl ly y
(b) Oxide-implant VCSEL
Ox O xi id de e -i - im m pl p la an nt t
Figure 4-6 Typical LIV curve for VCSELs with different oxide aperture size
Figure 4-7 Modulation response of VCSELs with difference oxide aperture
Modulation response (dB)
Frequency (GHz)
Modulation response (dB)
Frequency (Hz)
Aox= 8 µm Aox= 7 µm Aox= 6 µm
Figure 4-8 Resonance and 3dB frequency as a function of (I-Ith)1/2 for difference oxide aperture VCSELs
0.5 1.0 1.5 2.0 2.5 3.0
3 4 5 6 7 8 9 10 11 12 13
fr: Aox=8um f3dB
fr: Aox=7um f3dB
fr: Aox=6um f3dB
Frequency (GHz)
(I-Ith)-1 (mA-1)
Figure 4-9 Real and imaginary S11 parameter versus frequency from model and measured data (blue line is measured data, red line is simulated data)
1E10
real(S(1,1))real(S(3,3)) imag(S(1,1)) imag(S(3,3))
Frequency (Hz)
real(S(1,1))real(S(3,3)) imag(S(1,1)) imag(S(3,3))
Frequency (Hz)
Figure 4-10 Compared capacitances of oxide-only and oxide-implant VCSELs
0.2 0.4 0.6 0.8 1.0
0.3 0.6 0.9 1.2 1.5 1.8 2.1
Ca,oxide-implant
Cp,oxide-implant
Ca,oxide-only
Cp,oxide-only
Capacitance (pF)
(Bias current)-1 (mA-1) Implantation
process
Table 4-1 Equivalent circuit elements at different bias current for oxide-only VCSEL
Table 4-2 Equivalent circuit elements at different bias current for oxide-implant VCSEL
Chapter 5 Conclusion
We have performed experimental study of small signal modulation behavior of 850nm oxide-only and oxide-implant VCSELs. The simple planar VCSEL structure is fabricated with difference oxide aperture diameters.
We have found that implantation process gives VCSEL with excellent high speed performance without semi-isolating substrate. Modulation bandwidth of oxide-only VCSEL is limited caused to parasitic capacitance demonstrated by equivalent circuit modeling. The equivalent circuit model is established benefit for modifying VCSEL structure epitaxy and process. Otherwise, we search the best oxide aperture size of oxide-implant VCSEL for high speed operation as 7 µm.
Reference
[1] H. Soda, K. Iga, C. Kitahara, and Y. Suematsu, “GaInAsP/InP surface emitting injection lasers,” Japanese Journal of Applied Physics, vol.18, pp.2329-2330, 1979.
[2] K. Iga, F. Koyama, and S. Kinoshita, “Surface-emitting semiconductor lasers,”
IEEE Journal of Quantum Electronics, vol.24, pp.1845-1855, September 1988.
[3] H. E. Li, and K. Iga, “Vertical cavity surface emitting laser,” Springer, Berlin, New York, 2003.
[4] T. Honda, T. Shirasawa, N. Mochida, A. Inoue, A. Matsutani, T. Sakaguchi, F. Koyama, and K. Iga, "Design and fabrication processes consideration of GaN-based surface emitting lasers," Trans. IEICE, J81-C-II, pp.97-104, 1998.
[5] “2003 年光電半導體產業及技術動態調查報告,” PIDA, 2004.
[6] R. Brand, “10 gigabit Ethernet interconnection with wide area networks,”
10GEA, March 2003
[7] K. L. Lear, A. Mar, K. D. Choquette, S. P. Kilcoyne, R. P. Schneider jr., and K.
M. Geib, “High-frequency modulation of oxide-confined vertical cavity surface emitting lasers,” Electron. Letter, pp.457-458, February 1996.
[8] C. Carlsson, H. Martinsson, R. Schatz, J. Halonen, and A Larsson, “Analog modulation properties of oxide confined VCSELs at microwave frequencies,”
IEEE Journal of Lightwave Technology, vol.20, pp.1740-1749, September 2002.
[9] C. H. Chang, L. Chrostowski, and J. Chang-Hasnain, “Parasitics and design considerations on oxide-implant VCSELs,” IEEE Photonics Technology
Letters, vol.13, December 2001.
[10] A. K. Dutta Dutta, H. Kosaka, K. Kurihara, Y. Sugimasa, and K. Kasahara,
"High-speed VCSEL of modulation bandwidth over 7.0 GHz and its application to 100 m PCF datalink," IEEE Journal of Lightwave Technology, vol.16, pp.870-8755, May 1998.
[11] A. Larsson, C. Carlsson, J. Halonen, and R. Schatz, “Microwave modulation characteristics of 840nm oxide confined / proton implanted VCSELs,”
Microwave Photonics technical report, October 2001.
[12] A. Larsson, C. Carlsson, A. Haglund, J. Halonen, and R. Schatz,
“Microwave modulation characteristics of BCB-planarized oxide confined 850nm VCSELs,” Microwave Photonics technical report, June 2002.
[13] A. N. AL-Omari, and K. L. Lear, “Polyimide-planarized vertical cavity surface emitting lasers with 17.0GHz bandwidth,” IEEE Photonics Technology Letter, vol.16, pp.969-971, April 2004.