Eye Diagram for a Direct Modulated Semiconductor Laser in ATLAS


An eye diagram is a convenient way to visualise how the waveforms used to send multiple bits of data can potentially lead to errors in the interpretation of those bits. This is the so-called problem of intersymbol interference. An eye diagram is also a useful means for readily obtaining information regarding the timing jitter, voltage swing and transition time of the modulation data. The eye diagram operation in ATLAS is a post processing function where the eye diagram is created by taking the time domain signal and overlapping the traces for a certain number of symbols. The time domain signal in this article represents the output from a directly modulated semiconductor laser. The modulation of the laser drive current takes a pseudo-random form.


LASER Simulation

Firstly a Fabry-Perot semiconductor laser is created in ATLAS using the advanced material module Blaze; the laser structure created is shown in Figure 1. The laser considered consists of an InGaAsP active region with InP cladding layers. An important parameter in a heterostructure device is bandgap alingment. How the bandgap difference is distributed between the conduction and valance bands has a large impact on the charge transport. In order to align the bandgaps use may be made of the ALIGN parameter in the MATERIAL statement. This specifies the fraction of the bandgap difference which will appear as the conduction band discontinuity. Internally, Blaze creates the desired conduction band offset by modifying the electron affinity of the material for which the ALIGN parameter is specified. In this example ALIGN=0.6 has been used for the InP material, therefore 60% of the bandgap difference is assigned to the conduction band offset.

Figure 1. Cross sectional view of the
simulated semiconductor laser device.

Semiconductor lasers are simulated in ATLAS with the module Laser. The module allows coupled electrical and optical simulation of semiconductor lasers. Laser works in conjunction with Blaze to solve the two dimensional Helmholtz equation in order to calculate the transverse optical field profile. It also allows calculation of the carrier recombination rate, optical gain and laser output power. The modal gain spectra for several longitudinal cavity modes is also readily available. Laser simulation is first initiated by specifying an independent rectangular mesh for the solution of the Helmholtz equation covering the entire active region. Since light may not entirely be contained within the active region this rectangular grid is chosen to extend so as to include an area slightly bigger than the active region. The turn on characteristics are shown in Figure 2 where it is clearly evident that for forward bias voltages from around 1.2V the output dramatically increases. Figure 3(a) shows the side view of the laser structure for a forward bias of 1.4V, where light emission is clearly evident, whilst Figure 3(b) shows the transverse mode profile obtained from Figure 3(a) by a 1 dimensional cut line across the active region. From figure 3(b) good optical confinement is evident.


Figure 2. IV characteristic for simulated semiconductor laser.


Figure 3(a). Cross sectional view showing
optical intensity spatial properties.


Figure3(b). Light intensity profile obtained via a 1 dimensional
cutline through the centre of the semiconductor laser structure.


Pseudorandom Modulation

Modulation of the laser is performed by adjusting the drive current. A forward bias of 1.3V is chosen to represent logic level 0, whilst a forward bias of 1.4V is chosen to represent logic level 1. These two logic levels are separately applied several times, the occurrence of a particular logic level being pseudorandom. In this way a pseudorandom modulation waveform, figure 4, is established. Example Atlas syntax for such a modulation scheme is detailed below.

solve vanode=$logic_1 ramptime=2e-10 tstop=2.5e-9 dt=0.1e-11
solve vanode=$logic_0 ramptime=2e-10 tstop=3.0e-9 dt=0.1e-11
solve vanode=$logic_0 ramptime=2e-10 tstop=3.5e-9 dt=0.1e-11
solve vanode=$logic_0 ramptime=2e-10 tstop=4.0e-9 dt=0.1e-11
solve vanode=$logic_1 ramptime=2e-10 tstop=4.5e-9 dt=0.1e-11
solve vanode=$logic_0 ramptime=2e-10 tstop=5.0e-9 dt=0.1e-11
solve vanode=$logic_0 ramptime=2e-10 tstop=5.5e-9 dt=0.1e-11
solve vanode=$logic_1 ramptime=2e-10 tstop=6.0e-9 dt=0.1e-11

Figure 4. Anode voltage versus time for
pseudorandom modulation scheme.

Each statement in the above syntax represents a transient in ATLAS. Considering such a group of transient statements as that above, ATLAS solves for the first transient, where ramptime is the rise time of the pulse and tstop is the time at which ATLAS stops the transient. For this first transient the pulse width is almost equal to tstop. On solving for the next transient in the group, risetime is once again the rise time of the pulse, however the duration of this pulse is essentially the difference between the tstop of the first transient and that of the second transient. ATLAS then proceeds onto the next transient and so on. In order to record the solution data for the transient simulations it is noted that a log file must be opened so that only transient data is recorded in it.


Eye Diagram Creation

The output photon density versus time, Figure 5, shows the modulated output from the semiconductor laser. Another very useful representation of the modulation data in terms of examining inter symbol interference, timing jitter, voltage swing and transition time for example, is that of an eye diagram. The eye diagram is constructed by taking a large sample of the time domain signal and essentially splitting it up into several equal time segments, termed here as the period, which are then overlaid on each other. The ATLAS syntax used for this is detailed below:

eye.diagram inf=for_eye_diagram_a.log outf=eye_diagramd.log t.start=1.0e-9 period=1.5e-9


Figure 5. Light intensity output from the
pseudorandom modulated semiconductor laser.

In the above, inf=for_eye_diagram_a.log specifies the input log file that the eye diagram algorithm is to use i.e. the log file that the time domain information is stored in, outf=eye_diagramd.log is the output file that the eye diagram is to be stored in. The eye diagram algorithm requires the time at which the data in the input log file is to be used for the eye diagram, this is specified with t.start. The period i.e. time segment, that the eye diagram is to use is specified by period. Care must be taken to correctly chose this value. The resulting eye diagram is shown in Figure 6. As can be seen from the figure there is a large open eye in the figure indicative of a reasonable modulation scheme. Also of notice is that there exist a reasonable transition time, indicated by the slope of the walls of the eye diagram.


Figure 6. Eye diagram of the modulated output.



It has been shown that a pseudo-random modulation scheme is easily performed in ATLAS . Also of ease is the creation of an eye diagram. Since it is a post processing function it is extremely quick to perform experiments regarding the choice of the period.

Download pdf Version