Modal Bandwidth Enhancement Through Launch Condition Optimization for High Data Rate VCSEL Transmission Over Multimode Fibers

Abstract

Vertical-cavity surface-emitting laser (VCSEL)-based transmission over multimode fiber (MMF) has achieved data rates of 100G per lane and is progressing towards 200G/lane, which demands more modal bandwidth from MMF to ensure adequate transmission reach. We address the needs of higher modal bandwidth from the point of view of engineering VCSEL launch conditions. We explore the options for using subsets of 10 standard-based launch conditions by analyzing the measured encircled fluxes from commercial VCSEL transceivers over two options. By utilizing experimentally measured MMF data, we demonstrated a significant improvement in modal bandwidth with these options. The launch conditions also impact the wavelength dependence of modal bandwidth for VCSELs operating at wavelengths longer than 850 nm. We conducted detailed Monte Carlo simulation of the wavelength dependence of modal bandwidth over MMFs. For one launch condition option using a smaller area, the modal bandwidths are improved over the effective modal bandwidth (EMB), and favor very high data rate transmission by allowing the use of a smaller area photodetector.

1. Introduction

Multimode fibers (MMFs) have been extensively employed for short-reach communications, such as within buildings, campuses, or data centers [1]. MMFs, used with vertical-cavity surface-emitting lasers (VCSELs), form a system suitable for short-reach applications spanning distances up to 100 m or more [2]. VCSELs offer cost-effective manufacturing and lower power consumption compared to traditional single-mode lasers, making them an economical and energy-efficient choice for short-range communications. In contrast to single-mode fibers, MMFs boast a large core diameter and high numerical aperture, enabling the use of lower-cost VCSEL light sources and connectors between fibers. The single-lane data rates of VCSELs over MMFs have advanced significantly, progressing from 10 Gb/s to 25 Gb/s, and more recently to 100 Gb/s. The rise of AI and GPU-based computing has indeed catalyzed this acceleration in data rate increases, and the latest GPU architectures are now paving the way towards 200G/lane speeds [3,4], where the constraints on transmission distance become pronounced. Toward 200G/lane VCSEL–MMF transmission [5], the transmission reaches are reduced to 30–50 m. Effective modal bandwidth (EMB) higher than OM4 threshold is needed for 200G/lane VCSEL transmission to go over 50 m.

Various approaches have been investigated to maintain transmission reach over MMF. One method to extend the transmission distance of MMF-based links is to reduce the VCSEL spectral width by using single-mode (SM) VCSELs [6], which exhibit narrower spectral widths and consequently lower penalties from chromatic dispersion-related bandwidth limitations. However, SM VCSELs are not commercially available. Currently, all VCSEL-based transceivers employ multimode (MM) VCSELs, which support a few spatial transverse modes and typically have a specified laser linewidth of 0.6 nm. This study addresses this challenge by proposing a relatively simple approach to enhance system transmission capabilities without requiring significant modifications to the existing VCSEL–MMF ecosystem, such as the development of a new type of VCSEL or MMF with exceptionally high EMB.

In this paper, we address the needs of higher modal bandwidth for high data rate transmission of 100G/lane and higher, especially with regard to engineering VCSEL launch conditions. In VCSEL transmission over MMF, the EMB is a critical parameter gauging MMF performance. EMB quantifies the quality of an MMF and therefore its transmission capability. Since early 2000s, data rates began increasing to 10G and beyond, and therefore the need for a standardized measure of modal bandwidth became critical. To define EMB, comprehensive testing conditions, including 10 launch conditions, were introduced to ensure fibers can support higher data rates over short distances [7]. The standard has played a crucial role in guiding the development of increasingly sophisticated optical communication systems. However, the defined launch conditions were originally based on early-generation VCSELs. From a modern perspective, the requirement to cover all 10 launch conditions may now be viewed as overly restrictive. Advances in VCSEL technology and tailored optical designs for VCSEL–MMF coupling can result in launch conditions that vary within a much narrower range, enabling MMFs to achieve higher modal bandwidths. In this work, we examine the 10 standard launch conditions and investigate the feasibility of using a subset to further enhance the modal bandwidth attainable in VCSEL-based transmission. In Section 2, we explore the options for using a special subset of the 10 launch conditions. In Section 3, using actual MMF data, we show the improvement of modal bandwidth in two scenarios. The launch conditions also impact the wavelength dependence of modal bandwidth for VCSELs operating at wavelengths longer than 850 nm. In Section 4, we present a study of the wavelength dependence of modal bandwidth over OM3 and OM4 with different launch options, based on the Monte Carlo modeling approach adopted in Ref. [8]. In Section 5, we compare various options and discuss the benefits and trade-offs of employing specialized launch conditions. Finally, conclusions are presented in Section 6.

2. VCSEL Launch Conditions

2.1. 10 VCSEL Launch Conditions for Determining EMB

The development of standards for multimode fiber and VCSELs has evolved over several decades. MMFs with a 50-micron core have been predominantly used in data communication systems. These fibers are categorized as OM2, OM3, and OM4 based on their EMBs at a wavelength of 850 nm. The abbreviation ‘OM’ stands for ‘optical multimode’. The OM3 fiber has an EMB of 2000 MHz·km, while the OM4 fiber has a higher EMB of 4700 MHz·km at 850 nm. They were standardized in 2002 and 2009, respectively. EMB is a measure of the bandwidth of a multimode fiber, considering the distribution of modes excited by a VCSEL. It reflects the fiber’s capacity to transmit data effectively over a particular distance, accounting for intermodal dispersion.

The development of EMB metrics for MMFs is closely tied to advances in VCSEL technology. In the early days of MMF communications, LEDs were used as light sources. With the advent of commercial VCSELs, conventional MMFs were tested with VCSEL-based transceivers to achieve gigabit performance, as established by the “Gigabit Ethernet” standard (IEEE 802.3z, approved in 1998). At Telecommunications Industry Association (TIA), a restricted launch bandwidth task group was formed to address these requirements [9,10,11,12]. In EMB calculations, weight coefficients are applied to account for the distribution of power among different modes and their cumulative contributions to modal delays. By using 10 launch conditions, which are weighted sums of modal contributions, EMB can accurately predict the bandwidth and performance of a multimode fiber link, provided that the light source meets the specified beam quality criteria, such as emission area and numerical aperture (NA) [7]. These conditions simulate various scenarios of light coupling from VCSELs into multimode fibers. In the factory, MMFs undergo differential mode delay (DMD) measurements by launching single-mode light at various radial positions across the core and recording the output pulses. Using the resulting DMD chart, output pulses for 10 VCSEL launch conditions—also known as DMD weights—are assembled according to their respective weight functions. The modal bandwidth for each launch condition is then calculated to determine the EMB. The range of launch conditions ensures that EMB reflects the overall modal bandwidth capability of the fiber, rather than a single launch outcome.

The DMD weight functions associated with the 10 launch conditions are shown in Figure 1a, from which 10 calculated modal bandwidths, referred to as EMBc, are derived, from which EMB is further determined. The EMB can be further obtained. We also note that the integration of the DMD weight function over radius, which represents integrated power over radius, is referred to as encircled flux. The standard requires that at the 4.5 μm radius the integrated power is less than 30% of the total power, and at the radius of 19 μm radius the integrated power is over 86% of total power [7]. In Figure 1b we show the radius reaching 86% power vs. cumulative power at 4.5 μm radius, which illustrate the characteristic of different weights. Weights 1 and 2 have higher optical power concentrations at lower radius or near the center of the core. Therefore, the cumulative power at 4.5 μm is higher than other weights. On the other hand, weights 3, 8, 9, and 10 are weights occupying the middle radius region of the core, while weights 4, 5, 6, and 7 reach 86% power at a radius above 16 μm, which means the optical power is more concentrated in the outer region of the core with higher radius.

2.2. VCSEL Launch Conditions for High Data Rate Transmission

The 10 VCSEL launch conditions accommodate a wide range of conditions that would be supported by all VCSEL transceiver vendors. VCSELs emit light with several transverse modes. These modes have non-uniform intensity distribution, leading to a ring-like or donut-shaped beam profile. Early VCSELs typically supported more modes due to larger aperture sizes leading to a broader emission pattern as exhibited by two distinct ring shapes, sometimes referred to as double-donuts. Such pattern occurs when the VCSEL supports multiple high-order modes simultaneously, each contributing to the overall emission profile and covering a wider radial region when coupled into multimode fiber. Weights 6, 7, and 8 have peaks in two different radial regions, matching the emission pattern from such VCSELs. Over the years, more modern VCSELs have come to possess fewer modes, i.e., 3–4 modes, and emission pattern is more single-donut structured. VCSEL technology has advanced significantly since 1998. Modern VCSELs feature smaller aperture diameters (typically 4–5 µm, compared to 8–15 µm for pre-2000 devices) and reduced beam divergence, with full width at half maximum (FWHM) at the 1/e² level less than 30 degrees [9,10,11,12]. The beam parameter product has improved substantially, enabling modern VCSELs to reproducibly excite fewer modes in the fiber and to be able to confine launched light to a narrower radial region within the MMF core.

In addition to the VCSEL emission pattern, the launch condition of VCSEL light into the MMF is influenced by the optical components responsible for coupling the laser light into the fiber. Specifically, these components determine how the light from the VCSEL is distributed across the modes of the MMF. In a specific optical design, due to the passive nature of the alignment between the VCSEL and the MMF, variations in alignment conditions may result in launch conditions that fluctuate within a certain range but do not encompass all 10 launch conditions specified in the standard. Therefore, it is anticipated that the modal bandwidth can be enhanced for a subset of the 10 specified launch conditions. To support this perspective, we conducted measurements on several commercially available optical transceivers.

In Figure 2, we show the encircled flux measurement results from Transceiver 1, which is a commercially available 200G SR4 transceiver [13]. It operates with four channels over MMFs, with each fiber transmitting 50G four-level pulse amplitude modulation (PAM4) data. We used an encircled flux measurement instrument (MPX-1) from Ardent Photonics to acquire the data. In this case, the 86% radius is reached between 14–16 μm, while the cumulative optical power at 4.5 μm is below 0.18. This set of launch conditions falls into the general region of weights 3, 8, 9, and 10. A majority of the optical power concentrates in the middle radial region of the fiber core.

In Figure 3, we show the results from Transceiver 2, which is a commercially available 400G SR4.2 transceiver [14], also referred to as 400G BiDi transceiver. It operates with eight channels bidirectionally over MMFs at 850 nm and 910 nm wavelengths. From the measurements of eight individual channels, the 86% radius is reached between 17 and 18 μm, which is close to the high-end limit set by the standard at 19 μm. A majority of the optical power is distributed in the outer portion of the fiber core. For both Transceiver 1 and Transceiver 2, the light, regardless of whether it is coupled into the multimode fibers (MMFs), is concentrated in the central radial region or the outer radial region, remaining within areas where the alpha profile is more likely to achieve optimal performance during the manufacturing process.

The measured encircled fluxes for the two types of transceivers shown in Figure 2 and Figure 3 suggest that two approaches may be adopted to optimize launch conditions, in order to achieve potentially better modal bandwidth performance. In one approach, one may focus on the middle radial region of the core of the fiber with a majority of light residing in this area. In another approach, a majority of light resides in the more outer region of the core with 86% radius reached at above 16.5 μm. In Section 3, we study how the two approaches can lead to modal bandwidth improvements. We note that different transceiver makers made their design choice based on how the VCSEL light is coupled into the MMF—as long as the coupling is in compliance with the standard—meaning the launch conditions fall into the range regulated by the 10 standard-based launch conditions. In the literature, there have been considerations of using different launch conditions, such as in [15,16], which uses offset launch to create ring-shaped excitation, and in [17], which uses diffractive optical element to avoid launching too much light into the center of the fiber. The unique aspect explored by the current work is that we categorize the launch conditions by judging whether they fall into the coverage of a subset of standard-based launch conditions. Since, for the 10 standard-based launch conditions, all the calculations for modal bandwidths based on the 10 launch conditions are already part of the existing measurement system, it would be straightforward to obtain the modal bandwidth for a chosen subset of launch conditions, which is, in general, not feasible for a launch condition proposed without this context. The idea here enables the current study and differentiates it from prior studies. Based on the considerations outlined above, we selected weights 3, 8, 9, and 10 for evaluation in ‘Option 1’, and weights 4, 5, 6, and 7 for ‘Option 2’. To ensure robust fiber performance, we assessed modal bandwidth in conjunction with MMFs that meet existing grades of MMFs, such as OM3 and OM4, to ensure that the fibers meet a sufficiently high baseline quality standard.

3. Analysis of Improvement of Modal Bandwidth over Different Launch Condition Options

We analyzed data from several thousand reels of OM4 fibers with an EMB meeting 4700 MHz·km, from which we evaluated how different launch conditions can impact modal bandwidth. Different launch condition options, as delineated in Section 2, were used to calculate the increase in modal bandwidth. We then broke down the modal bandwidth increases and grouped them to show the percentage of the fiber associated with the bandwidth increase. The percentage of fibers falling into certain bandwidth improvement percentage bins for ‘Option 1’ is shown in Figure 4a and ‘Option 2’ in Figure 4b. As can be seen from this analysis, in both cases, a large portion of the fibers see some modal bandwidth improvements. We also show the cumulative percentage of the fibers from the end with more improvements. As can be seen from this analysis, in both cases, a large portion of the fibers see some modal bandwidth improvements. We also show the cumulative percentage of the fibers from the end having highest improvements. We observe that more than 21% of the OM4 can see 10% or more bandwidth improvement against EMB for ‘Option 1’, while more than 19% of the OM4 can see 10% or more bandwidth improvement for ‘Option 2’. The improvements are significant and can help achieve high data rate transmission when certain portions of the transmission require reach at the high end of the transmission specifications. It is understood that one mechanism of modal bandwidth improvement stems from using a subset of the 10 launch conditions, which is less restrictive than using all launch conditions. On the other hand, for each potential launch condition set, sufficient room should be given to accommodate the nature of passive alignment for VCSEL–MMF coupling. Although the study here shows two examples, different combinations are potentially feasible.

4. Wavelength Dependence of Modal Bandwidth Through Monte Carlo Simulation over Different Launch Conditions

4.1. Generation of Refractive Index Profiles for Monte Carlo Simulation

In the current work, we adopted the Monte Carlo simulation scheme in Ref. [8] to obtain the wavelength dependence of modal bandwidth in MMFs. Here, we briefly review the key aspects of the technical approach. The relative refractive index profile, or delta profile of an MMF with graded-index profile, is often referred to as alpha profile. It takes the form,

$$∆\left(\mathrm{r}\right)={∆}_0\left[1-{\left(\mathrm{r}/\mathrm{a}\right)}^{\mathsf{\alpha }}\right]$$

(1)

where ${\mathsf{\Delta }}_0$ is the relative refractive index in the center of the core defined as ${\mathsf{\Delta }}_0=(n_0^2-n_1^2)/\left(2n_0^2\right)$, $n_0$ is the refractive index in the center of the core, n₁ is the refractive index of the cladding, a is the core radius, and α is the profile shape parameter, referred to as alpha parameter or alpha shape parameter. The fiber has 1% core delta, ~50 µm core diameter. The MMF is optimized when the alpha parameter has a value of around 2.1, at which point the MMF’s modal bandwidth peaks at a wavelength near 850 nm. A trench with negative delta value is placed outside the core to enhance the bending performance of the fiber [18].

Due to the nature of MMF fabrication, each individual MMF differs from another. In the study, we generated 10,000 refractive index profiles to represent MMFs over a range of variation types that can occur in real fibers. Figure 5 illustrates how modal bandwidth is affected by the alpha and non-alpha factors. The bandwidth of MMF is peaked at a particular wavelength, as determined by the alpha value. However, the alpha value can vary over a range resulting in MMFs having modal bandwidth peak at different wavelengths. Therefore, to emulate actual fiber, we would alter the alpha value with the peak wavelength varying over a range of over 200 nm around 850 nm. The refractive index profile can also deviate from the ideal alpha profile, which is referred to as non-alpha error. It can affect the maximum bandwidth that can be achieved in Figure 5. In the modeling we expand such errors using Fourier series as shown here:

$$∆n\left(r\right)={\sum }_{i=1}^N\left[a\left(i\right)·\sin \left({\displaystyle \frac{i\pi r}{a}}\right)+b\left(i\right)·\cos \left({\displaystyle \frac{i\pi r}{a}}\right)\right].$$

(2)

In our actual generation of the Monte Carlo data, N is chosen to be 6. Another type of error that can occur is the center portion of the core refractive index, which can degrade system performance [19]. We introduce such errors within the first several microns of the core radius. One last type of error to consider is the variation of the trench. In this case, we generated random offsets in the Monte Carlo data set.

Using the refractive index profiles, the EMB of a particular fiber can be obtained from its DMD chart, which is the sequence of output pulses resulting from single-mode pulse launches at various radial offsets at the input. In reality, the DMD chart is measured using a DMD measurement bench, while in our case, it is calculated based on the refractive index profile [20]. In the Monte Carlo study, we generated 10,000 cases of refractive index profiles to cover a wide range of profile variabilities, with a mixture of all types of profile errors. The magnitude of each type of error was also randomized within a range to yield MMFs with EMB spread over a range of interest. The purpose of the Monte Carlo set is to generate refractive index profiles with various possibilities that can affect the modal bandwidth under different launch conditions. For each wavelength, we calculated the EMB from the profiles as described above. By combining all the wavelengths, we can generate the wavelength dependence curve of one individual MMF. The Monte Carlo simulation method described in Ref. [8] has been validated against IEC guidelines for the wavelength dependence of EMB for OM3 and OM4 fibers [21], showing good agreement. In the current work, we have applied the method to study the wavelength dependence of MMFs under specialized launch conditions.

4.2. Wavelength Dependence of Modal Bandwidth of MMFs

Using the data generated from the Monte Carlo simulation described in Section 4.1, we conducted a detailed study of wavelength dependence of modal bandwidth for several situations. First, we show the wavelength dependence of OM4 with two launch condition options: ‘Option 1’, using weights 3, 8, 9 and 10, and ‘Option 2’, using weighs 4, 5, 6, and 7. Among the 10,000 cases, fibers meeting the requirements were filtered out. Among them, the modal bandwidth value at 99.9 percentile or with lowest value at bottom 0.1 percentile was extracted for each wavelength of interest. The results are presented in Figure 6a. The data for OM4 and OM4 with ‘Option 2’ are essentially identical and therefore overlap. The modal bandwidth of ‘OM4 w/Option 1’ is notably higher than that of the other two cases at wavelengths above 850 nm, with improvements of 0.3 GHz·km or more in the long-wavelength regime. This situation can be understood with the illustration in Figure 7. For an ideal MMF optimized for 850 nm, the average pulse delay for various radial offsets across the fiber core exhibits minimal spread and remains nearly identical, resulting in essentially zero relative delay. However, at longer wavelengths, material dispersion causes the average delay to shift higher toward higher radial offset, which is shown as right tilt curve on the right side of Figure 7. The light spot with radius r2 is larger than that with radius r1, and the corresponding delay d2 is greater than d1. Since modal bandwidth is inversely proportional to delay, a launch condition at larger radius r2 with associated delay d2 will result in reduced modal bandwidth at longer wavelengths. The bandwidth improvement due to the use of a small launching spot or radius is quite significant once moving away from 850 nm. On the other hand, since the weights used in ‘Option 2’ are those reaching higher radius among the standard-defined 10 weights, using ‘Option 2’ does not improve modal bandwidth at longer wavelengths. The situation is also true for ‘Option 2’ for OM3 fiber as shown in Figure 6b. The data for OM3 and OM3 with ‘Option 2’ are essentially the same, so that they overlap. The modal bandwidth of ‘OM3 w/Option 1’ is notably higher than that of the other two cases at wavelengths above 850 nm, with improvements of 0.15 GHz·km or more in the long-wavelength regime. Note that OM3 is defined for MMFs with EMB of 2000 MHz·km at 850 nm.

We observe that for ‘Option 1’, the light in the fiber reaches 86% at a radius substantially lower than the required 19 um radius and fiber core radius of 25 μm. This would result in another benefit of enabling the use of higher bandwidth photodetector with smaller active area, which could be very helpful for 200G/lane VCSEL transmission. As the data rate reaches 200G, every component needs to be stretched for more bandwidth. The photodetector (PD) is a bottleneck because it needs to collect light from an MMF, which has a core diameter of 50 μm, much larger than standard single mode fiber. The bandwidth of a PD, such as a PIN diode used in VCSEL-based optical transceivers, is closely tied to its active area (detector area). A smaller detector area generally allows for higher bandwidth due to reduced capacitance. Since the existing photodetector used for VCSEL–MMF transmission has a detector area with a diameter of 30–40 μm, reducing such area to diameter of 20–30 μm with coupling optics could substantially increase the PD bandwidth. We view this as another benefit associated with certain launch options, such as ‘Option 1’, explored in this work.

5. Discussions

In this paper, we studied a new approach to improve the modal bandwidth of MMFs by tailoring the launch conditions for specific applications. To better understand the benefits of this approach, we discuss and highlight a few relevant aspects below.

• Special launch conditions versus other approaches for bandwidth improvement: Many different approaches have been considered for improving the high data rate transmission over MMFs [6]. At the VCSEL side, an SM VCSEL with narrow linewidth to reduce the chromatic dispersion-related bandwidth has been considered. However, the hurdle to develop and commercialize an SM VCSEL-based transceiver is high despite very active research. On the other hand, developing a new type of MMF with higher modal bandwidth is also a complicated matter, given the competition between SM and MM transmission. Adopting special launch conditions for emerging high-demand applications, such as 200G per lane transmission, could be an appealing approach, as it does not require the development of new types of VCSELs or MMFs. Regulating the launch conditions by a small group of transceiver makers for new 200G VCSEL transmission is much more feasible, without disturbing brown field or existing deployments.
• Benefits for 200G/lane transmission using ‘Option 1’ launch conditions: 200G/lane or 200G per wavelength VCSEL transmission is emerging [4,5]. A study group at IEEE under 802.3 is establishing a new generation of the standard at this new data rate to address data communications in AI applications for a transmission reach of 50 m or less. Modal bandwidth above OM4 threshold seems to be needed, as learned from early testing. Using the proposed launch conditions can ensure more fiber can meet higher modal bandwidth. As shown in Section 3, more OM4 fiber can meet higher modal bandwidth levels when the launch conditions are tailored for a specific subset of the existing 10 standard-based launch conditions. Another benefit of using ‘Option 1’ launch conditions is that it can enable the use of a small area photodetector with a detector area diameter of 20–30 μm, resulting in significant improvements to photodetector bandwidth.
• Benefit for wavelength division multiplexing (WDM) applications involving long wavelengths: In addition to the 850 nm wavelength window, which has been used dominantly for VCSEL-based transmission over MMF, WDM has also been adopted involving longer wavelengths. Since OM4 fiber is more 850 nm-optimized, the longer wavelength applications have been more often modal bandwidth-limited. The launch condition optimization studied in the current work offers one avenue to extract out more modal bandwidth, in addition to other approaches of using a new sub-type of MMFs. The ‘Option 1’ launch condition, as studied in Section 4, shows that it can offer significantly higher modal bandwidth.
• Trade-offs of using a subset of 10 known launch conditions: Utilizing a subset of the 10 widely adopted launch conditions allows us to leverage existing calculations and implement these conditions without developing new weighting functions. However, a key trade-off is ensuring that actual launch conditions align with those in the selected subset. The encircled flux measurements presented in Section 2 demonstrate the feasibility of this approach, but validation across a broader range of transceivers is still needed. Another limitation is the continued inclusion of weights 6, 7, and 8, which correspond to double-donut-shaped light distributions in the MMF. While such distributions were common with earlier VCSELs, modern devices rarely produce them, making their inclusion potentially outdated.

6. Conclusions

VCSEL-based transmission over multimode fiber (MMF) has achieved data rates of 100G per lane and is advancing toward 200G per lane. The VCSELs currently used in commercial transceivers are predominantly multimode VCSELs. However, achieving higher modal bandwidth is essential for supporting high data rate transmissions of 100G per lane and beyond, especially to meet the distance requirements of specific applications. In this study, we address this need by focusing on engineering VCSEL launch conditions, which requires only moderate adjustments to regulate VCSEL–MMF coupling in the development of new transceivers.

Using measured encircled flux data from commercially available 200G and 400G VCSEL transceivers, we identified two optimal launch condition options as subsets of the 10 standard launch conditions. By leveraging experimentally measured MMF data, we demonstrate significant improvements in modal bandwidth achieved with these launch conditions. Additionally, we examined the wavelength dependence of modal bandwidth for VCSELs operating at wavelengths longer than 850 nm across two grades of MMFs through Monte Carlo simulations. For launch conditions involving a smaller light spot than the fiber core area, modal bandwidths showed substantial improvements compared to EMB on the standard 10 launch conditions at longer wavelengths above 850 nm. Furthermore, this specific set of launch conditions offers an additional advantage for supporting very high data rate transmission by enabling the use of small-area photodetectors.