Optical Fiber

Role of Optical Fiber in short distance communication Janardhan N [1MJ04EC402] [email protected] Abstract With the continued search for the faster way to communicate things between the source and the destination has led man for a never ending hunger to his search for newer technologies. Many technologies with wide range of application oriented has shown up in the last decade. But the search for the much better technologies has and will never stop. Today we will put light on one such emerging technology, which is still in the rise among all. Yes, “Optical Fiber, in the field of communication.” Today let us discus an idea about the usage of Optical Fiber in the field of Integrated Circuits and in the field of Medical Research.

1. Introduction As in resent past, it is know for a lot about Fiber Optics. Yes, the same long, thin strands of very pure glass about the diameter of a human hair to coaxial cable used to transmit light signals over long distances communication, to a very short distance communication within the Integrated Circuits. One of the most important components in any optical fiber system is the optical fiber itself, since its transmission characteristics plays a major role in determining the performance of the entire system. Optical fiber refers to the medium and the technology associated with the transmission of information as light pulses along a glass or plastic wire or fiber. Optical fiber carries much more information than conventional copper wire and is in general not subject to electromagnetic interference and the need to retransmit signals. Most telephone company longdistance lines are now of optical fiber. In optical communication system, we have light as the carrier that transmits the data from source to destination and fiber itself as communication medium. Transmission on optical fiber wire requires repeaters at regular intervals depending upon the distance. The glass fiber requires more protection within and outer cable than other type of cables. Optical fibers are widely used in fiber-optic communication, which permits transmission over longer distances and at higher data rates than other

MVJCE

forms of communications. Fibers are used instead of metal wires because signals travel along them with less loss, and they are immune to electromagnetic interference.

2.1 Origin of optical fiber This idea is very simple. Let us fill up a container with water and shone a light into it. In a darkened room, then pull out the bung. The light shone out of the hole and the water gushed out. It is expected that the light would shine straight out of the hole and the water would curve downwards, as in the diagram. But the light stayed inside the water column and follows the curved path. Nature had found a way to guide light. What was expected and what actually happened here lead to the basic foundation of Optical Fiber. The basic requirements still remain the same today, a light source and a clear material (usually plastic or glass) for the light to shine through. The light can be guided around any complex path. Being able to guide light along a length of optic fiber has given rise to two distinct areas of use, light guiding and communications.

2.2 Modern day Optical Fiber Modern day optical fiber is oriented towards faster rate of communicating data between source and destination. Fiber might not to be in a line of sight, now light can pass through the complex loop as shown in the figure. Enormous resources were poured into the search for a material with sufficient clarity to allow the development of an optic fiber to carry the light over long distances. The early

1

results were disappointing. The losses were such that the light power was halved every three meters along the route. Within few years however, things have changed with the use of silica glass with losses comparable with the best copper cables. On the other hand, a fiber optic system using a glass fiber is certainly capable of carrying light over long distances. By converting an input signal into short flashes of light, the optic fiber is able to carry complex information over distances of more than a hundred kilometers without additional amplification. This is at least five times better than the distances attainable using the best copper coaxial cables.

Using the above example, value of 1.5 for the refractive index, this gives a speed of about 200 million meters per second.

2.2b How to provide data The angles of the rays are measured with respect to the normal. This is a line drawn at right angles to the boundary line between the two refractive indices, core and cladding region. The angles of the incoming and outgoing rays are called the angles of incidence and refraction respectively.

Advantages: This property of fiber to conduct even on bending made it more and more possessive towards new area of research.

2.2a Why is OFC, in such a huge demand Optical fiber had a property of commutating even when bent without much attenuation and on short versions of data communication had no or negligible data loss, which opted it more in medical and IC chip designing technology. It was indeed not necessary to add amplifiers on the output of every stage as in the above diagram, meaning no attenuation loss at all. Other important reason was its physical thin structure, smaller to accumulate in tiny areas compare to some other medium, i.e. fibers of 8 μm (315 millionths of an inch) diameter. Advantages: Conduction with or without amplifier at the later stage and its tiny structure had some commanding influence in the area of data communication.

As the refractive index is simply a ratio of the speed of light in a material to the speed of light in free space, it does not have any units. Lower refractive index => Higher would be the Speed.

MVJCE

The critical angle is well-named as its value is indeed critical to the operation of optic fibers. At angles of incidence less than the critical angle, the ray is refracted through the cladding region, outward. However, if the light approaches the boundary at an angle greater than the critical angle, then the light is actually reflected from the boundary region back into the first material. The boundary region simply acts as a mirror. This effect is called total internal reflection (TIR).

2.2c Types of optic fibers used There are two general categories of optical fiber: single-mode and multi-mode. We basically have Step indexed fiber optics and Graded index fiber optics. Multimode fiber was the first type of fiber to be commercialized. It has a much larger core than single-mode fiber, allowing hundreds of modes of light to propagate through the fiber simultaneously. Additionally, the larger core diameter of multimode fiber facilitates the use of lower-cost optical transmitters (such as light emitting diodes [LEDs] or vertical cavity surface emitting lasers [VCSELs]) and connectors. Single-mode fiber, on the other hand, has a much smaller core that allows only one mode of light at a time to propagate through the core. While it might appear that multimode fibers have higher capacity, in fact the opposite is true. Single mode fibers are designed to maintain spatial and spectral integrity of each optical signal over longer distances, allowing more information to be transmitted.

2

In step index, the core region will have a constant RI through out the core region and a stem increase in RI for the cladding region. In graded index, the RI will be high across the axis of the cable and decreases when moved towards the cladding region which supports higher order modes to propagate. Both Multi-mode and Single-mode has got advantage in their own fields. Advantages: Since both Medical and IC technology needs only one mode of data to propagate, step index would be better suited here and for the application where the data is to be transferred in parallel path, graded index would be preferred.

2.3a Role of Optical Fiber in Short distance communication Light is kept in the "core" of the optical fiber by total internal reflection. This causes the fiber to act as a waveguide. Fibers which support many propagation paths or transverse modes are called multimode fibers (MMF). Fibers which support only a single mode are called single mode fibers (SMF). Multimode fibers generally have a large-diameter core, and are used for long-distance communication links or for applications where low power must be transmitted. Single mode fibers are used for most communication links shorter than 20 meters. This makes the foundation for short distance communication between the source and destination, which is the most optimum in designing integrated circuits and other communicating devices. Present way of communicating data within the system is through communicating tracks, laid through printed circuit board or data cable through which the data is communicated between source and destination. The medium used for the transmission of information and data over distances has evolved from copper wire to optical fiber. It is quite likely that no wirebased information transmission systems will be installed in the future. The process of communicating data is taken over by optical fibers which are far better than the conventional copper wires, which has more power loss while propagating through and needs both voltage and current amplifiers

MVJCE

at all output points. As stated above, the communication between source and destination mainly requires either tracks or data cable which connect them together. Both come handy with cost and losses. Though PCB or tracks look like cost efficient, they are typically difficult to design and always require skilled technician to do it. Where as data cables are limited in both cost and distance they cover. Data cables are capable of communication effective data for shorter distances, but fails when it comes to transfer data to a longer distance. Any damage caused to either one port of the data cable makes it unusable for further use. There comes the new technology called optical fiber, which is the most efficient way of communicating data from one place to other. This was the famous quotation by the Scientist Stephen William Hawking “Optical Fiber will lead an example for all loss less communication in mere future” And now the worlds fastest calculating machine “ Blue Gene/L ”, Developed at Lawrence Livermore National lab, is capable of calculating 280.6 trillion calculations/sec which uses Optical fiber for its internal connection has done the miracle as stated above by Stephan Hawking Research and development of coherent optical fiber communications have been accelerated mainly because of the possibility of receiver sensitivity improvement reaching 20 dB, and partly because of the possibility of frequency-division multiplexing (FDM) with very fine frequency separation. In this paper, recent advances in the research on coherent optical fiber communication systems are reviewed, with emphasis on those reported in the past two years. Fibers that exhibit tightly controlled geometry tolerances will not only be easier and faster to splice but will also reduce the need for testing by ensuring Predictable, high-quality splice performance. This is particularly true when fibers are spliced by passive, mechanical, or fusion techniques for both single fibers and fiber ribbons. In addition, tight geometry tolerances lead to the additional benefit of flexibility in equipment choice. Now signal can be transferred as a single source of light or as a group of light sources. Step index fiber will be efficient for any serial communications like transfer of single data through out the path. Example transfer of acknowledgement signal from receiver to transmitter which is a single

3

bit and the transfer rate would be very fast so as to indicate transmitter to send next data.

Now transfer of data in parallel ports requires data bus as indicated above. But the same data can e treasured in a single string of fiber optics. That is by using Multi-mode step index fiber, where the data is transfused in multi modes configuration in such a way that N-signals could be transfused in a single string using multi mode configuration.

High Speed Optical Fiber Communication ICs Based on InP HEMT High-speed integrated circuit technology is the key to realizing large-capacity optical fiber communication systems. This paper describes the present status of 0. l-pm-gate InP HEMT ICs for the next-generation 40-Gbit/s/ch. systems. As an advanced IC technology, this paper also describes a 4O-Gbit/s OEIC that is monolithically fabricated with a uni-traveling-carrier photodiode and the 0.1-pm InP HEMTs. For the analog ICs, a distributed circuit configuration is attractive for widening the bandwidth. The conventional distributed amplifier, however, cannot ensure circuit operation from near DC because of the loss induced by the FET drain conductance. To cope with this, we devised a frequency dependent termination. In addition, a drain peaking line and a loss compensation circuit are also incorporated for loss compensation in high frequency region. These techniques make it possible to operate from near DC to the data rate and beyond. An InP

MVJCE

HEMT distributed baseband amplifier offers a 90GHz 3-dB-down bandwidth with 10-dB DC gain. Among the digital ICs, the decision IC (DEC) limits system speed because of its clocking operation with feedback action at the system data rate. To boost operation speed of the flip-flop (FE) circuit, which is a key element of the DEC and other digital ICs, we devised a super dynamic F/F (SD-F/F) [lo]. The key to increasing Operation speed is to minimize the effective logic swing to accelerate transition by introducing dynamic operation. A SD-F/F can operate 100 % faster than a conventional master-slave D-F/F. By using the SD-F/F, error-free operation of the packaged DEC was confirmed up to 50 Gbit/s, as shown in Fig. 4 (a) [I 11.

OEIC Technology The capability of monolithic integration with a photodiode is another great merit of the InP HEMT. As described in the previous section, the EDFA relaxes the gain requirement for the electrical amplifier in the optical receiver. Furthermore, a photodiode that has broad bandwidth and high saturation output power, such as the UTC-PD [19, 201, makes direct driving possible at the characteristic impedance of 50 R with a voltage swing of 1 Vp-p.

The combination of the EDFA and the monolithic OEIC with a UTC-PD and the 0.1-pm InP HEMTs has a high potential for realizing a simple, high-speed and highly functional optical receiver

4

without an electrical equalizing amplifier. The goal of the optical receiver is shown in Fig. 5. All high-speed signal processes needed for regenerating and retiming are executed in the chip, and the OEIC outputs demultiplexed low-speed electrical signals. Since the bandwidth limitations of the chip-to-chip interconnection and electrical amplifier are eliminated, higher-speed operation is expected compared with the conventional OEIC consisting of a photodiode and amplifiers. As a first step, we designed an optoelectronic decision IC (OE-DEC) and monolithically fabricated it with a UTC-PD and the InP HEMTs. The circuit consists of a UTC-PD, a bias circuit, a data suffer, a clock buffer, a core D-FF, and an output buffer and driver. A simple resistive divider circuit is adopted as the UTC-PD bias circuit. The data buffer consists of three stage differential amplifiers and generates the differential signal needed for internal circuit operation from the single ended electrical data signal converted in the UTC-PD and the bias circuit. The data buffer has a reference input terminal to adjust the slice level according to input optical power. A SD-FE is adopted as the core FE to achieve 40-GbiVs operation with a sufficient speed margin. The OEIC has single-ended electrical clock input and differential outputs that can be directly connected to an SCFL interface. In the present circuit, the load resistance was set to 50 S2 in order to compare the conventional discrete receiver and the monolithic OEIC in terms of circuit performance. Figure 7 shows a schematic cross-sectional view of the fabricated OEIC on an InP substrate. The hetero-epitaxitial layers of the UTCPD, Schottky diodes, and HEMTs grown by MOCVD were stacked, and the devices were fabricated in that order. The pulse response of the fabricated UTCPD was measured by electro-optic sampling (EOS). The full width at half maximum of the response signal was 4.8 ps and the responsively of the UTC-PD was 0.2 A/W. For several fabricated HEMTs, the typical transconductance was 1 S/mm, and the current gain cutoff frequency U;.> and the maximum oscillation frequency v,,) were 177 and 178 GHz, respectively. After on-wafer testing, the back side of the wafer was polished and then coated with anti-reflection film. Since back-side optical illumination is required for the optical interface of the OEIC, we developed a new OEIC package. The package structure is basically the same as that of the 4O-Gbit/s InP HEMT digital IC package 1 except for the inclusion of an optical input port. A set of aspherical lenses and a single-mode fiber was installed on the back side of the module. A

MVJCE

photograph of the OEIC module is shown in Fig. 8. Figure 9 shows the input and output waveforms of the OEIC module at 40 GbiVs when error-free operation was confirmed. Clear eye-opening was obtained at the output voltage of 950 mVp-p with the phase margin of43 degrees. The minimum averaged optical input power for correct circuit operation was 10.8 dE3m. This value is better than the 14.5 dBm confirmed for the discrete receiver configuration . As a feasibility study, the bit-error performance of an optical receiver consisting of two-stage EDFAs and the OEIC was examined for 40- Gbit/s RZ format optical signal. Figure 10 shows the biterror performance’s of the receiver. The average receiver sensitivity at the biterror-rate of was - 24.1 dE3m for the best performing channel, and - 23.6 dBm for the worst channel. No error-floor was observed, and the obtained sensitivity is, to our knowledge, the highest among 40- Gbit/s OEIC receivers reported to date. Optical communication plays a significant and increasing role in our society. The public demand for higher network speed requires an optical backbone network with larger capacity. Accompanying high transmission-rate optical communications system are severe technical specifications for optical devices and systems. Many popular optical devices could be represented with a digital filter model as described in this article. Use of well-developed signal processing techniques and algorithms to design these optical devices is a wise use of existing technology. As demonstrated in this article, signal processing could play an important role in the development of advanced optical communication systems. However, as demonstrated in the case of an electronic equalizer, some optical system characteristics may require special attention if signal processing techniques are to be applied successfully. Therefore, interdisciplinary cooperation between researchers in optics and signal processing will be crucial for optical communications to fully benefit from signal processing.

InP HEMT Technology A schematic cross-section of our 0.1-pm nAlAs/InGaAs/InP HEMT. The epitaxial layers were grown by metalorganic chemical vapor deposition (MOCVD). The electron channel is formed by a 15-

5

nmthick InGaAs layer beneath the T-shaped gate with a 0.1 - pm footprint. The carrier-supplying layer is an InAlAs layer with Si-delta doping. The most important feature of the devise structure is the InP gate-recess-etch stopper inserted into the InAlAs barrier layer. This drastically improves the uniformity of the threshold voltage ( V,h) [7]. The typical average Vrh is -0.6 V and the standard deviation is low at around 30 mV for a 3-inch wafer. The typical transconductance (gm) is 1 S/mm, and the current gain cut-off frequency U; .and maximum oscillation frequency
IC Design and Results The basic 40-Gbit/s optical sender (OS) and receiver (OR) configurations. The functions required for optical communication ICs are basically timedivision multiplexing, reshaping, retiming, regenerating, and time-division demultiplexing. Reshaping is performed by the Er-doped fiber amplifier (EDFA), photodetector (PD), preamplifier (Pre) and baseband amplifier (Base). The EDFA relaxes the noise and gain requirements for the preamplifier and baseband amplifier, but lower timing jitter is still required for both the decision IC (DEC) for regenerating and the clock extraction ICs, such as the exclusive OR IC (EXOR) and limiting amplifier (Limit.) for retiming. Broadband operation from near DC to the data rate with good eye opening is required for all ICs except for the clock extraction ICs. For the analog ICs, a distributed circuit configuration is attractive for widening the bandwidth. The conventional distributed amplifier, however, cannot ensure circuit operation from near DC because of the loss induced by the FET drain conductance. To cope with this, we devised a frequency dependent termination. In addition, a drain peaking line and a loss compensation circuit are also incorporated for loss compensation in high frequency region. These techniques make it possible to operate from near DC to the data rate and beyond. An InP HEMT distributed baseband amplifier offers a 90GHz 3-dB-down bandwidth with 10-dB DC gain. The measured S-parameters is for the amplifier. Among the digital ICs, the decision IC

MVJCE

(DEC) limits system speed because of its clocking operation with feedback action at the system data rate. To boost operation speed of the flip-flop (FE) circuit, which is a key element of the DEC and other digital ICs, we devised a superdynamic F/F (SD-F/F). The key to increasing operation speed is to minimize the effective logic swing to accelerate transition by introducing dynamic operation. A SD-F/F can operate 100 % faster than a conventional master-slave D-F/F. By using the SD-F/F, error-free operation of the packaged DEC was confirmed up to 50 Gbit/s. For high-frequency operation over 50 GHz, the distributed treatment is required even for the inside of the chip. A typical example is the 400-pm-long interconnection in the multiplexer IC (MUX), which corresponds to around two-tenths of the wavelength at 50 GHz. To reduce interconnection propagation delay time and to obtain an impedance-matched microstrip line, we introduced a two-metal-layer interconnection structure consisting of Au and a 2-pm-thick BCB film as an insulator. The MUX IC operated up to 80 Gbit/s by using the interconnection and impedance matching design shows the 80-Gbit/s output waveforms obtained in the on-wafer measurement. The fabricated IC chips were mounted onto our original package, which we call the chip-size cavity package [13, 141. This package supports operation up to 50-GHz. Table 1 summarizes the performances of the InP HEMT ICs. All ICs offer practical speed performance beyond 40-Gbit/s with good yield.

Questions of Strength One common misconception about optical fiber is that it must be fragile because it is made of glass. In fact, research, theoretical analysis, and practical experience prove that the opposite is true. While traditional bulk glass is brittle, the ultrapure glass of optical fibers exhibits both high tensile strength and extreme durability. How strong is fiber? Figures like 600 or 800 thousand pounds per square inch are often cited, far more than copper’s capability of 100 pounds per square inch. That figure refers to the ultimate tensile strength of fiber produced today. Fiber’s real, rather than theoretical; strength is 2 million pounds per square inch.

Advantages of Fiber Optics Why are fiber-optic systems revolutionizing telecommunications? Compared to conventional metal wire (copper wire), optical fibers are: Less expensive - Several miles of optical cable can be made cheaper than equivalent lengths of copper

6

wire. This saves your provider (cable TV, Internet) and you money. Thinner - Optical fibers can be drawn to smaller diameters than copper wire. Higher carrying capacity - Because optical fibers are thinner than copper wires, more fibers can be bundled into a given-diameter cable than copper wires. This allows more phone lines to go over the same cable or more channels to come through the cable into your cable TV box. Less signal degradation - The loss of signal in optical fiber is less than in copper wire. Light signals - Unlike electrical signals in copper wires, light signals from one fiber do not interfere with those of other fibers in the same cable. This means clearer phone conversations or TV reception. Low power - Because signals in optical fibers degrade less, lower-power transmitters can be used instead of the high-voltage electrical transmitters needed for copper wires. Again, this saves your provider and you money. Digital signals - Optical fibers are ideally suited for carrying digital information, which is especially useful in computer networks. Non-flammable - Because no electricity is passed through optical fibers, there is no fire hazard. Lightweight - An optical cable weighs less than a comparable copper wire cable. Fiber-optic cables take up less space in the ground. Flexible - Because fiber optics are so flexible and can transmit and receive light, they are used in many flexible digital cameras for the following purposes: Medical imaging - in bronchoscopes, endoscopes, laparoscopes Mechanical imaging - inspecting mechanical welds in pipes and engines (in airplanes, rockets, space shuttles, cars) Plumbing - to inspect sewer lines Because of these advantages, you see fiber optics in many industries, most notably telecommunications and computer networks. For example, if you telephone Europe from the United States (or vice versa) and the signal is bounced off a communications satellite, you often hear an echo on the line. But with transatlantic fiberoptic cables, you have a direct connection with no echoes.

integrated with a UTC-PD and the InP HEMTs was also confirmed to operate at 40 Gbit/s, and an optical receiver using the OEIC offered high receiver sensitivity. We believe that these IC and OEIC technologies based on InP HEMTs are promising for the construction of next- generation 4O-Gbit/ s/ch. optical fiber communication systems. Optical communication plays a significant and increasing role in our society. The public demand for higher network speed requires an optical backbone network with larger capacity. Accompanying high transmission-rate optical communications system are severe technical specifications for optical devices and systems. Many popular optical devices could be represented with a digital filter model as described in this article. Use of well-developed signal processing techniques and algorithms to design these optical devices is a wise use of existing technology. As demonstrated in this article, signal processing could play an important role in the development of advanced optical communication systems. However, as demonstrated in the case of an electronic equalizer, some optical system characteristics may require special attention if signal processing techniques are to be applied successfully. Therefore, interdisciplinary cooperation between researchers in optics and signal processing will be crucial for optical communications to fully benefit from signal processing. Reference 1.

K. -C. Wang, “High-speed circuits for lightwave communications,” World Scientific Publishing Co. Pte. Ltd., 1999.

2.

Y. Miyamoto, er. al., IEEE J . Solid-state Circuits.

3. C.K. Madsen and J.H. Zhao, Optical Filter 4.

Conclusion 40-Gbit/s ICs for next-generation optical fiber communication systems have been developed using 0.1 - pm InP HEMT technology. These ICs have sufficient speed margins for the 40-Gbit/s data rate. An optoelectronic decision IC monolithically

MVJCE

5.

Design and Analysis: A Signal Processing Approach. New York: Wiley, 1999. “Optoelectronics Circuit Collection”, Texas Instruments, http://wwws. ti.com/sc/psheets/sbea001/sbea001.pdf pdf files from www.ieeexplore.org

7

.

Integrated Circuit with CMOS Technology looks like this.

Integrated Circuit with Optical Fiber Technology looks like this.

This is the R&D Integrated Circuit, which is built up from Optical Fiber Technology. The size of all next generation IC’s will be or even smaller than this.

MVJCE

8