SOME STUDIES ON SEMICONDUCTOR OPTICAL AMPLIFIER AND KERR TYPE OPTICAL SWITCHING

Transcription

1 SOME STUDIES ON SEMICONDUCTOR OPTICAL AMPLIFIER AND KERR TYPE OPTICAL SWITCHING Thesis submitted to The University of Burdwan, Burdwan for the award of the degree of Doctor of Philosophy by Soma Dutta under the guidance of Prof. Sourangshu Mukhopadhyay DEPARTMENT OF PHYSICS THE UNIVERSITY OF BURDWAN, BURDWAN DECEMBER , Soma Dutta. All rights reserved.

2 Dedicated to My Family ii

3 The University of Burdwan Department of Physics Golapbag, Burdwan, West Bengal, India Phone: Fax: CERTIFICATE FROM THE SUPERVISOR This is to certify that the thesis entitled SOME STUDIES ON SEMICONDUCTOR OPTICAL AMPLIFIER AND KERR TYPE OPTICAL SWITCHING, submitted by Soma Dutta to The University of Burdwan, Burdwan, is a record of bona fide research work under my supervision. The contributions comprised in the thesis are done by her and it was not submitted to any other University or Institute for any degree or diploma. It is my pleasure to mention that Soma Dutta did her Ph. D. work with full sincerity and dedication. I wish every success in her life. Date: Place: Burdwan (Sourangshu Mukhopadhyay) Professor, Dept. of Physics, The University of Burdwan, Burdwan. iii

4 Acknowledgements I wish to express my deep sense of gratitude to my supervisor Prof. Sourangshu Mukhopadhyay for his invaluable guidance and inspiration throughout this course of research work. His advice, encouragement, patience and ample experience were invaluable and vital to the completion of this work. My wholehearted thanks to my laboratory colleague,, for their cooperation, adjustment and encouragement during this long period of research. I am also thankful to west Bengal state fund, for providing financial support during the course of this study. Above all, I would like to express my gratitude to my parents Mr. Mrinal Dutta and Mrs. Sujata Dutta, my younger sister Sayani Dutta, my husband Supravat Karak and all other family members and friends for their constant encouragement and love. The University of Burdwan, Burdwan December, 2012 Soma Dutta iv

5 Preface Optics has a strong and very potential role in information and data processing because of its inherent parallelism. It has several advantages over electronics in super-fast computation and data processing. Last few decades, several all optical data processors were proposed based on Boolean logic. Those optical systems and optical logic devices based on optical switches are found very much useful than electronic ones in connection to speed and many other aspects. Different types of all-optical methods have been proposed for implementation of the optical logic and arithmetic processors and devices. These types of optical systems require some optical switches like non-linear material based switches, electro-optic material based switches, optical bistable materials, optical filters, optical converters and beam splitters, Semiconductor Optical Amplifier (SOA) based optical switches etc. There are several types of encoding principle to implement the all-optical logic and arithmetic devices. These are the intensity encoding principle, polarization encoding principle, phase encoding and frequency encoding principle. Above all these encoding principles, frequency encoding principle is the most faithful and reliable one. Frequency is the basic characteristics of light. In optical computing and data processing therefore the most important point is the very high speed of processing. The interesting point is that this problem in long distance communication can be solved by using frequency encoding of light. One can encode and decode two different states of information by two different frequencies. These frequencies are remaining unaltered during the reflection, refraction, absorption etc when data is to be transmitted. Using this v

6 frequency encoding principle various types of all-optical logic operation are reported earlier. These operations are based on different types of optical switches like SOA based optical switches, optical modulator based switches, non-linear based switches etc. To use these types of optical switches different types of all-optical logic processors and devices are developed by scientists through all over the world. Observing the tremendous importance of optics in computation and communication I felt too much interested to contribute some research works in this area of optical computing and parallel processing. The current thesis encloses my research contributions. The work contained in this thesis is original and has been done by me under the guidance of my supervisor. The work has not been submitted to any other Institute for any degree or diploma. (Soma Dutta) The University of Burdwan Burdwan West Bengal India vi

7 CONTENTS Title Page Some studies on Semiconductor Optical Amplifier and KERR type optical switching Certificate Acknowledgements Preface Contents i iii iv v vii Chapter 1 A challenge of optical information processing with frequency encoded principle: An Introduction Advantages of optics over electronics in information processing, computing and data handling Optical logic, Arithmetic and algebraic 3 operations with optical switching technology 1.3 Various types of encoding for all optical logic 4 operations 1.4 Frequency encoding/ Wavelength encoding and 7 its advantages 1.5 Optical switches for conducting frequency encoded logic system Fundamentals characteristics of 9 SOA Using SOA as an Ultra fast 11 Nonlinear Medium Frequency conversion utilizing 13 Nonlinear Polarization Rotation (NPR) of SOA probe beam SOA based add/drop 15 multiplexer SOA based wavelength 17 converter SOA acting as a polarizationswitch (PSW) Objective of the thesis Conclusion 21 References 23 Chapter 2 A method of implementing frequency encoded alloptical latch with semiconductor optical amplifier 33 vii

8 Chapter Introduction Optical implementation of a NOT based latch Optical implementation of two-bit memory cell Conclusion 41 References 42 Some new approaches of conducting all-optical 44 frequency/wavelength encoded logic operations, programmable logic units and RS flip-flops with semiconductor optical amplifiers 3.1 Introduction Scheme of realization of frequency encoded optical 47 OR logic operation Principle of operation of optical OR 48 gate 3.3 Scheme of realization of frequency encoded Optical 51 AND gate Principle operation of optical AND 53 gate 3.4 Scheme of realization of frequency encoded Optical 55 NAND gate Principal operation of optical NAND 56 gate 3.5 Scheme of realization of frequency encoded Optical 59 R-S flip-flop Principle operation of optical R-S 60 flip-flop 3.6 Frequency encoded programmable logic unit scheme Optical OR logic gate controlled by 63 an optical signal Scheme of realization of 64 optical OR logic gate controlled by the light signal Optical AND logic gate controlled 67 by the light signal Scheme of realization of 68 optical AND logic operation controlled by the light signal Optical NAND logic gate controlled 71 by the light signal viii

9 Scheme of realization of 71 optical NAND logic operation controlled by the light signal Optical NOT logic gate controlled by the light signal Scheme of realization of 75 optical NOT logic operation controlled by the light signal Optical XOR logic gate controlled by the light signal Scheme of realization of 77 optical XOR logic operation controlled by the light signal Scheme of realization of frequency encoded programmable logic unit Important requirements for the 84 switching of SOA 3.8 Conclusion 85 References 86 Chapter 4 A new approach of implementing all-optical frequency/wavelength encoded clocked S-R flip-flop 4.1 Introduction Optical implementation of clocked S-R flip-flop Conclusion 94 References Chapter 5 Chapter 6 98 A new method for transmission of frequency encoded parallel data 5.1 Introduction A new method of frequency encoded parallel 100 data transmission through optical waveguide 5.3 Essential requirements for implementation of 103 the practical transmission of data 5.4 Conclusion 106 References 107 A new approach of developing Universal all-optical multiplexer with frequency encoded mechanism 109 ix

10 Chapter 7 Chapter Introduction Method of developing a frequency encoded 111 universal multiplexer with SOA 6.3 Method of developing double triggering 113 universal multiplexer 6.4 Practical realization Conclusion 117 References 119 Use of all-optical Kerr Cell for super fast conversion 121 of a binary number having a fractional part to its decimal counterpart and vice-versa 7.1 Introduction Optical tree architecture Non-linear material as an optical switch Optical conversion method of a binary number 126 having a fractional part to its equivalent decimal number 7.5 Optical conversion method of a decimal number 130 having a fractional part to its binary equivalent 7.6 Alternating approach of non-linear switch Conclusion 135 References 136 General conclusion and future scope of the work Introduction Conclusion of the thesis Future scope of the work in this area 141 x

11 CHAPTER-1 A challenge of optical information processing with frequency encoding principle: An Introduction 1.1 Advantages of optics over electronics in information processing, computing and data handling: Optics has already been established as a potential and promising candidate in information and data processing. Many all-optical logical algebraic and arithmetic processors have been proposed since the middle of the decade of seventy s [ ]. Several schemes of all optical logic gates, optical digital memory units, optical algebraic processors, image processors are proposed by the scientists around the globe since the decade of seventy. The choice of optical signal in replacement of conventional electronic signal in a data processor is mainly because of the inherent parallelism in optics, which leads a suparfast up gradation of computing technology. Different types of all-optical methods have been proposed for implementation of the optical logic and arithmetic processors and devices. These types of optical systems are required some optical switches like non-linear material based switches, electro-optic material based switches, optical bistable materials, optical filters, optical converters and beam splitters, Semiconductor Optical Amplifier (SOA) based optical switches etc. Again to support the increasing demand and rapid growth information, optics has been proved as proper alternative for very high speed communication with high bit rate and low bit error rate. In optical communication network the response time at the nodes is a very important issue for setting high speed communication. For managing the tremendously increasing day to day 1

12 data traffic, it is very necessary to enhance the transmission link capacity as well as the speed of the switching networks at the nodes. The realization of a network node with throughput at the order of 100 GB/s is not far away. SOA grating combination has been successfully used for 160 GHz wavelength conversion. Again using a cross-correlation system and non-linear polarization rotation 200 GBPS wavelength conversion at temporal resolution at 1.5 PS is also reported. Some logic gates are also implemented based on the four-wave mixing character of SOA with the mechanism of polarization shift keying and with tri-state operation logic [ ]. Communication and data processing is totally turned over to electronic communication and data processing when electron was found in nineteenth century. Actually twentieth century can be regarded as a century of electronics. But at the end of the last century scientists and technologists found some big problems of electronic systems which are relating to the speed of operation and many others also. These problems are written serially below. I. Electronic systems can not raise its speed greater than few GHz limit. II. Being a charged particle it has coulomb interaction with other charged particle or even among them. Because of these character an electron when behaves as an information carrier, it faces Von-Neumann bottle neck problems, problems related to cross-talk, interaction with external electric field, magnetic field and electromagnetic field and many others. III. Electronic massage can not be sent through a high magnetic field without interaction. 2

13 IV. It has also traffic jam problem in information broad ways with increase of data in the information banks. The problems can be solved with optics. In optical system photon is used as an optical signal carrier. It has many advantages over electronics. The advantages are- I. Photon is a basically charge less particle having no rest mass and obeys Bose- Einstein statistics and some times follows the classical passion distribution. For these reason problems like cross-talk, interaction with other charged particle or even among themselves can be removed. II. It was proved that photonic system can be extended an operation far more than THz limit. III. It has high degree of inherent parallelism. IV. In Electronic system I am forced to follow the only Boolean algebra. In photonic system one can code binary, ternary, quaternary, decimal, hexadecimal or even multi-valued data with optical signal. Different types of encoding system are there in photonic system. Except the above advantages there can be found many other points by which one can infer that optics is very much suitable as an information carrier over electronics. 1.2 Optical logic, Arithmetic and algebraic operations with optical switching technology: There has been significant research attempt in the general area of information processing by optical techniques over the last twenty years [ ]. Recently optical computation is the new addition part of this work. To perform numerical computations on one-dimensional or multidimensional data that are generally not images optical 3

14 computing is defined as the use of optical systems. In optical computation one can design, simulate, or analyze optical systems with optics and also can form the images. Optical computing systems have many prospective advantages, and also many disadvantages. Optical processors are fundamentally two-dimensional and parallel systems and can have high space-bandwidth and time-bandwidth products. Without any interaction optical signals can propagate through each other in separate channels, and can transmit in parallel channels without interference and crosstalk. Due to inherent parallelism of optics, in computation and arithmetic data processing it has made successful contributions. Several types of operation schemes for arithmetic data processing have already been reported by scientists over last few decades [ ]. Even though in wide and local area networks optical technologies are playing increasingly important roles. Several types of all-optical logic and algebraic operations are reported by many scientists. In recent years, some low-level all-optical logic gates mostly based on the nonlinearity of semiconductor optical amplifiers (SOA) are reported by several scientists [ ]. 1.3 Various types of encoding for all optical logic operations: The most important thing to implement some successful optical logic and arithmetic processors is how to encode your optical system. Several types of encoding systems are used to develop optical logic systems. These encoding systems are depends on intensity of light, polarization of light, phase change of light, frequency of light etc. Using these characters of light some popular optical encoding systems are made as for example Intensity encoding, Polarization encoding, Phase encoding and Frequency encoding principles etc. In intensity encoding principle different intensities of light is encoded as 4

15 logic state 1 and 0. If I intensity of light represents the logic state 1 then 2I intensity of light represents the logic state 0. In this intensity encoding method some non-linear materials are used as a switch. For some isotropic material the well established Kerr nonlinearity equation is n o 2 = n + n I (1.1) Where n is the refractive index (r.i) of the concerned non-linear material, n 2 is the non linear correction term, n 0 is a constant linear refractive index term and I is the intensity of light passing through the material. As for example of some non-linear material working as a optical switches are Pure silica glass (SiO 2 ), carbon di-sulfide (CS 2 ), gallium arsenide (GaAs), etc. This material shows this type of non-linearity and also follows this Kerr non-linearity equation. As for example for carbon di-sulfide the constant linear refractive index term n 0 = 1.62 and n 2 = m 2 /W, whereas for pure silica the values of constant linear refractive index term is n 0 = 1.46 and n 2 = m 2 /W. The refractive index (n) is changed when the intensity of incident light is changed and then the path of output light is also changed according to their changing r.i. Here any one can measure the amount of lateral change of channel due to refraction of light from NLM. This lateral change is measured by the term Δθ. Again it is seen that for use of input signal pulse LASER have a significant value of Δθ but continuous LASER does not have any significant value of Δθ. As for example for pure SiO 2 a laser pulse of duration 10 7 s and of power 100 mw incident light is made double then an angular separation of the refracted beam (from the NLM) becomes whereas the input incident angle fixed at 45 [ ]. In polarization encoding principle polarization of light is encoded as logic states 1 and 0. If un-polarized light represents the logic state 0 then polarized light 5

16 represents the logic state 1. To implement the Boolean logic operations the logic states are indicated by the intensity level of the light signal where nonlinear material is working as a switching element. For that reason it is required to maintain a constant intensity level of the light signal to represent any logic state. However the intensity of the signal is changed due to long distance communication of optical signal. Then the logic processor whose working principle is based on the intensity level of the signal fails to work properly in the detecting side. This problem may be avoided using the principle of polarization based encoding/decoding technique. Using of optical polarization based encoding systems for implementing logic family was already proposed by Lohmann et-al [ ]. In phase encoding principle two different phase of light is encoded as logic states 1 and 0. In optical phase encoding process, a co-sinusoidal light signal having, Acosωt represents the logic state 1 and with the help of a phase modulator introducing a phase π in a co-sinusoidal signal Acosωt represents the logic state 0. In the decoding process if the phase of the output wave is zero with respect to a reference signal 3Acosωt then the output bit will be 1 and it will be 0 if the phase of the output wave is π with respect to the reference signal 3Acosωt [ ]. To implement Boolean logic one can mention a frequency encoded technique which is more advantageous in contrast to above proposals. It is very well known that frequency of light remains unchanged during transmission of light signal. If the frequency n 1 represents the logic state 1 state and the frequency n 2 represents the logic state 0 then n 1 and n 2 will remain unaffected through out the transmission of data. Similarly, if ν 1 frequency of light represents the logic state 0 then ν 2 frequency of light represents the logic state 1. Several types of all-optical 6

17 logic operation and logic devices is already developed successfully using these all encoding techniques [ ]. 1.4 Frequency encoding/ Wavelength encoding and its advantages: Frequency is the fundamental character of light. So it remains unchanged in reflection, refraction, absorption etc. In optical computation photon is found to be a very suitable information carrier than electron not only in the connection of super fast speed but in many other aspects of information processing also. Thus these photonic systems can successfully replace the electronic systems. Again it is also seen that in case of optical data processing the conventional methodologies can not be followed always as it is done in electronics. Scientists and technologists are deeply involved in research to overcome the speed related difficulties to realize all-optical logic, arithmetic and algebraic operations with Boolean mechanism. There are found several popular reports on the development of optical logical systems where the logic gates are the basic building blocks. To implement of optical Boolean logic systems using the coding norms, as the presence of optical signal is equal to 1 and absence of optical signal is 0, face several problems. For these many alternative approaches are reported to encode 1 and 0 with optical signal. For example, several areas like modulation, spatial input/output encoding; symbolic substitution etc. which deal with different coding principles. Again it is also seen that many proposals are reported where the same coding mechanism (i.e. presence of optical signal is regarded as 1 and absence as 0) is followed to implement the optical logic gates in Boolean mechanism. In most of the cases, the presence of optical signal at the input or output of a logical system is encoded as 1 bit and the absence of the signal is regarded as 0 logic state, i.e. by intensity variation mechanism. As the intensity of 7

18 light decreases with the increase of optical path through a medium, so the intensity of light may drop down below the reference level of the concerned logic state 1 and may enter into the reference level of logic state 0. So this is not problem free encoding system. Similarly polarization encoding principle has some disadvantages. In this polarization encoding method two orthogonal polarized states of light are represented by 1 and 0 logic state respectively. During transmission and propagation the state of polarization may change at reflecting points and refracting points, so this reason may entree several problems in implementation of the optical logic gates. In wide range of data processing, Boolean system has some own limitations. That s why to implement Boolean logic gates lots of different proposals may be mentioned for the different coding norms. All the encoding norms and techniques mentioned above have some advantages but all of them widen the loss dependent problem. In compare to those proposals one can mention an advantageous proposal to implement Boolean logic which is called the frequency encoded technique. The frequency encoding principle can be used as a more reliable candidate than other encoding principles for the development of super-fast optical processors, as the frequency of a signal remains unchanged even after different optical transformations. To overcome all the above problems one can use the frequency of light for encoding a bit. If the presence of a specific frequency of light is treated as 1 state and then other specific one represents 0 state i.e. if 1 state is encoded by frequency ν 2 then that of the 0 state is done by another frequency by ν 1 where ν 1 and ν 2 remain unaltered throughout the transmission of data. Frequency encoding techniques is free from such transmission problems. This encoding technique has some more 8

19 advantages like very low BER effect, many data can be transmitted parallel through one optical channel and it is not attenuated in long distance communication [ ]. 1.5 Optical switches for conducting frequency encoded logic system: To develop the all-optical frequency encoded logic and arithmetic processors scientists are used some optical switches like SOA based wavelength converters; add/drop multiplexers, some optical filters, some optical prisms etc. The semiconductor optical amplifier (SOA) is one of the most frequently used devices for ultra fast signal processing. This is because the device technologies are already well established and it is commercially available. Ultra fast signal processing depends on how to take out the beneficial features of SOA [ ]. In this chapter I briefly review about the basic properties of SOA, their features and applications as an ultra fast nonlinear switch Fundamentals characteristics of SOA: The full form of SOA is semiconductor optical amplifier. The basic structure of an SOA is very simple and it has some similarity with semiconductor laser diodes. But SOA is different from semiconductor laser diodes for an anti-reflection (AR) coating on both facets and it is also sometimes used as the pointed structure of the waveguide close to the facets. These are to minimize the reflection at both surfaces. The active layer is surrounded as a gain medium and it can either be bulk, quantum well, or quantum dots. Using the following equation optical gain of the SOA with bulk or quantum well active layer can be achieved [1.68] g ω cn k = [ ( )] () 1 ( w) = D( k) Im χ ( k) k = 0 dk 9

20 k = ω = D k cnε ( ) hγ CV μ ( ω ( k) ω) 2 cv k = 0 h cv + h γ CV ( f ( k) f ( k) )dk c v (1.2) Here c is the velocity of light, ω is the light wave angular frequency, ε 0 is the vacuum dielectric constant, n is the refractive index and D(k) is the density of state expressed in terms of wave number k, ω cv (k) is the k-dependent transition frequency between conduction band (C) and valence band (V), γ CV is the dephasing time and f C, f V are Fermi Dirac distribution functions. How to comprehend large output saturation power, small noise and polarization insensitive gain is the main important thing in the development of SOA. The output power saturation takes place for the intense optical power due to reduced population inversion. It consumes the population inversion carriers. If I use the SOA at an output power above the saturation power, I cannot get good eye opening for random pattern modulation, known as the pattern effect. In general the equation of saturation power is given by [1.68] dω 1 ρ s = Ch ω (1.3) Γ g τ d Where d is the thickness of active layer, C is the fiber-chip coupling efficiency, w is the width of active layer, g d is differential gain, Γ is the optical confinement factor, τ is carrier lifetime and dω corresponds to the mode cross-section. Γ For a good semiconductor optical amplifier the most essential things are output saturation power, polarization intensive gain and low noise. The large saturation power depends on the design of the active layer, optical confinement factors and carrier lifetime. Polarization intensive gain also depends on the active layers. A waveguide with a thin active layer gives a small loss to the TE mode when compared with that for the TM mode 10

21 for polarization dependence. The rectangular cross section of the waveguide and the active layer give polarization independent waveguide loss. Because of degenerate heavyhole and light-hole bands the polarization dependence is small in the bulk active layer. In the quantum well active layer polarization-dependent optical gain occurs for the removal of degeneracy. For minimizing polarization dependence there are several types of designs of active layer. They consist of geometry and strain of active layer and the waveguide structure. Several types of SOA have been developed. The bulk active layer SOA may be the most suitable available device for large saturation power and polarization insensitive response SOA as an ultra fast non-linear optical switch: Semiconductor optical amplifier is acting as an ultra fast non-linear medium. The electron-hole pair density dependence of the nonlinear refractive index of GaAs is investigated by Lee et al. Absorption reduces and becomes negative for high excitation levels when increase the photo-excitation, correspondingly, the refractive index changes. This analysis was carried out based on the Kramers Kronig relation which is also applicable for SOA. The refractive index changes with the optical gain correspondingly because of the current injection. Slow response associated with carrier recombination is one of the most inconvenient features of the SOA for their ultra fast signal processing. So far two methods are reported to overcome this problem. One of them is the direct use of the ultrafast response by selecting only the fast response using a wavelength filter [ ] and the other one is the use of a symmetric Mach Zehnder configuration with SOA at each arm to cancel out the slow response component [1.76,-1.78]. 11

22 Four types of non-linearity s are found in SOA, which are cross gain modulation (XGM), cross phase modulation (XPM), self phase modulation (SPM), and four wave mixing (FWM). Cross gain modulation (XGM) is a result of gain saturation in SOA. It occurs when lights of two different wavelengths, a pump and a probe, are injected into the semiconductor optical amplifier when operated under gain saturation conditions; the available optical gain is distributed between the two wavelengths depending on their relative photon densities. The changes in the power level of the pump wavelength have an inverse effect on the gain available to the probe wavelength and results in data transfer. When the pump (λ 2 = nm) is not present, the gain to the probe wavelength (λ 1 = nm) is high so that the output power of the probe is very high. When the pump (λ 2 ) and probe (λ 1 ) are injected into the SOA at the same time, the pump power is so high that it saturates the gain of the SOA. The available gain to the probe (λ 1 ) will be reduced. So the output power of the probe is much lower. When the power of the pump is modulated with data, the gain of the probe is also modulated. Thus the output power of the probe is modulated. This results in transfer of data from pump to probe. Thus wavelength conversion is achieved. A cross-phase modulation (XPM) accompanies the cross gain modulation when two optical signals are simultaneously present in the SOA. An interferometer configuration can be used to convert the phase modulation to an intensity modulation. XPM in a semiconductor optical amplifier (SOA) used in an interferometer configuration has been used for all-optical wavelength conversion, optical demultiplexing and for optical clock recovery. The scheme has high conversion efficiency and high signal to 12

23 noise ratio. Generally a Mach-Zehnder or Michelsen interferometer configuration integrated on a single chip is used to convert phase modulation to intensity modulation. Semiconductor amplifiers are known for optical nonlinear effects, such as the four-wave mixing (FWM). FWM in SOA s has been used as a technique for performing wavelength conversion due to its good conversion efficiency and high speed response for wavelength division multiplexing (WDM) networks. Four-wave mixing (FWM) is a process by which optical signals at different (but closely spaced) wavelengths mix to produce new signals at other wavelengths, but with lower power. In the FWM process, light at two frequencies, ω 0 and ω 1, are injected into the amplifier. These injected signals are generally referred to as pump and probe beams. The pump and probe beams can be obtained from two single wavelength distributed feedback (DFB) lasers. The pump signal is of higher power than the probe signal. Consider the case when both the pump and the probe signals are CW. Propagation through the SOA results in the generation of two additional FWM signals with frequencies 2ω 0 ω 1 and 2ω 1 ω 0. The intensity of light at these wavelengths is measured using a spectrometer. The FWM signal at frequency 2ω 0 ω 1 has higher power if the pump signal strength (at frequency ω 0 ) is higher than that of the probe signal. If I 0 and I 1 are the intensities of the signals at frequencies ω 0 and ω 1, the intensities of the signals at frequencies 2ω 0 ω 1 and 2ω 1 ω 0 are proportional to I 2 0 I 1 and I 0 I 2 1 respectively Frequency conversion utilizing Nonlinear Polarization Rotation (NPR) of probe beam in SOA. Non-linear polarization rotation of the probe beam is the basic property of SOA. Optically induced nonlinear refractive index in a bulk SOA by highly intense pump 13

24 beams are responsible for this Non-linear polarization rotation of the probe beam. The intense pump beam can modify the optical properties of the SOA during the interaction of the intense pump beam with probe beam in nonlinear SOA which in turn modify the intensity of probe beam as well as its SOP. When a linearly polarized light is coupled in a SOA then its SOP will change after leaving the SOA. To measure the non-linear rotation in terms of intensity difference a polarization beam splitters (PBS) is present at the output end. The whole scheme is shown in figure-1.1. Now for the first case with suitable current the SOA has to be biased and also X and Y are the input pump beams are adjusted into the proper power level. Z is the linearly polarized weak intensity probe beam of frequency ν 0. This weak probe beam is coupled with the pump beams in SOA. The polarizer is adjusted in such a way that the pass axis of the polarization beam splitter (PBS) is crossed with respect to SOP of the linearly polarized probe beam (Z) when there is no input beams i.e. the absence of both the input pump beams X and Y. In this situation no light is obtained at the output end (O). Again no light is obtained at the output end (O) when only one input pump beam is present at the input (X/Y). So ν 0 frequency is obtained at the final output end (O) only when two pump beams are present at the input end (X, Y). It is to be noted that if only one pump beam has such intensity which is equal to sum of both input pimp beam intensity then one pump beam of such intensity can change the state of polarization of the probe beam of SOA. So, therefore using that pump beam of such intensity as a control beam one can get the transmitted probe beam to input end to output end. 14

25 X Y INPUT PUMP BEAMS SOA LINEARLY POLARIZED BEAM O PBS OPTICAL FILTER Figure-1.1: Frequency conversion of probe beam by NPR method of SOA SOA based add/drop multiplexer To achieve a successful routing of wavelengths, the ability of SOA to add and drop a specific wavelength channels in a wavelength-division multiplexed (WDM) network is also a great important function. This is an Add/Drop multiplexing unit of SOA. The function of an Add/Drop multiplexer (ADM) is to select one particular wavelength of light without interfering with the adjacent wavelengths. Several types of add/drop multiplexers are reported for developing several optical devices, some of them use grating filters and circulators and others use different light wave technology. The filters can be tuned by changing the biasing input current into the SOA. The tunable filter has the transfer function of the spectral width 0.9 nm around the selected wavelength. The selected specific wavelength is reflected by the filter and amplified by the multiple 15

26 quantum wells (MQW) and the circulator is used to drop the selected wavelength in a required direction. Others wavelength which comes parallel at the input along with the specific one will pass through the SOA made filter. So one can separate a particular wavelength of light using this add/drop multiplexer from a band of wavelengths. The system is shown in figure-1.2 schematically. As for example if one wants to get particular ν 1 frequency light from a stream of data of frequencies ν 1, ν 2, ν 3,.ν n then SOA has to be adjusted or tuned at the proper biasing current of ν 1 frequency then all the frequencies passes through the ADM only ν 1 frequency is reflected form the SOA and is collected by the circulator [1.65, 1.79]. Biasing terminal for ν 1 frequency MQW amplifier Grating filter Input signal λ 1 (ν 1 ), λ 2 (ν 2 ), λ 3 (ν 3 ) λ n (ν n ) Circulator (C) Reflected signal λ 1 (ν 1 ) ADD/DROP Multiplexer λ 2 (ν 2 ), λ 3 (ν 3 ) λ n (ν n ) Output signal λ 1 (ν 1 ) collected by the circulator Figure-1.2: SOA based ADD/DROP multiplexer 16

27 1.5.5 SOA based wavelength converter In modern communication those switching devices are be very much effective where one light signal is switched by another light signal. Wavelength conversion is based on the XGM character of SOA is a result of gain saturation phenomenon. A weak CW probe beam (at a specific wavelength) and a strong pump beam (at another specific wavelength) of light are injected jointly into the SOA. At a suitable biasing current in the amplifier the probe beam will be treated as a strong beam output from the SOA because of XGM character. Thus one can refer this incidence as wavelength conversion. There are two types of basic schemes used in XGM based wavelength conversion, one is copropagating and another one is counter propagating schemes. In the first case the pump and probe beams are injected from the same side of the SOA and in the second scheme pump and probe beams are injected in mutually opposite directions into the SOA. Copropagating scheme has better noise performance. A weak CW (continuous wave) probe light of wavelength λ 2 and a strong pump beam of wavelength λ 1 are injected into the input terminals of the SOA having an anti-reflecting surface at its input side for λ 2 and a highly reflecting surface for λ 1 at the output terminal. In this situation the strong pump beam transfers its total power to the weak probe beam and thus the weak probe beam being a stronger one and comes to the output terminal. The scheme is shown in figure If there lays no pump beams at the input side, no conversion is allowed [1.65, 1.66]. 17

28 λ 2 (ν 2 ) λ 1 (ν 1 ) Strong pump beam CW weak probe beam Semiconductor optical amplifier SOA λ 1 (ν 1 ) Output strong light beam of frequency ν 1 Figure-1.3: SOA based wavelength converter SOA acting as a polarization-switch (PSW) Semiconductor optical amplifier can work as a polarization switch. To design the non-linear polarization switching (PSW) one can use the properties of polarization SOA gain saturation [ ]. To implement this switching system two laser sources of different frequencies, one strained bulk SOA, a power meter, three polarization controllers, an attenuator and one polarization beam splitter (PBS) are needed. The whole scheme of the polarization switching is shown in Fig-1.4. The polarization controllers PC 1, PC 2, PC 3 are controlled the probe beam (this is a CW laser of ν 1 frequency), pump beam (highly intense beam of frequency ν 2 ) and output beam respectively. Now to reduced the power of input probe beam (-15 dbm) it is applied to one input terminal of SOA via an attenuator. The polarization direction of the input probe beam be 18

29 approximately 45 0 to the orientation of SOA layer because the orientation of linearly polarized probe beam is adjusted by PC 1 in that a way. The polarization beam splitter (PBS) is combined the output beam of SOA. The output beam from SOA is divided into two parts by the PBS, one part is the horizontal (H) and another part is the vertical polarization component (V). The vertical component of SOA output is received at port-1 and horizontal component at port-2 respectively. The two modes transverse electric field (TE) and transverse magnetic field (TM) components which are decomposed by the optical field of linearly polarized light due to absence of pump beam propagate through SOA independently. This propagation should be amplified by the biasing current in SOA. When the maximum gain of TE and TM modes are almost equal because of the biasing current value (162 ma) then under this condition PC 3 orients the state of polarization of output beam of SOA and the beam at the output port-1 becomes zero i.e. vertical component (V) of the output beam of SOA is absent and as a consequence maximum power is delivered at port-2. SOA have the property of polarization dependent gain saturation. So in the case the polarization dependent gain saturation character give rise to different refractive index change for TE and TM when highly intense pump beam is present. So one can say probe beam will appear at port-2 (ON-state) when the pump beam is absent. The probe beam will be suppressed in port- 2(OFF-state) when the pump beam of specific intensity is present and it is Obvious that the state of port-1 will be complementary with respect to port-2 i.e. power will develop at port-1 when the pump beam is present [1.65, 1.66]. 19

30 PUMPING CURRENT ATTENUA TOR PC1 SOA PC3 PBS H PORT 2 PROBE BEAM ν 1 PC2 ν PUMP BEAM ν 2 PORT 1 Figure-1.4: SOA acting as a polarization switch 1.6 Objective of my thesis: From the above discussion, it is clear from the study of optical switches and encoding systems that there exists considerable scope for investigation to improve the performance of the systems used in optical computation and communication. This dissertation contains seven chapters with a focus on some of the above issues. Chapter 1 gives a detail introduction on the need of optical frequency encoding and different types of optical switches. This chapter also discusses on some back ground review works in this area. Chapter 2 reports on the implementation of all-optical frequency encoded NOT latch using semiconductor optical amplifier based switches. Chapter 3 mainly deals with frequency encoded logic gates, R-S flip-flop and programmable logic unit with semiconductor optical amplifier based switches. 20

31 Chapter 4 presents a frequency encoded all-optical clocked S-R flip-flop based on semiconductor optical amplifier switches. Chapter 5 reports on the transmission on the frequency encoded parallel data. Chapter 6 deals with the frequency encoded universal all-optical multiplexer system. Chapter 7 reports on all-optical Kerr Cell for super fast conversion of a binary number having a fractional part to its decimal counterpart and vice-versa. Optical systems are very important for computation and communication and it will be very much essential and effective for the future communication. So the other objective of the thesis is to get a review work on the use of different types of optical switches for implementation of different types of optical logic systems. Optical systems with frequency encoding principle are very new concept and have many advantages over others. So, I make a review work on this topic. The objective is to know about the different types of optical switches which are very essential to develop optical logic and arithmetic systems. Again I also take a look of those optical devices which are already been established successfully. In this thesis I want to get an overview of all-optical frequency encoded logic operations. 1.7 Conclusion; By reviewing in this work in the area of frequency encoded optical logic and arithmetic operations I received the knowledge about the various kinds of optical operations and different optical systems and also got the idea of using light in high speed communication systems. I also reported the advantages and disadvantages of different types of optical encoding and decoding processes, especially the advantages of frequency encoding principle and different types of frequency encoded optical switches, devices and 21

32 methods are discussed here. The prime beauty of using frequency of light for encoding the logic states to implement the optical system is that the frequency is the fundamental character of light, so it is unchanged during reflection, refraction, absorption and transmission etc. this frequency encoding technique is more reliable than other encoding techniques to implement the optical logic and arithmetic operations. 22

33 References: 1.1 A. K. Datta, S. Mukhopadhyay, A. Basuray, Carry-less arithmetic operation of decimal numbers by signed digit substitution and its optical implementation, Optics Communications (The Netherlands), 88(2-3), 15 March, 1992, Debendra Nath Das, S. Mukhopadhyay, Image edge detection and enhancement by an inversion operation, Applied Optics, 37(35), 1998, A. Basuray, S. Mukhopadhyay, Hirak Kumar Ghosh, A. K. Datta, A tristate optical logic system, Optics Communications (The Netherlands), 85(2-3), 1 September, 1991, S. Mukhopadhyay, An optical conversion system: From binary to decimal and decimal to binary, Optics Communications (The Netherlands), 76(5-6), 15 May, 1990, S. Mukhopadhyay, A. Basuray, A. K. Datta, New technique of arithmetic operation using the positional residue system, Applied Optics (USA), 29(20), 1990, G. Della Giustina, G. Brusatin, M. Guglielmi, C. Palazzesi, E. Orsini, P. Prosposito, Complex epoxy based sol-gel structures by UV and soft lithography, Solid State Science,doi: /j.solidstatesciences (2010). 1.7 Stella, F. Casalboni, M. Cirillo, M. Merlo, C. Palazzesi, C. Pepe, G. P. Prosposito, P. Salvato, M; Integrating Superconductive and Optical Circuits; Appl. Phys. Lett. 92, 2008, R. S. Kaler, T. S. Kamal, A. K. Sharma, S. K. Arya and R. A. Aggarwal, Large Signal Analysis of FM-AM Conversion in Dispersive Optical Fibers for PCM systems 23

34 including Second Order Dispersion, Fiber and Integrated Optics Incorporating International Journal on Optoelectronics, 21 (3), 2002, Er. K. S. Bhatia, Dr. R. S. Kaler, T. S. Kamal and R. Kaler, Coded OFDM in Fiber- Optics Communication Systems with Optimum biasing of Laser, Signal Processing: An International Journal SPIJ, 5 (2), M.T.Fatehi, K.C.Wasmundt and S.A.Collins, Optical flip-flops and sequential logic circuits using a liquid crystal light valve, Appl. Optics, 23, 1984, M.S.Alam, Efficient trinary signed digit symbolic arithmetic, Opt. Lett, 19(5), 1994, Jun-feng Zhang, Yuping Chen, Feng Lu, Xianfeng Chen, Flexible wavelength conversion via cascaded second order nonlinearity using broadband SHG using broadband SHG in MgO-doped PPLN, Optics Express, 16(10), May,2008, L.Q. Guo and M.J. Connelly, All-optical AND gate with improved extinction ratio using signal induced nonlinearities in a bulk semiconductor optical amplifier, Optics Express,14(7), 2006, H.J.S.Dorren, Daan Lenstra, Yong Liu, Martin T. Hill, Giok-Djan Khoe, Nonlinear Polarization Rotation in Semiconductor Optical Amplifiers: Theory and Application to All-Optical Flip-FlopMemories, IEEE Journal of Quantum Electronics, 39(1), Jan-2003, J. W. Goodman, Introduction To Fourier Optics. New York: McGraw-Hill K. Preston, Jr., Coherent Optical Computers. New York: McGraw-Hill, 1972, Ch D. Casasent, Ed., Optical Data Processing-Applications (Topics in AppliedP hysics, Vol. 23). Berlin: Springer-Verlag,

35 1.18 S. H. Lee, Ed., Optical Information Processing Fundamentals (Topics in Applied Physics, Vol. 48). Berlin: Springer-Verlag, H. Stark, Ed., Applications of Optical Fourier Transforms.New York: Academic Press, J. W. Goodman, Operations achievable with coherent optical information processing systems, Proc. IEEE, 65, 1977, E. A. Mnatsakanyan, V. N. Morozov, and Yu. M. Popov, Digital data processing in optoelectronic devices (review), Sov.j. Quantum Electron., 9, 1979, J. N. Roy and S. Mukhopadhyay, A minimization scheme of optical space-variant logical operations in a combinational architecture, Opt. Commun. 119, 1995, A. A. Sawchuck and T. C. Strand, Digital optical computing, Proc. IEEE 72, 1984, A. Huang, Y. Tsunoda, J. W. Goodman, and S. Ishihara, Optical computation using residue arithmetic, Appl. Opt. 18, 1979, S. Mukhopadhyay, J. N. Roy, and S. K. Bera, Design of a minimized LED array for maximum parallel logic operations in optical shadow casting technique, Opt. Commun. 99, 1993, S. Mukhopadhyay, Binary optical data subtraction by using a ternary dibit representation technique in optical arithmetic problems, Appl. Opt. 31, 1992, M. T. Fatehi, K. C. Wasmundt, and S. A. Collins, Jr., Optical flipflops and sequential logic circuits using a liquid crystal light valve, Appl. Opt. 23, 1984,

36 1.28 P. Ghosh, D. N. Das, and S. Mukhopadhyay, All alternative approach of optical flip-flop with tristate logic, Indian J. Pure Appl. Phys. 36, 1998, W. Wu, S. Campbell, S. Zhou, and P. Yeh, Polarization encoded optical logic operations in photorefractive media, Opt. Lett. 18, 1993, C. Denz, T. Dellwig, J. Lembeke, and T. Tschudi, Parallel optical image addition and subtraction in a dynamic photo-refractive memory by phase code multiplexing, Opt. Lett. 21, 1996, S. Mukhopadhyay and D. N. Das, Image edge detection and enhancement by inversion operation, Appl. Opt. 37, 1998, S. Mukhopadhyay, P. Ghosh, and D. Dutta, A conversion scheme of decimal number to its residue for optical computation, Indian J. Pure Appl. Phys. 30, 1992, A. Basuray, S. Mukhopadhyay, H. K. Ghosh, and A. K. Datta, A tristate optical logic system, Opt. Commun. 85, 1991, T. Houbavlis, K. Zoiros, K. Vlachos, T. Papakyriakopoulos, H. Avramopoulos, F. Girardin, G. Guekos, R. Dall Ara, S. Hansmann, H. Burkhard, IEEE Photon. Technol. Lett. 11, 1999, H. Soto, D. Erasme, G. Guekos, IEEE Photon. Tech. Lett. 13, 2001, C. Bintjas, M. Kalyvas, G. Theophilopoulos, T. Stathopoulos, H. Avramopoulos, L. Occhi, L. Schares, G. Guekos, S. Hansmann, R. Dall Ara, 12, 2000, J.H. Kim, Y.M. Jhon, Y.T. Byun, S. Lee, D.H. Woo, S.H. Kim, IEEE Photon. Technol. Lett. 14, 2002,

37 1.38 U. Efron, Spatial light modulators for optical information processing Intl. optical computing conf.proc.spie-700, 1986, H.Hedcten and G.Mannesbery, optical fourier analysis of diffusely reflective screen dot pattern with optically addressed spatial light modulator, Opt.Engg.34(2), 1995, K.B.Nichols, B.E.Burke, B.F Aull, W.D.Goodhue, B.F.Granstorff, C.D Hyot and A.Vera slms using charge coupled device addressing and electro-absorption effects in GaAs/AIgaAs multiple quantum wells, Appl.Phys.Let.,52,1988, L.A.Lentine, H.Scoh Hinton, D.A.B Miller, J.E.Henry, J.E.Cunninghan, Symmetric self E-O effect devices: optical set reset latch, different logic gates and different modulator detector, IEEE J.Quan.Elect.QE 25, 1989, L.C.West, Picoseconds integrated optical logic, computer, 20(12), 1987, M.E.Prise, N.C.Craft, R.E La Marcha, M.M.Downs et.al., Module for optical logic circuits using symmetrical self electro optic devices., Applied Optics, 29(14), 1990, N.T.Pelekanos, B.Deveaud, J.M.Gerard, H.Hass et.al. All optical spatial light modulator with MHz modulation., Opt.Lett., 20(20), 1995, A. W. Lohmann, What classical optics can do for the digital optical computer, Applied Optics, 25, 1986, A. W. Lohmann, Polarization and optical logic, Applied Optics, 25, 1986, K. Mukherjee and P. Ghosh, A novel frequency encoded all optical CNOT gate exploiting difference frequency generation and implementation of fast binary adders using frequency encoding and nonlinear dielectric films., Optik-International Journal for Light and Electron Optics, DOI: /j.ijleo

38 1.48 T.Chattopadhyay, J,N.Roy, Polarization encoded all-optical quaternary successor with the help of SOA assisted Sagnac switch, Opt.Commun. 2011, doi: /j.optcom M.S. Alam, M.A.Karim and A. A. S. Awwal, Two dimensional register design using polarization encoded optical shadow casting, Opt.Commun., 106, 1994, T. Sato, K. Shiraishi, K. Tsuchida and S. Kawakami, Laminated polarization splitter with a large split angle., Appl. Phys.Lett. 61(22), 1992, T. Yatagai, Optical space-variant logic gate based on spatial encoding technique, Opt.Lett. 11(4), 1986, A.A.S. Awwal and M.A. Karim, Microprocessor design using polarization encoded optical shadow casting, Appl.Opt. 29(14), 1990, W.Wi, S.Cmpbell, S.Zhou and P.Yeh, Polarization encoded optical operations in photorefractive media, Opt.Lett. 18(20), 1993, Y.A.Zaghloul and A.R.M.Zaghloul, Unforced polarization based optical implementation of binary logic, Opt.Epress 14(16), 2006, Jianqi Zhang and Huan Xu, Optical computation based on nonlinear total reflection optical switch at the interface., Pramana-journal of physics, 72(3), 2009, Puntsri, K. Sittichivapak, S. Shalaby, H.M.H, throughtput/delay analusis of spectrally phase encoded optical CDMA over WDM networks, Communications B. Chakrabarty and S. Mukhopadhyay, Alternative approach of conducting phasemodulated all-optical logic gates, Opt.Eng. 48(3), 2009, H.E. Michel and A.A.S. Awwal, Artificial nearal networks using complex nembrs and phase encoded weights, Appl.Opt. 49(10), 2010, B71-B82. 28

39 1.59 L. Q. Guo and M. J. Connelly, A novel approach to all-optical wavelength conversion by utilizing a reflective semiconductor optical amplifier in co propagation scheme,opt. Communication, 17(281), Sept.2008, Sisir Kumar Garai, S. Mukhopadhyay, A novel method of developing all-optical frequency encoded memory unit exploiting nonlinear switching character of semiconductor optical amplifier, Optics and Laser Technology, 42(5), 2010, S.K.Garai,D.Samanta,S.Mukhopadhyay, All-optical implementation of inversion logic operation by second harmonic generation and wave mixing character of some nonlinear material, Opt.Optoelectronic Technol.,China6 (4), 2008, J.Wang,J.Sun,C.Luo,Q.Sun, Experimental demonstration of wavelength conversion between ps-pulses based cascaded sum and difference frequency generation (SFG+DFG) in LiNbO 3 waveguides, Opt.Express 13(19), 2005, K.Gallo,G.Assanto,G.Stegeman, Efficient wavelength shift over the erbium amplifier band width via cascaded second-order process lithium niobate waveguide, Appl. Phys. Lett.7, 1997, M.H.Chou,I.Brener,M.M.Fejer,E.E.Chaban,S.B.Christman, 1.5 mm band wavelength conversion based on cascaded second order nonlinearity in LiNbO 3 channel wavelength, IEEE Photon Technol.Lett.11, 1999, M. J. Connelly, Semiconductor Optical Amplifiers, Kluwer Academic publishers, N. K. Dutta and Q. Wang, Semiconductor Optical Amplifiers, World Scientific publishing Co. Pte. Ltd,

40 1.67 G. Papadopoulos, K. E. Zoiros, On the design of semiconductor optical amplifierassisted Sagnac interferometer with full data dual output switching capability, Optics and Laser Technology, 43(3), 2011, Hiroshi Ishikawa, Ultrafast All-Optical Signal Processing Devices, A John Wiley and Sons, Ltd, Publication, M. L. Nielsen, B. Lavigne, and B. Dagens, Polarity-preserving SOA-based wavelength conversion at 40 Gb/s using bandpass filtering, Electron. Lett., 39, 2003, Y. Liu, E. Tangdiongga, Z. Li, S. Zhang, H. de Waardt, G. D. Khoe, and H. J. S. Dorren, 80 Gb/s wavelength conversion using a semiconductor optical amplifier and an optical bandpass filter, Electron. Lett., 41, 2005, E. Tangdiongga, Y. Liu, J. H. den Besten, M. van Geemert, T. van Dongen, J. J. M. Binsma, H. de Waardt, G. D. Khoe, M. K. Smit, and H. J. S. Dorren, Monolithically integrated 80-Gb/s AWG-based all-optical wavelength converter, IEEE Photon. Technol. Lett., 18, 2006, Y. Liu, E. Tangdiongga, Z. Li, S. Zhang, H. de Waardt, G. D. Khoe, and H. J. S. Dorren, Error-free alloptical wavelength conversion at 160 gb/s using a semiconductor optical amplifier and an optical bandpass filter, J. Lightwave Technol., 24, 2006, Y. Liu, E. Tangdiongga, Z. Li, H. dewaardt, A. M. J. Koonen, G. D. Khoe, X. Shu, I. Bennion, and H. J. S. Dorren, Error-free 320-Gb/s all-optical wavelength conversion using a single semiconductor optical amplifier, J.Lightwave Technol., 25, 2007,

41 1.74 E. Tangdiongga, Y. Liu, H. dewaardt, G. D. Khoe, and H. J. S. Dorren, 320-to-40- Gb/s demultiplexing using a single SOA assisted by an optical filter, IEEE Photon. Technol. Lett., 18, 2006, H. Chayett, S. Ben Ezra, N. Shachar, S. Tzadok, S. Tsadka, and J. Leuthold, Regenerative all-optical wavelength converter based on semiconductor optical amplifier and sharp frequency response, presented at the Optical Fiber Commun. Conf., Los Angeles, CA, Feb. 2004, Paper ThS S. Nakamura, K. Tajima, and Y. Sugimoto, 10 ps all-optical switching in novel Mach Zehnder configuration based on band-filling nonlinearity of GaAs, Conference on Lasers and Electro-Optics (CELO 94), CThS2, K. Tajima, All-optical switch with switch-off time unrestricted by carrier lifetime, Jpn. J. Appl. Phys., 32, 1993, L1746 L Preecha P. Yupapin; Suebtarkul Suchat, Entangled photon generation using a fiber optic Mach-Zehnder interferometer incorporating the nonlinear effect in a fiber ring resonator, Journal of Nanophotonics, 01(01), 2007, Khanthanou Luangxaysana, Somsak Mitatha, Masahiro Yoshida, Noriyuki Komine, Preecha P. Yupapin, High-capacity terahertz carrier generation using a modified adddrop filter for radio frequency identification, Opt. Eng. 51 (8) Dorren H.J.S., Lenstra D., Liu Y, Hill M. T., Khoe G.D.(2003), Nonlinear Polarization Rotation in Semiconductor Optical Amplifiers: Theory and Application to All-Optical Flip-FlopMemories, IEEE Journal of Quantum Electronics, 39(1), 2003,

42 1.81 Sisir Kumar Garai, A scheme of developing frequency encoded tristate-optical logic operations using Semiconductor Optical Amplifier, Journal of Modern Optics (Taylor and Francis),57(6),2010, Sisir Kumar Garai and Sourangshu Mukhopadhyay, Method of all-optical frequency encoded decimal to binary and BCD, binary to gray and gray to binary data conversion using semiconductor optical amplifiers, Applied Optics, 50(21), 2011, S. K. Garai and S. Mukhopadhyay, Method of implementing frequency encoded multiplexer and demultiplexer systems using nonlinear Semiconductor Optical Amplifiers, Optics and Laser Technology, 41(8), 2009,

43 CHAPTER-2 A method of implementing frequency encoded all-optical latch with semiconductor optical amplifier ABSTRACT This chapter includes a method of implementing frequency encoded all-optical NOT latch unit. Optical logic gates based on the principle of frequency conversion of some non-linear materials play the key role for the implementation of a frequency encoded data processing system. Memory is one of the most important units to implement the logic operations. Using semiconductor optical amplifier based switches and frequency encoding principle make the latch unit to be more faithful and reliable one. Work reported in this chapter was published in: S. Dutta and S. Mukhopadhyay, An all optical approach of frequency encoded NOT based Latch using semiconductor optical amplifier J Opt 39 (1), 39 45,

44 2.1 Introduction: Increasing demand for a faster and reliable data processor has given birth of the concept of all optical super-fast computers. In last few decades, there are several proposed optical and photonic devices, which can be run with operation speed far above the conventional GHz limit. Many of those devices have been dedicated for performing logic, arithmetic and algebraic operations to achieve the goal of all-optical computer and data processor [ ]. Among all other components, memory is an essential one for any data processor. Different types of volatile and non-volatile memories have already been developed successfully for the present electronic data processing systems. Similarly optical memories are also of great importance for the development of optical computing technologies. Many scientists are involved in deep research for the realization of digital volatile and non-volatile optical memories. Though some successes have been achieved in the last few decades to develop optical memories, flip-flop, bi-stable multivibrators and latches, still the ultimate goal of developing an optical computer has not been fulfilled yet. In this chapter, a methodology of developing optical NOT based latch using a specific non-linear behavior of semiconductor optical amplifiers (SOA) [ ] accommodating with frequency encoding technique are reported. This all-optical latch system is developed by the use of one special type of semiconductor optical amplifier based switch that is wavelength converter. Several types of memory units are reported earlier [2.11, 2.12]. For using SOA based switch and frequency encoding principle a reliable, super-fast and not attenuated operation is achieved. 34

45 2.2 Optical implementation of a NOT based latch: Memory is the basic requirement to construct any electronic or optical processor. To develop a complete unit of digital optical memory, the first step is to develop a latch or a 1-bit memory unit as it can store a single bit. The proposed system described here is based on frequency encoding principle (fig.2.1). Here two different frequencies are used for encoding 1 and 0 i.e. if ν 1 frequency represents the state 0 then ν 2 represents the state 1. A is the input terminal and Q and Q are the two output terminals. To implement the optical NOT based latch logic some beam splitters (BS), and mirrors (M) and some SOA based wavelength converters (WC) are used at different position of the system. Here two frequency selecting filters are used, where one is ν 1 optical pass filter and other is ν 2 optical pass filter. When ν 1 frequency is applied at A, the light beam enters only to WC 1, but not in WC 4. Thus the ν 1 frequency of light beam which behaves as a week probe beam now falls on to the WC 1 from the input terminal A. Another proper constant beam strong pump light (CW) of ν 2 frequency is injected to the WC 1. For which the ν 1 frequency of light can be obtained which is again injected as a strong pump beam to the input terminal of WC 2. A constant weak probe beam of frequency ν 2 (CW) is applied at the input of WC 2, for which an output of ν 2 frequency is obtained at the terminal Q for the application of ν 1 at A. Here CP s (cross polarizer s) are used in the output of every WCs to block the unwanted light beam i.e. for this CP at the time of absence of pump beam or probe beam no light beam will come out from the WCs. This output light beam is feedback as a pump beam to the WC 3 and a constant weak light of ν 1 frequency is also applied on the other input of the WC 3 to get the output of at ν 1 frequency which is divided into two parts by a beam splitter (BS). One part is feedback to the input of the WC 1 and 35

46 the other part is sent to the output terminalq. Thus ν 1 is obtained at A, ν 2 at Q and ν 1 atq. Now ν 2 frequency of a week probe beam of light is given at input A then it falls directly to the WC 4, but it will not at all enter at WC 1 and a constantly supplied strong pump beam of ν 1 frequency is applied on the other input channel of the WC 4. In this situation the ν 2 frequency of light will come out from the output of WC 4 which is again treated as a strong pump beam to fall on the WC 5. A constant week probe beam of frequency ν 1 is kept in the input channel of WC 5 which helps to get the intense ν 1 frequency of light at the output of WC 5 (X). This is divided again into two more parts; one is feedback to the input of the WC 6 as a strong pump beam where the other part comes at the output X, which is ultimately connected with Q by BS and M. Again a constant week probe beam of ν 2 frequency given to the 2nd input of WC 6 which ensures the output of ν 2 frequency at the output Y which is ultimately connected withq. A portion of the output at Y is feedback to the input of WC 4. This is the overall connection of the whole latch unit. Now to describe the operation it can be said that when ν 1 frequency of light i.e. logic state (0) is applied at the input terminal A, the upper portion of the system described in figure-2.1, activated but lower half does not because of presence of the ν 1 optical pass filter which only allows the ν 1 frequency of light in the upper half of the unit and the ν 2 pass filter only passes the ν 2 frequency of light in the lower half of the system described in figure-2.1. So when ν 1 frequency of light i.e. logic state (0) is given to input terminal A the ν 2 frequency of light is obtained at output Q and ν 1 frequency of light at Q i.e. Q =1 and Q =0 when A=0. Similarly when ν 2 frequency of light i.e. logic state 1 is applied at the input terminal A, the lower portion of the system 36

47 takes the major role instead of upper portion and the ν 1 frequency of light is obtained at Q and ν 2 at Q i.e. Q =0 and Q =1 when A=1. Now when ν 1 frequency of light is applied into the input terminal A, then light beam passes through the optical ν 1 pass filter and is applied into the WC 1 as a weak probe beam. Here a strong pump beam is already present at the input terminal of WC 1, so the ν 1 frequency of light is achieved at the output of WC 1 which is applied as a strong pump beam to the WC 2. Presence of ν 2 frequency as a weak probe beam at the input of WC 2 the conversion is obtained and the ν 2 frequency is achieved at the output terminal of WC 2. This light beam is divided into two parts, one part is applied as a strong pump beam into the WC 3 and another part comes at the outputq. Where as in presence of ν 1 frequency at the input of WC 3 the conversion is obtained and the ν 1 frequency is achieved at the output of WC 3. Here the light beam is divided into two parts, one is feedback to the WC 1 and another part comes to theq. So Q =1 and Q =0 when A=0. The most important and interesting point here is that if the input ν 1 and ν 2 frequency of light are withdrawn, the system will continue to show the last attended values of Q and Q at the final output because of the feedback mechanisms. At this situation the feedback light will continue to excite the respective WCs. So, this system can behave as a frequency encoded optical one bit memory. 37

48 Probe beam Pump beam CP M M ν 2 ν 1 WC 1 ν 1 ν 2 WC 2 CP BS ν 2 M Q ν 1 pass filter M M M ν 2 ν 2 ν 1 WC 3 CP ν 1 M ν 1 BS M Q A ν 1 or ν 2 BS M ν 1 CP M M ν 1 ν 2 WC 4 CP M ν 2 ν 1 WC 5 BS M ν 1 M X ν 2 ν 2 pass filter M ν 2 M ν 1 WC 6 CP M BS ν 2 ν 2 Y M Figure-2.1: Frequency encoded optical one-bit memory cell based on latch logic. (WC, M, CP, BS are represents the wavelength converter, mirror, cross polarizer and beam splitter or combiner respectively). 2.3 Optical implementation of two-bit memory cell: Now slightly extending the one-bit memory cell circuit a two-bit memory cell (fig.-2.2) can be developed. Here I have attached two similar optical circuits of one-bit memory unit. Here A and B are the input terminals and X 1 and Y 1 are the output terminals. The block diagram of this scheme is shown in figure-2.2. Here A * and B * are the two unit memory cells (latches), Q and Q are the output terminals of A * block and X 1 and Y 1 are the output terminals of B * block. X 1 is then combined with Q and Y 1 is combined withq. So, an overall two-bit memory cell is developed. Now two inputs are 38

49 A and B and two outputs are Q and Q, if A=1(ν 2 ) and B=0(ν 1 ) then Q =0(ν 1 ) and Q =1(ν 2 ) whereas if A=0(ν 1 ) and B=1(ν 2 ) then Q =1(ν 2 ) and Q =0(ν 1 ). The states of the last output will be attended if there is no light signal present in the input terminal i.e. for the withdrawal of optical signal, the system to continue to show its last attended values at the output. The truth table of this memory unit is given in table 2.1. INPUTS OUTPUTS A B Q Q ν 2 (1) ν 1 (0) ν 1 (0) ν 2 (1) ν 1 (0) ν 2 (1) ν 2 (1) ν 1 (0) No light No light Last state attended Last state attended Table-2.1 Truth table of optical two bit memory cell 39

50 BLOCK-A * Q A Output terminals Q Input terminals BLOCK-B * X 1 B Y 1 Figure-2.2: Block diagram of frequency encoded optical two-bit memory cell based on NOT latch logic. Here each block comprise the memory or latch as given fig

51 2.4 Conclusion: Frequency encoded technique based optical NOT latch or memory cell has been proposed here. The whole system is all-optical one and is expected to support a high speed operation (far above GHz limit) which is a potential advantage of this mechanism. The coded information (0 or 1) of an output signal remains unaltered in reflection, refraction, absorption etc. due to this encoding principle. Therefore this technique will be very much useful in reliable optical communication. To achieve the faithful amplification the pump beam of WC should lie between 4dB to 10 db. The proposed system does not only offer a high speed operation but also offers noise free conversion to provide a high signal to noise (S/N) ratio. Using this latch logic and the frequency encoded technique one can implement many other operations like digital types of flip-flops, multiplexer, demultiplexer etc. with some modifications of the scheme. 41

52 References: 2.1 Kuladeep Roy Chowdhury, S. Mukhopadhyay, A new method of binary addition scheme with massive use of non-linear material based system, Chinese Optics Letters, 1(4), April 20, 2003, Partha Ghosh, Partha Pratima Das, S. Mukhopadhyay, New proposal for optical flipflop using residue arithmetic, ITCOM-2001, (4534B-22), SPIE Proceedings (Optoelectronic and Wireless Data Management, Processing, Storage and Retrieval), 4534, 8 November, 2001, Archan Kumar Das, S. Mukhopadhyay, General approach of spatial input encoding for multiplexing and De-multiplexing, Optical Engineering (USA), 43(1), 1 January, 2004, Kuladeep Roy Chowdhury, S. Mukhopadhyay, Binary optical arithmetic operation scheme with tree architecture by proper accommodation of optical nonlinear materials, Optical Engineering, 43(1),,1 January, 2004, Nirmalya Pahari, Debendra Nath Das, S. Mukhopadhyay, All-optical method for the addition of binary data by non-linear materials, Applied Optics, 43(33), 2004, Nirmalya Pahari, S. Mukhopadhyay, An all-optical R-S flip-flop by optical nonlinear material, Journal of Optics, 34(3), 2005, Sisir Kumar Garai, Debajyoti Samanta, S. Mukhopadhyay, All-optical implementation of inversion logic operation by second harmonic generation and wave mixing character of some nonlinear material, Optics and Optoelectronic Technology (China), 6(4), 2008,

53 2.8 L. Q. Guo and M. J. Connelly, A poincare approach to investigate nonlinear polarization rotation in semiconductor optical amplifiers and its applications to all-optical wavelength conversion, Proc. of SPIE Vol.6783, 2007, (1-5). 2.9 H. J. Dorren, Daan Lenstra, Yong Liu, Martin, T. Hill, Giok-Djan Khoe, Nonlinear polarization rotation in semiconductor optical amplifiers: theory and application to alloptical flip-flop memories, IEEE Journal of Quantum Electronics, 39(1), Jan-2003, M. A. Karim and A. S. Awwal, John Wiley and Sons, Optical Computing An Introduction INC., S.K. Garai, S. Mukhopadhyay, A method of optical implementation of frequency encoded different logic operations using second harmonic and difference frequency generation techniques in non-linear material, Opt. Int. J. Light Electron. Opt. (2008), doi: / j.ijleo , 121(8), 2010, Bikash Chakraborty, S. Mukhopadhyay, Alternative approach of conducting phase-modulated all optical logic gates, Optical Engineering, 48(3), 2009,

54 CHAPTER-3 Some new approaches of conducting all-optical frequency/wavelength encoded logic operations, programmable logic units and RS flip-flops with semiconductor optical amplifiers ABSTRACT In this chapter includes the frequency encoded logic operations, programmable logic units and also RS flip-flop. In conduction of parallel logic, arithmetic and algebraic operations, optics has already proved its successful role. Since last few decades a number of established methods on optical data processing were proposed and to implement such processors different data encoding/decoding techniques have also been reported. Memory is the most important criteria of any fundamental computation and also communication. Here in this chapter I want to explore the memory unit and logic operation as well as the programmable logic operations also. Using this programmable logic system any types of logic operation have been obtained just changing the controlling signal within a one system. Work reported in this chapter was published in: 1. S. Dutta, S. Mukhopadhyay, Alternating approach of implementing frequency encoded all optical logic gates and flip-flop using semiconductor optical amplifier, Optik-Int. J. Light Electron Opt.(2010), 122 (2011)

55 2. S.Dutta and S.Mukhopadhyay, Application of Semiconductor Optical Amplifier for development of ultra fast programmable unit Frontiers in Materials Science-2010, December-06, National Institute of Technology Durgapur, INDIA, Paper was presented (poster) by S. Dutta on 6 th December and awarded for the best paper. 3. S.Dutta and S.Mukhopadhyay, A new approach of all-optical frequency encoded programmable unit en2c, 21 st January

56 3.1 Introduction: All optical signal and data processing is especially attractive for high speed and high capacity computation to avoid the speed related problems in optoelectronic processing systems. Because of the inherent character of parallelism of light can show more strong and potential applications in information processing, computing, data handling and image processing. In optical computation photon is found to be a very suitable information carrier than electron not only in the connection of super fast speed but in many other aspects of information processing also. Thus these photonic systems can successfully replace the electronic systems. Again it is also seen that in case of optical data processing the conventional methodologies can not be followed always as it is done in electronics. There are found several popular reports on the development of optical logical systems where the logic gates are the basic building blocks. In previous chapter I have focused about the NOT based memory latch unit. In this chapter I have propose frequency encoded optical logic gates, programmable logic unit and RS flip-flop. Several types of optical mechanism are there for implementing several types of optical logic gates and optical flip-flops [ ]. Those systems are like non-linear material, polarization based, phase encoded based etc. But those systems have some problems due to long distance, interaction etc. In this chapter I have used frequency encoding principle to overcome all problems and make the operations with greater accuracy. Different types of frequency encoded logic systems are reported earlier using different kinds of semiconductor optical amplifier based switches [ ]. In this chapter I have proposed alternative and new approach of frequency encoded logic operations, RS flip-flop and programmable logic operations in a very simple way using simple kind of 46

57 SOA based switches [3.13]. Some logic gates with controlled light signals are combined to assemble a programmable logic unit. By controlling the light signals of a programmable logic unit one can operate the corresponding logic operation and this is the beauty of this system. 3.2 Scheme of realization of frequency encoded optical OR logic operation: Optical logic gates are the basic building block to implement any optical logical functions or operations. OR gate is one of the most important and a basic unit of integral logical system. To develop the system some ADD/DROP multiplexers, wavelength converters, mirrors (M) and beam splitters (BS) are used which is shown in fig-3.1. Here the input beam A, and B may have either the frequency of ν 1 or ν 2 (wavelength λ 1 or λ 2 respectively), where ν 1 frequency of light is encoded for 0 state and ν 2 frequency of light encoded for 1 state. Now a light beam of frequency ν 1 or ν 2 from point A falls on the 1 st ADM which is tuned at the biasing frequency ν 2 then only ν 2 frequency of light is reflected by the ADM and is captured by the optical circulator C 1 and on the other hand the ν 1 frequency of light passes through the ADM and falls on the BS and one part of light falls on the 1st wavelength converter (WC) as a pump beam and another part of light falls on the 2 nd wavelength converter (WC) as a pump beam, so for the case of the 1 st WC if the probe beam is present then the WC works and a converted strong probe beam is obtained at the output which is the input of the 3 rd ADM and in absence of the probe beam no output is obtained. Now in the 2 nd WC there is a constant weak probe beam so if there is found a strong pump beam then a converted strong beam is obtained at the output which is injected again to the 3 rd WC as a pump beam. The captured tuned frequency of light from the first circulator C 1 of 1 st input terminal A is injected to the 3 rd ADM with 47

58 the help of some mirrors and beam couplers. Here for the 2 nd input terminal the similar type of input light (ν 1 or ν 2 frequency) is injected to the 2 nd ADM which is also tuned with the biasing frequency (ν 2 ). So only this frequency of light is reflected and captured by the circulator C 2 and given to the 1 st WC as a weak probe beam. Another frequency of light passes through the 2 nd ADM and falls in the 3 rd WC as a weak probe beam. If the pump beam and probe beam are present then only a strong probe beam is obtained at the output otherwise absence of any probe or pump beam no output can be generated. The output injected to the 3 rd ADM. So the final output result of frequency encoded OR logic operation is obtained from the terminal Y. The scheme is shown in figure Principle of operation of optical OR gate: Now when the ν 1 frequency of light is applied to the first input terminal A the light passes through the 1 st ADM and it is divided into two parts by the BS, where one part is treated as a strong pump beam to the 1 st RSOA and other part is delivered to the 2 nd RSOA also as a strong pump beam. One constant weak probe beam of ν 2 frequency is delivered to the 2 nd RSOA, so that at the output it gives a strong light beam of ν 2 frequency which is again applied as a strong pump beam to the 3 rd RSOA. Here a light of ν 1 frequency is applied to the second input terminal B which passes through the 2 nd ADM and delivered as weak probe beam to the 3 rd RSOA. This light beam which is emerged an output beam of ν 1 frequency light and is applied to the 3 rd ADM. Here to select the proper frequency of light one optical filter is kept. Again the 3 rd ADM is tuned at its biasing frequency ν 1 so it reflects the output beam from 3 rd RSOA and it is collected by the circulator and combined with the output terminal Y with the help of mirrors. Thus ν 1 frequency of light is obtained an output result. In this case (A=B= ν 1 ) there is no probe 48

59 beam in the 1 st RSOA, so 1 st RSOA can not work and the output beam is obtained from the 1 st RSOA. In the second case ν 1 frequency of light is given at the input terminal A and ν 2 frequency of light is to the second input terminal B. Here the 1 st ADM passes the light and it is divided into two parts by the BS where one part is given to the 1 st RSOA as a strong pump beam and other part is given also as a strong pump beam to the 2 nd RSOA and for which the output result of ν 2 frequency of light is obtained from the 2 nd RSOA, which serves as a strong pump beam to the 3 rd RSOA but as there is no probe beam in the 3 rd RSOA because B=ν 2, so no ν 1 frequency of light comes to the 3 rd RSOA and it does not convert any signal so no output result is obtained from the 3 rd RSOA. As the second input terminal B receives ν 2 frequency of light so the 2 nd ADM (tuned at its biasing frequency ν 2 ) reflects the light beam and it is collected by the circulator C 2 as shown in the figure-3.1. This light is applied as a weak probe beam to the 1 st RSOA. So the 1 st RSOA works and the output result of ν 2 frequency of light is obtained, which is delivered to the 3 rd ADM and I get the ν 2 frequency of light at the output terminal Y. In the third case when A=ν 2 and B=ν 1, 1 st ADM reflects the light beam and it is collected by the circulator C 1 and applied to the 3 rd ADM. Now the beam passes through the 3 rd ADM and the result of ν 2 frequency of light is obtained at the output Y. The rest of circuit can not take part in this conversion. Now when both the two input takes ν 2 frequency of light the 1 st ADM reflects the light beam and it is collected by the circulator C 1 and delivers it to the 3 rd ADM, and passed through this ADM and the converted light beam of frequency ν 2 is obtained at the output terminal Y. Similarly in the 2 nd ADM the light is reflected by the ADM and collected by the circulator C 2 and is delivered as a probe beam to the 1 st RSOA, but as there is no pump beam in the 1 st RSOA it can not work and no converted 49

60 light beam is obtained from it. So that Y= ν 2 when A = ν 1 and B=ν 2 ; when A=ν 2 and B= ν 1 ; Y=ν 2 ; when A=ν 2 and B=ν 1 ; Y=ν 2 and finally when A=B=ν 2 ; Y= ν 2. This verifies the truth table of OR gate which is shown in table-3.1, if ν 2 is encoded as 1 and ν 1 by 0. A B Y ν 1 (0) ν 1 (0) ν 1 (0) ν 2 (1) ν 1 (0) ν 2 (1) ν 1 (0) ν 2 (1) ν 2 (1) ν 2 (1) ν 2 (1) ν 2 (1) Table-3.1: Truth table of optical logic OR gate 50

61 ν 2 ADM 1 RSOA 1 ν 1 A 1 ADM 3 C1 C3 Y RSOA 2 ν 2 2 C CIRCULATOR B ADM 2 RSOA 3 3 OPTICAL PASS FILTER 1, 2 ν 2 and 3-ν 1 optical pass filter C2 PROBE BEAM PUMP BEAM MIRROR A, B INPUT TERMINALS Y OUTPUT TERMINAL Figure-3.1: Frequency encoded optical OR gate 3.3 Scheme of realization of frequency encoded Optical AND gate: To implement optical AND gate two channels are taken A and B and they may have either ν 1 (corresponding wavelength λ 1 ) frequency of light or ν 2 (corresponding wavelength λ 2 ) frequency of light. This whole system is shown in figure-3.2. Now a beam of light of frequency ν 1 or ν 2 falls on the 1 st ADM which is tuned in the biasing 51

62 frequency ν 2 so only ν 2 frequency of light is reflected back by ADM and captured by the circulator C 1 and the ν 1 frequency passing through the ADM is injected to the 1 st wavelength converter (WC) as a weak probe beam. A constant ν 2 frequency of strong pump beam of light is given to the 1 st WC. If both the pump and probe beam is present in the RSOA (WC) then the converted strong probe beam i.e. ν 1 frequency of light is achieved at the output. Now the reflected ν 2 frequency of light from the 1 st ADM is captured by the circulator C 1 and is reflected by the mirror and is injected as a weak probe beam to the 2 nd RSOA. Here from the channel B the input beam of light of frequency ν 1 or ν 2 falls on the 2 nd ADM which is also tuned at its biasing frequency ν 2. So it reflects the ν 2 frequency of light and passes the ν 1 frequency of light through it. This light beam merges with the probe beam of light of 1 st RSOA with the help of mirror. The reflected ν 2 frequency of light from the 2 nd ADM is captured by the circulator C 2 and is given to the 3 rd RSOA as a strong pump beam and ν 1 frequency of weak probe constant beam light is given to the 3 rd RSOA. So when both pump or probe beam is present a converted strong probe beam is obtained at its output. This output beam is injected to the 2 nd RSOA as a strong pump beam with the help of mirrors. So if both pump or probe beam is present in the 2 nd RSOA one gets the output beam as converted weak probe beam into a strong one which is added with the output beam of light of 1 st RSOA using some properly oriented mirrors and these two light beams again are injected to the 3 rd ADM which is tuned at its biasing frequency ν 2. So it passes the ν 1 frequency of light and reflects the ν 2 frequency of light and it is collected by the circulator C 3. This reflected beam of light is ultimately added with the output beam from 3 rd ADM by the use of properly oriented mirrors. The whole system is shown in figure-3.2. Here optical filters 52

63 are used to select the proper frequency of light beam. Thus when A=ν 1 and B=ν 1 ; Y=ν 1, when A=ν 1 and B=ν 2 ; Y= ν 1, for A=ν 2 and B=ν 1 ; Y=ν 1 and finally when A=B=ν 2 ; Y=ν 2. This satisfies the truth table of AND gate Principle of operation of optical AND gate: The AND logic system is shown in figure-3.2. Now in the case of AND logic when two input beams are of ν 1 frequency then 1 st ADM passes the light beam and 2 nd ADM also passes the light beam and then they are combined together by mirror and beam splitter. The combined beams are serving as a weak probe beam to the 1 st RSOA. A constant strong pump beam is present in the 1 st RSOA so it gives a converted output light beam which is delivered to the 3 rd ADM and it passes through it. Thus a result of ν 1 frequency of light is obtained at the output end Y. The conversion of 2 nd and 3 rd RSOA can not take part due to the absence of the either pump or probe beam. Again when A=ν 1 and B=ν 2, 1 st ADM passes the light beam and the light is delivered as a probe beam to the 1 st RSOA and similarly due to the above operation the converted output light beam of ν 1 frequency is obtained at the output end Y. The rest of the circuit does not take part in the conversion process due to absence of either pump beam or probe beam. Now when A=ν 2 and B=ν 1 then 1 st ADM reflects the light beam and the light is collected by the circulator C 1 and delivered as a weak probe beam to the 2 nd RSOA. As the pump beam is absent here so conversion can not take part in 2 nd RSOA. Again as B= ν 1, so the 2 nd ADM passes this light beam and is combined with the output terminal of the 1 st ADM and serves as a probe beam to the 1 st RSOA. Thus a similar conversion is occurred in this time and as the output light beam of frequency ν 1 is obtained at Y. Finally when the ν 2 frequency of light is applied to both input terminal A and B, 1 st ADM reflects the light beam and is 53

64 delivered as a probe beam to the 2 nd RSOA. So, 1 st RSOA can not take part any role in this conversion. Now as B=ν 2 so the 2 nd ADM reflects the light beam and the light is collected by the circulator, which is delivered again as a strong pump beam to the 3 rd RSOA where probe beam of ν 2 frequency is already present. So the conversion takes place and the output light beam of ν 1 frequency is obtained and it is delivered as a pump beam to the 2 nd RSOA, for which the output light beam of ν 2 frequency from 2 nd RSOA comes and it is applied to the 3 rd ADM. This light is reflected back by the ADM and collected by the circulator C 3. The light from the circulator is connected to the output terminal Y. So then the ν 2 frequency of light is obtained at the output Y. Thus when A=ν 1 and B=ν 1 ; Y=ν 1, when A=ν 1 and B=ν 2 ; Y= ν 1, when A=ν 2 and B=ν 1 ; Y=ν 1 and finally when A=B=ν 2 ; Y=ν 2. This satisfies the truth table of AND gate which is shown in Table-3.2, if ν 1 is encoded as 0 and ν 2 as 1. A B Y ν 1 (0) ν 1 (0) ν 1 (0) ν 1 (0) ν 2 (1) ν 1 (0) ν 2 (1) ν 1 (0) ν 1 (0) ν 2 (1) ν 2 (1) ν 2 (1) Table-3.2: Truth table of optical logic AND gate 54

65 ν 2 ADM 1 RSOA 1 ν 2 A 1 ADM 3 C1 C3 Y RSOA 2 2 ν 2 ADM 2 C CIRCULATOR B OPTICAL PASS FILTER 1, 3 -ν 1 and 2-ν 2 optical pass filter C2 RSOA 3 PROBE BEAM 3 PUMP BEAM MIRROR A, B INPUT TERMINALS Y OUTPUT TERMINAL Figure-3.2: Frequency encoded optical AND gate 3.4 Scheme of realization of frequency encoded Optical NAND gate: NAND gate is the most important logic gate in the logics family as it is universal gate. Again here two input channels are taken A and B as sources of input light beams of frequency ν 1 or ν 2. The whole set up is much closed to the AND logic set up except three 55

66 main changes seen in this system. Here the output light from 1 st ADM is injected as a pump beam instead of the probe beam to the 1 st RSOA and the reflected beam of light from 1 st ADM is injected as a pump beam instead of a probe beam to the 2 nd RSOA. Finally reflected beam from the 2 nd ADM is applied as a pump beam to the 3 rd RSOA (figure-3.3). Thus the truth table of the NAND logic is developed and it is shown in table So when A=B=ν 1 then Y=ν 2, when A= ν 1 and B= ν 2 then Y= ν 2, for A= ν 2 and B= ν 1 ; Y= ν 2 and finally for A=B= ν 2 ; Y= ν 1. This supports the truth table of universal gate NAND gate Principle of operation of optical NAND gate: The diagram of frequency encoded NAND logic is shown in figure-3.3. Now in the case of optical NAND logic gate at first ν 1 frequency light is applied in both of the inputs A and B. So ν 1 frequency of light falls on the 1 st ADM from A, and as it is tuned at its biasing frequency ν 2, so ν 1 frequency light passes through the ADM and is given as a strong pump beam to the 1 st RSOA, one weak probe beam of ν 2 frequency is also given to the 1 st RSOA for which a converted strong light beam of ν 2 frequency will be generated. This is again delivered to the 3 rd ADM which is tuned also at ν 2 frequency, so the light beam reflected from the 3 rd ADM is captured by the circulator C 3 and connected with the output Y by the use of mirrors and the ν 2 frequency of light is obtained at the output Y. Now ν 1 frequency light is also given to the 2 nd ADM from B and the ADM is tuned at same frequency ν 2, so the ADM passes the light beam which is combined with the output beam of the 1 st ADM by the mirrors etc. Now when A=ν 1 and B=ν 2 then similar operation goes on in the 1 st ADM, but in this case of second input terminal B, the light beam falls on the 2 nd ADM and gets reflected from it and being captured by the circulator C 2 it falls 56

67 on the 3 rd RSOA as a pump beam. So the converted ν 1 frequency of light is obtained in presence of ν 1 probe beam at the output and it is again applied to the 2 nd RSOA as a probe beam. Due to absence of pump beam the conversion does not occur and hence the ν 2 frequency of light is obtained at the output Y for the operation of the 1 st section of the system. When A=ν 2 and B=ν 1 then 1 st ADM reflects the light beam and it is then captured by the circulator C 1. This is given to the 2 nd RSOA as a strong pump beam. No other conversion takes place now because of the absence of the probe beam. Here ν 1 frequency of light falls on the 2 nd ADM and passing through it, this delivers a strong pump beam to the 1 st RSOA and the conversion is occurred (due to presence of ν 2 frequency probe beam of light) and one receives the ν 2 frequency light at the output and it is then delivered to the 3 rd ADM. Thus the final ν 2 frequency light is found at the output terminal Y. Finally when A=B=ν 2 ; 1 st ADM reflects the light beam and is delivered to the 2 nd RSOA as a strong pump beam. Again the light beam also falls on the 2 nd ADM and gets reflected. It is captured then by the circulator C 3 and is given to the 3 rd RSOA as a strong pump beam and one gets the converted output beam of light of ν 1 frequency in the presence of a weak probe beam of ν 1 frequency. This converted beam of ν 1 frequency of light is delivered as a probe beam again to the 2 nd RSOA and one thus gets the converted output beam of ν 1 frequency of light which is delivered to the 3 rd ADM. This 3 rd ADM passes it and a ν 1 frequency of light is obtained at the output end Y. Thus surely the logical NAND output is obtained from Y from the system described in fig-3.3. Hence one finds Y=ν 2 when A=B=ν 1, again when A= ν 1 and B= ν 2 then Y= ν 2, for A= ν 2 and B= ν 1, Y= ν 2 and finally when A=B= ν 2 then Y= ν 1. This supports the truth table of a universal frequency encoded gate NAND gate which is shown in table-3.3. Here ν 1 is encoded as 0 and ν 2 as 1. 57

68 ν 2 ADM 1 RSOA 1 ν 2 A 1 ADM 3 C1 C3 Y RSOA 2 2 ν 2 ADM 2 C CIRCULATOR B OPTICAL PASS FILTER 1-ν 2 and 2, 3 ν 1 optical pass filter C2 RSOA 3 3 PROBE BEAM PUMP BEAM MIRROR A, B INPUT TERMINALS Y OUTPUT TERMINAL Figure-3.3: Frequency encoded optical NAND gate 58

69 A B Y ν 1 (0) ν 1 (0) ν 2 (1) ν 1 (0) ν 2 (1) ν 2 (1) ν 2 (1) ν 1 (0) ν 2 (1) ν 2 (1) ν 2 (1) ν 1 (0) Table-3.3: Truth table of optical logic NAND gate 3.5. Scheme of realization of frequency encoded Optical R-S flip-flop: Memory is the fundamental criteria for developing an all of optical processor. To realize this system with optics I have considered two input light channels R and S, each of them may take either ν 1 (0) or ν 2 (1) frequency of light. The outputs light channels are Q and Q respectively. The outputs are feedback to the input i.e. it is connected with input R and Q is connected with the other input S by the help of some mirrors and beam splitters. The whole system is shown in figure-3.4. Now a beam of light of frequency ν 1 or ν 2 falls on the 1 st ADM through point R. As the ADD/DROP multiplexer is tuned with its biasing frequency ν 2 so it reflects this frequency of light and passes the ν 1 frequency of light. This light is then introduced to the 1 st RSOA as a strong pump beam. There already presents a constant weak probe beam of ν 2 frequency. So if both pump and probe beam is present the converted strong probe beam of light with the respective frequency is obtained which again falls on the 3 rd ADM. The absences of any pump or probe beam in the RSOA makes the conversion stop. Here one optical filter is used to select the proper output light beam. The 3 rd ADM is tuned with its biasing frequency ν 1 i.e. it reflects only ν 1 frequency of light and passes all other frequencies. The reflected beam is captured by 59

70 the circulator C 3 and added with the output Q by the mirrors. Now the reflected light beam from the 1 st ADM is separated by the circulator C 1 and is introduced to the 2 nd RSOA as a strong pump light beam. A weak probe beam is also given to the 2 nd RSOA and the converted output light beam passes through the respective optical filter and merges with the output beam of 1 st RSOA. Ultimately the output from the 1 st RSOA is given to the 3 rd ADM. Same process is happened with the second input S and finally two outputs Q andq are obtained as shown in figure-3.4. Now when one applies the light beam in the two input channels i.e. when R=ν 1 and S=ν 2 then Q= ν 2 and Q = ν 1 respectively and when R= ν 2 and S=ν 1 then Q=ν 1 and Q = ν 2 but when light is withdrawn from the inputs the last state is obtained in the outputs Q andq. So this follows the truth table of optical RS flip-flop which is shown in table Principle of operation of optical R-S flip-flop: To implement the circuit diagram of an optical RS flip-flop the light beam is supplied first to both the input channel. This delivered light is of frequency ν 1 to the 1 st ADM through the input end R. As the 1 st ADM is tuned at its biasing frequency ν 2, so ν 1 frequency light beam passes through it and falls on the 1 st RSOA where a strong beam of ν 2 frequency is already present. So one gets the converted ν 2 frequency of light beam at the output of 1 st RSOA and it is again introduced to the 3 rd ADM (tuning frequency ν 1 ) and light passing through the ADM comes to the output end Q. This output beam of ν 2 frequency of light splits into two parts by the BS, one part is sent to the output end Q and another part is feedback to the second input terminal S. Again if a ν 2 frequency light is applied to the S terminal, a combined beam of ν 2 frequency of light is delivered to the 2 nd ADM (tuning frequency ν 2 ). This ADM reflects by it and sends to the 4 th RSOA by the 60

71 help of circulator C 2 and mirrors. Due to presence of a weak probe beam of ν 1 frequency of light in 4 th RSOA, the conversion is occurred and the converted output light beam of frequency ν 1 is obtained which is introduced to the 4 th ADM (tuning frequency ν 1 ) and is reflected by it. This light is captured by the circulator C 4 and is combined with the output and splitted by the BS. One part goes to the other output end Q where another part is feedback to the input terminal R and the process continues. Now when the R=ν 2 and S=ν 1 then reverse process is developed in this circuit and one gets Q= ν 1 and Q = ν 2 in the output ends. Now when no light passes through both the input channel i.e. for the absence of light in the input terminals, last state is attended i.e. If the last state is Q= ν 1 and Q = ν 2 for (R=ν 2 and S=ν 1 ) then the same result is obtained in the output terminals. As because the output terminals are feedback to the reverse input terminals, so always there is found a light present in the input terminals (even if no external input is applied) and that s why the conversion process continues. So certainly when R=ν 1 and S=ν 2 then Q= ν 2 and Q = ν 1 respectively and when R= ν 2 and S=ν 1, Q=ν 1 and Q = ν 2, but when there is no input light beam is applied at all i.e. when the inputs are withdrawn the last state is attended in the output terminals Q andq. So the truth table of an optical RS flip-flop is followed, this is shown in table

72 ν 2 ν 1 R ADM 1 RSO A1 1 ADM 3 BS C1 C3 Q RSO A2 2 ν 2 ν 1 S ADM 2 RSO A3 3 ADM 4 C2 C4 Q BS RSO A4 4 C CIRCULATOR OPTICAL PASS FILTER 1, 3 ν 2 and 2, 4-ν 1 optical pass R, S INPUT TERMINALS Q, Q OUTPUT TERMINALS PROBE BEAM BEAM SPLITTER PUMP BEAM MIRROR Figure-3.4: Frequency encoded optical RS flip-flop 62

73 R S Q Q ν 1 (0) ν 2 (1) ν 2 (1) ν 1 (0) ν 2 (1) ν 1 (0) ν 1 (0) ν 2 (1) No light No light Last state attended Last attended state Table-3.4: Truth table of optical frequency encoded RS flip-flop 3.6. Frequency encoded programmable logic unit scheme: In this programmable logic unit five logic gates, half adder and half subtractor are present. Here seven control signals are present to activate their corresponding logic gates (AND, OR, NAND, XOR, NOT), half adder and half subtractor unit. Here ν 3 frequency of light represent the control signal of OR gate, ν 4 frequency of light represent the control signal of AND gate, ν 5 frequency of light represent the control signal of NAND gate, ν 6 frequency of light represent the control signal of NOT gate and ν 7 frequency represents the control signal of XOR gate. Two input channels are present for application of the input light signal and two output channels are present for obtaining the output data. The whole scheme in block diagram is shown in figure In this chapter first I have explained the five controlled logic gates and then I integrate these logic units and make a programmable logic unit Optical OR logic gate controlled by an optical signal: Now in this chapter first I explain the OR gate logic controlled by the light signal. To implement this logic gate some beam splitters, mirrors, optical channels and SOA 63

74 based switches like add/drop multiplexer and wavelength converters have been used. The whole scheme is shown in figure-3.5. Actually the whole circuit diagram of optical logic gate is similar to the simple uncontrolled OR logic gates but the only difference in this circuit is in this scheme one controlling signal CS 1 is present. Only then OR logic gate can be operate when CS 1 is present. Absence of this CS 1 OR logic operation can not be operating whereas input signals A and B are present. This CS 1 is the controlling channel through which one selected frequency is introduced. This selected frequency is controlling the whole OR gate operation. In this operation I choose ν 3 frequency of a light as a controlling signal Scheme of realization of optical OR logic gate controlled by the light signal: Now if A=ν 1, B=ν 1 and CS 1 =ν 3 then ADM 1 passes the ν 1 frequency and it is divided into two parts. One part is applied as a strong pump beam to WC 1 and another part is applied also as a strong pump beam to WC 3. Whereas incase of WC 1, ν 3 frequency which is comes from controlled channel CS 1 is applied as a weak probe beam to WC 1. So ν 3 frequency is obtained at the output of WC 1. This ν 3 frequency light beam is applied as a strong pump beam to the WC 2. But absence of probe beam WC 2 can not be conducted and no output is obtained from WC 2. Now ν 1 frequency comes through the input terminal B and passes to the ADM 2 and it is applied as a weak probe beam to the WC 4. Again one part of CS 1 signal is applied to the WC 1 and another part is applied as a weak probe beam to the WC 3. Here already ν 1 frequency of light is present as a strong pump beam so conversion is occurred and ν 3 frequency of light is obtained at the output of WC 3. This ν 3 frequency of light is applied as a strong pump beam to WC 4. So conversion is obtained and ν 1 frequency of light is obtained at the output of WC 4 and it is applied to the ADM 3. 64

75 It is tuned at a ν 1 frequency so this light signal reflected by the ADM 3 and collected by the circulator C 3 and the final output ν 1 frequency of light is obtained at terminal Y. So, when A=ν 1, B=ν 1 and CS 1 =ν 3 then Y=ν 1. Now when A=ν 2, B=ν 1 and CS 1 =ν 3 then ADM 1 reflected the ν 2 frequency and it is collected by the circulator. This light beam is applied to the WC 4 as a weak probe beam. Here CS 1 (ν 3 ) light signal is applied as a strong pump beam and conversion is obtained and ν 2 frequency of light is obtained at the output of WC 4 which is directly applied to the ADM 3 and the light beam passes through the ADM 3. Again ν 1 frequency from input terminal B passes through the ADM 2 and it is applied as a weak probe beam to the WC 4 whereas pump beam is absent in WC 3 so conversion does not obtain. So absence of pump beam in WC 4 conversion does not obtain. So ν 2 frequency of light is obtained at the final output terminal Y. So, when A=ν 2, B=ν 1 and CS 1 =ν 3 then Y=ν 2. Now when A=ν 1, B=ν 2 and CS 1 =ν 3 then the ADM 1 passes the ν 1 frequency and it is applied as a strong pump beam to the WC 1. So ν 3 frequency of light from the CS 1 is applied as a weak probe beam to the WC 1. Probe and pump beam are both present so conversion is obtained from WC 1 and ν 3 frequency of light is obtained at the output of WC 1 which is again applied to the WC 2 as a strong pump beam. Here ν 2 frequency of light (from the input channel B) is applied as a weak probe beam to the WC 2 and conversion is occurred. So ν 2 frequency of light is obtained at the output of WC 2 which is applied to the ADM 3 and it passes the light. Final output ν 2 frequency of light is obtained at the output terminal Y. So, when A=ν 1, B=ν 2 and CS 1 =ν 3 then Y=ν 2. When A= ν 2, B=ν 2 and CS 1 =ν 3 then the ADM 1 reflects the ν 2 frequency and it is collected by the C 1 and is applied to the WC 4 as a weak probe beam. Here ν 3 frequency 65

76 from CS 1 is applied as a strong pump beam to the WC 4. So, ν 2 frequency is obtained at the output of WC 4 which is applied to the ADM 3. This light beam passes through the ADM 3 and the final output ν 2 frequency is obtained at the output terminal Y. Others WCs does not take part in this operation due to the absence of any one light beam either probe or pump beam. So, when A= ν 2, B=ν 2 and CS 1 =ν 3 then Y=ν 2. These all conditions are satisfied the truth table of OR logic operation which is shown in table-3.5. Here if controlling signal CS 1 (ν 3 ) is not present then whole logic operation will be stopped. INPUT CHANNELS OUTPUT CHANNEL CS 1 A B Y ν 3 ν 1 (0) ν 1 (0) ν 1 (0) ν 3 ν 1 (0) ν 2 (1) ν 2 (1) ν 3 ν 2 (1) ν 1 (0) ν 2 (1) ν 3 ν 2 (1) ν 2 (1) ν 2 (1) No light signal ν 1 (0) ν 1 (0) No Result No light signal ν 1 (0) ν 2 (1) No Result No light signal ν 2 (1) ν 1 (0) No Result No light signal ν 2 (1) ν 2 (1) No Result Table-3.5: Truth table of frequency encoded optical controlled OR gate 66

77 ν 2 ADM 1 WC 1 WC 2 ν 1 A ADM 3 C1 WC 4 C3 Y CS 1 (ν 3 ) WC 3 ν 2 ν 5 CONTROL LIGHT FREQUENCY ADM 2 WC 4 B C CIRCULATOR PROBE BEAM C2 PUMP BEAM MIRRORS AND BEAM SPLITTERS A, B INPUT TERMINALS Y OUTPUT TERMINAL Figure-3.5: Frequency encoded optical OR logic operation controlled by external light signal Optical AND logic gate controlled by the light signal: Now I have explained of operation of optical AND logic controlled by an optical signal. Whole operation is similar to optical AND logic operation but only difference is the controlling signal CS 2. This is an optical channel in which one particular frequency ν 4 67

78 is chosen for controlling light signal. This light signal controlled the whole operation. The scheme is shown in figure Scheme of realization of optical AND logic operation controlled by the light signal: In the first case, when A=ν 1, B=ν 1 and CS 2 =ν 4 then ν 1 frequency from input A passes through the ADM 1 and it is applied as a weak probe beam to input of WC 1. Here ν 4 frequency from controlled channel CS 2 is divided in to two parts by the beam splitters and one part is applied as a strong pump beam to the WC 2 and another part is applied as a weak probe beam to the input of WC 1. So, presence of both pump and probe beam in WC 1 the ν 1 frequency is obtained at the output of WC 1. This light beam passes through the ADM 3 and ν 1 frequency is obtained at the final output terminal Y. Due to the absence of either pump or probe beams into others wavelength converters they does not take any par in this operation. So when A=ν 1, B=ν 1 and CS 2 =ν 4 then Y=ν 1. When A=ν 2, B=ν 1 and Ct 2 =ν 4 then ν 2 frequency from input A is reflected by the ADM 1 and it is collected by the circulator C 1 and is applied to the WC 3 as a weak probe beam. Again one part of ν 4 frequency from CS 2 is applied as a strong pump beam to the input of WC 1. Now ν 1 frequency from input B passes through the ADM 2 and falls on the input of WC 1 as a weak probe beam. So presence of both pump and probe beam ν 1 frequency is obtained at the output of WC 1. This light beam passes through the ADM 3 and the ν 1 frequency light is obtained at the final output terminal Y as shown in the figure-3.6. Due to the absence of either pump or probe beams into others wavelength converters they does not take any part in this operation. So when A=ν 2, B=ν 1 and CS 2 =ν 4 then Y=ν 1. 68

79 In the 3 rd case, when A=ν 1, B=ν 2 and CS 2 =ν 4 then ν 1 frequency passes through the ADM 1 and is applied as a weak probe beam to input of the WC 1. Again one part of ν 4 frequency from CS 2 is introduced to the input of WC 1 as a weak probe beam. Presence of both pump and probe beam ν 1 frequency is obtained at the output of WC 1 which is passes through the ADM 3 and finally ν 1 frequency of light is obtained at the final output terminal Y. Due to the absence of either pump or probe beams into others wavelength converters they does not take any par in this operation. So when A=ν 1, B=ν 2 and CS 2 =ν 4 then Y=ν 1. When A=ν 2, B=ν 2 and CS 2 =ν 4 then ν 2 frequency from input A is reflected from the ADM 1 and collected by the circulator C 1. This light beam is applied as a weak probe beam to the WC 3. One part of ν 4 frequency from CS 2 is applied as a strong pump beam to WC 2. Here ν 1 frequency of probe beam is already present so ν 1 frequency is obtained at the output of WC 2 and it is applied as a weak probe to the WC 4. Again ν 2 frequency from input B is reflected from ADM 2 and is collected by the circulator C 2. This light beam is applied as a strong pump beam to the input of WC 4 so presence of both pump and probe beam in WC 4 the ν 1 frequency is obtained at the output of WC 4. This light beam is applied as a strong pump beam to the WC 3. In the presence of both pump and probe beam the ν 2 frequency of light is obtained at the output of WC 3 which is reflected by the ADM 3 and collected by the circulator C 3. So, ν 2 frequency is obtained at the final output terminal Y. When A=ν 2, B=ν 2 and CS 2 =ν 4 then Y=ν 2. The truth table of AND gate is shown in table

80 ν 2 ADM 1 WC 2 ν 2 A ADM3 CS 2 (ν 4 ) C1 C3 Y WC 3 WC 1 ν 2 ADM 2 C CIRCULATOR B ν 5 CONTROL LIGHT FREQUENCY C2 WC 4 PROBE BEAM PUMP BEAM MIRROR AND BEAM SPLITTERS A, B INPUT TERMINALS Y OUTPUT TERMINAL Figure-3.6: Frequency encoded optical AND logic operation controlled by external light signal 70

81 INPUT CHANNELS OUTPUT CHANNEL CS 2 A B Y ν 4 ν 1 (0) ν 1 (0) ν 1 (0) ν 4 ν 1 (0) ν 2 (1) ν 1 (0) ν 4 ν 2 (1) ν 1 (0) ν 1 (0) ν 4 ν 2 (1) ν 2 (1) ν 2 (1) No light signal ν 1 (0) ν 1 (0) No Result No light signal ν 1 (0) ν 2 (1) No Result No light signal ν 2 (1) ν 1 (0) No Result No light signal ν 2 (1) ν 2 (1) No Result Table-3.6: Truth table of frequency encoded optical controlled AND gate Optical NAND logic gate controlled by the light signal: To implement this logic operation some optical wavelength converters, mirrors, beam splitters and add/drop multiplexer are needed. This operation is similar to normal NAND operation but only difference is controlled light signal CS 3 (ν 5 ). This particular light frequency ν 4 is controlled the whole NAND logic operation. The whole scheme is shown in figure Scheme of realization of optical NAND logic operation controlled by the light signal: In this circuit arrangements A and B are the inputs and y is the output terminal. Here controlling light signal is CS 3 (ν 5 ). Now when A=ν 1, B= ν 1 and CS 3 =ν 5 then ν 1 light beam from A passes through the ADM 1 and is applied as a strong pump beam to the 71

82 WC 1. Again ν 5 light beam from CS 3 is divided into two parts one part is applied as a strong pump beam to the WC 3 and another part is also applied as a strong pump beam to the WC 4. In presence of ν 2 frequency light in WC 3 conversion is occurred and ν 2 frequency is obtained at the output of the WC 3 which is applied as a weak probe beam to the WC 1. Here in WC 1 conversion is obtained in presence of both pump and probe beam and ν 2 frequency is obtained at the output of WC 1. This light beam is applied to the ADM 3 and it is reflected from the ADM. Then it is collected by the circulator C 3 and ν 2 frequency is obtained at the final output terminal Y. Other WCs does not conducted in this operation due to absence of either pump or probe beam. So when A=ν 1, B= ν 1 and CS 3 =ν 5 then Y=ν 2. When A=ν 2, B=ν 1 and Ct 3 =ν 5 then ν 2 frequency from A is reflected by the ADM 1 and it is collected by the C 1 and is applied as a strong pump beam to the WC 2 as is describing in the figure-3.7. But absence of any probe beam WC 2 does not work. Again ν 1 frequency from B passes through the ADM 2 and is applied as a strong pump beam to the WC 1. Here ν 2 frequency from the output of WC 3 is present so conversion is happened and ν 2 frequency light is obtained at the output of WC 1 which is reflected by the ADM 3 and collected by the C 3. Now ν 2 frequency of light is obtained at the final output terminal Y. Other WCs does not conducted in this operation due to absence of either pump or probe beam. So when A=ν 2, B= ν 1 and CS 3 =ν 5 then Y=ν 2. Again when A=ν 1, B=ν 2 and CS 3 =ν 5 then ν 2 frequency from B is reflected by the ADM 2 and is collected by the C 2. This light beam is applied as a strong pump beam to the WC 5. Again ν 5 frequency from CS 3 is divided into two parts. Those are applied as strong pump beam to the WC 3 and WC 4 respectively. In WC 3 ν 2 frequency is present as a weak 72

83 probe beam and in WC 4 ν 1 frequency is present as a weak probe beam. So conversion is obtained in both WCs and ν 2 frequency is achieved at the output of WC 3 and ν 1 frequency at the output of WC 4. This ν 1 frequency is applied as a weak probe beam to the WC 5. In presence of both pump and probe beam ν 1 frequency is obtained at the output of WC 5 which is applied as a weak probe beam to the WC 2. But absence of pump beam conversion does not occurred. Again ν 2 frequency from WC 3 is applied as a weak probe beam to the WC 1 and ν 1 frequency from A is applied as a strong pump beam to the WC 1. So conversion is obtained and ν 2 frequency of light is obtained at the output of WC 1 which is introduced to the ADM 3 and reflected by it. This reflected beam is collected by the C 3 and ν 2 frequency of light is obtained at the final output terminal Y. So when A=ν 1, B= ν 2 and CS 3 =ν 5 then Y=ν 2. Now when A=ν 2, B=ν 2 and CS 3 =ν 5 then ν 2 frequency from A is reflected by the ADM 1 and is collected by the C 1. This light beam is applied as a strong pump beam to the WC 2. Again ν 2 frequency from B is reflected by the ADM 2 and is collected by the C 2. This light beam is applied as a strong pump beam to the WC 5. Here in WC 5 ν 1 frequency (from WC 5 ) is applied as a weak probe beam. So conversion is obtained and ν 1 frequency is obtained at the output of the WC 5 which is applied as a weak probe beam to the WC 2. In presence of both pump and probe beam in WC 2 ν 1 frequency is obtained at the output of WC 2 and this light beam passes through the ADM 3. So ν 1 frequency of light is obtained at the final output Y. So, when A=ν 2, B=ν 2 and CS 3 =ν 5 then Y= ν 1. This whole logic system can be operating only when this controlling signal is present. If I withdraw the signal then this NAND logic system can not operate. The truth table of NAND operation is shown in table

84 ν 2 ADM 1 WC 1 ν 2 A ADM3 CS 3 (ν 5 ) C1 WC 3 C3 Y WC 2 WC 4 ν 2 ADM 2 C CIRCULATOR B ν 7 CONTROL LIGHT FREQUENCY C2 WC 5 PROBE BEAM PUMP BEAM MIRROR A, B INPUT TERMINALS Y OUTPUT TERMINAL Figure-3.7: Frequency encoded optical NAND logic operation controlled by external light signal 74

85 INPUT CHANNELS OUTPUT CHANNEL CS 3 A B Y ν 5 ν 1 (0) ν 1 (0) ν 2 (1) ν 5 ν 1 (0) ν 2 (1) ν 2 (1) ν 5 ν 2 (1) ν 1 (0) ν 2 (1) ν 5 ν 2 (1) ν 2 (1) ν 1 (0) No light signal ν 1 (0) ν 1 (0) No Result No light signal ν 1 (0) ν 2 (1) No Result No light signal ν 2 (1) ν 1 (0) No Result No light signal ν 2 (1) ν 2 (1) No Result Table-3.7: Truth table of optical frequency encoded controlled NAND gate Optical NOT logic gate controlled by the light signal: To implement the NOT gate I have used some ADM s, WCs, mirrors and beam splitters. Here ν 6 is the control light frequency. Again A is the input terminal and Y is the output terminal. Here CS 4 is the controlling light signal terminal. This scheme is shown in figure Scheme of realization of optical NOT logic operation controlled by the light signal: Now when A=ν 1 and CS 4 =ν 6 then ν 1 frequency of light is applied through the input terminal A into the ADM 1 which is tuned at ν 2 frequency then ADM 1 passes this input ν 1 frequency of light and introduce as a strong pump beam to the WC 1. Here ν 6 frequency is applied to the WC 2 as a strong pump beam and ν 2 frequency of week probe 75

86 beam is already present in the input of WC 2 then the ν 2 frequency of light is obtained at the output of WC 2 and it is applied as a week probe beam to the WC 1. The WC 1 is activated and the ν 2 frequency of light is obtained at the output of WC 1 and after passing through the ADM 2 the ν 2 frequency of light is obtained at the output terminal Y. But if ν 6 frequency of light is absence then the total conversion process will be stopped. Now when A=ν 2 then ADM 1 reflects it and it is applied to WC 4 and again ν 6 frequency of light is present it serves as a strong pump beam to WC 3 where ν 1 frequency of week probe of light is present so conversion is obtained and the ν 1 frequency of light is obtained at the output of WC 3 which is applied as a week probe beam to the WC 4 and the ν 1 frequency of light is obtained at the output of WC 4. This light beam is applied to the ADM 2 which is tuned at ν 1 frequency so it reflects the light beam and the ν 1 frequency of light beam is obtained at the output terminal Y by using C 1 and mirrors. But absence of ν 6 frequency of light the whole process will be stopped. So when A=ν 1 then Y=ν 2 and A=ν 2 then Y=ν 1, only when ν 6 frequency of light is present. It satisfies the truth table of NOT gate which is shown in table-3.8 and ν 6 frequency of light serve as a control light signal. INPUT CHANNELS OUTPUT CHANNELS CS 4 A Y ν 6 ν 1 (0) ν 2 (1) ν 6 ν 2 (1) ν 1 (0) No light signal ν 1 (0) No Result No light signal ν 2 (1) No Result Table-3.8: Truth table of frequency encoded optical controlled NOT gate 76

87 ν 2 ν 1 A ADM 1 ADM 2 WC 1 Y C C CS 4 (ν 6 ) WC 2 ν 6 CONTROL LIGHT FREQUENCY WC 3 PROBE BEAM WC 4 PUMP BEAM C CIRCULATOR MIRRORS AND BEAM SPLITTERS A Y INPUT TERMINAL OUTPUT TERMINAL Figure-3.8: Frequency encoded optical NOT logic operation controlled by external light signal Optical XOR logic gate controlled by the light signal: Similarly to implement the optical XOR logic operation some ADMs, WCs, mirrors, beam splitters and optical channels are used. Here ν 7 frequency is used as a controlling light signal. This scheme is shown in figure Scheme of realization of optical XOR logic operation controlled by the light signal: In this operation system A and B are the input terminals. Y is the output terminal and CS 5 is the controlling light signal terminal. Now when I have introduced ν 1 frequency 77

88 at the A terminal, ν 1 frequency at the B terminal and ν 7 frequency at the CS 5 terminal then ν 1 frequency from A passes through the ADM 1 and divided into two parts by the use of beam splitters. One part is applied as a strong pump beam to the WC 3 and another part is applied as a weak probe beam to the WC 1. But due to absence of any strong pump beam and also probe beam in WC 1 and WC 3 respectively they are not working. Here ν 1 frequency from B passes through the ADM 2 and is divided into two parts. One part is applied as a strong pump beam to the WC 8 and another one is applied also as a strong pump beam to the WC 7. Due to absence of weak probe beam in WC 7 it does not work. Again ν 7 frequency from CS 5 is divided into two parts both parts are applied as a strong pump beam to WC 5 and WC 6 respectively as described in the figure-3.9. In presence of ν 2 frequency weak probe beam in WC 5 conversion process is obtained and ν 2 frequency is obtained at the output of WC 5. This ν 2 frequency light beam is divided into two parts one part is applied as a weak probe beam to the WC 2 and another one is applied as weak probe beam to the WC 8. In presence of ν 1 frequency weak probe beam in WC 6 conversion process is obtained and ν 1 frequency is obtained at the output of WC 6. So in presence of both pump and probe beam in WC 8 ν 2 frequency is obtained at the output of WC 8 which is applied as weak probe beam to the WC 3. Both pump and probe beam are present in WC 3. Then ν 2 frequency is obtained at the output of WC 3 which is applied as a strong pump beam to the WC 4. Output ν 1 frequency beam from WC 6 is divided into two parts both part are applied as a weak probe beam to WC 4 and WC 9 respectively. Both pump and probe beam are present in the WC 4. So, ν 1 frequency is obtained at the output of WC 4. This light beam passes through the ADM 3. So, ν 1 frequency is obtained at the final output terminal Y. It can be said that when A=ν 1, B=ν 1 and CS 5 = ν 7 then Y=ν 1. This is 78

89 satisfying the truth table of XOR logic operation. But in absence of CS 5 (ν 7 ), this whole operating system does not work. Now when ν 1 frequency is applied in terminal A, ν 2 frequency is applied in B and ν 7 frequency is applied in terminal CS 5 then ν 1 frequency passes through the ADM 1 and is divided into two parts. One part is applied as a weak probe beam to the WC 1 and another part is applied as a strong pump beam to the WC 3. Again ν 2 frequency from B is reflected from the ADM 2 and is collected by the circulator C 2. This light beam is divided into two parts one part is applied as a strong pump beam to the WC 1 and another one is applied also as a strong pump beam to the WC 9. So, presence of both pump and probe in WC 1 conversion is obtained and ν 1 frequency is obtained at the output of WC 1. This ν 1 is applied as a strong pump beam to the WC 2. Here ν 2 frequency from the output of WC 5 is applied as a weak probe beam. So, ν 2 frequency is obtained at the output of WC 2. This light beam is reflected by the ADM 3 and is collected by the C 3 and ν 2 frequency is obtained at the final output Y. So It can be said when A=ν 1, B=ν 2 and CS 5 = ν 7 then Y=ν 2. This is satisfying the truth table of XOR logic operation. But in absence of CS 5 (ν 7 ), this whole operating system does not work. Again ν 2 frequency is applied in terminal A, ν 1 frequency is applied in terminal B and ν 7 frequency is applied in terminal CS 5 then ν 2 frequency from A is reflected by the ADM 1 and it is collected by the C 1. This light beam is divided into two parts. One is applied as a weak probe beam to the WC 3 and another is also applied as a weak probe beam to the WC 7. Again ν 1 frequency from B passes through the ADM 2 and is divided into two parts as stated before. Both are applied as a strong pump beam to the WC 7 and WC 8. Here ν 2 frequency present as a weak probe beam in WC 7 so conversion is obtained 79

90 and ν 2 frequency is obtained at the output of WC 7. This ν 2 frequency is applied to the ADM 3 and is reflected by the ADM 3. This reflected beam is collected by the C 3 and ν 2 frequency is obtained at the final output terminal Y. So, when A=ν 2, B=ν 1 and CS 5 = ν 7 then Y=ν 2. This is satisfying the truth table of XOR logic operation. But in absence of CS 5 (ν 7 ), this whole operating system does not work. In the last consideration when ν 2 frequency is applied into both terminal A and B and ν 7 frequency at the terminal CS 5. Then ν 2 frequency from A is reflected by the ADM 1 and is collected by the C 1. As stated before this light beam is divided into two parts, both parts are applied as a weak probe beam to the WC 3 and WC 7 respectively. Again ν 2 frequency from B is reflected by the ADM 2 and is collected by the C 2. This light beam is divided into two parts and both are applied as a strong pump beam to the WC 1 and WC 9 respectively. In presence of both pump and probe beam in WC 9 ν 1 frequency is obtained at the output of WC 9 which is applied as a strong pump beam to WC 3. Here ν 2 frequency as a weak probe beam is present so conversion is obtained and ν 2 frequency is obtained at the output of WC 3. This frequency is again applied as a strong pump beam to the WC 4. Due to presence of ν 1 frequency probe beam in WC 4 conversion is obtained and ν 1 frequency is obtained at the output of WC 4 which is applied to the ADM 3. This light beam passes through the ADM 3 and ν 1 frequency is obtained at the final output terminal Y. So, when A=ν 2, B=ν 2 and CS 5 = ν 7 then Y=ν 1. This is satisfying the truth table of XOR logic operation which is shown in table-3.9. But in absence of CS 5 (ν 7 ), this whole operating system does not work. 80

91 A CS 5 (ν 7 ) C ADM 1 W C 5 W C 1 W C 3 W C 2 W C 4 C ADM 3 Y W C 6 C CIRCULATOR PROBE BEAM PUMP BEAM B C ADM 2 W C 7 MIRRORS and BEAM SPLITTERS W C 8 A, B INPUT TERMINALS W C 9 Y ν 7 OUTPUT TERMINAL CONTROLE LIGHT FREQUENCY Figure-3.9: Frequency encoded optical XOR logic operation controlled by external light signal 81

92 INPUT CHANNELS OUTPUT CHANNEL CS 5 A B Y ν 7 ν 1 (0) ν 1 (0) ν 1 (0) ν 7 ν 1 (0) ν 2 (1) ν 2 (1) ν 7 ν 2 (1) ν 1 (0) ν 2 (1) ν 7 ν 2 (1) ν 2 (1) ν 1 (0) No light signal ν 1 (0) ν 1 (0) No Result No light signal ν 1 (0) ν 2 (1) No Result No light signal ν 2 (1) ν 1 (0) No Result No light signal ν 2 (1) ν 2 (1) No Result Table-3.9: Truth table of frequency encoded optical controlled XOR gate Scheme of realization of frequency encoded programmable logic unit: To implement this programmable logic unit some optical channels, mirrors and beam splitters have been used. The whole block diagram is shown in figure Here in this figure two common input channel. This channels are A and B. Channel A is divided into five parts and is applied to the respective inputs of logic gates and B also is divided into five parts and is applied to the respective inputs of the logic gates. In this system five controlling terminals are CS 1, CS 2, CS 3, CS 4 and CS 5. The outputs of five logic gates are combined in one final output channel Y. So when I need to operate any one gate I have to apply respective frequency at the inputs A and B and the corresponding controlling signal then the output result is obtained. As for example, if the NAND logic operation has been operated then they have to give the corresponding frequency to the controlling terminal 82

93 CS 3 (ν 5 frequency) and just maintain the inputs according the NAND logic gate. If their inputs frequencies are present but controlling signal is not present then this system does not work. So, just controlling these control signal terminals one can get any types of logic operation in the output terminal. The whole truth table is shown in table INPUTS OUTPUT A B CS 1 CS 2 CS 3 CS 4 CS 5 Present Present Present OR Present Present Present AND Present Present Present NAND Present Present Present NOT Present Present Present XOR Table-3.10: Truth table of optical programmable logic system 83

94 CONTROL LIGHT SIGNAL TERMINALS CS 1 CS 2 CS 3 CS 4 CS 5 I N P U T A OR AND NAND NOT XOR T E R M I N A L S B Y OUTPUT TERMINAL Figure-3.10: Block diagram of frequency encoded optical programmable unit 3.7 Important requirements for the switching of SOA: To obtain the faithful operation the power of the input control light beam which, serves as a pump beams for the SOA should lie between 2 and 4 db. The performance of the data transfer depends on the pomp beam energy. Energy of each probe beam is to be maintained between -4 and -2 db. The wavelengths of the selected inputs are 1555 and 1550 nm corresponding to frequency ν 1 and ν 2 respectively. Frequencies of CW input signal should lie in C band ( nm). This wavelength range is favorable for optical communication. 84

95 3.8 Conclusion: To conclude, an all optical approach for the successful realization of high speed (far above GHz range) optical logic gates, memories and programmable logic operations have been proposed here. The potential advantage of these optical gates and flip-flop over many other established optical gates was the use of frequency encoding technique, for which the coded information (0, 1) in a signal remains unchanged in refraction, reflection, absorption etc. for a long distance transmission of data. The proposed system could offer also a noise free conversion to provide a high signal to noise (S/N) ratio. Using such frequency encoded technology. In this chapter not only about the logic operations have been discussed but also the programmable logic operations have been discussed. Using this programmable unit one can get any types of logic operation just only controlling the external controlling light signal. It is very helpful for any computation and communication systems. 85

96 References: 3.1 N. Pahari, S. Mukhopadhyay, An all-optical R S flip-flop by optical non-linear material, J. Opt., 34 (3), 2005, J.M. Jeong, M.E. Marhic, All-optical logic gates based on cross-phase modulation in a non-linear fiber interferometer, Opt. Commun., 85 (5 6), 1991, B.K. Jenkins, A.A. Sawchuk, T.C. Strand, R. Forchheimer, B.H. Soffer, Sequential optical logic implementation, Appl. Opt. 23 (19), 1984, T.A. Ibrahim, R. Grover, L.-C. Kuo, S. Kanakaraju, L.C. Calkoun, P.-T. Ho, All optical AND/NAND logic gates using semiconductor microresonators, IEE Photonics Technol. Lett., 15 (10), 2003, P. Ghosh, P.P. Das, S. Mukhopadhyay, New proposal for optical flip-flop using residue arithmetic, in: ITCOM-2001, (4534B-22), SPIE Proceedings (Optoelectronic and Wireless Data Management, Processing, Storage and Retrieval), 4534, 8 November, 2001, A.K. Das, S. Mukhopadhyay, General approach of spatial input encoding for multiplexing and De-multiplexing, Opt. Eng. (U.S.A.), 43, 2004, Y. Ichioka, J. Tanida, Optical parallel logic gates using a shadow-casting system for optical digital computing, Proc. IEE, 72 (7), 1984, K. E. Zoiros, M. K. Das, D. K. Gayen, H. K. Maity, T. Chattopadhyay, J. N. Roy, All-optical pseudorandom binary sequence generator with TOAD-based D flip-flops, Optics Communications, 284(19), 2011,

97 3.9 S.K. Garai, S. Mukhopadhyay, Method of implementing frequency encoded multiplexer and demultiplexer systems using nonlinear semiconductor optical amplifiers, Opt. Laser Technol. 41 (8), 2009, S.K. Garai, S. Mukhopadhyay, Method of implementation of all-optical frequency encoded logic operations exploiting the propagation characters of light through semiconductor optical amplifiers, J. Opt. (2009), doi: /s S.K. Garai, A. Pal, S. Mukhopadhyay, All-optical frequency encoded inversion operation with tristate logic using reflecting semiconductor optical amplifiers, Optik (2009), doi: /j.ijleo S.K. Garai, S. Mukhopadhyay, Amethod of optical implementation of frequency encoded different logic operations using second harmonic and difference frequency generation techniques in non-linear material, Optik Int. J. Light Electron. Opt. (2008), doi: /j.ijleo H.J. Dorren, D. Lenstra, Y. Liu, M.T. Hill, G.-D. Khoe, Nonlinear polarization rotation in semiconductor optical amplifiers: theory and application to all-optical flip-flop memories, IEEE J. Quantum Electron., 39, 2003,

98 CHAPTER-4 A new approach of implementing all-optical frequency/wavelength encoded clocked S-R flip-flop ABSTRACT In all optical networking and computing system, the role of all-optical flip-flops is very much essential. For signal synchronization with a reference clock and for storage of digital bits the flip-flop has no alternative. In this chapter I have proposed a method of developing an all optical frequency encoded clocked RS flip-flop using the non-linear character of semiconductor optical amplifiers. Frequency is the basic character of light and several encoding/decoding problems in computations and communications can be solved using the frequency encoding principle of optical data. The proposed system is alloptical and therefore it can extend a super fast speed of operation (far above THz limit). Work reported in this chapter was published in: S. Dutta, S. Mukhopadhyay, A new alternative approach of all optical frequency encoded clocked S R flip-flop exploiting the non-linear character of semiconductor optical amplifiers, Optik - Int. J. Light Electron Opt., 123 (2012)

99 4.1. Introduction: In the previous chapter the logic operations, RS flip-flop and programmable logic unit have been discussed. In this chapter I am interested how to make frequency encoded better memory unit and how to overcome the problems related with RS flip-flop. That s why here I have proposed clocked SR flip-flop with SOA based optical switches. Alloptical signal processing has drawn much attention of scientific communities because of its inherent parallelism and potential applications in high speed optical networks, optical computing systems etc. Again all-optical signal processing is especially useful to overcome the high bit rate problems in future communication systems. In recent years a lot of efforts have been seen in this area where optics is tactfully used for the processing of digital data. Several all-optical digital devices have been proposed which are supposed to run with the operation speed far above the GHz range [ ]. Many of those devices are dedicated for performing logic gates, flip-flops, optical buffers, and arithmetic operations to achieve the goal of all-optical computer/data processor. In particular, optical flip-flop can attract a special interest as because it can serve as an optical memory. Optical memory has also of great impact for the development of optical packet switches in networks. Now several types of optical flip-flops are proposed with different types of encoding mechanisms in last few years which are polarization encoding, phase encoding, intensity encoding and also frequency encoding [ ]. In this chapter a novel alternative approach of frequency encoded clocked SR flipflop based on SOA based switches has been proposed. Several types of all optical memory units have been reported earlier. Among these memory units some are 89

100 polarization encoded, some are phase encoded, some are intensity encoded and some are frequency encoded [ ]. There also some frequency encoded memory units which have already been reported by different scientists. MZI-SOA switches, add-drop multiplexers and PBSs were used there to implement the optical memory units and as well as the tristate logic gates. Also all-optical frequency encoded one bit memory unit and two- bit memory unit have been proposed earlier. These are based on SOA made switches, wavelength converters and add/drop multiplexers [ ]. The frequency encoded memory unit is not run by clock signal there. In digital optical communication and computation this frequency encoded flip-flop will take an important role. This type of encoding is chosen because of very high signal to noise (S/N) ratio and very low bit error problem Optical implementation of clocked S-R flip-flop: The frequency encoded optical RS flip-flop and the functions of SOA based switches are elaborately discussed in the previous chapters. Now to introduce a clock based memory operation I have propose a new concept of frequency encoded optical clocked RS flip-flop. To implement this clocked system some SOA based wavelength converters (WC), add/drop multiplexers (ADM), optical filters, mirrors (M) and beam splitters (BS) have been used. Now the proposed clock based RS flip-flop is given by a schematic diagram shown in fig-4.1 and the truth table is shown in table-4.1. Here one light source (serving as a clock signal) used as the clock (CLK) terminal. One optical filter which only passes the ν 2 frequency of light is placed in front of the clock (CLK). So this clock channel will only send the ν 2 frequency of light, when the CLK is 1. S and R are the two input terminals. Now if ν 2 (1) frequency of light is applied to the CLK and S= 90

101 ν 1 (0), and R= ν 2 (1), then ν 1 frequency of light from the terminal S falls on the ADM 1 which is tuned at frequency ν 2 and therefore the light passes through the ADM 1 and it is applied to the WC 1 as a pump beam. Again ν 2 frequency of light from CLK passes through the optical filter and is splitted into three parts by the two beam splitters. One part is applied to the WC 1 as a probe beam where as the second part is applied to the WC 4 also as a probe beam. Third part is applied to the WC 3 as the pump beam. Now for WC 1 both the pump and probe beam is present then the ν 2 frequency of light is obtained at its output which is directly applied to the input R 1 terminal to the unlocked RS flip-flop and the ν 1 frequency of light is obtained at the output terminalq (according to the principle of R-S flip-flop). This RS flip-flop operation is already discussed in our previous chapter. The figure of RS flip-flop is shown in figure-3.4 and truth table of this is shown in table For the WC 3 a constant probe beam (of frequency ν 1 ) is applied in its input then both pump and probe beam is present the ν 1 frequency of light is obtained from the output of the WC 3 which is divided again into two parts by the use of a BS. One part is applied as a probe beam to the WC 2, but due to absence of pump beam the conversion can not be obtained. Another part of the output from WC 3 is applied to the WC 5 as a probe beam. Again as ν 2 frequency of light is applied at the input terminal R and it falls on the ADM 2 (which is tuned also at ν 2 frequency), so it blocks the passage of ν 2 through it and reflects the light which is received by the optical circulator. It then serves as a pump beam to the WC 5. So in presence of both pump beam and probe beam the conversion is obtained at WC 5 and the ν 1 frequency of light is obtained at the output of WC 5 which goes to the S 1 terminal of RS flip-flop. Thus the ν 2 frequency of light is received at the output terminalq according to the principle of R-S flip-flop (table-3.4 of previous chapter). 91

102 Here absence of pump beam does not support any conversion WC 4. Thus when clock is ν 2 (1) and S=ν 1 (0) and R= ν 2 (1) the output of the whole flip-flop Q = ν 1 (0) and Q = ν 2 (1). When the clock is at the frequency ν 2 (1) and S= ν 2 (1), and R= ν 1 (0) then WC 2, WC 3 and WC 4 are active and take part in conversion process and input of the RS flip-flip is achieved of block A as, R 1 = ν 1 (0) and S 1 = ν 2 (1), which produces Q = ν 2 (1) and Q = ν 1 (0) respectively at the final outputs. Now when clock= ν 2 (1) but the light beams are withdrawn from the input terminals S and R then the clock system will not affect the WCs and they stop the conversion, but due to the feedback mechanism in the R-S flipflop, the outputs Q and Q will attend with its last achieved values. Similarly when CLK is set at o, i.e. when ν 1 frequency is applied at the clock terminal the filter will not pass any signal to the WCs. So instead of the presence of R-S inputs the WCs will not support the conversion process. Thus R 1 and S 1 will get no signal from R and S for this reason Q and Q will continue with its last attended values. The truth table of optical clocked SR flip-flop is shown in table-4.1. Thus it can seen that when the clock is applied to the flipflop the outputs (Q, Q ) give the result according to the truth table, but when no clock is applied or ν 1 (0) frequency is applied at the clock terminal the Q, Q outputs of the flipflop holds its last attended values. 92

103 S ADM 1 WC 1 R 1 Q C WC 2 CLK F WC 3 BLOCK-A R ADM 2 WC 4 S 1 Q C WC 5 PUMP BEAM BLOCK-A: OPTICAL RS FLIP-FLOP PROBE BEAM F OPTICAL ν 2 PASS FILTER C CIRCULATOR Figure-4.1: Optical clocked SR flip-flop, with SOA based switches 93

104 Clock S R Q Q ν 2 (1) ν 1 (0) ν 2 (1) ν 1 (0) ν 2 (1) ν 2 (1) ν 2 (1) ν 1 (0) ν 2 (1) ν 1 (0) ν 2 (1) No light No light Last state attended ν 1 (0) No light No light Last state attended Table-4.1: Truth table of optical clocked S-R flip-flop Last attended Last attended state state 4.3 Conclusion: The method of optical implementation of clocked SR flip-flop with all optical switching systems is illustrated here. The speed of operation can go far above the THz limit as the SOA based switches operate at this speed. A high speed operation of the proposed scheme is not the only a prospective advantage of the system, but the frequency encoding mechanism also offers a great support. The most important application of the system can be seen in digital communication, where the frequency encoded new data or an (old data depending on the applied clock) have been send in the communication channel. Even if the sender requires making a data to be continued for communication or a new data is to be introduced in the channel, he can easily do it by the use of the above system. The output signal from the clocked SR flip-flop can be sent to distant receiver 94

105 as it remains unaltered in reflection, refraction, absorption etc due to the nature of coding of bits (0 or 1) with frequency variation of light. Therefore this technique will be very much useful for conducting a reliable and faithful optical memory both in communication and computation. To achieve a good amplification the pump beam of WCs should lie between 4dB to 10 db. The proposed system does not only offer a high speed operation but it also offers an operation which provides a high signal to noise (S/N) ratio. The accommodation of frequency encoding process is the main reason for obtaining the high S/N ratio. For this reason bit error rate also goes down in comparison to conventional intensity based encoding processes. This optical clocked SR flip-flop and frequency encoding technique can be used for many other optical devices where clocked SR flipflop is an essential unit. 95

106 References: 4.1 M.T. Fatehi, K.C. Wasmundt, S.A. Collins, Optical flip-flops and sequential logic circuits using a liquid crystal light valve, App. Optics 23, 1984, M.T. Hill, H. de. Waardt, G.D. Khoe, H.J.S. Dorren, All-optical flip-flop based on coupled laser diodes, IEEE J. Quantum Elect. 37 (3), 2001, W. Wu, S. Campbell, S. Zhou, P. Yeh, Polarisation encoded optical logic operations in photorefractive media, Opt. Lett. 174, 1993, S. Dutta, S. Mukhopadhyay, All optical frequency encoding method for converting a decimal number to its equivalent binary number using tree architecture, Optik, 122, 2011, D. Samanta, S. Mukhopadhyay, A method of maintaining the intensity level of a polarization encoded light signal, J. Phys. Sci. (Vidyasagar Univ.) 11, 2007, B. Chakraborty, S. Mukhopadhyay, Alternative approach of conducting phase modulated all optical logic gates, Opt. Eng. 48 (3), 2009, S. Dutta, S. Mukhopadhyay, An all optical approach of frequency encoded NOT based Latch using semiconductor optical amplifier, J. Opt. 39 (1), 2010, S. Dutta, S. Mukhopadhyay, Alternating approach of implementing frequency encoded all-optical logic gates and flip-flop using semiconductor optical amplifier, Optik, 122, 2011, D. Samanta, S. Mukhopadhyay, Implementation of an optical S-R flip-flop with polarization encoded light signal, Optoelectron. Lett. 5, January 1, 2009, N. Mitra, S. Mukhopadhyay, A new scheme of an all-optical J-K flip-flop using nonlinear material, J. Opt. 37 (3), 2008,

107 4.11 S.K. Garai, S. Mukhopadhyay, A novel method of developing all-optical frequency encoded memory unit exploiting nonlinear switching character of semiconductor optical amplifier, Opt. Laser Technol. 42 (5), 2010, A. Mecozzi, Small-signal theory of wavelength converters based on cross-gain modulation in semiconductor optical amplifiers, IEEE Photon. Technol. Lett. 8, Z. Li, G. Li, Gates based on four-wave mixing in a semiconductor optical amplifier, IEEE Photon. Technol. Lett. 18, 2006, A. Kumpera, P. Honzathko, R. Slavik, Novel 160-GHz wavelength converter based on a SOA and a long period grating, Opt. Commun. 282, 2009, G. Raybon, U. Koren, B.I. Miller, M. Chien, M.G. Young, R.J. Capik, K. Dreyer, R.M. Derosier, A wavelength-tunable semiconductor amplifier/filter for add/drop multiplexing in WDM networks, IEEE Photon. Technol. Lett. 9, 1997,

108 CHAPTER-5 A new method for transmission of frequency encoded parallel optical data ABSTRACT Semiconductor optical amplifier (SOA) is a well known non-linear device which can exhibit Tera Hertz switching speed of operation. SOA based switching, therefore, has wide application in fiber optic communication. In this chapter a new concept of frequency encoded parallel data transmission with SOA for optical communication has been proposed. To achieve the transmission SOA is suggested for the generation of frequency encoded/decoded parallel data. It converts initially an intensity encoded optical data to frequency encoded one; whereas at the receiving end it again returns the intensity encoded data from frequency encoded one. Work reported in this chapter was published in: S. Dutta, S. Mukhopadhyay, A new approach of parallel data transmission through optical waveguide with SOA based frequency encoding/decoding technique, Optik - Int. J. Light Electron Opt., 123 (2012)

109 5.1 Introduction: In the previous chapters the several types of logic operations and clocked and unclocked memory units have been discussed. In this chapter I extend the area of work and interested to implement the frequency encoded parallel data transmission for the communication system. To support the increasing demand and rapid growth information, optics has been proved as proper alternative for very high speed communication with high bit rate and low bit error rate [ ]. In optical communication network the response time at the nodes is a very important issue for setting high speed communication. For managing the tremendously increasing day to day data traffic, it is very necessary to enhance the transmission link capacity as well as the speed of the switching networks at the nodes. The realization of a network node with throughput at the order of 100 Gb/s is not far away. SOA grating combination has been successfully used for 160 GHz wavelength conversion. Here a phase encoded signal is converted to amplitude modulated signal. Again using a cross-correlation system and non-linear polarization rotation 200 GBPS wavelength conversion at temporal resolution at 1.5 PS is also reported. Some logic gates are also implemented based on the four-wave mixing character of SOA with the mechanism of polarization shift keying [5.5] and with tri-state operation logic [5.6]. Habib Fathallah et al proposed the concept of a high bandwidth optical communication with fast optical frequency-hop code division multiple access (FFH-CDMA) system [5.7]. The encoding and decoding are done by all-fiber device. Here different frequencies of light are encoded as different logic state. Here I focus a principle for the successful realization of an efficient optical data transmission based on the frequency encoding principle, which can be used as a more reliable one than 99

110 other conventional encoding principle towards the achievement of super-fast optical communication and data processing [ ] A new method of frequency encoded parallel data transmission through optical waveguide: This communication system requires some specific frequencies (eight bit data string) which represent the bit 1 of each position of the data respectively. For example λ 1 represent the least significant bit (LSB) if it is 1 whereas λ 8 represents the most significant bit if it is 1 state. In this way all the other bits are represented by other frequencies. The absence of light signal represents the logic state 0. To implement the whole system which is shown in fig-5.1, eight SOA based wavelength converters, eight SOA based Add/Drop multiplexers, some beam couplers and mirrors are used. The conversion method of the SOA based wavelength converter is discussed earlier. Now using eight wavelength converters position-wise (e.g. for the first data bit WC 1 is used) the new mechanism of data transmission can be implemented. The pump beam of frequency ν 0 (corresponding wavelength λ 0 ) is used in the other input terminals of all the position-wise arranged WCs. The constant probe beams of light of wavelengths λ 1, λ 2, λ 3, λ 4, λ 5, λ 6, λ 7, λ 8 are also applied to the respective WCs position-wise. The outputs of the WCs are coupled by the beam couplers. This couplers output can be introduced to the input of the optical fiber. If an eight bit data represented as is to be sent through the fiber through this proposed encoder, the light intensity of wavelength λ 0 is applied as pump beams to WC 1, WC 4, WC 6 and WC 7 which produces the signal λ 1 λ 4 λ 6 and λ 7 in the coupled output and finally this coupled beam comprised of λ 1, λ 4, λ 6 and λ 7 are introduced to the fiber. Following the same process any eight bit data can be encoded 100

111 as frequency encoded data and this frequency encoded data can be transmitted through the fiber. In the receiving end eight Add/Drop multiplexers (ADM) are required for decoding of frequency encoded data to intensity encoded one. Add/Drop multiplexers (ADM) are arranged position-wise and each multiplexer is tuned for one particular frequency by the application of proper bias current. ADM 1 tuned at its biasing frequency ν 1, ADM 2 at ν 2, ADM 3 at ν 3 and all the other ADMs at ν 4 to ν 8 respectively. The output obtained from the receiving end of the optical fiber is applied to the input of ADM 8. Each ADM has one circulator to collect the selected frequency of light. When the light beam having frequency ν 1 to ν 8 applied to an ADM the respective frequency of light will be reflected from the respective ADM for which it is tuned. The other frequencies will easily pass through the ADM. Thus the reflected signals (having different frequencies) are collected by the optical circulators. The output from the circulators gives the same wavelengths in decoded data, which were sent originally at the input side. This data then is passed parallely and bitwise through 8 wavelength converters. The probe beams of the wavelength converters are at λ 0. These wavelength converters returns the original intensity encoded data with exactly position wise. In this output data all the 1 s are represented by presence of light at the wavelength λ 0 and 0 s by the absence of light. The whole process can be illustrated by an example. Let an eight bit data is used for encoding and decoding. For this light intensity of wavelength λ 0 is applied as pump beam to WC 1, WC 2, WC 6, WC 7 and WC 8 only. Constant probe beams are also applied to all the WCs. But the conversion takes place only in WC 1, WC 2, WC 6, WC 7, WC 8 whereas WC 3, WC 4, WC 5 are not working for conversion. Thus the output beam of wavelength λ 1 is obtained from WC 1, λ 2 from WC 2, λ 6 from WC 6, λ 7 from WC 7 and λ 8 101

112 from WC 8 and these light beams are coupled by the beam couplers. At the receiving end the coupled beam is introduced to the ADM 8. This ADM 8 reflects the light having λ 8 wavelength and passes λ 1, λ 2, λ 6 and λ 7 to the ADM 7 which reflects the light of frequency λ 7 wavelength and sends λ 1, λ 2 and λ 6 to the ADM 6 which again reflects the light of λ 6 wavelength. Similarly ADM 2 reflects the light having λ 2 wavelength and ADM 1 reflects the light having λ 1 wavelength. ADM 3, ADM 4, ADM 5 give no reflected light. These output lights of different frequencies from the ADMs are applied to the eight wavelength converters (WC 9, WC 10, WC 11, WC 12,..WC 16 ) as a pump beams and constant λ 0 probe beams are also applied to the WCs. So the output is obtained as λ 0 λ 0 λ 0 000λ 0 λ 0 i.e. the transmitted data is received at the output. Thus any eight bit data can be sent parallel with the system as described in fig.5.1. The system described in fig.5.1 enables of transmitting eight bit data. For a sixteen bit data transmission 16 ADMs are required for making the mechanism active. Any data of high number of bits can be sent in parallel following this mechanism. For another example a binary data string as ( ) can be taken to sent parallel through an optical fiber. In fig-5.2 the step by step result is described. In fig 5.2(a) the parallel data string is shown, whereas in fig 5.2(b) the bit wise intensity encoded parallel data string is depicted.in fig 5.2(c) the bit wise wavelength encoded data is shown, which is obtained from the series of wavelength converters as given in fig 5.1. This wavelength encoded data is sent through the optical fiber for parallel communication. The data obtained at the outlet of the fiber is again sent through the ADMs and wavelength converters, which are used for decoding the data. The wave length converters at the receiving end reconvert the frequency encoded data to bitwise 102

113 intensity encoded data, which is shown in fig 5.2(d). The received data string is shown in fig 5.2(e). 5.3 Essential requirements for implementation of the practical transmission of data: The essential requirements for setting a good response from a SOA based optical switch is pump beam for wavelength conversions should lie between 2 db to 4 db. An optical filter can be used just after the SOA converters which only select the desired probe beam frequency of light at the output. The 3 db bandwidth of the filter should be in the order of 1 nm. The performance of the data transfer depends upon the used pump beam energy. The intensity level maintained at the each probe beam should be between - 4 to -2 db. Again it is very important to mention that the wavelength of the both pump and probe beam should lie in C band ( nm). Based on all the above aspects some wavelengths in C band are proposed for the consideration of pump beam and eight different probe beams for encoding the bits of an eight bit byte. This is given in table-5.1. Pump beam λ nm Probe Beams (nm) λ 1 λ 2 λ 3 λ 4 λ 5 λ 6 λ 7 λ Table-5.1: Some proposed wavelengths in C band for encoding the eight different 1 bits of a byte and also for that of the probe beams 103

114 Input terminals for pump beam (λ 0 ) λ 8 λ 7 λ 2 λ 1 WC 8 WC 7 WC 2 WC 1 λ 8 λ 7 λ 2 λ 1 ADM 8 ADM 7 ADM 2 ADM 1 C λ 0 C λ 0 λ C 0 λ 0 C... WC 9 WC 10 WC 15 WC 16 Output string of bits (λ 0 represents 1 and no light represents 0) Represents Beam coupler Represents Optical fiber C Represents Circulator Figure-5.1: Frequency encoded data transmission method based on SOA switching. 104

115 (a) λ λ λ 3 λ λ 4 λ 5 λ 6 7 (b) λ 8 Intensity (c) Bit position Intensity λ 1 λ 3 λ 2 λ 4 λ 5 λ 7 Frequency encoded light signal at the input of the optical fiber (d) 1.0 λ 0 λ 0 λ λ 0 λ 0 λ 0 0 λ 0 λ 0 λ 6 λ Intensity (e) Position-wise output from the wavelength converters Figure-5.2 Graphical outputs of the frequency encoded data at different stages of the encoding/decoding system (a) Encoded intensity encoded eight bit data string, (b) Intensity distribution for the position-wise bits at different frequencies at initial stage, (c) Intensity distribution of the bit-wise coded light signals at the input of the optical fiber, (d) Intensity distribution of the position wise bits at the final output of the decoding system, (e) Decoded intensity encoded eight bit data string

116 5.4. Conclusion: The above method exhibits a method for parallel transmission of data bits in case of optical communication through fiber. The SOA takes the role of conversion of an intensity encoded data to a frequency encoded one by the exploitation of its wavelength conversion character. SOA can achieve the THz speed of operation for this conversion process. This proposed method can, therefore, ensures a very high speed optical communication over many other conventional communication techniques. If a data accommodates eight bits or sixteen bits, then all the bits can be sent in parallel through the optical fiber. It is important to mention here the wavelength all the wavelengths (λ 1 to λ 8 ) should be selected at the c-band for setting the best conversion efficiency as well as for low loss communication. The arrangement of SOA based wavelength converters and ADMs in fig 1 are given in such a way that the form of the originally intensity encoded data at the input of the fiber is maintained at the receiving end of the communication system. At the final output end all the data bits are represented either by 1 s if an intensity of light with wavelength λ 0 presents otherwise by 0 s for no light. 106

117 References: 5.1 A. Kumpera, P. Honzathko, R. Slavik, Novel 160-GHz wavelength converter based on a SOA and a long period grating, Opt. Commun. 282, 2009, J. Sakaguchi, T. Nishida, Y. Ueno, 200-Gb/s wavelength conversion using a delayedinterference all-optical semiconductor gate assisted by nonlinear polarization rotation, Opt. Commun. 282, 2009, C. Zhang, K. Qiu, B. Xu, Y. Ling, A novel all-optical label processing based on multiple optical orthogonal coded sequences for optical packet switching networks, Opt. Commun. 281, 2008, A. Argyris, D. Syvridis, L. Larger, V. Annovazzi-Lodi, P. Colet, I. Fischer, J. Garcia- Ojalvo, C.R. Mirasso, L. Pesquera, K. Alan Shore, Chaos-based communications at high bit rates using commercial fibre-optic links, Nature 438, 2005, Z. Li, G. Li, Gates based on four-wave mixing in a semiconductor optical amplifier, IEEE Photon. Technol. Lett. 18, 2006, S.K. Garai, A scheme of developing frequency encoded tristate-optical logic operations using Semiconductor Optical Amplifier, J. Mod. Opt. (Taylor and Francis) 57 (6), 2010, H. Fathallah, A. Rusch, S. Larochelle, Passive optical fast frequency-hop CDMA communication system, J. Lightwave Technol. 17, 1999, S.K. Garai, S. Mukhopadhyay, Method of implementing frequency encoded multiplexer and demultiplexer systems using nonlinear semiconductor optical amplifiers, Opt. Laser Technol. 41, 2009,

118 5.9 T. Durhuus, B. Mikkelsen, C. Joergensen, S. Lykke Danielsen, K.E. Stubkjaer, Alloptical wavelength conversion by semiconductor optical amplifiers, J. Lightwave Technol. 14, 1996, D.A.O. Davies, Small-signal analysis of wavelength conversion in semiconductor laser amplifiers via gain saturation, IEEE Photon. Technol. Lett. 7, 1995, E. Iannone, R. Sabella, L. De Stefano, F. Valeri, All-optical wavelength conversion in optical multicarrier networks, IEEE Trans. Commun. 44, 1996, K. Obermann, S. Kindt, D. Breuer, K. Petermann, C. Schmidt, S. Diez, H.G. Weber, Noise characteristics of semiconductor-optical amplifiers used for wavelength conversion via cross-gain and cross-phase modulation, IEEE Photon. Technol. Lett. 9, 1997, M.F.C. Stephens, D. Nesset, K.A. Williams, A.E. Kelly, R.V. Penty, I.H. White, M.J. Fice, Wavelength conversion at 40 Gbit/s via cross-gain modulation in distributed feedback laser integrated with semiconductor optical amplifier, Electron. Lett. 35, D.D. Marcenac, A.E. Kelly, D. Nesset, D.A.O. Davies, Bandwidth enhancement of wavelength conversion via cross-gain modulation by semiconductor optical amplifier cascade, Electron. Lett. 31, 1995,

119 CHAPTER-6 A new approach of developing Universal all-optical multiplexer with frequency encoding mechanism ABSTRACT Multiplexing and demultiplexing are essential in any networking system. Different encoding and decoding schemes have different advantages in optical multiplexing and demultiplexing for all optical systems. Frequency encoding technique is established as a faithful and reliable one among different encoding processes. Intensity encoded, frequency encoded and many other multiplexers were proposed earlier. In all those multiplexers the numbers of multiplexed input channels depend on the number of triggering channels. In this chapter a novel concept of frequency encoded universal multiplexer which can deal any number of input channels with the use of a single triggering channel has been described. 109

120 6.1. Introduction: In any information processing system as well as in communication networking system, multiplexer and demultiplexer are key elements. Different types of optical multiplexer and demultiplexer with various encoding mechanisms are reported by several scientists [ ]. Some frequency encoded multiplexers are reported with tri-state system [6.7]. In this system some specials types SOA based optical switches BSOA has been used. In previous studies the frequency encoded logic units, memory units and parallel transmission of data have been studied. In this chapter a novel approach of implementing frequency encoded universal triggered multiplexer using the semiconductor optical amplifier based optical switches has been proposed. To implement this system frequency encoded method but with Boolean logic system and SOA based wavelength converters and some optical pass filters have been used [ ]. Multiplexer means many into one, as it follows the combinational logic to connect one output channel with one of many input channels. This is selected by the signals in the triggering channels. In multiplexer, it has one or many signals in triggering channels and these triggering signals select the proper input channel to be connected with the output. A multiplexer having n numbers of triggering channels can accommodate 2 n input channels. In general Boolean triggering is used in conventional multiplexer and demultiplexrer systems. The novelty of this system is that any number of input channels can be accessed by a single triggering channel. So using this special multiplexer any one can access enormous number of input channels, if the triggering channels are increased. The frequency encoding mechanism and SOA made optical switches are exploited full to extend their advantages for implementation of such universal multiplexer. 110

121 6.2. Method of developing a frequency encoded universal multiplexer with SOA: To implement the universal optical multiplexing system I have used some optical pass filters (which passes selected one frequency of light and blocks others), some beam splitters (BS), beam couplers (BC), mirrors (M), optical channels and SOA based switches like wavelength converters (WC). The whole scheme is shown in figure-6.1. According to figure-6.1 T is the single triggering channel. Here any one of n numbers of frequencies ν 0, ν 1, ν 2, ν 3, ν 4,..ν n is used to select one from n inputs. There are n pass filters (PF) where each can select a specific frequency of light. In fig-6.1, block-1 is a ν 0 pass filter, block-2 is ν 1 pass filter, and block-3 is ν 2 pass filter and so on. Here I 1, I 2,.I n are the input channels and O is the output channel. Now only when triggering signal 0 i.e. ν 0 frequency is applied in the optical triggering channel and it is passed through the block-1 due to presence of optical pass filters which only passes the ν 0 frequency of light, then only the ν 0 frequency of light is applied into the WC 1 as a strong pump beam. Here the I 1 input channel which consumes a weak probe beam i.e. when the system is triggered by ν 0 frequency of light, the pump beam is present in only WC 1 and for this I 1 is connected to the output. Therefore the frequency of light applied at I 1 passes to the output cannel O. So the input signal of I 1 is transmitted directly to the output being controlled by the triggering signal ν 0. Similarly when ν 1 frequency of light is present in the triggering channel T then only I 2 input signal is transmitted to the output as ν 2 being filtered by the block-2 comes to WC 2 as a pump beam and sends I 2 (i.e. the frequency of I 2 ) to the output port O. In such a way a specific frequency (λ i, i=1.n) applied to the triggering channel T activates the specific WC i and helps I i to be transmitted to the output. These n numbers of frequencies in the triggering channel select n inputs 111

122 respectively. It is important to mention that when a specific WC works others remain nonfunctional because of absence of pump beams to those WCs. T (ν 0 to ν n ) BS M M n-1 n M M M s M WC 1 I 1 M s BC M O WC 2 I 2.. BC M s I n-1 WC n-1 BC M M s I n WC n BC M Figure-6.1: Architecture of all-optical single triggered universal multiplexer (M, BS, WC represents the mirror, beam splitters and wavelength converter). 112

123 6.3. Method of developing double triggering universal multiplexer: To implement a double triggered universal multiplexer some other sets of optical beam splitters, beam couplers, mirrors, wavelength converters and optical pass filters have been used. Here signals have been used in two triggering channels, where each triggering channel can accommodate n numbers of possible triggering signals as shown in figure-6.2. Here T 1 and T 2 are the triggering channels and blocks 1,2 n, and 1, 2.. n are the optical pass filters respectively in two channels. Block 1,2.. represents the ν 1, ν 2, ν 3. frequencies optical pass filters respectively and block- 1, 2.. also represents the ν 1, ν 2, ν 3.frequencies optical pass filters respectively. WC 11, WC 21..WC ij, WC nn and W 1, W 2,..W i,.. W n, and 1 W, W 2.. W j,. W n are the wavelength converters. I 11, I 21 I ij, I nn are the input channels and O is the output channel of the whole multiplexing system which collects all the output of the WCs. The whole scheme is shown in figure-6.2. Now when T 1 and T 2 both is ν 1 i.e. ν 1 frequency is applied into the triggering channels T 1 and T 2 then only block-1 and block-1 pass this frequency and they are connected to the wavelength converters W 1 and 1 W respectively as a pump beam. The probe beams of each W 1 and W 1 is ν 0. So the output of W 1 and W 1 are also ν 0. The combined output of W 1 and W1 is applied into the WC 11 as a strong pump beam. Here signal of I 11 input channel applied to the WC 11 (which is a weak probe beam of WC 11 ) and is transmitted to the output channel O. Applied strong pump beam of the wavelength converters which are coming from T 1 and T 2 must be in same frequency and have to satisfy the requirement of pump power of the WC s. If one frequency is different from other and it is applied to the WC as pump beam, then it does not work and no transmitted signal is obtained to the 113

124 output terminal. Again if T 1 = ν 1 and T 2 =ν 2 then block-1 passes the ν 1. In W 1, ν 1 frequency is applied as a strong pump beam and a constant weak probe beam of ν 0 frequency is already present there. So conversion is obtained from ν 1 to ν 0 and one gets the ν 0 frequency of light at the output terminal of W 1. The ν 2 frequency is passed through the block- 2 and it is also converted to the ν 0 frequency by the use of 2 W. These both ν 0 frequencies from W 1 and 2 W are applied as a strong pump beam jointly to the WC 21. I 21 the input channel of WC 21 takes the weak input probe beam as data. In presence of both pump and probe beams data of I 21 is transmitted to the output channel O in this multiplexer. Each of T 1 and T 2 triggering channels can accept n inputs. Thus overall the two channels jointly can handle n 2 number of input channels for multiplexing. The whole operation can be made clear by an example. Let T 1 takes ν i and T 2 takes ν j inputs then i th filter allows only to pass the ν i th frequency from T 1 and similarly j th filter allows the ν j th frequency from T 2 through it. Then ν i and ν j after passing through respective W i and W j are converted to ν 0 frequency. The joint power of both the ν 0 is adjusted in such a way that they meet the requirement of pump powers for conversion of wavelength when they fall on WC ij th wavelength converter. Under this situation WC ij th converter allows I ij th input to pass through it and it comes then to the final output terminal of the multiplexer. The advantage of the proposed universal multiplexer is described by the table-6.1. In this table the numbers of input channels multiplexed by the concerned number of triggering channels are shown. From this table it can be concluded that if there are m number of triggering channels in a multiplexer then it can access n m number of input channels, if n-array encoding is adopted. By the use of frequency encoding technique as 114

125 proposed in this scheme, the advantage of handling maximum no. of input channels with a fixed number of triggering channels can be achieved. T 1 T n n W 1 ν 0 ν 0 ν 0 ν 0 ν 0 W 2 W n W 1 W n... I 11 WC WC 11 WC 1n WC n1 I ij WC ij.... I nn I 1n WC nn..... O Output of Multipl exer I n1 Figure-6.2: Architecture of all-optical double triggered universal multiplexer. 115

126 Encoding Binary Trinary N-ary encoding schemes encoding encoding Number of 2 m 3 m.. n m input channels multiplexed by m no. of triggering channels Table-6.1: No. of input channels accessed by m number triggering channels for different encoding 6.4. Practical realization: To set the desired operation pump beam for wavelength conversions of SOA based optical switch should lie between 2 db to 4 db which is the essential requirement for realization of an experimental set up. The 3 db bandwidth of an optical pass filter, which is used for selecting the particular frequency of light, should be in the order of 1 nm to provide a larger number of inputs channels. The pump beam energy actually controls the performance of the data transfer so the selection of pump beam power is very important. The each probe beam should lie between -4 to -2 db for maintaining a 116

127 standard intensity level of light at the output of WCs. The important matter is to mention here is that the wavelengths of the both pump and probe beams should be in C band ( nm). Thus if 20 input frequencies are taken for T 1 and T 2 triggering channel both, the double triggering universal multiplexer then can accommodate 400 input channels Conclusion: The method of optical implementation of a universal multiplexer with all optical switching systems has been discussed in this chapter. The speed of operation of each SOA based wavelength converting switches is far above THz limit. So a high operational speed (above THz) is expected from the whole scheme. It is not only the one advantage of this proposed system but the frequency encoding mechanism also offers a great advantage. The output signal from the multiplexer can be sent to distant receiver as frequency remains generally unaltered in reflection, refraction, absorption etc for the coding of information (0 or 1) with different frequency of light. Therefore this optical technique will be very much useful for transmitting a reliable and faithful data for long distance transmission. To achieve a good amplification the pump beam of WCs should lie between 4dB to 10 db. Also the scheme offers a high signal to noise (S/N) ratio. The main reason for obtaining the high S/N ratio is due to the use of the frequency encoding process in this proposed system. Bit error rate also goes to a very small value in comparison to conventional intensity based or polarization based encoding processes. Increasing the number of triggering channels one can access more number of input channels for multiplexing if the span of frequency is restricted. The operation exploits the inherent parallelism of optics as far as practicable. It is important to mention that each 117

128 WC in fig-6.2 are only operative for transmitting its probe beam to its output if and only if the joint power of both ν 0 frequency coming after conversion from T 1 and T 2 satisfy the requirement of pump power for conversion operation by the WC. For this when only one WC is dedicated to take part in conversion mechanism, other WCs in the matrix of WCs are non operative. This multiplexer can be extended to develop a three or higher triggering based channel universal multiplexer. If it becomes an m channel multiplexer it then easily accommodates n m number of input channels for multiplexing. This is the prime advantage of the proposed scheme as a whole. 118

129 References: 6.1. S.K. Garai, S. Mukhopadhyay, Method of implementing frequency encoded multiplexer and demultiplexer systems using nonlinear semiconductor optical amplifiers, Optics and Laser Technology, 41, 2009, Jitendra Nath Roy, Anup Kumar Maiti, S. Mukhopadhyay, Designing of an alloptical time division multiplexing scheme with the help of non-linear material based treenet architecture, Chinese Optics Letters (China), 4(8), 2006, Jitendra Nath Roy, Anup Maiti, S. Mukhopadhyay, Exploitation of nonlinear material based tree-net architecture in all optical demultiplexing scheme, Journal of Optics, 36(1), 2007, C. Dragone, An N*N optical multiplexer using a planar arrangement of two star couplers, Photonics Technology Letters, 3 (9), 06 August 2002, Unnikrishnan Gopinathan, Thomas J. Naughton, and John T. Sheridan, Polarization encoding and multiplexing of two-dimensional signals: application to image encryption, Applied Optics, 45(22), 2006, Jose M. Castro, David F. Geraghty, Seppo Honkanen, Christoph M. Greiner, Dmitri Iazikov, and Thomas W. Mossberg, Optical add-drop multiplexers based on the antisymmetric waveguide Bragg grating, Applied Optics, 45(6), 2006, Ashish Pal and Sourangshu Mukhopadhyay, An alternative approach of developing a frequency encoded optical tri-state multiplexer with Broad area semiconductor optical amplifier (BSOA), Optics and Laser Technology, 44, 2012,

130 6.8. H.J. Lee, M. Sohn, K. Kim and H.G. Kirn, Wavelength dependent performance of a wavelength converter based on cross-gain modulation and birefringence of a semiconductor optical amplifier, IEEE Photon. Technol. Lett., 11, 1999, L. Deming, N.J. Hong and L. Chao, Wavelength conversion based on cross-gain modulation of ASE spectrum of SOA, IEEE Photon. Technol. Lett, 12, 2000, H. Yu, D. Mahgerefteh, P.S. Cho and J. Goldhar, Improved transmission of chirped signals from semiconductor optical devices by pulse reshaping using a fiber Bragg grating filter, J. Lightwave Technol., 17, 1999,, D.D. Marcenac, A.E. Kelly, D. Nesset and D.A.O. Davies, Bandwidth enhancement of wavelength conversion via cross-gain modulation by semiconductor optical amplifier Cascade, Electron. Lett., 31, 1995, X. Zheng, F. Liu and A. Kloch, Experimental investigation of the cascadability of a cross-modulation wavelength converter, IEEE Photon. Technol. Lett., 12, 2000,

131 CHAPTER-7 Use of all-optical Kerr Nonlinearity for super-fast conversion of a binary number having a fractional part to its decimal counterpart and vice-versa ABSTRACT The role of optical tree architecture is an important approach for conversion of an optical data from binary to decimal and vice-versa. In this chapter I have proposed a new concept of converting of a binary number having some fraction (fractional value) to its equivalent decimal counterpart and its vice-versa. To perform this operation optical tree and some nonlinear material based switches are used properly. Work reported in this chapter was published in: S. Dutta and S. Mukhopadhyay, All optical frequency encoding method for converting a decimal number to its equivalent binary number using tree architecture, Optik - Int. J. Light Electron Opt., 122 (2011)

132 7.1. Introduction: In the previous chapters the frequency encoded all optical systems has been discussed. Those systems are implemented by encode the state of working using frequency of light. Several types of encoding systems are popular for implementing of optical systems. Optical tree architecture is one of them. In connection to the new developments of several all-optical data processing techniques, the role of optical tree architecture can be mentioned specially [ ]. This tree has already been used to convert a position-wise encoded optical decimal data to its binary counterpart and from binary data to its decimal counterpart. Not only are these conversions but also there several other conversions which are possible to be conducted by this tree architecture for the need of all-optical data processing and computing. Here in this chapter I have proposed a modification of optical tree architecture by which a binary data having a fractional part can be converted to its respective equivalent decimal value and from decimal number having a fractional part to its equivalent binary form. Already optical tree has been used for conversion of different forms of data, but this type of conversion of binary number having fractional part to its decimal counterpart and its vice-versa is a completely new concept. The beauty of this optical tree architecture is no use of any kind of switches. In tree architecture only optical channels, beam splitters and mirrors have been used. That s why power conservation is so high and the system will be so fast. 7.2 Optical tree architecture: The schematic diagram of an optical tree-architecture is shown in figure-7.1. Here a light beam (preferably a laser) emitted from a point source (A) passes through the some optical switches to break into two different light beams into BC and BD. Each of these 122

133 two beams breaks into two more parts individually as BC to CE and CF and BD to DG and DH. Proceeding in this way ultimately eight spots (from I to P) are obtained from a single beam AB. By arranging another splitting arrangement in the output plane one can get 16 spots. To use this circuit some optical channels (A 0, A 1, A 2 ) which control the passing of the light from one main channel to a sub channel have been used. The control channels also carries light beam. A 2 control channel is connected to the switch B, A 1 channel is connected to C and D switches, where as A 0 channel is connected to E, F, G, H switches. The function of the switches can be illustrated by an example. If light in the control channel A 2 is present then it activates the switch B to pass the light from AB to BC (upper) sub channel otherwise the light of AB channel will go through the (lower) sub channel BD. In the same way the other switches function. In this connection here this optical system has been used for the conversion of a binary number having a fractional part to its equivalent decimal number. In the present analysis all the switches are alloptical in nature. For the conversion of decimal to binary data different arrangements of mirrors, beam splitters and optical channels are used [7.1]. 123

134 C E I J A B D F G K L M N H O P A 2 A 1 A 0 Figure-7.1 Optical tree architecture for binary to decimal conversion 7.3. Non-linear material as an optical switch: In this system only non-linear type of material has been used as a switch. Now the function of some isotropic non-linear material (NLM) has been discussed as an all-optical switch. This Kerr non-linearity equation for some isotropic material is well established n o 2 = n + n I (1.4) where n is the refractive index (r.i) of the concerned non-linear material, n 2 is the nonlinear correction term, n 0 is the constant linear refractive index term and, I is the intensity of light passing through the material. As for example of some non-linear materials is pure silica glass (SiO2), gallium arsenide (GaAs), carbon di-sulfide (CS 2 ), etc. The path 124

135 of output light is depended on the intensity of the input light. If the intensity of incident light is increased or decreased the refractive index is also changed according this light intensity for that reason the path of output light is also changed. It s a very simple switching system. According to this equation when a beam AB of some fixed intensity I falls in the point O (O is a point in the boundary of linear material (LM) and non-linear material (NLM) it passes through the DE channel as shown in fig.7.2. Now when one beam of intensity I in AB channel and other beam of intensity I in CB channel fall jointly on O, then according to the equation the non-linear refractive index of the material becomes n=n 0 + n 2 I, and for this reason light beam ultimately will pass through the FG channel of fig.7.2. This can be compared as an optical switch. Many all-optical logic devices have been proposed based on this principle [ ]. G E F D NLM LM O B A M C Figure7.2: Non-linear material as an optical switch. 125

136 7.4. Optical conversion method of a binary number having a fractional part to its equivalent decimal number: The optical method of conversion of binary to decimal number is an established one. The next important function is the implementation of a system for transformation of fractional binary number to its decimal counterpart. Here modified tree architecture can be used for such conversion. The significant difference of this scheme from that of the described earlier is the method of placement and position of NLM based switches, beam splitters (BS) and mirrors (M). Placing these components in a properly different way the modified system can be developed. The whole scheme is shown in fig.-7.3(a). Here the control channels (B 4 B 3 B 2 B 1 B 0 ) carry the bits of fractional binary number. The whole things of the system can be made clear by an example. Let the fractional binary is So B 4 =B 2 =B 1 =1 and B 3 =B 0 =0 and for these type of input signal light beam from constant light source (CLS) first passes through the upper channel from NLM 1 then it comes to the lower channel from NLM 2 and again passing through the upper channel from NLM 4 it ultimately exits through the channel number 7. This indicates that the conversion of gives the position-wise encoded decimal number 0.7. In this way any binary number having fractional part can be converted to the respective decimal number by the use of the system described in a figure-7.3(a). Now the whole scheme can be integrated together to convert a binary number having both fractional and nonfractional parts to its equivalent decimal value. Here the block A consist the system of non-fractional part. A 3 A 2 A 1 A 0 are the input channels i.e. the channels for placing the binary non-fractional inputs. On the other hand block-b carries the same system as described in figure-7.3(a) i.e. fractional part, where B 4 B 3 B 2 B 1 B 0 are the input channels 126

137 for placing the binary fractional inputs. By the joint action of the two blocks (A and B) one can convert any binary number (having both fractional and non-fractional parts) to its equivalent decimal value which is encoded position-wise. The whole block diagram is shown in figure-7.3(b) For example if are applied to this integrated system 3.6 is obtained at the output. All the converted fractional decimal numbers from its equivalent binary numbers are shown in table-7.1. FRACTIONAL NUMBER DECIMAL CORRESPONDING EQUIVALENT FRACTIONAL BINARY NUMBER Table-7.1: Converted numbers of fractional decimal to its equivalent binary code 127

138 9 NLM NLM 2 NLM 4 NLM CLS NLM 1 NLM O U T P U T S NLM 9 1 NLM 3 NLM 5 0 B 4 B 3 B 2 B 1 B 0 REPRESENTS MIRROR REPRESENTS BEAM SPLITTERS Figure-7.3(a): All optical system for converting fractional binary number to its decimal counterpart 128

139 CONVERSION METHODE FOR BINARY TO DECIMAL A 0 A 1. CONVERSION METHODE FOR BINARY (FRACTIONAL) TO DECIMAL B 0 B 1 BLOCK-A A 2 A 3 BLOCK-B B 2 B OUTPUTS Figure-7.3(b): All optical integrated system for converting a binary number having both fractional part and non-fractional part to its equivalent decimal number 129

140 7.5. Optical conversion method of a decimal number having a fractional part to its binary equivalent: Now I also propose a concept of converting a decimal number having a fractional part to its binary equivalent in an all optical process. Here the decimal number has two parts as a whole. First part is the non fractional where as second part is the fractional. Second part extends the conversion of fractional part. The proposed system for conversion of the fractional digit to its binary equivalent is shown in figure-7.4(a). Here beam combiners have been used for coupling two light beams into a single one and also for breaking a single beam into two parts. These are represented by BS in the figure. The mirrors (M) are used for the reflection of light beam. Using these BS s the converted result is obtained at the output channel B 4 B 3 B 2 B 1 B 0 of the system shown in fig. 7.4(a). By the use of the modified tree structure B 4 B 3 B 2 B 1 B 0 ultimately show the converted equivalent binary number of the fractional decimal contribution which lies from 0 to 0.9. To develop this system the channels marked by 0,1,2,3,..,9 are used for applying the fractional part of the decimal number. If 0.8 is to be converted the light is to be applied at the channel marked by 8, and binary converted output is obtained in the channels marked by B 4 B 3 B 2 B 1 B 0.As for example-when the light goes through the channel 7, no light comes channels B 0 and B 3 because of the absence of BS in the respective channel. So B 0 =B 3 =0, but B 4,B 2,B 1 channels get the necessary light reflected from the respective BS s. Hence B 4 =B 2 =B 1 =1. Ultimately B 4 B 3 B 2 B 1 B 0 =10110 is obtained for the application of light in the input channel marked by 7. So it is the equivalent binary number of decimal value 0.7. Similarly any decimal number from 0 to 0.9 can be converted to its binary value. The non-fractional part and fractional part of a system can be combined 130

141 together to develop a single integrated system from which a decimal number having both the non fractional part and the fractional part to its binary equivalent value can be converted. The scheme is shown in figure-7.4(b). As for example- if it is required to convert of the decimal number a.b to its equivalent binary number (where a is the non fractional decimal digit and b is the fractional decimal digit), the light beams are to be placed in the a th channel of the block A and to b th channel of the block B in the system. Here block A and block B comprise the system of non-fractional part and the system of fractional part respectively. The A 3 A 2 A 1 A 0.B 4 B 3 B 2 B 1 B 0 comes as the result of the converted binary number from its decimal counterpart. If the decimal number is 3.7 then surely has been received at the output stage i.e. A 1 =A 0 =B 4 =B 2 =B 1 =1 and A 3 =A 2 =B 3 =B 0 =0. 131

142 M BS 9 M BS BS BS BS 8 M BS BS BS BS BS 7 6 BS BS BS 5 M BS 4 M BS BS M BS BS BS 3 2 BS BS BS M BS BS BS 1 0 B 4 B 3 B 2 B 1 B 0 OUTPUTS Figure-7.4(a) All-optical Conversion system from fractional decimal number to its binary counter part 132

143 CONVERSION SYSTEM FOR NONFRACTIONAL PART (0 TO 9) BLOCK-A CONVERSION SYSTEM FOR FRACTIONAL PART (0 TO 9) BLOCK-B A 3 A 2 A 1 A 0. B 4 B 3 B 2 B 1 B 0 OUTPUTS Figure-7.4(b): The block diagram of complete system for conversion of decimal number having both fractional part and non-fractional part to its binary equivalent 7.6. Alternating approach of non-linear switch: In this chapter non-linear material has been used as an optical switch for conducting the intensity encoded data based operation. In this approach the optical switching operation is realized by the use of non-linear character of semiconductor optical amplifier (SOA). Here the channel shifting operation in the tree architecture is 133

144 conducted by SOA. As usual SOA based wavelength converter converts the wavelength of the light which is described before in the introduction section. One prism is placed in front of it. So λ 1 wavelengths of light first passing through the SOA switch goes through the prism and exist from a specific (channel 1). Now if a (λ 2 wavelength) of weak probe beam is applied to the input side of the SOA switch in addition to the existing pump beam then a converted light beam of wavelength λ 2 is obtained from the output of the SOA switch and after passing through the prism it comes out through another channel (channel 2). Thus two light beams come in two different channels of the output in the system described in fig-7.5. When a light of wavelength λ 1 is applied at the input side of the SOA, it ultimately exits from channel-1 and when light of λ 1 and λ 2 wavelengths both falls in the input side of SOA, light is achieved from another channel marked as channel- 2. This type of switch can be used in replacement of the non-linear material switch. It is already seen that for the conversion of a decimal number (having some fractional part) to the equivalent binary number and its vice-versa the systems proposed in fig-7.4(a) and fig-7.3(a) will be useful. The conventional non-linear switches used in these two systems may be replaced by the SOA based switches as described above to get a faithful operation. The output power of the digits of the converted data becomes stronger here sufficiently. 134

145 Biasing current λ 2 λ 1 INPUTS SOA based wavelength converter Optical prism λ 2 OUTPUTS λ 1 Figure-7.5: SOA based intensity encoded switch 7.7. Conclusion: In this chapter the method of conversion of a binary number having a fractional part to its equivalent decimal number and its vice-versa with all optical switching system has been discussed. The speed of operation is real time. The scheme may be extended both vertically as well as horizontally for the conversion of a higher valued binary number to its equivalent decimal number and from decimal to binary. The inherent parallelism of optics is exploited in the scheme as far as practicable to obtain superfast operation speed. To get more accuracy in the conversion of the fractional binary number one may use more number of control channels in an extended system. As a whole the total system is all optical one and hence the advantages of using optics are achieved. A suitable non-linear material like SOA and a suitable diode laser should be used, for getting low optical power consumption and for reliable operational result. 135

146 References: 7.1 An optical conversion system: From binary to decimal and decimal to binary, Optics Communications (The Netherlands) 76(5 6), 1990, S. Mukhopadhyay, J. N. Roy, and S. K. Bera, Design of a minimized LED array for maximum parallel logic operations in optical shadow casting technique, Opt. Commun. 99, 1993, Parallel Distributed Processing: Explorations in the Microstructure of Cognition, D. E. Rumelhart and J. L. McClelland, Eds., Vols. 1 and 2, MIT Press, Boston M. M. Mirsalehi, History of optical parallel processing, R. A. Mayer, Ed., Encyclopedia of Laser and Optical Tech., Academic Press Inc., New York N. Peyghambarian and H. M. Gibbs, Optical bistability for optical signal processing and computing, Opt. Eng. 24(1), 1985, R. Tripathi, G. S. Pati, and K. Singh, Non-linear processing and fractional-order filtering in a joint fractional Fourier transform correlator: performance evaluation in multiobject recognition, Appl. Opt. 40(17), 2001, S. D. Smith, I. Janossy, H. A. Mackenzie, J. G. H. Mathew, J. J. E. Reid, M. R. Taghizadeh, F. A. P. Tooley, and A. C.Walker, Nonlinear optical circuit elements, logic gates for optical computers: the first digital optical circuits, Opt. Eng. 24(4), 1985, N. Pahari, D. Das and S. Mukhopadhyay, All-optical method for the addition of binary data by non-linear materials, Applied Optics, 43(33), 2004,

147 7.9 J. N. Roy, A. K. Maiti, D. Samanta and S. Mukhopadhyay, Tree-net architecture for integrated all-optical arithmetic operations and data comparison scheme with optical nonlinear material, Optical Switching and Networking, 4, 2007, N. Mitra and S. Mukhopadhyay, A new scheme of an all optical J-K Flip-flop using non-linear material, Journal of Optics., 37(3), 2008, H. L. Minh, Z. Ghassemlooy and W. P. Ng, All-optical.ip.opbased on asymmetric Mach Zehnder switch with a feedback loop and multiple forward set/reset signals, Opt. Eng. 46, 2007, Mitra and S. Mukhopadhyay, A method of developing all-optical mono-stable multivibrator system exploiting the Kerr non-linearity of medium, Optik International Journal for Light and Electron Optics, 122, 2011, Srivastava, S. Medhekar, Switching of one beam by another in a Kerr type nonlinear Mach Zehnder interferometer, Optics and Laser Technology, vol. 43, no. 1, 2011, pp Srivastava, S. Medhekar, Switching behavior of a nonlinear Mach Zehnder interferometer: Saturating nonlinearity, 43(7), 2011, Medhekar, Rajkamal, S. Konar, Successive uptapering and stationary self-trapped propagation of a laser beam in a saturating nonlinear medium Laser and Particle Beams, 13(4), Medhekar, S. Konar, Rajkamal, Self tapering and uptapering of a self guided laser beam in an absorbing/gain medium with nonlinearity, Pramana-journal of Physics, 44(3), 1995,

148 CHAPTER-8 General conclusion and future scope of the work ABSTRACT Here in this chapter I have included an overall conclusion of the whole thesis, where the advantages and limitations of the proposed schemes are discussed. Accordingly I have discussed also the remaining scope of work in this field where some contribution may be done. 8.1 Introduction: The need of optical computation and communication has stimulated new scientific research for flexible, efficient and super fast optical techniques and optical switches. Optical systems based on semiconductors optical amplifiers switches have shown the potential of obtaining the cheap and easy techniques to implement the super fast optical computation and communication systems. In this dissertation, I have undertaken a systematic study to improve the all-optical logic and arithmetic operations using a problem free encoding method. The major conclusions from the dissertation are summarized below: 8.2 Conclusion of the thesis Chapter-1 contains the brief past history of optical computation and communication systems as well as the current status in this field. In these chapter advantages of optics over electronics has been described. The most important property of the frequency encoding principle is that, frequency is the primary character of the wave and during its propagation throughout the communicating media it can protect its uniqueness irrespective of the absorption, reflection, 138

149 transmission. Over all other conventional encoding techniques this is the most prospective advantage of the frequency encoding technique and these advantages are described in this chapter. Chapter 2 reports on the implementation of all-optical frequency encoded NOT latch using semiconductor optical amplifier based switches. Here the advantages of frequency encoded NOT latch has been described. Using frequency encoding principle this NOT latch scheme will be more reliable and faster comparatively other techniques. Using this latch logic and the frequency encoded technique one can implement many other operations like digital types of flip-flops, multiplexer, demultiplexer etc. with some modification of the scheme. Chapter 3 mainly deals with frequency encoded logic gates, R-S flip-flop and programmable logic unit with semiconductor optical amplifier based switches. Logic gates are the building blocks of any memory unit. The potential advantage of these optical gates and flip-flop over many other established optical gates was the use of frequency encoding technique, for which the coded information (0, 1) in a signal remains unchanged in refraction, reflection, absorption, etc. for a long distance transmission of data. The proposed system can offer also a noise free conversion to provide a high signal to noise (S/N) ratio. Chapter 4 presents a frequency encoded all-optical clocked S-R flip-flop based on semiconductor optical amplifier switches. The most important application of the system can be seen in digital communication, where one can send the frequency encoded new data or an (old data depending on the applied clock) in the communication channel. Even if the sender requires making a data to be 139

150 continued for communication or a new data is to be introduced in the channel, he can easily do it by the use of the above system. The output signal from the clocked S R flip-flop can be sent to distant receiver as it remains unaltered in reflection, refraction, absorption, etc. due to the nature of coding of bits (0 or 1) with frequency variation of light. Therefore this technique will be very much useful for conducting a reliable and faithful optical memory both in communication and computation. This optical clocked S R flip-flop and frequency encoding technique can be used many other optical devices where clocked S R flip-flop is an essential unit. Chapter 5 reports on the transmission on the frequency encoded parallel data. The SOA takes the role of conversion of an intensity encoded data to a frequency encoded one by the exploitation of its wavelength conversion character. SOA can achieve the THz speed of operation for this conversion process. This proposed method can, therefore ensure a very high speed optical communication over many other conventional communication techniques. If a data accommodates eight bits or sixteen bits, then all the bits can be sent in parallel through the optical fiber. Chapter 6 deals with the frequency encoded universal all-optical multiplexer system. Here to implement this universal some wavelength converters and some ADD/DROP multiplexers has been used. Using frequency encoding technique will be very much useful for transmitting a reliable and faithful data for long distance transmission. Also the scheme offers a high signal to noise (S/N) ratio. The main reason for obtaining the high S/N ratio is due to the use of the frequency encoding process in this proposed system. Bit error rate also goes to a very small 140

151 value in comparison to conventional intensity based or polarization based encoding processes. Increasing the number of triggering channels one can access more number of input channels for multiplexing if the span of frequency is restricted. The operation exploits the inherent parallelism of optics as far as practicable. Chapter 7 reports on all-optical Kerr Cell for super fast conversion of a binary number having a fractional part to its decimal counterpart and vice-versa. The speed of operation is real time. The scheme may be extended both vertically as well as horizontally for the conversion of a higher valued binary number to its equivalent decimal number and also from decimal to binary. The inherent parallelism of optics is exploited in the scheme as far as practicable to obtain super fast operation speed. Increasing the controlling channel the can be extended. Again this conversion system with non-linear material can be replaced with the SOA based switches. This is the advantageous point of view of this whole scheme. 8.3 Future scope of the work in this area Although the present dissertation reports a detailed study on semiconductor optical amplifier and kerr type optical switching for developing some systems in the area of optical computation and communication to improve efficiency and speed of the proposed systems, but several aspects (as mentioned below) could not be taken up, which are worth for further investigations. There are possibilities to implement all these proposed works physically and experimentally. 141

152 All other building blocks which are not been proposed yet can be implemented in future works. Implementing the optical frequency encoded register and counter based on semiconductor optical amplifier based switches. Implementation of some all-optical frequency encoded multiplier, tristate optical processors, optical softwares, optical nano processors etc. 142

153 List of published papers Section A: Journal papers 1. S. Dutta and S. Mukhopadhyay, All optical frequency encoding method for converting a decimal number to its equivalent binary number using tree architecture, Optik 122 (2011) S. Dutta and S. Mukhopadhyay, An all optical approach of frequency encoded NOT based Latch using semiconductor optical amplifier, J Opt 39 (1) (2010) S. Dutta, S. Mukhopadhyay, Alternating approach of implementing frequency encoded all optical logic gates and flip-flop using semiconductor optical amplifier, Optik 122 (2011) S. Dutta and S.Mukhopadhyay, All-optical approach for conversion of a binary number having a fractional part to its decimal equivalent and vice-versa, Optics and Photonics Letters, 3(1) (2010) S. Dutta, S. Mukhopadhyay, A new approach of parallel data transmission through optical waveguide with SOA based frequency encoding/decoding technique, Optik 123 (2012) S. Dutta, S. Mukhopadhyay, A new alternative approach of all optical frequency encoded clocked S R flip-flop exploiting the non-linear character of semiconductor optical amplifiers, Optik 123 (2012)

154 Section B: Conference papers 1. S.Dutta and S.Mukhopadhyay, Optical conversion system: from any fractional decimal number to its equivalent binary form, 16th West Bengal State Science & Technology Congress, The University of Burdwan, Bardhaman, India (28th February -1st March, 2009). Paper was presented (oral) by S. Dutta on 28th February. 2. Soma Dutta, S. Mukhopadhyay, All optical frequency encoding method for converting a decimal number to its equivalent binary number using tree architecture, International Conference on Optics & Photonics (ICOP-2009), XXXIV Symposium of the Optical Society of India, Central Scientific Instruments Organisation (CSIO), Sector -30 C, Chandigarh, India, October 30 November 1, Paper was presented (poster) by S. Dutta on 31st December. 3. Soma Dutta, S. Mukhopadhyay, Method of developing an all optical frequency encoded clocked R-S flip-flop exploiting the nonlinear character of semiconductor optical amplifiers, International Conference on Radiation Physics and its Applications (ICRPA 2010), organized by Department of Physics, The University of Burdwan; held at Science Centre, Golapbag, Bardhaman, 16th and 17th January, Paper was presented (poster) by S. Dutta on 17th January. 4. S.Dutta and S.Mukhopadhyay, A new approach of all-optical frequency encoded conversion from a binary data to the decimal one using SOA based switches, International Conference on Computing and Systems 2010 (ICCS 10), organized by the Department of Computer Science, The University of Burdwan, 144

155 held at Science Centre, Golapbag, Bardhaman, 19th and 20th November, Paper was presented (oral) by S. Dutta on 20th November. 5. S.Dutta and S.Mukhopadhyay, Application of semiconductor optical amplifier for development of ultra fast programmable unit, Frontiers in Materials Science- 2010, December- 06, National Institute of Technology Durgapur, INDIA, Paper was presented (poster) by S. Dutta on 6th December and awarded for the best paper. 6. S.Dutta and S.Mukhopadhyay, A new approach of all-optical frequency encoded programmable unit, en2c, 21st January S.Dutta and S.Mukhopadhyay, All-optical frequency encoded programmable adder and subtractor unit, National Workshop on Quantum Perspective of advanced Material (QPAM-11), Organized by the dept. of physics and technophysics, Vidyasagar University, 23rd to 25th March, Paper was presented (poster) by S.Dutta on 24th March. 8. S.Dutta and S.Mukhopadhyay, A new approach of all-optical frequency encoded memory unit based on semiconductor optical amplifier, International conference on laser, materials science & communication (ICLMSC-2011), Paper was presented (poster) by S. Dutta on 9th December,

156 Author's personal copy Optik 122 (2011) Contents lists available at ScienceDirect Optik journal homepage: All optical frequency encoding method for converting a decimal number to its equivalent binary number using tree architecture Soma Dutta, Sourangshu Mukhopadhyay Department of Physics, The University of Burdwan, Burdwan , West Bengal, India article info abstract Article history: Received 21 May 2009 Accepted 5 November 2009 Keywords: Optical computing Optical frequency encoding principle Optical tree architecture Semiconductor optical amplifier (SOA) Optical nonlinearity In any kind of computing and data processing system the use of binary numbers are found very much suitable and reliable. On the other hand several natural representations have been realized using decimal numbers. So conversion of a decimal number to its binary equivalent and vise-versa are of great importance in the field of computation technology. There lie already a number of established methods regarding such conversion processes. Again optical tree architecture is one of the most promising systems for realizing the optical conversion of any decimal number to its equivalent binary. Here in this communication the authors propose a new method for optical conversion of a decimal number to its binary equivalent using tree architecture based system and frequency encoding principle. In frequency encoding system, frequency of light is used for encoding of decimal digits or binary bits instead of intensity variation. For example 0 and 1 bits of binary number are coded by two different frequencies of light signal, instead of representing the presence of light as 1 and absence by 0. The proposed conversion process has multifaceted advantages in communication, as well as in data processing. To implement the above conversion some characteristic features of semiconductor optical amplifier (SOA) have been used massively. The wavelength conversion property, cross gain modulation and some nonlinear properties of SOA are exploited to get the frequency encoded response. The proposed system carries all the basic advantages of optical processing as well as those of frequency encoding also Elsevier GmbH. All rights reserved. 1. Introduction Optics has a strong and very potential role in information and data processing because of its strong inherent parallelism. It has several advantages over electronics in superfast computation and data processing. Last few decades, several all optical data processors were proposed based on the Boolean logic. Those optical systems and optical logic devices based on optical switches are found very much useful than electronic ones in connection to speed and many other aspects. There lie several types of optical switches. Semiconductor optical amplifier (SOA) is one of them which can be used as a potential high speed optical switch. Again it is well established that like electronic computation, in optical computation also the conversion from decimal number to binary number or vise-versa is very much important as data represented in binary are found most suitable for computation. Here in this communication the authors propose a new concept to implement all optical conversion from decimal to binary with optical tree architecture using the frequency encoding principle [1,2]. The advantages of Corresponding author. addresses: soma.dtta@gmail.com (S. Dutta), sourangshu2004@yahoo.com (S. Mukhopadhyay). frequency encoding are that as the frequency is a fundamental character of any signal, so it remains unchanged in reflection, refraction, absorption, etc. during transformation of signal. The high reflecting and frequency diverting properties of optical add and drop multiplexer (ADM) and high wavelength conversion property of reflected semiconductor optical amplifier (RSOA) because of its four waves mixing character have been exploited here to implement the above conversion [3,4]. 2. Frequency encoding principle Frequency is the basic characteristics of light. In optical computing and data processing therefore most important point is the very high speed of processing. In most of the cases the presence of optical signal at input/output end are encoded as 1 and absence of optical signal as 0. But for long distance in communication intensity of optical signal may significantly change, it may dropdown below the respective reference level. The interesting point is that this problem can be solved by using frequency of light which is the fundamental characteristics of light. One can encode and decode two different states of information by two different frequencies [5,6]. Here if 1 state is represented by a frequency 2 then 0 state is represent by another frequency 1. These frequencies are remaining unaltered /$ see front matter 2010 Elsevier GmbH. All rights reserved. doi: /j.ijleo

157 Author's personal copy 126 S. Dutta, S. Mukhopadhyay / Optik 122 (2011) Fig. 2. Transfer of power from pump beam to a weak probe beam by wavelength converting mode of RSOA (here HR and AR represent the high reflection and antireflection coating). Fig. 1. Add/drop multiplexer (ADM) by RSOA (here 3 is selected for reflection back). during the reflection, refraction, absorption, etc. when data is to be transmitted. Here in this proposal we encode and decode the state of information 0 by a beam of wavelength 1 (which is corresponding to 1 ) and the state of information 1 by another wavelength 2 (which is corresponding to frequency 2 ). 3. Switching operation of semiconductor optical amplifier (SOA) Semiconductor optical amplifier (SOA) is an optoelectronic device that under suitable operating conditions (i.e. at proper bias current) can amplify an input optical signal. Semiconductor optical amplifier SOAs can be classified into two main types, the Fabry Perot (FP-SOA) where reflections from the end facets are significant whereas in the traveling wave SOA (TW-SOA) type reflections are negligible. Anti-reflecting coating can be used to create SOAs for facet reflection. In optical transport action network these SOAs can be successfully used. Many of the functional applications of SOA are based on its nonlinearity. SOAs can be used as different types of optical switches which are based on four wave mixing, wavelength conversion, Add/drop multiplexing (ADM), etc. Here in this communication the authors use the add/drop multiplexing character of SOA for selecting the proper frequency from a band of frequencies, not disturbing others. The wavelength conversion property of SOA can be used for the conversion of the selected frequency or wavelength to another desired one. There are specifically of some types of ADM. Here in this paper the WDM add/drop multiplexer is particularly used. A mixer of frequencies of pulse or continuous signal ( 0 to 9 ) falls on the ADM which is tuned at any particular frequency by the application of the proper bias current. The ADM then passes all the frequencies not allowing that particular frequency to pass through it. One can receive that particular frequency of light at output by the use of an optical circulator. This is shown in Fig. 1. The reflecting semiconductor optical amplifier (RSOA) on the other hand is a wavelength converter [7,8]. In this switch a particular frequency of week light signal is given as a probe beam and another signal of another frequency is given as a pump beam to the input channels of RSOA. If only those two specific light signals are present in the input channels then one can get the power of the pump beam is delivered to the week probe beam at output, i.e. the probe beam is exchanged by a power factor. The scheme is shown in Fig. 2. Using the ADM a specific frequency of light is selected and using RSOA the selected frequency is amplified [9 12]. switches are required here for converting a decimal digit to its binary counter-part. Some beam splitters (BS) and mirrors (M) are used in proper positions of the tree to execute this conversion. Some optical channels are used here to get an output. Light beams are splited by the BS s and reflected by the M s and go through the optical channels. In Fig. 3 the tree architecture is shown. Here any decimal digit (positionally represented) is converted to its respective binary counter-part. For example if the decimal value (which is supposed to be converted) is 5, then one have to apply light to the input channel of the tree marked by 5. At that time one can receive the output 0101, i.e. light will come at X 2 and X 0 but no light will come at X 3 and X 1. The conversion scheme is shown in Fig Conversion method Here 0 to 9 the ten frequencies are taken for representing 10 decimal digits. A light beam having frequency ( 0 to 9 ) fall on the ADM 1 (the 1st add drop multiplexer) where this ADM 1 is tuned for reflecting the light of frequency 1 by the application of proper bias current to it. So, only 1 frequency of light will not pass through the 1st SOA whether all other will easily pass through it. The 1 frequency will come out from the ADM 1 by the optical circulator (C). Similarly nine other ADMs are used in series which are tuned for different other frequencies ( 2 to 9 ), respectively. So, those selected frequencies of light will come from 10 different channels of the 10 SOA blocks in the series of ADMs. Now using some optical beam splitters BS s and mirrors M s one can convert a decimal number to an appropriate binary number follow the principle of conversion by tree architecture. For this conversion one can use now the frequency encoding principle. Here the light frequency is used to code signal bit. If it is to convert the decimal number 5 one should use the 5 light frequency at initial input. As for example 101 is the equivalent binary number of the decimal number 5. So first the 5 light frequency is selected and applied as initial input for conversion. It passes through the ADM 1, ADM 2, ADM 3, and ADM 4 but cannot pass through ADM 5 as it is tuned for the frequency 5.So 5 cannot pass through the ADM 5 and the 5 frequency is received from the respective circulator. Then passing through the properly 4. Optical tree architecture For optical conversion from a decimal to binary number and its vise-versa are very important. Several types of conversion methods are well known in computation process. Optical tree architecture is one of them. It is also an established optical technique. No Fig. 3. Optical tree architecture for conversion from decimal to its binary counterpart.

158 Author's personal copy S. Dutta, S. Mukhopadhyay / Optik 122 (2011) Fig. 4. An integrated scheme of converting a frequency encoded decimal number to its binary counter-part. oriented BSs and Ms (as shown in Fig. 4) one can get the binary output from the respective channels. These outputs are A 3 A 2 A 1 A 0. In the particular case one can get 5 signal at A 2 and A 0 and no light at A 3 and A 1 which indicates Now these outputs A 3 A 2 A 1 A 0 treated as inputs of the wavelength converter switches (the RSOAs). Here these outputs are used pump beams and another constant 1 frequency light beam is given as probe beam to all the wavelength converters. For different input frequencies 0 to 9 (representing different decimal inputs) the A 3 A 2 A 1 A 0 are converted in such a way that the bits (A 3 or A 2 or A 1 or A 0 ) takes that particular frequency when it is 1 and it takes no light when it is 0. For example when the decimal input 7 is applied for conversion, it indicates 7 frequency is applied. Therefore A 3 A 2 A 1 A 0 is 0111 which is binary equivalent of 7. So here A 3 represent 0 by consuming no light in the channel whereas A 2 A 1 A 0 takes the light of frequency 7 to represent 1. Now for conversion of decimal 6, 6 is applied as decimal input. The binary equivalent of 6 is To represent it the A 2 and A 1 takes light of 6 frequency to represent 1 and no light to represent 0. Therefore for different conversion 1 bit are represented by different frequencies. To get rid of this problem, the outputs of A 3, A 2,A 1 and A 0 are passed through four different RSOA s for conversion of light frequencies to a specific frequency 1 to represent 1. So when 9 is to be converted, the bits of the converted binary number are represented by frequency 1, by the use of RSOA. Thus the final outputs are B 3 B 2 B 1 B 0, which comes after the operations from RSOAs. Therefore one can get 1 frequency of light in B 3 and B 0 and no light B 2 and B 1. In the same way for application of all the decimal inputs the 1 bits of the converted binary number is represented the 1 frequency of light and 0 bits by no signal (Fig. 4). 6. Conclusion The whole operation system is all optical one and extends a very high speed operation (far above GHz limit). This method can be extended also for converting a decimal number of higher values (greater than 9) by enlarging the tree architecture. The potential advantage of the method over any other electronic and optical one is the frequency depending input and output encoding. For that reason the coded information (1 or 0) in a signal remain unchanged in reflection, refraction, transmission, absorption, etc. For the same converters based on the intensity based encoding/decoding of decimal numbers, the most disadvantageous point is the fluctuation of intensity of light during transition which may alter the reference level of bit value. Again due to the fluctuation of the light intensity the direction of the output may be changed in case of implementation of the conversion by Kerr type nonlinear materials switch based optical switch with intensity based encoding, decoding mechanism. So in the case of nonlinear material based optical switches it always requires constant intensity light source. Whereas in frequency based encoding/decoding system this disadvantage is removed. Here the coded frequency remains unaltered in all the above conditions. Therefore, this proposed system does not only give a high speed operation, but also offers a trouble free and noise free conversion. References [1] S.K. Garai, S. Mukhopadhyay, A method of optical implementation of frequency encoded different logic operations using second harmonic and difference frequency generation techniques in non-linear material, Opt. Int. J. Light Electron. Opt. (2008), doi: /j.ijleo [2] S. Mukhopadhyay, An optical conversion systems from binary to decimal and decimal to binary, Opt. Commun. (The Netherland) 76 (May (5 6)) (1990) [3] S.K. Garai, D. Samanta, S. Mukhopadhyay, All-optical implementation of inversion logic operation by second harmonic generation and wave mixing character of some nonlinear material, Opt. Optoelectr. Technol. China 6 (August (4)) (2008) [4] J.P.R. Lacey, M.A. Summerfield, S.J. Madden, Tunability of polarization-intensive wavelength converters based on four wave mixing in semiconductor optical amplifiers, J. Lightwave Technol. 16 (12) (1998) [5] M.J. Connelly, Semiconductor Optical Amplifiers, Kluwer Academic publishers, [6] N.K. Dutta, Q. Wang, Semiconductor Optical Amplifier, World Scientific Publishing, Singapore, [7] L.Q. Guo, M.J. Connelly, A novel approach to all-optical wavelength conversion by utilizing a reflective semiconductor optical amplifier in co propagation scheme, Opt. Commun. 281 (September (17)) (2008) [8] Q. Guo, M.J. Connelly, All-optical and gate using induced nonlinear polarization rotation in a bulk in a bulk semiconductor optical amplifier, in: Technical Digest: Optical Amplifiers and Their Applications, The Optik Society of America, Washington, DC, 2005 (Press no. SuB9). [9] L.Q. Guo, M.J. Connelly, A poincare approach to investigate nonlinear polarization rotation in semiconductor optical amplifiers and its applications to all-optical wavelength conversion, Proc. SPIE 6783, (1 5) (2007). [10] H.J. Dorren, D. Lenstra, Y. Liu, M.T. Hill, G.-D. Khoe, Nonlinear polarization rotation in semiconductor optical amplifiers: theory and application to all-optical flip-flop memories, IEEE J. Quantum Electron. 39 (January (1)) (2003) [11] H. Soto, D. Erasme, G. Guekos, Cross-polarization modulation semiconductor optical amplifiers, IEEE Photon. Technol. Lett. 11 (1999) [12] L.Y. Lin, J.M. Wiesenfeld, J.S. Perino, A.H. Gnauck, Polarization-intensive wavelength amplifier, IEEE Photon. Technol. Lett. 10 (7) (1998)

159 J Opt 39(1) : J Opt 39 (1) : RESEARCH ARTICLE An all optical approach of frequency encoded NOT based Latch using semiconductor optical amplifier Soma Dutta Sourangshu Mukhopadhyay Received: 01 September 2009 / Accepted: 28 January 2010 Optical Society of India 2009 Abstract In case of super-fast optical computation and communication, frequency encoding techniques are found to be very promising and reliable one. Optical logic gates based on the principle of frequency conversion of some nonlinear materials play the key role for the implementation of a frequency encoded data processing system. Again semiconductor optical amplifier has already been established successfully for frequency conversion. In frequency encoding system, different frequencies of light signal are used for representation of binary bits 0 or 1 instead of intensity variation. For example 0 and 1 bits of Boolean logic can be coded by two different frequencies of light signa and respectively. In this communication, we propose the method of developing an optical memory or a NOT based latch. Several types of phase encoded, polarization encoded and intensity encoded optical memories have been reported earlier, including latch also, whereas this proposal have been planned to develop an all optical latch logic using frequency encoded principle and it offers a reliable and faithful processing rather than other established techniques. Key words Optical computation Non-linear materials Semiconductor optical amplifier. Introduction Increasing demand for a faster and reliable data processor has given birth of the concept of all optical super-fast computers. In last few decades, there are several proposed optical and photonic devices, which can be run with operation speed far above the conventional GHz limit. Many of those devices have been dedicated for performing Soma Dutta( ) Sourangshu Mukhopadhyay Department of Physics, The University of Burdwan, Golapbag, Burdwan, pin , West Bengal, India soma.dtta@gmail.com 13

160 40 J Opt 39(1) : logic, arithmetic and algebraic operations to achieve the goal of all-optical computer and data processor [1-6]. Among all other components, memory is an essential one for any data processor. Different types of volatile and non-volatile memories have already been developed successfully for the present electronic data processing systems. Similarly optical memories are also of great importance for the development of optical computing technologies. Many scientists are involved in deep research for the realization of digital volatile and non-volatile optical memories. Though some successes have been achieved in the last few decades to develop optical memories, flip-flop, bi-stable multivibrators and latches, still the ultimate goal of developing an optical computer has not been fulfilled yet. In this communication we have proposed a methodology of developing optical NOT based latch using a specific non-linear behavior of semiconductor optical amplifiers (SOA) accommodating with frequency encoding technique [7 11]. In most of the encoding cases, presence of optical signal at the input or output of a logical device is encoded as 1 logic state and absence of optical signal as 0 logic state. As the intensity of light decreases with the increase of optical path of a light signal, so the intensity of light may drop down below the reference level of logic state 1 and may enter into the reference level of logic state 0. To overcome this problem one can use the frequency of light signal to encode a logic bit. Here the presence of a specific frequency of light is treated as 1 state and then other specific one represents 0 state i.e. if 1 state is encoded by a frequency then 0 state is done by another frequency. So and will remain unaltered throughout the transmission of data as frequency is a fundamental characteristic of a light signal and it remains unchanged under reflection, refraction, and absorption etc. Thus the frequency encoded technology can be more promising and reliable one for the successful realization of super-fast computation and also for communication with optics over many other conventional techniques [12 13]. There lies also many other encoding techniques like polarization encoding, phase encoding etc. [14 15]. But all these encoding principles have the same problem of changing the encoded state during transmission. Frequency encoding technique is free from such problems. Semiconductor optical amplifier based optical switching Semiconductor optical amplifier (SOA) is generally based on specially GaAs material. It is used for developing several optoelectronic devices which under suitable operating condition can amplify an input signal. SOA can be classified into two main types; one is the Fabry Perot SOA (FP-SOA), where reflections from the end surfaces are important the signal undergoes many passes through the amplifier. On the other side in Traveling Wave (TW-SOA) the reflection is negligible a signal undergoes only a single passes through the amplifier. This type of SOA can be used in optical transparent networking. The non-linearity of SOA can also be used successfully in many functional applications, which are caused by carrier density induced by the amplifier s input signal. There are four types of non-linearities in SOA, which are cross gain modulation (XGM), cross phase modulation (XPM), self phase modulation (SPM), and four wave mixing (FWM). Here in this communication the authors exploit the cross gain modulation character of SOA. Changing carrier density of the amplifier will affect all of the input signals entered into the SOA. Carrier lifetime depends upon the temporal response of the carrier density. A weak CW probe light of wavelength 1 and a strong pump beam of wavelength 2, with a small-signal harmonic modulation at angular frequency, are injected to the input terminals of the SOA. The strong pump beam transfer its total power to the weak probe beam and then the weak probe beam becomes strong and comes out to the output terminal the SOA acts as a wavelength converter [16]. It transfers 13

161 J Opt 39(1) : information from one wavelength to another signal at a different wavelength. The scheme is shown in Fig.1. There are two basic schemes used in XGM based wavelength converters; where one is the co-propagating and the other is counter-propagating schemes. In this communication we use the co-propagating type XGM wavelength converter. This type of converters having anti-reflecting coating at the front surface which provides no reflection for 1 but supports transmission for 1 and 2 and a highly-reflecting surface at the output which provides a very good reflection for 2 and good transmission for 1 wavelength. If 1 does not exist at one input probe terminal, this conversion is not allowed. Now for the conversion process the roles of the above coatings are very much important. This coating basically ensures the obtaining of 1 signal at the output. Thus this SOA behaves as perfect optical switch. The wavelengths of the pump and probe inputs are generally selected as 1555 nm ( 1 ) and 1550 nm ( 2 ) corresponding to frequency and respectively when GaAs is used as the concerned SOA. Fig.1 A schematic diagram of semiconductor optical amplifier (SOA) used for wavelength conversion. Antireflection and high-reflection coatings are for 1 and 2 wavelengths respectively. Optical implementation of a NOT based latch: Memory is the basic requirement to construct any electronic or optical processor. To develop a complete unit of digital optical memory, the first step is to develop a latch or a 1-bit memory unit as it can store a single bit. The proposed system described here is based on frequency encoding principle (Fig.2). Here two different frequencies are used for encoding 1 and 0 i.e. if frequency represents the state 0 then represents the state 1.A is the input terminal and Q and Q are the two output terminals. To implement the optical NOT based latch logic some beam splitters (BS), and mirrors (M) and some SOA based wavelength converters (WC) are used at different position of the system. Here two frequency selecting filters are used, where one is optical pass filter and other is optical pass filter. When frequency is applied at A, the light beam enters only to WC 1, but not in WC 4. Thus the frequency 13

162 42 J Opt 39(1) : Fig.2 Frequency encoded optical one-bit memory cell based on latch logic. (WC, M, CP, BS are represents the wavelength converter, mirror, cross polarizer and beam splitter or combiner respectively). of light beam which behaves as a week probe beam now falls on to the WC 1 from the input terminal A. Another proper constant beam strong pump light (CW) of frequency is injected to the WC 1. For which one can obtain the frequency of light which is again injected as a strong pump beam to the input terminal of WC 2. A constant weak probe beam of frequency (CW) is applied at the input of WC 2, for which an output of frequency is obtained at the terminal Q for the application of at A. Here CP s (cross polarizer s) are used in the output of every WCs to block the unwanted light beam i.e. for this CP at the time of absence of pump beam or probe beam no light beam will come out from the WCs. This output light beam is feedback as a pump beam to the WC 3 and a constant weak light of frequency is also applied on the other input of the WC 3 to get the output of at frequency which is divided into two parts by a beam splitter (BS). One part is feedback to the input of the WC 1 and the other part is sent to the output terminal Q. Thus for at A one can get at Q and at Q. Now frequency of a week probe beam of light is given at input A then it falls directly to the WC, but it will not at all enter at WC and a constantly supplied strong pump beam of frequency is applied on the other input channel of the WC. In this situation the frequency of light will come out from the output of WC which is again treated as a strong pump beam to fall on the WC. A constant week probe beam of frequency is kept in the input 13

163 J Opt 39(1) : channel of WC which helps to get the intense frequency of light at the output of WC (X). This is divided again into two more parts; one is feedback to the input of the WC as a strong pump beam where the other part comes at the output X, which is ultimately connected with Q by BS and M. Again a constant week probe beam of frequency given to the 2nd input of WC which ensures the output of frequency at the output Y which is ultimately connected with Q. A portion of the output at Y is feedback to the input of WC. This is the overall connection of the whole latch unit. Now to describe the operation it can be said that when frequency of light i.e. logic state (0) is applied at the input terminal A, the upper portion of the system described in figure-2, activated but lower half does not because of presence of the optical pass filter which only allows the frequency of light in the upper half of the unit and the pass filter only passes the frequency of light in the lower half of the system described in Fig.2. So when frequency of light i.e. logic state (0) is given to input terminal A one can get the frequency of light at output Q and frequency of light at Q i.e. Q=1 and Q = 0 when A=0. Similarly when frequency of light i.e. logic state 1 is applied at the input terminal A, the lower portion of the system takes the major role instead of upper portion and one can ensure the frequency of light at Q and at Q i.e. Q=0 and Q =1 when A=1. The most important and interesting point here is that if the input and frequency of light are withdrawn, the system will continue to show the last attended values of Q and Q at the final output because of the feedback mechanisms. At this situation the feedback light will continue to excite the respective WCs. So, this system can behave as a frequency encoded optical one bit memory. Optical implementation of two-bit memory cell Now slightly extending the one-bit memory cell circuit one can develop a two-bit memory cell (Fig.3). Here we attached two similar optical circuits of one-bit memory unit. Here A and B are the input terminals and X 1 and Y 1 are the output terminals. The block diagram of this scheme is shown in Fig.3. Here A* and B* are the two unit memory cells (latches), Q and Q are the output terminals of A* block and X 1 and Y 1 are the output terminals of B* block. X 1 is then combined with Q and Y 1 is combined with Q. So, an overall two-bit memory cell is developed. Now two inputs A and B and two outputs Q and Q, if A=1( ) and B=0( ) then one can obtain Q=0( ) and Q =1( ) whereas if A=0( ) and B=1( ) then Q=1( ) and Q =0( ). The states of the last output will be attended if there is no light signal present in the input terminal i.e. for the withdrawal of optical signal, the system to continue to show its last attended values at the output. The truth table of this memory unit is given in table 1. Table 1 Truth table of optical two bit memory cell. Input Output A B Q Q (1) (0) (0) (1) (0) (1) (1) (0) 0 0 Last Last state state attended attended 13

164 44 J Opt 39(1) : Fig.3 Block diagram of frequency encoded optical two-bit memory cell based on NOT latch logic. (A and B receive the frequency encoded inputs, Q and Q give the frequency encoded outputs). Here each block comprise the memory or latch as given Fig.2. Conclusion To conclude, we have proposed here a frequency encoded technique based optical NOT latch or memory cell. The whole system is all-optical one and is expected to support a high speed operation (far above GHz limit) which is a potential advantage of this mechanism. The coded information (0 or 1) of an output signal remains unaltered in reflection, refraction, absorption etc. due to this encoding principle. Therefore this technique will be very much useful in reliable optical communication. To achieve the faithful amplification the pump beam of WC should lie between 4dB to 10 db. The proposed system does not only offer a high speed operation but also offers noise free conversion to provide a high signal to noise (S/N) ratio. Using this latch logic and the frequency encoded technique one can implement many other operations like digital types of flip-flops, multiplexer, demultiplexer etc. with some modifications of the scheme. References 1. K. R. Chowdhury and S. Mukhopadhyay, A new method of binary addition scheme with massive use of nonlinear material based system, Chinese Opt Letts, 1(4), (2003) 13

165 J Opt 39(1) : P. Ghosh, P. P. Das and S. Mukhopadhyay, New proposal for optical flip-flop using residue arithmetic, ITCOM-2001, (4534B-22), SPIE Proceedings (Optoelectronic and Wireless Data Management, Processing, Storage and Retrieval), 4534, (2001) 3. A. K. Das and S. Mukhopadhyay, General approach of spatial input encoding for multiplexing and Demultiplexing, Opt Engineering (USA), 43(1), (2004) 4. K. Roy Chowdhury and S. Mukhopadhyay, Binary optical arithmetic operation scheme with tree architecture by proper accommodation of optical nonlinear materials, Opt Engineering, 43(1), (2004) 5. N. Pahari, D. N. Das and S. Mukhopadhyay, All-optical method for the addition of binary data by non-linear materials, Appl Opt, 43(33), (2004) 6. N. Pahari and S. Mukhopadhyay, An all-optical R-S flip-flop by optical nonlinear material, J of Opt, 34(3), (2005) 7. S. K. Garai, D. Samanta and S. Mukhopadhyay, All-optical implementation of inversion logic operation by second harmonic generation and wave mixing character of some nonlinear material, Optics and Optoelectronic Technology (China), 6(4), 39 42(2008) 8. L. Q. Guo and M. J. Connelly, A novel approach to all-optical wavelength conversion by utilizing a reflective semiconductor optical amplifier in co propagation scheme, Opt. Communication, 281(17), (2008) 9. L. Q. Guo and M. J. Connelly, A poincare approach to investigate nonlinear polarization rotation in semiconductor optical amplifiers and its applications to alloptical wavelength conversion, Proc. of SPIE 6783, (1 5) (2007) 10. H. J. Dorren, D. Lenstra, Y. Liu, Martin, T. Hill, G. D. Khoe, Nonlinear polarization rotation in semiconductor optical amplifiers: theory and application to all-optical flip-flop memories, IEEE Journal of Quantum Electronics, 39(1), (2003) 11. M. A. Karim and A. S. Awwal, John Wiley and Sons, Optical Computing An Introduction INC. (1992) 12. S. Dutta and S. Mukhopadhyay, All optical frequency encoding method for converting a decimal number to its equivalent binary number using tree architecture, Presented and Published in the proceedings of International Conference Of Optics and Photonics (ICOP-2009) in chandigar, 31st OCT. to 2nd NOV (2009) 13. S. K. Garai, S. Mukhopadhyay, A method of optical implementation of frequency encoded different logic operations using second harmonic and difference frequency generation techniques in non-linear material, Opt Int J Light Electron Opt, doi: / j.ijleo (2008) 14. D. Samanta and S. Mukhopadhyay, A method of maintaining the intensity level of a polarization encoded light signal, Journal of Physical Sciences (Vidyasagar University), 11, (2007) 15. B. Chakraborty and S. Mukhopadhyay, Alternative approach of conducting phase-modulated all optical logic gates, Opt Engineering, 48(3), (2009) 16. M. J. Connelly, Semiconductor Optical Amplifiers, Kluwer Academic publishers (2002) 13

166 Author's personal copy Optik 122 (2011) Contents lists available at ScienceDirect Optik journal homepage: Alternating approach of implementing frequency encoded all-optical logic gates and flip-flop using semiconductor optical amplifier Soma Dutta, Sourangshu Mukhopadhyay Department of Physics, Burdwan University, Golapbag, Burdwan , West Bengal, India article info abstract Article history: Received 18 December 2009 Accepted 16 June 2010 Keywords: Optical computation Non-linear optics Semiconductor optical amplifier Fiber optics In conduction of parallel logic, arithmetic and algebraic operations, optics has already proved its successful role. Since last few decades a number of established methods on optical data processing were proposed and to implement such processors different data encoding/decoding techniques have also been reported. Currently frequency encoding technique is found be a promising as well as a faithful mechanism for the conversion of all-optical processing as the frequency of light remains unaltered after refection, refraction, absorption, etc. during the transmission of light. There are already proposed some frequency encoded optical logic gates. In this communication the authors propose a new and different concept of frequency encoded optical logic gates and optical flip-flop using the non-linear function of semiconductor optical amplifier Elsevier GmbH. All rights reserved. 1. Introduction All optical signal and data processing is especially attractive for high speed and high capacity computation to avoid the speed related problems in optoelectronic processing systems. Because of the inherent character of parallelism of light can show more strong and potential applications in information processing, computing, data handling and image processing. In optical computation photon is found to be a very suitable information carrier than electron not only in the connection of super fast speed but in many other aspects of information processing also. Thus these photonic systems can successfully replace the electronic systems. Again it is also seen that in case of optical data processing the conventional methodologies cannot be followed always as it is done in electronics. Scientists and technologists are deeply involved in research to overcome the speed related difficulties to realize all-optical logic, arithmetic and algebraic operations with Boolean mechanism. There are found several popular reports on the development of optical logical systems where the logic gates are the basic building blocks. In this present communication, we focused on the successful realization of some efficient optical logic gates based on the frequency encoding principle, which can be used as a more reliable candidate than other encoding principles for the development of super-fast optical processors, as the frequency of a signal remains unchanged even after different optical transformations. In most of the cases, the presence Corresponding author. Tel.: addresses: soma.dtta@gmail.com (S. Dutta), sourangshu2004@yahoo.com (S. Mukhopadhyay). of optical signal at the input or output of a logical system is encoded as 1 bit and the absence of the signal is regarded as 0 logic state. As the intensity of light decreases with the increase of optical path through a medium, so the intensity of light may drop down below the reference level of the concerned logic state 1 and may enter into the reference level of logic state 0. To overcome this problem one can use the frequency of light for encoding a bit. If the presence of a specific frequency of light is treated as 1 state and then other specific one represents 0 state i.e. if 1 state is encoded by frequency 2 then that of the 0 state is done by another frequency by 1 where 1 and 2 remain unaltered throughout the transmission of data. There lies also many other encoding techniques like polarization encoding, phase encoding, etc. All these encoding principles have the same problem of changing the encoded state during transmission. Frequency encoding techniques is free from such transmission problems. There are reported some frequency encoded optical logic systems [1 3]. Here in this communication we propose an alternative and new approach for the implementation of some all-optical logic gates and flip-flops using frequency encoding principle [4 8]. To implement these logic gates and flipflops different methods are adopted, which show the switching advantages of implementation. Like logic gates different types of memories are also the essential part of any computing system. Already different types of volatile and non-volatile memories have been successfully developed for electronic data processing systems. On the other hand optical memories are also of great importance for the development of optical computing technology. There are found many reports where the schemes for developing all-optical memories, and flip-flops are reported [9 14]. Here in this communication the authors also pro /$ see front matter 2010 Elsevier GmbH. All rights reserved. doi: /j.ijleo

167 Author's personal copy S. Dutta, S. Mukhopadhyay / Optik 122 (2011) pose a method of developing all optical frequency encoded RS flip-flop based on some all optical frequency encoded logic gates with the potential uses of the switching character of semiconductor optical amplifier. 2. Frequency encoding principle We have already discussed above that the frequency of light is the fundamental character of light and it is unaltered after the reflection, refraction, absorption, etc. In optical computing and data processing the most important point is the very high speed of processing. In most of the cases the presence of optical signal at input/output end is encoded as 1 and absence of optical signal as 0. The interesting point is that the problem of variation of intensity of light in data processing due to increasing distance can be solved by using the frequency encoding principle of light. One can encode and decode two different states of information by two different frequencies. Here if 1 state is represented by a frequency 2 then 0 state is represented by another frequency 1. These frequencies remain generally unaltered during transmission. Here in this communication we encode and decode the state of information 0 by a beam of wavelength 1 (which is corresponding to 1 ) and the state of information 1 by another wavelength 2 (which is corresponding to another frequency 2 ) for the implementation of optical logic gates and based on this logic gates we proposed a method of implementing all optical frequency encoded RS flip-flop. 3. Semiconductor optical amplifier based switches (SOA) Semiconductor optical amplifier is generally based on GaAs material. It is used for developing several optoelectronic devices, which under suitable operating conditions can amplify an input optical signal [15 19]. SOA can be classified into two main types, one is the Fabry Perot SOA (FP-SOA), where reflection from the end surface is important i.e. the signal undergoes many passes through the amplifier, on the other side in case of Traveling Wave SOA (TW- SOA) the reflection is negligible i.e. a signal undergoes only a single passes through the amplifier. These behaviors of SOA can be used in optical transparent networking and digital optical processing. The non-linearity of SOA can also be used in many functional applications which are caused by carrier density induced by the amplifiers signal. There are found four types of non-linearity s in SOA, which are cross gain modulation (XGM), cross phase modulation (XPM), self phase modulation (SPM), and four wave mixing Fig. 1. Reflecting semiconductor optical amplifier (RSOA). (FWM). Here in this communication the authors exploit the cross gain modulation character of SOA. A weak CW probe light of wavelength 1 and a strong pump beam of wavelength 2 are injected to the input terminals of the SOA having an anti-reflecting surface at the input for 1 and a highly reflecting surface for 2 at the output terminal. The strong pump beam transfer its total power to the weak probe beam and then the weak probe beam becomes stronger and comes out to the output terminal i.e. the SOA acts as a proper wavelength converter. The scheme is shown in Fig. 1. This type of SOA switch is called RSOA. The switching action of RSOA can be used to develop optical logic gates and flip-flop. Again optical ADD/DROP multiplexer is a frequency selecting network. It is tuned in a particular biasing frequency and it reflects that particular frequency of light through it and passes all other frequencies of light. The whole scheme is shown in Fig Scheme of realization of frequency encoded optical OR logic operation Optical logic gates are the basic building block to implement any optical logical functions or operations. OR gate is one of the most important and a basic unit of integral logical system. To develop the system some ADD/DROP multiplexers, wavelength converters, mirrors (M) and beam splitters (BS) are used which is shown in Fig. 3. Here the input beam A, and B may have either the frequency of 1 or 2 (wavelength 1 or 2 respectively), where 1 frequency of light is encoded for 0 state and 2 frequency of light encoded for 1 state. Now a light beam of frequency 1 or 2 from point A falls on the 1st ADM which is tuned at the biasing frequency 2 then only 2 frequency of light is reflected by the ADM and is captured by the optical circulator C 1 and on the other hand the 1 frequency of light passes through the ADM and falls on the BS and one part of light falls on the 1st wavelength converter (WC) as a pump beam and another Fig. 2. ADD/DROP multiplexer.

168 Author's personal copy 1090 S. Dutta, S. Mukhopadhyay / Optik 122 (2011) Fig. 3. Frequency encoded optical OR gate. part of light falls on the 2nd wavelength converter (WC) as a pump beam, so for the case of the 1st WC if the probe beam is present then the WC works and one can get a converted strong probe beam at the output which is the input of the 3rd ADM and in absence of the probe beam one cannot get any output. Now in the 2nd WC there is a constant weak probe beam so if there is found a strong pump beam one can get then a converted strong beam at the output which is injected again to the 3rd WC as a pump beam. The captured tuned frequency of light from the first circulator C 1 of 1st input terminal A is injected to the 3rd ADM with the help of some mirrors and beam couplers. Here for the 2nd input terminal the similar type of input light ( 1 or 2 frequency) is injected to the 2nd ADM which is also tuned with the biasing frequency ( 2 ). So only this frequency of light is reflected and captured by the circulator C 2 and given to the 1st WC as a weak probe beam. Another frequency of light passes through the 2nd ADM and falls in the 3rd WC as a weak probe beam. If the pump beam and probe beam are present then only one can get a strong probe beam at the output otherwise absence of any probe or pump beam one cannot generate any output. The output injected to the 3rd ADM. So one can get the final output result of frequency encoded OR logic operation from the terminal Y. The scheme is shown in Fig Principle of operation of optical OR gate Now when the 1 frequency of light is applied to the first input terminal A the light passes through the 1st ADM and it is divided into two parts by the BS, where one part is treated as a strong pump beam to the 1st RSOA and other part is delivered to the 2nd RSOA also as a strong pump beam. One constant weak probe beam of 2 frequency is delivered to the 2nd RSOA, so that at the output it gives a strong light beam of 2 frequency which is again applied as a strong pump beam to the 3rd RSOA. Here one can apply a light of 1 frequency to the second input terminal B which passes through the 2nd ADM and delivered as weak probe beam to the 3rd RSOA. This light beam which is emerged an output beam of 1 frequency light and is applied to the 3rd ADM. Here to select the proper frequency of light one optical filter is kept. Again the 3rd ADM is tuned at its biasing frequency 1 so it reflects the output beam from 3rd RSOA and it is collected by the circulator and combined with the output terminal Y with the help of mirrors. Thus one can get an output result of 1 frequency of light. In this case (A = B = 1 ) there is no probe beam in the 1st RSOA, so 1st RSOA cannot work and one cannot get any output from the 1st RSOA. In the second case 1 frequency of light is given at the input terminal A and 2 frequency of light is to the second input terminal B. Here the 1st ADM passes the light and it is divided into two parts by the BS where one part is given to the 1st RSOA as a strong pump beam and other part is given also as a strong pump beam to the 2nd RSOA and for which one can get the output result of 2 frequency of light from the 2nd RSOA, which serves as a strong pump beam to the 3rd RSOA but as there is no probe beam in the 3rd RSOA because B = 2,sono 1 frequency of light comes to the 3rd RSOA and it does not convert any signal so one cannot get any output result from the 3rd RSOA. As the second input terminal B receives 2 frequency of light so the 2nd ADM (tuned at its biasing frequency 2 ) reflects the light beam and it is collected by the circulator C 2 as shown in Fig. 3. This light is applied as a weak probe beam to the 1st RSOA. So the 1st RSOA works and one can get the output result of 2 frequency of light, which is delivered to the 3rd ADM and we get the 2 frequency of light at the output terminal Y. In the third case when A = 2 and B = 1, 1st ADM reflects the light beam and it is collected by the circulator C 1 and applied to the 3rd ADM. Now the beam passes through the 3rd ADM and one can get the result of 2 frequency of light at the output Y. The rest of circuit cannot take part in this conversion. Now when both the two input takes 2 frequency of light the 1st ADM reflects the light beam and it is collected by the circulator C 1 and delivers it to the 3rd ADM, and passed through this ADM and we get the converted light beam of frequency 2 at the output terminal Y. Similarly in the 2nd ADM the light is reflected by the ADM and collected by the circulator C 2 and is delivered as a probe beam to the 1st RSOA, but as there is no pump beam in the 1st RSOA it cannot work and we cannot obtained any converted light beam from it. So that one can get Y = 2 when A = 1 and B = 2 ; when A = 2 and B = 1 ; Y = 2 ;

169 Author's personal copy S. Dutta, S. Mukhopadhyay / Optik 122 (2011) Table 1 Truth table of optical logic OR gate. A B Y 1 (0) 1 (0) 1 (0) 2 (1) 1 (0) 2 (1) 1 (0) 2 (1) 2 (1) 2 (1) 2 (1) 2 (1) Table 2 Truth table of optical logic AND gate. A B Y 1 (0) 1 (0) 1 (0) 1 (0) 2 (1) 1 (0) 2 (1) 1 (0) 1 (0) 2 (1) 2 (1) 2 (1) when A = 2 and B = 1 ; Y = 2 and finally when A = B = 2 ; Y = 2. This verifies the truth table of OR gate, if 2 is encoded as 1 and 1 by 0 (Table 1). 6. Scheme of realization of frequency encoded optical AND gate To implement optical AND gate we take two channels A and B and they may have either 1 (corresponding wavelength 1 ) frequency of light or 2 (corresponding wavelength 2 ) frequency of light. Now a beam of light of frequency 1 or 2 falls on the 1st ADM which is tuned in the biasing frequency 2 so only 2 frequency of light is reflected back by ADM and captured by the circulator C 1 and the 1 frequency passing through the ADM is injected to the 1st wavelength converter (WC) as a weak probe beam. A constant 2 frequency of strong pump beam of light is given to the 1st WC. If both the pump and probe beam is present in the RSOA (WC) then one can get the converted strong probe beam i.e. 1 frequency of light at the output. Now the reflected 2 frequency of light from the 1st ADM is captured by the circulator C 1 and is reflected by the mirror and is injected as a weak probe beam to the 2nd RSOA. Here from the channel B the input beam of light of frequency 1 or 2 falls on the 2nd ADM which is also tuned at its biasing frequency 2. So it reflects the 2 frequency of light and passes the 1 frequency of light through it. This light beam merges with the probe beam of light of 1st RSOA with the help of mirror. The reflected 2 frequency of light from the 2nd ADM is captured by the circulator C 2 and is given to the 3rd RSOA as a strong pump beam and 1 frequency of weak probe constant beam light is given to the 3rd RSOA. So when both pump or probe beam is present one can get a converted strong probe beam at its output. This output beam is injected to the 2nd RSOA as a strong pump beam with the help of mirrors. So if both pump or probe beam is present in the 2nd RSOA one gets the output beam as converted weak probe beam into a strong one which is added with the output beam of light of 1st RSOA using some properly oriented mirrors and these two light beams again are injected to the 3rd ADM which is tuned at its biasing frequency 2.Soit passes the 1 frequency of light and reflects the 2 frequency of light and it is collected by the circulator C 3. This reflected beam of light is ultimately added with the output beam from 3rd ADM by the use of properly oriented mirrors. The whole system is shown in Fig. 4. Here optical filters are used to select the proper frequency of light beam. Thus when A = 1 and B = 1 ; Y = 1, when A = 1 and B = 2 ; Y = 1, for A = 2 and B = 1 ; Y = 1 and finally when A = B = 2 ; Y = 2. This satisfies the truth table of AND gate. 7. Principle of operation of optical AND gate The AND logic system is shown in Fig. 4. Now in the case of AND logic when two input beams are of 1 frequency then 1st ADM passes the light beam and 2nd ADM also passes the light beam and then they are combined together by mirror and beam splitter. The combined beams are serving as a weak probe beam to the 1st RSOA. A constant strong pump beam is present in the 1st RSOA so it gives a converted output light beam which is delivered to the 3rd ADM and it passes through it. Thus a result of 1 frequency of light is obtained at the output end Y. The conversion of 2nd and 3rd RSOA cannot take part due to the absence of the either pump or probe beam. Again when A = 1 and B = 2, 1st ADM passes the light beam and the light is delivered as a probe beam to the 1st RSOA and similarly due to the above operation one can get the converted output light beam of 1 frequency at the output end Y. The rest of the circuit does not take part in the conversion process due to absence of either pump beam or probe beam. Now when A = 2 and B = 1 then 1st ADM reflects the light beam and the light is collected by the circulator C 1 and delivered as a weak probe beam to the 2nd RSOA. As the pump beam is absent here so conversion cannot take part in 2nd RSOA. Again as B = 1, so the 2nd ADM passes this light beam and is combined with the output terminal of the 1st ADM and serves as a probe beam to the 1st RSOA. Thus a similar conversion is occurred in this time and as the output light beam of frequency 1 is obtained at Y. Finally when the 2 frequency of light is applied to both input terminal A and B, 1st ADM reflects the light beam and is delivered as a probe beam to the 2nd RSOA. So, 1st RSOA cannot take part any role in this conversion. Now as B = 2 so the 2nd ADM reflects the light beam and the light is collected by the circulator, which is delivered again as a strong pump beam to the 3rd RSOA where probe beam of 2 frequency is already present. So the conversion takes place and one can get the output light beam of 1 frequency and it is delivered as a pump beam to the 2nd RSOA, for which the output light beam of 2 frequency from 2nd RSOA comes and it is applied to the 3rd ADM. This light is reflected back by the ADM and collected by the circulator C 3. The light from the circulator is connected to the output terminal Y. So then one can get the 2 frequency of light at the output Y. Thus when A = 1 and B = 1 ; Y = 1, when A = 1 and B = 2 ; Y = 1, when A = 2 and B = 1 ; Y = 1 and finally when A = B = 2 ; Y = 2. This satisfies the truth table of AND gate which is shown in Table 2, if 1 is encoded as 0 and 2 as Scheme of realization of frequency encoded optical NAND gate NAND gate is the most important logic gate in the logics family as it is universal gate. Again here we take two input channels A and B as sources of input light beams of frequency 1 or 2. The whole set up is much closed to the AND logic set up except three main changes seen in this system. Here the output light from 1st ADM is injected as a pump beam instead of the probe beam to the 1st RSOA and the reflected beam of light from 1st ADM is injected as a pump beam instead of a probe beam to the 2nd RSOA. Finally reflected beam from the 2nd ADM is applied as a pump beam to the 3rd RSOA (Fig. 5). Thus the truth table of the NAND logic is developed and it is shown in Table 3.SowhenA = B = 1 then Y = 2, when A = 1 and B = 2 then Y = 2, for A = 2 and B = 1 ; Y = 2 and finally for A = B = 2 ; Y = 1. This supports the truth table of universal gate NAND gate. 9. Principle of operation of optical NAND gate The diagram of frequency encoded NAND logic is shown in Fig. 5. Now in the case of optical NAND logic gate at first 1 frequency light is applied in both of the inputs A and B.So 1 frequency of light falls on the 1st ADM from A, and as it is tuned at its biasing frequency

170 Author's personal copy 1092 S. Dutta, S. Mukhopadhyay / Optik 122 (2011) Fig. 4. Frequency encoded optical AND gate. 2,so 1 frequency light passes through the ADM and is given as a strong pump beam to the 1st RSOA, one weak probe beam of 2 frequency is also given to the 1st RSOA for which a converted strong light beam of 2 frequency will be generated. This is again delivered to the 3rd ADM which is tuned also at 2 frequency, so the light beam reflected from the 3rd ADM is captured by the circulator C 3 and connected with the output Y by the use of mirrors and one can get the 2 frequency of light at the output Y. Now 1 frequency light is also given to the 2nd ADM from B and the ADM is tuned at same frequency 2, so the ADM passes the light beam which is combined with the output beam of the 1st ADM by the mirrors, etc. Now when A = 1 and B = 2 then similar operation goes on in the 1st ADM, but in this case of second input terminal B, the light beam falls on the 2nd ADM and gets reflected from it and being captured by the circulator C 2 it falls on the 3rd RSOA as a pump beam. So one can get the converted 1 frequency of light in presence of 1 probe beam at the output and it is again applied to the 2nd RSOA as a probe beam. Due to absence of pump beam the conversion does not occur and hence the 2 frequency of light is obtained at the output Y for the operation of the 1st section of the system. When A = 2 and B = 1 then 1st ADM reflects the light beam and it is then captured by the circulator C 1. This is given to the 2nd RSOA as a strong pump beam. No other conversion takes place now because of the absence of the probe beam. Here 1 frequency of light falls on the 2nd ADM and passing through it, this delivers a strong pump beam to the 1st RSOA and the conversion is occurred (due to presence of 2 frequency probe beam of light) and one receives the 2 frequency light at the output and it is then delivered to the 3rd ADM. Thus the final 2 frequency light is found at the output terminal Y. Finally Table 3 Truth table of optical logic NAND gate. A B Y 1 (0) 1 (0) 2 (1) 1 (0) 2 (1) 2 (1) 2 (1) 1 (0) 2 (1) 2 (1) 2 (1) 1 (0) when A = B = 2 ; 1st ADM reflects the light beam and is delivered to the 2nd RSOA as a strong pump beam. Again the light beam also falls on the 2nd ADM and gets reflected. It is captured then by the circulator C 3 and is given to the 3rd RSOA as a strong pump beam and one gets the converted output beam of light of 1 frequency in the presence of a weak probe beam of 1 frequency. This converted beam of 1 frequency of light is delivered as a probe beam again to the 2nd RSOA and one thus gets the converted output beam of 1 frequency of light which is delivered to the 3rd ADM. This 3rd ADM passes it and a 1 frequency of light is obtained at the output end Y. Thus one can ensure the logical NAND output from Y from the system described in Fig. 5. Hence one finds Y = 2 when A = B = 1, again when A = 1 and B = 2 then Y = 2, for A = 2 and B = 1, Y = 2 and finally when A = B = 2 then Y = 1. This supports the truth table of a universal frequency encoded gate NAND gate which is shown in Table 3. Here 1 is encoded as 0 and 2 as Scheme of realization of frequency encoded optical R S flip-flop Memory is the fundamental criteria for developing an all of optical processor. To realize this system with optics we consider two input light channels R and S, each of them may take either 1 (0) or 2 (1) frequency of light. One can get the output in the channels Q and Q, respectively. The outputs are feedback to the input i.e. Q is connected with input R and Q is connected with the other input S by the help of some mirrors and beam splitters. Now a beam of light of frequency 1 or 2 falls on the 1st ADM through point R. As the ADD/DROP multiplexer is tuned with its biasing frequency 2 so it reflects this frequency of light and passes the 1 frequency of light. This light is then introduced to the 1st RSOA as a strong pump beam. There already presents a constant weak probe beam of 2 frequency. So if both pump and probe beam is present one can get the converted strong probe beam of light with the respective frequency which again falls on the 3rd ADM. The absences of any pump or probe beam in the RSOA makes the conversion stop. Here one can use one optical filter to select the proper output light beam. The 3rd ADM is tuned with its biasing frequency 1 i.e. it

171 Author's personal copy S. Dutta, S. Mukhopadhyay / Optik 122 (2011) Fig. 5. Frequency encoded optical NAND gate. Table 4 Truth table of optical frequency encoded RS flip-flop. R S Q Q 1 (0) 2 (1) 2 (1) 1 (0) 2 (1) 1 (0) 1 (0) 2 (1) 0 0 Last state attended Last state attended reflects only 1 frequency of light and passes all other frequencies. The reflected beam is captured by the circulator C 3 and added with the output Q by the mirrors. Now the reflected light beam from the 1st ADM is separated by the circulator C 1 and is introduced to the 2nd RSOA as a strong pump light beam. A weak probe beam is also given to the 2nd RSOA and the converted output light beam passes through the respective optical filter and merges with the output beam of 1st RSOA. Ultimately the output from the 1st RSOA is given to the 3rd ADM. Same process is happened with the second input S and finally we get the two outputs Q and Q. The whole scheme is shown in Fig. 6. Now when one applies the light beam in the two input channels i.e. when R = 1 and S = 2 then Q = 2 and Q = 1 respectively and when R = 2 and S = 1 then Q = 1 and Q = 2 but when light is withdrawn from the inputs one can get the last state attended in the outputs Q and Q. So this follows the truth table of optical RS flip-flop which is shown in Table 4 and the figure of whole scheme of optical RS flip-flop is shown in Fig Principle operation of optical R S flip-flop To implement the circuit diagram of an optical RS flip-flop the light beam is supplied first to both the input channel. This delivered light is of frequency 1 to the 1st ADM through the input end R. As the 1st ADM is tuned at its biasing frequency 2,so 1 frequency light beam passes through it and falls on the 1st RSOA where a strong beam of 2 frequency is already present. So one gets the converted 2 frequency of light beam at the output of 1st RSOA and it is again introduced to the 3rd ADM (tuning frequency 1 ) and light passing through the ADM comes to the output end Q. This output beam of 2 frequency of light splits into two parts by the BS, one part is sent to the output end Q and another part is feedback to the second input terminal S. Again if a 2 frequency light is applied to the S terminal, a combined beam of 2 frequency of light is delivered to the 2nd ADM (tuning frequency 2 ). This ADM reflects by it and sends to the 4th RSOA by the help of circulator C 2 and mirrors. Due to presence of a weak probe beam of 1 frequency of light in 4th RSOA, the conversion is occurred and one can get the converted output light beam of frequency 1 which is introduced to the 4th ADM (tuning frequency 1 ) and is reflected by it. This light is captured by the circulator C 4 and is combined with the output Q and splitted by the BS. One part goes to the other output end Q where another part is feedback to the input terminal R and the process continues. Now when the R = 2 and S = 1 then reverse process is developed in this circuit and one gets Q = 1 and Q = 2 in the output ends. Now when no light passes through both the input channel i.e. for the absence of light in the input terminals, last state is attended i.e. If the last state is Q = 1 and Q = 2 for (R = 2 and S = 1 ) then one gets the same result in the output terminals. As because the output terminals are feedback to the reverse input terminals, so always there is found a light present in the input terminals (even if no external input is applied) and that s why the conversion process continues. So certainly when R = 1 and S = 2 then Q = 2 and Q = 1 respectively and when R = 2 and S = 1, Q = 1 and Q = 2, but when there is no input light beam is applied at all i.e. when the inputs are withdrawn one can ensure the last state attended in the output terminals Q and Q. So the truth table of an optical RS flip-flop is followed. 12. Important requirements for the switching of SOA To obtain the faithful operation the power of the input control light beam which, serves as a pump beams for the SOA should lie between 2 and 4 db. The performance of the data transfer depends on the pomp beam energy. Energy of each probe beam is to be maintained between 4 and 2 db. The wavelengths of the selected inputs are 1555 and 1550 nm corresponding to frequency 1 and 2, respectively. Frequencies of CW input signal should lie in C band ( nm). This wavelength range is favorable for optical communication.

172 Author's personal copy 1094 S. Dutta, S. Mukhopadhyay / Optik 122 (2011) Fig. 6. Frequency encoded optical RS flip-flop. 13. Conclusion To conclude, we have proposed an all optical approach for the successful realization of high speed (far above GHz range) optical logic gates and memories. The potential advantage of these optical gates and flip-flop over many other established optical gates was the use of frequency encoding technique, for which the coded information (0, 1) in a signal remains unchanged in refraction, reflection, absorption, etc. for a long distance transmission of data. The proposed system could offer also a noise free conversion to provide a high signal to noise (S/N) ratio. Using such frequency encoded technology based optical logic gates and flip-flops one can implement many other digital operations like multiplexer, demultiplexer, multivibrators, etc. References [1] S.K. Garai, S. Mukhopadhyay, Method of implementing frequency encoded multiplexer and demultiplexer systems using nonlinear semiconductor optical amplifiers, Opt. Laser Technol. 41 (8) (2009) [2] S.K. Garai, S. Mukhopadhyay, Method of implementation of all-optical frequency encoded logic operations exploiting the propagation characters of light through semiconductor optical amplifiers, J. Opt. (2009), doi: /s [3] S.K. Garai, A. Pal, S. Mukhopadhyay, All-optical frequency encoded inversion operation with tristate logic using reflecting semiconductor optical amplifiers, Optik (2009), doi: /j.ijleo [4] S.K. Garai, S. Mukhopadhyay, A method of optical implementation of frequency encoded different logic operations using second harmonic and difference frequency generation techniques in non-linear material, Optik Int. J. Light Electron. Opt. (2008), doi: /j.ijleo [5] Y. Ichioka, J. Tanida, Optical parallel logic gates using a shadow-casting system for optical digital computing, Proc. IEE 72 (7) (1984) [6] J.M. Jeong, M.E. Marhic, All-optical logic gates based on cross-phase modulation in a non-linear fiber interferometer, Opt. Commun. 85 (5 6) (1991) [7] B.K. Jenkins, A.A. Sawchuk, T.C. Strand, R. Forchheimer, B.H. Soffer, Sequential optical logic implementation, Appl. Opt. 23 (19) (1984) [8] T.A. Ibrahim, R. Grover, L.-C. Kuo, S. Kanakaraju, L.C. Calkoun, P.-T. Ho, Alloptical AND/NAND logic gates using semiconductor microresonators, IEE Photonics Technol. Lett. 15 (10) (2003) [9] P. Ghosh, P.P. Das, S. Mukhopadhyay, New proposal for optical flip-flop using residue arithmetic, in: ITCOM-2001, (4534B-22), SPIE Proceedings (Optoelectronic and Wireless Data Management, Processing, Storage and Retrieval), vol. 4534, 8 November, 2001, pp [10] A.K. Das, S. Mukhopadhyay, General approach of spatial input encoding for multiplexing and De-multiplexing, Opt. Eng. (U.S.A.) 43 (2004) [11] K.R. Chowdhury, S. Mukhopadhyay, Binary optical arithmetic operation scheme with tree architecture by proper accommodation of optical nonlinear materials, Opt. Eng. 43 (2004) [12] N. Pahari, D.N. Das, S. Mukhopadhyay, All-optical method for the addition of binary data by non-linear materials, Appl. Opt. 43 (33) (2004) [13] N. Pahari, S. Mukhopadhyay, An all-optical R S flip-flop by optical non-linear material, J. Opt. 34 (3) (2005) [14] K.R. Chowdhury, S. Mukhopadhyay, A new method of binary addition scheme with massive use of non-linear material based system, Chin. Opt. Lett. 1 (April (4)) (2003) [15] M.J. Connelly, Semiconductor Optical Amplifiers, Kluwer Academic publishers, [16] L.Q. Guo, M.J. Connelly, A novel approach to all-optical wavelength conversion by utilizing a reflective semiconductor optical amplifier in co propagation scheme, Opt. Commun. 281 (2008) [17] L.Q. Guo, M.J. Connelly, A poincare approach to investigate nonlinear polarization rotation in semiconductor optical amplifiers and its applications to all-optical wavelength conversion, Proc. SPIE 6783 (1 5) (2007) [18] H.J. Dorren, D. Lenstra, Y. Liu, M.T. Hill, G.-D. Khoe, Nonlinear polarization rotation in semiconductor optical amplifiers: theory and application to all-optical flip-flop memories, IEEE J. Quantum Electron. 39 (2003) [19] M.A. Karim, A.S. Awwal, Optical Computing An Introduction, John Wiley and Sons, Inc., 1992.

173 Optics and Photonics Letters Vol. 3, No. 1 (2010) c World Scientific Publishing Company DOI: /S ALL-OPTICAL APPROACH FOR CONVERSION OF A BINARY NUMBER HAVING A FRACTIONAL PART TO ITS DECIMAL EQUIVALENT AND VICE-VERSA SOMA DUTTA and SOURANGSHU MUKHOPADHYAY Department of Physics, The University of Burdwan Burdwan, pin , West Bengal, India soma.dtta@gmail.com sourangshu2004@yahoo.com Received 25 October 2010 Optics is found as a very potential candidate in information processing and computing. Several all optical methods have been proposed for implementation of all optical logic and arithmetic devices during the last few decades. In this regard, the role of optical tree architecture can be mentioned as an important approach for conversion of an optical data from binary to decimal and vice-versa. In this communication, the authors propose a new concept of conversion of a binary number having some fraction (fractional value) to its equivalent decimal counterpart and its vice-versa. To perform this operation, optical tree and some nonlinear material based switches are used properly. Keywords: Non-linear optics; optical computation; optical tree-architecture. 1. Introduction Optics has already been established as a potential and promising candidate in information and data processing. Many all-optical logical, algebraic and arithmetic processors have been proposed since the middle of the decade of seventy. 1 5 Several schemes of all optical logic gates, optical digital memory units, optical algebraic processors, image processors etc were proposed by scientists and technologists The choice of optical signal in replacement of conventional electronic signal in data processor is mainly because of the inherent parallelism in optics, which can lead to a super fast up-gradation of computing technology. The major developments conducted in this area are also discussed in the 1st and 2nd International Conferences on Photonic Computing. 12,13 In connection to the new developments of several all-optical data processing techniques, the role of optical tree architecture can be mentioned specially. This tree has already been used to convert a position-wise encoded optical decimal data to its binary counterpart and from binary data to its decimal counterpart. 14 Not only these conversions but there are also several other conversions which are possibly conducted by the use of this tree architecture for 51