All-Optical Logic Gates : Designs , Classification , and Comparison

The paper reviews the current status and designs of all-optical gates. Various schemes with and without semiconductor optical amplifiers are discussed and compared. The optical gates are classified according to their design structures. It is divided into two major divisions that is, nonsemiconductor optical amplifier based gates and semiconductor optical amplifier based gates. In nonsemiconductor optical amplifier based gates, different schemes have been proposed to create non-linearity which is discussed. The semiconductor optical amplifier based gates of different design structures are discussed to show the probe pulse that is modulated in different ways to obtain results.


Introduction
Today the demand for high bandwidth has rapidly increased to obtain the speed limit of electronic devices.The general purpose of all-optical signal processing is still on the horizon.Nowadays, prototype of all-optical logic gates at high bitrate are coming out from the laboratories.The researches are going forward in this field to make it possible.However, in optical signal processing the digital gates have complicated and cumbersome electrooptic conversion.To make alloptical systems, it is necessary that entire components which are used in optical networks such as add-drop multiplexer, packet synchronization, clock recovery, address recognition, and signal regeneration, and so forth should be all-optical elements.To make the dream come true the basic requirement is optical gates.Gates are the key elements to realize all-optical functions.Thus, to realize digital gates into alloptical logic gates at the same platform, it is necessary to develop several basic designs.It is impractical to design an optical component with some gates using ultrafast nonlinear interferometer (UNI), some gates with SOA and some with high nonlinear fiber (HNLF).So the design is only successful when all the gates are implemented with same technique.Alloptical logic gates are core logic unit to implement various alloptical systems for optical signal processing.To design optical gates it is necessary to implement a nonlinear medium which modulates the signal to produce the desired results.The nonlinearity may be generated in numerous ways such as using nonlinear loop mirror, nonlinear fiber, photonic crystal, filter, waveguide, thyristor, acoustic waves, or semiconductor optical amplifier.Therefore, there are many researches going on to realize all-optical signal processing systems which are already discussed in various papers.It may be classified in multiple ways according to the design structures.All-optical gates are divided in two basic structures as in Figure 1 which are without SOA and with SOA.

All-Optical Gates
All-optical gates may be constructed using the nonlinearity effect which is introduced without SOA or with SOA.Numerous ways of all-optical gates without SOA using length of the fiber, waveguide, circulator, filters, acoustic-optic waves, and changing the refractive index of the optical waveguide have been discussed in the first part.Gates constructed with SOA are discussed in the second part.

All-Optical
Gates without SOA.The nonlinearity in silica fiber arises from the nonlinear index of refraction.The change in nonlinear refractive index gives rise to an intensity dependent phase of the optical field.The effect of nonlinear interaction between two copropagating signals in the fiber can be expressed by the change in electric field on one of the signals caused by the other after propagating through some distance in fiber.The intensity dependent refractive index of silica medium gives rise to three effects, self-phase modulation (SPM), cross gain modulation (XGM), and four wave mixing (FWM).Self-phase modulation (SPM) occurs when intensity modulated signal travels through an optical fiber.The peak of the pulse travels slower than the wings.Due to this the wavelength of a pulse is stretched at the leading edge of the pulse and compressed on the trailing edge.Therefore, the trailing edge acquires a "blue shift" and leading edge acquires a "red shift." This modulates the signal and leads to broadening of the pulse.
Cross gain modulation (XGM) is another way in which intensity fluctuations affect the phase of a signal.Chromatic dispersion plays a significant role in gain modulation of the signal.Thus, the intensity fluctuation in the signal power of one channel propagating in the fiber modulates the phase of the other channel.
Four wave mixing (FWM) is a third-order nonlinearity and analogous to intermodulation distortion in the optical system.It is produced when beating between two channels at different frequencies modulates the signal phase at that frequency, generating new tones as side bands.The power of the side bands is always less than the signal power [1][2][3].

Dispersion Shifted Fiber/High Nonlinear Fiber (DSF/HNLF).
The first design of Figure 2 consists of length of a fiber which introduces nonlinearity in the propagating signal.The length of the fiber introduced in the designs is of three types, dispersion shifted fiber (DSF), high nonlinear fiber dispersion shifted fiber (HNLF-DSF), and high nonlinear fiber (HNLF).The first design of Figure 2(a) introduced DSF that adds a constant shift due to the selfphase modulation (SPM) between the counterpropagating data at the output [4][5][6].The counterpropagating data produces a cross phase modulation (XPM) which changes a pump power level and produces a constructive and destructive interference at the output.Both the data of same wavelength is used to avoid the four wave mixing (FWM).
In case of HNLF-DSF (Figure 2 signals is present and launched into the HNL-DSF, it appears at the output, but if both the signals are present, no significant power appears at the output due to the state of polarization between the given data.By coupling these two output signals, one can achieve the desired results depending on the strength of the XGM [7,8]. In the design (Figure 2(c)), all-optical logic gates are realized on the nonlinear polarization rotation (NPR) in HNLF.The polarization of light depends upon the intensity and relative polarization of the signal.When pump and probe signals travel through the HNLF, it introduces a nonlinear phase shift due to the SPM and XPM.The polarization of a probe signal changes and different gates can be realized [9][10][11][12].This type of gate design using the length of fiber to produce a phase shift makes the design bigger and bad competitor.

Waveguide Configuration.
In the design of Figure 3(a) two data A and B of different wavelengths are generated through microelectronic and mechanical system (MEMS) external cavity tunable laser.Both the data are coupled through mirror into the Fabry Perot chip (FP-chip).FP-chip basically has multimode wavelength output.As data passes through the cavity the signal experiences the nonlinearity effect.Due to the nonlinearity of the signals FP-chip is optically locked.The band pass filter (BPF) selects the wavelength which results as a gate operation [13].
The design consists of lights of different wavelengths injecting into the Si-wire waveguide with different peak power as in Figure 3(b).While travelling through the waveguide they experience two photon absorption which gives rise to the cross gain modulation.By adjusting the proper power of the pump and probe pulse we can get the results [14].
In Figure 3(c) two phase encoded data streams of different wavelengths are generated by two clock wave lasers and modulated with 33% duty cycle to produce RZ data.Both the data are copropagated with a continuous wave and fed into the chalcogenide As 2 s 3 (ChG) waveguide.The ChG waveguide offers broadband and flexible wavelength operation with ultrafast nonlinear response due to the kerr nonlinear index coefficient [15].Due to a nondegenerate FWM process new wavelength is generated which is fed into a tunable band pass filter (TBPF) to extract the gate output.
The design without pump is used in periodically poled lithium niobate (PPLN) waveguide to produce gate output [16][17][18].Here the waveguide is used to produce sum frequency generation (SFG) depending on the guide length.Two data are injected into the waveguide in which SFG occurs under the quasi-phase matching condition.When both of the data are the same the signal is depleted during the generation of sum frequency wave and finally output will be zero.If any one of the data is high, only one data will be depleted, and simultaneously another will still exist at the output.

Circulator.
In Figure 4, two data of different wavelengths are generated which are passed through first 2 × 2 coupler and second 2 × 2 coupler.Now both pump and probe are passed through Fabry Perot laser diode (FP-LD).The unique property of FP-LD is that if a single mode beam with slightly higher wavelength of longitudinal mode of FP-LD is injected, the beam will experience the gain while the other modes of FP-LD are suppressed.When second beam is injected to the FP-LD with a detuning range higher than the previous beam, the first beam will be suppressed while the second beam will experience a gain and also FP-LD will be locked by the second beam.This is a "gain modulation" technique in which pump power should be more than probe signal [19].

Optical Channel-Dropping Filter.
Novel design of alloptical logic AND/OR designed by using dark bright soliton conversion is shown in Figure 5.Here the dark (D) and bright (B) solitons represent the input logic "0" and "1", respectively.The input of stage-1 optical channel dropping filter (OCDF) is dark soliton (logic "0") which is a control light pulse.In stage-1 OCDF optical filter, the dark soliton is converted into dark and bright solitons.Next the data A is fed into stage-2 A/D filters, and then data B is fed into stage-3 OCDF filters.Stages are divided according to the sequence of the input given.OCDF is composed with two sets of coupled waveguides which produce a phase shift of  with respect to the input signal [20].

Multilayer
Waveguide.Several all-optical devices using optical nonlinearity have been proposed and implemented.
In the design a multilayer planer waveguide with nonlinear guided film is taken as shown in Figure 6.The waveguide is divided in three sections L 1 , L 2, and L 3 which corresponds to nonlinear three branch output, nonlinear double trapped waveguide, and linear two branch input sections, respectively.Branching angle between inputs is  and output is .The refractive index of the upper three arms and lower two arms is  0 and   , respectively.By changing the nonlinearity of the output branch and launching the input power accordingly, the optical gate can be verified [21,22].
2.1.6.Double Heterostructure Optical Thyristor.In Figure 7 authors demonstrate a monolithically integrated vertical cavity laser with depleted optical thyristor (VCL-DOT) structure which can configure into many optical logic functions using a simple operating technique by changing the condition of the driving voltage.As in Figure 7 is a bistable PnpN active region device.If the forward bias is applied to the thyristor, the s-shaped current-voltage characteristics are divided in three distinct stages, forward blocking region, negative resistance region, and forward conduction region.The forward conduction region is known as ON-state and the forward blocking region is known as OFF-state for the optical thyristor.Boolean optical gates can be realized by connecting these thyristors in series or in parallel and changing the reference voltage [23].

Acousto Optical Tunable Filter (AOTF).
Operation of gates is based on the switching of TE and TM polarized data A TE and B TM and vice versa as in Figure 8.The polarized data are then modulated in pulse position modulation (PPM).
Then the polarized data are excited through SAW transducer.
Pulses pass through acoustic-optic waveguide that determine the wavelength to be transmitted.The Boolean logic gates are verified by the temporal displacement of the output pulse [24].

All-Optical
Gates with SOA.Again the gates are divided according to the interferometer techniques such as ultra-high nonlinear interferometer (UNI), sagnac interferometer (SI), Michelson interferometer (MI), Mach-Zehnder interferometer (MZI), and delay interferometer (DI) to implement the nonlinearity in SOA.In the following sections the nonlinearity of SOA may be used in several ways.Different design structures and categories of SOA-based all-optical gates have been investigated in this section.SOA is a small size nonlinear amplifier that offers advantages to be integrated to produce a subsequent system essential in optical communication system.The SOAs exhibit low power consumption and their single mode waveguide structures make them particularly appropriate for use with single mode fiber [3].At present, SOA is the most developed optical amplifier that makes a rapid progress towards optical signal processing.The nonlinearity effect in SOA makes it a promising module for optical logic gates.The three nonlinearity effects that is cross gain modulation (XGM), cross phase modulation (XPM), and four wave mixing (FWM) make it possible to use it as nonlinear medium for gates.
In XGM data pulses at one wavelength, modulates the carrier density and at the same time results as a gain variation indentation in inverted copy of the clock pulse injected into the SOA as shown in Figure 9(a).Due to the modulation of a carrier density there is a gain compression in the pump signal that produces a chirping of the converted signal.The SOA is operated under the high optical intensity to reduce the gain recovery time.The problem related to XGM is at longer wavelength extinction ratio penalty associated with it.This phenomenon can be easily accommodated at high bit rate.
The chirp of the converted signal is used as an advantage byincluding the SOA in an interferometer configuration that converts this XPM into an intensity modulation.This can be done by SOA, incorporated with interferometer configuration.To obtain a complete extinction in an interferometer a phase shift of  is needed as in Figure 9(b), which can be achieved with gain compression in SOA.The phase shift is independent of wavelength, so the conversion to a longer wavelength has no problem with XPM.The disadvantage of an interferometer structure is that if the phase shift increases more than , it impairs the extinction ratio which may be controlled by changing the bias condition of SOA.The interferometer configuration may be defined in two ways, copropagation and counterpropagation.In copropagation, filter is required because pump and probe travel in the same direction to filter the probe signal with pump.But in counterpropagation both travel in opposite directions, so the filter is not required.
In FWM two signals of different wavelengths are injected into the SOA.On passing through SOA there is an intensity beating which arises due to the difference in frequency modulated signals in SOA.If the frequency separation is small the carrier density will be modulated.If the frequency separation is large, the modulated carrier will set up a moving grating in the active strip of SOA.The grating scatters the input signal and produces the sidebands which are located at the lower and higher frequency between the input signals.
The power of the side bands is usually less as compared to the signal power as in Figure 9(c).It is a process which depends on the phase of the optical signal instead of their intensity.It is a polarization dependent phenomenon and capable of handling intensity modulation, phase modulation, and frequency shift keying signal.As it depends upon distance between the signals and converted wavelengths, therefore the conversion efficiency is adequately affected.Therefore the scheme is not used in all-optical network.The application of FWM is used in dispersion management by optical phase conjugate.The process produces a mirror image of the original signal which is oppositely chirped in a spectral domain [25].

Ultranonlinear Interferometer (UNI) Configuration.
The concept of operation of the UNI gates relies on polarization rotation of the incoming signals to be switched in the presence of a switched pulse in SOA.It is divided in two copropagating UNI gates and counterpropagating UNI gates.
(1) UNI Copropagating Gates.According to the design structure UNI copropagating gates can be divided in two categories.In the first design, the data and clock directly inter into SOA and on the other hand the data is first modulated through delay interferometer (DI) and then send to SOA.  shift between two orthogonal polarized components of the signal [26][27][28][29][30][31][32][33][34][35][36][37].Here the clock pulse is orthogonally polarized and delayed after passing through polarization maintained fiber (PMF).The phase and amplitude of modulated data with higher power are copropagated through SOA.If both the data are present or absent the differential phase shift between the probe signals includes destructive interference; therefore output will be zero.If only one data is present, the phase change in the probe is adjusted to introduce constructive interference at the output.Another technique of dual SOA-UNI gate is shown in Figure 10(b).It may be constructed with two SOAs in UNI based elements [38].
(ii) All-Optical Conical Logic Unit.The operational principal of the design is divided into two stages as shown in Figure 11 [39,40].In stage-1 the carrier wavelength of DPSK signals is sent simultaneously to delay interferometer (DI).DI is an asymmetric Mach-Zehnder interferometer (MZI) with differential delay and tunable phase shift in both the to get Buffer and NOT operation at both the output ports, respectively.The output of DI is then launched into stage-2.In stage-2 both the data are injected into the SOA, out of which one acts as a pump and other as a probe.After passing through SOA, there is a cross gain modulation (XGM) and the output that is minterms ( 2 ) or AND operation is selected accordingly at the output of tunable band pass filter (TBPF).Again, the full set of minterms are combined together with original data directly to obtain maxterm ( 2 ) or OR operation.
(2) UNI Counterpropagating Gates. Figure 12 shows the counterpropagation in UNI gates where the clock and data signal propagate in opposite directions.Therefore the pump signal passes through SOA causing the carrier depletion in SOA.
The carrier depletion leads to gain saturation in SOA.Due to this, there is a marked intensity reduction of an incoming probe signal, which leads to no pulse existence for an output signal [41,42].If two SOAs are used in parallel, the output of first can be used to construct the multifunctional logic gates.

Sagnac Interferometer (SI)
Gates.The design as in Figure 13 consists of optical fiber loop with SOA placed asymmetrically.This gate is using the principle of TOAD (terahertz optical asymmetric demultiplexer).The offset position of SOA is controlled to obtain short switching window using nonlinearity in SOA.The sagnac interferometer gate consists of a 2 × 2 coupler which is used to join input port and output port.To maintain the polarization state of the fiber, polarization controller (PC) is used with polarization maintained fiber (PMF).The clock signal propagates through the coupler and splits in two equal parts with phase difference of /2.One will travel in clockwise (CW) direction and the other will travel in counter-clockwise direction (CCW) in the fiber loop.The time asymmetry between CW and CCW is maintained at Δ. Orthogonally polarized data enters through the polarization selective coupler (PSC) into the fiber loop.After passing through SOA there is a XPM between two counterpropagating probe signals.If any data is present due to differential phase shift, two probes interfere constructively on looping back to the coupler resulting as one.If both the data are present or absent the two probes interfere destructively and the output become zero [43][44][45][46][47][48][49][50].

Michelson Interferometer (MI)
Gates.The MI arrangement as depicted in Figure 14 is half or folded version of MZI of the counterpropagation scheme.The SOA is placed at the upper and lower arm of MI.When data of same wavelengths A and B passes through both SOAs in the opposite directions of clock wave signal, the refractive index produces phase variation in the medium of SOA.This modulates the clock wave signal incorporating the phase modulation at the output terminal.The circulator is used to recombine the reflected clock wave from both the SOAs.When constructive interference between two interferometer paths is maintained, the circulator output produces a converted signal.And destructive interference between two interferometer paths is maintained; the circulator produces no signal.It comprises a simple structure utilizing only one coupler, smaller in size, and requires less signal power than MZI [51][52][53].

Mach-Zehnder Interferometer (MZI) Configuration.
There are several possibilities for realizing the optical gates utilizing XPM in SOA based interferometer configuration.The gates comprise two SOAs located in the two paths or two arms, in which phase to amplitude modulation can be obtained when a relative phase difference is introduced in the interferometer.This phase difference may be produced in various ways, such as using 2 × 2 coupler, inserting a phase shifter in both the arms, or using different values of SOA, and so forth.
(1) MZI Copropagation.The MZI configuration may be divided into two categories that are MZI in copropagation and MZI in counterpropagation.In MZI copropagation, probe and pump propagate in the same direction.Thus, in such configuration band pass filter is required to filter the probe pulse from pump.
(i) MZI Copropagating Gates.Copropagation MZI is a suitable device for high speed all-optical demultiplexing.It operates on the principle of phase change, caused by the ray of light propagating through the 3 dB coupler.Gates consist of a symmetrical MZI with two SOAs placed in the upper and lower arm of the interferometer, as shown in Figure 15.In order to perform the operation two data streams enter in the upper and lower arms of interferometer through multiplexers.A continuous clock pulse through mode lock laser (MLL) enters through upper arm of first 3 dB coupler.When both pulses pass through SOA, the XGM will take place due to gain saturation in SOA.Data and clock pulses of different wavelengths are injected into SOA, operated under the gain saturation condition, where the available optical gain is distributed between two wavelengths depending on their relative photon densities.Therefore, the data is transferred in the clock pulse but in inverted form.It happens in both, that is, lower and upper arms of interferometer.After passing through first 3 dB coupler, the phase difference of /2 is created between upper and lower arms of clock pulse, which is travelling through the interferometer.When saturated pulse passes through second 3 dB coupler the total phase shift becomes .If both the data have same value, they will cancel because of  phase difference; therefore, at transmission port zero will appear.If both the data have different value then it will not cancel; therefore, one will appear at transmission port [54][55][56][57][58][59][60][61][62][63][64].
(ii) MZI Copropagating Push-Pull Gates.The principle of operation of push-pull configuration is shown in Figure 16.Two data streams A and B of same wavelengths are entered through upper and lower arm of MZI.The data A in the upper arm is ahead of one bit period to data B which is travelling in the lower arm.Similarly, lower arm data B is one bit period ahead to upper arm data A. This creates a switching window for data streams.Both the data are copropagated with clock pulse which enters through 3 dB coupler.When data A is 1 and B is zero, the pulse from data A splits into two parts.One pulse is pushed to the upper SOA and other is delayed by the switching window.Thus, the upper SOA is switched before the lower SOA.Therefore, the MZI is unbalanced and clock wave is switched at the T-port.If data A is zero and data B is one, then lower SOA is switched and again clock wave appears at T-port.If both the data are the same, then the SOAs are equally affected by the injected pulse.Consequently, the respective push and pull pulse temporarily coincide with each other and zero phase difference is introduced between the two arms.Thus, no switching occurs at T-port [65][66][67][68][69][70].
(2) MZI Counterpropagating Gates.In counterpropagation the clock and data pulse propagate in opposite direction through MZI as in Figure 17.If any of the data is one, there is a XPM between the clock and data pulse inside SOA that creates the differential phase shift between the two clock components which unbalance the MZI and the clock pulse exits at the T-port.However, if both the data are the same, the total phase shift will become  and clock pulse is cancelled at T-port.In this configuration no filter is required as data and clock pulse counterpropagate through the arm of MZI.Thus, at T-port there is no data pulse to filter [71][72][73][74].

Delayed Interferometer (DI)
Gates.The delay interferometer (DI) in Figure 18, the clock, and two data pulses of same wavelengths are injected into the SOA simultaneously [75,76].The data pulses will include a XGM, which leads to invert the clock signal in SOA.Then, the clock enters into DI and splits into two signals propagating through the two arms of DI with equal amplitude which interferes at the output.One arm of DI is delayed by Δ relative to the other arm.Therefore, at the output there is a relative phase difference between the two signals that interferes and produces an output.The XPM phenomenon produces a time-varying phase change on the clock pulse.The delay Δ determines the speed and the signal-to-noise ratio needed for good signal quality.A low Δ would restrict the speed of the data processing and also degrade the signal to noise ratio.

Conclusion
As discussed in the paper, all-optical gates of different designs have their utility but still the competitors are digital gates, which are compact and easy in coupling.If one goes for speed, optical gates are leading but in the case of extinction ratio (ER) the digital gates are still superior.The different designs use different nonlinear elements to create modulations.In Tables 1 and 2, different design structures are compared according to the polarization sensitivity and integration capacity.The nonlinear elements used in optical gates are length of the fiber, crystal waveguide, circulators, filters, thyristor, acoustic tunable filter and SOAs as discussed above in different schemes.Some optical gates such as HNLF/DSF, sagnac interferometer and delay interferometer configurations are bigger that make it inconvenient as compared to digital gates.Some gates using circulators, optical channeldropping filters, thyristor, and acoustic tunable filter have been reported less number of logical gates but still they are compact in size.The MZI copropagation waveguide structure design reported all types of logical gates with high operating speed and good extinction ratio that makes them superior than others.

Figure 1 :
Figure 1: Classification of all-optical logic gates.

Figure 5 :
Figure 5: Design of gate consisting of A/D filter.

Figure 8 :
Figure 8: Design of gate using acoustic-optics tunable filter.

Table 1 :
Comparison between non-SOA based gates.

Table 2 :
Comparison between SOA based gates.