A photonic crystal fiber based highly sensitive sensing mechanism is proposed, in the terahertz frequency band, able to detect a wide range of analytes, such as toxic or non-toxic chemicals and illicit drugs. The proper optimization of the PCF structure increases the light-matter interaction in the core, which results in a high relative sensitivity of about 94.0% with negligible confinement loss at the optimum frequency. Few liquids, chemicals, and drugs are considered to justify the sensing mechanism: a relative sensitivity of 99.60% can be achieved for the maximum porosity of core while ketamine was the analyte. Other fiber properties are also analyzed to check the feasibility of the proposed fiber with standard fiber and have obtained good performance. Therefore, the sensor may find applications to sense a wide range of analytes, non-toxic and toxic chemicals, as well as illicit drugs for example, in the THz region.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Optical fibers are dielectric in nature and have large information carrying capacity . It uses photonic crystals to form the cladding around the core, known as photonic crystal fiber (PCF), having exceptional properties [2–4] such as low loss, high birefringence, single mode transmission which cannot be achieved in standard fiber. Another flexibility of PCF is that the core of it can be made porous. The above features lead PCFs in many areas [5–7] such as sensing, spectroscopy, biomedicine, telecommunication, and so on. Among the above areas, sensing attracts much attention nowadays and a large amount of research has been carried on it. Various types of analytes can be sensed such as liquids, non-toxic and toxic chemicals, illegal drugs and so on. Among them, toxic chemicals and illicit drugs get much more emphasis because they are poisonous or harmful for health. For example, benzene, a volatile organic compound which forms blood carcinogen and exposure to this chemical is harmful for human health . Besides that, a few highly toxic poisons have particular antidotes and cyanide is one of them and inhalation of it can give rise to drastic poisoning which may lead to the death of human beings . Moreover, the abuses of illicit drugs such as amphetamine and ketamine cause mental disorder, short-term memory loss, hypnotic effects [10–12]; hence, an efficient mechanism is very important for the detection of analytes especially toxic chemicals and illicit drugs.
Various techniques are available for the detection of the above-mentioned analytes (liquids, chemicals and drugs): benzene is detected thorough spectroscopic detection technology by using scanning tunneling microscope  and ionization detected Raman spectrum which offers double-resonance capabilities . There are few methods such as potentiometric , electrometric, radiochemical, chromatographic for the detection of cyanide . As an example, a simple, fast and selective identification of cyanide in water can be achieved by pervaporation–UV photodissociation–potentiometric detection approach . In addition, various illegal drugs are detected with the help of testing saliva, blood or urine and among them, the most accurate one is testing blood . But these techniques take a long time for analysis, require sophisticated instruments with skilled users and need complex systematic approach. Recently, PCF has become an encouraging technology for sensing various types of analytes or chemicals because of its strong interaction property of light-matter , excellent design as well as controlling flexibility and so on. Terahertz (THz) frequency can be a right place to work because it is affluent in new explorations and needs appropriate waveguide for sensing mechanism. As a result, PCF is a proper platform in THz region as it has been enhanced as a flexible approach for sensing non-toxic and toxic chemicals, illegal drugs [19–21].
Several literatures are available, in THz frequency ranges, to detect various analytes or chemicals by using PCF. For example, an octagonal photonic crystal fiber (O-PCF) was designed by Ahmed et al., in 2015, that achieved sensitivities up to 45.05%, 46.87% and 47.35%, respectively, for water, ethanol and benzyne but the sensitivities are very low . Sensitivity increased to 68.87% in PCF structure as proposed by Sultana and co-authors  but the use of nine elliptical air holes in the core makes the structure complex. In 2018, Paul et al. designed microstructure quasi photonic crystal fiber for chemical sensing application in terahertz  and achieved higher relative sensitivity of 78.8%, 77.8% and 69.7% for ethanol, benzene and water, respectively, but the proposed structure is very complex. Besides that, a new design of chemical sensor having circular cladding with rotated hexa-core in PCF was designed by Sen et al. in 2019  in terahertz region achieving relative sensitivity of 77.16% for benzene but confinement loss is very high as 2.33×10−3 dB/m and 2.84×10−2 dB/m at 1 THz for ethanol and water.
Moreover, PCF based sensor was proposed by Islam et al. and shows moderate sensitivity of 85% for the detection of water, ethanol and benzene . The core area is not appropriately used due to the presence of rectangular air holes  and kagome type cladding increases the device fabrication complexity. A suspension type cladding was introduced in a PCF structure to reduce this complexity but the structure is consisted of elliptical and circular air holes . Here, some portion of the core is also unused as elliptical or circular air holes cannot cover the maximum area of the hexagonal core which limits the stronger interaction of light-matter and results in limited relative sensitivity. Overall, an optimum chemical sensing mechanism is urgently needed in a simple PCF platform that can overcome the above limitations (limited light-matter interaction in the core, low sensitivity, complex structure, high confinement loss and so on).
In this paper, a porous core photonic crystal fiber (PC-PCF) based sensing mechanism is proposed and the device enhances the optical properties to eliminate the above-mentioned shortcomings. This sensor works based on the principle of modified total internal reflection (MTIR) since the core is filled with analytes having a higher refractive index than cladding. In current sensor, maximum portion of core is used for filling analytes so that maximum light can interact with the analytes instead of the background material. Besides, the suspension type cladding makes the structure simpler compared to the kagome type structure . Therefore, the proper optimization of core as well as whole structure leads to the enhanced interaction of light with sensing analytes in the core, hence, increases the detection of sensitivity. The effect of core porosity on the sensitivity of the device is also studied and extremely high sensitivity of 99.60% can be obtained for the design having the least number of air holes. Other than sensitivity, birefringence, effective area, confinement loss, numerical aperture and effective material loss are also demonstrated. Fabrication feasibility of the proposed structure is studied and obtained acceptable tolerance with the available fabrication techniques.
The organization of the remaining paper is as follows: section 2 includes the design of the proposed sensor; numerical method is in section 3; section 4 presents results, discussion and fabrication feasibility; lastly, there is conclusion in section 5.
2. Design of the proposed sensor
The schematic diagram of the proposed PCF based sensing mechanism is shown in Fig. 1. The slots inside the hexagonal core of the PCF are used to make it porous. There are 7 air holes inside the core, among them, the central one is of hexagonal type and the others are of trapezium type shape and the core is suspended in the cladding. Higher sensitivity can be obtained in the fiber having highly polarized characteristics but maintaining polarization state is not easy in circular air holes , hence, asymmetric air holes are used in the current design. Asymmetric air holes are basically responsible for high birefringence which may help to maintain the polarization state in a better way  than circular air holes as reported in the literature [25–27]. The pitch (p), the center to center distance between two air holes, is same everywhere inside the core and is chosen as 52.32 µm. The breadth (b) of the hexagonal core and the height (h) of the core are optimized as 370 µm and 400 µm, respectively. The uniform width (w) of all the air holes are considered as 48.32 µm with the struts width (w1) of 4.0 µm. In the cladding region, the distance between parallel struts is denoted by d with the value of 215.38 µm and thickness of the struts (d1) in the cladding is kept as 6.0 µm.
Several polymer materials such as Topas, Zeonex, Teflon, Polymethayl-methacrylate (PMMA) are well known as a bulk material for different fiber-based sensors in the THz regime. In our case, a cyclic olefin co-polymer (COC) named topas is used as the background material due to its several interesting properties [21,28,29] such as constant refractive index of 1.53 over wide frequency region  which is desirable for low material dispersion. It also shows good chemical and heat resistant with low material absorption loss, attractive water vapor barrier properties, high transparency in THz region, immense strength, high glass transition temperature , and insensitiveness to humidity. Moreover, PCF with topas material is already fabricated and characterized as discussed by Nielsen and co-authors .
Three types of analytes are used, as an example, in this work to cover the sensing performance of the proposed device over a wide range. Two samples are taken from each type of analytes: first type is liquids such as water (n=1.33), benzene (n=1.366) ; second type of analyte is toxic chemicals (sodium cyanide (n=1.45), hydrogen cyanide (n=1.26) ); and lastly, illegal drugs such as ketamine (n=1.562), amphetamine (n=1.518)  and the above mentioned samples cover analyte refractive index from 1.26 to 1.562 but the device is not limited within this range. Besides that, THz frequency is chosen to operate the designed structure and the above analytes are selected without considering any relation with frequency.
There are few commonly used methods [30–32] to fill the analytes in microstructure optical fibers such as selective infiltration of PCF by collapsing air holes, splicing single mode fiber and femtosecond laser. For example, Huang and co-authors  proposed a selective-filling technique by using a syringe filled with chemical. They put chemical in one end of the sensing fiber then pressure helps to fill the air holes automatically. In addition, Cordeiro et al.  demonstrated another filling technique by lateral access to the air holes which is very simple and reproducible. In this case, the analytes could be filled in the core by feeding the channel slowly by a hypodermic needle. Filling can also be done by depositing a layer at one end of fiber and thus it can block the analyte from entering the cladding .
3. Numerical method
In order to analyze and characterize the proposed PCF based sensor, full vector finite element method (FEM) is used while PML works as an absorbing boundary. The amount of interaction of the light with the analytes is measured by sensitivity; hence, it is necessary to calculate the relative sensitivity, for understanding the sensing performance, of the proposed device and can be expressed according to the Beer Lambert law  as35] as
Besides sensitivity, other properties of the fiber need to be calculated. Birefringence is one of them which indicated the effective mode index difference between orthogonal polarization states of core modes and is defined  as20] 24] as 21] 36] as
4. Results and discussion
For examining the accuracy of the simulated results, convergence tests are performed based on the stability analysis of relative sensitivity, R and confinement loss, Lc. In this case, sodium cyanide (NaCN) is chosen as analytes for the calculation purposes and the results are shown in Figs. 2(a) and 2(b): R is plotted along the right side of vertical axis and Lc is in the left side of vertical axis.
PML layer thickness is varied from 1% to 24% of the total radius (750 µm) of the fiber while other parameters are constant. It is seen from Fig. 2(a) that Lc values are oscillating in nature initially and then becomes flat over 6% of PML and sensitivity is nearly stable after 2% of PML, hence, 10% is used throughout the simulation. Further, the mesh size parameter, m is varied from 0.1 to 1.5 by keeping other parameters constant and it is observed (Fig. 2(b)) that Lc and relative sensitivity does not change within the range of m, hence, chosen as m=1 where the lowest mesh size is (λ/6m) for background material/analytes and largest size is (λ /4m) for the air. Therefore, it can be said that the results are stable with the considered PML thickness as well as mesh.
Figure 3 represents the field profile of the PCF sensor for the three analytes (benzene, water, NaCN) for x-polarization and for the rest of the three analytes (HCN, ketamine, amphetamine) for y-polarization. Red color indicates the highest light confinement in the core. In the field profiles, the horizontal and vertical arrows indicate the orientation of electric field for the x- and y-polarized modes, respectively. The propagation of only fundamental modes is observed in the core area and it is necessary to confine maximum light inside the core to get better performance.
Since the proposed device is designed in THz frequency region (as discussed earlier), the relative sensitivity, confinement loss, birefringence and other parameters are plotted with respect to the frequency in x-axis to see the impacts of frequency variation on them, simplify the explanations, and make the presentation clearer. Relative sensitivity with respect to frequency in THz region is illustrated in Figs. 4(a) and 4(b) for x- and y-polarization, respectively. For the simplicity of explanation, two liquids (benzene, water), two toxic chemicals (NaCN, HCN), and two illegal drugs (amphetamine, ketamine) are considered here and in the rest of explanation. Frequency is varied from 0.8 THz to 3.0 THz and all the characteristics are calculated at constant design parameters. Relative sensitivity is calculated with the help of Eq. (1) and Eq. (2). From Fig. 4(a), it is obtained that relative sensitivity for ketamine is the highest as its refractive index is the highest among these six analytes. For the liquids (benzene, water), sensitivity increases along 0.8 THz to 1.4 THz and gradually decreases afterwards. Interaction of light increases gradually with the increase in frequency and it is maximum at 1.4 THz. As frequency increases, some useful power is leaked outside the core as well as the background materials and thus sensitivity is reduced. Maximum sensitivity is observed at 1.4 THz for the liquids and the corresponding relative sensitivities are 92.51% and 91.92% for benzene and water, respectively, for x-polarization.
On the other hand, the relative sensitivities are maximum at around 1.0 THz for ketamine, amphetamine and NaCN and the values are 95.36%, 94.88% and 93.89%, respectively, which means light interaction is maximum at this frequency for these three analytes. But for HCN, sensitivity is maximum at 2.0 THz and its value is 90.75%. The reason behind that the index of HCN is far from rest of the five analytes. Therefore, the reason of maximum light interaction occurrence at different frequencies for different analytes is the different refractive indices of the analytes. For the above-mentioned reason, 1.4 THz is selected as optimum frequency for the liquids (benzene, water); 1.0 THz for those of ketamine, amphetamine and NaCN; 2.0 THz for the poisonous HCN. Similarly, Fig. 4(b) show the values of relative sensitivities for benzene and water are 93.95% and 93.7%, respectively, at 1.4 THz for y-polarization whereas for ketamine, amphetamine, NaCN are 94.87%, 94.78%, 94.42%, respectively, at around 1.0 THz. At 2.0 THz, the relative sensitivity for HCN is 93.4% for y-polarization. Since both polarizations provide a similar trend and y-polarized mode shows a bit greater relative sensitivity than x-polarization, therefore, only y-polarized case will be discussed in the rest of the discussion.
In addition to the relative sensitivity, other fiber properties such as birefringence, effective area, confinement loss, numerical aperture and effective material loss are also calculated to see the additional features and feasibility of the proposed PCF with standard fiber. In this case, the frequency range of 0.8 THz to 3.0 THz is selected by considering the variation of optimum frequencies of the corresponding maximum sensitivities. It is noted that sensitivities are not directly measured from those properties. For example, birefringence is necessary for maintaining polarization states which is desirable for better sensing performance. Hence, asymmetry in the structure is introduced to achieve better birefringence as well as better sensitivity. Similarly, high effective cross-section area provides a greater area for light-analyte interaction which is highly required for better sensing performance.
Firstly, birefringence as a function of frequency (THz) are observed as shown in Fig. 5(a). In order to cover a broad sensing range, six analytes are considered as benzene, water, NaCN, HCN, ketamine, and amphetamine. The observed birefringence value is nearly flat after 1.2 THz across the next considered frequency range because the effective mode index difference between the two polarization modes is nearly same over the specified frequency region. It is noted that the use of analytes in the core reduces the refractive index difference between core and cladding region as well as the birefringence of the PCF structure. The birefringence values for benzene and water are 0.002 and 0.0032, respectively, at the optimum frequency of 1.4 THz and for those of ketamine, amphetamine and NaCN are ∼ 0.001, 0.0012 and 0.004, respectively, at 1.0 THz. Moreover, the birefringence is ∼0.007 for hydrogen cyanide (HCN) at 2.0 THz. According to the above explanations, overall birefringence of the device is not only depended on the structure but also on the analytes and similar performance is observed as shown in . The birefringence is also calculated without analytes (air instead of chemicals) in the core and obtained values are 0.01, 0.012, 0.014, 0.0146 at the frequencies of 1.0 THz, 1.4 THz, 2.0 THz, 2.4 THz respectively. Therefore, the improvement of birefringence as well polarized state, due to asymmetric air holes, might be observed after comparing the above calculated birefringence values (∼1.0×10−2) with the existing literatures (∼10−5 , ∼10−4 , 1.65×10−3 ) having circular airholes [25–27].
Secondly, effective area (Aeff) with respect to the frequency (THz) is illustrated in Fig. 5 (b) for the above-mentioned analytes. It is perceived that Aeff decreases gradually with the increase of frequency. In the region of porous core, light confinement happened in a compact way with the increase in frequency which causes Aeff to decrease. The values of Aeff at optimum frequency (1.4 THz) are 94300 µm2 and 96500 µm2, respectively, for benzene and water. Similarly, in the case of ketamine, amphetamine, and NaCN, the effective areas are 97900 µm2, 99900 µm2 and 10400 µm2, respectively, at their optimum frequency (1.0 THz). Moreover, HCN shows the Aeff of 88900 µm2 at 2.0 THz.
The characteristics of CL versus frequency (THz) are obtained for the same three types of analytes as shown in Fig. 6(a). It is noted that confinement loss decreases slowly with the increase of frequency because the light confinement is stronger in the higher frequency regions. At 1.4 THz, the values of Lc are 9.4×10−12 cm-1 and 6.12×10−11 cm-1 for benzene and water respectively. Again, for ketamine, amphetamine and NaCN, the values of Lc are 1.46×10−10 cm-1, 7.27×10−9 cm-1, 8.68×10−9 cm-1, respectively, at their optimum frequencies. On top of that, the value of Lc is 1.003×10−14 cm-1 for poisonous hydrogen cyanide at 2.0 THz.
The variation of numerical aperture (NA) with frequency is observed in Fig. 6 (b), for liquids, toxic chemicals and illegal drugs. Here, sensing efficiency is not directly measured from the above NA curves but this fiber property is important to analyze because broad sensing application requires a large numerical aperture . It is noticed from Fig. 6(b) that NA decreases moderately with increasing frequency because the effective mode indices difference between cladding and core reduces with the increase in frequency. The obtained values of NA are 0.37, 0.363, 0.465, 0.27, 0.476 and 0.472, respectively, for benzene, water, NaCN, HCN, ketamine and amphetamine at their corresponding optimum frequencies. Although NA for all analytes shows its maximum values at 0.8 THz but 1.0 THz, 1.4 THz and 2.0 THz are also chosen as optimum frequencies (depending on the analytes used as discussed above) by considering other properties, specially sensitivity, of the device.
Another loss mechanism, EML as a function of frequency at the optimized condition is shown in Fig. 7(a) and six analytes as benzene, water, NaCN, HCN, ketamine, amphetamine are considered for the discussion purposes. The nature of the EML curve obtained from Fig. 7(a) is almost flat within the considered frequency region and EML is the highest while the analyte is ketamine as well as lowest for HCN. This is because of the variation of refractive indices of the analytes — as the refractive index of HCN is the lowest among the analytes, so the scattering of light is higher in other region rather than the core. The values of EML are 0.0125 cm-1, 0.0118 cm-1, 0.0149 cm-1, 0.01 cm-1, 0.0175 cm-1 and 0.0165 cm-1 for benzene, water, NaCN, HCN, ketamine and amphetamine, respectively, at their corresponding optimum frequencies.
It is also important to discuss analytes sensing range of this device which is not limited to the specified range (1.26 to 1.562 in the above figures): analytes can be beyond this range. To clarify this issue, five arbitrary analytes having refractive indices of 1.20, 1.23, 1.40, 1.58, 1.62 are considered and their corresponding relative sensitivities are calculated. It is found from Fig. 7(b) that frequencies for maximum sensitivities are red shifted with the increase of the analyte RI. The obtained maximum sensitivities and their respective frequencies are 93.057% at 2.80 THz, 93.28% at 2.40 THz, 94.16% at 1.20 THz, 95.37% at 0.80 THz and 95.93% at 0.70 THz, respectively, for the analyte RI of 1.20, 1.23, 1.40, 1.58 and 1.62 (Fig. 7(b)). Therefore, sensitivity is higher for analytes having large refractive indices and vice-versa.
Furthermore, fiber’s ability to sense liquid is exhibited by considering the fabrication tolerance of ±2%: the optimum design parameters (w, w1 and d1) are varied and then sensing performance is observed. It is noted that geometrical parameters of PCFs can be controlled within ±1% during fabrication as discussed in the literature such as Russell et al.  and others .
Figures 8 and 9 show the relative sensitivity variation as a function of frequency for the ±2% variation of major optimized structural parameters (w, w1 and d1) while only one type of analyte is considered in order to simplify the analysis. Since the variation of w is responsible for the variation of the pitch (p), it’s effect is not considered in this case. The effects of w and w1are shown in Figs. 8 (a) and 8(b), respectively, while ketamine and NaCN are used as analytes. A similar trend is observed for the variation of w and w1 but the effect of w on sensitivity is more than that of w1. The values of relative sensitivities at 1.0 THz are nearly 96.79%, 94.87% and 92.94% for the w tolerance of +2%, optimum, and -2% while ketamine was analyte. It is noted that relative sensitivity increases with the increase of air holes width because the effective area for light-analyte interaction is increased and details can be seen subsection 4.1. The values of relative sensitivity slightly increase for reducing w1 which is the reverse case of varying w. Overall, negligible effect is observed in the case of w1 variation (relative sensitivity are 94.35%, 94.42% and 94.60%, respectively, for +2%, optimum and – 2% w1 at around 1.0 THz) as shown in Fig. 8 (b).
In addition, relative sensitivity variation is analyzed with the change of d1 while benzene is considered as analyte and the effects are shown in Fig. 9 (a). It is noticed that very small change is observed due to the variation of d1: the relative sensitivities are 93.951%, 93.953% and 93.958%, respectively, for +2% to – 2% of design parameter (d1) at 1.4 THz. Lastly, all of the three parameters (w, w1, d1) are also varied simultaneously by considering amphetamine as analyte and the effects are shown in Fig. 9 (b). It is observed that the sensitivity is tuned by around ±2% while all three parameters are varied by ±2% from its optimized values — the corresponding sensitivities for +2%, optimum and –2% variations are 96.53%, 94.78% and 93.06%, respectively, at 1.0 THz. In summary, based on the above discussion it can be said that the structure can tolerate well for ±2% variation of structural parameters.
4.1 Effects of core porosity on the sensitivity of the proposed sensor
Moreover, the effect of changing porosity of air holes in the core on relative sensitivity is illustrated in this subsection. In order to do that, the core of the proposed PC-PCF sensor is modified from seven air holes (Fig. 10 (a)) to five air holes (Fig. 10 (b)) and then three air holes (Fig. 10 (c)) with the same height (h) and breadth (b) of the hexagonal core. In all the cases, the struts width (w1) of the background material is kept same as 4.0 µm.
By following the above discussion, the air holes width (w) and pitch (p) are modified as w=68.84 µm and p=72.84 µm, respectively, for the core having five air holes as shown in Fig. 10 (b). The simulation is performed over the same frequency ranges from 0.8 THz to 3.0 THz. As the total core diameter remains same, the width of the air holes increases with the decrease of the number of air holes, thus the light interacts more, in the core, with the analytes compared to the background material (topas); hence, high sensitivity is expected.
The effect of having seven air holes in the core is already discussed above in Fig. 4 (b) and now Fig. 11 (a) explains the relative sensitivities with respect to the frequency (THz) while five air holes are considered in the core. Six analytes (two liquids, two toxic chemicals, and two illegal drugs) are considered, in this case, for discussion purposes. The values of relative sensitivities are 95.88%, 95.55%, 96.48%, 96.61%, 97.10% and 96.96%, respectively, for benzene, water, NaCN, HCN, ketamine and amphetamine, at their corresponding optimum frequencies. Similarly, the relative sensitivity variation with frequency (THz) is obtained for the same analytes as shown in Fig. 11 (b) where three air holes (in the core) are considered by keeping the same diameter but the expected modified values of w and p are 118.0 µm and 122.0 µm respectively (can be seen in Fig. 10 (c)). In this case, the values of relative sensitivities for benzene, water, NaCN, HCN, ketamine and amphetamine are 98.36%, 97.93%, 99.14%, 97.4%, 99.60% and 99.56%, respectively, at their corresponding optimum frequencies. It is seen from the curves that the relative sensitivity as a function of frequency increases, for every anatyte, with the increase of the porosity in the core which is expected and the reason is that more lights in the core can be interacted with analytes.
Furthermore, we considered few recent related sensors available in the literature to see the performance of the proposed device as shown in Table 1 and comparison is made based on several parameters such as relative sensitivity, analyte detection range, optimum frequency, EML, CL. It is seen from Table 1 that liquids (ethanol, benzene, water) are sensed by Sen et al.  and Paul et al.  with limited relative sensitivity of around 70-75% and moderately high confinement loss of order 10−2 to 10−8 dB/m. Relative sensitivity is further improved by Islam and co-workers  with comparatively higher loss. For the identification of toxic chemicals, one sensor is proposed by Islam and colleagues  which provides relative sensitivity of nearly 88% and EML of 0.023 cm-1 at 1.8 THz by considering KCN as analyte. In the case of illegal drugs, another sensor suggested by Tahhan and co-authors  achieves relative sensitivity of 82% and confinement loss of 2.58×10−15 (cm-1) at 1.0 THz while the analyte was ketamine. According to the above discussion, it is seen that few sensors have higher relative sensitivity but higher loss and vice-versa. The sensing range is limited too. Hence, the proposed sensor shows the best performances among all the literature (shown in Table 1) in terms of relative sensitivities as well as other fiber properties such as low confinement loss, EML and useful for the detection of most of the liquid chemical analytes.
5. Feasibility of fabrication
It is important to consider fabrication feasibilities of the presented PCF based sensor with respect to the available fabrication technologies. Extrusion, 3D-printing, capillary stacking, sol-gel, stack and drilling are the most usual available fabrication technologies [40–46] nowadays. It is noted that sol-gel  and capillary methods are mostly applied to fabricate circular air holes only . In addition, suitability of the techniques used for PCF fabrication depends on the type of structure such as symmetrical or asymmetrical. Extrusion  and 3D-printing [42–44] technologies have the capability to fabricate any complex structures (rectangular, square, elliptical) with symmetrical and asymmetric air holes [41–46]. For example, Atakaramians et al.  used extrusion technique to fabricate complex structures (spider-web and rectangular shaped PCF) having large dimensions. Moreover, suspension type core structure is also fabricated by Liu and his colleague by implementing extrusion and drawing by 3D-printing process .
There are two types of air holes in our proposed PCF based sensor which are of hexagonal and trapezium structure. Therefore, it is mandatory to choose such a fabrication technology which can fabricate these two types of air holes. Since our proposed sensor has asymmetric air holes in suspended type core, so it is preferred to have 3D-printing or extrusion method to fabricate it. The proposed structure has large dimensions which is normal in many fibers designed in THz frequency range theoretically and experimentally [21,24,46]: it may have advantages in terms of easy fabrication but may need additional setup or synchronization issues with existing facilities which may lead to extra cost. PCF with large dimensions can be fabricated as discussed in . Moreover, the sharp edges of air holes, in this structure, may become smoother to some extent during fabrication and the individual area of the air holes may get reduced a bit which may slightly reduce sensitivity but structure having square or rectangular air holes are successfully fabricated as discussed above . So, we hope that our proposed sensor could be fabricated with the available fabrication techniques. Moreover, some postprocessing on fabricated fiber can be done in order to achieve the desired performance.
A porous core PCF based sensing strategy is proposed and analyzed for the detection of a wide range of chemical analytes in THz regime. Three types of analyte (benzene, water; NaCN, HCN; ketamine and amphetamine) are considered as an example to see the performance of the device. The corresponding frequency for maximum sensitivity is depending on the type of analytes over the frequency ranges: obtained relative sensitivities are about 93.95% and 93.7% at 1.4 THz for benzene and water, respectively, whereas 94.87%, 94.78% and 94.42% at around 1.0 THz for ketamine, amphetamine, NaCN and 93.4% at 2.0 THz for HCN. The fiber shows negligible confinement loss of 9.4×10−12 cm-1, 6.12×10−11 cm-1, 8.68×10−9 cm-1, 1.0×10−14 cm-1, 1.46×10−10 cm-1 and 7.27×10−9 cm-1 while benzene, water, NaCN, HCN, ketamine and amphetamine are the analytes respectively. Moreover, the practical realization of the fiber is tested for ±2% variation of the optimum design parameters and achieved acceptable performance. Therefore, the claimed PCF based sensor may be suitable for detecting a wide range of analytes including non-toxic and toxic chemicals as well as illegal drugs at THz frequency regime.
The authors acknowledge the Departments of Electronics & Telecommunication Engineering, Department of Electrical & Electronic Engineering and Office of the Research & Extension of Rajshahi University of Engineering & Technology, Bangladesh, due to their supports. Anika Rahman thanks Dr. Abdul Khaleque for his supervision during the work.
The authors declare no conflicts of interest.
1. G. Keiser, Optical Fiber Communications (Wiley Encyclopedia of Telecommunications, 2003).
2. T. A. Birks, J. C. Knight, and P. S. J. Russell, “Endlessly single-mode photonic crystal fiber,” Opt. Lett. 22(13), 961–963 (1997). [CrossRef]
3. S. A. Razzak, Y. Namihira, M. A. Hossain, and A. Khaleque, “Designing birefringence of index-guiding non-hexagonal photonic crystal fibers,” J. Opt. 40(2), 56–64 (2011). [CrossRef]
4. J. C. Knight, T. A. Birks, P. S. J. Russell, and J. P. De Sandro, “Properties of photonic crystal fiber and the effective index model,” J. Opt. Soc. Am. A 15(3), 748–752 (1998). [CrossRef]
5. M. T. Rahman and A. Khaleque, “Ultra-short polarization splitter based on a plasmonic dual-core photonic crystal fiber with an ultra-broad bandwidth,” Appl. Opt. 58(34), 9426–9433 (2019). [CrossRef]
6. M. S. Islam, S. Rana, M. R. Islam, M. Faisal, H. Rahman, and J. Sultana, “Porous core photonic crystal fibre for ultra-low material loss in THz regime,” IET Commun. 10(16), 2179–2183 (2016). [CrossRef]
7. M. M. Rahman, A. Khaleque, M. T. Rahman, and F. Rabbi, “Gold-coated photonic crystal fiber-based polarization filter for dual communication windows,” Opt. Commun. 461, 125293 (2020). [CrossRef]
8. H. Bahadar, S. Mostafalou, and M. Abdollahi, “Current understandings and perspectives on non-cancer health effects of benzene: a global concern,” Toxicol. Appl. Pharmacol. 276(2), 83–94 (2014). [CrossRef]
9. A. H. Hall and B. H. Rumack, “Clinical toxicology of cyanide,” Ann. Emergency Med. 15(9), 1067–1074 (1986). [CrossRef]
10. B. Sinner and B. M. Graf, “Ketamine,” In Modern Anesthetics (Springer, 2008), pp. 313–333.
11. J. W. Dotson, D. L. Ackerman, and L. J. West, “Ketamine abuse,” J. Drug Issues 25(4), 751–757 (1995). [CrossRef]
12. P. H. Connell, “Amphetamine psychosis,” Br. Med. J. 1(5018), 582 (1957). [CrossRef]
13. J. I. Pascual, J. J. Jackiw, Z. Song, P. S. Weiss, H. Conrad, and H. P. Rust, “Adsorbate-substrate vibrational modes of benzene on Ag (110) resolved with scanning tunneling spectroscopy,” Phys. Rev. Lett. 86(6), 1050–1053 (2001). [CrossRef]
14. P. Esherick, A. Owyoung, and J. Pliva, “Ionization-detected Raman studies of the 1600 cm−1 Fermi dyad of benzene,” J. Chem. Phys. 83(7), 3311–3317 (1985). [CrossRef]
15. B. Vallejo-Pecharromán and M. L. de Castro, “Determination of cyanide by a pervaporation–UV photodissociation–potentiometric detection approach,” Analyst 127(2), 267–270 (2002). [CrossRef]
16. H. B. Singh, N. Wasi, and M. C. Mehra, “Detection and determination of cyanide—A review,” Int. J. Environ. Anal. Chem. 26(2), 115–136 (1986). [CrossRef]
17. S. R. Tahhan and H. K. Aljobouri, “Sensing of Illegal Drugs by Using Photonic Crystal Fiber in Terahertz Regime,” J. Opt. Commun. 1, (ahead-of-print) (2020). (DOI: https://doi.org/10.1515/joc-2019-0291)
18. A. K. Paul, A. K. Sarkar, and A. Khaleque, “Dual-core photonic crystal fiber plasmonic refractive index sensor: a numerical analysis,” Photonic Sens. 9(2), 151–161 (2019). [CrossRef]
19. S. Sen and K. Ahmed, “Design of terahertz spectroscopy based optical sensor for chemical detection,” SN Appl. Sci. 1(10), 1215 (2019). [CrossRef]
20. B. K. Paul, K. Ahmed, D. Vigneswaran, F. Ahmed, S. Roy, and D. Abbott, “Quasi-photonic crystal fiber-based spectroscopic chemical sensor in the terahertz spectrum: Design and analysis,” IEEE Sens. J. 18(24), 9948–9954 (2018). [CrossRef]
21. M. S. Islam, J. Sultana, K. Ahmed, M. R. Islam, A. Dinovitser, B. W. H. Ng, and D. Abbott, “A novel approach for spectroscopic chemical identification using photonic crystal fiber in the terahertz regime,” IEEE Sens. J. 18(2), 575–582 (2018). [CrossRef]
22. K. Ahmed and M. Morshed, “Design and numerical analysis of microstructured-core octagonal photonic crystal fiber for sensing applications,” Sensing and Bio-Sensing Research 7, 1–6 (2016). [CrossRef]
23. J. Sultana, M. S. Islam, K. Ahmed, A. Dinovitser, B. W. H. Ng, and D. Abbott, “Terahertz detection of alcohol using a photonic crystal fiber sensor,” Appl. Opt. 57(10), 2426–2433 (2018). [CrossRef]
24. M. S. Islam, J. Sultana, A. Dinovitser, K. Ahmed, B. W. H. Ng, and D. Abbott, “Sensing of toxic chemicals using polarized photonic crystal fiber in the terahertz regime,” Opt. Commun. 426, 341–347 (2018). [CrossRef]
25. M. Al Mahfuz, M. R. Hasan, M. R. Momota, A. Masud, and S. Akter, “Asymmetrical photonic crystal fiber based plasmonic sensor using the lower birefringence peak method,” OSA Continuum 2(5), 1713–1725 (2019). [CrossRef]
26. A. K. Ghunawat, S. Sharma, S. Sahu, and G. Singh, “Highly Sensitive Octagonal Photonic Crystal Fiber for Ethanol Detection,” In Optical and Wireless Technologies (Springer, 2020), pp. 457–466.
27. R. Hao, Z. Li, G. Sun, L. Niu, and Y. Sun, “Photonic crystal fiber with high birefringence and low confinement loss,” Optik 124(21), 4880–4883 (2013). [CrossRef]
28. K. Nielsen, H. K. Rasmussen, A. J. Adam, P. C. Planken, O. Bang, and P. U. Jepsen, “Bendable, low-loss Topas fibers for the terahertz frequency range,” Opt. Express 17(10), 8592–8601 (2009). [CrossRef]
29. I. K. Yakasai, A. Rahman, P. E. Abas, and F. Begum, “Theoretical assessment of a porous core photonic crystal fiber for terahertz wave propagation,” J. Opt. Commun. 1, (ahead-of-print) (2018).
30. B. T. Kuhlmey, B. J. Eggleton, and D. K. Wu, “Fluid-filled solid-core photonic bandgap fibers,” J. Lightwave Technol. 27(11), 1617–1630 (2009). [CrossRef]
31. Y. Huang, Y. Xu, and A. Yariv, “Fabrication of functional microstructured optical fibers through a selective-filling technique,” Appl. Phys. Lett. 85(22), 5182–5184 (2004). [CrossRef]
32. C. M. Cordeiro, E. M. Dos Santos, C. B. Cruz, C. J. de Matos, and D. S. Ferreira, “Lateral access to the holes of photonic crystal fibers–selective filling and sensing applications,” Opt. Express 14(18), 8403–8412 (2006). [CrossRef]
33. M. S. Islam, J. Sultana, A. A. Rifat, A. Dinovitser, B. W. H. Ng, and D. Abbott, “Terahertz sensing in a hollow core photonic crystal fiber,” IEEE Sens. J. 18(10), 4073–4080 (2018). [CrossRef]
34. C. M. Cordeiro, M. A. Franco, G. Chesini, E. C. Barretto, R. Lwin, C. B. Cruz, and M. C. Large, “Microstructured-core optical fibre for evanescent sensing applications,” Opt. Express 14(26), 13056–13066 (2006). [CrossRef]
35. Y. L. Hoo, W. Jin, C. Shi, H. L. Ho, D. N. Wang, and S. C. Ruan, “Design and modeling of a photonic crystal fiber gas sensor,” Appl. Opt. 42(18), 3509–3515 (2003). [CrossRef]
36. M. I. Hasan, S. A. Razzak, G. K. M. Hasanuzzaman, and M. S. Habib, “Ultra-low material loss and dispersion flattened fiber for THz transmission,” IEEE Photonics Technol. Lett. 26(23), 2372–2375 (2014). [CrossRef]
37. W. H. Reeves, J. C. Knight, P. S. J. Russell, and P. J. Roberts, “Demonstration of ultra-flattened dispersion in photonic crystal fibers,” Opt. Express 10(14), 609–613 (2002). [CrossRef]
38. A. Khaleque, E. G. Mironov, and H. T. Hattori, “Analysis of the properties of a dual-core plasmonic photonic crystal fiber polarization splitter,” Appl. Phys. B 121(4), 523–532 (2015). [CrossRef]
39. M. B. Hossain, E. Podder, A. A. M. Bulbul, H. S. Mondal, M. Raihan, and M. T. Islam, “Identification of Cyanide within Hollow Core Photonics Crystal Fiber,” (10th International Conference on Computing, Communication and Networking Technologies (ICCCNT) IEEE, 2019), p. 1–4.
40. R. T. Bise and D. J. Trevor, “Sol-gel derived microstructured fiber: fabrication and characterization,” In Optical Fiber Communication Conference (p. OWL6). Optical Society of America (2005).
41. A. Ghazanfari, W. Li, M. C. Leu, and G. E. Hilmas, “A novel freeform extrusion fabrication process for producing solid ceramic components with uniform layered radiation drying,” Addit. Manuf. 15, 102–112 (2017). [CrossRef]
42. A. M. Cubillas, S. Unterkofler, T. G. Euser, B. JM Etzold, A. C. Jones, P. J. Sadler, P. Wasserscheid, and P.St J. Russell, “Photonic crystal fibres for chemical sensing and photochemistry,” Chem. Soc. Rev. 42(22), 8629–8648 (2013). [CrossRef]
43. H. Ebendorff-Heidepriem, J. Schuppich, A. Dowler, L. Lima-Marques, and T. M. Monro, “3D-printed extrusion dies: a versatile approach to optical material processing,” Opt. Mater. Express 4(8), 1494–1504 (2014). [CrossRef]
44. W. Talataisong, R. Ismaeel, S. R. Sandoghchi, T. Rutirawut, G. Topley, M. Beresna, and G. Brambilla, “Novel method for manufacturing optical fiber: extrusion and drawing of microstructured polymer optical fibers from a 3D printer,” Opt. Express 26(24), 32007–32013 (2018). [CrossRef]
45. Z. Liu, C. Wu, M. L. V. Tse, and H. Y. Tam, “Fabrication, characterization, and sensing applications of a high-birefringence suspended-core fiber,” J. Lightwave Technol. 32(11), 2113–2122 (2014). [CrossRef]
46. S. Atakaramians, S. Afshar, H. Ebendorff-Heidepriem, M. Nagel, B. M. Fischer, D. Abbott, and T. M. Monro, “THz porous fibers: design, fabrication and experimental characterization,” Opt. Express 17(16), 14053–14062 (2009). [CrossRef]