Abstract
In the free-space optical (FSO) communication system, alignment and coupling are key issues. In this work, we adopt a PIN photodiode board as the new receiver to address this question. Firstly, with rapid thermochemical vapor deposition (RTCVD) and solar cell technology, the PIN photodiode board is manufactured. Then, using scanning electron microscope (SEM) and transmission electron microscope (TEM), the microphotographs of the PIN photodiode are taken. After that, the PIN board is arranged as a new receiver in the FSO system to do a bit error rate (BER) experiment. In total, we have carried out 4 groups of experiments. The BERs of the ordinary receiver are as follows: (10−8.5, 10−8, 10−8, and 10−7.9) and that of the new receiver is (10−9.2, 10−9.1, 10−9.1, and 10−9), respectively. It means the BER of the new receiver is lower. In other words, the new receiver performs better.
1. Introduction
FSO communication has attracted growing interests of scholars and researchers because of its advantages such as fast transmission rate and wide space coverage. In FSO, the optical antennas, as a part of receiving system, usually including a series of lenses, focus optical signals into an optical fiber. Optical fiber transmits the optical signals to PIN/APD diodes, where the optical signal is converted to an electrical signal.
In such systems, the coupling of light-into-fiber is the key issue. Many documents have studied this question. In [1], it focuses on neighbor discovery using high-directional transceivers over the same communication channel. In [2], an adaptive digital combination algorithm for coherent FSO communication based on binary phase-shift keying (BPSK) and orthogonal phase-shift keying (QPSK) modulation is proposed to eliminate the time-consuming and complex estimation process of random time-varying channel fading. In [3], the performance of FSO communication system using a PIN photodiode receiver for M-element phase-shift keying (PSK) is evaluated under the situation of strong non-Kolmogorov atmospheric turbulence. In [4], the average capacity of free-space optical communication over a Malaga atmospheric turbulent channel with pointing error and path loss is studied for intensity modulation/direct detection (IM/DD) and heterodyne detection. In [5], the performance of multiband phase shift keying (PSK) SIM communication system using PIN photodiode receivers is evaluated under strong non-Kolmogorov atmospheric turbulence with Gauss beam as an excitation source.
In our previous work, we have studied this question and have published several papers. In [6], a new type of conical optical receiver is proposed, in which a special tapered structure can improve the coupling efficiency by enlarging the optical receiving area. In [7], the conical array is proposed as a new receiver to improve efficiency and the sample of the conical arrays was fabricated in the laboratory. In [8], the conical fiber arrays were fabricated in the laboratory and test experiment under microvibration environment is done to compare its performances with an avalanche photodiode (APD). In [9], in order to analyze the loss of free space optical receiver caused by vibration, coordinate systems are established on the surface of receiving lens and receiving optical fiber, respectively. Then, using Gauss optical theory, the coupling efficiency equation is obtained. In [10], the error rate and coupling characteristics of the new tapered fiber have been studied. In [11], a PIN photodiode array is proposed to receive space light directly.
On the whole, in above approaches, for the receiving system, the problem of too small acceptance area has not been solved well. Then, the alignment and track become a great challenge. Especially, the problem will become worse when the distance is long.
In this work, we attempt to address this question from a new way. We adopt a particular PIN board as the receiver, which is manufactured with solar cell technology. Since the PIN board can convert the optical signal into electrical signal directly, then the coupling of light-into-fiber is not needed. Because the receiving area of PIN board is large, the difficulty in alignment is reduced greatly. In above approaches, some of them have studied the PIN photodiode receiver. However, our work is different. The feature of our work is adopting solar cell technology to produce the PIN board. With the vigorous development of solar energy industry, the solar cell technology becomes more and more advanced. In this paper, we work with the technicians of Zhejiang Anxun Solar Energy Technology Co., Ltd. to produce the PIN board. After the sample of the PIN board is produced, its I–V curves are tested. Then, it is arranged in FSO system to do the experiment. The experimental result shows its BER is lower than that of the traditional receiver.
2. Typical and New Spatial Optical Communication System
Figure 1(a) shows a typical FSO system. Laser beams are coupled into fiber through the antenna and converted into electrical signal in the PIN detector [12–14]. Figure 1(b) shows the new FSO system, in which the PIN board receives optical signal and converts the optical signal into electrical signal directly. In the new system, the optical antenna system is not needed.
(a)
(b)
3. Manufacture of PIN Board
3.1. The Making Process of the Sample
In recent years, with the high-speed development of the solar energy industry, the producing process of solar cells becomes more and more sophisticated [15, 16]. In order to produce the PIN board, we conduct the research with the technicians of Zhejiang Anxun Solar Energy Technology Co., Ltd. This company is located in Yiwu City, Zhejiang Province, China. It is mainly engaged in research, manufacturing, sales, and after-sales service of crystalline silicon solar cells. After several months of effort, the p-Ge/i-Ge layer is grown on the phosphorus-doped n-type Si wafer with the solar cell technology and RTCVD technology. It means the sample is fabricated successfully. RTCVD is a mature technology (for more details, refer [17–20]). The manufacture process includes the following steps:(a)Detection and clean. The phosphorus-doped n-type Si wafer is the carrier of PIN cells. It is necessary to detect the parameters such as minority electron lifetime and resistivity to eliminate the unqualified silicon wafers. Then, the wafers are wet cleaned with HF (HF: DI = 6 : 1).(b)Growth of Ge buffer layers on the phosphorus-doped n-type Si wafer. We use the equipment ECOPIA-RTCVD-100 to do this work; it is the product of ECOPIA Company, South Korea. More information can be found on its website and the product manual. In this process, the main parameters are as follows: temperature: 350°C; pressure: 20 Torr; source gas: GeH4 (20% in H2); and thickness of Ge buffer layers: ∼110 nm.(c)Growth of high temperature i-Ge layers. In this process, the main parameters are as follows: temperature: 500°C; pressure: 20 Torr; source gas: GeH4 (20% in H2) at 30 sccm with 20 slm H2 as carrier gas; and thickness of i-Ge layers: ∼1.4 μm.(d)Growth of boron-doped p-type Ge layers. The doping concentration is about 1018 cm−3. In this process, the main parameters are as follows: temperature: 500°C; pressure: 20 Torr; source gas: GeH4 (20% in H2) at 30 sccm with 20 slm H2 as carrier gas; and thickness: ∼0.32 μm.(e)Growth of synthesis of boron-doped Ge layer. In this process, 100 ppm B2H6 in H2 is used with 0.025 of the dopant number. The main parameters are as follows: temperature: 500°C; pressure: 20 Torr; source gas: GeH4 (20% in H2) at 30 sccm with 20 slm H2 as carrier gas.(f)In this stage, the p-Ge/i-Ge layer is grown on Si wafer. Then, in order to check the hierarchical structure, we use both SEM and AFM to measure the thickness of the p-Ge/i-Ge layer. To ensure the diameters of the mesas ranging from 100 mm to 120 mm, we use a load-locked BMR (HiEtch) high-density plasma etch system to etch the sample [21, 22]. The etch system consists of chamber (operating at 2 MHz), an inductively coupled plasma (ICP), and an additional RF bias (13.56 MHz).(g)Deposition of Si3N4 of 4000 Å on the n-Ge/i-Ge layer. For this purpose, plasma-enhanced chemical vapor deposition (PECVD) is used. After that, contact metallurgy of Ni (300 nm)/Au (300 nm) is deposited by electron-beam evaporation. After fabrication, the contacts are treated by rapid thermal annealing at 500℃ for 30 s in an N2 environment.(h)Silk-screen printing. In this stage, PIN junctions can generate current under light. It is necessary to fabricate positive and negative electrodes on the surface of PIN cells. There are many ways to make electrodes. Among them, silk-screen printing is the most common way. Screen printing is to print predetermined graphics on the substrate by means of imprinting. The equipment consists of three parts: silver-aluminum slurry printing on the back of the battery, aluminum slurry printing on the back of the battery, and silver slurry printing on the front of the battery.(i)Rapid sintering. In this stage, the PIN cells need to be sintered quickly in a sintering furnace to burn off the organic resin binder, leaving almost pure silver electrodes which are tightly bonded to the wafer. The sintering furnace is divided into three stages: presintering, sintering, and cooling. The purpose of the presintering stage is to decompose and burn the high-polymer binder, and the temperature rises slowly in this stage. During the sintering stage, various physical and chemical reactions are completed to form a resistance film structure so that it has resistance characteristics, and the temperature reaches a peak value in this stage. During the cooling stage, the resistance film structure is fixed on the substrate.
In above steps, (b), (c), (d), (e), and (f) belong to RTCVD technology, while (g), (h), and (i) belong to common technology of solar cells. In fact, the RTCVD technology is similar to both low-pressure chemical vapor deposition (LPCVD) and plasma-enhanced chemical vapor deposition (PECVD). Both LPCVD and PECVD are common technologies in solar cell industry.
3.2. The Discussion about the Sample
After the new sample of PIN board is produced, we have measured it. Its shape is about 10 × 8 cm2. Figure 2 shows the surface morphology by using scanning electron microscope (SEM). Figure 3 shows the cross-sectional image by using transmission electron microscope (TEM). The inset of Figure 3 is the enlarged image of the interface.
For the ideal diode, the forward-biased current can be expressed as follows [23, 24]:
In above equations, n is known as the ideal factor, and it is a quantity parameter; k is the Boltzmann constant; T is the absolute temperature; V is the positive bias voltage; and Is is the reverse saturation.
In view of the Shockley theory, at a low voltage, n is about 1.0. At a higher voltage, it is up to 2.0. However, this theory is not applicable to the situation when n is greater than 2.0. For such case, equation (1) needs to be modified as follows [25, 26]:where Rp is the parallel resistance and Rs is the series resistance. When Rp is towards infinity, Rs is towards 0.
In such case, equation (3) reduces to equation (1).
The parallel resistance (Rp) can be deduced near the I–V curve origin, where (V << (Eg/q)):
In general, the series resistance is much smaller than parallel resistance, namely, Rs << Rp. Then, we can evaluate the series resistance at a high voltage without taking the parallel resistance into account. The high voltage means the voltage exceeds turn-on limit, namely, V > (Eg/q). In such case, the I–V curve becomes linear, and Rs is given by
For our pin photodiode, we have done the experiment to test its I–V characteristics at dark. The result is shown in Figure 4, in which both Rp and Rs are marked. From this figure, we can see the current is allowed to flow only in positive direction and the reverse current is close to zero.
At the same time, we have measured the photocurrent (Iph) under the illumination of different wavelengths: 1.10 μm, 1.31 μm, and 1.55 μm. As shown in Figure 5, the responsive photocurrents of 1.10, 1.31, and 1.55 μm are 0.38, 0.23, and 0.19 A/W, respectively. Indeed, Iph could be roughly estimated by the Hecht formula [27–29]:where μτ, d, and Isat are the effective life, i-layer thickness, and the saturation photocurrent, respectively. Vb = V + Vi, where V is the applied voltage and Vi is the built-in voltage of the photodiode. If there is even electric field in the device, and only one trapping level is in consideration, we can use equation (6) to estimate Iph for the first approximation.
4. The Experiment
In order to study the BER performance of the PIN board in the FSO system, we do the following experiment. As shown in Figure 6(a), it adopts an ordinary receiving system in FSO [22]. In the experiment, the bit error tester produces signals and radiates them. After the signals go through the vibration deflector, they are coupled into fiber by the receiving lens. Then, the optical signals are converted into electrical signals in PIN [23]. In the next step, after the electrical signals are processed by the signal processor, they come back to the bit error tester. In the experiment, the vibration environment is used to simulate the alignment error.
(a)
(b)
While, in Figure 6(b), we adopt a new receiving system in FSO, where the PIN board receives the optical signals and converts them into electrical signals directly.
At present, the error testing technology has become more and more sophisticated. The bit error tester produces data and then receives it to calculate the error rate. There are many types to choose from. In this experiment, we adopt the CMR-2048V bit error rate tester, which is produced by Beijing Wangyuan Communication Co., Ltd. It has a laser transmitter with adjustable wavelength. In the previous work, we have used this equipment and platform to do experiments. For more details, refer [9, 10].
In the experiment, we control the frequency and amplitude of the vibrating mirror through the rotary axis control system. The experiment consists of 4 steps. First, the frequency is set to 50 Hz and the amplitude is set to 25 urad, 50 urad, 75 urad ,…, 300 urad. In each test, the time is set to 3 minutes and the laser wavelength is set to 850 . The data rate is set to 10 Mbps. Later, the frequency is set to 100 Hz to do the experiment again. In the third step, the amplitude is set to 50 urad and the frequency is set to 10 Hz, 20 Hz, 30 Hz,…, 120 Hz. In the fourth step, the amplitude is set to 100 urad to do the experiment again. Finally, the results are shown in Figures 7–10.
From Figure 7, we can see when the frequency is set to 50 Hz, the amplitude is changeable, and lg(BER) of the ordinary receiver is up to nearly −8.5. It means the BER is about 10−8.5. “lg(BER)” means “log10(BER).” However, lg(BER) of the new receiver is nearly −9.2. It means its BER is about 10−9.2. This result shows the BER of the new receiver is lower. In the experiment, when the amplitude is 0, it means the vibrating mirror does not work. In this case, the lg(BER) of both new and ordinary receivers is about −11.
From Figure 8, we can see lg(BER) of the ordinary receiver is up to nearly −8. It means the BER is about 10−8. However, lg(BER) of the new receiver is nearly −9.1. It means BER is about 10−9.1. Then, we can draw the similar conclusion that BER of the new receiver is lower. Compared with Figure 7, the difference is that when the frequency is set to 100 Hz, lg(BER) of the ordinary receiver is up to nearly −8, and it is higher than the value in the frequency of 50 Hz. It means the BER is higher in a higher frequency. For the new receiver, we can draw the similar conclusion.
In Figure 9, the amplitude is set to 50 urad, and the frequency is changeable from 0 to 120 Hz. lg(BER) of the ordinary receiver is up to nearly −8. It means the BER is about 10−8. However, lg(BER) of the new receiver is nearly −9.1. It means the BER is about 10−9.1. In Figure 10, the amplitude is set to 100 urad and the frequency is changeable from 0 to 120 Hz. lg(BER) of the ordinary receiver is up to nearly −7.9. It means the BER is about 10−7.9. However, lg(BER) of the new receiver is nearly −9. It means the BER is about 10−9. Both Figures 9 and 10 show the BER of the new receiver is lower.
All of the above figures show that when the vibrating mirror works, the BER of the new receiver is lower. When the vibrating mirror does not work, the BERs of both the ordinary and new receiver are almost the same. In the ordinary receiving system, the optical signals should be coupled into fiber by the receiving lens. However, the diameter of optical fiber is too small. It causes errors. In the vibration environment, this disadvantage is even obvious. In the new receiver, this shortcoming has been avoided.
Here, we would like to make the comparison with other studies. In [30], the performance evaluation of an FSO link has been performed using an array of photodetectors. From the results presented in [30], it can be concluded that as the number of photodetectors increased, there is a significant decrease in BER. Specifically, when single APD, array of 2 APD, array of 4 APD, and array of 8 APD are used at the receiver end of FSO link, the BERs are about (10−8.2, 10−8.5, 10−9.2, and 10−9.5). For the new receiver in our work, we have carried out 4 groups of experiments. The experimental BERs are (10−9.2, 10−9.1, 10−9.1, and 10−9), respectively. It shows that the performance of our new receiver is equivalent to the array of 4 APD in [30].
In [31], the numerical simulation analysis showed that the NRZ coding and Mach–Zehnder modulation give the optimized overall performance. In this case, the BER is about 10−9.1. This value is equivalent to that of our new receiver. It means the performance of our new receiver reaches the optimized overall performance in [31].
In addition, in the experiment, the size of the PIN board is about 10 × 8 cm2. In fact, we can connect several boards in series and in parallel to increase the receiving area. In this way, the difficulty in alignment could be further reduced.
The experiment is done in the dark room to avoid the effect of external light. However, in actual communication, the external light is ubiquitous. Then, how to avoid its influence needs further research. For this problem, we would like to propose two methods: one is the compensation method, and the other is fixed wavelength PIN receiver. For the first method, we detect the external light firstly and then deduct it out. For the second method, PIN receiver can only receive a certain kind of light, and it responds little to other light. Of course, it needs further study to judge whether such methods are effective. In the next step, we plan to study the relation between the size of the PIN board and the receiving efficiency.
5. Conclusion
In this work, we propose to use the PIN board as the new FSO receiver. Firstly, we use RTCVD and solar cell technology to produce the PIN board. Then, the PIN board is arranged as a new receiver in the FSO system to do a BER experiment. In the experiment, the vibrating mirror is set between the transmitter and receiver. We control the frequency and amplitude through the control system to simulate the different vibration environments. In total, we have carried out 4 groups of experiments. All the experimental results show that the new receiver performs better.
Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.
Conflicts of Interest
The authors declare that there are no conflicts of interest.
Acknowledgments
The author HU Qinggui acknowledges the financial support from the National Natural Science Foundation of China (Grant no. 61275080), Solar Energy Integration Technology Popularization, and Application Key Laboratory of Sichuan Province (2018TYNSYS-Z-01). This project was supported by the National Natural Science Foundation of China (Grant no. 61275080) and the National Technology Innovation Fund (Grant no. 13C26212201166).