Technical Articles | Frontline engineers explain to you: Tips for testing the dark current of passive device PD

soso

Technical Articles | Frontline engineers explain to you: Tips for testing the dark current of passive device PD [Copy link]

I saw an article this morning, written by a front-line engineer. I hope someone who needs it will share it here:

Tektronix will launch a series of technical articles on optical communication testing. This is the first lecture, explaining the PD dark current test of passive devices. The author of this article is an engineer friend Duan Gong from Tektronix agent "Ketai Testing". He accidentally heard a few friends in optoelectronics talking about PD dark current and how to test it, and wanted to share some experience and understanding in this regard.

Dark current, as explained by Wikipedia, is a tiny current that is still generated in a light sensor (such as a photomultiplier tube, photodiode, and photocoupler) when no photons pass through it. In non-optical components, it is called reverse bias leakage current, which exists in all diodes. The reason for the formation of dark current is the random generation of electrons and holes in the depletion layer of the component.

There are several key points in this explanation: 1) lightless environment, reverse bias, leakage current; 2) any diode has dark current; 3) dark current belongs to the thermal noise of the component, which is randomly generated and cannot be eliminated. In simple terms: first, this current is not generated by external photons, but by thermal noise inside the component; second, any diode has a theoretical characteristic of forward conduction and reverse cutoff, but in reality, it is impossible for a diode component to be truly cutoff in the reverse direction (the reverse saturation current is 0). Finally, the dark current cannot be completely eliminated and can only be reduced by cooling with TEC or liquid nitrogen.

Generally speaking, dark current is very small, basically in the uA and nA level. In the industrial field, dark current test is a must-test item. This test indicator is mainly used to determine whether the diode component is broken down and whether there is a problem with the wafer process. So how can we accurately and reliably measure such a small dark current? Some people say that a general multimeter or ammeter can be used. Is it really possible?

The sisters in "Sisters Who Make Waves" are very hardworking, working on details frame by frame, and persisting in practicing until perfection no matter how tired, hard, or torturous it is. The daily life of engineers is full of challenges, and any detail bug becomes a stumbling block for progress. In fact, the test of low current is not as simple as it is said, and there are still many difficulties to overcome. Here are a few of the main difficulties.

Difficulty 1: How to overcome the input voltage drop caused by the ammeter?

When the multimeter measures currents above mA, the internal resistance of the ammeter can be basically ignored. However, the magnitude of small currents is basically at the uA or even nA level. At this time, the internal resistance of the ammeter cannot be ignored. The internal resistance of the ammeter will cause a voltage drop, which is called "input voltage drop (voltage burden)". The size of this indicator will directly affect the measurement accuracy of the current:

For example: Assume Vs=0.7V, Is=100uA, Ifs=200uA, Rs=10KΩ, and the input voltage drop is 200mV:

Then we can calculate IM = (0.7V-0.2V (100uA/200uA))/10KΩ = 60uA

In the ideal case, IM = 0.7/10KΩ = 70uA, so the test error = 14%;

If the input voltage drop is reduced to 200uV, the entire test IM = 69.99uA and the test error = 0.01%;

Conclusion: The above simple example shows that the voltage drop at the input end of the ammeter will directly affect the measurement accuracy of the ammeter. The larger the voltage drop at the input end, the greater the measurement error of the current, and the smaller the voltage drop at the input end, the smaller the measurement error.

Difficulty 2: How to add a suitable reverse bias when measuring current?

Commonly used multimeters can only solve the measurement problem, but many dark current tests currently require a reverse bias voltage (Bias Voltage). Why do we need to add a bias voltage? On the one hand, the bias voltage can accelerate the migration process of electrons and holes, reduce the recombination rate of electrons and holes, and thus improve the quantum efficiency and response time; however, the reverse voltage cannot be increased indefinitely, and excessive bias voltage may cause reverse breakdown of the diode, etc. On the other hand, many diodes are avalanche diodes such as APDs, which themselves require a certain bias voltage to reach working conditions and form an avalanche effect. Looking at the current ammeters and multimeters, none of them have the function of providing a bias voltage, so a voltage source must be added to the ammeter's circuit, which will make the test system complicated and introduce more interference conditions, resulting in the entire dark current test accuracy cannot be guaranteed.

So what equipment do related industries (such as LED/PD industry) use for dark current (with bias) testing? Through research and visits to several industries, it is found that there are currently two main options for dark current testing:

(1) SMU source measurement unit. On the one hand, by using its voltage source function, it can complete the reverse bias scan. On the other hand, by using its measurement function, it can complete the small current test. The advantage of this solution is that the voltage scan range is large, up to several hundred volts, and the current measurement function can basically meet the nA level test requirements. The disadvantage is that the unit price of SMU is relatively high, and the cost performance is relatively not that high.

(2) High-precision DMM or picoammeter. Both products are measuring devices that can be used for dark current testing. The current test accuracy can even reach the pA level. The advantages of the products are moderate prices and high measurement accuracy. However, the disadvantages of these two products are: 1) They cannot provide bias voltage and can only complete dark current testing in a bias-free environment; 2) The input voltage drop of a high-precision multimeter is relatively high, which will affect the test accuracy of small currents.

At present, 5G infrastructure is in full swing, and the optical communication industry is experiencing explosive growth, with increasing bandwidth and speed. Whether it is passive optical networks such as FTTx, optical fiber cables, or active optical transceiver modules, optical chips, etc., higher and higher requirements are put forward for the sensitivity of the PD end. Then, the improvement of sensitivity will inevitably put more and more stringent requirements on dark current. By consulting many specifications, a considerable part of the dark current test requirements clearly require dark current ≤1nA, and some even require ≤several hundred pA. At the same time, the bias voltage is required to be between 5-15V, and some voltage requirements are ≥100V, which is basically powerless for DMM and picoammeter.

Is there an instrument that can provide both bias voltage scanning and low current testing? The answer is yes, such as Keithley's 6487. Let's first look at several important indicators of this bias picoammeter:

● 10 fA resolution

● ＜200uV burden voltage

● Support voltage scanning and Analog output

● Scanning voltage range 0-505V;

For the three indicators 1, 2 and 4, the requirements of the dark current test mentioned earlier in the article are fully met, and the voltage source supports synchronous scanning and is greater than 100V. It also has an analog output function, which can not only draw IV curves and test high-resistance devices, but can also be applied to typical optoelectronic applications such as fiber alignment and PD on-wafer testing.

PD On-Wafer Testing

High Resistance Testing

If you happen to be testing and evaluating the dark current of a diode or PD, or you are doing fiber coupling or PD on-wafer testing and other related industry applications, it is recommended that you learn about Keithley's 6487 or 6482/2502 (dual-channel) biased picoammeter.