E. 參考文獻
3. Comparison of Sample A and Sample B
In order to compare the performance of Sample A (high doping density 1x1017cm-3) and
Sample B (low doping density 5x10
16cm-3), we summarized those experimental results for analysis. Figure 13 shows the I-V curves under different temperatures. The solid lines are the dark current at 60Kand 70K,while the dash line is the photocurrent at 30K. Because the doping density in the well in Sample A is higher than that in Sample B, the dark current and photocurrent of Sample A are higher than Sample B. However, the background limitedperformance temperature TBLIP of Sample B (70K) is larger than that of Sample A (60K).
We also show the responsivity comparison in Fig. 14. Sample B has better responsivity for both forward and reverse biases than Sample A. This may be due to the impurity scattering in the well. When the electrons are excited by the photons into the second miniband, and oscillate between two blocking layers, these electrons will be scattered by the impurities in the wells into the first miniband, and emit the LO phonons. It is expected that the absorption rate of photons and the scattering rate of electrons in the superlattice are both proportional to the doping density in the well. Their effects on the photocurrent will cancel each other. According to our experimental results, the device with low doping density is better than the high doping density one because of the decrease of impurity scattering.
The operational voltage of Sample A ranges from 1.3V to 2.0V, under forward bias and from -0.9V to -2.0V under reverse bias. The operational voltage of Sample B ranges from 1.3V to 2.2V under forward bias, and from -1.7V to -2.5V under reverse bias. Because the voltage drop in the superlattice of Sample B is higher than Sample A, the operational voltage of Sample B is higher than Sample A.
Sample B has better quantum efficiency under both forward and reverse biases than Sample A.
The detectivity of Sample A and B under several positive biases at 60K and at 80K are shown as Fig. 15. The detectivity of Sample A is lower than Sample B at 60K and 80K.
As a consequence, Sample B (doping density 5x1016cm-3) has better background limited performance temperature TBLIP,responsivity, quantum efficiency, and detectivity.
-3 -2 -1 0 1 2 3
Fig. 10 The current-voltage curves of Sample B. The solid lines are the dark current at temperature 40Kto100K,while the dash line is the photocurrent at temperature 30K.
11
Fig. 11 The spectral response of Sample B measured (a) at several positive biases and (b) at several negative biases.
11
-3 -2 -1 0 1 2 3 1E-10
1E-9 1E-8 1E-7 1E-6 1E-5 1E-4
I-V comparison
Sample A 60K Sample A Photocurrent
Sample A 70K
Sample B 60K Sample B
70K Sample B Photocurrent
C u rren t(A)
Voltage(V)
Fig. 13 The I-V curves of Sample A and B. The solid lines are the dark current at temperature 60Kand 70K,while the dash line is the photocurrent at temperature 30K.
7 8 9 10 11
Fig. 14 The spectral response comparison. Sample B has better responsivity for both forward and reverse biases than Sample A.
0.8 1.0 1.2 1.4 1.6 1.8 2.0 10
710
810
910
10Positive bias at 60K
Sample B
Sample A
D etectivity (cm x H z
0.5/ W )
Voltage(V) (a)
0.6 0.8 1.0 1.2 1.4 1.6
10
710
810
9Positive bias at 80K
Sample B
Sample A
D etectiv ity (c m x H z
0.5/ W )
Voltage(V) (b)
Fig. 15 The detectivities of Sample A and B under several positive biases (a) at 60K and (b) at 80K.
4. 1-D Detector Array
In the section, we will show some experimental results of 1-D detector array. The sample structure is the same as Sample A (as shown in Fig. 5). New process for 1-D array is used, so some fabrication results are proposed first.
(1) Fabrication Result
Figure 16 shows the SEM picture of the trench with the optical isolation structure which is fabricated by PECVD deposition silicon nitride (Si3N4) and sputtering titanium (Ti) on the wall of the trench. It is noted that the thicknesses of the Si3N4 and Ti on the wall are about 1500Å and 400Å when the deposition thickness is 2000Å on the top surface of the sample. Because the skin depth of Ti at the wavelength 9um is about 300Å, so we sputter thickness 5000Å on the top surface of the sample and thickness 1000Å on the wall of the trench to ensure that light would be reflected by the metal-insulator-semiconductor (MIS) structure.
In order to couple the normal incident light, the V-groove coupling scheme is fabricated. Figure 17 shows the AFM picture of several V-grooves, and the tilted angel is about 41o. These facets can direct the normal incident light into the device though refraction, as shown in Fig. 18 (a). In order to analyze the couple efficiency of different tilted angles, we assume the incident light is TM mode and calculate the electrical field ratio of the refraction component to the incident light versus the tilted angle. The result is shown in Fig. 18 (b). We can see that the electrical ratio is 34.3% when the tilted angle is 41o. The maximum is 39.5% if the tilted angle is 54o. Finally, the fabricated detector array is shown in Fig. 19.
(2) Current-voltage Characteristic
Fig. 20 shows the I-V curves of 1-D detector array under different temperatures. The solid lines present the dark current from 20Kto80K,while the dash line is the photocurrent at 25K under the room temperature background radiation upon the detector. The dark current begins to dominate when temperature is above 50K. The background limited performance temperature TBLIP is up to 50K for the forward bias smaller than 1.4V, and reverse bias smaller than 1.3V.
(3) Spectral Response
The spectral response of 1-D detector array under several positive and negative biases is shown in Fig. 21. The peak responsivity is about 10.3mA/W under +2.6V, at wavelength 8.95um and 2.85mA/W under -2.2V, at wavelength 9.03um.
With the responsivity taken as 10.3mA/W under +2.6V, and 2.85mA/W under -2.2V, we estimated the quantum efficiency to be 0.14% at the voltage +2.1V and 0.04% at the voltage -2.5V.
(4) Specific Detectivity
The peak detectivity D* at 60K is 8.02×108cmHz1/2/W under +1.5V at wavelength 9.02μm, and 2.15×108cmHz1/2/W under 1.4V at wavelength 9.04μm at 80K. The decrease of detectivity with
the increase of temperature is caused by the rapid increasing dark current.
Fig. 16 The SEM picture of the trench with the PECVD deposition Si3N4 and sputtering Ti on the wall of the trench. The thicknesses of the Si3N4 and Ti on the wall are about 1500Å and 400Å when the deposition thickness is 2000Å on the top surface of the sample.