Results & Discussion
3.3 Ion Implantation
As we all know, the conductivity of all the devices can be improved by ion implantation because it could supply more carriers in the channeal [30][31]. So, we also experiment a series of the research incorporation to the ion implantation. The condition of the ion implantation was the same for the following two experiments which the energy E=15 KeV and the dose D=5E15. After ion implantation, we all treat the device for thermal annealing to active the dopants. The temperature of thermal anneal was set at T=950 ℃ for 30 min. It was the same with the previous experiment for the comparison.
3.3.1 Only Source/Drain Ion Implantation
It add one lithography process of the Mask #PR to block the channel being implanted. We use Photo Resist (PR) for hard mask. After implantation, the other processes were all the same with the previous experiment. Fig 3.21, 3.24, 3.27 and 3.22, 3.25, 3.28 show the Id-Vd of the Si0.89Ge0.11 and Poly-Si nanowire for the same length respectively. Using the previous way, the current I=2.63 µA in L=4 µm, I=2.35 µA in L=5 µm, and I=2.12 µA in L=6 µm for Si0.89Ge0.11 nanowire and I=1.47 µA in
L=5 µm, I=1.57 µA in L=5 µm, I=1.77 µA in L=6 µm, By calculation, the current of the Si0.89Ge0.11 nanowire is I=7.256 A/V‧cm, 7.201 A/V‧cm, 6.732 A/V‧cm and the current of the Poly-Si nanowire is I=0.915 A/V‧cm, 0.918 A/V‧cm, 0.962 A/V‧cm. It was evident that the current after normalizing is higher than the without any treatment respectively. Fig 3.23, 3.26 and 3.29 show the normalized Id-Vd at Vg= -10 V. It’s clear that the SiGe nanowire has higher current than Poly-Si one. And the current is larger than previous one due to the contact resistance is reduced by the S/D ion implantation. In fig 3.26 and 3.29 show a little zero current at Vd=-1~1 V and this is because there is still Schottky barrier between the Al/SiGe contact. Table 4. summary the current after S/D ion implantation then anneal at T=950 ℃ for 30 min. Each term of the current was improved by the process due to reduce the contact resistance.
3.3.2 S/D and Channel Ion Implantation
The process of ion implantation was done after the Source/Drain of the SiGe were formed in the Mask #03. There was no sacrificial layer grown on the SiGe channel before being ion implantation. The other processes were all the same with the previous experiment. Fig 3.30, 3.33, 3.36, 3.39 show the Id-Vd of the Si0.89Ge0.11 nanowires and Fig 3.31, 3.34, 3.37, 3.40 show the Id-Vd of the Poly-Si nanowires. It is obvious that the current of the Poly-Si nanowire is larger than Si0.89Ge0.11 ones. So we should normalize the current to compare them. Fig 3.32, 3.35 show two normalized Id-Vd diagram fixed at Vg = -10 V. It was clear that the Si0.89Ge0.11 has higher current than Poly-Si at any drain voltages.
The current is much improved in SiGe nanowire. In 3.38 and 3.41, we show both the Vg = 10 V and –10 V at the same time. The Si0.89Ge0.11 has higher current than
Poly-Si both at Vg=10 V and –10 V at any drain voltages. The SiGe nanowire is tend to p-type, so the more negative bias we supply, the more larger current we’ll get. As we known, the nanowire used as a sensor by dip specific PH solution as the applied gate voltage. For a fixed Vd, the variance range of SiGe current is more widespread from Vg=10 V to –10 V than Poly-Si and this means that the SiGe nanowire is more sensitive than Poly-Si. Fig 3.42, 3.43 show the Id-Vd of the Si0.89Ge0.11 and Poly-Si nanowires we measured from –1 V to 1 V with the different Vg change from –20 V to 20 V. It shows the very sensitive properties of the nanowires, and the SiGe nanowire is more sensitive than Poly-Si nanowire.
If we assume that the area of the nanowire are the same with the present values, we will get the current per unit length of the Si0.89Ge0.11 nanowire is I=99.1 A/V‧cm, 104.59 A/V‧cm, 106.8 A/V‧cm, and 103.17 A/V‧cm and the current of the Poly-Si nanowire is I=63.3 A/V‧cm, 71.54 A/V‧cm, 73.27 A/V‧cm, and 64.96 A/V‧cm. Both the SiGe and Poly-Si were improved, but it was obvious that the improvement of the SiGe is lower than Poly-Si one. Table 5. summary the current after S/D and channel ion implantation then anneal at T=950 ℃ for 30 min. Each term of the current was improved by the process. Table 6. list all the average current per unit length we normalized, and table 7 show the comparison of the SiGe and Poly-Si nanowire. The SiGe/Poly-Si ratio is 9.537 in the initial, and is reduced to 7.578 after S/D implantation and 1.515 after S/D, channel implantation. The improvement properties of the Si0.89Ge0.11 is getting lower. What is wrong ?
3.4 Discussion
After Source/Drain and Channel being ion implantation, both the current of the Si0.89Ge0.11 and Poly-Si nanowires were both improved. But the improvement of Si Ge is lower than Poly-Si. It might divide into two parts to be discussed: (1)
Contact Resistance, (2) Implant Doses.
(1) Contact Resistance :
Fig 3.44 and 3.45 show contact resistance-length figure of the Si0.89Ge0.11 and Poly-Si nanowires after S/D implantation. We could read out the contact resistance is R=71.57 Ω for Si0.89Ge0.11 and R=31.702 Ω for Poly-Si. The contact resistace of the Si0.89Ge0.11 is twice larger than the Poly-Si. This might be due to the Al electrode contact pad does not form good ohmic contact with SiGe film. Choose a comportable metal electrode as contact pad such as or TiN/W for the SiGe film could be tried.
Dawei Wang, etc. have experimented that the Ni forms good ohmic contact for the boron nanowires [38]. The contact electrode for the Si0.89Ge0.11 might not be the optimum.
(2) Implant Doses
The reduction of the improvement of the SiGe nanowire in our S/D, channel implantation experiment is more serious. The most possible reason might due to the Boron doses in the channel are too much. The nanowire after the channel implantation is almost a boron conducting wire. In the Yi Cui, etc. show it is possible to incorporate high dopant concentrations in the nanowire and to approach the metallic regime [39].
They fabricated the Si nanowire with the SiH4:B2H6 ratios of 1000:1 and 2:1, respectively. Fig 3.46 show the I-V curves of the metallic properties of different nanowire diameters. There are too many dopants in the channel that the improvement by SiGe would be vague.
Chapter 4 Conclusion
In our thesis, We have successfully fabricated the SiGe nanowire on silicon wafer with the conventional lithography process. The electrical properties were measured by 4156C and the structure of the SiGe nanowire on the sidewall spacer were observed by SEM.
The sensitivity is the most important key issue for the nanowire as a sensor, and the higher conductance means that the nanowire is more sensitive. We have improved the current by using SiGe material. With the Ge concentration increases, the current will increase, too. After Source/Drain ion implantation, we can get the higher current due to reduce the contact resistance. When the channel was implanted, the current becomes larger than without any treatment.
The improvement of the SiGe nanowire was lower than Poly-Si after implantation. We have proved that the contact resistance in Al/SiGe is higher than Al/Poly-Si nanowire. In the channel implantation, the reduction of the improvement of SiGe might due to the too higher Boron doses we implanted and then conducting properties was dominated by boron.