A p-type Si substrate (100) was used in this study. The resistivity of silicon substrate was about 1~10Ω-cm.
1. Standard RCA clean and wet oxidation to grow 5000Å thick bottom oxide by SVCS Furnace system. The structure is shown in Fig. 2-1.
2. Mask#1: Define active area. TEL CLEAN TRACK MK-8 and Canon FPA-3000I 5+stepper lithography system were employed to transfer pattern onto oxide layer. Then, dry etching 3000Å was carried out with TEL5000R.I.E system to form oxide step. It is ready for sidewall spacer formation. The structure is shown in Fig.
2-2.
3. Standard RCA clean, α-Si layer was deposited on bottom oxide by Vertical furnace system. This α-Si layer deposited 200A、400A and
16
600A separately for the stack structure. Then, α-Si served as seed layer for SiGe film deposition. The structure is shown in Fig. 2-3.
4. Standard RCA clean, and then SiGe was deposited by ANELAVA SiGe UHV-CME.The structure is shown in Fig. 2-4.
5. Mask#2: Define S/D region and form spacer sidewall. Dry etching was carried out with TCP9400 SE poly etcher to define S/D region on the active layer and the SiGe/Si nanowires were defined on the spacer. The structure is shown in Fig. 2-5.
6. Mask#3: Remove unwanted sidewall spacer. TCP 9400 SE poly etcher was employed to remove unwanted spacer, which would have resulted in short circuit between two nanowire devices if not removed. The structure is shown in Fig. 2-6.
7. Boron-fluoride(BF249+
) was implanted into SiGe nanowires by E500HP implanter. The implantation dose is focused on 1×1015 ions/cm2 and energy was focused on 50keV.The structure is shown in Fig. 2-7.
8. Annealing in Furnace at 950℃ for 30min to activate dopants.
9. 4000Å ~5000Å Aluminum deposited by AST Peva 600I. The structure is shown in Fig. 2-8.
10. Mask#4: Define aluminum contact pad. Al pads are formed by wet etching (HNO3:CH3COOH:H3PO4:H2O=2:9:50:10). The structure is shown in Fig. 2-9.
11. Aluminim sintering at 400℃ in N2 ambient for 30 min.
12. The device view from top position. The structure is shown in Fig.
2-10.
17 2.2 Functionalization
First, we used amino-propyl-trimethoxy-silane (APTMS) to modify the surface of SiNW. The native oxide was around SiNW, so the APTMS to SiNW oxide surface resuled in a surface terminating in both -NH2 and -SiOH groups. The modification was shown in Fig. 2-11. After APTMS modification, the surface of nanowire was terminated by amine groups. In our experiment, amine group was prone to be positive charge. It is like that SiNW had the positive gate bias, so the conductanece of P-type SiNW decreased. Next, we used bis-sulfo-succinimidyl substrate (BS3) to bond on the APTMS. BS3 treatment resulted in negative charge, so the conductance of P-type nanowies increased. In this study, we focused on sensitivity (S) and conductance with different stack structures.
2.3 Measurement of electric characteristics
HP4156A was used in this study to measure the electric characteristics os nanaowire sensors. Drain voltage(VD) was from -10V to 10V and step was 100mV, and back gate voltage(VG) was 0V.The electric measurement of electric characteristics was performed at every stage of surface modification , and the average conductance was then extracted from ID-VD characteristics with VD=4~6V .
18 2.4 Define the sensitivity S
First, we measured the I-V curve for no treatment devices, and then we defined the current was I0. Second, We dripprd the APTES to the surface of SiGe nanowires , and the measured the I-V curve. Then we defined the current was I. The sensitivity was
S=
. Thus, we focused on this definition to measure sensitivity.19
Chapter 3
The Characteristics of SiGe/Si Stacked Structures
3.1 Motive of the experiment
According to our group’s previous research, we focused on Ge density、annealing temperature and oxidation time. We based on our group’s previous research, and then we changed the thickness of amorphous Si, which was the purpose of increasing current. When we increased the thickness of amorphous Si, the thickness of SiGe relatively was thin.However, the amorphous Si had the higher resistance than SiGe, so the thinner SiGe had the high current.
In previous research, we focused on annealing temperature 950°C.
Ge would diffuse over 1000°C and sensitivity would decrease because there was not enough energy to repair the defects at 900°C. Hence, we anneal at 950°C in order to obtain better quality and higher sensitivity.
We found out the best combination in the experiment. In this section, we only discussed the characteristics of SiGe/Si stack structure with on treatment.
20 7% imaged Fig.3-2, which showed the width of nanowire was 72nm. The structure of 200Å 14% imaged Fig.3-3, which showed the width of nanowire was 77nm. The structure of 200Å 20% imaged Fig.3-4, which showed the width of nanowire was 65nm. The structure of 400Å 7%
imaged Fig.3-5, which showed the width of nanowire was 67nm. The structure of 400Å 14% imaged Fig.3-6, which showed the width of nanowire was 73nm. The structure of 600Å 7% imaged Fig.3-7, which showed the width of nanowire was 70nm. The structure of 600Å 14%
imaged Fig.3-8, which showed the width of nanowire was 72.5nm.
3.2 The different stacked structures with different Ge concertrations
In this section, we discussed the structures in this thesis . We used the stacked structures, which was SiGe/Si. The bottom part was amorphous Si with different thicknesses and the top part was SiGe with different Ge concentrations. The following sections were discussed the stacked structures and Ge concentrations separately.
21
3.2.1 Comparing with different stacked structures
In this section, we focused on the stacked structures. The stacked structures were divided into the different depths of amorphous Si, which were 200Å 、400Å and 600Å .The purpose was the amorphous Si with high resist, so the current flowed into the path of SiGe. When the amorphous Si was thick, the current of SiGe nanowire was enhanced. The conclusion was in Fig. 3-9, which was focused on Si0.93Ge0.07 .
3.2.2 Comparing with different Ge concentrations
In this section, we focused on the different Ge concentrations. The Ge concentrations were divided into 7%, 14% and 20%. We fixed the thickness of amorphous Si on 200Å , and the conclusion was in Fig. 3-10.
And then the amorphous Si on 400Å and 600Å were shown in Fig. 3-11 and Fig. 3-12. In this conclusion, we found that whatever the thickness of amorphous Si, only the Ge concentration enhanced the performance of SiGe nanowire devices. If Ge concentration increased, the the performance of SiGe nanowires was increased. The Ge had high mobility, so that the high Ge concentration enhanced high current.
22
3.2.3 Combining different stacked structures and Ge concentrations
In this section, we discussed the different stacked structures with different Ge concentrations. We focused on amorphous Si 200Å 、400 Å and 600 Å with Ge 7% and 14% concentrations. The conclusion was shown in Fig. 3-13.The 600Å 14% had the higher current than others.
Because the structure had high Ge concentration and SiGe on the amorphous Si 600Å was relatively thinner than other structures. We showed the conductance of all structures in Fig.3-14 and found out the structure with amorphous Si 600Å and Si0.86Ge0.14 had the high conductance.
3.3 Electrical response after dripping APTMS and BS3
The APTMS and BS3’s bond structures contained positive and negative electricity respectively, Therefore, the conductance would decrease after drip APTMS on p-type SiGe nanowire, but it would be increased when the bonds of BS3 occur after it dripped on APTMS.
3.4The Sensitivity of SiGe nanowire with no treatment
Before the sensitivity, we discussed the conductance with the
23
structure of 200Å 7%, which was shown in Fig.3-15. We obviously the APTMS decreased and the BS3 increased, because they were positive charge and negative charge. According to the conclusions, we could confirm our nanowire had the characteristic of bio-sensor.
First, we dripped APTMS on SiGe nanowires , and then we found the sensitivity S1
=
. Second, we dripped BS 3 on the SiGe nanowire , which was dripped APTMS already, and then we found the sensitivity S2=
(I0: dripped water current, I1:dripped APTMS current, I2:dripped BS3 current).The conclusion was shown in Fig. 3-16.The sensitivity of dripped WATER was 0.644%. The sensitivity of dripped APTMS was -6.54%. The sensitivity of dripped BS3 was 2.49%.The conclusion was that the APTMS was positive charge, so APTMS could accumulate the surface of p-type nanowire. And the BS3 was negative charge, so BS3 could deplete the surface of p-type nanowire.
We also showed the conductance and sensitivity with different stack structures to prove our nanowire with the characteristic of bio-sensor. The 400Å 7% was shown in Fig.3-17. The 600Å 7% was shown in Fig.3-18.
3.4.1 The sensitivity of different Ge density with the same width
First, we defined that the sensitivity was the part of dripped APTMS.
Thus the sensitivity was S1
=
. In the view of the APTMS was24
the linker, the APTMS was the important part.
In this section, we had to compare Si nanowire with SiGe nanowure and knew what density of Ge was the best. In the conlusion, we were known that the nanowires with Ge ingredients had the highest sensitivity.
So we confirmed that SiGe nanowire was better that the Si nanowire.
In previous research, when Ge density increased, the sensitivity decreased. It was that the higher amount of Ge would increase high vacancies of surface which would not easy for objects under test to bond and decreased sensitivity. We confirmed the conclusion. It was shown in Fig.3-16, which was focused on amorphous Silicon 200Å . The sensitivity of poly Si was 2.873%. The sensitivity of Ge 7% was 3.019%. The sensitivity of Ge 14% was 5.4% The sensitivity of Ge 20% was 2.2%.Thus, we focused on the density of Ge were 7% and 14% in next sections.
3.4.2 The sensitivity of the same width with different Ge densities
As the result of previous discussions, we focused on the density of Ge, which were 7% and 14%.Thus this section, we fixed the Ge density.
We had to know what thickness of amorphous Si was the best stacked structure. According to previous research, we assumed that amorphous Si 600Å enhanced the high sensitivity, because the amorphous Si 600 Å held the thin SiGe layer and the high Ge density. In view of the important factors, the stacked structure which was amorphous Si 600 Å
25
had the high sensitivity. The conclusion which was fixed on Ge 7% was shown in Fig.3-17. The sensitivity of amorphous Si 200A was 4.379%.
The sensitivity of amorphous Si 400A was 4.487%. The sensitivity of amorphous Si 600A was 5.845%. And another which was fixed on Ge 14% was shown in Fig.3-18. The sensitivity of amorphous Si 200A was 4.61%. The sensitivity of amorphous Si 400A was 5.073%.The sensitivity of amorphous Si 600A was 7.95%.
3.5 Summary of the characteristics of SiGe/Si stacked structure
In this chapter, we only focused on the the characteristics of SiGe/Si stacked structure without any treatment. However, we could clean out the following conclusions. When the thickness of SiGe was narrowed and the Ge density was 14%, the sensitivity was increased at most. In Fig.3-19, we could find that the slope of Si0.86Ge0.14 was higher than the slope of Si0.93Ge0.07. It was mean that the high concentration and amorphous Si 600Å could have the high sensitivity with no treatment. And in the Fig.3-20, we would know that the raise of sensitivity between 7% of Ge density and 14% of Ge density. It was shown that the thickness of amorphous was 600Å, which raised about 29.8%, because the narrowed SiGe layer and the high Ge density were in the structure. In the next chapter, we would do some treatments of oxidation to increase sensitivity more.
26
Chapter 4
The Treatment of Oxidation for SiGe/Si Stacked Structure
4.1 The Introduction of Oxidation Treatment
In previous research, we produced non-homogeneous structure. With the help of oxidation, we could find the better quality of deposited SiGe layer with oxidation and then high density of Ge would be separated on surface. Besides, base on Auger analysis, we could obviously realize that Ge only remained on the surface of nanowires, and the sensitivity happened to be decreased when Ge’s density increased. We assumed that the higher amount of Ge would increase high vacancies of surface which would not easy for objects under test to bond and decreased sensitivity.
Above of all, we used the characteristic of SiGe oxidation, the oxidation process with nitrogen could repair the vacancies of surface and improved the quality of oxide in order to increase sensitivity.
In this section, we discussed the treatment of oxidation for SiGe/Si stacked structure. We treated N2/O2 ratio for SiGe/Si stacked structure.
27
We focused on no oxidation, N2 0%, N2 13.3% and N2 100% for 3min and the oxidation temperature was 900°C. In this treatment, we would find the best one to treat different time, which was divided to 3min、5min and 7min.
4.2 The Influence of SiGe Nanowire with proportion of Nitrogen and Oxygen
To process non-homogeneous structure was the intention of this thesis. We obtained high Ge concentration during oxidation, but the quality was not good enough, because the bonding energy of surface reduced and the sensitivity fell down.
As the conclusions, we tried to add the content of nitrogen during oxidation in order to repair the defects and improve the quality of surface.
In previous research, we add content of nitrogen to 13%, 40% and 66.6%, and we testified 13% of nitrogen would be best. However, in this thesis, we would focus on 13% of nitrogen. In addition, we would discuss 100%
of nitrogen, which was pure nitrogen. We would compare the repair of pure nitrogen to 13% of nitrogen.
4.3 The Performance of Stacked Structures after Oxidation
Before the discussion of oxidation, we focused on N2 100%, which was that we only flowed nitrogen without oxygen at 900°C. We would compare the oxidation without oxygen to the oxidation with oxygen. Next,
28
the treatment of oxidation was divided into N2 0% and N2 13%, which were the ratio of nitrogen and oxygen.
First, we focused on the characteristics which were no oxidation, N2
0% and N2 100%. In this part, we discussed the performance of SiGe through repair of nitrogen. The conclusion was shown in Fig. 4-1, which was focused on 200Å 14%. And another structure’s conclusion was shown in Fig. 4-2, which was 400A 14%. In the conclusion, we obviously found out the current increased after oxidation and repair of nitrogen. And the current of N2 0% was higher than the current of N2 100%, because the effect of oxidation was weak. In this condition, it made Ge reject less, so that the current was not obvious.
Second, we would add N2 13% to measure the I-V curve. In previous research, we found out that N2 13% was the best treatment, so we added the treatment on our stacked structures. We identically exploited 200Å
29
concentration of Ge went up, the sensitivity raised. And then the 13% of nitrogen increased at most. The sensitivity of 13% of nitrogen was 5.819%. The 200Å 14% was shown in Fig. 4-6, and the sensitivity of 200Å 14% was 8.94% in 13% of nitrogen. The 400Å 7% was shown in Fig. 4-7, and the sensitivity of 400Å 7% was 10.4% in 13% of nitrogen.
The sensitivity of 13% of nitrogen was 5.819%. The 400Å 14% was shown in Fig. 4-8, and the sensitivity of 400Å 14% was 12.03% in 13%
of nitrogen. The sensitivity of 13% of nitrogen was 5.819%. The 600Å 7% was shown in Fig. 4-9, and the sensitivity of 600Å 7% was 11.679%
in 13% of nitrogen. The sensitivity of 13% of nitrogen was 5.819%. The 600Å 14% was shown in Fig. 4-10, and the sensitivity of 600Å 14% was 4.493% in 13% of nitrogen.
We could see that the 13% of nitrogen was the best. Because nitrogen could repair the defects, the concentration of Ge on the surface raised and the quality became better after the oxidation of SiGe nanowire.
Further, the width of SiGe nanowire became narrow, so the raise of sensitivity was obvious excluding the structure of 600Å 14%.
We also observed that the 100% of nitrogen was less than pure oxygen. This was because when we added too much nitrogen, the effect of oxidation was weak. In this condition, it made Ge reject less, so that the raise of the sensitivity was not obvious.
In upward conclusions, we found out the sensitivity of 600Å 14%
decreased and the sensitivity of 600Å 7% was raised less than the sensitivity of 400Å 14%. Because the process limited, the width of 600Å 14% and 600Å 7% were about 65nm to 70nm. It was mean that the SiGe
30
layer only was 5nm to 10nm. Thus, when it was through 3min oxidation, it was oxidized all very likely. The nanowire was only poly silicon, so its sensitivity was close to silicon nanowire. We could confirm the conclusion by EDS, Which was shown in Fig. 4-11. We perceived the Ge information was not many. And then, we found that when the Ge density was high, the oxidation rate would increase. It was confirmed in the paper[49], which explained the SiGe oxidation with the different Ge densities. Overall, we could explained this condition due to the process limited, the SiGe oxidation rate was faster than Si oxidation rate and the high Ge density had the fastest oxidation rate.
4.5 The comparison between N2 100% and N2 0%
In this section, we would compare 100% of nitrogen to 13% of nitrogen. In previous discussions, we knew that both of 100% of nitrogen and 13% of nitrogen could enhance the sensitivity, so we had to realize which one was the best. In the Fig. 4-12, we obviously found out that the structure of 200Å 7%, which was shown that the 100% of nitrogen increased sensitivity about 28% and the 13% of nitrogen increased sensitivity about 92%. And the structure of 200Å 14% was shown that the 100% of nitrogen increased sensitivity about 18.6% and the 13% of nitrogen increased sensitivity about 99%. This was because when we added too much nitrogen, the effect of oxidation was weak. In this condition, it made Ge reject less, so that the raise of the sensitivity was
31 not obvious.
4.6 The Raise of Sensitivity After N2 13% of Oxidation
We already discussed the 13% of nitrogen, which was the best treatment of oxidation, so we focused this treatment to calculate the raise of sensitivity. We compared every stacked structures including poly Si nanowire.
First, we fixed on Si0.93Ge0.07. The result was in Fig. 4-13. We obviously obtained that the sensitivity of poly Si raised 62%, the sensitivity of amorphous 200Å raised 93%、the sensitivity of amorphous 400Å raised 119% and the sensitivity of amorphous 600Å only raised 93%. The raise percentage of sensitivity was defined as the next equation.
The raise percentage of sensitivity(%)=
Si was the sensitivity of SiGe before dipping chemical molecules, Sf
was the sensitivity of SiGe nanowire after dipping chemical molecules.
The value showed the percentage change of sensitivity after the APTMS and BS3 modified. The Fig. 4-14 was the percentage change of Si0.93Ge0.07 after APTMS modified.
Excluding the structure of amorphous 600Å , the high concentration of Ge and the narrow width of nanowire was raised the sensitivity at most.
Second, we fixed on Si0.86Ge0.14. The result was in Fig. 4-15. We obviously obtained that the sensitivity of poly Si raised 62%, the
32
sensitivity of amorphous 200Å raised 101%、the sensitivity of amorphous 400Å raised 138% and the sensitivity of amorphous 600Å was decreased 33%. The reason of decreased amorphous 600Å was explained in front section. Identically, the high concentration of Ge and the narrow width of nanowire was raised the sensitivity at most. The Fig. 4-16 was the percentage change of Si0.86Ge0.14 after APTMS modified. Next, we fixed on amorphous Si, so the Fig. 4-17, Fig. 4-18 and Fig. 4-19 were fixed on
sensitivity of amorphous 200Å raised 101%、the sensitivity of amorphous 400Å raised 138% and the sensitivity of amorphous 600Å was decreased 33%. The reason of decreased amorphous 600Å was explained in front section. Identically, the high concentration of Ge and the narrow width of nanowire was raised the sensitivity at most. The Fig. 4-16 was the percentage change of Si0.86Ge0.14 after APTMS modified. Next, we fixed on amorphous Si, so the Fig. 4-17, Fig. 4-18 and Fig. 4-19 were fixed on