Chapter 3: Experimental and material
3.10 Preparation of the AFP sample in vitro or in vivo
AFP (AFP, α-fetoprotein) is a protein [6] [7] which in humans is encoded by the afp gene [8][9] . AFP is produced by the yolk sac and the liver during fetal life.
Like any elevated tumor marker, elevated AFP by itself is not diagnostic, only sug-gestive. Tumor markers are used primarily to monitor the result of a treatment (e.g.
chemotherapy). If levels of AFP go down after treatment, the tumor is not growing.
In the case of babies, after treatment AFP should go down faster than it would nor-mally. A temporary increase in AFP immediately following chemotherapy may indi-cate not that the tumor is growing but rather that it is shrinking (and releasing AFP as the tumor cells die). AFP-L3, an isoform of AFP which binds Lens culinaris
ag-45
glutinin, can be particularly useful in early identification of aggressive tumors asso-ciated with hepatocellular carcinoma (HCC). AFP is the main tumor marker (some-times with HCG) used to monitor testicular cancer, ovarian cancer, and malignant teratoma in any location: values of AFP over time can have significant effect on the treatment plan. For this reason, AFP is a valuable detection target for our FET bio-sensor. All AFP is purchase from Blossom Biotechnologies Inc. Taipai, Taiwan, and the all are reagent-grade quality. All the AFP samples is store in the -20 ℃.
46
Chapter 4 Result and Discussions
4.1 Physical and electrical properties of the back-gate SiNB-FET
The back-gated silicon nanobelt field effect transistors (NWFETs) on sili-con-on-insulator wafer were successfully fabricated. The width of nanobelt is about 60nm (figure 4-1 top). It is as same as the traditional metal oxide semiconductor (MOSFETs), the values of Id can be well controlled by the applied gate voltage (Fig.
4-1(b) (c). In this case, an n-channel FET is applied a gate voltage (Vg) from nega-tive to posinega-tive direction to induce neganega-tive charge in the channel. When the gate voltage was smaller, p-n junction was blocked the electrical current from drain to source, and there was a very low leakage current (Ioff) in the level of picoampere (pA). When the gate voltage is high enough, construct the conducting inversion layer and ‘open’ the channel, the drain-to-source current (Id) increase outstanding and the gate voltage is called ‘threshold voltage (Vth). The threshold voltage of FET was about -1.5 V. The current (Ion) reach different saturation dependant on different applied drain-to-source voltage (Vd). The ON/OFF current ratio is about 5 orders of magnitude. The Id-Vd curve also depends on controlled gate voltage (Vg), and back-gate SiNB-FET biosensors were characterized the physical and electrical prop-erties by 4156C semiconductor analyzer.
47 60nm. (b) Electrical properties of SiNB FET sensor. The threshold voltage of FET we fabrication is about -1.5 V, The ON/OFF current ratio is about 5 orders.
The Id-Vd curve dependant on controllable gate voltage (Vg), and the current (Ion) reach different saturation dependant on different drain-to-source voltage (Vd) we applied.
48
4.2 Real-time Detection of the PBS pH
4.2.1 Hepatocellular carcinoma will cause blood acidosis cause
According to the reported [1], hepatocellular carcinoma (HCC) and orther can-cercan alter the pH of thephysical body fluid. Because the tumor, in our paper, it stand for HCC, have a more actively and effective metabolism behavior, in order to replenish such high energy consume, more violent respiration is necessary. Such violent respiration will cause the tumor cell release more CO2 to blood. The CO2 in blood may enter into the erythrocyte, then converted to CO3 by carbonic anhydrase , and release to blood. This phenomenon will cause the blood pH decrease. In addition, many different type of HCC will released AFP, this is a acidic protein (pI = 5.5), this protein also cause blood pH decreased. For these reasons, we can think the pH of blood as a indicator of health, so we want to use the bioFET to detection the pH of saline.4.2.2 Detection response of the solution pH
Amineand oxide-functionalized SiNBs exhibit linearly pH-dependented current over a large dynamic range and could be understood in terms of the change in surface charge during protonation and deprotonation. The APTES coated process has de-scribed on previous chapter, and here we show the measurement result. At first, the PBS solution with pH 6.4 was injected to the detection region. After current was in equilibrium, the PBS solution with pH was from 6.6 to 7.4 with 0.2 in steps was then injected in sequence, and drain current of the NB FET sensor continued measur-ing at a constant gate voltage. As ahown in Fig. 4-2, the NB FET sensor exhibited de-creasing current with the pH of solution, and the single to noise ratio was >> 3,
49
showing an excellent sensing ability. The drain current versus pH is shown in Fig. 4-3, which exhibits a linearly dependence over the pH 6.4 to 7.4 range. This result sug-gests that the NB FET could be acted as the pH sensor to screen the serum pH in the future. In previous report [2], the steady state pH sensing show the logarithmic de-pendence on the target molecular concentration, (in this case, target molecular is hy-drogen ion). The net charge density on the sensor surface could be obtained based on first-order chemical kinetics of bond dissociation for the particular type of surface functionalization schemes was, in this case,–OH and –NH2.This is the main reason why the linear response could be obtained over a wide range of pH in our experiment.
Fig. 4-2 Detecting the solution pH from 6.6 to 7.4. Drain current of sensor de-creased continually dependent on the PBS we pumped, and the single /noise ratio >>
3.
50
Fig. 4-3 Drain current of the NB FET sensor as a function of solution pH. Drain current and pH value has a liner relationship. The net charge density on the NB sur-face was obtained from charge bring chemical groups.
4.2.3 Effect of the solution ion concentration to the NBFET
The solution pH was dependent to the hydrogen ion concentration, and hence the electrical response was also correlates to the hydrogen ions concentration due to the change of surface charge. Many reports have been proposed that the ion concentration of the PBS was an important factor for bioFET sensitive. This is because that the de-tection of solution pH was by monitoring the bioFET drain current, which was influ-enced by the electrical charge of the sensor surface. Therefore, the relationship of drain current and electrical charge is dependence, and it is varied with the ion con-centration of PBS.
The charge of solution-based molecules and macromolecules is screened by dis-solved solution counterions: a negative species such as streptavidin or DNA will be surrounded by positively charged ions due to electrostatic interactions. On a certain length scale, termed the Debye length (ìD), the number of net positive charges ap-proaches the number of negative charges on the protein or DNA. The result is a screening effect such that the electrostatic potential arising from charges on the
pro-51
tein or DNA decays exponentially toward zero with distance. Thus, for optimal sens-ing, the Debye length must be carefully selected for NWFET measurements because molecules binding to the devices are removed from the sensor surface by _2-12 nm.
Here we reported a result to demonstrate the effect of ion concentration of PBS. We prepared PBS with three different ion concentrations, denoted 0X, 1X, and 10X, re-spectively. The PBS having pH from 4 to 8 was injected into the sensor, and moni-tored the electrical response of the bioFET sensor. The response of sensor was ob-served to have a relationship with ion concentration of PBS. The respond of sensor is smaller upon increasing the PBS concentration, that is, less sensitive. (Fig. 4-4)
Fig. 4-4 Drain current of sensor as a function of PBS concentrations with vari-ous pH solution. Concentrated of the PBS, the respond of sensor is decrease. Three different ion concentrations PBS, there are 0.X, 1X, and 10X respectivity. And we pumped pH value 4 to 8 PBS, in the figure are use different colors to express.
52
4.3 Real-time detection of HBV x gene fragments
4.3.1 Certification of DNA immobilization by fluorescence image
In order to confirm successful immobilization, DNA with a fluorescence labeled in the 3’ terminal was employed. We used the fluorescence microscope to observe the DNA immobilization. Figure 4-5 shows the fluorescence image of the fluores-cence-labeled DNA immobilized on the 60 nmsilicon nanobelt. This is a strong proof to demonstrate the DNA exactly immobilized on the silicon nanobelt.Fig. 4-5 Fluorescence image of the fluorescence-labeled DNA immobilized on the 60 nm-silicon nanobelt.The DNA which functional fluorescence compound in its 3’ end. We use the fluorescence microscope to confirm the immobilized situation.
4.3.2 Effect of PBS to the bioFET DNA sensor
The ions in PBS have the potential to influence the response of FET biosensor.
This phenomenon may cause a bed result of sensing, for example false positive or false negative. Figure 4-6 is the experiment result to prior rectify the phosphate
ef-53
fect to our bioFET DNA sensor. A pick will appearance and following slight drain current decrease when the PBS path through the sensing zone of bioFET DNA sen-sor. (The drain current value has about 4 percent decreased)
Time (s)
Fig. 4-6 Typical electrical response of PBS injection into the BioFET sensor.
Drain current value has about 4 percent decreased, this would caused by silicon oxide and bare APTES coated on the sensing zone.
4.3.3 Real-time detection of X gene DNA fragments with various concentrations
In this section, HBV X gene DNA fragments with various concentrations includ-ing 1fM, 10fM, 100fM, and 1pM, was injected into the sensor and monitored their responses.The relationship between DNA concentration and drain current was shown in Fig. 4-7. The more concentrate DNA solution injected into the sensor, the more drain current decreased was monitored. This result was caused by the high
54
level negative charge contained in the DNA sequence. The increased negative charge hindered the conducting channel of bioFET DNA sensor, and hence decreased the drain current.
Fig. 4-7 Electrical responses of HBV X gene DNA sample with various concen-trations. 1pM sample has about 70% drain current decreasing (green), 100fM sam-ple has about 39% drain current decreasing (blue), 10fM samsam-ple has about 10%
drain current decreasing (red), and 1fM sample has about 5% drain current decreas-ing (black), but the 1fM sample result is less statistic significance. In the end, we also test the specificity of our bioFET DNA sensor
4.3.4 Determination the sensing limit of bioFET DNA sansor
The normalized current shift exhibits a good linearity with logarithmic DNA con-centration (Figure 4-8). The linear fitting for the calibration curve is y = 0.2105* log
55
(x) - 0.088, with a correlation coefficient of 0.9926. The detection limit of this bio-sensor, which is defined as DNA concentration that gives a signal intensity which is 3 times the standard deviation of the blank, is estimated to be 2.4 fM. This result demonstrated that our sensor is potential in diagnosing HBV because of its excellent senssitivity.
y = 0.2105*log(x) - 0.088 R
2= 0.9926
DNA concentration (fM)
10
-110
010
110
210
310
4N o rm al iz ed cu rren t sh if t
0.0 0.2 0.4 0.6 0.8
y = 0.2105 * log(x) - 0.088 R
2= 0.9926
Fig. 4-8 Detection limit of the DNA sensor. The detection limit is very close to 1fM level, and the detection range of detection is about four orders.
56
4.3.5 Determination of the specificity sensing
Two important factors are most concerned in fabricating a good sensor. In the prior section, we demonstrated the sensitivity of our sensor. In this section, the property of specific binding was examined. The specificity is a essential topic and a big challenge of sensing technology. In faced, most of all sensors, no matter what molecules they sense sensing, encountered the false-positive problem. We employed various sequences of DNA fragments, including 1-base, 5-bases, and all-bases mis-match DNA fragments, to demonstrate the specificity property of the sensor. Figure 4-9 shows the electrical responses of the mismatch DNA injected into the NB sensor.
The black curve is the control sample of HBV X gene DNA fragment (complemen-tary), and the current was decreased largely. The current shifts of 5-bases and 15-bases mismatched target DNA are almost invisible, while the 1-base mismatched sample exhibited a current decrease. This result implies the 1-base mismatched tar-get DNA nonspecific binds to the capture DNA probe in a small quantity, but the difference can be distinguished from the complementary sample. Similar result has been reported that the 3-mismatched mutant genes were nonspecific interaction to the nanowire sensor but can be quantified and distinguished from the wild type genes.
57
Time (s)
0 50 100 150 200
N o rm al iz ed cu rren t
0.6 0.7 0.8 0.9 1.0 1.1
complementary 1 mis-match 5 mis-match all mis-match
Fig. 4-9 Response of the mismatch DNA of the DNA sensor. Detection response of complementary, and 1-base, 5-bases, and 15-bases mismatched target DNA sam-ples. DNA sensor has no response to all of bases mismatch (green curve) and five bases (blue curve) mismatch. DNA sensor can distinguish just one mismatch DNA fragments. (red curve), compared to the complementary sequence, the drain current of one mismatch sequence is about ten percent higher.
58
4.4 Real-time detection of the liver cancer maker AFP
4.4.1 Detection the different concentration AFP DNA fragments sample solution
AFP (AFP, α-fetoprotein) is a protein [7][8], synthesis by the yolk sac and the liv-er during fetal life. In humans, AFP produce levels decrease gradually aftliv-er birth, reaching adult levels by eight to twelve months. Normal adult alpha-fatoprotein le-vels are low, but still detectable. In normal, alpha fetoprotein has no obvious func-tion, and the level maintain in a steady-state, but like other elevated cancer maker, it is a index for the program of hepatocyte tumor or liver cancer. When the health of liver goes to worsen, the AFP level would increase. On the contrary, if level of AFP goes down after treatment, the tumor is not growing. For these reasons, we chose AFP as the target molecular for trace the aggressive of liver cancer. In figure 4-10, we coated the anti-AFP antibody on the sensing zone of our bioFET, which is spe-cific captured to the AFP. Because AFP carries negative charges, the drain current will decrease when alpha fetoprotein is captured on the detection region. We pre-pared 3 ng/ml (black circle), 15 ng/ml (red triangle), 30 ng/ml (blue square), 50 ng/ml (green triangle), 100 ng/ml (brown square), 300 ng/ml (light blue crucifix), and 600 ng/ml (olive green hexagon), and we can see that there is a relationship be-tween drain current value and concentration of AFP. Figure 4-10 also shows our bioDET AFP sensor can be saturated when the concentration of alpha feto-protein is higher than 100 ng/ml, hence the detection range of our bioFET AFP sensor is about two orders. In figure 4-11 we can see the current value and AFP concentration are logarithmic dependent, except the two saturated concentration level.
59
Time (s)
0 50 100 150 200 250
N o rm al iz ed cu rren t
0.2 0.4 0.6 0.8 1.0
3 ng/ml 15 ng/ml 30 ng/ml 50 ng/ml 100 ng/ml 300 ng/ml 600 ng/ml
Fig. 4-10 Detection response with various AFP concentrations from 3 to 600 ng/ml. The drain current decreased level is dependent on the AFP concentration in the sample solution.
60
Fig. 4-11 Normalized current shift as a function of AFP concentration. The drain current value of bioFET AFP sensor and AFP concentration are logarithmic depen-dent, except the two saturated concentration level.
4.4.2 Detection the real mouse AFP level
Blood plasma is not like PBS, the plasma contained a good deal of DNA, RNA, protein, and several other compounds. These impurities, in generally, influence or obstruct the sensitivity and specificity of biosensor by erode, block, destroy, cap-ping, nonspecific binding, or change the electrical properties of biosnesor. There-fore, the detection result of biosensor may have a significant difference between targeting molecular which dissolve in PBS and in its physiology environment. For this reason, we also detected AFP level in the real mice. In figure 4-12, because the
61
AFP level is not the same [9] [10] in the infant mice and adult mice, the drain cur-rent decrease level is diffecur-rent. The adult mice AFP sample decreased the drain current value about twenty percent, while forty percent for the infant mice AFP sample. Compare to the figure 4-8, the AFP level of adult and infant mice, was es-timated to be about 15 and 50ng/ml, respectively. We picked this result to contrary to literature which study the value about mice [11-16], and foung the AFP level of real mice was very similar to that of in PBS. This result indicates the influence of impurity is not very obvious in our bioFET AFP sensor.
Time (s)
Fig. 4-12 Detecting response of AFP concentration from adult and infant mice.
The AFP level of adult mice and baby mice, is about 15ng/ml, and 50ng/ml, and cause the drain current decrease 20% and 40 %, respectively.
62
4.4.3 Real-time detection of the AFP with various PBS concentra-tions
The relationship of drain current and electrical charge there are not always in con-stant, it’s varied with ion concentration of PBS. Because of within the ion concentra-tion charge, the debye length change, and this phenomenon decide the ability of sensing of our bioFET DNA sensor. Figure 4-13 shows the influence of PBS ion concentration by our sensor detection. The PBS with 10X, 2X, 1X, and 0.1X in concentrations was injected into the sensor in sequence. As shown in Fig. 4-13, the bioFET AFP sensor in10X PBS environment was less sensitive, and the bioFET AFP sensor in 0.1X PBS was the most sensitive. The relationship between drain current change and PBS was logarithmic dependence. This result is consistent to other study [17], and intuitively obvious because screening by the ions suppresses the overall charge effective in modulating the sensor response. The results also show that opti-mizing the ion concentration of PBS in experiments is the critical importance in studying biosensors, because the same target molecules might result in different sen-sitivity with different concentration electrolyte.
63
Fig. 4-13 Influence of PBS ion concentration by our sensor detection. We pr e-pared the 10X, 2X, 1X, 0.1X PBS, and we pumped 10X, 2X, 1X, 0.1X, and back to 10X PBS in sequence. And monitor the drain current value. As the result, we can see the bioFET AFP sensor in10X PBS environment is most less sensitivity , and the bioFET AFP sensor in 0.1X PBS is most sensitivity. And the relationship between drain current change and PBS is logarithmic dependent.
64
4.5 Life time identification of the self-assembled monolayer
4.5.1 Identification of the ion impurities trapped in SOI wafer
In previous experiments described above we used PBS as buffer solution, the PBS buffer was,however, might contain problems. The PBS solution contained sodium ions, which was reported to affect device performance due to ions trapping pheno-menon into the silicon.[18]. The sodium ions might alter or destroy the crystal lattice, resulting in current decreased and unstable of the device. Here we reported the experimental result of the electrical performance about the PBS soaped effect, which was caused by the sodium ions in the PBS buffer. As shown in the Figure 4-13, the current was increased dramatically after 0.5 day PBS soaping, which might be caused by the diffusion of sodium ions into oxide film. This result was similar to the well-known mobile ion charge trapped in the oxide film, resulting in the left-shift of the ID-VG curve. Therefore, The drain current was increased as biased at a constant gate voltage. After two day immersing, the drain current was continuous-ly decreased with days. In addition, the decreased current ratio of the sensor was dependant to the PBS concentration. This result illustrated that the current path of the NB FET was obstructed by the sodium ions.
65
Fig.4-14 Drain current shifts of the NB FET sensors as a function of PBS soap-ing days. The prior increase drain current caused is by the parallel currents through PBS solution. After 0.5 days and later, drain current decreased as the daytime in-crease. And high level ion concentration causes the drain current degeneration much serious. It believed caused by more sodium ion trapping into the SOI wafer crystal lattice.