Chapter 1 Introduction
1.4 Plasma treatment and REFET
Thus, knowing from the above described, the sensing properties of pH-ISFET are
depending on various materials and the situation of oxide/electrolyte interface.
Whatever the materials covered on the gate layer are, the main purpose of using the PVC membrane or Nifon is to reduce the influence of channel region and suppress ion interfering effects[12][13]. Here, the purposed plasma treatment method is finding other way to do so, and it’s also reducing the complex process to glue the PVC membrane or Nifon on the gate. And membranes ZrO2 and TiO2 are used in this work, because of the stable electrical characteristic compared to the Si3N4 [14] , and SiO2
[14]. And the higher sensitivity then Si3N4 and SiO2 [15] films will be presented in this work.
According to the theory of site-binding model, the sensitivity of pH-ISFET is related to the influence of interface between oxide and electrolyte. The plasma treatment is proposed for trying to recover the dangling bonds in this work to reduce sensitivity. The manner of plasma treatment is usually breaking the bond between the atoms, making more dangling bonds on the sensing membrane surface during longer time treatment and recovering dangling bonds during shorter time treatment. Here, the combination between plasma radicals and surface atoms for shorter treating time presents lower sensitivity than without plasma treatment, and the higher sensitivity for longer treating time. Whatever, a result of either making more trap state (dangling bond) of H+-ions or combination between plasma radicals and surface atoms (recover dangling bonds) can be seen from the relationship between sensitivity and surface plasma treating time by TiO2 sensing membrane with NH3 plasma treatment. In this work, we are trying to find the best process windows for TiO2 and ZrO2 sensing membranes
Finally, the coplanar structure can form a coupled ISFETs or just a REFET by the difference which eliminates the unnecessary interferences such as temperature effects, ion perturbations and lightening conditions under detection.
1.4 Thesis organization
In the first chapter, a brief history of ISFET and the theory developed by those great people was introduced. And the reason why we use plasma to treat the sensing membrane surface is also addressed in chapter 1. The detailed theory, including the band diagram of ISFET, site-binding model introduced by Yates et al , and the sensitivity of ISFET, are described in chapter 2. In this chapter, the brief introduce of REFET is presented. In the next section, the entire experiment procedures and measurement setup is presented in detail. The various kinds of sensing membrane are used to produce the coplanar structure ISFET-REFET pairs. In the last two chapters, some thoughts about the results are proposed and the conclusions are presented, too.
Finally, some works are presented to do in the future.
1.5 References
[1] Yong Jiang, Wulin Song, Changsheng Xie, Aihua Wang, Dawen Zeng, Mulin Hu,
“Electrical conductivity and gas sensitivity to VOCs of V-doped ZnFe2O4 nanoparticles” , Material Letters, 60(2006) p.1374-1378
[2] M. Zagnonia, A. Golfarelli, S. Callegari, A. Talamelli,V. Bonora, E. Sangiorgi, M.
Tartagni, “A non-invasive capacitive sensor strip for aerodynamic pressure measurement” , Sensors and Actuators: A. Physical 123-124 complete(2005) p240-248
[3] P. Bergveld ,“Development of an ion sensitive solid-state device for
neurophysiological measurements＂, IEEE Trans.Biomed. Eng.,vol. BME-17 (1970) p.70
[4] D.E. Yates , S. Levine and T.W. Healy ,“Site-binding model of the electrical
double layer at the oxide/wafer interface ＂, J. Chem. Soc. Faraday Trans. , 70 (1974) p.1807-1818
[5] J. Lyklema, “The electrical double layer on oxides” , Croatica Chem. Acta, 43 (1971) p.249
[6] T. Hiemstra, W. H. van Riemsdijk, and G. H. Bolt, “Multisite Proton Adsorption Modeling at the Solid/Solution Interface of (Hydr)oxides: A New Approach” , J.
Colloid and Interface Sci. 133(1989) p91
[7] R. E. G. van Hal, J. C. T. Eijkel, P. Bergveld, “A general model to describe the electrostatic potential at electrolyte oxide interfaces” , Colloid interface Sci.
68(1996) p.31-62
[8] Alexey P. Soldatkinet al, “Analysis of the potato glycoalkaloids by using of enzyme biosensor based on pH-ISFETs” , Talanta 66(2005) p.28-33
[9] Henning Matthiesen, “In situ measurement of soil pH” , J. of Arch. Sci.
31(2004)p.1373-1381
[10] Arshak Poghossian et al, “Chemical sensor as physical sensor: ISFET-based flow-velocity,flow-direction and diffusion-coefficient sensor” , Sensors and Actuators B 95(2003) p.384-390
[11] P. Estrela et al, “Field effect detection of biomolecular interactions” , Electrochimica Acta 50 (2005) p.4995-5000
[12] Z.M. Baccar, N. Jaffrezic-Renault , C. Martelet , H. Jaffrezic, G. Marest, A.
Plantier
“Sodium microsensors based on ISFET/REFET prepared through an
ionimplantationprocess fully compatible with a standard silicon technology”
Sensors and Actuators B 32 (1996) 101-105
[13]Michal Chudy, Wojciech Wro´ blewski, Zbigniew Brzo´zka “Towards REFET”
Sensors and Actuators B 57 (1999) 47–50
[14]Fukuzawa Y, “Machining characteristics of insulating ceramics by electrical discharge machine” INDUSTRIAL CERAMICS 21 (3): 187-189 SEP-DEC 2001
[15]Jung-Chuan Chou, “Sensitivity and hysteresis effect in Al2O3 gate pH-ISFET”
Materials chemistry and physics 71(2001) 120-124
Chapter 2
Theory & Principle
2.1 Operation theory of ISFET
The operation theory of an Ion Selective Field Effective Transistor (ISFET) is similar to a MOSFET. Considering the following structure of a MOSFET
Al_b | Si | SiO2 | Al_t
Al_b : the back side of silicon coated Al as electrode
Al_t : the top side of silicon coated Al as S/D/G electrical contact
Before the different materials contact to each other, the flat-band voltage is build. [1]
When these materials contact to each other and form a MOSFET structure, which result in the potential differences in between these materials has been presented in the band diagram. Through the band diagram, we can obtain the flat band voltage as the E.q.
OX SS ms
FB C
Q V = q1Φ −
The same properties of ISFET are presented as following. The electrolyte layer was inserted in between oxide and metal layer, and we take the SiO2 film as the sensing layer to detect the specific ions in electrolyte. Forming the following structure,[2]
Al_b |1 Si |2 SiO2 |3 Electrolyte |4 M |5 Al_t
The couple layer M | Al_t was taken as the reference electrode. The reference electrode is not the key subject in this study, the considered couple layer presented here is used to simplify the model of reference electrode. Figure 2.1 shows the above ISFET structure band diagram. Considering the above structure as cell, then the applied voltage of the cell can be written as follows
b
Due to the equilibrium at interfaces 1 and 5 , Eq.(1) reduced to ) Where the electrochemical potentials have been considered as chemical and electrical contributions. Because of the electrical contributions, the reference electrode part must be considered. The following Eq. is the definition of reference electrode.
)
1 M ( M Si
ref q
E =− Φ + φ −φ (3) Here Eref was named “reduced absolute electrode potential” by Trasatti[R].
Substituting the Eq.(3)into Eq(2) gives
Si The difference terms solution bulk (φsol) and silicon bulk(φSi) , can be separated as follows As can be seen in band diagram (Fig2.1), each term on the right hand side can be interpreted
ψ0
φ
φbSol − dSol =− (6)
Where ψ0 is the potential drop in the electrolyte at the oxide/electrolyte interface The potential drop across the oxides (Vox) and the silicon surface potential (ψ0) are presented above From Eqs (4)-(8) and (9) and (10) the following expression for the flat-band voltage is obtained as below Taking the perfect interface of oxide/solution and oxide/silicon into account, meaning that the condition of interface is not concluded in this study, the following Eq. is obtained Thus, the flat-band voltage, suited with the EOS structure, is obtained
OX The ψ0 term, which is presented in Eq(13), determines the operation of the EOS structure as chemical sensor. It depends mainly on the solution pH, in the case of oxide material???. In particular, the solution pH at which ψ0 = 0 is called the pHpzc
(point of the zero charge)
2.2 The site-binding model and the sensitivity of ISFET
According to the above detail, the parameter of ψ0 plays an important role of the sensitivity. Obviously, the interface condition on oxide/solution must be considered by the combination of the oxide-ion. Yates et al [3] introduced the site-binding model based on the adsorbed counter ion form interfacial ion pairs with discrete charged surface groups. The direction of plane was considered [4], but did not make the same consideration in this study. And the influence of the porosity of the layer [5] was not concluded here.
Considering the oxide surfaces as amphoteric, meaning that the surface hydroxyl groups can be natural, whatever positively charged (protonized) and negatively charged (deprotonized). The charging mechanism of an oxide is the result of equilibrium between the AOH surface sites and the H+-ions in the bulk of the solution.
And the surface dissociation reactions are [6] :
+ From above reaction we can get the following thermodynamic equations :
+
where νi is the surface activity and μoi is the standard chemical potential of species i.
Following Eq. (1) and (2) and the definition of the activity between surface and
bulk solution
where the K values are dimension less intrinsic dissociation constants. From the above Eqs , showing that the K values are real constants independent of the ionization state of the oxide surface. Then the surface charge density, νi , is obtained as follows
)
Taking the pH variations in the oxide/electrolyte interface into account, then the surface charge density versus the pH variation on the surface can be calculated as following definition
0 [ ] βint βint is called intrinsic buffer capacity, depending on the activity of surface H+-ions . And thus we can finally find the expression for the intrinsic buffer capacity
+ Because of the βint is dependent on the variation of surface charge density and activity of surface H+-ions, we may consider the intrinsic buffer capacity as the parameter for the sensing material. And that , there are several parameters affecting the active surface groups, e.g. the valence of the metal ion. Hiemstra et al.[7] introduced a multisite complexation model (MUSIC) to describe the charging mechanism and to estimate the value of intrinsic dissociation constants of the active surface groups from physical parameters. But these factors are specific for any particular oxides having different reactive groups are present on different oxides. The general expression for all types of oxides can not be achieved. Every oxide should be treated separately.
Because of the amorphous type of the sensing layer on the ISFET in this experiment, the MUSIC model is not suitable here.
Next, according to the charge neutrality, an equal but opposite charge is built up, σDL , in the electrolyte solution side of the double layer. Thus there will be, something like capacitor (Fig2.2), built up in the oxide/electrolyte system. We can obtain the following Eq. by such equilibrium in Boltzmann equation:
) Where φx is the potential at any distance x with respect to the bulk of the solution; ci(x) and ci0 are the molar concentrations of species i at a distance x and in the bulk of the solution respectively and zi is the magnitude of the charge on the ions. And the combination of the Boltzmann and Poisson equation the related charge density with the potential is obtained as follows:
0 Considering that the ions adsorbed on the oxide/electrolyte interface as a couple layer ,
inner layer (Stern layer) and outer layer (diffuse layer) , made the potential drop on Instead Eq(12) in Eq(11) , and differentiating and rearranging Eq(11) , gives the following equation :
To simplify the above equation as follows
)
There will be seen easily , the differential capacitance can be distinguish into two parts , the first term is the contribution of the called Stern layer , the second term is the contribution of the diffuse layer. Then the following equation will be obtained:st
From now, the appearance of sensitivity will be discovered. Considering the activity between surface and bulk solution , the Eq(3) is repeated here.
kT
From above equation, we can obtain the following expression ,kT Taking the surface charge density into account , the variation of σ0 versus the potential drop will be presented as follows,
dif
Combination of (8) and (18) leads to an expression for the sensitivity of the electrostatic potential towards changes in +
HS The next expression is given by the combination of (19) and (17)
kT dif Finally, rearranging of (20) gives a general expression for the sensitivity of the electrostatic potential to changes in the bulk pH:
δ α
The sensitivity parameter α is dimensionless and the value varies between 0 and 1 depending on the intrinsic buffer capacity and the differential capacitance. Where the maximum value of sensitivity is about 60 mV/pH. In the experiment of this study, the higher sensitivity about 70 mV/pH is presented, however , the theory of the higher value of sensitivity are still discussing. Does the lower sensitivity of the ISFETs be useless ? According to the introduction of REFET , the principle of REFET will be presented as the following section.
2.3 The Principles of REFET
Because the requirements of a stable reference electrode for the potentiometric sensors to do proper functioning, the same meaning as the ISFET , the so called reference field effect transistor (REFET) is developed. The major characteristic of REFETs is the lowest sensitivity for the detection under such an environment we appointed. A pH REFET is developed in such a thought, making the lower sensitivity of pH. Considering the theory of ISFET is described above, the sensitivity parameter α relates the differential capacitance and the intrinsic buffer capacity , as pointed to Eq(22).
From the eq(22) where the lower sensitivity appeared , the lower intrinsic buffer capacity will also be obtained. The issue of intrinsic buffer capacity and it’s relation to the effective sites (NS) on the interface between oxide/electrolyte is described and is observed in the eq(8). An assumption of the lower sensitivity was proposed by recover the effective sites (NS) on the interface between oxide/electrolyte. Plasma treatment is one of solutions proposed to recover the dangling bonds on the interface between oxide/electrolyte. As can be seen in the final result of the experiment, the sensitivity was reduced successfully by such post plasma treatment on the sensing material surface.
In the experiment, we will discuss the effects of the plasma treatment. Which include the nitridation of NH3 plasma and the increase of the site density of dangling bonds on the sensing membrane surface.
2.4 References
1. Neamen, Donald A , “Semiconductor physics and devices :basic principles”, McGraw-Hill , 2003
2. Luc Bousse , “Single electrode potentials related to flat-band voltage measurements on EOS and MOS structures” J. Chem. Phys. , 76 , (1982) p.5128-p.5133
3. D.E. Yates , S. Levine and T.W. Healy ,“ Site-binding model of the electrical double layer at the oxide/wafer interface ”, J. Chem. Soc. Faraday Trans. , 70 (1974) p.1807-1818
4. Fabien Gaboriaud , and Jean-Jacques Ehrhardt , “Effect of different crystal faces on the surface charge of colloidal goethite (α-FeOOH) particles: An experimental and modeling study” , Geochimica et Cosmochimica Acta, 67(2003) p. 967-983
5. J. Lyklema , “The electrical double layer on oxides” , Croatica Chem. Acta, 43 (1971) p249
6. R. E. G. van Hal, J. C. T. Eijkel, and P. Berveld , “A general model to describe the electrostatic potential at electrolyte oxide interfaces” , Adv. Colloid Interface Sci. 68 (1996) p.31-62
7. T. Hiemstra, J. C. M. de Wit, and W. H. van Riemsdijk, “Multisite proton adsorption modeling at the Solid/Solution interface of (Hydr) oxides: A New Approach II. Application to various important (Hydr)oxides”, J. Colloid and Interface Sci. , 133(1989) p. 105-116
Chapter 3 Experiment
3.1 Introduction
ISFET has the same manufacturing process as the conventional MOSFET. The difference in MOSFET and ISFET procedure is the process of gate electrode. The ISFET take the gate membrane as a sensing layer immersed in the pH-solution [1], and the reference electrode is placed overhead the sensing layer as the gate voltage controller. Furthermore, the strong development of IC industry assists the procedure of ISFET more easily, but there still have a lot of problems confused us. Purposed plasma treatment on sensing layer may find a way out of the confused issue.
Furthermore, applying the successful integration-circuit technology, the ISFET devices have potential advantages over conventional ion selective glass electrodes in their rapid response, low cost, small size, high input impedance and low output impedance.
3.2 Procedures of ISFET
All procedures of experiment are done in NDL (National Nano Device Laboratory) and NFC (Nano Facility center), similar to the manufacturing process of MOSFET [2].
The corresponding diagram of ISFET is shown in Figure 3.1. The sensing layers titanium dioxide and zirconium dioxide are deposited onto the SiO2 gate ISFET which prepared by E-gun and Sputter in Nano Facility center. Before every step, besides after sensing membrane deposited onto SiO2 gate, the initial clean immersed in
H2SO4+H2O2 about 5 minutes and dipped in HF solution were done. The fabrication parameters are listed in Table 3.1, and the fabrication procedures are listed as follows:
1. RCA clean .
2. Wet oxidation 6000 Å .
Temperature = 1050℃ for 65 min . 3. Mask -Ⅰ. S / D definition .
4. BOE etch wet oxide .
5. Dry oxidation for screening 300 Å . Temperature = 1050℃ for 12 min . 6. S / D implantation .
5e15 (1/cm2)，25Kev (Phosphorus) 7. N-type annealing .
Temperature = 950℃ for 30 min . 8. PECVD - oxide for 1 μm .
9. Mask - Ⅱ. contact hole & gate region definition .
10. BOE etch PECVD - oxide for 1 μm (contact hole region) .
PECVD - oxide for 1 μm+ wet oxide for 6000 Å (gate region) . 11. Dry oxidation 100 Å ( gate oxide ) .
Temperature = 850℃ for 60 min . 12. Sensing layer deposition .
* Deposit titanium dioxide by E-gun and zirconium dioxide by sputter, respectively
13. Plasma treatment
*We use the NH plasma to treat the sensing membrane surface by
several time interval.
14. Annealing in pure O2 .
Temperature = 850℃ for 60 min .
15.Thermal coating of Al ( electrical contact ) 5000 Å . 16.Sintering Al electrode
Temperature = 400℃ for 30 min .
3.3 Experiment detail
3.3.1 Gate region formation
At the beginning, The RCA clean has to be done. RCA clean was used to reduce possible pollution such as particles, organics, diffusion ions and native oxide. For the good performance in this work, we take the RCA clean carefully. And for a good efficiency, the p-type wafers were purchased from CARTINA. The device structure was definite before wet oxidation. Following above step, the SiO2 was grown by thermal wet oxidation 600nm for blocking layer under the S/D ion implantation. The density and the energy of S/D implant is 5E15 (1/cm2) and 25Kev with phosphorus dopant, respectively. After S/D implantation, following a 950℃ 30 minutes N+
anneal in N2 gas performed to activate the dpants. Noted that, in the following experiment, the temperature here, must be lower than the annealing temperature to prevent dopant redistribution.
After all, the PE-oxide film about 1μm was deposited onto the original 600nm thickness oxide, which protect the structure of a pH-ISFET blocking the ions diffuse when immersed in electrolyte. Thus, a stable electrical characteristic can be obtained.
Next, the gate region was definite by lithography. Before deposit sensing membrane,
the 100A think oxide grown in oven as gate oxide, making the sensing layer adhesion more tightly.
3.3.2 Sensing layer deposition
Methods of deposited sensing membrane as gate material are different. And the deposition methods decide the electrical characteristic of pH-ISFET affected by the inner cave [3]. Before this work, the research about method of deposit sensing membrane has been done by the group I belong to. Thus, in this work, titanium dioxide is prepared by E-gun under the degree of vacuity about 10-6 and zirconium dioxide are deposited by sputter to reduce the inner cave under the degree of pressure about 10-3 (here, the sputter is pumped till the degree of pressure about 10-6 first, then the argon plasma is used to sputter the target material). To prevent the sensing membrane react with acid/base, the hard mask is used in such a thought. The detailed parameters of sputter are listed in Table 3.1 (c). Table 3.1 (b) lists the parameters of the dual E-gun system.
3.3.3 Plasma treatment
Subsequently, the PECVD system is used to produce plasma to treat the sensing layer. The plasma system used in this work, located in the Nano Facility center in National Chiao-Tung University, Hsinchu. Three kinds of plasma are used to do such a treatment in the first round of the experiment. The NH3 plasma is used to treat sensing membrane surface for several time interval. The PECVD system is a couple of parallel planes which is DC-biased. The one is that RF power increases, it will
Subsequently, the PECVD system is used to produce plasma to treat the sensing layer. The plasma system used in this work, located in the Nano Facility center in National Chiao-Tung University, Hsinchu. Three kinds of plasma are used to do such a treatment in the first round of the experiment. The NH3 plasma is used to treat sensing membrane surface for several time interval. The PECVD system is a couple of parallel planes which is DC-biased. The one is that RF power increases, it will