Chapter 2 Theory
2.1.2 The pH sensitivity of ISFET
2.1.2 The pH sensitivity of ISFET
A general expression for the pH sensitivity , pHB
δ δΨ0
, which is the change of the bulk pH over a change of the insulator-electrolyte potential , Ψo , is given . The expression is derived from a separate treatment of both sides of the double layer ( gate insulator and the electrolyte ) . The site-dissocuation model introduced by Yates et al [2] in 1974 describes the charging mechanism of an oxide as the result of an equilibrium between the AOH surface sites and the H+ ions in the bulk of the solution . The surface reaction are [2] :
AOH ⇔ AO-+ HB+ and AOH2+ ⇔ AOH + HB+
where B refers to the bulk .
with the following thermodynamic equations :
μoAOH + kT ln νAOH = μoAO- + kT ln νAO- + μoHs+ + kT ln aHs+ …….…..(2)
and μo AOH +
2 + kT ln νAOH2+= μoAOH + kT ln νAOH + μoHs+ + kT ln aHs+ ....……...(3)
where νi is the surface activity and μoi is the standard chemical potential of species i .
Equation (2) and (3) can be rearranged to :
where the Ki values are dimensionless intrinsic dissociation constants ; νi is the number of sites per unit area . The relationship between the surface activity of H+ ,αHS+ , and the bulk activity of H+ , αHB+ , is given by the Nernst equation [3]:
where q is the elementary charge , k is the Boltzmann constant and T is the absolute temperature . The surface charge density , σ0 , is given by
σ0 = q ( νAOH2+ - νAO– ) = q Ns (Θ+ -Θ- ) ………..……….……..(5) where Ns is the density of the available sites ; Θ+ and Θ- are the fractions of Ns
carrying charge , i.e. , AOH2+ and AO- , respectively . The fractions Θ+ and Θ- are calculated from the Eq. (2) and Eq. (3) and substitutes in Eq. (5) to give
σ0 = q ‧Ns‧ + +
where [B] is the number of negatively charged groups minus the number of positively charged groups per unit area . pHpzc , pH at the point of zero charge , is defined as the pH were both fractions are equal and [B] is zero . The change in the number of charged groups as a result of an infinitesimal increase in pHs is the intrinsic buffer capacity , βint :
the charge in the electrolyte is equal but opposite to the charge on the oxide surface . The Gouy-Chapman-Stern model is used to describe the electrolyte side of the double layer . This model involves a diffuse layer of charge in the solution starting at a distance X2 from the surface . This distance X2 is the plane of closest approach for the center of the ions in the solution ( Stern layer ) . The charge in the diffuse layer is [4] ( Fig. 2-1 ) magnitude of the charge on the ions . the integral capacitance , Ci , is often denoted as K . The ability of the electrolyte to store charge in response to a change in the electrostatic potential is the differential capacitance [4]
Ψ0
=
-
combination of both sides of the double layer ( Eqs. (7) and (9) ) yields Substitution of Eq.(4) in Eq.(10) and rearrangement gives the general expression for the sensitivity of the electrostatic potential to changes in the bulk pH :
where α is a dimensionless sensitivity parameter . The value of α varies between 0 and 1 depending on the intrinsic buffer capacity and the differential capacitance . The site-binding theory and the Gouy-Chapman-Stern model were used in the derivation of this model , but other theories can be used as well to determine the intrinsic buffer capacity [5] , the differential capacitance and thus the sensitivity parameter α .
2.2 The ReFET
A pH reference field effect transistor (pH-ReFET) is just an ISEFT , which is less sensitive to pH . Most of the attempts to creat a ReFET are based on covering the gate oxide of an ISFET with an additional ion insensitive membrane .
Several ReFETs have been made based on different approaches , such as using a
buffered hydrogel as insensitive membrane [6] or with an ion-blocking parylene gate [7] . The first publication concerning this approach is from Matsuo , who deposited a parylene film on the Si3N4 gate of an ISFET [8] . Although the goal to make pH sensitivity decrease as possible is obtained , the stability of FET was not as expected , probably due to its pinholes on this very thin layer . Once time of measurement increases , the characteristic of ReFET will gradually get worse because of the trapping charges in the film and the loss of the ability to isolate H+ from oxide beneath . Therefore , other membranes have also been deposited , usually thicker layers , such as Teflon [9] .
Recently , PVC membrane is another choice , especially in chemical detection and analysis [10,11] . However , there are still some problems about the utility of PVC membrane . The PVC has usually not been considered for ReFETs as it typically shows cation permselectivity [12] . This behavior is common to many polymeric membranes , and means that the membranes are permeable for cations . That will affect dramatically the electrical characteristic and the stability of ISFET because of the directly contact between sensing film and electrolyte to be analyzed .
In the thesis , we develop a ReFET which can be fabricated with ISFET at the same time . To measure ISFET-ReFET pairs , a differential measurement setup shown in Fig. 2-2 to maintain both IDS and VDS constant is taken into our experiment . The upper current source forces a constant current IDS through the ISFET , while the lower current sinker forces an identical current through the fixed resistor RDS to give a constant voltage drop . The qRE we implement here is a conventional reference electrode . Under the condition of both IDS and VDS are given constant , and according to Eq. (1) , VG is varies with the change of threshold voltage , VT , which depends on the pH value of electrolyte that we analyze . For an ideal ReFET , VT does not vary and so VG(RE) is constant ; the difference , ( VG(IS) – VG(RE) ) , forms the pH-dependent
signal .
2.3 Reference
[1] P. Bergveld and A. Sibbald ,“ Analytical and Biomedical Applications of ISFETs ＂, Elsevier , Amsterdam , (1988)
[2] 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
[3] P. Bergveld ,“ Development of an ion-sensitive solid-state device for neuro- physiological measurements ＂, IEEE Trans. Biomed. Eng. , BME-17 (1970) p.70.
[4] A.J. Bard and L.R. Faulkner ,“ Electrochemical Methods Fundamentals and Applications ＂, Wiley , New York , (1980)
[5] 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 Interface Sci , 133 (1989) p.91-104
[6] P. A. Comte , J. Janata ,“ A field-effect transistor as a solid-state reference electrode ＂, Anal. Chim. Acta 101 (1978) p.247-252
[7] T. Matsuo , H. Nakajima ,“ Characteristics of reference electrodes using a polymer gate ISFET ＂, Sens. Actuators 5 (1984) p.293-305
[8] T. Matsuo , M. Esashi , in : Proceedings of the 153rd Meeting of the Electro- chemical Society Extended Abstract (1978) p.202-203
[9] H. Nakajima , M. Esashi , T. Matsuo ,“ The cation concentration response of polymer gate ISFET ＂, J. Electrochem . Soc. 129 (1982) p.141-143
[10] M. J. Madou , S. R. Morrison , Chemical Sensing with Solid State Devices , Academic Press , San Diego , CA , USA , (1989)
[11] A. Errachid , J. Bausells , N. Jaffrezic-Renault ,“ A Simple REFET for pH detection in differential mode ＂, Sens. Actuator B60 (1999) p.43-48
[12] A. van den Berg ,“ Ion sensors based on ISFET’s with synthetic ionophores ＂, ph.D Thesis , Univ. of Twente , the Netherlands , (1988)
Chapter 3 Experiment
3.1 Procedures of experiment
All procedures of experiment are done in NDL (National Nano Device Laboratory) and NFC (Nano Facility center) . The corresponding crosssection graph is illustrated in Fig . 3-1 .
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 (P) 7. N-type annealing .
Temperature = 950 for 30 min .℃ 8. PECVD - oxide for 1 μm .
9. Mask - Ⅱ. contact hole & gate region definiton .
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 α deposition .
13. Mask - Ⅲ. sensing layer α definition . 14. Etch of sensing layer α .
15. Sensing layer β deposition .
16. Mask - Ⅳ. sensing layer β definition . 17. Etch of sensing layer β .
18. Annealing in pure O2 .
Temperature = 850 for 60 min .℃ 19. Deposition of Ti / Pt .
20. Mask - Ⅴ. Ti / Pt region definition . 21. Pt annealing .
Temperature = 400 for 30 min .℃
22. Thermal coating of Al ( back side ) 5000 Å .
3.2 Experiment details
3.2.1 Gate region definition
First of all , in order to remove particles、metal ions and native oxide , RCA clean must be done in our experiment . P-type wafers were purchased from CARTINA (Table. 3-1) . Then we deposit wet oxide (6000 Å) for the define of S/D region . The oxide is also used for blocking layer during S/D implantation . The density and the energy of S/D implant is 5E15 (1/cm2) and 25 KeV in phosphorous . After implantation , N-type annealing is required for activating the dopants . That is done at 950 for 30 min . ℃
In standard MOSFET process, we do not have to deposit oxide for 1 μm by PECVD , however , it is necessary for protecting structure in using as a pH-ISFET [1] . During our measurements , the sensing region is immersed in electrolyte for a long time . The ions in electrolyte may diffuse into structure of ISFET and influence the electric characteristics . So we must deposit a thick passivation layer to avoid the influences of electrolyte [2] . Following above , we grow dry oxide for 100 Å as gate oxide after etch of PE-oxide . This layer is not only the critical structure which is significant for the electric characteristics of FET but also the key of improving adhesion toward our sensing layers .
3.2.2 Deposition of sensing layers
Subsequently , the sensing layer which is the most significant part of ISFET is deposited . We have six kinds of sensing layers which are deposited with LPCVD , E-gun or Sputter respectively in NFC . Nitride has been grown by LPCVD , and HfO2 , ZrO2 have been deposited by Sputter . E-gun is used for depositing Hf02 , TiO2 , and Al2O3 . Because the LPCVD-nitride (low pressure chemical vapor deposition) is a good material as sensing film for its high sensitivity and low drift . There are lots of researches about it [3] . And the PE-oxide (plasma enhanced chemical vapor deposition) has unstable sensitivity and drift characteristics because of being not compact structure as LP-nitride . ReFETs will be choosen through their sensitivity . All parameters of sensing layers deposition is shown in Table. 3-2 .
In this thesis , ISFET-ReFET pairs have been combined with external differential measurement setup , and that will accomplish the goal of improving long-term drift and stability .
3.2.3 S / D contact area deposition
Following sensing layer deposition , LP-nitride layer is etched by HDP-RIE (High Density Plasma Reactive Ion Etch) (Table. 3-3). And the etch of the other sensing films in our experiment is done by 49 % HF .
In deposition of S/D contact area , Pt layer of 1000 Å is chosen . However , adhesion between Pt and silicon is very bad. Ti is a good adhesion layer between silicon and Pt . So the double layer of Ti/Pt is formed by E-gun [4]. At last , 150ml HNO3, 450ml HCl and 600ml water are mixed for the wet etching of Pt . Finally , Al is deposited on the backside of the silicon by thermal coater .
3.3 Measurement system
3.3.1 Electrical characteristics measurement
In our experiment , HP-4156 is used to measure the electrical characteristics of the ISFETs . The system of measurement is shown in Fig. 3-2 . All the measurements must be done in the dark box at 25 ℃ , because of the influences of light [5] . In the setup of HP-4156 , substrate is grounded and the reference electrode is sweeping to different voltages .
Before measurement , we have to glue a plastic container right on the top of the sensing area which must contact with buffer solution . In the plastic container , we add the buffer solutions with different pH values with the dropper . At first , adding buffer solution of pH 7 for several minutes is to balance the interface potential between electrolyte and oxide . The pH buffer solutions that we use are purchased from Riedel-deHaen and the pH-value is 1 , 3 , 5 , 7 , 9 , 11 , 13 respectively .
Because the electric potential of the pH-solution is always floating [6] , the disturbance from the surroundings would induce the electric potential variance of the solution . A reference electrode must be inserted into the electrolyte for provide a constant potential . An ideal reference electrode for use as the ISFET gate terminal should provide [7] :
1) an electrical contact to the solution from which to define the solution potential .
2) an electrode / solution potential difference that does not vary with solution composition .
3.3.2 Differential sensing measurement
We use external circuit to make a differential potentiometric measurement with an instrumentation amplifier between the ISFET and the ReFET , which are both oxide-based FETs and electrically identically devices [8] . In our experiment , we still make differential sensing with a conventional electrode (qRE) to provide a constant potential . Since both ISFET and ReFET operate under the same conditions , changes in temperature and solution potential will affect both equally and then the non-ideal influences of ISFET can be alleviated . Between measurement with next one , in order to reach the accuracy of measurement, the container has to be washed by the next pH-solution after measuring previous pH value.
3.3.3 Drift measurement
On measurement of drift characteristic , Id-Vg curves at pH 7 will be extracted for every particular period we set in stress program . Then under the condition of
constant Id , we plot the diagram of Vg versus time . Drift characteristic of ISFET changes fast initially and then keeps stable several hours later . Using differential measurement of ISFET-ReFET pair , the goal to reduce this non-ideal effect can be obtained . Total measurement time is 25870 seconds and Vg depends on the Id current that we keep constant .
3.4 References
[1] U. Guth
,
“ Investigation of corrosion phenomena on chemical microsensors ＂, Electrochimica Acta 47 (2001) p.201-210[2] George T. Yu ,“ Hydrogen ion diffusion coefficient of silicon nitride thin films ＂, Applied Surface Science 202 (2002) p.68-72
[3] P. Hein ,“ Drift behavior on ISFET with nitride gate insulator ＂, Sensors and Actuators B , 13-14 (1993) p.655-656
[4] I-Yu Huang ,“ Fabrication and characterization of a new planar solid-state
reference electrode for ISFET sensors ＂, Thin Solid Films 406 (2002) p.255-261 [5] Hung-Kwei Liao,“ Multi-structure ion sensitive field effect transistor with a
metal light shield ＂, Sensors and Actuators B61 (1999) p.1-5
[6] P. Bergveld ,“ How electrical and chemical requirements for REFETs may coincide ＂, Sensors and Actuators 18 (1999) p.309-327
[7] Paul A. Hammond , Danish Ali , and David R. S. Cumming ,“ Design of a Single-Chip pH Sensor Using a Conventional 0.6-μm CMOS Process ＂, IEEE Sensors Journal , vol. 4 , No. 6 December (2004)
[8] P. A. Hammond , D. R. S. Cumming , D. Ali ,“ A Single-Chip pH Sensor Fabricated by a Conventional CMOS Process ＂, IEEE (1991)
Chapter 4
Results and Discussions
4-1 Drift of different kinds of ISFET-ReFET pairs
In our work , we choose five kinds of CMOS-compatible materials which are silicon nitride (Si3N4) 、 Hafnium oxide (HfO2) 、 Titanium oxide (TiO2) 、 Zirconium oxide (ZrO2) 、 and Aluminium oxide (Al2O3) as sensing layers of ISFET . Sputter and E-gun are used for deposition of HfO2 film . So we have twelve kinds of ISFET-ReFET pairs for demonstrate the drift characteristics of sensing layers we choose here . Table. 4-1 shows the composition of ISFET-ReFET pairs . In different kind of ISFET-ReFET pair , we keep Id constant respectively . The drift characteristics curves of every kind of ISFET-ReFET pairs in our work are shown in Fig. 4-1 ~ Fig. 4-12 . The relative drift after ten thousand seconds and the correction coefficient are listed in Table. 4-2 .
We can find that the drifts of silicon nitride film and E-gun HfO2 film are very unstable . But we can correct the shortcomings by using differential measurement setup . The correction coefficient means that the drift improving percentage after ten thousand seconds . If the ISFET-ReFET pair has compatible drift characteristics , the result of improving drift characteristics is remarkable .
4-2 To choose the proper ReFET
According to our results about ISFET-ReFET pairs , we can obtain the drift characteristics of every kind of sensing films we choose . First , we must choose a sensing layer of ReFET that has similar drift characteristics with the sensing layer of ISFET we used here . Because the ISEFT-ReFET pair has compatible drift
characteristics , we can obtain better result of improving drift characteristics . And then we take the sensitivities of sensing layers into consideration in order to define ReFET . Even if the sensitivity of ISFET-ReFET pair is not quite high after differential measurement , we still can distinguish the voltage difference due to different pH value of buffer solution .
4-3 Conclusions
In sample-1、sample-10 and sample-11 , the correction coefficient is quite high because of the compatibility of drift characteristics of two kinds of sensing layers . That is to say we can obtain a great improvement of drift characteristics . If the difference of drift between two kinds of sensing layers is large , the result of improving drift is limited (sample-3、sample-5 and sample-12) . Once the difference of drift between two kinds of sensing layers is quite large , we even may not obtain any improvement (sample-4 and sample-7) . According to our results about drift , we can choose the most appropriate ReFET that has the most compatible drift characteristics with ISFET we use here and even predict the effect of drift to sensitivity of ISFET-ReFET pair . Once the drift characteristics can be reduced as possible , the stability of ISFET-ReFET pair will get better naturally to solving a major problem of ISFET .
According to our results , the S-HfO2-Si3N4 pair has the best result of drift characteristic (0.19mv/hr after 10000 s) . And the TiO2-ZrO2 pair has the best of drift correction up to 98 % .
Chapter 5 Future Work
5-1 Future work
According to our results , we can have a data base to choose proper ReFET for the ISFET we used . We can find that some kind of ISFET-ReFET pair has great improvement to drift characteristics . In order to realize miniaturization , we can integrate reference electrode with ISFET-ReFET pair on the same chip in the future . Combining with MEMS technology , we also can fabricate a multi-channel sensor for sensing different kinds of materials at the same chip .
Besides , the external differential measurement circuit is a subject that we have to study in the future . Even we can focus on compensation of the non-ideal effect at the same time .
Figure 1-1 Schematic representation of MOSFET (a) , ISFET (b) , and electronic diagram (c) .
Figure 2-1 Potential profile and charge distribution at oxide / electrolyte interface
Sensor effect e.g. pH shift
Uat
Ua2 = Uat Ua1 = Um + Uat
Signal difference
Ua1-Ua2 = Um
ReFET ISFET
Um
Common mode disturbance : undefined metal-
liquid interface : leakage current : temperature
Uat
Figure 2-2 Differential measurement setup of an ISFET-ReFET pair
`
silicon
(a) RCA clean bare silicon
Wet oxide
(b) wet oxidation
(c) maskⅠ
silicon
(d) screening oxidation
Figure 3-1 Experimental process
silicon
(e) implantation
silicon
PE - oxide
(f) passivation layer deposition
silicon
(g) maskⅡ
silicon
(h) dry oxidation
Figure 3-1 Experimental process
silicon
(i) sensing layers deposition and mask Ⅲ
silicon
(j) Ti / Pt deposition and mask Ⅳ
silicon
(k) backside Al deposition
Figure 3-1 Experimental process
Diameter (mm): 100+/-0.5 Type / Dopant : P / Boron
Orientation : <100>
Resistivity (ohm-cm):1-10 Thickness (μm) ：505-545
Grade : Prime
Table 3-1 Specifications of wafers
TiO2 HfO2 Al2O3
Density 4.26 13.3 3.9 Z-ratio 0.4 0.36 0.336
Tooling 50.47 65 50
Current (mA) 1 ~ 60 1 ~ 60 1 ~ 50
Rate (Å/sec) 1.2 0.5 1.8
Pressure (Torr) 5*10-6 5*10-6 5*10-6
E – gun
Table 3-2-a Parameters of sensing layers deposition with E – gun
parameters of HfO2 sputter parameters of ZrO2 sputter
power : 200 W power : 200 W
Ar / O2 : 24 / 8 ( sccm ) Ar / O2 : 24 / 8 ( sccm ) Density : 13.9 Density : 6.51
Acoustic impendance : 24.53 Acoustic impendance : 14.72 Tooling factor : 0.533 Tooling factor : 0.533
Rate : 0.01 Å / s Rate : 0.01 Å / s pre sputter 60W for 10 min pre sputter 60W for 10 min
Pressure : 7.6×10-3 Pressure : 7.6×10-3
Sputter
Table 3-2-b Parameters of sensing layers deposition with Sputter
parameters of LP-nitride deposition NH3 : 17 sccm
SiH2Cl2 : 85 sccm 1000 Å for 13 min Temperature : 850 ℃
Pressure : 180 mT
LPCVD
Table 3-2-c Parameters of sensing layers deposition with LPCVD
HDP-RIE
Process pressure : 10 mTorr flow rate of CHF3 : 40 sccm
flow rate of Ar : 40 sccm ICP power : 600 W Bias power : 150 W etch time of LP-nitride : 21 sec
Table 3-3 Recipe of HDP-RIE for LP-nitride
silicon
Figure 3-2 The system of measurement
sample L R sample L R
#1 S-HfO2 LP nitride #7 S-ZrO2 LP nitride
#2 E-HfO2 LP nitride #8 S-HfO2 S-ZrO2
#3 E-TiO2 LP nitride #9 S-HfO2 E-Al2O3
#4 E-Al2O3 LP nitride #10 E-TiO2 S-ZrO2
#5 E-HfO2 S-HfO2 #11 E-TiO2 E-Al2O3
#6 E-HfO2 E-TiO2 #12 S-ZrO2 E-Al2O3
Table 4-1 Table of composition of ISFET – ReFET pairs
0 5000 10000 15000 20000 25000 30000
1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7
VG (V)
Time(s)
S-HfO2
Figure 4-1-a drift characteristic of sputter HfO2 (sample-1)
0 5000 10000 15000 20000 25000 30000 1.0
1.1 1.2 1.3 1.4 1.5 1.6 1.7
VG (V)
Time (s)
LP-nitride
Figure 4-1-b drift characteristic of LP-nitride (sample-1)
0 5000 10000 15000 20000 25000 30000
0.00 0.05 0.10 0.15 0.20
VG (V)
Time (s)
sample-1
Figure 4-1-c drift characteristic of sample-1
0 5000 10000 15000 20000 25000 30000 0.6
0.8 1.0 1.2 1.4 1.6 1.8
VG (V)
Time (s)
E-HfO2
Figure 4-2-a drift characteristic of E-gun HfO2 (sample-2)
0 5000 10000 15000 20000 25000 30000
0.6 0.8 1.0 1.2 1.4 1.6 1.8
VG(V)
Time(s)
LP-nitride
Figure 4-2-b drift characteristic of LP-nitride (sample-2)
0 5000 10000 15000 20000 25000 30000
Figure 4-2-c drift characteristic of sample-2
0 5000 10000 15000 20000 25000 30000
0.8
Figure 4-3-a drift characteristic of E-gun TiO2 (sample-3)
0 5000 10000 15000 20000 25000 30000
Figure 4-3-b drift characteristic of LP-nitride (sample-3)
Figure 4-3-b drift characteristic of LP-nitride (sample-3)