The purpose of this study is to give detailed information about fabrication and characterization of resistance temperature detector by smart mask design. The effects of annealing on both structural and electrical properties were investigated. Changes in micro strain, lattice parameter and grain size values were observed by means of annealing. It has been shown that the structural changes cause a decrease in resistivity and sheet resistance values. Thousands of sensors can be fabricated according to the substrate size of thin film RTD produced. It is almost impossible to fabricated thousands of sensors with the same resistance value. But by the agency of the smart mask design proposed in this study, it is possible to adjust each RTD to the same resistance value by using the resistance adjustment points. Maximum TCR value was found to be around 3.84x103 ppm/0C. This value is very close to the standard TCR value 3.85x103 ppm/0C used for industrial applications.
Research Article
Fabrication and Characterization of Resistance Temperature Detector By Smart Mask Design
https://doi.org/10.21203/rs.3.rs-949037/v1
This work is licensed under a CC BY 4.0 License
You are reading this latest preprint version
Today's technology has been improved continuously. All electronic circuits are produced in a smaller size. One spends a lot of energy in discovering sustainable materials and using them in direction of our goals and objectives[1][2]. New generation technologies, especially cell phones and spacecraft, require smaller electronic systems in terms of both weight and economy[3]. It is getting harder to produce smaller electronic circuit elements. Errors arising from this production increase at the same rate. Many systems have been developed in recent years and these systems can be remotely controlled. For example, many systems such as automobiles and industrial machines act automatically independent of the user. Almost all industrial systems use more than a few sensors growth per year. As a result, the demand of this industry directly affects the sensor markets and causes 10%. Out of this sensors types the temperature sensors are developing faster and faster [4–7] Temperature is the most frequently measured environmental quantity [8]. Physical, electronic, chemical, mechanical and biological systems are affected by temperature [9]. A temperature sensor is also of great importance in many different production stages that we use in our daily lives. It has a critical importance for many industries such as food processing, machinery, chemistry and computers. In addition, temperature sensors are used in many kitchen appliances in our houses [10, 11].
Temperature sensors are basically divided into types. The first is non-contact type and the second is contact type sensors. The non-contact type sensors work based on the thermal emission of electromagnetic radiation. RTD- Resistance Temperature Detector, thermocouple and thermistors are known as contact type sensors[12]. The thermocouples were invented by Thomas Seebeck in 1821; It is formed as a result of joining both ends of two different metals or metal alloys by welding. The temperature difference between different metals or alloys creates an electromotive force. This force, known as the Seebeck effect, is proportional to the temperature difference and this electromotive causes a voltage in the order of millivolts at the cold point ends[13]. This phenomenon is the working logic of thermocouples. The increase in the resistance of a metal with temperature is the basis of resistive temperature sensor RTDs (PTC). The resistances of the elements which is made of metal conductors are directly proportional to the temperature. Like RTDs, thermistors (NTC) are temperature sensitive resistors[14]. The thermistors are made of semiconductor materials and the resistance of semiconductor materials is inversely proportional to temperature. Among temperature sensors, RTDs are more advantageous than their competitors due to their small size and more reliable results. RTDs are more expensive than their competitors. In addition, some errors arising from the production stages increase the cost. Thanks to the smart mask design in this study, it will contribute to the reduction of these errors.
The aims of this study are given as (a) effects of different annealing temperatures on structural properties (b) investigation of resistivity changes with temperature (c) smart mask design and investigation of properties (d) effects of platinum layer thickness (e) Reducing production errors with smart mask design.
2.1. Fabrication
Alumina wafers, which are 500 µm thick, were cleaned by RCA cleaning process. It was dried with nitrogen gas. Lift-off lithography method was used to produce a resistance temperature detector. The 3D schematic diagram of the application stages of the Lift-off lithography method is shown in figure 1. To create the pattern, ma-N 1410 Photoresist was coated at 3000 rpm. The photoresist thickness is approximately 1 µm. UV exposure dose was about 600 mJ/cm2. Then, prebake was done at 100 degrees for 3 minutes. After photoresist coating and prebake processes, ma-D 533/S developer was used to remove the pattern trace. After the photoresist coating process, hard back was made for 3 minutes and the substrate was immediately loaded into the RF magnetron sputtering system. The pressure of the vacuum chamber was reduced to 6.0 x 10−4 Pa. A 99.9999 pure platinum target with a radius of 4 inches was used to deposit the platinum layer on the substrate. In order to deposit the platinum layer more firmly on the alumina wafer surface, the substrate temperature was adjusted to 90 0C. Pre-sputter was done at 300 watts for 10 minutes in order to get rid of the dirt on the target surface. Following the pre-sputtering, the 1-1.05 µm thick platinum thin films were deposited onto alumina wafers at 300 watts for 25 min. Then, mr-Rem 700 remover was used to remove photoresist residues. To measure the thickness and examine the structural properties of the temperature resistance detectors, the platinum layer was deposited on the same alumina substrate following the same procedure. Then, the thickness and uniformity of the platinum layer were measured by ellipsometry. Platinum thin films were divided into 4 parts and annealed under nitrogen at 200 0C, 400 0C, 6000C and 800 0C degrees for 150 minutes and cooled very slowly. Crystallographic structures were investigated by Rigaku X-ray diffractometry (XRD). The sheet resistance was analyzed by signatone 4-point probe. The resistance change with temperature was measured between 0 0C and 450 0C degrees by means of a closed probe station in which the semiconductor characterization system 4200 SCS was embedded. TCR values were calculated with increments of 50 0C from 0 to 450 0C degrees.
The platinum films were deposited using the RF magnetron sputtering method. The X-ray diffraction patterns of these films at different annealing temperatures are shown in figure 2. The peak at 400 belongs to cubic Platinyum (111). The XRD data show consistency with the ICDD Data Card No:87-0640. Among these annealed samples, it was found that the samples annealed at 800 degrees were oriented along (111) with the best crystal structure. Orientation occurs more easily in thinner film structures (< 100 nm). The orientation (111) about 1 µm thickness was also shown in this study. The (111) orientation of metal thin film structures is necessary to minimize the surface energy[15, 16].
Annealing temperature dependency of the change in Grain size and micro-strain values are shown in figures 3 (a) and (b). Pt thin films annealed at 200 0C have lower grain size values. Low grain size caused "high density of grain-boundary" and "the density of grain boundary". For this reason, the resistivity was high at 200 0C. With the increase of annealing temperature, grain size increases and micro strain values decrease.
Figure 4 shows the relationship between XRD peaks and micro strain. There is no strain in the ideal crystal structure, so the reflective planes show uniform d0 spacing. In the presence of a uniform tensile strain, the spacing of the reflective planes becomes greater than d0 and the diffraction line shifts to lower angles. This situation is clearly observed in XRD measurement. These changes helped the Pt film structure to become more stable. As a result, it contributed to make the resistivity value more stable. Non-uniform / uniform strain causes the grain to grow. This bending is effective in increasing the grain size[17].
The variation of the lattice parameters (a0) with the annealing temperature is shown in figure 5. The lattice parameter decreased with the annealing temperature. The nature of the residual stress of most annealed Pt films examined was found to be tensile. This situation has been demonstrated in many studies[18–20]. This situation is directly related to the substrate; because the substrate thermal expansion coefficient is 7.6 x 10− 8 cm/°C [21] and the platinum thermal expansion coefficient is 2.26 x 10− 5 cm/°C [19]. As a result of this great mismatch between the expansion coefficients of the platinum film and the substrate is responsible for the variation of the tensile stress and the lattice parameter [22].
2.1 Properties and Smart Mask Design
The smart mask design designed for the production of temperature sensors is shown in figure 6. The calculation of the resistances of metal thin films is directly related to the length and thickness of the thin film. The following relation can be used for the calculation of the resistance [23].
(1)where R, ρ, L and A are resistor resistance, platinum resistivity, is the length and is the platinum layer thickness, respectively. The thickness of the platinum layer is very important in the production of temperature sensors. When a platinum layer is deposited on a 6-inch subsrate, the thickness of the platinum layer one produce will vary everywhere. The thickness in the middle part of the temperature sensors produced on the same substrate is thicker than the edge parts. This situation directly affects the resistance values of RTDs. The platinum thickness of these sensors in the edge is 1 and 1.05 µm in the middle part. The platinum layer thickness difference between these sensors is only 0.05 µm. The total resistance values resulting from this thickness difference will be discussed in the next sections.
The sheet resistance and resistivity variation depending on different annealing temperatures are shown in figure 7 (a) and (b). The samples were annealed under nitrogen gas for 150 minutes. After annealing, each sample was slowly cooled. Sheet resistances were measured at 23 degrees. Resistivity was calculated using sheet resistance and platinum film thickness from the equation 2.
ρ = Rs.t (2)
Rs is the sheet resistance t is the thickness of the platinyum layer. As shown in figure 7 (b), as the annealing temperature increased, the resistivity decreased and became more stable between 600 and 800 degrees. The dramatic decrease in sheet resistance after the first annealing temperature is due to the correction of defects in the structure by annealing. The possible reason for the annealing temperature decrease of sheet resistance is due to grain size growth and vacancy coalescence. In addition, micro-strain effects should not be ignored.
There are 2 different resistance adjustments on this mask design and they are named S and K on the mask design as seen in Fig. 6. Temperature sensors should have a certain value according to their usage areas. The resistances of the most used temperature sensors are Pt100, Pt500 and Pt 1000. Due to the thickness of the platinum layer, all of the thousands of sensors produced have different resistances. In order to produce all sensors at the same resistance value, one can adjust the resistance value by making some cuts in the S and K parts.
S part
As seen in Fig. 8, some cuts change the resistance value. The path that the current follows is getting longer. Cutting perpendicular to the current flow in the resistor is known to increase the resistance rapidly to a value close to the target value. A cut parallel to the current flow slows the rate of resistance increase to reach the target resistance. In this section, one can apply thousands of different resistance corrections. Phillip Sandborn and Peter A. Sandborn have examined this issue in detail in their studies [5]. We could not examine this part because we could not make smooth and linear cuts in the S part.
K part
Figure 9 clearly shows that there are two paths that the current can follow. The paths followed by the current were shown in different colors. The red line represents the resistor 1 (R1) and the black line the represents (R2) and the path of the current is shown in the figure 9. Due to the parallel connection of R1 and R2, the equivalent resistance given by Rt [24].
(3)It is well known that the equivalent resistance value of resistors connected in parallel is lower than each resistance of each resistor. The values of R1 and R2 are 8.47Ω and 2.97Ω respectively for a temperature sensor with a thickness of approximately 1.05 um. Uniformity is very critical for microelectronics. This is especially important for temperature sensors. We have previously stated that the total resistor resistance is directly related to the thickness of the platinum film. In the case of parallel connection R1and R2, its contribution to the total resistor resistance is 2.19 Ω. When a cut is made on the R2 resistor, the path is completely blocked, the total resistance value will increase to approximately 6.27 Ω. It should be noted that the contribution from connecting R1 and R2 in parallel is 2.19 Ω. There are 9 resistance adjustment points to adjust the resistance value. The theoretical calculation of the resistance change and the experimental results in case of cutting the 9 point, taking into account the thickness difference, are shown in Fig. 10. The real image of the temperature resistance detector under the optical microscope is also given in figure 11.
The linewidth of the temperature sensor is 10 µm. The total uncut length of the trimming points is 45150 µm. The total resistance of the sensor with a 1.05 µm meter thick platinum layer is 473 Ω. Therefore, the thickness is very important when fabrication an industrial scale. When the platinum thickness is 1 µm, the total resistance of the sensor is 497 Ω. So, The total resistance of the sensors with a thickness difference of 0.05 µm varies about 24 Ω. In the previous section, it was stated how important the uniformity on a 6-inch substrate is when depositing the platinum layer. For industrial applications, this is of critical importance. Thanks to this smart mask design, it can complete the resistance of each sensor about 500 Ω. But if the resistance of the produced sensors exceeds 500 Ω, It is not possible to reduce the resistance value. Since the platinum material is very expensive[25], it offers the opportunity to prevent production errors thanks to these trimming points.
When the trimming points of the fabricated sensors are cut, the total resistance change is given in Fig. 12. When temperature sensor is produced on a 6 inch substrate, the initial resistance of the sensors is between 4.73 Ω and 4.97 Ω. The desired resistance value in this study is 500 ohms. Therefore, the platinum layer thickness is about 1 µm. Crystallization level of the platinum layer, structural defects, voids may be found in the layer. While making resistance correction with trimming points, the resistance values may deviate slightly from the target resistance value. By using trimming points for the temperature sensors of two different thicknesses, the resistance values were adjusted to approximately 500 Ω. It is understood from figure 13 that the temperature resistance gradients of the two sensors are different from each other. One of the reasons for this is the small differences that occur when adjusting the initial resistor value. Another reason is that the TCR values are different.
The temperature dependent resistance and TCR values of two different thickness sensors are shown in Figure 13 (a) and (b). By increasing temperature, the change in resistance per temperature of the sensor with a platinum thickness of 1 micrometer increases more. There are two reasons. The first is the resistivity and the second is the temperature coefficient of resistance (TCR). The resistivity value was calculated from equation 2. Here, it is clearly seen that the RS value is a function of the thickness. Therefore, the temperature showed its effect on the resistance curve at high temperatures. Naturally, the Rs value is also responsible for the change in the TCR value. The temperature coefficient of resistance can be defined as the rate of change of resistance per degree change in the temperature from a substance’s original temperature. However, the TCR value is affected by many variables. Pattern quality, sputter condition, annealing conditions, resistivity can be given as examples of these variables. TCR values were found to be around 3.84x103 at 0 0C. This value is very close to the standard TCR value (3.85x103 ) [26] used for industrial applications [27]. As a result, the variation of the temperature sensor with temperature for both sensors is quite linear. Therefore, both sensors can be used for industrial applications.
In this study, a temperature sensor was fabricated with Lift-Off Lithography method. A sputter was used to deposit the platinum layer. According to the XRD diffraction results of the annealed platinum layer at different temperatures, it was observed that the grain size increased. The micro strain and lattice parameter (a0) decreased with annealing. It was observed that these structural changes caused a decrease in resistivity and sheet resistance values. When one wants to produce a temperature sensor with a certain resistance value on a 6 inch substrate, fabricated each sensor by smart mask design are demonstrated in minimizing errors caused by non-uniformity from the platinum layer. In addition, since platinum is a rare and expensive material, this mask will be an example for industrial establishments in particular. Thanks to the smart mask design, the desired target resistance can be adjusted with the trimming points on the temperature sensors whose production has been completed. Even if one cannot produce the resistance value, one can adjust these sensors according to their sensitivity and use them wherever one wants. As a result, it has been shown that the temperature resistance change is linear even when the trimming points are cut. In conclusion smart mask design has advantage for mass production produce RTD. It will be used very frequently the future microelectronic applications.
Acknowledgements: Authors would like to thank Assist.Prof.Dr. Erhan Budak from Bolu Abant Izzet Baysal University. Department of Chemistry - for his valuable advices. Authors would like to thank, Umutcan Gürer and Ozan Yılmaz in the Institute of graduate studies at Bolu Abant Izzet Baysal University-- for their help during experimental studies.
Funding: This work is supported by the Scientific and Technological Research Council of Turkey (TUBITAK) and Arcelik A.S under TEYDEB 1505 (Grant Number: 5170059) - Technology and Innovation Support Programs Presidency and also supported in part by the Presidency of Turkey, Presidency of Strategy and Budget under Contract Number: 2016K12- 2834.
Conflicts of interest: The authors have no conflicts of interest to declare that are relevant to the content of this article
Availability of data and materials: The data presented here are available upon request ([email protected])
Code availability: Not applicable
Ethics approval: Not applicable
Consent to participate: Not applicable.
Consent for publication: Not applicable.
- Hussain A, Subasi O, Lazoglu I (2020) A novel sensor using photo-interrupter for measuring static friction coefficient. J Mech Sci Technol 34:2333–2339. https://doi.org/10.1007/s12206-020-0509-5
- Meister S, Grundhöfer L, Stüve J, Groves RM (2021) Imaging sensor data modelling and evaluation based on optical composite characteristics: Investigation of data quality for inline inspection. Int J Adv Manuf Technol 116:3965–3990. https://doi.org/10.1007/s00170-021-07591-5
- Hussain AM, Hussain MM (2016) CMOS-Technology-Enabled Flexible and Stretchable Electronics for Internet of Everything Applications. Adv Mater. https://doi.org/10.1002/adma.201504236
- Tiggelaar RM, Sanders RGP, Groenland AW, Gardeniers JGE (2009) Stability of thin platinum films implemented in high-temperature microdevices. Sensors Actuators A Phys. https://doi.org/10.1016/j.sna.2009.03.017
- Sandborn P, Sandborn PA (2008) A random trimming approach for obtaining high-precision embedded resistors. IEEE Trans Adv Packag. https://doi.org/10.1109/TADVP.2007.908030
- Kim J, Kim J, Shin Y, Yoon Y (2001) A Study on the Fabrication of an RTD (Resistance Temperature Detector) by Using Pt Thin Film. Korean J Chem Eng. https://doi.org/10.1007/BF02707199
- Qian Y, Luo Z, Liu Z et al (2014) Application of RTD sensor in the real time measurement and wireless transmission. In: Proceedings - 2014 4th International Conference on Instrumentation and Measurement, Computer, Communication and Control, IMCCC 2014
- Mansoor M, Haneef I, Akhtar S et al (2015) Silicon diode temperature sensors - A review of applications. Sensors Actuators, A Phys
- Makinwa KAA (2010) Smart temperature sensors in standard CMOS. In: Procedia Engineering
- Courbat J, Briand D, de Rooij NF (2008) Reliability improvement of suspended platinum-based micro-heating elements. Sensors Actuators A Phys. https://doi.org/10.1016/j.sna.2007.04.006
- Dankoco MD, Tesfay GY, Benevent E, Bendahan M (2016) Temperature sensor realized by inkjet printing process on flexible substrate. Mater Sci Eng B Solid-State Mater Adv Technol. https://doi.org/10.1016/j.mseb.2015.11.003
- Turkani VS, Maddipatla D, Narakathu BB et al (2019) Nickel based RTD fabricated via additive screen printing process for flexible electronics. IEEE Access. https://doi.org/10.1109/ACCESS.2019.2904970
- Kedzierski MA (1993) Principles and methods of temperature measurement. Exp Therm Fluid Sci. https://doi.org/10.1016/0894-1777(93)90045-k
- Feteira A (2009) Negative temperature coefficient resistance (NTCR) ceramic thermistors: An industrial perspective. J. Am. Ceram. Soc
- Witt F, Vook RW (1968) Thermally induced strains in cubic metal films. J Appl Phys. https://doi.org/10.1063/1.1656671
- Witt F, Vook RW (1969) Thermally induced strains in diamond cubic, tetragonal, orthorhombic, and hexagonal films. J Appl Phys. https://doi.org/10.1063/1.1657457
- Nasiri-Tabrizi B (2014) Thermal treatment effect on structural features of mechano-synthesized fluorapatite-titania nanocomposite: A comparative study. J Adv Ceram. https://doi.org/10.1007/s40145-014-0090-4
- Moret MP, Devillers MAC, Tichelaar FD et al (2003) Damage after annealing and aging at room temperature of platinized silicon substrates. Thin Solid Films. https://doi.org/10.1016/S0040-6090(03)00453-X
- Sreemany M, Sen S (2006) Effect of substrate temperature and annealing temperature on the structural, electrical and microstructural properties of thin Pt films by rf magnetron sputtering. Appl Surf Sci. https://doi.org/10.1016/j.apsusc.2006.05.040
- Ehrlich A, Weiß U, Hoyer W, Geßner T (1997) Microstructural changes of Pt/Ti bilayer during annealing in different atmospheres - An XRD study. Thin Solid Films. https://doi.org/10.1016/S0040-6090(96)09452-7
- Inokuma T, Taketa Y, Haradome M (1987) The Relationship Between the Electrical Properties of Thick Film Resistors and the Thermal Expansion Coefficient of the Substrates. Act Passiv Electron Components. https://doi.org/10.1155/1987/65302
- Solliard C, Flueli M (1985) Surface stress and size effect on the lattice parameter in small particles of gold and platinum. Surf Sci. https://doi.org/10.1016/0039-6028(85)90610-7
- Sandborn PAM, Sandborn PA (2007) Using embedded resistor emulation and trimming to demonstrate measurement methods and associated engineering model development. Int J Eng Educ
- Suresh KS (2013) Electric circuit analysis
- Schuettler M, Stiess S, King BV, Suaning GJ (2005) Fabrication of implantable microelectrode arrays by laser cutting of silicone rubber and platinum foil. In: Journal of Neural Engineering
- Clayton WA (1988) Thin-Film Platinum for Appliance Temperature Control. IEEE Trans Ind Appl. https://doi.org/10.1109/28.2875
- Mailly F, Giani A, Bonnot R et al (2001) Anemometer with hot platinum thin film. Sensors Actuators A Phys. https://doi.org/10.1016/S0924-4247(01)00668-9