**Some Contributions at the Technology of Electrochemical Micromachining with Ultra Short Voltage Pulses**

Richard Zemann, Philipp Walter Reiss, Paul Schörghofer and Friedrich Bleicher *Vienna University of Technology, Institute for Production Engineering and Laser Technology, Austria* 

#### **1. Introduction**

The tendency to make progressively smaller and increasingly complex products is no longer an exclusive demand of the electronics industry. Many fields such as medicine, biomechanical technology, the automotive, and the aviation industries are searching for tools and methods to realize micro- and nanostructures in various materials. The microstructuring of very hard materials, like carbides or brittle-hard materials, pose a particularly major challenge for manufacturing technology in the near future. For these reasons the Institute for Production Engineering and Laser Technology (IFT) of the Vienna University of Technology is working in the field of electrochemical micromachining with ultra short voltage pulses (µPECM) in nanosecond duration. With the theoretical resolution of 10 nm, this technology enables high precision manufacturing. [Kock M.]. A question, which can illustrate the motivation to do this research work in this field, is: "Which parameters have to be set at a production machine and which framework conditions have to be managed to reach a desired result?" To answer this question for the materials nickel and steel (1.4301), the IFT has done experimental work.

#### **2. Electrochemical micromachining**

Basically, the term machining stands for the removal of material. Furthermore, micromachining is the production of very small scaled shapes and parts in the range of 100 µm – 0,1 µm. DIN 8580 is the classification of all manufacturing processes. Figure 1 illustrates DIN 8590 for ablation, which is a part of DIN 8580.

Ablation is a non-mechanical separation of material. It can be divided into chemical, thermal and electrochemical methods. For example water jet cutting is not yet assigned to either ablation methods or to cutting methods. Electrochemical micromachining (ECM) uses electrochemical reactions to treat a metal work piece. These reactions are for example processes in an electrolyser or a battery. In electrolysers the chemical reaction is driven by an externally applied voltage, whereas in a battery a voltage is created by a chemical reaction. As depicted in figure 1, the group of electrochemical processes are assigned to

Some Contributions at the Technology

**3.1 Method and procedure** 

of Electrochemical Micromachining with Ultra Short Voltage Pulses 5

Fig. 3. Comparison of the electrochemical micromachining methods in the field of resolution

**3. Electrochemical micromachining with ultra short voltage pulses (µPECM)** 

Electrochemical micromachining with ultra short voltage pulses was developed at the Fritz-Haber-Institute of the Max-Planck-Corporation. Furthermore this innovative method for micromachining was published for the first time in the beginning of 2000. Other universities and companies working on similar topics can be found in Germany, Poland, Korea, and Austria. Since late 2010 the Institute for Production Engineering and Laser Technology (IFT) at the Vienna University of Technology has been working with this method as well. The IFT is striving to deliver machining strategies, new material–electrolyte combinations and production parameters for the industrial applicability. The machining technology of µPECM is based on the already well-established fundamentals of common electrochemical manufacturing technologies. The major advantage of the highest manufacturing precision is derived from the extremely small working gaps that are achievable through ultra short voltage pulses. This describes the main difference to common electrochemical technologies. As previously stated general advantage of electrochemical machining technologies is that the treatment of the work piece takes place without any mechanical forces or thermal influences. Therefore, no abrasive wear of the tool occurs and aspect ratios of >100 are possible which sets the basis for extremely sharp-edged geometries. There is no

These days appropriate electrolytes have already been found for several nonferrous metals such as nickel, tungsten, gold etc., as well as alloys like non-corroding steel 1.4301. Nevertheless, a main research focus for the Institute will be the search for new materialelectrolyte combinations to expand the field of application for this technology and to enhance its manufacturing productivity. This needs to be accomplished in order to fulfil the requirements of industrial production because in industries such as the automotive sector the production rate is very important. At the Nano-/Micro-Machining-Center of the IFT, an assortment of high quality measuring devices is available. Based on the technology of µPECM and on the use of high end measuring devices, specimens and parts in the micrometer range are to be manufactured and analyzed in order to investigate material

unintentional rounding of edges and no burring on the part.

removal rates and the accuracy of resulting work piece geometries.

ablation, which is a non-cutting technology. Cutting technologies for the realization of microstructures, like high speed cutting, induce mechanical stress, and thermal technologies, like laser ablation, induce thermal stress upon the work piece. Due to the fact that electrochemical technologies have none of these disadvantages, they are of interest to many industrial cases. No stress is induced in the work piece, therefore the structure of the work piece remains unchanged. Another advantage is that there is no machining force necessary and thus it is possible to machine areas which are difficult to reach. Pulsed electrochemical micromachining (PECM) as well as electrochemical micromachining with ultra short pulses (µPECM) belong to the electrochemical micromachining methods. Figure 2 shows the voltage-current curve of metal dissolution. This curve is segmented in active dissolution, passivity and trans-passive dissolution. PECM is positioned in the trans-passive section of the curve (2) whereas µPECM is positioned in the active metal dissolution area (1). Once a voltage of εP is reached, the current slopes down rapidly. The current remains low until the end of the passive section. At further increase of the voltage the current rises again to the trans-passive section. Machines, which are working with technologies in the range of active metal dissolution are more precise but obtain lower removal rates as others working in the trans-passive range.

Fig. 1. Classification of ablation (DIN 8590)

Fig. 2. Schematic illustration of current-voltage curve for metals: The three characteristic sections are: active dissolution, passivity and trans-passive dissolution

Figure 3 shows the main differences of the electrochemical micromachining methods. The conventional ECM uses direct current as energy source. Whereas both PECM and µPECM, use pulsed energy sources, the major difference between these technologies is the pulse width. While the PECM uses pulse widths from milli- to microseconds, the electrochemical micromachining with ultra short pulses uses pulse widths from micro- to picoseconds.

For PECM the removal rate is dependent on the current density distribution. µPECM directly controls the working gap by locally charging and discharging the so called electrochemical double layers. This leads to the advantage of µPECM, that the spatial confinement of electrochemical reactions and the thereby produced resolution is very high.

### **3. Electrochemical micromachining with ultra short voltage pulses (µPECM)**

#### **3.1 Method and procedure**

4 Cutting Edge Research in New Technologies

ablation, which is a non-cutting technology. Cutting technologies for the realization of microstructures, like high speed cutting, induce mechanical stress, and thermal technologies, like laser ablation, induce thermal stress upon the work piece. Due to the fact that electrochemical technologies have none of these disadvantages, they are of interest to many industrial cases. No stress is induced in the work piece, therefore the structure of the work piece remains unchanged. Another advantage is that there is no machining force necessary and thus it is possible to machine areas which are difficult to reach. Pulsed electrochemical micromachining (PECM) as well as electrochemical micromachining with ultra short pulses (µPECM) belong to the electrochemical micromachining methods. Figure 2 shows the voltage-current curve of metal dissolution. This curve is segmented in active dissolution, passivity and trans-passive dissolution. PECM is positioned in the trans-passive section of the curve (2) whereas µPECM is positioned in the active metal dissolution area (1). Once a voltage of εP is reached, the current slopes down rapidly. The current remains low until the end of the passive section. At further increase of the voltage the current rises again to the trans-passive section. Machines, which are working with technologies in the range of active metal dissolution are more precise but obtain lower removal rates as others working

Fig. 2. Schematic illustration of current-voltage curve for metals: The three characteristic

Figure 3 shows the main differences of the electrochemical micromachining methods. The conventional ECM uses direct current as energy source. Whereas both PECM and µPECM, use pulsed energy sources, the major difference between these technologies is the pulse width. While the PECM uses pulse widths from milli- to microseconds, the electrochemical micromachining with ultra short pulses uses pulse widths from micro- to picoseconds. For PECM the removal rate is dependent on the current density distribution. µPECM directly controls the working gap by locally charging and discharging the so called electrochemical double layers. This leads to the advantage of µPECM, that the spatial confinement of electrochemical reactions and the thereby produced resolution is very high.

sections are: active dissolution, passivity and trans-passive dissolution

in the trans-passive range.

Fig. 1. Classification of ablation (DIN 8590)

Electrochemical micromachining with ultra short voltage pulses was developed at the Fritz-Haber-Institute of the Max-Planck-Corporation. Furthermore this innovative method for micromachining was published for the first time in the beginning of 2000. Other universities and companies working on similar topics can be found in Germany, Poland, Korea, and Austria. Since late 2010 the Institute for Production Engineering and Laser Technology (IFT) at the Vienna University of Technology has been working with this method as well. The IFT is striving to deliver machining strategies, new material–electrolyte combinations and production parameters for the industrial applicability. The machining technology of µPECM is based on the already well-established fundamentals of common electrochemical manufacturing technologies. The major advantage of the highest manufacturing precision is derived from the extremely small working gaps that are achievable through ultra short voltage pulses. This describes the main difference to common electrochemical technologies. As previously stated general advantage of electrochemical machining technologies is that the treatment of the work piece takes place without any mechanical forces or thermal influences. Therefore, no abrasive wear of the tool occurs and aspect ratios of >100 are possible which sets the basis for extremely sharp-edged geometries. There is no unintentional rounding of edges and no burring on the part.

These days appropriate electrolytes have already been found for several nonferrous metals such as nickel, tungsten, gold etc., as well as alloys like non-corroding steel 1.4301. Nevertheless, a main research focus for the Institute will be the search for new materialelectrolyte combinations to expand the field of application for this technology and to enhance its manufacturing productivity. This needs to be accomplished in order to fulfil the requirements of industrial production because in industries such as the automotive sector the production rate is very important. At the Nano-/Micro-Machining-Center of the IFT, an assortment of high quality measuring devices is available. Based on the technology of µPECM and on the use of high end measuring devices, specimens and parts in the micrometer range are to be manufactured and analyzed in order to investigate material removal rates and the accuracy of resulting work piece geometries.

Some Contributions at the Technology

manufacturing process labelled

electrochemical double layer is formed. [Schuster R.]

of Electrochemical Micromachining with Ultra Short Voltage Pulses 7

usage. However, a more complex machine structure would give the possibility to reach the highest precision requirement. Figure 5 shows a view inside the IFT´s machine. The whole machining process takes place in a basin filled with an electrolyte solution that has to be adequately adapted to the work piece material used. At the bottom of this electrolyte basin a hole for the connection of work piece and machine can be found. It is important that the basin is well sealed, so that no leakage can occur. The basin is made of Teflon, which has resistance against the electrolytes used in the experiments. Even when filling the basin, caution is required due to the fact that once in contact with the electrolyte, the surface of the material could begin to react. To protect the work piece surface from the influence of the electrolyte-solution, a cathodic protection-current is applied by the backing electrode which is immersed in the electrolyte. At the IFT, a tungsten wire is the preferred tool for the electrochemical micromachining with ultra short voltage pulses. With the basin filled as

needed, the process of work piece calibration can be performed.

Fig. 5. View inside the electrochemical machine with all important parts for the

The measurement process for finding the work piece surface coordinate is executed automatically by the machine. Therefore a tool potential is necessary to detect the electrical short circuit thru a contact between work piece and tool. Another possible measurement process is to match the local coordinate systems of the work piece with the global coordinate system of the machine structure. With the result of this measurement process and three positioning screws on the plate, whereon the electrolyte basin is mounted, it is now possible to get the necessary congruence between these two coordinate systems. Then the manufacturing program, which conforms to a standard CNC-program, is started. The tool moves along the pre-programmed paths and selectively ablates material due to the principle, that is based on the finite time constant for double layer charging, which varies linearly with the local separation between the electrodes. During nanosecond pulses, the electrochemical reactions are confined to electrode regions in close proximity. [Schuster R.]. To view the manufacturing process and get optical magnification, a USB–camera is used. Similar to conventional electrochemical manufacturing methods the µPECM process uses an oppositional electric voltage for the work piece and the tool. At the phase boundaries between the tool and the electrolyte and also between the work piece and the electrolyte, an

Due to the multidisciplinary nature of this technology, intensive cooperation with other institutes of the Vienna University of Technology in the fields of electro-technical engineering, high frequency technology and electrochemistry is established. The goal of this research will be to elevate this technology to an appropriate level of possible industrial usage by enhancing the manufacturing accuracy and the process efficiency for current components. Therefore a profound knowledge of material science, electrochemistry, and production technology for extremely small dimensions will be required. The necessary expertise in these fields will be provided by the cooperating institutes and interested companies.

To accomplish these improvements in the technology of electrochemical micromachining with ultra short pulses it will be necessary to merge several research projects which are currently dealing with the topics of piezo-driven nano-positioning devices and the development of high precision machine structures for different types of machines. Table 1 shows all the relevant adjustable parameters for µPECM. In addition to the proper choice of the electrical process parameters like the amplitude of the pulses, the pulse width, the voltages at the tool, and the work piece, the right choice of electrolyte is probably the most important aspect for this process.


Table 1. Adjustable parameters which have an influence on the process

In figure 4, the relevant parameters of the applied voltage pulses are illustrated. The duty cycle is the sum of the pulse width and the pause time. A pulse width of 100 ns and a pause time of 800 ns conforms a pulse–pause ratio of 1/8.

Fig. 4. Pulse-pause ratio of the applied voltage pulse, with pulse width p, length of pause, amplitude A, tool voltage T, applied pulsed voltage signal U(t)

Due to the fact that µPECM is one of the latest elaborated removal technologies, there are no fully developed machines available in the market. All the institutes and companies, which investigate these fields, work with machines in laboratory stage. The machine at the IFT is simple constructed and very easy to maintain, consequently it is adequate for industrial

Due to the multidisciplinary nature of this technology, intensive cooperation with other institutes of the Vienna University of Technology in the fields of electro-technical engineering, high frequency technology and electrochemistry is established. The goal of this research will be to elevate this technology to an appropriate level of possible industrial usage by enhancing the manufacturing accuracy and the process efficiency for current components. Therefore a profound knowledge of material science, electrochemistry, and production technology for extremely small dimensions will be required. The necessary expertise in these fields will be

To accomplish these improvements in the technology of electrochemical micromachining with ultra short pulses it will be necessary to merge several research projects which are currently dealing with the topics of piezo-driven nano-positioning devices and the development of high precision machine structures for different types of machines. Table 1 shows all the relevant adjustable parameters for µPECM. In addition to the proper choice of the electrical process parameters like the amplitude of the pulses, the pulse width, the voltages at the tool, and the work piece, the right choice of electrolyte is probably the most

**Adjustable parameters for the process abbraviations** 

In figure 4, the relevant parameters of the applied voltage pulses are illustrated. The duty cycle is the sum of the pulse width and the pause time. A pulse width of 100 ns and a pause

Fig. 4. Pulse-pause ratio of the applied voltage pulse, with pulse width p, length of pause,

Due to the fact that µPECM is one of the latest elaborated removal technologies, there are no fully developed machines available in the market. All the institutes and companies, which investigate these fields, work with machines in laboratory stage. The machine at the IFT is simple constructed and very easy to maintain, consequently it is adequate for industrial

A p T I ppr D E

provided by the cooperating institutes and interested companies.

important aspect for this process.

pulse width voltage at the tool

pulse–pause ratio diameter of the tool electrolyte solution

amplitude of the pulses

time of 800 ns conforms a pulse–pause ratio of 1/8.

amplitude A, tool voltage T, applied pulsed voltage signal U(t)

current through the backing electrode

Table 1. Adjustable parameters which have an influence on the process

usage. However, a more complex machine structure would give the possibility to reach the highest precision requirement. Figure 5 shows a view inside the IFT´s machine. The whole machining process takes place in a basin filled with an electrolyte solution that has to be adequately adapted to the work piece material used. At the bottom of this electrolyte basin a hole for the connection of work piece and machine can be found. It is important that the basin is well sealed, so that no leakage can occur. The basin is made of Teflon, which has resistance against the electrolytes used in the experiments. Even when filling the basin, caution is required due to the fact that once in contact with the electrolyte, the surface of the material could begin to react. To protect the work piece surface from the influence of the electrolyte-solution, a cathodic protection-current is applied by the backing electrode which is immersed in the electrolyte. At the IFT, a tungsten wire is the preferred tool for the electrochemical micromachining with ultra short voltage pulses. With the basin filled as needed, the process of work piece calibration can be performed.

Fig. 5. View inside the electrochemical machine with all important parts for the manufacturing process labelled

The measurement process for finding the work piece surface coordinate is executed automatically by the machine. Therefore a tool potential is necessary to detect the electrical short circuit thru a contact between work piece and tool. Another possible measurement process is to match the local coordinate systems of the work piece with the global coordinate system of the machine structure. With the result of this measurement process and three positioning screws on the plate, whereon the electrolyte basin is mounted, it is now possible to get the necessary congruence between these two coordinate systems. Then the manufacturing program, which conforms to a standard CNC-program, is started. The tool moves along the pre-programmed paths and selectively ablates material due to the principle, that is based on the finite time constant for double layer charging, which varies linearly with the local separation between the electrodes. During nanosecond pulses, the electrochemical reactions are confined to electrode regions in close proximity. [Schuster R.]. To view the manufacturing process and get optical magnification, a USB–camera is used.

Similar to conventional electrochemical manufacturing methods the µPECM process uses an oppositional electric voltage for the work piece and the tool. At the phase boundaries between the tool and the electrolyte and also between the work piece and the electrolyte, an electrochemical double layer is formed. [Schuster R.]

Some Contributions at the Technology

sufficiently recharged. [Zemann R.]

double layer (right)

of Electrochemical Micromachining with Ultra Short Voltage Pulses 9

the metal surface. If the voltage pulse width is very short, the erosion takes place very closely to the tool (Rshort), since the ohmic resistance of the electrolyte prevents ablation at areas further away from the tool (Rlong) due to the double layer capacitor not being able to be

The right illustration in figure 8 shows schematically the two different charging curves of the double layers at the work piece for Rshort and Rlong. At smaller distances between the tool and the work piece, the charging curve is steeper; this leads to the formulas (1) and (2).

Fig. 8. Applied voltage pulse (left) and time variable voltage curve in the electrochemical

τ time constant for double capacitor charging

U(t) applied voltage with dependence on time τ time constant for double capacitor charging

accuracy is regarded as a principal target.

CDL capacitance of the electrochemical double layer

UDL charging voltage of the electrochemical double layer

Relektrolyte resistance of the electrolyte

R • C electrolyte DL

( t / ) U U t • 1 e DL

Another important influence on the charge of the double layers has the pulse width and the choice of the electrolyte. Small working gaps between the tool and the work piece of less than 1 µm are produced with pulse widths of less than 100 nanoseconds and lead to a very high resolution of the machined structure. Even more accurate machining can be achieved with pulse widths of less than 1 nanosecond and by separating the processing pulse into a pre-pulse and a main pulse, which is a future research topic for the IFT. In order to elaborate on the research work concerning the technology of using ultra short voltage pulses, the relevant demands of industry, basically increasing the material removal rate, has to be considered as a main goal. Subsequently, an increase in the already high machining

Another major advantage of this technology is the possibility to reverse the process electrically. This means that not only the work piece can be machined, but also the tool itself can be defined as the work piece and be machined to its ideal geometry without any further set-up. Regarding all these functionalities, the requirements for precise micromachining are

 

(1)

(2)

Figure 6 shows the detailed structure of the double layer. The double layer consists of a rigid, outer Helmholtz layer (OHL) and a diffuse area. The inner Helmholtz layer (IHL) is a part of the OHL. In the diffuse area the hydrated metal ions are versatile. The functionality of the OHL can be understood basically as a kind of a plate capacitor, with a plate separation of half of the atom radius. [Hamann C.H.]

Fig. 6. Simplified Stern-Graham-Model of the electrochemical double layer [Hamann C.H.]

Fig. 7. Schematic illustration of the electrochemical double layers as capacitors and the electrolyte as electrical resistor between tool and work piece (left) and the equivalent circuit diagram (right) with U(t) as energy source, CDL as capacitance of the double layers and Relectrolyte as the ohmic resistor of the electrolyte.

The left section of figure 7 shows the schematic illustration of the tool, the work piece in the electrolyte basin, and the electrochemical double layers illustrated as plate capacitors. The electrolyte has comparable characteristics to a linear ohmic resistor with a value that is dependent on the length of the current path. The length of the current path is equal to the distance between the tool and the work piece. The right section of figure 7 shows the equivalent circuit diagram in a simplified version of the left illustration in figure 7. Through charging and discharging the electrochemical double layer, metal ions are solvated out of

Figure 6 shows the detailed structure of the double layer. The double layer consists of a rigid, outer Helmholtz layer (OHL) and a diffuse area. The inner Helmholtz layer (IHL) is a part of the OHL. In the diffuse area the hydrated metal ions are versatile. The functionality of the OHL can be understood basically as a kind of a plate capacitor, with a plate

Fig. 6. Simplified Stern-Graham-Model of the electrochemical double layer [Hamann C.H.]

Fig. 7. Schematic illustration of the electrochemical double layers as capacitors and the electrolyte as electrical resistor between tool and work piece (left) and the equivalent circuit diagram (right) with U(t) as energy source, CDL as capacitance of the double layers and

The left section of figure 7 shows the schematic illustration of the tool, the work piece in the electrolyte basin, and the electrochemical double layers illustrated as plate capacitors. The electrolyte has comparable characteristics to a linear ohmic resistor with a value that is dependent on the length of the current path. The length of the current path is equal to the distance between the tool and the work piece. The right section of figure 7 shows the equivalent circuit diagram in a simplified version of the left illustration in figure 7. Through charging and discharging the electrochemical double layer, metal ions are solvated out of

Relectrolyte as the ohmic resistor of the electrolyte.

separation of half of the atom radius. [Hamann C.H.]

the metal surface. If the voltage pulse width is very short, the erosion takes place very closely to the tool (Rshort), since the ohmic resistance of the electrolyte prevents ablation at areas further away from the tool (Rlong) due to the double layer capacitor not being able to be sufficiently recharged. [Zemann R.]

The right illustration in figure 8 shows schematically the two different charging curves of the double layers at the work piece for Rshort and Rlong. At smaller distances between the tool and the work piece, the charging curve is steeper; this leads to the formulas (1) and (2).

Fig. 8. Applied voltage pulse (left) and time variable voltage curve in the electrochemical double layer (right)

$$
\pi = \mathsf{R}\_{\text{electrolyte}} \bullet \mathsf{C}\_{\text{DL}} \tag{1}
$$

τ time constant for double capacitor charging Relektrolyte resistance of the electrolyte

CDL capacitance of the electrochemical double layer

$$\mathbf{U}\_{\rm DL} = \mathbf{U}(\mathbf{t}) \mathbf{\hat{}} \left(\mathbf{1} - \mathbf{e}^{\left(-\mathbf{t} \;/\tau\right)}\right) \tag{2}$$

UDL charging voltage of the electrochemical double layer

U(t) applied voltage with dependence on time

τ time constant for double capacitor charging

Another important influence on the charge of the double layers has the pulse width and the choice of the electrolyte. Small working gaps between the tool and the work piece of less than 1 µm are produced with pulse widths of less than 100 nanoseconds and lead to a very high resolution of the machined structure. Even more accurate machining can be achieved with pulse widths of less than 1 nanosecond and by separating the processing pulse into a pre-pulse and a main pulse, which is a future research topic for the IFT. In order to elaborate on the research work concerning the technology of using ultra short voltage pulses, the relevant demands of industry, basically increasing the material removal rate, has to be considered as a main goal. Subsequently, an increase in the already high machining accuracy is regarded as a principal target.

Another major advantage of this technology is the possibility to reverse the process electrically. This means that not only the work piece can be machined, but also the tool itself can be defined as the work piece and be machined to its ideal geometry without any further set-up. Regarding all these functionalities, the requirements for precise micromachining are

Some Contributions at the Technology

**3.2 Tooling** 

of Electrochemical Micromachining with Ultra Short Voltage Pulses 11

The favoured material used for the tool is tungsten. Tungsten can be easily treated with NaOH as electrolyte and has preferable mechanical properties like a Mohs hardness of 7,5 and a Young´s modulus of 410 GPa. For the experimental work wires with a diameter of 75 and 150 µm were used. The first tooling step is, to cut the tungsten wire manually to a length of 15 – 20 mm. The wire is fixed with a collet in the toolholder and should protrude far enough to produce the necessary geometries, mostly that is about 4 – 5 mm. The toolholder has to be protected from the acid to prevent corrosion, which is performed by a layer of Lacomit. It is a dark red fluid, once hardened it isolates the toolholder against the electrolyte. This red fluid functions as a barrier between the electrolyte and the toolholder. Only the top of the upper part of the tungsten wire is free of Lacomit to treat the work piece.

Figure 9 shows two toolholders with the different diameters of tool wire.

Fig. 9. Tools ready for manufacturing. The left tool has a diameter of 75 µm and the right

Fig. 10. Tungsten wire with a diameter of 150 µm, untreated with the end after manual

cutting (left) and the finished end after electrochemical flattening (right).

As mentioned before the tool/wire is cut off manually. Due to the mechanical characteristics of tungsten it is possible that the cut end splits. If that happens the split section and the

tool a diameter of 150 µm, both with Lacomit layer.

usual cut end of the tool (figure 10, left) has to be removed.

met. Possible tasks that can be performed with this machining centre include: tooling, milling, turning, sinking, and measuring.

Characteristics of the µPECM process with ultra short voltage pulses:


Table 2 shows that electrochemical micromachining with ultra short voltage pulses has several advantages compared to other nano- and micromachining technologies. For example the theoretical dissolution range and the aspect ratio are outstanding, whereas in case of the removal rate, µPECM is not competitive against technologies like high speed cutting. For material removal, µPECM is mainly used for post-processing and for producing surfaces with hydrophobic and hydrophilic characteristics at the moment.


LIGA is the acronym for lithography (LI), electroforming (G) and molding (A) FIB focussed ion beam milling EDM electric discharge machining

Table 2. Comparison of nano- and micromachining methods [Kock M.]

#### **3.2 Tooling**

10 Cutting Edge Research in New Technologies

met. Possible tasks that can be performed with this machining centre include: tooling,

Table 2 shows that electrochemical micromachining with ultra short voltage pulses has several advantages compared to other nano- and micromachining technologies. For example the theoretical dissolution range and the aspect ratio are outstanding, whereas in case of the removal rate, µPECM is not competitive against technologies like high speed cutting. For material removal, µPECM is mainly used for post-processing and for producing surfaces

**treatable** 

active materials

etch-able, evaporable materials

galvanic removable materials

dielectrics

materials

**materials category removal** 

electrochem. micromachining

chemical

mechanical/ thermal method

thermal

thermal

polymers cutting method \*\*\*

method \*\*

method \*\*

method \*\*

method \*\*

**rate** 

\*

\*\*

High aspect-ratio >100 (only limited thru the young's modulus of the material)

Characteristics of the µPECM process with ultra short voltage pulses:

Small working gaps between tool and work piece (< 1 µm)

with hydrophobic and hydrophilic characteristics at the moment.

**µPECM** limit: 10 nm > 100 electrochem.

**Laser ablation** ~ µm ~ 1 metals and

**cutting** ~ µm ~1 metals and

**FIB** ~ 30 nm ~ 10 conducting

LIGA is the acronym for lithography (LI), electroforming (G) and molding (A)

Table 2. Comparison of nano- and micromachining methods [Kock M.]

**EDM** ~ µm ~ 10 metals thermal

**aspect ratio** 

milling, turning, sinking, and measuring.

No mechanical process forces

 Manufacturing of hard materials Very small edge-rounding

High quality measuring function

No thermal load

No tool wear

No burring

**high speed** 

FIB focussed ion beam milling EDM electric discharge machining

High precision (theoretical resolution of 10 nm)

Adjustable roughness of the work piece surface

**theoretical dissolution range** 

**Lithography** >10 nm ~ 1

**LIGA** ~ 100 nm ~100

The favoured material used for the tool is tungsten. Tungsten can be easily treated with NaOH as electrolyte and has preferable mechanical properties like a Mohs hardness of 7,5 and a Young´s modulus of 410 GPa. For the experimental work wires with a diameter of 75 and 150 µm were used. The first tooling step is, to cut the tungsten wire manually to a length of 15 – 20 mm. The wire is fixed with a collet in the toolholder and should protrude far enough to produce the necessary geometries, mostly that is about 4 – 5 mm. The toolholder has to be protected from the acid to prevent corrosion, which is performed by a layer of Lacomit. It is a dark red fluid, once hardened it isolates the toolholder against the electrolyte. This red fluid functions as a barrier between the electrolyte and the toolholder. Only the top of the upper part of the tungsten wire is free of Lacomit to treat the work piece. Figure 9 shows two toolholders with the different diameters of tool wire.

Fig. 9. Tools ready for manufacturing. The left tool has a diameter of 75 µm and the right tool a diameter of 150 µm, both with Lacomit layer.

As mentioned before the tool/wire is cut off manually. Due to the mechanical characteristics of tungsten it is possible that the cut end splits. If that happens the split section and the usual cut end of the tool (figure 10, left) has to be removed.

Fig. 10. Tungsten wire with a diameter of 150 µm, untreated with the end after manual cutting (left) and the finished end after electrochemical flattening (right).

Some Contributions at the Technology

D tool diameter in µm a working gap in µm

working gap can be calculated via formula (3).

B measured width of the groove in µm

Fig. 12. Sketch of the produced groove

of Electrochemical Micromachining with Ultra Short Voltage Pulses 13

3000 mV to 2100 mV in 300 mV steps. After measuring the width of every groove, the

The diagram in figure 13 shows that a smaller pulse width reduces the working gap. The optical estimation shows that grooves made with lower pulse widths have much better optical qualities (figure 13, left). This outcome can be explained by the localization of the manufacturing reactions. Smaller voltage pulses lead to a spatial confinement of the electrochemical reactions so that the working gap shrinks and the geometry gets more precise which is confirmed in figure 13, right. As a consequence, the pulse width is the most important parameter for the machining precision. Dependent on the machine, the minimal pulse width of p = 80 ns is further used in the experiments to produce grooves in high quality. The adjusted electrochemical parameters for this experiment are indicated in table 3.

Fig. 13. Illustration of grooves (left) - from top downwards different pulse widths were used.

Diagram of the appurtenant working gaps over pulse widths (right).

D 2a B (3)

The flattening process is performed directly in the µPECM machine. Due to the fact that the spatial resolution and pulse width are linearly related: the higher the pulse width, the higher the spatial resolution [Kock M.], the flattening process is split into two parts to produce a tool with high quality. Another advantage of this sequential machining is that the machining time is reduced. At first a large pulse width (i.e. 400 ns) is used to increase the removal speed of the cut end. Afterwards a smaller pulse width (i.e. 80 ns) is used to create a sharp edged tool with a glossy surface. Only with such tools it is possible to produce geometries with sharp edges on the work piece. Figure 11 illustrates the difference of the radius on the tool´s top for small and large pulse widths.

Fig. 11. Influence of the pulse width on the radius on the top of the tool

#### **3.3 Manufacturing of nickel**

Nickel is a hard (Mohs hardness: 3,8) and ductile metal with a silvery-white and slightly golden shine. Nickel is apart from chrome and molybdenum an important element for the refinement of steel. The ferromagnetic metal is corrosion-resistant. Nickels protective oxide surface resists most acids and alkalis. The corrosion-resistance is one of the most important characteristics of parts in laboratory environments or health care, therefore nickel is the common material in those branches. For the electrochemical manufacturing of nickel the electrolyte hydrochloride acid (HCl) is used. HCl deactivates the passive surface of nickel and renders the material processable. The following experiments were done to find the optimal processing parameters for the manufacturing of products and special surfaces made of nickel. To evaluate the outcome of the experiments, the produced structures were measured with a high-end optical measuring device. Also optical considerations through a light microscope helped to evaluate the following characteristics of the produced surfaces:


#### **3.3.1 Pulse width (p) and amplitude (A)**

In the first experiment the pulse width and the amplitude of the pulse were varied in order to see which effects the adjustment of these parameters cause. The experimental setup is a block with five parallel grooves. Every groove is made with different pulse widths from 400 ns to 80 ns. A sketch of the groove geometry is illustrated in figure 12. Overall four of these blocks with different amplitudes were manufactured. The range of the amplitudes was from 3000 mV to 2100 mV in 300 mV steps. After measuring the width of every groove, the working gap can be calculated via formula (3).

$$\mathbf{D} + \mathbf{2a} = \mathbf{B} \tag{3}$$

D tool diameter in µm

12 Cutting Edge Research in New Technologies

The flattening process is performed directly in the µPECM machine. Due to the fact that the spatial resolution and pulse width are linearly related: the higher the pulse width, the higher the spatial resolution [Kock M.], the flattening process is split into two parts to produce a tool with high quality. Another advantage of this sequential machining is that the machining time is reduced. At first a large pulse width (i.e. 400 ns) is used to increase the removal speed of the cut end. Afterwards a smaller pulse width (i.e. 80 ns) is used to create a sharp edged tool with a glossy surface. Only with such tools it is possible to produce geometries with sharp edges on the work piece. Figure 11 illustrates the difference of the

radius on the tool´s top for small and large pulse widths.

**3.3 Manufacturing of nickel** 

shape / geometry

 shine of the surface edge rounding

topology (smoothness of the bottom surface)

**3.3.1 Pulse width (p) and amplitude (A)** 

Fig. 11. Influence of the pulse width on the radius on the top of the tool

Nickel is a hard (Mohs hardness: 3,8) and ductile metal with a silvery-white and slightly golden shine. Nickel is apart from chrome and molybdenum an important element for the refinement of steel. The ferromagnetic metal is corrosion-resistant. Nickels protective oxide surface resists most acids and alkalis. The corrosion-resistance is one of the most important characteristics of parts in laboratory environments or health care, therefore nickel is the common material in those branches. For the electrochemical manufacturing of nickel the electrolyte hydrochloride acid (HCl) is used. HCl deactivates the passive surface of nickel and renders the material processable. The following experiments were done to find the optimal processing parameters for the manufacturing of products and special surfaces made of nickel. To evaluate the outcome of the experiments, the produced structures were measured with a high-end optical measuring device. Also optical considerations through a light microscope helped to evaluate the following characteristics of the produced surfaces:

In the first experiment the pulse width and the amplitude of the pulse were varied in order to see which effects the adjustment of these parameters cause. The experimental setup is a block with five parallel grooves. Every groove is made with different pulse widths from 400 ns to 80 ns. A sketch of the groove geometry is illustrated in figure 12. Overall four of these blocks with different amplitudes were manufactured. The range of the amplitudes was from


Fig. 12. Sketch of the produced groove

The diagram in figure 13 shows that a smaller pulse width reduces the working gap. The optical estimation shows that grooves made with lower pulse widths have much better optical qualities (figure 13, left). This outcome can be explained by the localization of the manufacturing reactions. Smaller voltage pulses lead to a spatial confinement of the electrochemical reactions so that the working gap shrinks and the geometry gets more precise which is confirmed in figure 13, right. As a consequence, the pulse width is the most important parameter for the machining precision. Dependent on the machine, the minimal pulse width of p = 80 ns is further used in the experiments to produce grooves in high quality. The adjusted electrochemical parameters for this experiment are indicated in table 3.

Fig. 13. Illustration of grooves (left) - from top downwards different pulse widths were used. Diagram of the appurtenant working gaps over pulse widths (right).

Some Contributions at the Technology

of Electrochemical Micromachining with Ultra Short Voltage Pulses 15

Fig. 15. Image of grooves made with 0,5M HCl and 1M HCl for A = 2400 mV (left). Working

A = varied p = varied T = 200 mV E = varied

To investigate the influence of the current through the backing electrode, the current was varied between 500 µA and 4000 µA. The results in figure 16 (left) show an increased processing time at higher currents. The minimal working gaps are in the range of 2000 to 3000 µA, as illustrated in figure 16 (right). Because of the optical criteria and the working gap a current of I = 2000 µA was used for further experiments. The illustration in figure 17 shows the difference between a high-quality and a low-quality groove. The electrochemical

Fig. 16. Processing time at different currents (left) and working gap at different currents (right)

gap over pulse width at different electrolyte concentrations for A = 3000 mV (right)

I = 1000 µA ppr = 1/8 D = 75 µm

Table 5. Adjustments for the experiment of figure 15

**3.3.3 Current through the backing electrode (I)** 

parameters for this experiment are shown in table 6.


Table 3. Adjustments for the experiment of figure 13

Figure 14 shows that similar to the pulse width the reduction of the amplitude causes a reduction of the working gap. At a pulse width of p = 80 ns an amplitude of less than 3000 mV does not lead to a removal of material, due to the fact that the double layers cannot be sufficiently charged with the provided energy. Equally the provided energy of 2400 mV amplitude and 100 ns pulse width is not sufficiently for production. The overview of the production parameters for these experiments is mentioned in table 4.

Fig. 14.Working gaps over amplitude at different pulse widths.


Table 4. Adjustments for the experiment of figure 14

#### **3.3.2 Electrolyte-concentration**

The concentration of the electrolyte is a very important parameter for the electrochemical processing. In the equivalent circuit diagram of the electrochemical cell, the electrolyte is equal to an ohmic resistor. For this experiment hydrochloric acid (HCl) in three different concentrations was used to explore the correlation between the electrolyte-concentration and the working gap. The diagram in figure 15 shows that the reduction of the electrolyte concentration leads to smaller working gaps. This outcome can be explained by the reduced conductivity of the electrolyte and the following localization of the reactions.

A reduction of the concentration increases the resistance because of the lack of ions in the aqueous solution. In such solutions ions are the charge carriers and therefore responsible for the electric conductivity. The illustration in figure 15 shows the optical differences of changed electrolyte concentrations. The processing parameters for this experiment are indicated in table 5.

A = 3000 mV p = varied T = 200 mV E = 1M HCl

Figure 14 shows that similar to the pulse width the reduction of the amplitude causes a reduction of the working gap. At a pulse width of p = 80 ns an amplitude of less than 3000 mV does not lead to a removal of material, due to the fact that the double layers cannot be sufficiently charged with the provided energy. Equally the provided energy of 2400 mV amplitude and 100 ns pulse width is not sufficiently for production. The overview of the

A = varied p = varied T = 200 mV E = 0,2M HCl

The concentration of the electrolyte is a very important parameter for the electrochemical processing. In the equivalent circuit diagram of the electrochemical cell, the electrolyte is equal to an ohmic resistor. For this experiment hydrochloric acid (HCl) in three different concentrations was used to explore the correlation between the electrolyte-concentration and the working gap. The diagram in figure 15 shows that the reduction of the electrolyte concentration leads to smaller working gaps. This outcome can be explained by the reduced

A reduction of the concentration increases the resistance because of the lack of ions in the aqueous solution. In such solutions ions are the charge carriers and therefore responsible for the electric conductivity. The illustration in figure 15 shows the optical differences of changed electrolyte concentrations. The processing parameters for this experiment are

I = 1000 µA ppr = 1/8 D = 150 µm

production parameters for these experiments is mentioned in table 4.

Fig. 14.Working gaps over amplitude at different pulse widths.

I = 1000 µA ppr = 1/8 D = 75 µm

conductivity of the electrolyte and the following localization of the reactions.

Table 4. Adjustments for the experiment of figure 14

**3.3.2 Electrolyte-concentration** 

indicated in table 5.

Table 3. Adjustments for the experiment of figure 13

Fig. 15. Image of grooves made with 0,5M HCl and 1M HCl for A = 2400 mV (left). Working gap over pulse width at different electrolyte concentrations for A = 3000 mV (right)


Table 5. Adjustments for the experiment of figure 15

#### **3.3.3 Current through the backing electrode (I)**

To investigate the influence of the current through the backing electrode, the current was varied between 500 µA and 4000 µA. The results in figure 16 (left) show an increased processing time at higher currents. The minimal working gaps are in the range of 2000 to 3000 µA, as illustrated in figure 16 (right). Because of the optical criteria and the working gap a current of I = 2000 µA was used for further experiments. The illustration in figure 17 shows the difference between a high-quality and a low-quality groove. The electrochemical parameters for this experiment are shown in table 6.

Fig. 16. Processing time at different currents (left) and working gap at different currents (right)

Some Contributions at the Technology

**3.3.5 Pulse-pause ratio** 

recommended.

of Electrochemical Micromachining with Ultra Short Voltage Pulses 17

A = 3000 mV p = 80 ns T = varied E = 0,5M HCl

The pulse-pause ratio is an important parameter that influences the electrochemical reactions. To ensure a precise and fast dissolution of the material, the ratio of pulse time to pause time should be correctly chosen. Every single pulse that charges the electrochemical double layer dissolves a monolayer of atoms from the material into the electrolyte solution. Due to the fact, that one monolayer of atoms is a very small amount of material the pulses must be applied with very high frequency to solvate the material in a reasonable rate. If the ratio is too high, the process time is unnecessarily lengthened as these rates obey an exponential law (Butler-Volmer equation). To find an appropriate pulse-pause ratio, five grooves with a different ppr-parameter were produced. Figure 19 shows a decreased removal rate at higher pulse-pause ratios for the drilling and milling processes. All of these grooves have the same working gap with negligible deviations in the range of maximal 5 µm. There is great potential to speed up the process by reducing the pulse-pause ratio without losing much precision. The used parameters for the experiment are specified in table 8. Considering the optical estimations, a pulse-pause ratio between 1/6 and 1/8 is

A = 3000 mV p = 80 ns T = 100 mV E = 0,5M HCl

In this experiment the maximum possible drilling depth should be found. The drilling process works without any problems to a depth of 140 µm. All over the removal speed slows down slightly. At a depth of 140 µm the drilling speed slows down rapidly and the experiment has to be stopped. An explanation is that in this depth the exchange of

I = 2000 µA ppr = 1/8 D = 75 µm

I = 2000 µA ppr = 1/8 D = 75 µm

Table 7. Adjustments for the experiment of figure 18

Fig. 19. Removal rate over pulse-pause ratio

**3.3.6 Drilling with µPECM** 

Table 8. Adjustments for the experiment of figure 19

Fig. 17. Image of grooves with I = 2000 µA (above) and I = 500 µA (below)


Table 6. Adjustments for the experiment of figure 17.

#### **3.3.4 Tool voltage (T)**

For successful application of ultra short voltage pulses for electrochemical machining, the electrochemical conditions, e.g. the average electric potentials of the tool (T) and the work piece have to be precisely controlled. These potentials are independently adjusted by a lowfrequency bipotentiostat and a platinium backing electrode. [Kock M.]

To investigate the influence of T, seven grooves with different tool voltages were produced. The production parameters for this manufacturing are indicated in table 7. After the measurement and evaluation of the working gap via formula (3), the results show that between -100 mV and + 100 mV the working gap reaches a minimum (figure 18, left). The optical appearance of these grooves has also the highest quality (figure 18, right). Another advantage is that the processing time decreases with lower tool voltages. For the further experimental work a tool voltage of +100 mV was used.

Fig. 18. Working gap at different tool voltages (left), image of grooves with T = 600 mV, 0 mV and -600 mV


Table 7. Adjustments for the experiment of figure 18

#### **3.3.5 Pulse-pause ratio**

16 Cutting Edge Research in New Technologies

For successful application of ultra short voltage pulses for electrochemical machining, the electrochemical conditions, e.g. the average electric potentials of the tool (T) and the work piece have to be precisely controlled. These potentials are independently adjusted by a low-

To investigate the influence of T, seven grooves with different tool voltages were produced. The production parameters for this manufacturing are indicated in table 7. After the measurement and evaluation of the working gap via formula (3), the results show that between -100 mV and + 100 mV the working gap reaches a minimum (figure 18, left). The optical appearance of these grooves has also the highest quality (figure 18, right). Another advantage is that the processing time decreases with lower tool voltages. For the further

Fig. 18. Working gap at different tool voltages (left), image of grooves with T = 600 mV, 0

E = 0,5M HCl

Fig. 17. Image of grooves with I = 2000 µA (above) and I = 500 µA (below)

A = 3000 mV p = 80 ns T = 200 mV

I = varied ppr = 1/8 D = 75 µm

frequency bipotentiostat and a platinium backing electrode. [Kock M.]

Table 6. Adjustments for the experiment of figure 17.

experimental work a tool voltage of +100 mV was used.

**3.3.4 Tool voltage (T)** 

mV and -600 mV

The pulse-pause ratio is an important parameter that influences the electrochemical reactions. To ensure a precise and fast dissolution of the material, the ratio of pulse time to pause time should be correctly chosen. Every single pulse that charges the electrochemical double layer dissolves a monolayer of atoms from the material into the electrolyte solution. Due to the fact, that one monolayer of atoms is a very small amount of material the pulses must be applied with very high frequency to solvate the material in a reasonable rate. If the ratio is too high, the process time is unnecessarily lengthened as these rates obey an exponential law (Butler-Volmer equation). To find an appropriate pulse-pause ratio, five grooves with a different ppr-parameter were produced. Figure 19 shows a decreased removal rate at higher pulse-pause ratios for the drilling and milling processes. All of these grooves have the same working gap with negligible deviations in the range of maximal 5 µm. There is great potential to speed up the process by reducing the pulse-pause ratio without losing much precision. The used parameters for the experiment are specified in table 8. Considering the optical estimations, a pulse-pause ratio between 1/6 and 1/8 is recommended.

Fig. 19. Removal rate over pulse-pause ratio



#### **3.3.6 Drilling with µPECM**

In this experiment the maximum possible drilling depth should be found. The drilling process works without any problems to a depth of 140 µm. All over the removal speed slows down slightly. At a depth of 140 µm the drilling speed slows down rapidly and the experiment has to be stopped. An explanation is that in this depth the exchange of

Some Contributions at the Technology

Fig. 21. Averaged groove depth over dwelling time

(SEM) at different resolutions

of Electrochemical Micromachining with Ultra Short Voltage Pulses 19

Fig. 22. Images of the microstructure, photographed with a scanning electron microscope

electrolyte is not sufficient, so the dissolved metal ions saturate the electrolyte in the drilled hole and prevent any further metal dissolution. This can be disabled by an alternately up and down movement of the tool to realize a kind of flushing (pulsed mechanical movement). In figure 20 the removal speed over drilling depth is shown. Table 9 indicates the drilling parameters for the process.

Fig. 20. Removal speed over drilling depth


Table 9. Adjustments for the experiment of figure 20

#### **3.3.7 Dwelling time**

For this experiment the tool was positioned 4 µm above the nickel surface and remained at this position for different time periods. At the first position the dwelling time was 0 seconds. On each position the dwelling time was doubled to finally 640 seconds. The longer the pulses are applied, the more material is removed (figure 21). At 0 seconds only a scratch was produced. At higher dwelling times the holes are deeper. Finally, the removal rate decreases and a maximum gap will be developed. The electrical resistance between tool and work piece grows with the distance of them, until finally no more reaction/dissolution is possible. A referential groove was produced for the measurement. It is very important to adjust an optimized machine feed rate, because longer dwelling times lead to enlarged working gaps. The Adjustments for this experiment are illustrated in the table 10.


Table 10. Adjustments for the experiment of figure 21

electrolyte is not sufficient, so the dissolved metal ions saturate the electrolyte in the drilled hole and prevent any further metal dissolution. This can be disabled by an alternately up and down movement of the tool to realize a kind of flushing (pulsed mechanical movement). In figure 20 the removal speed over drilling depth is shown. Table 9 indicates

the drilling parameters for the process.

Fig. 20. Removal speed over drilling depth

**3.3.7 Dwelling time** 

in the table 10.

A = 3000 mV p = 80 ns T = 100 mV

I = 2000 µA ppr = 1/8 D = 75 µm

A = 3000 mV p = 80 ns T = 100 mV

I = 2000 µA ppr = 1/8 D = 75 µm

Table 10. Adjustments for the experiment of figure 21

For this experiment the tool was positioned 4 µm above the nickel surface and remained at this position for different time periods. At the first position the dwelling time was 0 seconds. On each position the dwelling time was doubled to finally 640 seconds. The longer the pulses are applied, the more material is removed (figure 21). At 0 seconds only a scratch was produced. At higher dwelling times the holes are deeper. Finally, the removal rate decreases and a maximum gap will be developed. The electrical resistance between tool and work piece grows with the distance of them, until finally no more reaction/dissolution is possible. A referential groove was produced for the measurement. It is very important to adjust an optimized machine feed rate, because longer dwelling times lead to enlarged working gaps. The Adjustments for this experiment are illustrated

Table 9. Adjustments for the experiment of figure 20

E = 0,5M HCl

E = 0,5M HCl

Fig. 21. Averaged groove depth over dwelling time

Fig. 22. Images of the microstructure, photographed with a scanning electron microscope (SEM) at different resolutions

Some Contributions at the Technology

of Electrochemical Micromachining with Ultra Short Voltage Pulses 21

was A = 2800 mV and p = 100 ns. The production with shorter pulse widths with the tool diameter of 150 µm was not possible. The energy applied by shorter pulse widths or lower amplitudes was not sufficient to recharge the double layer in order to realize material removal. By increasing the amplitude it was possible to finish grooves made with a pulse width of 80 ns, but the overall result was not favorable. The overview of the used

parameters for the experiment shown in figure 24 is illustrated in table 12.

Fig. 24. Working gap over pulse width for A = 2800 mV

Table 12. Adjustments for the experiment of figure 24

**3.4.2 Current through the backing electrode (I)** 

A = varied p = varied T = 100 mV

I = 1500 µA ppr = 1/8 D = 150 µm

This experiment was performed to show the influence of the cathodic protection-current on the process. The applied current protects the work piece in the electrolyte from corrosion or any other reactions. Eight grooves with the same dimensions as in the experiment before were made with I from 4000 to 500 µA. An obvious trend of how the cathodic protectioncurrent influences the process could not be observed from the series of grooves. The results show that I from 3000 to 4000 µA achieves the smallest working gap and the best surface condition. Figure 25 shows two grooves with an obvious optical difference. Topology of the ground, sharpness of the edges, and form of the groove is much better with I = 3000 µA. Therefore, I has to be fixed at 3000 µA for the next attempts. All other electrochemical parameters for this experiment are indicated in table 13. During this phase of the experiments, the choice of which of the parameters to fix was dedicated by the optical

assessment and the working gap measurement and not yet by the removal rate.

E = 3% HF/3M HCl

#### **3.3.8 Part production (micro injection mould)**

The manufactured microstructure in figure 22 has an overall diameter of less than 50 µm, is 15 µm deep, and approximately shaped like a gearwheel. This microstructure was manufactured in 4 hours, with an electrolyte concentration of 0,2M HCl. The tool for this experiment (figure 23) was made out of a tungsten wire with diameter D = 150 µm by successively reducing the diameter in the tooling basin to < 5 µm. The magnification of 45 in a light microscope was not sufficient to examine the structure; therefore, a scanning electron microscope has to be used. The experiment shows that the production of a micro injection mould in a range < 100 µm is possible with the IFT´s machine.

Fig. 23. Image of the tool to produce the micro injection mould with a top of D < 5µm.


Table 11. Adjustments for the experiment to produce a micro injection mould

#### **3.4 Manufacturing of steel (1.4301)**

1.4301 steel is the most widely used non corroding steel and it has a very broad scope of application. The need of micro-structuring of such a standard material is continually growing. A solution of hydrofluoric acid and hydrochloric acid was used as electrolyte. The exact designation of this electrolyte solution is 3% HF/3M HCl. As previously mentioned, four criteria were used for the optical consideration of the grooves. These are:


The experiments on 1.4301 were the same as on nickel with the difference that the electrolyte was not changed.

#### **3.4.1 Pulse width (p) and amplitude (A)**

Grooves with a length of 200 µm and a depth of 20 µm were manufactured. Thereon the amplitudes and the pulse widths were varied and the optical consideration of the grooves was performed to classify the results. The spatial resolution is almost linearly related to the pulse width. [Kock M.]. Figure 24 confirms this as the working gap shrinks with the reduction of the pulse width. The combination with the highest manufacturing precision

The manufactured microstructure in figure 22 has an overall diameter of less than 50 µm, is 15 µm deep, and approximately shaped like a gearwheel. This microstructure was manufactured in 4 hours, with an electrolyte concentration of 0,2M HCl. The tool for this experiment (figure 23) was made out of a tungsten wire with diameter D = 150 µm by successively reducing the diameter in the tooling basin to < 5 µm. The magnification of 45 in a light microscope was not sufficient to examine the structure; therefore, a scanning electron microscope has to be used. The experiment shows that the production of a micro injection

Fig. 23. Image of the tool to produce the micro injection mould with a top of D < 5µm.

A = 3000 mV p = 80 ns T = 100 mV E = 0,2M HCl

1.4301 steel is the most widely used non corroding steel and it has a very broad scope of application. The need of micro-structuring of such a standard material is continually growing. A solution of hydrofluoric acid and hydrochloric acid was used as electrolyte. The exact designation of this electrolyte solution is 3% HF/3M HCl. As previously mentioned,

The experiments on 1.4301 were the same as on nickel with the difference that the electrolyte

Grooves with a length of 200 µm and a depth of 20 µm were manufactured. Thereon the amplitudes and the pulse widths were varied and the optical consideration of the grooves was performed to classify the results. The spatial resolution is almost linearly related to the pulse width. [Kock M.]. Figure 24 confirms this as the working gap shrinks with the reduction of the pulse width. The combination with the highest manufacturing precision

**3.3.8 Part production (micro injection mould)** 

mould in a range < 100 µm is possible with the IFT´s machine.

I = 2000 µA ppr = 1/8 D < 5 µm

**3.4 Manufacturing of steel (1.4301)** 

topology/ smoothness of the bottom surface

**3.4.1 Pulse width (p) and amplitude (A)** 

shape/ geometry

 shine of the surface edge rounding

was not changed.

Table 11. Adjustments for the experiment to produce a micro injection mould

four criteria were used for the optical consideration of the grooves. These are:

was A = 2800 mV and p = 100 ns. The production with shorter pulse widths with the tool diameter of 150 µm was not possible. The energy applied by shorter pulse widths or lower amplitudes was not sufficient to recharge the double layer in order to realize material removal. By increasing the amplitude it was possible to finish grooves made with a pulse width of 80 ns, but the overall result was not favorable. The overview of the used parameters for the experiment shown in figure 24 is illustrated in table 12.

Fig. 24. Working gap over pulse width for A = 2800 mV


Table 12. Adjustments for the experiment of figure 24

#### **3.4.2 Current through the backing electrode (I)**

This experiment was performed to show the influence of the cathodic protection-current on the process. The applied current protects the work piece in the electrolyte from corrosion or any other reactions. Eight grooves with the same dimensions as in the experiment before were made with I from 4000 to 500 µA. An obvious trend of how the cathodic protectioncurrent influences the process could not be observed from the series of grooves. The results show that I from 3000 to 4000 µA achieves the smallest working gap and the best surface condition. Figure 25 shows two grooves with an obvious optical difference. Topology of the ground, sharpness of the edges, and form of the groove is much better with I = 3000 µA. Therefore, I has to be fixed at 3000 µA for the next attempts. All other electrochemical parameters for this experiment are indicated in table 13. During this phase of the experiments, the choice of which of the parameters to fix was dedicated by the optical assessment and the working gap measurement and not yet by the removal rate.

Some Contributions at the Technology

Fig. 27. Removal rate over pulse–pause ratio

**3.4.4 Drilling with µPECM** 

experiment are illustrated in table 15.

A = 2800 mV p = 100 ns T = 0 mV

I = 3000 µA ppr = varied D = 150 µm

To this point in the series of experiments all grooves were manufactured with an adjusted depth of 20 µm. This experiment was done to show how the manufacturing depth influences the process. Figure 28 shows that at a depth between 125 – 175 µm the speed of removal rapidly reduces from above 35 to less than 10 µm per minute. A possible explanation is that the electrolyte is not sufficiently available in the drilled hole. The electrolyte is sated in such depth, so the transport of new solved ions out of the bore slows down and the removal speed reduces. After the depth of around 425 µm was reached, the process was stopped, because it was no longer possible to manufacture the work piece. To prepare sufficient electrolyte solution in such depth and thus realize better transport of the solved ions out of the bore, the mechanical movement of the tool inside the drilled hole could be pulsed to get a kind of flushing and reach higher depths. The manufacturing parameters of this

Table 14. Adjustments for the experiment of figure 26 and 27

E = 3% HF/3M HCl

evaluation.

of Electrochemical Micromachining with Ultra Short Voltage Pulses 23

The best combination of the optical quality of the surface and the removal rate was detected from a pulse–pause ratio of 1/7. The consequence was to fix this parameter for the next experiments. Based on the optical result, the pulse–pause ratio of 1/5 was not viewed in the

Fig. 25. Image of grooves made with I = 3000 µA above, respectively I = 500 µA below


Table 13. Adjustments for the experiment of figure 25

#### **3.4.3 Pulse-pause ratio**

The idea of this experiment was a variation of the pulse–pause ratio from 1/5 to 1/11. Figure 26 shows the manufactured grooves of the ppr experiment. The manufacturing parameters of this process are illustrated in table 14. For this experiment the voltage at the tool was zero. An experiment with the potential at the tool has shown that a very low voltage leads to the best results in case of the optical considerations. But these low tool voltage could bring up some problems.

When the drilling depth is higher, it can happen that the positive ions from the work piece treatment deposit at the tool. This deposition starts with a slight change of the tool geometry and can lead to a kind of ion based short circuit bridge between tool and work piece. Such a short circuit disrupt the manufacturing process. For the further experimental work the tool voltage was set at 100 mV to avoid any unwanted occurances.

Fig. 26. Grooves produced for the pulse–pause ratio experiment

Figure 27 shows that the higher the pulse–pause ratio, the lower the removal rate. If within a period of time fewer pulses are applied, the charging and discharging of the electrochemical double layer also occurs less frequently. This is the obvious explanation for the low manufacturing speed of the groove made with a ppr of 1/11. For this ratio the manufacturing process was stopped because economic material removal could not be realized.

Fig. 25. Image of grooves made with I = 3000 µA above, respectively I = 500 µA below

I = varied ppr = 1/8 D = 150 µm

Table 13. Adjustments for the experiment of figure 25

voltage was set at 100 mV to avoid any unwanted occurances.

Fig. 26. Grooves produced for the pulse–pause ratio experiment

process was stopped because economic material removal could not be realized.

**3.4.3 Pulse-pause ratio** 

voltage could bring up some problems.

A = 2800 mV p = 100 ns T = 100 mV E = 3% HF/3M HCl

The idea of this experiment was a variation of the pulse–pause ratio from 1/5 to 1/11. Figure 26 shows the manufactured grooves of the ppr experiment. The manufacturing parameters of this process are illustrated in table 14. For this experiment the voltage at the tool was zero. An experiment with the potential at the tool has shown that a very low voltage leads to the best results in case of the optical considerations. But these low tool

When the drilling depth is higher, it can happen that the positive ions from the work piece treatment deposit at the tool. This deposition starts with a slight change of the tool geometry and can lead to a kind of ion based short circuit bridge between tool and work piece. Such a short circuit disrupt the manufacturing process. For the further experimental work the tool

Figure 27 shows that the higher the pulse–pause ratio, the lower the removal rate. If within a period of time fewer pulses are applied, the charging and discharging of the electrochemical double layer also occurs less frequently. This is the obvious explanation for the low manufacturing speed of the groove made with a ppr of 1/11. For this ratio the manufacturing The best combination of the optical quality of the surface and the removal rate was detected from a pulse–pause ratio of 1/7. The consequence was to fix this parameter for the next experiments. Based on the optical result, the pulse–pause ratio of 1/5 was not viewed in the evaluation.

Fig. 27. Removal rate over pulse–pause ratio


Table 14. Adjustments for the experiment of figure 26 and 27

#### **3.4.4 Drilling with µPECM**

To this point in the series of experiments all grooves were manufactured with an adjusted depth of 20 µm. This experiment was done to show how the manufacturing depth influences the process. Figure 28 shows that at a depth between 125 – 175 µm the speed of removal rapidly reduces from above 35 to less than 10 µm per minute. A possible explanation is that the electrolyte is not sufficiently available in the drilled hole. The electrolyte is sated in such depth, so the transport of new solved ions out of the bore slows down and the removal speed reduces. After the depth of around 425 µm was reached, the process was stopped, because it was no longer possible to manufacture the work piece. To prepare sufficient electrolyte solution in such depth and thus realize better transport of the solved ions out of the bore, the mechanical movement of the tool inside the drilled hole could be pulsed to get a kind of flushing and reach higher depths. The manufacturing parameters of this experiment are illustrated in table 15.

Some Contributions at the Technology

right)

of Electrochemical Micromachining with Ultra Short Voltage Pulses 25

A = 2800 mV p = 100 ns T = 100 mV E = 3% HF/3M HCl

The goal of the last experiment was to produce a micro structure with the knowledge of the described experimental work. So, the emblem of the Institute for Production Engineering and Laser Technology was chosen to be machined in a small steel plate. The first step, as in all other experiments, was to provide an appropriate tool to produce a high quality result. To manufacture grooves with a maximum width of 30 µm a tool diameter of about 20 µm is necessary. In a special tooling basin the diameter reduction from 150 µm to 20 µm was

Fig. 30. Tool before (diameter 150 µm - left) and after the tooling process (diameter ≈20 µm -

Figure 31 shows the result seen through a light microscope with forty-five-fold magnification and table 17 illustrates the used processing parameters. To get an idea of the dimensions of the emblem, a human hair was attached for comparison. The total removal time to produce this logo was 03:04:44 (hh:mm:ss). The groove 0-1 has an adjusted length of 322,5 µm and an adjusted depth of 30 µm. The manufacturing time was 11,02 minutes and

E = 3% HF/3M HCl

the width is 26,3 µm. This leads to a removal rate of 0,027 106 µm³/min.

A = 2300 mV p = 80 ns T = 100 mV

I = 3000 µA ppr = 1/7 D ≈ 20 µm

Table 17. Adjustments for the manufacturing of the Institute's logo

I = 3000 µA ppr = 1/7 D = 150 µm

Table 16. Adjustments for the experiment of figure 29

**3.4.6 Manufacturing of the Institute´s logo with µPECM** 

realised. Figure 30 shows the result of the tooling process.

Fig. 28. Removal speed over drilling depth


Table 15. Adjustments for the experiment of figure 28

#### **3.4.5 Dwelling time**

Figure 29 shows the effect of the dwelling time during the process. In this experiment a tool with a diameter of 150 μm was positioned 4 μm above the work piece´s surface. The parameters of the experiment are shown in table 16. The tool was stopped at eight different positions. On the first position the dwelling time was about 0 s, and afterwards it was doubled on each position from 5 s to 640 s. With the maximum depth of around -10 μm at the longest dwelling time this experiment confirmed the relevance of the dwelling time for the manufactured geometry. If the manufacturing feed rate is chosen too low, the precision of the manufactured geometry shrinks - caused by the time-dependent development of the working gap. This is one of the effects, which has to be controlled in industrial usage of the µPECM technology. Table 16 gives a overview of the process parameters for the dwelling time experiment.

Fig. 29. Averaged groove depth over dwelling time


Table 16. Adjustments for the experiment of figure 29

24 Cutting Edge Research in New Technologies

A = 2800 mV p = 100 ns T = 100 mV E = 3% HF/3M HCl

Figure 29 shows the effect of the dwelling time during the process. In this experiment a tool with a diameter of 150 μm was positioned 4 μm above the work piece´s surface. The parameters of the experiment are shown in table 16. The tool was stopped at eight different positions. On the first position the dwelling time was about 0 s, and afterwards it was doubled on each position from 5 s to 640 s. With the maximum depth of around -10 μm at the longest dwelling time this experiment confirmed the relevance of the dwelling time for the manufactured geometry. If the manufacturing feed rate is chosen too low, the precision of the manufactured geometry shrinks - caused by the time-dependent development of the working gap. This is one of the effects, which has to be controlled in industrial usage of the µPECM technology. Table 16 gives a overview of the process parameters for the dwelling

I = 3000 µA ppr = 1/7 D = 75 µm

Table 15. Adjustments for the experiment of figure 28

Fig. 29. Averaged groove depth over dwelling time

Fig. 28. Removal speed over drilling depth

**3.4.5 Dwelling time** 

time experiment.

#### **3.4.6 Manufacturing of the Institute´s logo with µPECM**

The goal of the last experiment was to produce a micro structure with the knowledge of the described experimental work. So, the emblem of the Institute for Production Engineering and Laser Technology was chosen to be machined in a small steel plate. The first step, as in all other experiments, was to provide an appropriate tool to produce a high quality result. To manufacture grooves with a maximum width of 30 µm a tool diameter of about 20 µm is necessary. In a special tooling basin the diameter reduction from 150 µm to 20 µm was realised. Figure 30 shows the result of the tooling process.

Fig. 30. Tool before (diameter 150 µm - left) and after the tooling process (diameter ≈20 µm right)

Figure 31 shows the result seen through a light microscope with forty-five-fold magnification and table 17 illustrates the used processing parameters. To get an idea of the dimensions of the emblem, a human hair was attached for comparison. The total removal time to produce this logo was 03:04:44 (hh:mm:ss). The groove 0-1 has an adjusted length of 322,5 µm and an adjusted depth of 30 µm. The manufacturing time was 11,02 minutes and the width is 26,3 µm. This leads to a removal rate of 0,027 106 µm³/min.


Table 17. Adjustments for the manufacturing of the Institute's logo

Some Contributions at the Technology

process can be done.

**5. Prospects** 

**6. References** 

Germany

1, Weinheim, Germany

of Electrochemical Micromachining with Ultra Short Voltage Pulses 27

Caused by the complexity of this technology, the variation of one of the adjustable parameters could significantly affect the result. Therefore at this point of research it is not definitely possible to give tangible instructions on how to reach requested results. It is very much experience necessary to interpret the proceedings at the machine correctly and to enhance the manufacturing process. Due to the multidisciplinary nature of this technology, intensified cooperation with other experts and an extensive research study has to be done; before a reasonable forecast for the processing parameters of a specific manufacturing

E = 0,2M HCl

E = 3% HF/3M HCl

A = 3000 mV p = 80 ns T = 100 mV

I = 2000 µA ppr = 1/8 D = 75 µm

A = 2800 mV p = 100 ns T = 100 mV

I = 3000 µA ppr = 1/7 D = 150 µm

producer can find the printed serial number, due to the small size of it.

Dissertation – Freie Universität Berlin, Berlin, Germany

Table 19. Adjustments to achieve appropriate results working on steel (1.4301)

In the course of the experiments, it was also tried to treat carbide metal by electrochemical micromachining with ultra short pulses. The work piece used for experimental work was a K40FF. This carbide metal consists of a 12% cobalt matrix with 88% tungsten-carbide as stengthener. The electrolytes used were 3% HF/3M HCl and 2M NaOH. Both electrolytes were found to be unsuitable in combination with this carbide metal. A major challenge is to find new material-electrolyte combinations to apply electrochemical micromachining with ultra short pulses. The IFT has some tangible visions to realize treatment of carbide metal. A prospectively area for application of this technology could be protection of plagiarism. Technical devices and parts could be branded with the µPECM technology so, that only the

Buhlert, M. (2009). *Elektropolieren*, Eugen G. Leuze Verlag, ISBN 978-3-87480-249-9, Saulgau,

Hamann, C.H. & Vielstich, W. (2005). *Elektrochemie*, 4. vollständig überarbeitete und

Kirchner, V. (2001). *Elektrochemische Mikrostrukturierung mit ultrakurzen Spannungsimpulsen*,

aktualisierte Auflage, WILEY-VCH Verlag GmbH & Co. KGaA, ISBN 3-527-31068-

Table 18. Adjustments to achieve appropriate results working on nickel

Fig. 31. Logo of the Institute in comparison to a human hair (diameter ≈ 50 µm)

#### **4. Conclusion**

The technology of electrochemical micromachining with ultra short voltage pulses has successfully displayed the many applications especially for prototype building or for the manufacturing of special products where there is no other technology which can combine a very high manufacturing precision for special materials without any mechanical forces or thermal influences. [Zemann R.] In principal, it can be applied to all electrochemically active materials, including semiconductors. [Schuster R.] Also, the use of applicable effects on process accuracy and material removal rate of difficult to machine materials offers a wide range of possible applications for µPECM technologies in the future. The occurring electrochemical problems are tradable and topics at the IFT, as well as the micromachining of many different materials like nickel, tungsten, titanium, non-corroding steels, or hard metals. As already mentioned, the machine at the IFT is simple constructed and very easy to maintain, so it is adequate for industrial use. However, a more complex machine structure would enable to reach highest precision requirements, but needs more maintenance and a higher financial investment. The experiments on the IFT´s machine proved that electrochemical micromachining is achievable for SME's. With the parameter sets in table 18 and 19 appropriate results were manufactured. Appropriate results means, that with these parameters, the grooves deliver adequate working gaps and optical results – geometry, topology, sharpness of the edges, and shine of the ground. Other parameters would perhaps reach higher removal rates, but on the other side lose quality with regard to precision.


Table 18. Adjustments to achieve appropriate results working on nickel

Caused by the complexity of this technology, the variation of one of the adjustable parameters could significantly affect the result. Therefore at this point of research it is not definitely possible to give tangible instructions on how to reach requested results. It is very much experience necessary to interpret the proceedings at the machine correctly and to enhance the manufacturing process. Due to the multidisciplinary nature of this technology, intensified cooperation with other experts and an extensive research study has to be done; before a reasonable forecast for the processing parameters of a specific manufacturing process can be done.


Table 19. Adjustments to achieve appropriate results working on steel (1.4301)

#### **5. Prospects**

26 Cutting Edge Research in New Technologies

Fig. 31. Logo of the Institute in comparison to a human hair (diameter ≈ 50 µm)

The technology of electrochemical micromachining with ultra short voltage pulses has successfully displayed the many applications especially for prototype building or for the manufacturing of special products where there is no other technology which can combine a very high manufacturing precision for special materials without any mechanical forces or thermal influences. [Zemann R.] In principal, it can be applied to all electrochemically active materials, including semiconductors. [Schuster R.] Also, the use of applicable effects on process accuracy and material removal rate of difficult to machine materials offers a wide range of possible applications for µPECM technologies in the future. The occurring electrochemical problems are tradable and topics at the IFT, as well as the micromachining of many different materials like nickel, tungsten, titanium, non-corroding steels, or hard metals. As already mentioned, the machine at the IFT is simple constructed and very easy to maintain, so it is adequate for industrial use. However, a more complex machine structure would enable to reach highest precision requirements, but needs more maintenance and a higher financial investment. The experiments on the IFT´s machine proved that electrochemical micromachining is achievable for SME's. With the parameter sets in table 18 and 19 appropriate results were manufactured. Appropriate results means, that with these parameters, the grooves deliver adequate working gaps and optical results – geometry, topology, sharpness of the edges, and shine of the ground. Other parameters would perhaps reach higher removal rates, but on the other side lose quality with regard to precision.

**4. Conclusion** 

In the course of the experiments, it was also tried to treat carbide metal by electrochemical micromachining with ultra short pulses. The work piece used for experimental work was a K40FF. This carbide metal consists of a 12% cobalt matrix with 88% tungsten-carbide as stengthener. The electrolytes used were 3% HF/3M HCl and 2M NaOH. Both electrolytes were found to be unsuitable in combination with this carbide metal. A major challenge is to find new material-electrolyte combinations to apply electrochemical micromachining with ultra short pulses. The IFT has some tangible visions to realize treatment of carbide metal. A prospectively area for application of this technology could be protection of plagiarism. Technical devices and parts could be branded with the µPECM technology so, that only the producer can find the printed serial number, due to the small size of it.

### **6. References**


**2** 

Branko L. Dokic

*Bosnia and Herzegovina* 

**CMOS and BiCMOS Regenerative Logic Circuits** 

Schmitt triggers with standard CMOS logic circuits are described, first. Mathematical models for calculating basic parameters and their limits are presented. Most of the chapter is dedicated to different solutions for CMOS and BiCMOS Schmitt logic circuits in monolithic integrated circuits. Two types of inverters with entirely different topologies are described. Also, solutions for Schmitt triggers with voltage-controlled thresholds are described. Beside inverters, NAND and NOR Schmitt logic circuits are analyzed. Basic circuit is inverted Schmitt trigger with three pairs of CMOS transistors. Expansion of the number of inputs is reached in a similar way as in standard CMOS and BiCMOS logic circuits. It is shown that voltage transfer characteristics depend, beside voltage supply and parameters of transistors, on the number of logical circuits' inputs. NAND and NOR Schmitt circuits, in which voltage hysteresis in transfer characteristic is generated only through one input, are also described. Analytic models and SPICE simulations are used for analysis of static and dynamic parameters and conditions for work stability and reliability. Areas of reliability, influence of

technology and electrical parameters of transistors and their limits are analyzed.

1996, Dokic, 1988). Today, some of them (Dokic, 1984) are treated as conventional.

interferences. An example of this kind of application is given in Fig.1.

noise amplitude of which is greater than the voltage hysteresis.

Concerning the field of application, in literature there are different solutions of Schmitt triggers (Zou et al, 2008, Al-Sarrawi, 2008, Katyal et al, 2008, Lo et al, 2010). In this chapter, solutions with fundamental applications in digital integrated circuits – Schmitt logic circuits are described. The author published most of these solutions (Dokic, 1983, Dokic 1984, Dokic

The term regenerative is used because every change of state is followed by a regenerative process – positive feedback. Owning to that, transfer characteristic has shape of a hysteresis, like in Schmitt trigger. That is why the term Schmitt logic circuits is most commonly used. Unlike conventional logic circuits, where the output level is uniformly determined for the input voltage value, for Schmitt logic circuits, in certain extent, it is not uniformly determined. In fact, due to hysteresis, in the area of the input voltages between two logic thresholds, logic state at the output depends, beside the input voltage value, also on the previous state. Due to that Schmitt circuits can be used as filters for low frequency

Whenever the value of the input signal passes the value of the threshold voltage ்ܸ of the standard logic circuit, a change of the logic state at the output appears. Therefore, the changes of the input voltage created by noise are transferred to the output as glitches. The change of the logic state at the output of the Schmitt logic circuit can appear only after the

**1. Introduction** 

*University of Banja Luka, Faculty of Electrical Engineering* 


## **CMOS and BiCMOS Regenerative Logic Circuits**

Branko L. Dokic

*University of Banja Luka, Faculty of Electrical Engineering Bosnia and Herzegovina* 

#### **1. Introduction**

28 Cutting Edge Research in New Technologies

Kock, M. (2004). *Grenzen der Möglichkeiten der elektrochemischen Mikrostrukturierung mit* 

Schuster, R., Kirchner V., Allongue, P. (2000). *Electrochemical Micromachining*, SCIENCE Vol

Zemann, R. (2010). Electrochemical Milling, *Annals of DAAAM for 2010 & Proceedings of the* 

Germany

Vienna, Austria

289, sciencemag, 7. July 2000, p. 98-101

*ultrakurzen Spannungspulsen*, Dissertation - Freie Universität Berlin, Berlin,

*21st International DAAAM Symposium "Intelligent Manufacturing & Automation: Focus on Interdisciplinary Solutions"*, 20-23rd October 2010, Zadar, Croatia, B. Katalinic, ISSN 1726-9679, ISBN 978-3-901509-73-5, S. 843 – 844, DAAAM International,

> Schmitt triggers with standard CMOS logic circuits are described, first. Mathematical models for calculating basic parameters and their limits are presented. Most of the chapter is dedicated to different solutions for CMOS and BiCMOS Schmitt logic circuits in monolithic integrated circuits. Two types of inverters with entirely different topologies are described. Also, solutions for Schmitt triggers with voltage-controlled thresholds are described. Beside inverters, NAND and NOR Schmitt logic circuits are analyzed. Basic circuit is inverted Schmitt trigger with three pairs of CMOS transistors. Expansion of the number of inputs is reached in a similar way as in standard CMOS and BiCMOS logic circuits. It is shown that voltage transfer characteristics depend, beside voltage supply and parameters of transistors, on the number of logical circuits' inputs. NAND and NOR Schmitt circuits, in which voltage hysteresis in transfer characteristic is generated only through one input, are also described. Analytic models and SPICE simulations are used for analysis of static and dynamic parameters and conditions for work stability and reliability. Areas of reliability, influence of technology and electrical parameters of transistors and their limits are analyzed.

> Concerning the field of application, in literature there are different solutions of Schmitt triggers (Zou et al, 2008, Al-Sarrawi, 2008, Katyal et al, 2008, Lo et al, 2010). In this chapter, solutions with fundamental applications in digital integrated circuits – Schmitt logic circuits are described. The author published most of these solutions (Dokic, 1983, Dokic 1984, Dokic 1996, Dokic, 1988). Today, some of them (Dokic, 1984) are treated as conventional.

> The term regenerative is used because every change of state is followed by a regenerative process – positive feedback. Owning to that, transfer characteristic has shape of a hysteresis, like in Schmitt trigger. That is why the term Schmitt logic circuits is most commonly used. Unlike conventional logic circuits, where the output level is uniformly determined for the input voltage value, for Schmitt logic circuits, in certain extent, it is not uniformly determined. In fact, due to hysteresis, in the area of the input voltages between two logic thresholds, logic state at the output depends, beside the input voltage value, also on the previous state. Due to that Schmitt circuits can be used as filters for low frequency interferences. An example of this kind of application is given in Fig.1.

> Whenever the value of the input signal passes the value of the threshold voltage ்ܸ of the standard logic circuit, a change of the logic state at the output appears. Therefore, the changes of the input voltage created by noise are transferred to the output as glitches. The change of the logic state at the output of the Schmitt logic circuit can appear only after the noise amplitude of which is greater than the voltage hysteresis.

CMOS and BiCMOS Regenerative Logic Circuits 31

integrated circuits with mixed signals contain Schmitt triggers (Young, 2010, Chien, 2011, Li,

The high threshold ��� and the low threshold ��� are defined by the supply voltage ���,

1 / *DD tp tn n p*

��� and ��� are threshold voltages of nMOS and pMOS transistors, respectively, and the

, / <sup>2</sup>

where �� and �� are the mobility of the electrons and the holes, ��� is oxide dielectric constants, ��� the oxide thickness and � and � are width and length of the transistor's

The fact that basic parameters depend on the ratio of the resistance ����� yields a wide choice of their absolute values. These values range from several tens of �� (�� � �� � ��, where �� represents output resistance of the inverter ��), to several hundreds of ��. On the other hand, their ratio can vary within a broad range. Since it is always true: �� � ���, then �� � ��. The other limitations do not exist, it is possible for the ratio to be ����� ≪ 1. This

*V VVkk*

*n p*

*k k*

*T*

/ <sup>2</sup> *n ox n nn ox k WL t* 

*V*

*V V RR TH T* 1 /1 2 (4)

*V V R R VR R TL T* 1/ / 12 12 *DD* (5)

*V V V VR R H TH TL DD* 1 2 / (6)

(7)

(8)

/

*p ox p pp ox k WL t* 

ratio of the resistors ����� and the threshold voltage �� of the inverter ��. Namely,

The basic Schmitt trigger consists of two CMOS inverters and two resistors (Fig.2).

2009, Hard & Voinigesku, 2009, Wang et al, 2008, Arrabi 2011).

Fig. 2. Schmitt trigger (a) and it's transfer characteristic (b).

**2. Schmitt trigger with logic circuits** 

and voltage hysteresis is given by:

constants of transistors are given by:

where �� is given by:

channel.

Fig. 1. Transfer characteristic of Schmitt logic circuit (a) and outputs of standard and Schmitt circuit to an input with noise addition (b).

Schmitt trigger is able to hold it's logic state for all changes of the input voltage which are ��� � �� � ���, where ��� and ��� are the high and the low threshold of the Schmitt trigger. Fig.1 shows the ability of the Schmitt trigger to filter the noise, which, in this case, do not influence the output of the circuit. At the same time at the output of the standard circuit there are two pulse glitches which create system errors.

The hysteresis within the transfer characteristic causes increase of the static noise immunity (Fig.1a). Thus:

$$V\_{\rm NIL} = V\_{\rm TH} \,\,\,\,\,\,\tag{1}$$

$$V\_{\rm NIH} = V\_{\rm DD} - V\_{\rm TL} \tag{2}$$

Comparing Schmitt triggers to the standard circuits, the increase of the static noise immunity appears if the threshold voltage �� of the standard gate lies between the thresholds of the Schmitt circuit, i.e.:

$$V\_{TL} < V\_T < V\_{TH} \tag{3}$$

The transfer characteristic, concerning the noise immunity, is optimum if the thresholds ��� and ��� are symmetric around ���/2. The larger the value of the voltage hysteresis becomes, the noise immunity increases further. The transfer characteristic of the CMOS Schmitt gates is almost perfectly symmetric around ���/2.

There is another advantage of Schmitt circuits compared to the standard ones. Because of the positive feedback, the transfer characteristic is ideal, which means that values of noise margin and noise immunity are equal. Schmitt triggers are used to shape pulses or convert signals that change slowly into pulse signals with short rise and fall times, which is necessary where synchronizing circuits are used. These are the reasons to use Schmitt triggers so often, both as an independent integrated circuit and as a part of a MSI or VLSI circuit. In the second case, Schmitt trigger is almost always used as an input circuit. Also,

Fig. 1. Transfer characteristic of Schmitt logic circuit (a) and outputs of standard and Schmitt

Schmitt trigger is able to hold it's logic state for all changes of the input voltage which are ��� � �� � ���, where ��� and ��� are the high and the low threshold of the Schmitt trigger. Fig.1 shows the ability of the Schmitt trigger to filter the noise, which, in this case, do not influence the output of the circuit. At the same time at the output of the standard circuit

The hysteresis within the transfer characteristic causes increase of the static noise immunity

Comparing Schmitt triggers to the standard circuits, the increase of the static noise immunity appears if the threshold voltage �� of the standard gate lies between the

The transfer characteristic, concerning the noise immunity, is optimum if the thresholds ��� and ��� are symmetric around ���/2. The larger the value of the voltage hysteresis becomes, the noise immunity increases further. The transfer characteristic of the CMOS Schmitt gates

There is another advantage of Schmitt circuits compared to the standard ones. Because of the positive feedback, the transfer characteristic is ideal, which means that values of noise margin and noise immunity are equal. Schmitt triggers are used to shape pulses or convert signals that change slowly into pulse signals with short rise and fall times, which is necessary where synchronizing circuits are used. These are the reasons to use Schmitt triggers so often, both as an independent integrated circuit and as a part of a MSI or VLSI circuit. In the second case, Schmitt trigger is almost always used as an input circuit. Also,

*V V NIL TH* , (1)

*V VV NIH DD TL* (2)

*V VV TL T TH* (3)

circuit to an input with noise addition (b).

thresholds of the Schmitt circuit, i.e.:

is almost perfectly symmetric around ���/2.

(Fig.1a). Thus:

there are two pulse glitches which create system errors.

integrated circuits with mixed signals contain Schmitt triggers (Young, 2010, Chien, 2011, Li, 2009, Hard & Voinigesku, 2009, Wang et al, 2008, Arrabi 2011).

#### **2. Schmitt trigger with logic circuits**

The basic Schmitt trigger consists of two CMOS inverters and two resistors (Fig.2).

Fig. 2. Schmitt trigger (a) and it's transfer characteristic (b).

The high threshold ��� and the low threshold ��� are defined by the supply voltage ���, ratio of the resistors ����� and the threshold voltage �� of the inverter ��. Namely,

$$V\_{TH} = V\_T \left(1 + R\_1 \, / \, R\_2\right) \tag{4}$$

$$V\_{TL} = V\_T \left(1 + R\_1 \, / \, R\_2 \right) - V\_{DD} R\_1 \, / \, R\_2 \tag{5}$$

and voltage hysteresis is given by:

$$V\_H = V\_{TH} - V\_{TL} = V\_{DD} R\_1 \, / \, R\_2 \tag{6}$$

where �� is given by:

$$V\_T = \frac{V\_{DD} + V\_{tp} + V\_{tn}\sqrt{k\_n / k\_p}}{1 + \sqrt{k\_n / k\_p}}\tag{7}$$

��� and ��� are threshold voltages of nMOS and pMOS transistors, respectively, and the constants of transistors are given by:

$$k\_n = \frac{\mu\_n \varepsilon\_{\alpha x}}{\mathfrak{D}t\_{\alpha x}} \mathcal{W}\_n \;/\ L\_n \;/\ \ k\_p = \frac{\mu\_p \varepsilon\_{\alpha x}}{\mathfrak{D}t\_{\alpha x}} \mathcal{W}\_p \;/\ L\_p \tag{8}$$

where �� and �� are the mobility of the electrons and the holes, ��� is oxide dielectric constants, ��� the oxide thickness and � and � are width and length of the transistor's channel.

The fact that basic parameters depend on the ratio of the resistance ����� yields a wide choice of their absolute values. These values range from several tens of �� (�� � �� � ��, where �� represents output resistance of the inverter ��), to several hundreds of ��. On the other hand, their ratio can vary within a broad range. Since it is always true: �� � ���, then �� � ��. The other limitations do not exist, it is possible for the ratio to be ����� ≪ 1. This

$$\begin{array}{c} V\_o = V\_{T2} \end{array} \tag{9}$$

$$I\_{Dn} = I\_{Dp} + \frac{V\_{DD} - V\_o}{R} \tag{10}$$

$$k\_n \left(V\_i - V\_{tn}\right)^2 = k\_p \left(V\_{DD} + V\_{tp} - V\_i\right)^2 + \frac{V\_{DD} - V\_o}{R} \tag{11}$$

CMOS and BiCMOS Regenerative Logic Circuits 35

The resistance � should be within the range from several hundreds of Ω up to several �Ω. Sensivity of the threshold change decreases if � increases. If � � ��Ω, this sensitivity is

Fig. 7. Schmitt trigger with transmission gate instead of the resistor (a) and the dependency

Instead of resistor � the transmission gate can be used (Fig.7a). The control input of the transmission gate should be in the logic state which keeps it on all the time. In this case, TG acts as a resistor whose resistance depends on the type of TG and the supply voltage. In Fig.7b the dependency of threshold voltages on the supply voltage ��� is shown, if the inverters are CD4069, and TG is CD4066 (full line) or CD4016 (broken line). The resistance of the TG CD4066 is lower, thus the voltage hysteresis of the Schmitt trigger is wider in this

If NAND and NOR logic circuits are used instead of input inverters in Fig.3 and 7a the

The basic circuit is the Schmitt trigger – inverter with three pairs of CMOS transistors (Fig.8). This solution has been initially proposed in (Dokic, 1984), which is the most

of the high and low threshold voltages on the supply voltage (b).

case, than when CD4016 is used.

**3. Schmitt trigger – inverter** 

NAND and NOR Schmitt trigger are obtained.

The dependency of the thresholds ��� and ��� on � and ��� is shown in Fig.6.

Fig. 6. The high and the low threshold voltages as functions of � and ���.

very low.

which leads to:

$$V\_o = V\_{DD} - k\_p R \left[ \frac{k\_n}{k\_p} \left( V\_i - V\_{tn} \right)^2 - \left( V\_{DD} + V\_{tp} - V\_i \right)^2 \right] \tag{12}$$

Practically, inverters are symmetric, or almost symmetric, circuits. That is why we, in further text, consider such a case: �� � �� and ��� � ����� of all transistors. Then (12) can be written as:

$$V\_o^\dagger = V\_{DD} \left[ \mathbf{1} + k\_p R \left( V\_{DD} + \mathbf{2} V\_{tp} \right) \right] - \mathbf{2} k\_p R \left( V\_{DD} + V\_{tp} - V\_{tn} \right) V\_i \tag{13}$$

and the condition of change of the output voltage becomes:

$$\overline{V\_o}^{'} = V\_{\text{DD}} \not\mid \mathbf{2} \tag{14}$$

When �� � � ���/2, then �� � ���, so, taking into account (13) and (14), the high voltage of the Schmitt trigger is given by:

$$V\_{TH} = \frac{V\_{DD}}{2} + \frac{V\_{DD}}{4k\_pR\left(V\_{DD} + V\_{tp} - V\_{tn}\right)}\tag{15}$$

#### **2.2 Determining the low threshold voltage** ���

Equivalent circuit used to determine ��� is shown in Fig.5b. Then:

$$I\_{Dp} = I\_{Dn} + \boldsymbol{V\_o}^{\cdot}/R \tag{16}$$

Between points F and G both transistors are saturated, so:

$$k\_p \left(V\_{DD} + V\_{tp} - V\_i\right)^2 = k\_n \left(V\_i - V\_{tn}\right)^2 + V\_o^\cdot / R \tag{17}$$

Presuming that the transistors are symmetric, (17) leads to:

$$V\_o^{\prime} = k\_n R \left(V\_{DD} + V\_{tp} - V\_i\right) \left(V\_i - V\_{tn}\right) \tag{18}$$

Combining (14) and (18) and replacing �� � ���, low threshold voltage is given by:

$$V\_{TL} = \frac{V\_{DD}}{2} - \frac{V\_{DD}}{4k\_n \left(V\_{DD} + V\_{tp} - V\_{tn}\right)R} \tag{19}$$

Voltage hysteresis is:

$$V\_H = V\_{TH} - V\_{TL} = \frac{V\_{DD}}{2k\_p \left(V\_{DD} + V\_{tp} - V\_{tn}\right)R} \tag{20}$$

From the previous analysis, the following conclusions are derived:


<sup>2</sup> ' <sup>2</sup> *<sup>n</sup> o DD p i tn DD tp i*

Practically, inverters are symmetric, or almost symmetric, circuits. That is why we, in further text, consider such a case: �� � �� and ��� � ����� of all transistors. Then (12) can be written as:

' 1 22 *V V kRV V kRV V V V o DD <sup>p</sup> DD tp p DD tp tn i*

� � ���/2, then �� � ���, so, taking into account (13) and (14), the high voltage of the

2 4

'

*p DD tp tn*

*kRV V V*

*DD DD*

'

*V V <sup>V</sup>*

Combining (14) and (18) and replacing �� � ���, low threshold voltage is given by:

*V V <sup>V</sup>*

*<sup>V</sup> VV V*

 2 4 *DD DD*

*kV V V R*

*kV V V R*

*n DD tp tn*

2 *DD*

*p DD tp tn*

(13)

/ 2 *V V o DD* (14)

*I I VR Dp Dn o* / (16)

<sup>2</sup> <sup>2</sup> ' *kV V V kV V V R p DD tp i n i tn o* / (17)

' *V kRV V V V V o n DD t <sup>p</sup> i i tn* (18)

(12)

(15)

(19)

(20)

*<sup>k</sup> V V kR V V V V V*

*p*

*k*

and the condition of change of the output voltage becomes:

*TH*

Equivalent circuit used to determine ��� is shown in Fig.5b. Then:

Between points F and G both transistors are saturated, so:

Presuming that the transistors are symmetric, (17) leads to:

*TL*

*H TH TL*

From the previous analysis, the following conclusions are derived:

 the thresholds are symmetric relative to ���/2; ���, ��� and �� are inversely proportional to �.

**2.2 Determining the low threshold voltage** ���

which leads to:

When ��

Schmitt trigger is given by:

Voltage hysteresis is:

The dependency of the thresholds ��� and ��� on � and ��� is shown in Fig.6.

Fig. 6. The high and the low threshold voltages as functions of � and ���.

The resistance � should be within the range from several hundreds of Ω up to several �Ω. Sensivity of the threshold change decreases if � increases. If � � ��Ω, this sensitivity is very low.

Fig. 7. Schmitt trigger with transmission gate instead of the resistor (a) and the dependency of the high and low threshold voltages on the supply voltage (b).

Instead of resistor � the transmission gate can be used (Fig.7a). The control input of the transmission gate should be in the logic state which keeps it on all the time. In this case, TG acts as a resistor whose resistance depends on the type of TG and the supply voltage. In Fig.7b the dependency of threshold voltages on the supply voltage ��� is shown, if the inverters are CD4069, and TG is CD4066 (full line) or CD4016 (broken line). The resistance of the TG CD4066 is lower, thus the voltage hysteresis of the Schmitt trigger is wider in this case, than when CD4016 is used.

If NAND and NOR logic circuits are used instead of input inverters in Fig.3 and 7a the NAND and NOR Schmitt trigger are obtained.

#### **3. Schmitt trigger – inverter**

The basic circuit is the Schmitt trigger – inverter with three pairs of CMOS transistors (Fig.8). This solution has been initially proposed in (Dokic, 1984), which is the most

$$V\_1 = V\_{DD} - V\_{tn0} - A\_n \left(V\_i - V\_{tn1}\right) \tag{21}$$

$$A\_{\pi} = \sqrt{\frac{W\_{n1} / L\_{n1}}{W\_{n0} / L\_{n0}}} \quad \text{and} \quad V\_{tn1} \le V\_i \le V\_1 + V\_{tn} \tag{22}$$

$$V\_{IN} = V\_{tn} + \frac{V\_{DD} - V\_{tn}}{1 + A\_n} \tag{23}$$

$$k\_n \left(V\_i - V\_1 - V\_{t\text{tr1}}\right)^2 = k\_{p\epsilon} \left(V\_{DD} + V\_{tp} - V\_i\right)^2\tag{24}$$

$$V\_{TH} = V\_{tn} + \frac{V\_{DD} + V\_{tp} - V\_{tn} + B\_1 \left(V\_{DD} - V\_{tn}\right)}{1 + B\_1 \left(1 + A\_n\right)}\tag{25}$$

$$B\_1 = \sqrt{k\_n \;/\; k\_{p\epsilon}} = \sqrt{2k\_n \;/\; k\_p} \tag{26}$$

$$V\_2 = A\_p \left( V\_{DD} + V\_{tp} - V\_i \right) - V\_{tp} \tag{27}$$

$$A\_p = \sqrt{\frac{\mathcal{W}\_p \, / \, L\_p}{\mathcal{W}\_{p0} \, / \, L\_{p0}}} \tag{28}$$

$$V\_{IP} = V\_{DD} + V\_{tp} - \frac{V\_{DD} + V\_{tp}}{1 + A\_p} \tag{29}$$

$$k\_{nc} \left(V\_i - V\_{tn}\right)^2 = k\_p \left(V\_2 + V\_{tp} - V\_i\right)^2\tag{30}$$

$$V\_{TL} = V\_{tn} + \frac{A\_p \left(V\_{DD} + V\_{tp} - V\_{tn}\right) - V\_{tn}}{1 + B\_2 + A\_p} \tag{31}$$

$$B\_2 = \sqrt{k\_{ne} \,/\, k\_p} = \sqrt{k\_n \,/\left(2k\_p\right)}\tag{32}$$


$$A\_n = \sqrt{\mathcal{V}\mathcal{V}\_n \not\supset \mathcal{V}\mathcal{V}\_{n0}} \; \_\mathcal{A} = \sqrt{\mathcal{V}\mathcal{V}\_p \not\supset \mathcal{V}\mathcal{V}\_{p0}} \tag{33}$$

$$V\_{TL} = V\_{tn} + \frac{V\_{DD} + V\_{tp} - V\_{tn}\sqrt{\left(k\_n \,/\, \text{2}\right) / k\_p}}{1 + \sqrt{\left(k\_n \,/\, \text{2}\right) / k\_p}}\tag{34}$$

$$I\_{Dn} = I\_{Dp} + I\_{Dp1} \tag{35}$$

$$dV\_o \, / \, dV\_i = -1\tag{36}$$

$$R\_{p1} \approx \frac{1}{2k\_{p1} \left(V\_{DD} + V\_{tp1}\right)}\tag{37}$$

$$V\_{TH} = \frac{V\_{DD}}{2} + \frac{k\_{p1}}{k\_p} \frac{V\_{DD} + V\_{tp1}}{2\left(V\_{DD} + V\_{tp} - V\_{tn}\right)} V\_{DD} \tag{38}$$

$$R\_{n1} \approx \frac{1}{2k\_{n1} \left(V\_{DD} - V\_{tn1}\right)}\tag{39}$$

$$V\_{TL} = \frac{V\_{DD}}{2} - \frac{k\_{n1}}{k\_n} \frac{V\_{DD} - V\_{tn1}}{2\left(V\_{DD} + V\_{tp} - V\_{tn}\right)} V\_{DD} \tag{40}$$

$$V\_H = \frac{k\_1}{k} \frac{V\_{DD} - V\_t}{V\_{DD} - \mathfrak{D}V\_t} V\_{DD} \tag{41}$$

$$V\_o \left(V\_{\rm DD}\right) \le V\_{\rm T2} = V\_{\rm DD} / \,\,\,\,\,\,\,\text{and}\,\,\,V\_o \left(0\right) \ge V\_{\rm T2} = V\_{\rm DD} / \,\,\,\,\,\,\tag{42}$$

$$k\_n \nmid k\_{v1} > 2 \text{ and } k\_v \nmid k\_{n1} > 2 \tag{43}$$

CMOS and BiCMOS Regenerative Logic Circuits 49

*<sup>n</sup> ne ne <sup>k</sup> k k m p*

where *p* marks the position of the first active input (for example, if ൌ͵, the inputs of the

<sup>ᇱ</sup> are at ܸ, and ܯଷ, ܯଷ

1 0 / 1 *A k k Amp ne ne n n*

1 1 / 1 *B k k B nm p e ne pe*

From eq. (25), replacing ܣ by ܣ and ܤଵ by ܤଵ we obtain the high threshold voltage of the

1 11 1

To calculate ்ܸ the circuit in Fig.21a can be replaced by an equivalent one shown in Fig.9b.

ܯଶ is on. ܯ and ܯଵ (݅ ൌ ͳǡ ǥ ǡ ݊) with active input can be replaced by equivalent

1/2 *B k k B mn* <sup>2</sup>*e ne pe* / 1 2

From eq. (36), replacing ܣ by ܣ and ܤଶ by ܤଶ, we obtain the low threshold voltage of the

<sup>2</sup> 1

*p*

*A n V V B mn V*

*A n B mn*

1/2 1/2 2 1/2 1/2

*p DD tp tn*

ܯ ,... , <sup>ᇱ</sup>

*B nm p A m p*

<sup>1</sup> 1

1/2

1/2

*DD tp DD n tn*

*V V B nm p V A m p V*

(46)

(47)

1/2 1/2

ᇱ are on and can be replaced by an equivalent

*kk m ne n* / 2 (49)

*pe pe*<sup>1</sup> *<sup>p</sup> k k nk* (50)

1/2 *A k k An pe pe p p* 1 2 / (51)

(52)

(53)

1/2 1/2

1 1

*n*

ᇱ can be replaced by the equivalent transistors

(45)

<sup>ᇱ</sup> are at ܸ). Now, eq.(22) and (26)

(48)

ܯ ,... , <sup>ᇱ</sup>

ܯ and ܯଵ, respectively, whose constants ݇ are given by:

where ܣ and ܤଵ are given by eqs. (22) and (26), respectively.

1

1

transistors ܯ and ܯଵ, respectively, with the constants ݇:

*TL*

where ܣ and ܤଶ are given by eqs. (27) and (32), respectively.

*V*

The ݇ ratios of ܯଵ to ܯ and ܯ to ܯଵ, respectively, are given by:

<sup>ᇱ</sup> and ܯଶ, ܯଶ

the transistors ܯଵ, …, ܯ and ܯଵ

transistors ܯଵ, ܯଵ

become, respectively:

*m*-inputs NAND Schmitt circuit:

*TH*

Namely, ܯଶ is off. ܯଵ, …, ܯ, ܯଵ

*V*

one ܯ with the constant ݇:

*m*-inputs NAND Schmitt circuits:

Fig. 21. Principle schematics of m-input NAND and NOR Schmitt circuits (Dokic, 1996).

The transistors ܯ and ܯ provide feedback to effect rapid change of the output voltage and the transfer characteristic has a shape of the hysteresis curve. Hence the circuits in Figs. 21a and 21b are *m*-input NAND and NOR Schmitt circuits, respectively.

#### **6.1 NAND circuit analysis**

Parallel or series transistors can be replaced by one transistor such as the conventional NAND and NOR circuits (Dokic, 1982). In this way, NAND and NOR Schmitt circuits can be replaced by an equivalent Schmitt trigger-inverter (Fig.9) by dc analysis. It will be shown by analyzing an *m*-input NAND Schmitt circuit.

Assume that *n* inputs are active (at ܸ), where ͳ݊݉, and that the other *m-n* inputs are at ܸ. Let the input voltage ܸ increase from zero to ܸ. For Ͳܸ ்ܸு the NAND circuits in Fig.21a can be replaced by the equivalent circuit in Fig.9a. ܯ and *m-n* pMOS transistors at ܸ ൌ ܸ are off. Therefore, the equivalent pMOS transistor ܯ consists of *n* pairs of pMOS transistors ܯ, ܯ <sup>ᇱ</sup> with active inputs. Hence ܯ constant ݇ is given by:

$$k\_{\rm pe} = nk\_p \;/\; 2 \tag{44}$$

Transistor ܯଵ (Fig.9a) needs to be replaced with one equivalent transistor which consists of series nMOS transistors of a conventional NAND gate.

The equivalent constant ݇ of series transistors depends on the number of transistors and position of the first active input (Dokic, 1982). Consequently, for the case of *n* active inputs

Fig. 21. Principle schematics of m-input NAND and NOR Schmitt circuits (Dokic, 1996).

21a and 21b are *m*-input NAND and NOR Schmitt circuits, respectively.

**6.1 NAND circuit analysis** 

pairs of pMOS transistors ܯ, ܯ

by analyzing an *m*-input NAND Schmitt circuit.

series nMOS transistors of a conventional NAND gate.

The transistors ܯ and ܯ provide feedback to effect rapid change of the output voltage and the transfer characteristic has a shape of the hysteresis curve. Hence the circuits in Figs.

Parallel or series transistors can be replaced by one transistor such as the conventional NAND and NOR circuits (Dokic, 1982). In this way, NAND and NOR Schmitt circuits can be replaced by an equivalent Schmitt trigger-inverter (Fig.9) by dc analysis. It will be shown

Assume that *n* inputs are active (at ܸ), where ͳ݊݉, and that the other *m-n* inputs are at ܸ. Let the input voltage ܸ increase from zero to ܸ. For Ͳܸ ்ܸு the NAND circuits in Fig.21a can be replaced by the equivalent circuit in Fig.9a. ܯ and *m-n* pMOS transistors at ܸ ൌ ܸ are off. Therefore, the equivalent pMOS transistor ܯ consists of *n*

Transistor ܯଵ (Fig.9a) needs to be replaced with one equivalent transistor which consists of

The equivalent constant ݇ of series transistors depends on the number of transistors and position of the first active input (Dokic, 1982). Consequently, for the case of *n* active inputs

<sup>ᇱ</sup> with active inputs. Hence ܯ constant ݇ is given by:

/ 2 *pe p k nk* (44)

the transistors ܯଵ, …, ܯ and ܯଵ ܯ ,... , <sup>ᇱ</sup> ᇱ can be replaced by the equivalent transistors ܯ and ܯଵ, respectively, whose constants ݇ are given by:

$$k\_{ne} = k\_{ne1} = \frac{k\_n}{m - p + 1} \tag{45}$$

where *p* marks the position of the first active input (for example, if ൌ͵, the inputs of the transistors ܯଵ, ܯଵ <sup>ᇱ</sup> and ܯଶ, ܯଶ <sup>ᇱ</sup> are at ܸ, and ܯଷ, ܯଷ <sup>ᇱ</sup> are at ܸ). Now, eq.(22) and (26) become, respectively:

$$A\_{ne} = \sqrt{k\_{n\epsilon 1} / k\_{n0}} = A\_n \left( m - p + 1 \right)^{-1/2} \tag{46}$$

$$B\_{1e} = \sqrt{k\_{ne} \,/\, k\_{pe}} = B\_1 \Big[ n \left( m - p + 1 \right) \right]^{-1/2} \tag{47}$$

where ܣ and ܤଵ are given by eqs. (22) and (26), respectively.

From eq. (25), replacing ܣ by ܣ and ܤଵ by ܤଵ we obtain the high threshold voltage of the *m*-inputs NAND Schmitt circuit:

$$V\_{TH} = \frac{V\_{DD} + V\_{tp} + B\_1 \left[ n \left( m - p + 1 \right) \right]^{-1/2} \left[ V\_{DD} + A\_n \left( m - p + 1 \right)^{-1/2} V\_{tn} \right]}{1 + B\_1 \left[ n \left( m - p + 1 \right) \right]^{-1/2} \left[ 1 + A\_n \left( m - p + 1 \right)^{-1/2} \right]} \tag{48}$$

To calculate ்ܸ the circuit in Fig.21a can be replaced by an equivalent one shown in Fig.9b. Namely, ܯଶ is off. ܯଵ, …, ܯ, ܯଵ ܯ ,... , <sup>ᇱ</sup> ᇱ are on and can be replaced by an equivalent one ܯ with the constant ݇:

$$k\_{ne} = k\_n \,/\, (2\,\text{m})\tag{49}$$

ܯଶ is on. ܯ and ܯଵ (݅ ൌ ͳǡ ǥ ǡ ݊) with active input can be replaced by equivalent transistors ܯ and ܯଵ, respectively, with the constants ݇:

$$k\_{pe} = k\_{pe1} = nk\_p \tag{50}$$

The ݇ ratios of ܯଵ to ܯ and ܯ to ܯଵ, respectively, are given by:

$$A\_{p\epsilon} = \sqrt{k\_{p\epsilon 1} \,/\, k\_{p\cdot 2}} = A\_p n^{1/2} \tag{51}$$

$$B\_{2e} = \sqrt{k\_{ne} \,/\, k\_{pe1}} = B\_2 \left( mm \right)^{-1/2} \tag{52}$$

From eq. (36), replacing ܣ by ܣ and ܤଶ by ܤଶ, we obtain the low threshold voltage of the *m*-inputs NAND Schmitt circuits:

$$V\_{TL} = \frac{A\_p n^{1/2} \left(V\_{DD} + V\_{tp}\right) + B\_2 \left(mm\right)^{-1/2} V\_{tn}}{1 + A\_p n^{1/2} + B\_2 \left(mm\right)^{-1/2}}\tag{53}$$

where ܣ and ܤଶ are given by eqs. (27) and (32), respectively.

CMOS and BiCMOS Regenerative Logic Circuits 51

Therefore, the threshold voltages depend on exactly the same parameters as the thresholds of the NAND circuit except that in the NOR circuit, ��� does not depend on the position of

Fig. 23. SPICE values of ��� and ��� for two-input NAND circuit versus ������ = ������ for various numbers and combinations of active inputs at ��� = �� and for optimum

In many applications when NAND and NOR gates are used in an MSI or LSI circuit there is a Schmitt trigger at the external input only. So, for example, a Schmitt trigger action in the clock input of counter provides pulse shaping that allows unlimited clock input pulse rise and fall times. These circuits are made by a conventional NAND or NOR gate and Schmitt trigger on their external input. CMOS NAND or NOR gate and Schmitt trigger action at one input (Fig.24) is a better solution. The Schmitt trigger is an integral part of the gate. The advantages of these circuits, compared with the conventional ones, are: smaller number of

Fig.24 illustrates the principle schemes of the two input NAND and NOR logic circuits with hysteresis when the input �� is active. When the input �� is active only, the transfer characteristic is without hysteresis. These circuits consist of the Schmitt trigger on Fig.8 and one pair of CMOS transistors (�� and ��). Multiple inputs are made in a conventional way

Consider the NAND circuit in Fig.24a. When the input �� only is active, and �� = 1, the transistor �� is off, and �� is on. Then the transfer characteristic is determined by the Schmitt trigger. If the input �� only is active the Schmitt trigger does not operate, so that the transfer characteristic will be the same as that of the CMOS inverter made by the transistors

geometry ratio of nMOS and pMOS transistors, that is at ����� = 2.

**7. CMOS gates with regenerative action at one of inputs** 

transistors, smaller area of the chip and higher switching speed.

(by adding one pair of CMOS transistors to each input).

The NOR gate will be described more fully.

the first active input *p*.

**7.1 Principle schemes** 

�� and ��.

The threshold voltages ்ܸு and ்ܸ depend on supply voltage ܸ, the ratio of the constants ݇Ȁ݇, ݇Ȁ݇, i.e. ݇Ȁ݇, number of inputs *m*, and number of active inputs *n*. Besides, ்ܸு depends on the position of the first active input *p*.

In Fig.22 two-input Schmitt NAND gate and it's transfer characteristics at ܸ ൌ ͷܸ, for ݇ ൌ ݇ ൌ ʹ݇ ൌ ʹ݇, ݅ ൌ Ͳǡ ǥ ǡͶ and ܸ௧ ൌ െܸ௧ ൌ ͳܸ, are shown. The high threshold depends on the number and the combination of active inputs, and the low only on the number of active inputs. The voltage thresholds, as function of channel widths ratio of cascode transistors and of transistors ܯ and ܯ in the circuit of positive feedback loop, are shown in Fig.23.

Fig. 22. Two-input NAND Schmitt trigger (a) and it's transfer characteristic (b) at ܸ ൌ ͷܸ.

#### **6.2 NOR circuit**

As the NOR circuit is obtained from the NAND one through the interchange of the *p*channel and *n*-channel transistors and a power supply polarity change, the previous analysis can be applied analogously to this circuit. In this way we obtain:

$$V\_{TH} = \frac{V\_{DD} + V\_{tp} + B\_1 \left(mm\right)^{1/2} \left(V\_{DD} + A\_n n^{1/2} V\_{tn}\right)}{1 + B\_1 \left(mm\right)^{1/2} \left(1 + A\_n n^{1/2}\right)}\tag{54}$$

$$V\_{TL} = \frac{A\_p \left(m - p + 1\right)^{-1/2} \left(V\_{DD} + V\_{tp}\right) + B\_2 \left[n \left(m - p + 1\right)\right]^{1/2} V\_{tn}}{1 + A\_p \left(m - p + 1\right)^{-1/2} + B\_2 \left[n \left(m - p + 1\right)\right]^{1/2}} \tag{55}$$

The threshold voltages ்ܸு and ்ܸ depend on supply voltage ܸ, the ratio of the constants ݇Ȁ݇, ݇Ȁ݇, i.e. ݇Ȁ݇, number of inputs *m*, and number of active inputs *n*. Besides, ்ܸு

In Fig.22 two-input Schmitt NAND gate and it's transfer characteristics at ܸ ൌ ͷܸ, for ݇ ൌ ݇ ൌ ʹ݇ ൌ ʹ݇, ݅ ൌ Ͳǡ ǥ ǡͶ and ܸ௧ ൌ െܸ௧ ൌ ͳܸ, are shown. The high threshold depends on the number and the combination of active inputs, and the low only on the number of active inputs. The voltage thresholds, as function of channel widths ratio of cascode transistors and of transistors ܯ and ܯ in the circuit of positive feedback loop,

(a) (b) Fig. 22. Two-input NAND Schmitt trigger (a) and it's transfer characteristic (b) at ܸ ൌ ͷܸ.

As the NOR circuit is obtained from the NAND one through the interchange of the *p*channel and *n*-channel transistors and a power supply polarity change, the previous

1

<sup>1</sup> 1 1

1/2 1/2

(54)

(55)

*n*

*DD tp DD n tn*

*B mn A n*

1 1

*p DD tp tn*

 

2

*V V B mn V A n V*

*A m p V V B nm p V <sup>V</sup> A m p B nm p*

11 1

1/2 1/2

1/2 1/2 2 1/2 1/2

analysis can be applied analogously to this circuit. In this way we obtain:

*p*

*TH*

*V*

*TL*

depends on the position of the first active input *p*.

are shown in Fig.23.

**6.2 NOR circuit** 

Therefore, the threshold voltages depend on exactly the same parameters as the thresholds of the NAND circuit except that in the NOR circuit, ��� does not depend on the position of the first active input *p*.

Fig. 23. SPICE values of ��� and ��� for two-input NAND circuit versus ������ = ������ for various numbers and combinations of active inputs at ��� = �� and for optimum geometry ratio of nMOS and pMOS transistors, that is at ����� = 2.

#### **7. CMOS gates with regenerative action at one of inputs**

In many applications when NAND and NOR gates are used in an MSI or LSI circuit there is a Schmitt trigger at the external input only. So, for example, a Schmitt trigger action in the clock input of counter provides pulse shaping that allows unlimited clock input pulse rise and fall times. These circuits are made by a conventional NAND or NOR gate and Schmitt trigger on their external input. CMOS NAND or NOR gate and Schmitt trigger action at one input (Fig.24) is a better solution. The Schmitt trigger is an integral part of the gate. The advantages of these circuits, compared with the conventional ones, are: smaller number of transistors, smaller area of the chip and higher switching speed.

#### **7.1 Principle schemes**

Fig.24 illustrates the principle schemes of the two input NAND and NOR logic circuits with hysteresis when the input �� is active. When the input �� is active only, the transfer characteristic is without hysteresis. These circuits consist of the Schmitt trigger on Fig.8 and one pair of CMOS transistors (�� and ��). Multiple inputs are made in a conventional way (by adding one pair of CMOS transistors to each input).

Consider the NAND circuit in Fig.24a. When the input �� only is active, and �� = 1, the transistor �� is off, and �� is on. Then the transfer characteristic is determined by the Schmitt trigger. If the input �� only is active the Schmitt trigger does not operate, so that the transfer characteristic will be the same as that of the CMOS inverter made by the transistors �� and ��.

The NOR gate will be described more fully.

$$V\_{\mathcal{g}^\text{s}} = V\_i - V\_\mathbf{1} = V\_{\text{tr}} \tag{57}$$

$$V\_{TH} \approx \frac{V\_{DD} + \sqrt{\mathcal{W}\_{n1} / \mathcal{W}\_{n0}} \left(V\_{\hbar} + V\_{BE}\right)}{1 + \sqrt{\mathcal{W}\_{n1} / \mathcal{W}\_{n0}}} \tag{58}$$

$$V\_2 = V\_{BE2} - V\_{tp0} - \sqrt{V\_{p1} \left/ V V\_{p0}} \left( V\_{DD} + V\_{tp1} + V\_i \right) \tag{59}$$

$$V\_i = V\_2 + V\_{tp} \tag{60}$$

$$V\_{TL} \approx \frac{\sqrt{\mathcal{W}\_{p1} / \mathcal{W}\_{p0}} \left(V\_{DD} + V\_{tn}\right) + V\_{BE}}{1 + \sqrt{\mathcal{W}\_{p1} / \mathcal{W}\_{p0}}} \tag{61}$$

CMOS and BiCMOS Regenerative Logic Circuits 57

depend on, like those of the standard ones, the number of inputs, number of active inputs

Al-Sarawi, S.F. Low Power Schmitt Trigger Circuits, *Electronics Letter*, vol. 38, 29th August

Arrabi, S. A 90nm CMOS Data Flow Processor Using Fine Grained DVS for Energy Efficient

Bundalo, Z., Dokic, B., Non-inverting Regenerative CMOS Logic Circuits, *Microelectronics* 

Dokic, B., CMOS Schmitt triggers, *IEE Procedings - Part G*, vol. 131, No. 5, October 1984, pp

Dokic, B., CMOS Regenerative Logic Circuits, *Microelectronics Journal* vol. 14, No.5, 1983., pp

Dokic, B., CMOS NAND and NOR Schmitt Circuits, *Microelectronics Journal* 27, No. 8,

Dokic, B. et al, CMOS Gates with Regenerative action at One of Inputs, *Microelectronics* 

Dokic, B. and Bundalo, Z., Regenerative Logic Circuits with CMOS Transistors, *Int. J-*

Dokic, B., Influence of Series and Parallel Transistors on DC Characteristics of CMOS Logic

Dokic, B., Three-State BiCMOS Buffers with Positive Feedback, *Proc. 20th Int. Conf. on* 

Hard, A. and Voinigesku, S. P., A 1GHz Bandwidth Low-Pass ∆∑ ADC with 20-50 GHz

Chien, H., Switch-controllable Dual-hysteresis Mode Bistable Multivibrator Employing

Katyal, V. et al, Adjustable Hysteresis CMOS Schmitt Triggers, *Circuits and Systems*, 2008

Lo, Y. et al, Current-input OTRA Schmitt Trigger with Dual Hysteresis Modes, *Int. J. Circ.* 

Pengfei, L. et al., A Delay-locked Loop Synchronization Scheme for High-Frequency

Wang, C. et al., A 570kbps ASK Demodulator without External Capacitors for Low Frequency Wireless Bio-implants, *Microelectronics Journal* 39, 2008, pp 130-136.

Adjustable Sampling Rate, *IEEE Journal of SSC*, vol. 44 No. 5, May 2009, pp 1401-

Single Differential Voltage Current Converter, *Microelectronics Journal 42*, 2011, pp

Multiphase Hysteretic DC-DC Converters, *IEEE Journal of SSC* vol. 44, No. 11,

Circuits, *Microelectronics Journal*, vol. 13, No. 2, 1982, pp 25-30.

*Microelectronics (MIEL '95)*, vol. 2, 1995, pp 509-511.

ISCAS 008. IEEE Int. Symposium on, pp 1938-1941.

*Theor. Appl. 2010*, vol. 38, pp 739-746.

November 2009, pp 1401-1414.

Operation from 0.3V to 1.2V, *IEEE SSC Magazine*, Spring 2011, vol. 3, No. 2, pp 52-

197-202. The paper has been copied to Express Information PEAVT, No. 21, 1985,

21-30. The paper has been copied to Express Information PEAVT, No. 29, 1984, pp

and the position of the first active input.

2002, pp 1009-1010.

*Journal*, vol. 16, No. 5, 1985, pp 5-17.

pp 4-14, The Acadamy of Sciences of USSR.

4-14, The Acadamy of Sciences of USSR.

*Journal*, vol. 19, No. 3, 1988, pp 17-20.

*Electronics*, 1985, vol. 58, No. 6, pp 907-920.

November, 1996, pp 757-766.

**10. References** 

58.

1414.

745-753.

Limitation of application of the circuit in Fig.27 at lower supply voltages is it's decrea-sed logic amplitude of the output voltage ��� = ��� � 2���. The low voltage Schmitt trigger is shown in Fig.29. Transistors ��� and ��� through the inverter � hold output voltages at �� = ��� and �� = 0.

Transistor ��� delivers another improvement. It removes deformation from the transfer characteristic of the standard circuit before the change from high to low level (Fig.28). This deformation is a consequence of the knee-effect of characteristic of bipolar transistor ��. In Fig.30 SPICE analysis of the output of the low voltage Schmitt trigger designed in 0���� BiCMOS process for equal constants �� and �� of all nMOS and pMOS transistors is shown.

Fig. 30. The output of the low voltage Schmitt trigger to a triangular input at ��� = 2���.

#### **9. Conclusion**

Schmitt logic circuits are produced as independent integrated circuits as well as input circuits of standard MSI/VLSI and ASIC integrated circuits. The author expects that the overview of CMOS and BiCMOS Schmitt triggers will be useful to both engineers who design digital integrated circuits and to those who design digital systems with integrated circuits.

Common characteristics of the described solutions are:


Overall, the Schmitt inverter (section 3) yields the best characteristic. Its operating stability does not depend on transistor geometries ratio, nor on tolerance of technology. Symmetry of all transistors is optimum, concerning noise immunity and propagation time. While in standard circuits every input is joined by one pair, in Schmitt logic circuits every input is joined by two pairs of CMOS transistors. Taking into account that one pair closes positive feedback loop, it follows that *m*-input NAND and NOR Schmitt circuits are comprised of 2*m*+1 pairs of CMOS transistors. Transfer characteristic of NAND and NOR Schmitt circuits depend on, like those of the standard ones, the number of inputs, number of active inputs and the position of the first active input.

#### **10. References**

56 Cutting Edge Research in New Technologies

Limitation of application of the circuit in Fig.27 at lower supply voltages is it's decrea-sed logic amplitude of the output voltage ��� = ��� � 2���. The low voltage Schmitt trigger is shown in Fig.29. Transistors ��� and ��� through the inverter � hold output voltages at

Transistor ��� delivers another improvement. It removes deformation from the transfer characteristic of the standard circuit before the change from high to low level (Fig.28). This deformation is a consequence of the knee-effect of characteristic of bipolar transistor ��. In Fig.30 SPICE analysis of the output of the low voltage Schmitt trigger designed in 0���� BiCMOS process for equal constants �� and �� of all nMOS and pMOS transistors is shown.

Fig. 30. The output of the low voltage Schmitt trigger to a triangular input at ��� = 2���.

Common characteristics of the described solutions are:

Schmitt logic circuits are produced as independent integrated circuits as well as input circuits of standard MSI/VLSI and ASIC integrated circuits. The author expects that the overview of CMOS and BiCMOS Schmitt triggers will be useful to both engineers who design digital integrated circuits and to those who design digital systems with integrated

threshold voltages ��� and ��� are almost symmetric around voltage ���/2, so the

Overall, the Schmitt inverter (section 3) yields the best characteristic. Its operating stability does not depend on transistor geometries ratio, nor on tolerance of technology. Symmetry of all transistors is optimum, concerning noise immunity and propagation time. While in standard circuits every input is joined by one pair, in Schmitt logic circuits every input is joined by two pairs of CMOS transistors. Taking into account that one pair closes positive feedback loop, it follows that *m*-input NAND and NOR Schmitt circuits are comprised of 2*m*+1 pairs of CMOS transistors. Transfer characteristic of NAND and NOR Schmitt circuits

transfer characteristics are optimum, concerning noise immunity; basic parameters (���, ��� and ��) do not depend on technology.

�� = ��� and �� = 0.

**9. Conclusion** 

circuits.


**3** 

*Croatia* 

*University of Zagreb,* 

*Faculty of Textile Technology,* 

**A New Pre-Wet Sizing Process – Yes or No?** 

As one of the most complex steps in fabric production, sizing plays a very important role in the weaving process. The primary purpose of the sizing process is to obtain the warp threads that can successfully be woven without major damages which occur during the yarn passage through sliding metal parts of the weaving machine (Lord, 2003). It applies to the improvement of physical and mechanical parameters of warp threads, primarily to increase strength and abrasion resistance and thus to reduce the number of warp breaks to a minimum in order to achieve the maximum efficiency of weaving machines and energy savings. Also, the goal of sizing is to keep the fibers in the yarn in a position where they were before sizing, with minimal yarn deformations during weaving. The success of the weaving process depends on the complexity of several factors including the characteristics of the desired material, the sizing process, the sizing ingredients and yarn properties, but also the extensive knowledge of a textile technologist (chemistry, rheology, electronics, mechanical engineering, physics, mechanics, mathematics, etc...), which makes this process more difficult and more important for the overall process of making woven fabric (Adanur, 2001). Today's achievements in all engineering branches enable an exceptional progress of the sizing processes to achieve a very high quality of sizing that meets the needs of today's modern weaving. However, the sizing costs, despite the complete automation of the regulation and control of the most important sizing process parameters, are still very high. Their reduction is possible by reducing the consumption of sizing agents and energy, as well as by modernization and development of machinery and technology, and all without consequence on the quality of the sized yarn. The choice of sizing agents plays the most important role in meeting all the requirements placed on the sizing process and sized yarn as well as optimizing and keeping the sizing process conditions and the size pick-up constant (Kovačević et al., 2006). Even today the optimization of the size pick-up applied to the yarn presents a major problem in the sizing process, despite the high degree of automation and high quality sizing agents. Influential parameters in the optimization of size pick-up are defined with the substance balance that enters and exits the size box (Equation 1). The requirement of keeping the size pick-up optimized and constant can be achieved by continuous measuring and keeping temperature and yarn moisture, size concentration in the size box constant, as well as automatic regulation of squeezing force and sizing speed

**1. Introduction** 

(Pleva & Rieger, 1992; Soliman, 1995).

Ivana Gudlin Schwarz and Stana Kovačević

*Department of Textile Design and Management,* 


## **A New Pre-Wet Sizing Process – Yes or No?**

Ivana Gudlin Schwarz and Stana Kovačević

*University of Zagreb, Faculty of Textile Technology, Department of Textile Design and Management, Croatia* 

#### **1. Introduction**

58 Cutting Edge Research in New Technologies

Young, F., A High-speed Differential CMOS Schmitt Trigger with Regenerative Current

Zou, Z., et al., A Novel Schmitt Trigger with Low Temperature Coefficient, *Circuits and Systems*, 2008, APCCAS 2008, IEEE Asia Pacific Conference on, pp 1398-1401.

121-127.

Feedback and Adjustable Hysteresis, *Analog Integrated Circuits Process*, 2010, 63: pp

As one of the most complex steps in fabric production, sizing plays a very important role in the weaving process. The primary purpose of the sizing process is to obtain the warp threads that can successfully be woven without major damages which occur during the yarn passage through sliding metal parts of the weaving machine (Lord, 2003). It applies to the improvement of physical and mechanical parameters of warp threads, primarily to increase strength and abrasion resistance and thus to reduce the number of warp breaks to a minimum in order to achieve the maximum efficiency of weaving machines and energy savings. Also, the goal of sizing is to keep the fibers in the yarn in a position where they were before sizing, with minimal yarn deformations during weaving. The success of the weaving process depends on the complexity of several factors including the characteristics of the desired material, the sizing process, the sizing ingredients and yarn properties, but also the extensive knowledge of a textile technologist (chemistry, rheology, electronics, mechanical engineering, physics, mechanics, mathematics, etc...), which makes this process more difficult and more important for the overall process of making woven fabric (Adanur, 2001). Today's achievements in all engineering branches enable an exceptional progress of the sizing processes to achieve a very high quality of sizing that meets the needs of today's modern weaving. However, the sizing costs, despite the complete automation of the regulation and control of the most important sizing process parameters, are still very high. Their reduction is possible by reducing the consumption of sizing agents and energy, as well as by modernization and development of machinery and technology, and all without consequence on the quality of the sized yarn. The choice of sizing agents plays the most important role in meeting all the requirements placed on the sizing process and sized yarn as well as optimizing and keeping the sizing process conditions and the size pick-up constant (Kovačević et al., 2006). Even today the optimization of the size pick-up applied to the yarn presents a major problem in the sizing process, despite the high degree of automation and high quality sizing agents. Influential parameters in the optimization of size pick-up are defined with the substance balance that enters and exits the size box (Equation 1). The requirement of keeping the size pick-up optimized and constant can be achieved by continuous measuring and keeping temperature and yarn moisture, size concentration in the size box constant, as well as automatic regulation of squeezing force and sizing speed (Pleva & Rieger, 1992; Soliman, 1995).

A New Pre-Wet Sizing Process – Yes or No? 61

the pre-wetting box and size temperature in the size box with integrated heaters and thermostats, which indirectly warm up the water and size through the walls of the boxes. Thread tension was measured during the sizing process at the box entry, while warp moisture was measured at all important places: at the box entry, between two boxes - the pre-wetting box and size box, at the size box exit and after the dryer. Drying the sized yarn is preformed by contact, moving it across the two heated cylinders of the contact dryer. It is also possible to regulate and keep the sizing speed constant using the winder of the sized

Fig. 1. Laboratory sizing machine (constructed on Faculty of Textile Technology, University of Zagreb, Croatia): 1 - creel for cross wound bobbins, 2 – moisture contact measuring device, 3 - thread tension measuring device, 4 - pre-wetting box, 5 - size box, 6 - rollers for immersing yarn into water and size, 7 - rollers for water and size squeezing, 8 - regulation of the pressure of the last squeezing roller, 9 - contact dryer, 10 - winder of the sized yarn

Sizing is a process which can be carried out on yarns of different raw material composition, and in this case it was 100% cotton carded ring spun yarn with a nominal count of 20 tex

> Parameters Yarn Ttn (tex) 20.00 Ttr (tex) 18.55

> T (twists/m) x 913.04

H (number of protruding fibers) x 20428.00

A (number of cycle) x 80.76

F (cN) x 281.69

CV (%) 16.01 Thin places 11.50 Thick places 95.50 Neps 208.00

CV 5.20

CV 2.30

CV 9.39

CV 6.78

**3. Testing methods and materials (yarn and size)** 

and with certain properties shown in Table 1.

Parameters of unevenness (400 m/min)

and dried yarn, as well as the speed regulator (Gudlin Schwarz et al. 2010, 2011).

$$\text{Sp} = \frac{\text{W}\_{\text{Sp}} - \text{W}\_{\text{H}}}{\frac{100}{\text{C}} - 1 - \frac{\text{W}\_{\text{Sp}}}{100}} \quad \text{(\%)}\tag{1}$$

Where: Sp – size pick-up, WH - warp moisture at the box entry (%),WSp - warp moisture at the box exit (%), C - size concentration in the box (%)

Both, sizing conditions and yarn parameters affect yarn size pick-up. If the yarn is in some sections "more closed" with fewer interspaces among fibers or with more twists, the absorption of size in this section will be lower, resulting in a lower size pick-up on the yarn, despite constant sizing conditions. During the sizing process two types of forces are needed to be overcome: the forces of surface tension (wetting) and the forces of diffusion (Goswami, 2004). Penetration of the liquid (wetting) occurs in two phases:


Also a big unknown in size pick-up optimization is the sizing of wet warp, as well as the entire pre-wet sizing process. All previous knowledge of pre-wet sizing points to the obtainment of outstanding results, relevant physical-mechanical properties, reduction of consumption in sizing agents and energy, and an increase in weaving productivity (Hyrenbach, 2002; Rozelle, 1999, 2001; Sherrer, 2000). Therefore, this chapter aims to bring knowledge of the pre-wet sizing process by its analysis and by making a comparative analysis of the standard sizing process. The goal is to prove that there are a number of justified reasons for a new technological process, and to highlight the advantages and also disadvantages and possible improvements of better physical-mechanical properties of the yarn, reduction of size and energy consumption and increase in weaving productivity (Gudlin Schwarz et al. 2010, 2011).

#### **2. Sizing machine**

The pre-wet sizing process is still a rather unexplored area and not confirmed by scientific research. The most important reason why this research area has remained unexplored and with such a poor representation of this topic in scientific work is aggravated laboratory samples processing. Thanks to the laboratory sizing machine (Fig. 1) designed and constructed at the Faculty of Textile Technology, University of Zagreb, Croatia (whose segments are sizing box and dryer – which represents two consensual patents No.: PK20070247 and PK20070248, registered at the Croatian State Intellectual Property Office) both sizing processes - standard sizing process and pre-wet sizing process were able to be carried out. It consists of a creel for cross wound bobbins, with the possibility of tension regulation, two boxes – a box for pre-wetting with hot water and a size box, and a dryer. The pre-wetting box consists of a pair of immersion rollers and a pair of rollers for squeezing out excess water. The size box consists of a working box with two pairs of immersion rollers and two pairs of rollers for size squeezing, as well as a pre-box that allows to keep size levels in the working box constant, namely continuous size circulation from the working box to the pre-box with natural flow, and from the pre-box to the working box using a pump. During the sizing process it is possible to keep water temperature constant in

W W Sp % 100 W 1 C 100

Where: Sp – size pick-up, WH - warp moisture at the box entry (%),WSp - warp moisture at

Both, sizing conditions and yarn parameters affect yarn size pick-up. If the yarn is in some sections "more closed" with fewer interspaces among fibers or with more twists, the absorption of size in this section will be lower, resulting in a lower size pick-up on the yarn, despite constant sizing conditions. During the sizing process two types of forces are needed to be overcome: the forces of surface tension (wetting) and the forces of diffusion (Goswami,

1. penetration of the liquid into the capillary spaces between the fibers in the yarn (during which two forces have to be overcome - the difference between the pressure of enclosed air and surrounding liquid, and tension forces of the interface between fiber and water) 2. penetration of liquids into the fiber, i.e. extrusion of air bubbles and filling those spaces

Also a big unknown in size pick-up optimization is the sizing of wet warp, as well as the entire pre-wet sizing process. All previous knowledge of pre-wet sizing points to the obtainment of outstanding results, relevant physical-mechanical properties, reduction of consumption in sizing agents and energy, and an increase in weaving productivity (Hyrenbach, 2002; Rozelle, 1999, 2001; Sherrer, 2000). Therefore, this chapter aims to bring knowledge of the pre-wet sizing process by its analysis and by making a comparative analysis of the standard sizing process. The goal is to prove that there are a number of justified reasons for a new technological process, and to highlight the advantages and also disadvantages and possible improvements of better physical-mechanical properties of the yarn, reduction of size and energy consumption and increase in weaving productivity

The pre-wet sizing process is still a rather unexplored area and not confirmed by scientific research. The most important reason why this research area has remained unexplored and with such a poor representation of this topic in scientific work is aggravated laboratory samples processing. Thanks to the laboratory sizing machine (Fig. 1) designed and constructed at the Faculty of Textile Technology, University of Zagreb, Croatia (whose segments are sizing box and dryer – which represents two consensual patents No.: PK20070247 and PK20070248, registered at the Croatian State Intellectual Property Office) both sizing processes - standard sizing process and pre-wet sizing process were able to be carried out. It consists of a creel for cross wound bobbins, with the possibility of tension regulation, two boxes – a box for pre-wetting with hot water and a size box, and a dryer. The pre-wetting box consists of a pair of immersion rollers and a pair of rollers for squeezing out excess water. The size box consists of a working box with two pairs of immersion rollers and two pairs of rollers for size squeezing, as well as a pre-box that allows to keep size levels in the working box constant, namely continuous size circulation from the working box to the pre-box with natural flow, and from the pre-box to the working box using a pump. During the sizing process it is possible to keep water temperature constant in

the box exit (%), C - size concentration in the box (%)

with a liquid.

(Gudlin Schwarz et al. 2010, 2011).

**2. Sizing machine** 

2004). Penetration of the liquid (wetting) occurs in two phases:

 Sp H Sp

(1)

the pre-wetting box and size temperature in the size box with integrated heaters and thermostats, which indirectly warm up the water and size through the walls of the boxes. Thread tension was measured during the sizing process at the box entry, while warp moisture was measured at all important places: at the box entry, between two boxes - the pre-wetting box and size box, at the size box exit and after the dryer. Drying the sized yarn is preformed by contact, moving it across the two heated cylinders of the contact dryer. It is also possible to regulate and keep the sizing speed constant using the winder of the sized and dried yarn, as well as the speed regulator (Gudlin Schwarz et al. 2010, 2011).

Fig. 1. Laboratory sizing machine (constructed on Faculty of Textile Technology, University of Zagreb, Croatia): 1 - creel for cross wound bobbins, 2 – moisture contact measuring device, 3 - thread tension measuring device, 4 - pre-wetting box, 5 - size box, 6 - rollers for immersing yarn into water and size, 7 - rollers for water and size squeezing, 8 - regulation of the pressure of the last squeezing roller, 9 - contact dryer, 10 - winder of the sized yarn

#### **3. Testing methods and materials (yarn and size)**

Sizing is a process which can be carried out on yarns of different raw material composition, and in this case it was 100% cotton carded ring spun yarn with a nominal count of 20 tex and with certain properties shown in Table 1.


A New Pre-Wet Sizing Process – Yes or No? 63

To test the samples before and after sizing, standardized methods were used. Thus, the real count of yarn was tested according to HRN ISO 2060:2003, while yarn unevenness was tested on an Unevenness tester 80, type B, Keisokki Company. The breaking properties (breaking force, elongation at break, work to rapture and tenacity of yarns) were tested on a Statimat M made by Textechno according to ISO 2062. Yarn hairiness was tested before and after sizing by recording the fibers protruding from the yarn structure using a Zweigle G 565 hairiness meter according to ASTM D 5674-01, while the twists were tested by means of a MesdanLab Twist tester according to ISO 17202. Abrasion resistance tests were performed on a Zweigle G 551 abrasion tester before and after sizing, where each of 20 types of thread loaded with a weight of 20g was subjected to the abrasion process until thread breakage. The movement of the cylinder coated with emery paper (fineness 600): left - right and its rotation around its axis achieves certain abrasion intensity in the yarn and emery paper. During the process the yarn weakens, and at the moment when the mass of the weights hung on the yarn overcomes the yarn strength, a break occurs, and the number of roller movements until breaking the yarn is

The determination of the size pick-up can be performed in several ways, but for the purposes of this study the gravimetric method was used. The implementation process of this method is as follows: before sizing the samples were dried to absolutely dry, after which they were weighed, and then returned to climatic conditions and sized; after sizing the samples were again dried to absolutely dry and weighed (Kovačević et al., 2002). The

*g g g*

Sp 100 %

The most prominent mechanical properties of yarn are primarily breaking properties, which include: breaking force, elongation at break, work to rapture and tenacity. These parameters show us some of the most important yarn characteristics for the weaving process, and the positive impact of sizing to those properties is also one of the most relevant role of the sizing process. Sizing pursued to a greater increase in breaking force and at the same time in a less decrease in elongation at break, which in turn depends on the size pick-up and size distribution in the yarn (Gudlin Schwarz et al. 2011; Kovačević &

Figure 2 shows an F-E diagram of unsized yarn and yarns sized with two different size recipes R1 and R2, where the differences in force (F) and elongation (ε) between the yarns

The values of breaking force of the tested samples are shown in Figure 3a, where a very small difference between the sized yarns occurs, with an average increase of almost 40%

U G G

G

S U

(2)

amount of size pick-up is calculated using Equation 2:

**4. Testing methods and materials (yarn and size)** 

subjected to different processing methods can be clearly seen.

Where: Sp (%) – amount of size pick-up GS (g) – mass of absolutely dry sized yarn GU (g) – mass of absolutely dry unsized yarn

**4.1 Yarn breaking properties** 

Penava, 2004).

recorded.


Table 1. Properties of unsized yarn; where : Ttn – nominal count (tex), Ttr – real count (tex), T – twist (twist/m), H – hairiness (No. of protruding fibers longer of 1mm from the yarn surface), A – abrasion resistance (No. of cycles), F – breaking force (cN), ε – elongation at break (%), W – work to rapture (cN×tex), σ – tenacity (cN/tex), CV – coefficient of variation, x - mean value

As it was mentioned above, for a successful sizing process right choice of sizing agents and size preparation are of great importance, depending on yarn (fiber) type, origin of the sizing agent, different sizing auxiliaries, and the requirements of the sizing process itself. (Vassallo, 2005; Zhu, 2003). In the presented example and based on these needs, two different recipes with different concentrations were used in both sizing processes, as shown in Table 2.


Table 2. Characteristics of the sizing agents and auxiliaries, and size recipes

Sizing conditions are exactly defined and held constant during both sizing processes, and are shown in Table 3.


Table 3. Sizing conditions

Parameters Yarn Ε (%) x 3.83

W (cNx cm) x 292.75

(cN/tex) x 15.19

Table 1. Properties of unsized yarn; where : Ttn – nominal count (tex), Ttr – real count (tex), T – twist (twist/m), H – hairiness (No. of protruding fibers longer of 1mm from the yarn surface), A – abrasion resistance (No. of cycles), F – breaking force (cN), ε – elongation at break (%), W – work to rapture (cN×tex), σ – tenacity (cN/tex), CV – coefficient of variation,

As it was mentioned above, for a successful sizing process right choice of sizing agents and size preparation are of great importance, depending on yarn (fiber) type, origin of the sizing agent, different sizing auxiliaries, and the requirements of the sizing process itself. (Vassallo, 2005; Zhu, 2003). In the presented example and based on these needs, two different recipes with different concentrations were used in both sizing processes, as

Compounds Recipe Concentration

Sizing conditions are exactly defined and held constant during both sizing processes, and

Temperature of water in the pre-wetting box 65°C Size temperature in the size box 75°C Sizing speed 3 m/min Pressure on the last pair of the rollers for squeezing excess size 19.1 N/cm2 Temperature on the cylinders of the contact dryer 140°C Output moisture 6 %

Table 2. Characteristics of the sizing agents and auxiliaries, and size recipes

Thread tension between creel for cross wound bobbins and pre-wetting

2. Sizing agent based on polyvinilalcochol (PVA) 3. Sizing auxiliaries composed of natural fats and waxes with a specific emulsifier system

x - mean value

shown in Table 2.

1. Water

are shown in Table 3.

Sizing condition

Table 3. Sizing conditions

box

CV 7.26

CV 13.08

CV 6.78

Recipe 1 - R 1 7.5%

Recipe 2 - R 2 5.0%

40 cN

To test the samples before and after sizing, standardized methods were used. Thus, the real count of yarn was tested according to HRN ISO 2060:2003, while yarn unevenness was tested on an Unevenness tester 80, type B, Keisokki Company. The breaking properties (breaking force, elongation at break, work to rapture and tenacity of yarns) were tested on a Statimat M made by Textechno according to ISO 2062. Yarn hairiness was tested before and after sizing by recording the fibers protruding from the yarn structure using a Zweigle G 565 hairiness meter according to ASTM D 5674-01, while the twists were tested by means of a MesdanLab Twist tester according to ISO 17202. Abrasion resistance tests were performed on a Zweigle G 551 abrasion tester before and after sizing, where each of 20 types of thread loaded with a weight of 20g was subjected to the abrasion process until thread breakage. The movement of the cylinder coated with emery paper (fineness 600): left - right and its rotation around its axis achieves certain abrasion intensity in the yarn and emery paper. During the process the yarn weakens, and at the moment when the mass of the weights hung on the yarn overcomes the yarn strength, a break occurs, and the number of roller movements until breaking the yarn is recorded.

The determination of the size pick-up can be performed in several ways, but for the purposes of this study the gravimetric method was used. The implementation process of this method is as follows: before sizing the samples were dried to absolutely dry, after which they were weighed, and then returned to climatic conditions and sized; after sizing the samples were again dried to absolutely dry and weighed (Kovačević et al., 2002). The amount of size pick-up is calculated using Equation 2:

$$\mathrm{Sp} = \frac{\mathrm{G\_S}(\mathrm{g}) - \mathrm{G\_U}(\mathrm{g})}{\mathrm{G\_U}(\mathrm{g})} \cdot 100 \text{ (\%)}\tag{2}$$

Where: Sp (%) – amount of size pick-up GS (g) – mass of absolutely dry sized yarn GU (g) – mass of absolutely dry unsized yarn

#### **4. Testing methods and materials (yarn and size)**

#### **4.1 Yarn breaking properties**

The most prominent mechanical properties of yarn are primarily breaking properties, which include: breaking force, elongation at break, work to rapture and tenacity. These parameters show us some of the most important yarn characteristics for the weaving process, and the positive impact of sizing to those properties is also one of the most relevant role of the sizing process. Sizing pursued to a greater increase in breaking force and at the same time in a less decrease in elongation at break, which in turn depends on the size pick-up and size distribution in the yarn (Gudlin Schwarz et al. 2011; Kovačević & Penava, 2004).

Figure 2 shows an F-E diagram of unsized yarn and yarns sized with two different size recipes R1 and R2, where the differences in force (F) and elongation (ε) between the yarns subjected to different processing methods can be clearly seen.

The values of breaking force of the tested samples are shown in Figure 3a, where a very small difference between the sized yarns occurs, with an average increase of almost 40%

A New Pre-Wet Sizing Process – Yes or No? 65

2,6 2,8 3,0 3,2 3,4 3,6 3,8 4,0

(c) (d) Fig. 3. Diagrams of breaking properties of the unsized yarn and the yarns sized with recipe 1

Yarn hairiness and abrasion resistance are parameters which are extremely important for the weaving process, which are greatly improved by a successful sizing process. Hairiness, i.e. the number of protruding fibers, is reduced by sizing, and the abrasion resistance is increased, which affects the reduction in friction resulting from the thread passing through the metal elements of the weaving machine and, therefore, the number of thread breaks in

Figure 4a shows the values of the tested hairiness for the unsized yarn and the yarns sized with both processes and with both recipes. Yarn hairiness reduction sized with both processes is very similar, and amounts to 78% for the yarns sized with the standard process, and 81% for the yarns sized with the pre-wet sizing process compared to the unsized yarn. The value diagram of the abrasion resistance of the unsized yarn and the yarn sized with both procedures and with both recipes is shown in Figure 4b. It is interesting that the yarns sized with both processes but with a higher concentration of size (R1) show good results in terms of increasing the abrasion resistance compared to the unsized yarn. The yarns sized with the standard process recorded an increase by even 68%, while the yarns sized with the pre-wet process showed almost half an increase by only 36%. Regarding the samples sized with a smaller size concentration (R2) with both processes, a notable decrease in abrasion resistance compared to the unsized yarn is recorded, namely by 4% for the yarns sized with the standard process, and by 14% for the yarn sized with pre-wet process. This phenomenon, in spite of the size pick-up which strengthens the yarn, is attributed to the

**T (cN/tex)**

**ε (%)**

(a) (b)

U R1 / S R1 / W R2 / S R2 / W

U R1 / S R1 / W R2 / S R2 / W

**W (cNxcm)**

**F (cN)**

U R1 / S R1 / W R2 / S R2 / W

U R1 / S R1 / W R2 / S R2 / W

**4.2 Yarn hairiness and abrasion resistance** 

the weaving process (Gudlin Schwarz, 2011).

and 2 by the standard sizing process and the pre-wetting sizing process

compared to the unsized yarn. The only yarn that shows a deviation from the others in the form of a small increase in breaking force (of only 32%) is the yarn sized with R2 and the pre-wet sizing process.

Fig. 2. F-E diagram of unsized yarn and yarns sized with recipes 1 and 2, by standard sizing process and pre-wetting sizing process; where: U - unsized yarn, R1 - yarn sized with recipe 1, R2 - yarn sized with recipe 2; S – standard sizing process, W – pre-wetting sizing process

Generally, by using the sizing process the elongation at break reduces and therefore represents the disadvantage of this process. The values of the elongation at break of the tested yarns are shown in Figure 3b, where the results are divided into two groups with almost identical values: the yarns sized with the standard process, which shows a decrease of almost 20%, and those sized with the pre-wet sizing process, with a decrease of almost 25%, compared to the unsized yarn.

In Figure 3c, which shows the results of work to rapture, are presented the uniform values within one sizing process, where a larger increase by 12% is recorded in the yarns sized with the standard process than in the yarns sized with the pre-wet sizing process, where the values increase by 4% for the yarn sized with R1, while the yarn sized with R2 records even a slight drop by only 3% compared to the unsized yarn.

Tenacity is a parameter that brings into relation yarn finesses and force, and the values obtained are shown in Figure 3d. The values of the sized yarns are quite consistent with an increase by nearly 33% for the yarn sized with the standard process, and by 31% for the yarn sized with the pre-wet sizing process.

compared to the unsized yarn. The only yarn that shows a deviation from the others in the form of a small increase in breaking force (of only 32%) is the yarn sized with R2 and the

U R1 / S R1 / M R2 / W R2 / W

0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 **ε (%)**

Fig. 2. F-E diagram of unsized yarn and yarns sized with recipes 1 and 2, by standard sizing process and pre-wetting sizing process; where: U - unsized yarn, R1 - yarn sized with recipe 1, R2 - yarn sized with recipe 2; S – standard sizing process, W – pre-wetting

Generally, by using the sizing process the elongation at break reduces and therefore represents the disadvantage of this process. The values of the elongation at break of the tested yarns are shown in Figure 3b, where the results are divided into two groups with almost identical values: the yarns sized with the standard process, which shows a decrease of almost 20%, and those sized with the pre-wet sizing process, with a decrease of almost

In Figure 3c, which shows the results of work to rapture, are presented the uniform values within one sizing process, where a larger increase by 12% is recorded in the yarns sized with the standard process than in the yarns sized with the pre-wet sizing process, where the values increase by 4% for the yarn sized with R1, while the yarn sized with R2 records even

Tenacity is a parameter that brings into relation yarn finesses and force, and the values obtained are shown in Figure 3d. The values of the sized yarns are quite consistent with an increase by nearly 33% for the yarn sized with the standard process, and by 31% for the yarn

pre-wet sizing process.

0,0

sizing process

25%, compared to the unsized yarn.

sized with the pre-wet sizing process.

a slight drop by only 3% compared to the unsized yarn.

0,5

1,0

1,5

2,0

2,5

**F (cN)**

3,0

3,5

4,0

4,5

Fig. 3. Diagrams of breaking properties of the unsized yarn and the yarns sized with recipe 1 and 2 by the standard sizing process and the pre-wetting sizing process

#### **4.2 Yarn hairiness and abrasion resistance**

Yarn hairiness and abrasion resistance are parameters which are extremely important for the weaving process, which are greatly improved by a successful sizing process. Hairiness, i.e. the number of protruding fibers, is reduced by sizing, and the abrasion resistance is increased, which affects the reduction in friction resulting from the thread passing through the metal elements of the weaving machine and, therefore, the number of thread breaks in the weaving process (Gudlin Schwarz, 2011).

Figure 4a shows the values of the tested hairiness for the unsized yarn and the yarns sized with both processes and with both recipes. Yarn hairiness reduction sized with both processes is very similar, and amounts to 78% for the yarns sized with the standard process, and 81% for the yarns sized with the pre-wet sizing process compared to the unsized yarn.

The value diagram of the abrasion resistance of the unsized yarn and the yarn sized with both procedures and with both recipes is shown in Figure 4b. It is interesting that the yarns sized with both processes but with a higher concentration of size (R1) show good results in terms of increasing the abrasion resistance compared to the unsized yarn. The yarns sized with the standard process recorded an increase by even 68%, while the yarns sized with the pre-wet process showed almost half an increase by only 36%. Regarding the samples sized with a smaller size concentration (R2) with both processes, a notable decrease in abrasion resistance compared to the unsized yarn is recorded, namely by 4% for the yarns sized with the standard process, and by 14% for the yarn sized with pre-wet process. This phenomenon, in spite of the size pick-up which strengthens the yarn, is attributed to the

A New Pre-Wet Sizing Process – Yes or No? 67

sensitivity, and thus susceptibility to extension and deformation. Yarn unevenness also affects extension properties to a great extent, where thin and thick places represent weaker yarn parts which are more sensitive to tension, especially in the wet state (Gudlin Schwarz, 2010). Similarly, the values shown in Figure 5 indicate that the elongation is higher during

R1 / S R1 / W R2 / S R2 / W

Fig. 5. Diagram of extension of yarns sized with recipe 1 and 2 by standard sizing process

As stated in the introduction, optimizing the size pick-up to achieve the maximum utilization of the sizing process represent the biggest challenge in the whole process (Kovačević et al., 2002). The obtained results indicate that the amount of size pick-up and its

Figure 6 shows the value of the amount of size pick-up on the yarn. A small difference between the yarns sized with R1 by both sizing processes is easily observable. Significant differences are evident in the yarn sized with R2 in both processes, where an almost equal difference in reduced size pick-up (an average of 50%) between the standard sizing process

The yarn sized with lower size concentrations showed very good results in all important properties in terms of no significant deviations (in spite of a lower amount of size pick-up on the yarn) in relation to the yarn sized with a higher size concentration. This phenomenon is particularly interesting for the yarn sized with the pre-wet sizing process, where the amount of size pick-up on the yarn is considerably lower (due to water filling the interior of the yarn and the different distribution of the size pick-up on the yarn), than on the yarn sized with the standard process, indicating a reduced consumption of sizing agents and

distribution determine many features of sized yarn properties.

(R2/S) and the pre-wet sizing process (R2/W) is maintained.

resulting in great savings (Sejri et al., 2008, 2011).

sizing with a lower size concentration (R2) in both sizing processes.

0,5

and pre-wet sizing process

**4.4 Size pick-up of yarn** 

0,8

1,1

1,4

**E (%)**

1,7

2,0

yarn extension that occurs during sizing, and it is greater in sizing with a lower size concentration (R2).

Fig. 4. Diagram of hairiness and abrasion resistance of the unsized yarn and the yarns sized with recipe 1 and 2 by the standard sizing process and pre-wet sizing process

#### **4.3 Yarn extension**

In the sizing process warp tension is a very important and unavoidable parameter, which in turn causes extension (visible even at a minimum tension) and thus the deformation shown by changes of mechanical properties. The appearance of yarn extension during the sizing process is unfortunately a reality that can not be avoided despite the minimum warp tension in segments when the warp is the most sensitive and that is in the wet state. The sensitivity of wet yarn begins with the entry into the size box and lasts until the exit from the dryer. The greater the yarn tension and its length in those segments, the higher is the yarn extension. During the pre-wet sizing process the yarn length in the wet state additionally increases between the pre-wetting box and the size box, which further increases tension

yarn extension that occurs during sizing, and it is greater in sizing with a lower size

U R1 / S R1 / W R2 / S R2 / W

U R1 / S R1 / W R2 / S R2 / W

(b) Fig. 4. Diagram of hairiness and abrasion resistance of the unsized yarn and the yarns sized

In the sizing process warp tension is a very important and unavoidable parameter, which in turn causes extension (visible even at a minimum tension) and thus the deformation shown by changes of mechanical properties. The appearance of yarn extension during the sizing process is unfortunately a reality that can not be avoided despite the minimum warp tension in segments when the warp is the most sensitive and that is in the wet state. The sensitivity of wet yarn begins with the entry into the size box and lasts until the exit from the dryer. The greater the yarn tension and its length in those segments, the higher is the yarn extension. During the pre-wet sizing process the yarn length in the wet state additionally increases between the pre-wetting box and the size box, which further increases tension

with recipe 1 and 2 by the standard sizing process and pre-wet sizing process

(a)

concentration (R2).

0

3000 6000 9000

40

60

80

100

**A (No. of cycles)**

**4.3 Yarn extension** 

120

140

**H (No. of portruding fibres)**

sensitivity, and thus susceptibility to extension and deformation. Yarn unevenness also affects extension properties to a great extent, where thin and thick places represent weaker yarn parts which are more sensitive to tension, especially in the wet state (Gudlin Schwarz, 2010). Similarly, the values shown in Figure 5 indicate that the elongation is higher during sizing with a lower size concentration (R2) in both sizing processes.

Fig. 5. Diagram of extension of yarns sized with recipe 1 and 2 by standard sizing process and pre-wet sizing process

#### **4.4 Size pick-up of yarn**

As stated in the introduction, optimizing the size pick-up to achieve the maximum utilization of the sizing process represent the biggest challenge in the whole process (Kovačević et al., 2002). The obtained results indicate that the amount of size pick-up and its distribution determine many features of sized yarn properties.

Figure 6 shows the value of the amount of size pick-up on the yarn. A small difference between the yarns sized with R1 by both sizing processes is easily observable. Significant differences are evident in the yarn sized with R2 in both processes, where an almost equal difference in reduced size pick-up (an average of 50%) between the standard sizing process (R2/S) and the pre-wet sizing process (R2/W) is maintained.

The yarn sized with lower size concentrations showed very good results in all important properties in terms of no significant deviations (in spite of a lower amount of size pick-up on the yarn) in relation to the yarn sized with a higher size concentration. This phenomenon is particularly interesting for the yarn sized with the pre-wet sizing process, where the amount of size pick-up on the yarn is considerably lower (due to water filling the interior of the yarn and the different distribution of the size pick-up on the yarn), than on the yarn sized with the standard process, indicating a reduced consumption of sizing agents and resulting in great savings (Sejri et al., 2008, 2011).

A New Pre-Wet Sizing Process – Yes or No? 69

Fig. 7. Microscopic longitudinal-section image of the yarn sized with the standard sizing process (A) and with the pre-wetting sizing process (B); SEM microscopy JSM – 6060LV

Fig. 8. Microscopic cross-sectional image of the yarn sized with the standard sizing process (A - view of the entire yarn, B, B1, - enlarged representation of the periphery of the yarn, C -

enlarged representation of the centre of the yarn); SEM microscopy JSM – 6060LV

Fig. 6. Diagram of size pick-up (Sp) on the yarns sized with recipe 1 and 2 by the standard sizing process and the pre-wet sizing process

#### **4.4.1 Distribution of size pick-up on yarn**

As already mentioned, the pre-wetting sizing process differs from the standard sizing process in the construction of the sizing machine, where another pre-wet box with hot water is installed in the front of size box. The importance of the pre-wetting box is in soaking the yarn in hot water (60-70°C) before entering the size box, which enables the dissolving and removal of grease and other impurities and additives present in the raw yarn. Furthermore, in the phase of pre-wetting, it comes to wetting the yarn in water, i.e. to fill interstitial spaces in the interior of the yarn with water, and after squeezing the excess water the yarn remains wet and partially filled with water. As such, it enters the size box with much higher humidity than it is the case with the yarn in the standard sizing process where it enters the size box dry. Therefore, the contact of the retained water in the yarn with size leads to very rapid mutual bonding, allowing faster and easier penetration and diffusion of the size into the yarn. However, the size concentration in the interior of the yarn is lower than the size concentration in the size box, because the water remained in the interior of the yarn after wetting diluted it. Therefore, the greater part of the size remains on the surface of the yarn. When sizing the dry yarn, penetration of size into the interstitial yarn spaces is not as rapid as in the case of wet yarn, and thus the inner part remains almost unfilled with size, while around or on its periphery a solid size coat is formed, unlike the yarns sized with the pre-wetting sizing process, where the size pick-up on the yarn periphery does not form such an intensive solid size coat (Fig. 7-9) (Gudlin Schwarz, 2011; Johnen, 2005).

R1 / S R1 / W R2 / S R2 / W

Fig. 6. Diagram of size pick-up (Sp) on the yarns sized with recipe 1 and 2 by the standard

As already mentioned, the pre-wetting sizing process differs from the standard sizing process in the construction of the sizing machine, where another pre-wet box with hot water is installed in the front of size box. The importance of the pre-wetting box is in soaking the yarn in hot water (60-70°C) before entering the size box, which enables the dissolving and removal of grease and other impurities and additives present in the raw yarn. Furthermore, in the phase of pre-wetting, it comes to wetting the yarn in water, i.e. to fill interstitial spaces in the interior of the yarn with water, and after squeezing the excess water the yarn remains wet and partially filled with water. As such, it enters the size box with much higher humidity than it is the case with the yarn in the standard sizing process where it enters the size box dry. Therefore, the contact of the retained water in the yarn with size leads to very rapid mutual bonding, allowing faster and easier penetration and diffusion of the size into the yarn. However, the size concentration in the interior of the yarn is lower than the size concentration in the size box, because the water remained in the interior of the yarn after wetting diluted it. Therefore, the greater part of the size remains on the surface of the yarn. When sizing the dry yarn, penetration of size into the interstitial yarn spaces is not as rapid as in the case of wet yarn, and thus the inner part remains almost unfilled with size, while around or on its periphery a solid size coat is formed, unlike the yarns sized with the pre-wetting sizing process, where the size pick-up on the yarn periphery does not form such an intensive solid size coat (Fig. 7-9)

2

sizing process and the pre-wet sizing process

**4.4.1 Distribution of size pick-up on yarn** 

(Gudlin Schwarz, 2011; Johnen, 2005).

3

4

5

**Sp (%)**

6

7

8

9

Fig. 7. Microscopic longitudinal-section image of the yarn sized with the standard sizing process (A) and with the pre-wetting sizing process (B); SEM microscopy JSM – 6060LV

Fig. 8. Microscopic cross-sectional image of the yarn sized with the standard sizing process (A - view of the entire yarn, B, B1, - enlarged representation of the periphery of the yarn, C enlarged representation of the centre of the yarn); SEM microscopy JSM – 6060LV

A New Pre-Wet Sizing Process – Yes or No? 71

Existing knowledge as well as the results of the conducted research shows that there is no big difference in the sized yarn properties obtained with the two processes, relevant for further process of making woven fabrics, while some properties of the yarn sized with the pre-wet sizing process are even noticeably better. Each of these processes brings certain advantages and disadvantages. The standard sizing process is a generally well- known, accepted and ubiquitous process in the textile industry. But of great significance is the fact that the replacement of the standard process in terms of upgrading and installing a part of the sizing range necessary for the implementation of the pre-wet sizing process does not require complex procedures or large financial expenditure. All these indicators are of great importance for the new sizing process, which give this process a priority over the standard sizing process in view of the possibility to reduce costs (size, water and energy costs - both for the sizing process and desizing process), with no negative impact on the properties of the sizing process nor on the quality of the sized yarn, including exceptional significance for

The knowledge and results shown in this chapter came from the extensive research carried out within the project "Advanced Technical Textiles and Processes", code: 117-0000000-1376, Faculty of Textile Technology, University of Zagreb, Croatia, conducted with the support of

Adanur S. (2001). *Handbook of weaving*, Technomic Publishing Company, ISBN: 1- 58716-013-

Goswami, B.C.& Anandjiwala R.D., Hall D.M. (2004). *Textile Sizing,* Marcel Dekker Inc.,

Gudlin Schwarz I., Kovacevic S. & Dimitrovski K. (2011). Comparative Analysis of the

Gudlin Schwarz I., Kovačević S. & Dimitrovski K. (2011). Analysis Of Changes In

Gudlin Schwarz I., Kovačević S., Dimitrovski K. & Katović D. (2010). Istraživanje parametara pređa škrobljenih postupkom s prednamakanjem, *Tekstil*, 59, 279-286 Hyrenbach H. (2002). Partical experience with the prewetting proces in sizing, *Melliand* 

Johnen A. (2005). Experiences in wet-in-wet sizing, *Melliand International*, 11, March, 34-36 Kovačević S. & Penava Ž. (2004). Impact of Sizing on Physico-mechanical Properties of Yarn,

Kovačević S. Grancarić A.M. & Stipančić M. (2002). Determination of the Size Coat; *Fibres &* 

Kovačević S., Penava Ž. & Oljača M. (2006). Optimisation of Production Costs and Fabric

Standard Sizing Process and the Pre-wet Sizing Process, *Fibres & Textiles in Eastern* 

Mechanical And Deformation Yarn Properties by Sizing, *Textile research journal*, 81,

the Ministry of Science, Education and Sports of the Republic of Croatia.

**5. Conclusion** 

**6. Acknowledgment** 

**7. References** 

the environmental aspect of the overall process.

7, Basel, Switzerland.

*Europe*, 19, 4 (87), 131-137

5, 545-555

Basel, ISBN: 0-8247-5053-5, New York

*International,* 8, December, 251-252

*textiles in Eastern Europe*, 10, 3, 63-67

*Fibres & Textiles in Eastern Europe*, 48, 4, 32-36

Quality, *Fibres & Textiles in Eastern Europe*, 14, 2, 40-48

Fig. 9. Microscopic cross-sectional image of the yarn sized with the pre-wetting sizing process (A - view of the entire yarn, B, B1, - enlarged representation of the yarn periphery, C, C1, C2 - enlarged representation of the yarn centre); SEM microscopy JSM – 6060LV

#### **5. Conclusion**

70 Cutting Edge Research in New Technologies

Fig. 9. Microscopic cross-sectional image of the yarn sized with the pre-wetting sizing process (A - view of the entire yarn, B, B1, - enlarged representation of the yarn periphery, C, C1, C2 - enlarged representation of the yarn centre); SEM microscopy JSM – 6060LV

Existing knowledge as well as the results of the conducted research shows that there is no big difference in the sized yarn properties obtained with the two processes, relevant for further process of making woven fabrics, while some properties of the yarn sized with the pre-wet sizing process are even noticeably better. Each of these processes brings certain advantages and disadvantages. The standard sizing process is a generally well- known, accepted and ubiquitous process in the textile industry. But of great significance is the fact that the replacement of the standard process in terms of upgrading and installing a part of the sizing range necessary for the implementation of the pre-wet sizing process does not require complex procedures or large financial expenditure. All these indicators are of great importance for the new sizing process, which give this process a priority over the standard sizing process in view of the possibility to reduce costs (size, water and energy costs - both for the sizing process and desizing process), with no negative impact on the properties of the sizing process nor on the quality of the sized yarn, including exceptional significance for the environmental aspect of the overall process.

#### **6. Acknowledgment**

The knowledge and results shown in this chapter came from the extensive research carried out within the project "Advanced Technical Textiles and Processes", code: 117-0000000-1376, Faculty of Textile Technology, University of Zagreb, Croatia, conducted with the support of the Ministry of Science, Education and Sports of the Republic of Croatia.

#### **7. References**


**Part 2** 

**Control Systems, Automation** 

