CALIBRATION OF THE NOZZLES OF 3D PRINT HEADS

(1)

CALIBRATION OF THE NOZZLES

OF 3D PRINT HEADS

Tanmay Mondal

Master Thesis May 2017

Department of Physics and Mathematics

University of Eastern Finland

(2)

Tanmay Mondal Calibration of the nozzles of 3D print heads, 65 pages University of Eastern Finland

Master’s Degree Programme in Photonics Supervisors Prof. Jyrki Saarinen

Early Stage Researcher Bisrat Assefa

Abstract

Additive Manufacturing or 3D Printing is the process of fabricating three-dimensional objects by adding layers upon layers. Different manufacturing techniques like molding, casting, forming, joining, machining etc. has been evolved over the last few decades. All these processes have certain limitations such as flexibility of different shapes, high cost, low quality and long production time. Additive manufacturing is an advanced process which overcomes most of the limitations of the old manufacturing techniques. Though, it has been difficult to manufacture imaging optics with 3D printing technology without post processing such as cutting, grinding, polishing etc.

LUXeXceL invented and developed Printoptical^c technology, which can produce quality optics without post processing. A computer aided 3D design of the optics is sliced into several layers and fed to the 3D printer. Those layers are then printed one upon another with UV curable material by the 3D printer and the final 3D object is UV-cured. This technology has cost advantage and single optics can be produced within hours.

In the experimental part of the work, the nozzles of the print heads of the 3D printer were calibrated. The calibration was needed to improve the surface profile accuracy of the printed optics. Once the droplet size from each nozzle is known, the surface profile accuracy can be improved by using so-called compensation layers or by controlled movements of the print heads. Two different methods were used to calibrate the nozzles. The surface profile was measured across the printed lines by profilometers. Finally, Matlab was used to analyze the data. Results of these two different approaches were compared with each other and conclusions were drawn.

Possible solutions to improve the surface profile accuracy of the 3D printed optics, were also discussed based on the results of the nozzle calibration.

(3)

Preface

This Master of Science Thesis is part of the research work of 3D printed photonics at the University of Eastern Finland, Joensuu. It is supervised by Professor Jyrki Saarinen and authored by Tanmay Mondal.

At first, I would like to express my gratitude to my supervisors Prof. Jyrki Saarinen and early stage researcher Bisrat Assefa for their constant guidance and instructions during the span of thesis work. I would like to express my deepest gratitude to LUXeXceL for providing me the opportunity to work with the 3D printer. I want to specially thank Dr. Pertti Silfsten for constantly guiding me.

I also thank Olympus for encouraging me to work in this thesis. I am especially thankful to Markku Pekkarinen for constantly guiding me during printing and Olli Ovaskainen for helping me in measurements.

I would like to express my regards to our course coordinator Noora Heikkil¨a for providing information, assistance and guidance throughout the entire span of master degree. I am highly obliged to the Department of Physics and Mathematics, the University of Eastern Finland for providing the required facilities for the thesis work.

I am grateful to my dear friends Rajannya Sen, Anusmita Addy and Atri Halder for being at my side and helping me to complete this two years of journey in Finland.

My earnest gratitude to all my seniors for guiding me in all possible ways during these two years.

Joensuu, May 2017 Tanmay Mondal

(4)

Chapter I

Introduction

Printing is the process of producing text, images or objects from a known design.

The most common printing familiar to us is printing of text or images on paper.

A printing ink is required to produce the design on the paper. The print head of the printer sprays the ink on the paper according to the requirement of the master template. As technologies improved, various kind of printing methods were introduced. The most demanded printing technology in recent times is the printing of three-dimensional objects.

The revolution in the world of printing or manufacturing is brought by the concept of manufacturing three-dimensional objects. Manufacturing is a process of producing predefined objects from a raw material which has a larger value than that of the raw material. The raw material can be of a solid or liquid form and have irregularities in dimensions, surface profiles, shapes, optical properties etc. Manu- facturing techniques are applied to the raw material to obtain predefined shapes, defined dimensions and desired surface profile. [1]

There are different manufacturing techniques like molding, casting, forming, joining, machining etc, that has been evolved over the last few decades. All these processes have certain limitations such as the flexibility of different shapes, high cost, low quality and high production time. During the time of overcoming the limitations of manufacturing, a new method called as Additive Manufacturing (AM) was introduced to the world.

Additive manufacturing is an advanced process which overcomes most of the limitations of the old manufacturing techniques. Instead of making frames or shaping and measuring, this process produces the desired object directly by adding small

(7)

defined layers of the raw material one after another. The flexibility on the shape of printed objects has dramatically increased because of the AM. In Subtractive Manufacturing (SM), a 3D object is made by gradually cutting down and shaping a bigger block of material. This process produces more waste material than AM.

Therefore, efficiency of AM is higher than SM in terms of material usage. [2]

As AM and 3D printing defines the same process, from now on the process will be referred as 3D printing in the thesis. A 3D printer can fabricate three-dimensional objects directly from the 3D design. At first, a Computer Aided Design (CAD) of the 3D object is made and fed to the 3D printer and from that design, the 3D object is fabricated directly. In the year of late 1980’s the concept of 3D printing technology came to the industry and that time it was called as Rapid Prototyping (RP) technology. In real terms, the first patent for 3D printing was issued for stereolithography apparatus in the year of 1986. In the year of 1984, the co-founder of 3D Systems Corporation, Chuck Hull developed a 3D printing prototype system based on stereolithography, where layers are added by curing photopolymers with Ultraviolet (UV) light lasers. But the quality of the printed object was not up to the mark. After that 3D printing technology received the attention of the researchers and industries. Over the last few years, the technology has developed a lot and the development is still in progress. [3]

The optical industries can benefit from the 3D printing technology in a number of ways. Previously it used to take several hours of hard work just to produce a single lens. To make a lens manually first one need to cut the glass in a cylindrical form, then to slice the cylinder into proper width, then to grind and polish the lens.

After that the lens was sent for testing where it’s various optical properties like focal length, the radius of curvature etc. were tested. The workload and time in the production of optical components can be dramatically reduced by 3D printing technology. One just needs to make the 3D CAD design of the optical component and then with the help of software the whole design is sliced into several layers. The number of the sliced layer depends on the height of the object to be printed and the layer thickness. Then these sliced layer files are uploaded to the 3D printer in a format that the printer can access. Then the 3D printer prints those layers one upon another and UV light cures it. The materials that are used to produce optical components are UV curable, which means as liquid photopolymer is exposed to UV light, it hardens into a transparent solid. A number of optical components, like

(8)

different lenses, beam splitters, gratings, base layers, prisms, reflectors etc. can be printed in a short time with 3D printing technology.

There are many other advantages of 3D printing technology other than low workload and low time consumption. Coatings of the object can also be done through a 3D printer and in advanced 3D machines, joining of two components can also be possible. However, 3D printing technology is still developing and has some limitations. Production of complex structures is still a challenge. The surface quality of the optical components produced by the 3D printer is still not good enough for imaging applications.

The surface profile accuracy of the printed optics should be less than 5 µm for imaging application. Our previous analysis showed that we have achieved a surface profile accuracy of ±10 µm for a small lens (diameter of 5 mm). To improve the surface profile accuracy, we need to know the exact droplet size from each nozzle of the print head.

In this thesis, calibration of the nozzles of the 3D print head was done and the the droplet size from each nozzle was determined. Two different methods were used for calibration: the direct weighting method and the line scanning method. It was showed that the line scanning method takes much less time than the direct weighting to calibrate the nozzles. After calibration, different methods were suggested, which can improve the surface profile accuracy using the knowledge of nozzle calibration.

In Chapter II different additive manufacturing technologies are discussed. In Chapter III the different 3D printing technologies for printing optics are discussed.

In Chapter IV configuration of the print head system is explained. In Chapter V different devices that are used for our weight measurement and data analysis are discussed. In the Chapter VI different methods for calibration of nozzles are discussed in details and in Chapter VII the result of the measurements are provided and compared. The conclusions and discussions are given in Chapter VIII.

(9)

Chapter II

3D printing

In this chapter different methods, which are used for 3D printing are discussed.

These different techniques are based on raw material that is used for printing, application of the 3D printed object etc. The technology used for printing the same material can also be different and thus the name of the 3D printing technology can be different.

2.1 Methods

Fabrication of three-dimensional objects by adding layer upon layers is called as 3D printing. A CAD design is fed to the computer and then the design is sliced into thin layers. Then the 3D printer produces those layers one after another, thus producing the whole 3D object. There are a number of different 3D printing methods that are available and are discussed shortly here. [4, 5]

2.1.1 Stereolithography (SLA)

Stereolithography (SLA) is a 3D printing technique where the object is manufactured by exposing a chamber of photopolymer by a UV laser. There is an elevator platform present in the chamber which is filled up by the liquid photopolymer. At first, a CAD design is uploaded to the computer and sliced into several layers. Then the UV laser polymerizes the first layer according to the design of sliced first layer. After fabrication of each layer, a resin-filled blade moves across the surface re-coating with fresh material. Then the elevator platform moves down according to the thickness of the second layer and the liquid photopolymer is exposed to the UV laser again and the second layer is formed on the first layer. This process is repeated for each

(10)

layer until the whole design is completed. After that, the whole design is immersed in a chemical to remove the excess resin and finally cured in a UV oven. [6]

Figure 2.1: Schematic diagram of Stereolithography (SLA)

Figure 2.1 shows the schematic diagram which shows the steps for the Stere- olithography. Unfortunately, this process can not be applied for printing optical components as it creates borders for each layer and increases surface roughness.

2.1.2 Selective Laser Sintering (SLS)

Selective Laser Sintering (SLS) was first developed and patented by Dr. Carl Deckard and Dr. Joe Beaman at the University of Texas at Austin in mid-1980s.

This process uses a high power laser to fuse powders of different materials, such as metals, plastic, ceramic, glass etc into the predefined 3D design. [7]

At first, the support stage is covered with the desired powder, that is a powder bed is made. A high power laser then sinters the powder selectively according to the design of the 3D model and binds the powder into a solid structure. After the first layer is fabricated, the support layer is moved downwards to a distance equal to the thickness of the second layer in the design. Then in the similar process, the second layer is created and so on until the whole 3D object is developed. It is easy to separate the constructed object from its surroundings as the object is surrounded

(11)

Figure 2.2: Schematic diagram of Selective Laser Sintering (SLS)

Figure 2.2 shows the schematic diagram of Selective Laser Sintering (SLS). As this process can be used easily to produce 3D objects from a CAD design, it is widely used all over the world. However, the final object contains borders due to sintering and requires post-processing to achieve better surface quality. Also, SLS takes a longer time to fabricate and not a good approach for producing optical components. [8]

2.1.3 Fused Deposition Modelling (FDM)

Fused Deposition Modelling (FDM) was first developed by S. Scott Crump in late 1980’s and was commercialized by Stratasys in 1991. [9]

At first, a CAD design is made and sliced into several layers using slicing software and the design is uploaded to the FDM machine. A coil of plastic filament is unwounded and through the driving motor sent to the molten paste chamber to an extrusion nozzle. The nozzle is heated to the temperature above the melting point of the material. The nozzle can turn on or off the flow the molten material selectively and can move both vertically and horizontally. The nozzle produces the object layer by layer according to the design and the molten material cools down naturally to

(12)

produce the solid object of the 3D design. [10]

Figure 2.3: Schematic diagram of Fused Deposition Modelling (FDM)

Figure 2.3 shows the schematic diagram of Fused Deposition Modelling (FDM).

Though the FDM is simple to utilize, it has some limitations. Good surface quality objects can not be produced, mass production is not possible with this kind of technology and it takes a long time as the molten material cools down naturally. [11]

2.1.4 Laminated Object Manufacturing (LOM)

Laminated Object Manufacturing (LOM) uses a roll of film of paper, plastic or metals to produce objects. This method was developed by Helisys Inc. [12]

A roll of paper or plastic is unwounded on to a platform. A heated roller heats the film and the optical system traces the profiles and the first layer is deposited.

After that, the stage moves downwards and new a layer of the film comes on top of it and in the same way the second layer is created. After the fabrication of layer is completed, the laser cutter is used to cut it into shape.

(13)

Figure 2.4: Schematic diagram of Laminated Object Manufacturing (LOM) Figure 2.4 shows the schematic diagram of LOM. This process is not efficient to produce good quality objects and during the fabrication huge amount of smoke is produced. The accuracy of dimensions of produced 3D objects is not good. [13]

2.2 Materials

The 3D printing technology comes with the use of a wide range of materials depending on the use of the output product and the application domain. Based on the materials, different 3D printing techniques are also selected.

2.2.1 Metals

The main metals that are used in 3D printing are

• Steel

• Gold

(14)

• Silver

• Titanium

• Stainless steel

Even compounds can also be used in 3D printing using different fusing processes. [14]

2.2.2 Thermoplastics

Thermoplastics or polymers are largely used in 3D printing as raw material and they are also low-cost materials. The main thermoplastics are

• Polycarbonate

• Polyvinyl alcohol (PVA)

• Acrylonitrile butadiene styrene (ABS)

• Polylactic acid (PLA)

Among these ABS has the widest use, though the use of PLA is also increasing due to its flexibility, availability, both rigid and soft finishes. PVA is used as support material in 3D printing while the use of polycarbonate is still in development. [14]

2.2.3 Medical and biochemical materials

The method of 3D printing is taking a huge part in medical applications. Use of Bio-ink or bone material in 3D printing is a hot topic nowadays. Bone cells combined with a compound material made of calcium phosphate, zinc and silicon are printed as new bones. Later in the printed material is dissolved leaving only the new bone. [14]

2.2.4 Glass

The glass is the best material to produce optical components. At first, the glass is ground to powder and used as raw material in 3D printing process. [14]

(15)

2.2.5 Other materials

There are also many other materials that are coming up fast in 3D printing technology. Recently, building materials are also used to manufacture buildings with the 3D printing process. Conductive carbomorph are used 3D printing to print circuits and batteries on it. Foods are being produced in different shapes using 3D printing technologies. [15]

2.2.6 Support material

Other than 3D printing material, another type of material is also used called the support material. Sometimes it is necessary to hold the layers in the proper place or to keep the base material out of the way by a support structure. The materials used in this support structure are known as support materials. Once the printing is done on the support structure, the support material is removed from the 3D structure by using different methods such as chemical bath, cutting etc. Some of the methods like Stereolithography, FDM uses support materials. [16]

A wide range of raw materials can be used in 3D printing. Different material or different colors can be used at the same time, which makes 3D printing more attractive. [17]

2.3 Advantages

The 3D printing technology has many advantages over traditional manufacturing.

The complex structures can be produced by designing that complex structure in CAD and then uploading that CAD design into the 3D printer. The complex structure can be made faster and cheaper but in an efficient way with 3D printing.

Modification of the previous structure can be done easily by modifying the CAD file accordingly.

Less production time is required in 3D printing than other manufacturing techniques. The dimensions of the 3D object defined by the CAD file are achieved more accurately in 3D printing.

From the point of view of 3D manufacturing industries, it is needed to increase the production rate. That is, industries want to produce more in less time. Here, 3D printing has more advantage than any other 3D manufacturing techniques, as

(16)

3D printing can produce more objects in less time. This technique can be used for mass production. [18]

The 3D printing technology does not remove materials unlike those subtractive manufacturing methods such as milling, grinding, polishing etc. Therefore, 3D printing produces almost no waste material and saves huge cost during mass production.

2.4 Challenges

There are a number of advantages of 3D printing technology. However, there still remains some challenges that 3D printing technology needs to overcome. The optics produced by 3D printing has poor surface quality, which requires post processing or coating. For imaging application, the typical limit on surface roughness is 5 nm.

During previous measurements we have found that, it is 24 nm for an evaluation area of 62×47µm² of an unpolished 3D printed flat surface [19]. The surface profile accuracy of imaging optics should be less than 5 µm. Our latest results prove that we start having good enough surface roughness (rms roughness close to 5 nm for evaluation are 62×47 µm²) [19], but surface profile accuracy is not good enough (±10 µm) for imaging application.

(17)

Chapter III

3D printing for optics

Fabrication of optical components has always been a challenging and troublesome process. Glass is the basic material used for fabricating optical components. Manu- facturing of optical components goes through several steps such as cutting of glass slab into proper size, grinding, polishing, testing etc. Each of these processes requires expert for handling and usually takes several days or weeks to produce a single optical component.

Many of these drawbacks can be overcome by 3D printing technology. It does not require experts for cutting, grinding, polishing or testing. Only one expert CAD designer can print the whole optical component with the 3D printer. One optical component can be printed in a couple of hours. Also, several components can be printed simultaneously with the 3D printer. [20]

3.1 Technologies

Fabrication of optical components using 3D printing is challenging. At first, the 3D CAD design is sliced into several numbers of layers and then fed to the 3D printer.

An optically transparent material is used as the ink of the 3D printer, which comes out through a nozzle of the inkjet print head as droplets, according to the design of the first layer. The print head is capable of moving in all three directions and it is controlled by a computer. After printing of first layer, the second layer is printed upon the first layer by the 3D printer. After printing each layer, the print head moves upwards according to the thickness of the next layer. After completion of printing the whole 3D design, the 3D object is moved to the another location where it is hardened. [20]

(18)

Based on the type of the inkjet printer, printing is divided into two classes as Continuous inkjet printing and Drop-on-Demand (DOD) inkjet printing. [21]

3.2 Continuous Inkjet (CIJ) Printing

Continuous inkjet printing is the technology where droplets are ejected from the nozzles continuously. This method can also be divided into Binary deflection printing and Multiple deflection printing.

3.2.1 Binary deflection inkjet printing

In Binary deflection inkjet printing, a stream of piezoelectrically charged ink is di- rected to the nozzle through a gunbody by high-pressure pump creating a continuous stream of ink.

Figure 3.1: Schematics of Binary deflection inkjet printing

Figure 3.1 shows the schematic diagram of Binary deflection inkjet printing process. As the charged ink passes through a pair of charged electrode, it breaks into droplets. Then those droplets pass through a high voltage deflection plate. Be- cause of high voltage all the charged droplets deflect to the Gutter, which stores and recycles the ink. The uncharged droplets pass straight and hit the substrate and prints. [22–24]

(19)

3.2.2 Multiple deflection inkjet printing

Multiple deflection inkjet printing operates in an exactly opposite way of Binary deflection inkjets. Here the uncharged droplets are stored in the Gutter and recycled later.

Figure 3.2: Schematics of Multiple deflection inkjet printing

The charged droplets are deflected to the suitable position on the substrate by high voltage deflection plate. The schematic process of Multiple deflection inkjet printing is shown in Figure 3.2. [22]

Continuous inkjet printing has high-velocity droplets, which allows long distance printing between print head and substrate. The CIJ printers are easy to use, reliable, of low cost and can run for long hours. Continuous inkjet printing is used to print date codes, bar codes, logos etc. Though CIJ printing is not capable of printing good quality optics as the user do not have control over droplets and probability of printing at an undesirable position is high thus increasing the surface roughness. [24]

3.3 Drop-On-Demand (DOD) inkjet printing

In Drop-On-Demand (DOD) inkjet printing the droplet is delivered whenever it is needed. There are many different DOD technologies but thermal DOD inkjet and piezoelectric DOD inkjet are the most popular. [24]

(20)

3.3.1 Thermal Drop-On-Demand (DOD) inkjet printing

Thermal DOD inkjet printing is one of the most popular methods in the printing industry today. It uses a heater/resistive layer which is located near and back of the nozzle.

Figure 3.3: Schematics of thermal Drop-On-Demand (DOD) inkjet printing When a droplet is required, a pulse of voltage is transferred to the resistive layer which heats the layer to the critical temperature for bubble nucleation. The bubble then expands because of the excess heat and forces the ink to form a droplet out of the nozzle as shown in Figure 3.3. After the heat is removed the bubble collapses and the ink comes back on the resistive layer. As the voltage is not applied now, the resistive layer is not heated and no bubble is formed. The whole process takes about 10µs. [25, 26]

3.3.2 Piezoelectric Drop-On-Demand (DOD) inkjet printing

Piezoelectric DOD inkjet printing works on the principle of inverse piezoelectric effect of piezoelectric ceramic materials. When a voltage is applied, the material changes its shape. Because of the applied pulse, the material contracts and squeezes the material out of the nozzle as droplets. There are three main types of piezoelectric inkjet printing available depending on the actuator and material: squeeze mode, bend mode and push mode. [19], [25], [27]

(21)

3.3.2.1 Squeeze mode actuator

The squeeze mode actuator works on the principle of squeezing the confined volume of the ink and creating pressure. The actuator of the print head is made of a piezoelectric ceramic tube, which is polarized radially and there are two electrodes present at either side of the tube.

Figure 3.4: Operating principle of squeeze mode actuator Drop-On-Demand (DOD) inkjet printing

When a droplet is needed, a voltage pulse is applied to the transducer with proper polarity, the transducer contracts creating a pressure on the enclosed volume of ink. Thus, a droplet of ink comes out of the nozzle (Figure 3.4). Though, when the pressure is released and the tube expands to its original volume, a small amount of ink comes back to the tube. To overcome this drawback, the pressure is released slowly. [25]

3.3.2.2 Bend mode actuator

Figure 3.5 shows the operating principle of bend mode actuator of piezoelectric ceramics. One side of the pressure chamber consists of conducting diaphragm with a deflection plate which is made of piezoelectric ceramics. The plate has a conductive coating on the outer surface of it. The pressure chamber also has an ink inlet and an outlet nozzle. At the time of droplet requirement, a voltage pulse is applied to the plate, which bends the diaphragm inwards creating a pressure to the enclosed volume of ink.

(22)

Figure 3.5: Operating principle of bend mode actuator Drop-On-Demand (DOD) inkjet printing

This pressure produces a droplet coming out of the nozzle. The size of the output droplet depends on the pulse duration, the amount of applied voltage and diameter of the nozzle. [25], [28]

3.3.2.3 Push mode actuator

Figure 3.6 shows the operating principle of push mode actuator type DOD process.

Figure 3.6: Operating principle of push mode actuator Drop-On-Demand (DOD) inkjet printing

(23)

When a droplet is required, a voltage pulse is applied and the piezoelectric ceramic rod (shown in Figure 3.6) expands. This expansion pushes the diaphragm creating pressure on the ink chamber and a droplet comes out of the nozzle. [25], [29]

3.4 Printoptical

^c

Technology of LUXeXceL

LUXeXceL has developed a new printing method for optics called the Printoptical^c Technology which can produce high-quality optics without any postprocessing. At first, they started to produce optics for illumination based applications in 2009. [30]

Figure 3.7: Schematic diagram of Printoptical^c Technology of LUXeXceL.

[31] (Substrate→ , Droplets→1 , UV lamp→2 , UV light→3 , Single4 layer→5 and Multiple layers→)6

Using the Printoptical^c Technology one can have control over the surface roughness of the optical components. This is the main reason for the use of Printoptical^c Technology in the printing of different optical components such as lenses, gratings, prisms, beam splitters, micro lenses, LED optics etc.

A UV curable material is used as the ink of the printer which has a refractive index of 1.53 for wavelength 587.6 nm [32]. The UV curable material has a high surface tension and thus two droplets or two layers can merge with each other with-

(24)

out making any lines. Because of this property of the material, it is possible to print continuous layers and decrease the surface roughness. The modified industrial inkjet printer works on the principle of DOD printing technology. The print head is controlled piezoelectrically. The print head has a resolution of 1440 dpi (dots per inch). From this resolution, the size of a single droplet comes about 17µm and the volume of a single droplet comes about 7 picoliters. [33] The size or volume of the droplet can be controlled by changing the applied voltage to the piezoelectric.

The UV curable transparent polymer material is jetted on the substrate as droplets and these droplets form the first layer (Figure 3.7→ ). After that, the5 second layer is formed in the same way and so on. There is one UV lamp attached to the print head (Figure 3.7→ ) which cures the layer. As the UV falls on the3 UV curable material, it polymerizes. After printing of each layer, the UV lamp polymerizes each layer. The polymerization also is the type of free radical polymerization which forms the polymer by successive addition of free radicals. Because of this property, the layers combines with each other forming a larger single optical component. The individual droplets merge with each other because of the high surface tension of the polymer and sufficient settling time. Once the whole design is finished, the whole design is cured again with UV and finally forms a hard and smooth optical component. [31], [34]

The Printoptical^c Technology is capable of printing high-quality optics without any continuous supervision on the design. The research is going on to increase the dpi to 2880 (droplet size about 12µm) from the current position, which will certainly improve the quality of produced optics. [34]

(25)

Chapter IV

Configuration of print head system

The configuration of the print head is important to understand the printing process and the configuration is explained schematically in Figure 4.1.

Figure 4.1: Schematic diagram of a single print head (Side 0 and Side 1), showing nozzle distribution

The print head system consists of three individual print heads having similar nozzle distribution. Each head has two sides: side 0 and side 1. There are 500 nozzles present in each side, therefore 1000 nozzles in each head and 3000 nozzles in

(26)

the print head system. The lateral distance between any two consecutive nozzle in the same side is 140 µm and the lateral distance between two consecutive nozzles of side 0 and side 1 is 70 µm. The distance between any two consecutive alternate nozzle of the same side is 280µm.

Each head has printing resolution of 23.5 µm along printing direction (Y-axis) and 23.3 µm across the printing direction (X-axis).

(27)

Chapter V

Measurement devices

In this chapter, we discuss different devices that we have used for our measurement and surface profile analysis.

5.1 SCALTEC SBC 31 balance

SCALTEC SBC 31 is a high precision digital balance. We use it in this work to measure the weight of the droplets. The balance is installed on a platform with bubble indicator and leveling feet to make it perfectly horizontal.

Figure 5.1: SCALTEC SBC 31 balance

The weighting pan is protected by a glass frame as even small amount of air flow

(28)

Table 5.1

Specification of SCALTEC SBC 31 balance. [35]

Accuracy class I Max. weighting capacity 220 g

Min. capacity 0.01 g Scale interval (d) 0.1 mg Verification scale interval (e) 0.01 g

Tare range (subtractive) -220 g

can affect the result. The U-shaped black frame is used to get rid of static electricity from the sample.

A sample is placed on the sample platform inside the glass box and the balance shows the weight in grams with an accuracy of 0.1 mg. There is a Tare button present in this system, which is used to define the zero point. One should take the measurement value when the display becomes steady. The technical specification of the balance is given in Table 4.1. The system should be operated in an environment with temperature range +15^◦C to +25^◦C. [35]

5.2 Mitutoyo SJ-210 Portable Surface Roughness Tester

Mitutoyo SJ-210 is a Portable Surface Roughness Tester which is easy to use and provides two-dimensional surface roughness measurement. With proper alignment setup, it can be used for both horizontal and vertical surface measurement.

Figure 5.2: Mitutoyo SJ-210 Portable Surface Roughness Tester

(29)

Table 5.2

Specification of Mitutoyo SJ-210 Portable Surface Roughness Tester. [36]

Drive speed Measuring: 0.01”/s (0.25 mm/s), 0.02”/s ( 0.5 mm/s), Returning: 0.03”/s (0.8 mm/s)

Max. evaluation length 12.5 mm

Detecting method Differential inductance

Measuring range (350 µm) (-200 µm to +150 µm) Material of stylus Diamond

Stylus tip radius 5 µm Measuring force 4 mN

To use the tester first theStartbutton is pressed and the stylus head moves out.

The surftest USB toolis opened in the computer and then the head is placed on the sample. One can choose the length to scanned by the software, pressing the menu Meas.Cond., up to 12.5 mm. To start the scan, the Start measurement menu is pressed. Now the head retraces back as it scans the surface. The diamond head stylus having 10 µm diameter measures the surface accurately. The stylus produces a force of 4 mN on the surface (Table 4.2), which is small to damage the surface as the stylus moves on it. The output of the measurement can be saved as an Excel file which clearly shows the Ra, Rq and Rz roughness parameters with roughness data.

Here, Ra defines the absolute value of the average deviation of the roughness from the mean line over one sampling length. The parameter Rq or Root Mean Square Roughness provides the standard deviation of the distribution of surface heights. In International ISO system, the Rz parameter is defined as the difference in height between five highest peaks and five lowest valleys along the assessment length of the profile. [36]

One has to be careful of stylus tip as it is sensitive. To close the tester, first press the Esc/Guide button to make the screen off. Then press and hold the Power/Data and Start/Stop button together to make the stylus back to the safety position. It is important to close the device properly after use to protect the stylus. [37]

(30)

5.3 Dektak 150 surface profiler

Dektak 150 surface profiler is one of the most powerful surface profiler device capable of measuring two-dimensional surface profile. The profile measurement is based on contact stylus profilometry. It is a computer controlled high performance and cost effective device with high measurement repeatability, a wide range of sample flexibility, low noise floor and optimum modularity.

Figure 5.3: Dektak 150 surface profiler [38]

Operating principle of the profiler is user-friendly. At first put the test sample on the circular table such that the sample is just underneath of the stylus. By clicking the down menu, bring the stylus to the test surface. The stylus will at first touch the surface and then it will move up a little. Now the surface can be seen on the computer screen through the camera attached near to the stylus. One has to be careful while bringing the stylus down. Make sure that the stylus do not touch the bare table, but touches the sample. As the profiler scans only in one direction, the sample is aligned such that the lines or layer separations are perpendicular to the scanning direction. The alignment is done by moving sidewise and looking through the camera, but do not move the sample up: it can damage the stylus. [38]

(31)

Now select the first point to start the measurement. Give suitable parameters to the software like scan length, scan duration, maximum height etc. as shown in Figure 5.4.

Figure 5.4: Software operation of Dektak 150 surface profiler

For our measurement, the maximum scan length was 3.8 cm with 400 s scan duration and 524 µm height (Table 4.3). The sample had a length of 7.5 cm. So, the measurement is done in two parts. The maximum number of data points per scan was 120,000. As the stylus moves on the sample, at each position it finds surface variation, it produces graphs with peaks and valleys. For proper measurement, the specification of the device is important, which is given in Table 4.3. [38]

The output of the measurement can be saved as .data file of .csv file. It is better to export the file as .csv file, which can be accessed easily with Excel or Matlab.

(32)

Table 5.3

Specification of Dektak 150 surface profiler. [38]

Stylus sensor Low-inertia sensor (LIS 3) Stylus force 1 to 15 mg with LIS 3

0.03 to 15 mg with N-Lite sensor option Stylus Options Stylus radius options from 50 nm to 25 µm;

High Aspect Ratio (HAR) tips 10 µm× 2 µm and 200 µm× 20 µm

Scan Length Range 55 mm standard; up to 200 mm with stitching option Data Points Per Scan 120,000 maximum

Max. Sample Thickness Up to 90 mm, depending on configuration

Max. Wafer Size 150 mm (200 mm with Advanced Automation Package) Step Height Repeatability ≤6A (D150);^◦ ≤4A (D150+ option);^◦

1 sigma on 0.1 µm step

Vertical Range 524µm (max. 1 mm optional) Vertical Resolution 1A max. (at 6.55^◦ µm range)

Sample Viewing 640 × 480-pixel (1/3 ”-format) camera, USB;

fixed magnification, 2.6 mm HFOV (166X with 17 ” monitor);

optional manual zoom, variable 0.67 to 4.29 mm HFOV (644X to 100X with 17 ” monitor)

(33)

5.4 Glossmeter

Glossmeter is a device which measures the specular reflection from a surface. Spec- ular reflection is defined as the reflection where the incident angle (θi) and reflected angle (θr) is equal to each other.

Figure 5.5: Glossmeter

The laser source (Class 3B laser) and the detector are fixed in the black box (Figure 4.5), while the base can move in both X and Y direction. The height from the base to laser source is fixed. Two powerful magnets are used to keep the sample in position. Two-dimensional scan of the sample is performed with this Glossmeter.

The scan area can be varied by varying the scan length and scan width. The resolution of the scan can also be varied. The more resolution, the more time it takes to complete the scan.

(34)

Chapter VI

Methods of calibration

The main objective of the work is to calibration the nozzles of the print head of the 3D printer. The calibration is necessary as the error in single nozzle droplets will additively increase because of additive manufacturing. There are two ways that are used to calibrate nozzles in this scope of work. One is the direct weighting method which is based on the measurement of the weight of the nozzle output directly.

Another method is the line scanning method based on the printing of lines and then measurement of surface profile, specular reflection from the surface and also by image processing.

6.1 Direct weighting method

In direct weighting method, the weight of a single droplet from each nozzle is measured. We need to find out exactly how much a droplet from each nozzle deviates from each other so that we can put proper compensation layer to improve the surface profile accuracy. This compensation layer can be provided during 3D printing.

As the droplets are small (in nanograms), printing and weighting a single droplet can produce errors in measurement. To overcome this error, 2,000,000 droplets were printed for each nozzle and then weighted. We used white antistatic one-time use cups for printing the droplets on it.

At first, the weight of the empty cup (W¹) is measured. After that, by enabling a single nozzle 2,000,000 droplets were printed on the cup. Again the weight of the cup with printed droplets on it is measured (W²). The weight of 2,000,000 droplets (W³) can be found out by subtracting these two weights. From the weight W³ and

(35)

W4 = W2−W1

2,000,000 = W3

2,000,000 (6.1)

To check the repeatability of the measurement, the weight of the same nozzle was calculated more than once and the results were also compared with each other.

The direct weighting method is time-consuming, while the line scanning method takes the least time to calibrate nozzles. It takes about 15 minutes just to print 2,000,000 droplets from a single nozzle, while printing of all 500 lines takes about 10 minutes.

6.2 Line scanning method

The line scanning method is based on printing lines on the substrate and then analysis of the sample with profilometers, Matlab, Scanning Electron Microscope (SEM) and Glossmeter. Also, the photograph of the lines printed sample taken by the camera of the 3D printer is analyzed. Each print head consists of 2 sides and each side has 500 nozzles. A maximum of 500 lines can be printed at a time to obtain better resolution, ease of calculation and better results.

First, a uniform layer of the UV curable material is printed on the substrate.

Without this uniform layer, the lines will merge with each other without producing distinct lines because of the surface tension of the UV curable material and surface free energy. The surface free energy (σs) of the solid is related to surface tension of the liquid (σl), contact angle (θ) and interfacial tension (σsl) between solid and liquid by Young’s equation,

σs =σsl+σl·cosθ (6.2)

For Poly-methyl methacrylate (PMMA) substrates the surface tension of the liquid is more than the interfacial tension between substrate and liquid. [39]

Then all the print heads and sides were disabled while keeping only one side enabled and single layer lines (500 lines) were printed on the uniform layer as shown in the following figures.

(36)

Figure 6.1: Schematic diagram of single layer lines on uniform layer

Figure 6.1 shows the schematic diagram of single layer lines printed on the uniform layer. If a small part of Figure 6.1 is magnified, then the top view of the printed lines will look similar to the Figure 6.2.

Figure 6.2: Top view of single layer lines on uniform layer

(37)

nozzles of the same side is theoretically 140µm.

Figure 6.3: Side view of single layer lines on uniform layer

Figure 6.3 shows the side view of the printed single layer lines. The height of the PMMA substrate is 3 mm. In Figure 6.3, Y-axis defines the direction of printing and Z-axis denotes the height.

Once the printing is done, several measurement techniques were applied to the printed lines such as surface profile measurement, analysis of images of the lines taken by printer ccd and gloss measurement.

6.2.1 Surface profile measurement

At first, the surface profile of the printed lines was measured by stylus profilometer.

The surface profile measurement gave a graph having peaks at each position of lines.

Therefore, there were 500 peaks for 500 lines. In the obtained graph the base of the peaks were not in the same plane as well as the height of all peaks were different depending on the inaccuracy of the nozzle output. So a Matlab code was developed which placed the baseline of all the peaks to zero line while keeping the area under each peak same as in the original file. Now the area under each peak was calculated by the same Matlab code.

To find the volume of each droplet, the length of a single droplet along Y-axis (Figure 6.2) should be determined. Along Y-axis or along the printing direction,

(38)

each nozzle prints 1088 dots/inch. The Y-length of each peak or each droplet can be found out by the following calculation. There were 1088 dots present in 1 inch, which means that 1088 dots were present in 24.5×10⁻³ µm. Therefore theY-length of 1 droplet was ²⁴^.⁵1088^×¹⁰⁻³ µm = 23.34µm.

Now the area (A) of each peak was multiplied by 23.34 µm, which gave the volume (V) of each droplet. From the calculated volume one can also determine the layer thickness. As each pixel had a dimension of 23.5×23.34 µm², then the layer thickness (h) is defined by the following equations.

V =A×23.34µm (6.3)

Also,

V = 23.5µm×23.34µm×h (6.4)

Therefore,

h= A×23.34µm

23.5µm×23.34µm (6.5)

The gap between two consecutive lines can also be calculated roughly from the original data graph of surface profile in Matlab. Again if the density (d) of the UV curable liquid or material is known, then the mass (m) of each droplet can also be calculated from

m=V ×d (6.6)

After completion of surface profile measurement, another set of 500 lines were printed from another side of the head and the data analysis of surface profile was performed to obtain all the parameters.

6.2.2 Analysis of image taken by printer camera

There is a CCD camera present beside the print heads, which is used to calibrate the print heads. It can also be used to take the image of any printed optics.

(39)

Figure 6.4: Photograph of single layer lines taken by 3D printer camera

After printing the lines, an image of the printed substrate is taken by the camera and a part of that image is shown in Figure 6.4.

From the image, the gap between two consecutive lines can be calculated and the result can be compared with the previous results obtained by surface profile measurement. First zoom in the image such that the pixels can be seen clearly, shown in Figure 6.5.

Figure 6.5: Zoomed in photograph of single layer lines

(40)

It is known that the distance between two consecutive nozzles of the same side is 140 µm (Figure 4.1). So the distance between two center pixels of two consecutive line will also be 140 µm as shown in Figure 6.5. One can approximately count the number of pixels between the center of two lines, then again count the number of pixels between two lines. Then the gap (g) between two lines can be determined by

g = 140×n1

n² µm (6.7)

Where n¹ is the number of pixels present the in gap and n² is the number of pixels between the center of two lines.

(41)

Chapter VII

Experimental results

In this chapter, experimental results are discussed in details. All the lines on the base layer were printed at the speed of 100 mm/s. The resolution along the printing direction (Y-axis) was 1080 dpi (dots per inch), that is 23.5 µm and 1088 dpi across the printing direction (X-axis), that is 23.3 µm. The base layer was made from 40 layers printed at the speed of 125 mm/s. All the experiments were performed by using both sides of the Head 2.

7.1 Direct weighting method

Firstly, 2,000,000 droplets from each nozzle of the Head 2, Side 0 were printed on antistatic cups (one for each nozzle) and their weight was measured by SCALTEC SBC 31 balance (Figure 5.1). Then the average weight of a single droplet was calculated. To check the repeatability of the whole process, the test was repeated for each nozzle on two consecutive days. The part of the results is shown in the Figure 7.1.

It is evident from the Figure 7.1 that the results deviate from each other quite a lot, though they are obtained from the same nozzles. This is because of the inaccuracy of the balance. The resolution of the balance was 100 µg, while the measured average weight of 2,000,000 droplets was about 6 mg. Therefore, the calculated average weight of a single droplet was in nanograms. If the weight of a single droplet is about 3.025 ng, then the weight of 2,000,000 droplets will be 6050 µg. But the balance will show the weight as 7000µg, as it can not resolve the weight of 50µg.

It takes about 15 minutes to print 2,000,000 droplets from a single nozzle. Each

(42)

Head consists of two Sides, each having 500 nozzles (total 1000 nozzles in each Head)(Figure 4.1). Therefore it will take about 250 hours just to print 2,000,000 droplets from each of 1000 nozzles of each Head.

To overcome this lack of proper measurement device and to reduce the time consumption, another method called the line scanning method was introduced.

1 2 3 4 5 6 7 8 9 10 11 12

2 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6 3.8 4

Nozzle number ( 1 to 12 )

Weight per droplet (ng)

Day 1 Day 2

Figure 7.1: Comparison of weights of single droplet from the same nozzles of the Head 2, Side 0

7.2 Line scanning method

Firstly, single layer lines were printed on a 40 layered uniform base and the surface profile was scanned along the cross section of the lines. To check the repeatability of the whole process, single layer lines from each side of the Head 2 were printed once on two separate PMMA substrates and the results were compared using the line scanning method.

7.2.1 Feasibility test of the concept of the line scanning method

To check the accuracy of the line scanning method, three different and independent measurements were performed. These three independent measurements were portable stylus measurement, photograph analysis and gloss measurement. Finally,

(43)

7.2.1.1 Portable stylus measurement

At first, the surface profile of the printed lines (from the Head 2, Side 0) were scanned along the cross section of the lines by the Mitutoyo portable surface roughness tester (Figure 5.2) and the data was analyzed by Matlab. The surface profile scan was performed on several randomly chosen positions of each sample.

Figure 7.2 shows a small section of the actual surface profile of printed single layer lines measured along the cross section of the lines. A Matlab code was developed to calculate the area under each peak. At first, the code determines the base line of the peaks. Then, by subtracting the base line values from the actual profilometer data, all the peaks are shifted to zero base level (Figure 7.3), while keeping the area under each peak fixed.

0 0.5 1 1.5 2 2.5

−1

−0.5 0 0.5 1 1.5 2 2.5

X−axis ( mm )

Z−axis ( µm )

Original profilometer data of surface profile Base line

Figure 7.2: Surface profile of the cross section of single layer lines measured by the Mitutoyo portable surface roughness tester

Here, we were interested in the area under peaks, rather than the peak values.

Because, if the droplet spreads then the peak value will decrease while the area under the peak will remain fixed. The droplets spreads because of several parameters: time gap for UV curing, surface properties and viscosity of each droplet. According to the hardware configuration of the 3D printer, the droplets which are printed later, are UV cured first and the droplet which are printed at the starting of printing process, are cured later. This process allows those droplets, which are printed at

(44)

the beginning, more time to spread than the droplets which are printed later. The surface irregularities of the substrate also allows the droplets to spread differently at different positions. The viscosity of each droplet may change during the printing process, therefore the spread of each droplet can be varied.

0 0.5 1 1.5 2 2.5

0 0.5 1 1.5 2 2.5 3

X−axis ( mm )

Z−axis ( µm )

Figure 7.3: Base line corrected single layer lines

The Figure 7.3 is the base line corrected profile of the Figure 7.2. Here we are more interested in the area under each peak than the volume or mass of each droplet because the volume or mass can be obtained just by multiplying with constants and we are not interested in absolute values but the difference between figures from each nozzle. From the corrected profile, the average gap between two lines was calculated and it was around 69µm. Based on our previous work and analysis of line thickness, it was expected to have the width of each line about 70µm. As the distance between any two nozzle is 140 µm, therefore theoretically the average gap between two lines should be about 70 µm. Therefore, our measured average gap between two lines matches the theoretical value.

7.2.1.2 Photograph analysis

After the surface profile measurements, the photograph of the same sample (Figure 7.2) was analyzed.

(45)

Figure 7.4: Photograph of single layer lines printed by the Head 2, Side 0 From the photograph showed in Figure 7.4, it can be analyzed that the gaps between the lines are quite consistent and also the lines are quite uniform. The average gap between lines are also roughly calculated from the image and the average gap between two lines was about 69µm. The average gap analyzed from the photograph also matches the theoretical value and the results of line scanning method.

The photograph was also read and plotted by Matlab and the average gap between two lines was measured again. The Figure 7.5 shows the Matlab plotted image of the same photograph in the Figure 7.4. The measured average gap between two lines was ca. 68µm.

Y−axis

X−axis

300 400 500 600 700 800 900 1000

100 150 200 250 300 350 400 450 500 550 600

Figure 7.5: Matlab plotted image of the photograph of single layer lines in Figure 7.4

(46)

7.2.1.3 Gloss measurement

Finally, the gloss measurement device was used to obtain the surface profile of the printed lines in terms of specular reflection. The measured gloss data was then plotted in Matlab and analyzed to obtain the gaps between lines. Figure 7.6 shows the gloss map of the same sample (Figure 7.2 and Figure 7.4) of single layer lines.

Figure 7.6: Gloss map of single layer lines of the sample in Figure 7.2

Although, some lines were visible in the gloss map (Figure 7.6), it was difficult to calculate the gap between lines. The gloss has maximum values while incident light was reflected back to the detector from both the top surface of the lines and the gaps between lines. Unlike the surface profile, the peaks in the gloss map correspond to those flat areas of the sample.

From the analysis of surface profile and photograph of single layer lines, it can be concluded that the gap between lines was consistent in both measurements. These results proves the accuracy of line scanning method.

7.2.2 Single layer lines

For the final measurements, Dektak 150 stylus profilometer was used to scan the lines instead of the Mitutoyo portable surface roughness tester because the Portable roughness tester can scan only a small length (up to 12.5 mm) but the Stylus profilometer can scan up to 3.8 cm. To analyze all 500 peaks, the length that had to

(47)

be scanned was ca. 7 cm. The measurement was done by Stylus profilometer by scanning the length twice from two opposite sides of the sample and then combining those results in Matlab.

Single layer lines from each side of the Head 2 (Side 0 and Side 1) were printed on the base layer. As each side contains 500 nozzles, each sample contains 500 single layer lines. After that, the sample was scanned across the lines at randomly chosen positions by the Dektak 150 stylus profilometer. Finally, the average droplet size from each scan was calculated. Figure 7.7 shows part of the original profilometer data printed by the Head 2, Side 0.

0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

2000 4000 6000 8000 10000 12000 14000

X−axis (mm)

Z−axis (nm)

Original stylus profilometer data Base line

Edge effect

Figure 7.7: Stylus profilometer data for single layer lines printed by the Head 2, Side 0

There is a increase in height of the base layer in the profile at the beginning shown in Figure 7.7, which is also present at the other end of the base layer. This is called as the edge effect of 3D printing. The edge effect was measured for 120 base layers printed on the glass substrate and shown in Figure 7.8.

(48)

Figure 7.8: Surface profile of 120 base layers printed on glass substrate by Head 2 The average height of the bump is ca. 20 µm. The edge effect increases as the number of layers increases. It was calculated that ca. 170 nm extra height adds up per layer because of the edge effect.

Figure 7.9: Schematic diagram explaining the edge effect

The droplets in the middle regions of the printed optics can spread in all directions but the droplets at the edge regions can only spread inside the printed region, not in all directions. Because of this nonuniform spreading at the edges, the height

(49)

Figure 7.9). The extra height at the edges increases about 0.17µm per layer because of the edge effect. There is also another bump (ca. 4 µm) present in the middle region. The reason behind the bump in the middle region, is explained later in section 6.2.3 below Figure 7.26.

To subtract the edge effect of base layers from the printed lines, the base line correction was done and the data was replotted in Figure 7.10. The base line correction also removes any nonuniformity on surface profile of the substrate or the base layers.

From the baseline corrected data, the area under each peak was calculated for all 500 lines ( = Nozzles). The scanning across the lines was performed 9 times and each from randomly chosen positions and finally, average area per peak was calculated from those measurements. The measurement was repeated for the Head 2, Side 1 and also for another set of sample (Sample 2: Head 2, Side 0 and Sample 2: Head 2, Side 1). The results of these two samples (printed on two consecutive days) were compared with each other to check the repeatability of the droplet size.

0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

0 500 1000 1500 2000 2500 3000

X−axis (mm)

Z−axis (nm)

Figure 7.10: Base line corrected stylus profilometer data of Figure 7.7 for single layer lines printed by the Head 2, Side 0

(50)

2 3 4 5 6 7 8 9 10 11 12 13 14 15 75

80 85 90 95

Sample 1

Nozzle number Area per droplet ( µm2 )

+10%

−10%

Average

±3.4%

±2.5%

±5% ±1.5%

±6.7%

Figure 7.11: Eight scanning results of lines printed on the sample 1 (Head 2, Side 0)

2 3 4 5 6 7 8 9 10 11 12 13 14 15

78 83 88 93 98

Nozzle number Area per droplet ( µm2 )

Sample 2

−10%

+10%

Average

±3.7% ±3.2%

±3.1%

±3.6%

±4.2%

Figure 7.12: Nine scanning results of lines printed on the sample 2 (Head 2, Side 0)

From the calculations and comparisons of the two samples (Figure 7.11 and Figure 7.12), it can be concluded that there was maximum±10% variation on droplet

(51)

size for the same nozzles. As shown in the graphs (Figure 7.11 and Figure 7.12), the variation in droplet size for nozzle 2 is±3.4% in sample 1 and±3.7% in sample 2. For nozzle 13, the variation in droplet size is±1.5% in sample 1 but it is±4.2% in sample 2. So, the deviation between these eight measurements of Sample 1 (Figure 7.11) and nine measurements of Sample 2 (Figure 7.12) seems quite significant (average

±1.5 %).

The deviation in droplet size from the same nozzle can be caused by instability of the droplet size from nozzles. To check the repeatability of the measurement, the two samples were printed on two separate days. There might be a change in temperature of print heads or printing material or other 3D printing conditions in this two consecutive days, which caused the variation in droplet size. Also, these eight (Sample 1) and nine (Sample 2) measured places of each sample was randomly chosen. This variation in scanning positions can be the cause of the variations of droplet size. If the scanning positions in both samples are same, then the variation in droplet size in two samples will be similar. So, accuracy control of the scanning positions is needed to obtain good results.

The average droplet size of sample 1 matches sample 2 by 95% in volume when fitted together (Figure 7.13).

100 105 110 115 120 125 130 135 140 145

82 84 86 88 90 92 94

Nozzle number

Area per droplet ( µm2 ) ^+5%

−5%

Average

Figure 7.13: Averaged droplet sizes, when results from the samples 1 and 2 are fitted together (the constant difference about 5% removed from the results)

(52)

Although there were similar kind deviations in nine measurements, which can be seen if the average of sample 1 and sample 2 were shifted to fit together (Figure 7.13).

The size of a single droplet from all 1000 nozzles of the Head 2 was obtained from the analysis of single layer lines from both sides of the head 2 and presented in Figure 7.14. This complete picture of droplet sizes helps to know the exact size of the droplets and the deviations could be compensated by adding proper compensation layers.

0 100 200 300 400 500 600 700 800 900 1000

80 85 90 95 100

Sample 1, Head 2, Side 0

Nozzle number ( 1 to 1000 ) Area per droplet ( µm2 )

+10%

−10%

Average

Figure 7.14: Average size of droplets from all 1000 nozzles of the Head 2

It can be concluded that the average droplet size in two samples is not consistent (varies about±1.5 % with each other) for each nozzle itself (Figure 7.11 and Figure 7.12) while the droplet size from nozzle to nozzle varies (±10%).

The results of scanning single layer lines show quite a lot variation (±10%) on droplet size from nozzle to nozzle and also variation more than ±5% for a single nozzle (Shown in Figure 7.11, nozzle number 11) along printing direction. Printing of optical components usually consists of several hundreds of layers. We know that there is an averaging of droplets comes in account due to the surface tension of the UV curable ink and multiple numbers of printed layers.

As the scanning positions were chosen randomly, sometimes the stylus head moved across the middle of a droplet giving more size of the droplet than when

(53)

the size of the droplet is bigger (scanning the middle region of the droplet) in line 1 but minimum (scanning in between two droplets) in line 2 for scanning position 1. Though, the size of the droplet is minimum (scanning in between two droplets) in line 1 but maximum (scanning the middle region of the droplet) in line 2 for scanning position 2.

Figure 7.15: Variation of droplet size based on scanning positions

7.2.3 20 layer lines

To account for these effects with scanning positions, 20 layer lines were printed and scanned with Dektak 150 surface profiler. The radius of the stylus head was not small enough (12.5 µm) to obtain the proper resolution for the measurement of 20 layer lines so, the lines from each side of the head were printed in two different sections. At first 20 layer lines from the odd nozzles were printed and then on another substrate 20 layer lines from the even nozzles were printed and scanned by the stylus. After that, both the results were combined to obtain the average area per droplet for all 500 nozzles of each side.

For each sample of even nozzles and odd nozzles, 10 scans, each from different positions were performed and the average size per droplet from the 10 measurements were calculated and plotted.

CALIBRATION OF THE NOZZLES OF 3D PRINT HEADS