Materials and Design 190 (2020) 108580 Contents lists available at ScienceDirect Materials and Design j ourna l homepage: www.e lsev ie r .com/ locate /matdes The challenge of upscaling paraffin wax actuators Arne Mann, Thiemo Germann ⁎, Mats Ruiter, Peter Groche ⁎ Institute for Production Engineering and Forming Machines, Technische Universität Darmstadt, Germany H I G H L I G H T S G R A P H I C A L A B S T R A C T • Presentation of a new concept of power- ful phase change actuators in mem- brane design • Description of the manufacturing method and challenges of the joining process • Characterization shows actuating forces of up to 65 kNwith high reproducibility • Compared to existing systemswith sim- ilar working principle systems, the working density is increased by a factor of 7 • Demonstrator application with passive actuator, compensating the loss of pre- tension due to thermal loads of a bolted joint ⁎ Corresponding authors. E-mail addresses: germann@ptu.tu-darmstadt.de (T. G groche@ptu.tu-darmstadt.de (P. Groche). https://doi.org/10.1016/j.matdes.2020.108580 0264-1275/© 2020 The Authors. Published by Elsevier Ltd a b s t r a c t a r t i c l e i n f o Article history: Received 23 November 2019 Received in revised form 14 February 2020 Accepted 17 February 2020 Available online xxxx Keywords: Paraffin wax Actuator Phase change material Machinery property adjustment Automation Higher levels of automation necessitate active spacer and adjusting elements generating high stroke forces. For these, the multitude of applications inside a manufacturing system requires a space- and cost-effective design. Conventional actuator concepts struggle with these demands. A new and efficient actuator concept for establish- ing closed-loop control circuits is needed. This article presents a newactuator concept, based on the phase change material paraffin wax. Although paraffin wax actuators are a convenient solution for microactuators, high force macroscopic actuators are not established yet. On a macroscopic scale the design of the actuator housing and themanufacturing process are challenging. The presented concept consists of a closed housing, which surrounds the phase change material. A compact actuator design without sealed moveable parts is realized. Thus, the actu- ators provide high axial forces. Compared to existing solutions an increase in performance by a factor of 7 could be achieved. The essential actuator structure is introduced, characterized and the challenges inmanufacturing are discussed. A possible application is demonstrated by a thermal compensation element activated by energy from the surroundings. © 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/). ermann), . This is an open access article under 1. Introduction Smaller batches as well as shorter delivery andmanufacturing times require a continuous innovation of the manufacturing technologies. Currently, industry 4.0 and the consequent digitalization of the indus- trial plants are promising approaches for higher level manufacturing systems [1]. More and more manufacturing processes are equipped the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). http://crossmark.crossref.org/dialog/?doi=10.1016/j.matdes.2020.108580&domain=pdf http://creativecommons.org/licenses/by-nc-nd/4.0/ http://creativecommons.org/licenses/by-nc-nd/4.0/ https://doi.org/10.1016/j.matdes.2020.108580 mailto:germann@ptu.tu-darmstadt.de mailto:groche@ptu.tu-darmstadt.de https://doi.org/10.1016/j.matdes.2020.108580 http://creativecommons.org/licenses/by-nc-nd/4.0/ http://www.sciencedirect.com/science/journal/ www.elsevier.com/locate/matdes Fig. 2. Energy density of different actuator materials (adapted by the author) [10]. 2 A. Mann et al. / Materials and Design 190 (2020) 108580 with sensor systems [2]. Process information is recorded and the col- lected data can contribute to an improved process understanding based on white or grey box models [3]. The next step in manufacturing research is using the deeper process understanding and the determined cause and effect relationships to establish controlled manufacturing processes, as shown by Alwood et al. [4]. The objective of automation at this level is to control product features with a closed loop control [5]. Several different actuator principles have been implemented in closed loop control circuits for the main drives as well as the auxiliary drives used for clamping and adjustment operations. Formain drives inmachine tools, actuators using hydraulic, pneumatic or electric drives are well established. In contrast, attractive and economical actuator concepts for auxiliary drives (optional process drive for e.g. automatization or adjust- ment) seem to be of short supply. Fig. 1 shows a classification of actuator tasks in dependency on the process' main drive actuation speed. Applica- tions with the need for higher speeds are often necessary to counteract the impact of stochastic process disturbances. Examples of stochastic dis- turbances are fluctuating material properties or states of lubrication. In these cases, a rapid process adjustment is required. Countermeasures ini- tiated by actuatorsmay lead to a direct compensation of the process devi- ations or a compensation in a following process step using closed loop or feed forward controls [4]. Since they also rendermanually operatedmachinery adjustments un- necessary, higher levels of automation in manufacturing [4] need control circuits beyond closed loop control of main drives. The targeted adjust- ment actuations of machine parameters are carried out with an actuating speed below the process cycle frequency (Fig. 1). Themain cause for nec- essarymachinery adjustments is givenbygradual shifts in systemproper- ties. Examples can be the repositioning of a tool that no longer generates the target geometry of a component due towear. The auxiliary drivesmay also compensate the deflection of deepdrawing tools [6], setting the guid- ing clearance or adjusting the tool mounting condition [7]. Other exam- ples are process shifts due to thermal expansion of machine components or deterioration of the fluid media's mechanical properties [8]. For this kind of application with low actuation speeds and power re- quirements, the actuators must be able to provide forces of several kilonewtons whereas typically small displacements in the millimetre range and reaction times of several seconds are sufficient. In the field ofmanufacturing process control, thermal expansion is yet an uncommon actuator principle. It seems however promising, especially when phase change materials are used. According to Ogden et al., who compared the energy densities of different actuator materials, phase change materials have the highest energy density compared to shape memory alloys, electromagnetic, thermopneumatic, bimetallic, electro- static and piezoelectric actuator materials (Fig. 2). The high energy den- sity is due to the high thermal expansion coefficient combined with a low compressibility [9]. A drawback of thermal actuators which is counteracting the common use in manufacturing processes is their low Fig. 1. Overview of necessary actuator dynamics for different applications in manufacturing industry. energy efficiency. However, by using the lost heat of processes or ma- chines, the low-efficiency of the thermal actuators is ofminor importance. The selected material for this investigation is paraffin wax. Paraffin wax is a phase change material, which has a significant volume increase when changing from solid to liquid state. Its small compressibility qual- ifies it to be an especially suited phase changematerial. Paraffinwax is ac- tivated by applying heat [10]. For instance, the paraffin wax used in this publication (SIGMA-ALDRICH paraffin wax mp 58–62 °C) shows a free volume expansion of approximately 15% at 75 °C compared to 25 °C [11]. The activation time depends on the time required for heat supply to and distributionwithin the paraffinwax. The low thermal conductivity of the paraffin wax (0.21 W/mK [12]) leads to a rather inert actuator be- haviour. Thismeans for short response times, e.g. for process adjustments (compare Fig. 1), the amount of paraffin is to be kept low (miniaturized actuators). Additionally the thermal conductivity can be increased by ad- ditives [13], enabling faster response times. Fig. 3. Volume expansion data of SIGMA-ALDRICH paraffin wax mp 58–62 °C [11]. 3A. Mann et al. / Materials and Design 190 (2020) 108580 For current actuator designs, there are different options to convert the expansion of the phase change material into a force transmission. Ogden et al. divide the possible mechanical transmission elements in pistons, membranes, composites and direct contact [10]. Today paraffin wax is commonly used for miniaturized phase change actuators. Anotherfield of application of paraffinwax is the integration as a func- tional material in smart structures. Recent examples are temperature re- sponsive surfaces, which enable for passive and active droplet motion control aswell as anti-fogging control onmiscellaneous surfaces. The par- affin wax determines the slipperiness of the material by its temperature dependent structure [14]. The thermal properties of paraffin wax are fur- thermore employed in realizing artificialmuscles. Their ability towet car- bon nanotubes leads to the use as thermal activated fibers [15]. Further investigations show the high potential of tailored fillers by blending par- affin wax with other polymers [16]. An overview of the current research of twisted nanofibers is provided by Mirvakili and Hunter [17]. Another approach is the use of the paraffins volume expansion asMcKibbenmus- cles. A tailored braid of intertwined steel and cottonwith a paraffin filling generates a contraction by heating the paraffin wax [18]. Nevertheless the first applications of paraffin wax known to the au- thorswere used as temperature sensitive closures in sealingmeans [19]. Today's often employed macroscopic paraffin actuators based on a pis- ton design can be ascribed to Sherwood, who invented the electrically heated paraffin wax actuator in the 1960s [20]. The use of a piston is known from hydraulics and pneumatics. In- side piston-cylinder systems, relative movements between the pis- ton and the cylinder take place due to a change of volume in the chamber between piston and cylinder. The contact area between cyl- inder and piston has the function of sealing the active fluid media and often also of guiding the piston during its movement. In this ap- plication, ring seals are often used. Tibbitts presents a commonly used cylinder piston design for a paraffin wax actuator [21]. In com- parison with the actuator concepts described above, the actuator housing is significantly larger. The cylindrical housing has a length of 58 mm and a stroke of 25 mm, whereby the paraffin chamber is sealedwith an O-ring fixed with a screw thread. The actuator reaches a maximum force of 160 N. Another cylinder piston design actuator presented by Kabei et al. has a paraffin chamber length of 90 mm and a diameter of 2 mm [22]. The paraffin is sealed in a silastic tube. Under free expansion the actuator reaches a maximum stroke between 9 and 10 mm, which decreases slightly under a load of m = 1000 g. The subsequently presented patent describes the com- mercially used actuator design. The cylinder-piston actuator design is similar to the one proposed by Kabei et al., which seals the paraffin with a membrane or an elastomer insert [23]. The housing is joined by a mechanical bond. The actuator reaches a maximum force of 1500 N. During the design of such a system, the sealing technology and the loads during use are important design criteria, as the sealing and guiding function can only be maintained within narrow geomet- ric tolerances. Since the cylinder is designed as an open component, Fig. 4. Different types of paraffin actuators: Polymer stencil design (a) [10], the necessary stiffness must be generated by a large wall thickness. In conclusion, the cylinder-piston system seems to be complex, al- though the piston design can realize large forces and large displace- ments at the same time. In addition to ring seals, membranes are used in piston-cylinder paraffin actuators for sealing the cylinder [23]. Using amembrane design, themembrane seals the paraffinwax in a cavity and deflects during an actuator stroke. Since the amount of paraf- fin significantly determines the response behavior of the actuators, par- affin wax phase change actuators with larger membrane designs are uncommon today. The maximum force of actuators with a membrane is determined, among other factors, by the properties of the membrane and its bonding to the paraffin cavity. The force transmission and the displacement are caused by an elastic deformation of the membrane. The achievable maximum force limits the usage to specific applications as valves and pumps [10]. The operating behaviour of an actuator in membrane design results by the interaction of the individual system components. The use of phase change materials within microactuators, mircopumps and microvalves is quite common. Due to the high en- ergy density and the good availability, paraffin wax is often the cho- sen material. Typically, the paraffin wax is encapsulated within a solid housing with a flexible part, e.g. a membrane, to apply the actu- ation [10]. In a review paper Ogden et al. describe three material classes, which are typically used for microscopic phase change actu- ators (compare Fig. 4). The most common design in microfabrication bases on silicon and glass. They both exhibit high stiffness and yield strength, which predestines them for high-force or high-pressure applications. A strong sealing was achieved with anodic bonding or fusion of different layers. Examples are the concept of Klintberg et al. [24,25] or Carlen and Mastrangelo [26]. Alternatively, polydi- methylsiloxane PDMS or polymers in general are employed. Their advantage lies in simple processing and, for polymers, high stiffness as well as yield strength. This design is typical for microvalves and micropumps as shown in Bóden et al. [27] or Svenson et al. [28]. The third material class consists of metals, especially stainless steels. It is characterized by appropriate high mechanical properties but ex- hibits difficulties if a direct bonding is targeted. Typically, a poly- meric bonding agent is used as an intermediate layer [10]. Examples are summarized by Ogden et al. [10] or Lehto [29]. Despite the described challenges, there aremacroscopic actuator de- signs, which are commonly used today. A typical application of macro- scopic paraffin wax actuators are thermostatic working elements such as thermal switches and mixing taps [10]. These actuators in piston de- sign can achieve high displacements, but they have restrictedmaximum forces. Due to the typical usage as thermostatic working elements, a maximum force was not specified as a target development parameter [30]. Thereby those approaches are not sufficient for the presented ap- plication, i.e. adjusting machinery properties (compare Fig. 1), since it requires a high stroke force of several kilonewton. silicon glass design (b) [31], piston design (c) [23] (graphics adapted). Fig. 5. Design of the phase change actuator in a cross-sectional view. Table 1 Geometric parameters of the actuator housing Inner cup (I) Outer cup (O) Diameter di in mm 27.00 28.75 Thickness bottom tb in mm 1.00 1.00 Thickness wall tw in mm 0.80 0.80 Radius rI, rO in mm 1.00 0.50 Material [−] DP600 DP600 4 A. Mann et al. / Materials and Design 190 (2020) 108580 To evaluate the efficiency of the presented actuator concepts a com- parative figure is needed. Introducing thework density, Krulevitch et al. enable a useful comparison between different types and sizes of actua- tors. It is defined as the work output per unit volume Wd [32]. It is de- fined by the filling volume V, the maximum deflection dmax and the actuating force at this point Fmax. Wd ¼ Fmaxdmax 2 1 V ð1Þ Ogden et al. presented a comparison of known actuator concepts. Summarising the investigation results, the commonly used membrane design actuators achieve a work density between 7.2 ∗ 103 J/m3 [33] and 93 ∗ 103 J/m3 [34]. Due to the stiffer structure of the housingwithin a piston design, the achievable work density is considerably higher (260 ∗ 103 J/m3 [30]). Still, there is a gap by an order of magnitude be- tween current actuators and the potential of paraffin wax (compare Fig. 2). Therefore, upscaling the structure of microactuators is a promis- ing approach. The objective of this research is to realize amacroscopic paraffinwax actuator for high axial forces. An upscaling of the silicon glass design is not feasible due to process limits of anodic bonding even though the high stiffness and yield strength provide the necessary basis for macro- scopic actuators. Therefore, PDMS actuators are no applicable solutions, as their housing stiffness is not sufficient for the intended force range of the actuator. The third housing material class defined by Ogden et al. is metal. Thematerial properties of themetal housingmight be promising. However, since the design target of piston actuators are high displace- ments, whereas the stroke force is subordinated, the piston design is not sensible for upscaling [30]. Another design is an actuator structure out of a stack of metal stencils and polyimide foils, which creates a very high stiffness. Therefore, this design is e.g. the one of choice for micropumps [34]. However, upscaling this design to utilize a macro- scopic amount of paraffinwax results in an inapplicable housing. There- fore, a new approach is necessary to design amacroscopic phase change actuator which provides high actuation forces within a relatively small displacement and limited housing space. This study presents a new design of macroscopic phase change ma- terial actuators, which provides solutions for the addressed application. 2. The closed laser-welded phase change material actuator Based on the aforementioned application areas the requirements of a targeted paraffin actuator can be defined. In order to affect the proper- ties of manufacturing machines, a sufficient actuation force is the most important parameter. Compared to that, a short reaction time is of less importance (compare Fig. 1). The compensation of geometrical changes due to thermal stress or wear requires a high stiffness of the actuator, though the displacement can be rather small [8]. Under the described boundary conditions, the membrane principle appears to be promising. Due to the closed design, themaximum forces are limited by the used materials and joining technologies. The design presented below is intended to withstand high internal pressures due to a closed sheet metal housing and a metallurgical high-strength joint created through laser welding. Fig. 5 illustrates the design of the phase change actuator concept for high forces. The design of the described actuator concept is mainly af- fected by the necessity to implement a metallurgical bond to join the housing. Through the metallurgical bond, the housing withstands the working pressure and a sealing of the housing is guaranteed. The reali- zation of the bonding by laser welding minimizes the amount of heat applied. Still the housing needs to be sealed during thewelding process, due to the increasing pressure resulting from the heat transfer from the weld area into the paraffinwax. For this reason, a cutting seal is included in the actuator housing design. The setup of the actuator is axisymmetric and consists of two deep drawn cups (inner and outer cup) with the target geometrical dimen- sions defined in Table 1. The cup diameters are designed to ensure that the inner cup fits into the outer cup. For each cup, specific deep drawing tools are implemented and the steel grade DP600 with thick- ness of 1 mm is used. A hydraulic press realizes the deep drawing pro- cess during which an additional counter punch for a plane cup bottom is applied. The deep drawing process is followed by a turning process, which reduces the cup height to a constant height all over the perime- ter. At the inner cup's tip, a chamfer is positioned. The inner cup is filled with liquid paraffin wax (SIGMA-ALDRICH paraffin wax mp 58–62 °C). During the cooling-down, additional paraffin wax is poured into the cup to compensate the lacking volume in the cup due to shrinkage. Fill- ing the actuator in its entirety is achieved by a continuous refill up to a surplus. At the end of the filling process, excess paraffin wax above the cup edge is removed. Inside the outer cup, a brass sheet is placed (thick- ness 50 μm). Both cups have a fixed height, which leads to a weld area between the outer cup's tip and the radius of the inner cup (Fig. 5). Be- tween both cups results a contact line at the inner cup's tip and the outer cups bottom, which is positioned opposite to the weld area and represents the sealing. Due to an axial force during the joining process, the chamfer of the inner cup cuts into the brass sheet and establishes a sealing function as a cutting seal. The axial force can control the bearable sealing pressure. Both cups are joined by welding. Thewelding of the filled actuator cups described in this paragraph is the most critical part of the production chain. For the welding process, an YLS-300-S2T laser source with 3 kW maximum output is used. For a positive outcome of the welding process, the key parameter is the en- ergy per unit length. Due to the round shape of the actuators, a station- ary welding point is combined with a rotating clamping of the actuator. The feed speed is defined by the rotations per second and the shape of the actuator. Still there are various other parameters to adjust as the focus position, the welding angle or the axial locking force. The energy per unit length is discussed in detail while the other parameters have been fixed as shown in Table 2; the results are shown in Fig. 6. Table 2 Process parameters of the laser welding process of the actuator housing. Process parameter Welding angle in ° 30.0 Axial locking force in kN 2.0 Paraffin wax filling volume in % 99.0 Diameter weld spot in mm 1.2 Overlap weld seam in % 0.75 Fig. 7. Structure of the test bench and measuring points of the parameters. 5A. Mann et al. / Materials and Design 190 (2020) 108580 The process window describe by the energy per unit length is char- acterized by three typical cases (compare Fig. 6). The first is defined with an insufficient energy per unit length. It occurs when the feed speed is high in relation to the laser power. The outcome is an unsteady weld seam,which only partly joins the components. This unsteadyweld seam is due to an insufficient weld pool, either the inner or the outer cup is not sufficiently melted. Predominantly, the outer cup shows a molten edge, whereas the inner cup is not affected. If the energy per unit length is too high, the resulting weld seams are equally unaccept- able. This second case can be subdivided in three cases by the increasing laser power (compare Fig. 6). If a low laser power and feed speed is used, the process time is long enough for the paraffin to build up a sig- nificant amount of liquid phase. The resulting volume expansion lifts the cups of the actuator against the axial locking force. Therefore, liquid paraffin wax emerges and inhibits the welding. Contrary a very high laser output effects an immediate spraying of the paraffin through the molten actuator cups. This effect results from a local pressure increase at the weld line due to the high amount of applied heat. After the solid- ification the weld seam still shows points of separation.With a medium high laser power hybrid forms of both processes where detectable. The third case is the right relation of feed speed and laser power. In this sit- uation strongweld seams are achievable. Case 3 in Fig. 6 defines thepro- cess window for the manufacturing of macroscopic paraffin wax actuators. 3. Experimental characterization of the phase change actuator In the following section an examination procedure for the character- ization of the novel actuator is presented. Therefore, the used test setup and test cycle are explained. It is used for the determination of achiev- able forces. Additionally, a force-displacement characterization will be Fig. 6. Process window of the actuator manufactu discussed. Subsequently the hysteresis behaviour is characterized. Since the material behaviour of paraffin wax is strongly temperature- depended four different temperatures close to the melting point are considered for the characterization. The examination takes place in a combined tensile compression test machine (Zwick Roell 100) using an upsetting tool for applying a com- pression force F to the actuators top and bottom contact area. During the heating, the positions of both pressure plates arefixed by anoptical path control, which prevents axial movements. The resulting axial compres- sion force F is recorded. Due to the optical measurement at the upper and lower pressure plate, the mechanical compliance of the test ma- chine is negligible, as a repeatable testing with a pre force of F0 = 500 N is applied prior to the heating. A heater band (270 W; 86.12 mW/ mm2) is used to heat the actuator around its housing shell (compare Fig. 7). The target temperature θt is set and the heater band is powered continuously until the target temperature is reached. The temperature is kept at the target temperature θt by a control for the heating interval th. The control temperature for the target temperature θt is measured between heater band and actuator housing in an additional aluminum ring serving as heat conductor and sensor mount. The ring includes three evenly distributed thermocouples. The average temperature of these thermocouples is labelled as the heater band temperature θhb and is used as a feedback signal in the closed loop control. The alumi- num ring is slotted to prevent a radial support of the actuator. An addi- tional thermocouple is placed at the bottom centre of the actuator and measures the actuator temperature θa. ring according to the energy per unit length. Fig. 9. Complete actuator characterization ha = 7 mm, actuator temperature θa and compression force F, the temperature characterization ranges are marked, start- temperature θ0 = 30 ° C 6 A. Mann et al. / Materials and Design 190 (2020) 108580 The tested target temperatures θt are θt = 50 °C, 60 °C, 70 °C, 80 °C. Each temperature trial consists of four cycles. Hereby, each temperature trial for a fixed actuator height and temperature consists out of four heating and cooling intervals. Fig. 8 exemplifies the testing procedure. Every temperature trial consists of the target temperature θt, the heating interval th, the cool down temperature θc and the number of repetitions n (n = 4). The experimental characterization starts at room temperature. At first, the heater band temperature θhb is set to the cool down temperature θc. After reaching the cool down tempera- ture, the heater band temperature is set to the target temperature and the characterization starts. The heating lasts for th = 1200 s. Fig. 8 shows an increasing heater band temperature θhb, which oscil- lates above the target temperature θt. After the heating phase, the free cooling starts. When the cooling temperature θc is reached, the next cycle starts at this point. A complete characterization procedure of an actuator with the com- pression force F and the actuator temperature θt is presented in Fig. 9. The target temperatures θt are tested one after the other (θt = 50 °C, 60 °C, 70 °C, 80 °C). Each target temperature is set four times. The com- plete experimental characterization took place over a period of about 11.5 h, for the sake of clarity the experimental progress is given instead of the time. At the beginning, the target temperature leaps to θt = 50 °C. Fig. 9 shows the corresponding actuator temperature θa. The test setup con- sists of highly heat conductive parts, therefore a temperature difference between actuator and target temperature remains even after the heating time th= 1200 s. After the heating time, the target temperature is reduced to θt = 40 °C for the cool down (for θt = 70 °C and θt = 80 °C the cool down temperature is set to θt = 45 °C). The cooling is clearly visible by a decreasing actuator temperature θa. Hereafter, three further cycles can be seen. The described procedure is repeated for the follow- ing target temperatures θt = 60 °C, 70 °C, 80 °C. Furthermore, the corre- sponding compression force F is displayed. For θt = 50 °C the change in force is minimal, but a clear trend of force F corresponding to the actu- ator temperature θa is visible. For θt = 60 °C, the force increases rapidly, following the actuator temperature θa. After an increase in force, the gradient of the force graph decreases. The decreasing force gradient is due to the heater band temperature θhb, which reached the target Fig. 8. Example for the test method, two cycles for ha = 7 mm, target θt and heater band temperature θhb. temperature θt and is constant from this moment on. The heat supply is reduced, as the target temperature θt only has to be maintained. The steady-state of the heat supply results in the force F approaching a lim- iting value. The same qualitative behavior can be observed at the actua- tor temperature θa with a time offset. When the cooling starts, a rapid decline in the compression force F and actuator temperature θa is ob- served. The compression force decreases until it reaches aminimum, al- though the actuator temperature θa is decreasing further. The course of the actuator temperature θa mainly corresponds to the course of the compression force F. The heating time is chosen in such a way that the compression force F shows a minimal gradient. The temperature distri- bution in the actuator seems to have reached an almost steady state. Furthermore, the qualitative course of the graph is similar for all cycles and tested temperatures. The heating and cooling rate-limiting factor of the actuator is the thermal conductivity of the paraffinwax. To enhance the reaction or cooling time there are two possibilities. Additives for the paraffin, i.e. copper wool, improve the thermal conductivity within the actuator. Alternatively, the length of the thermal pathswithin the paraf- fin wax could be reduced by the integration of thermal conductive structures, i.e. stringers, in the housing. Preliminary investigation con- firms the potential of those structures. Therefor these countermeasures are subject of subsequent research. Table 3 summarizes averagemaximum forces Fmax for the tested tar- get temperatures θt and standard deviations for all tests. The small stan- dard deviations indicate a good reproducibility. The compression force F increases with increasing θt. Table 3 Summary of the maximum compression forces Fmax, standard deviations σ and heating times th Temperature θt in °C Average maximum force Fmax in kN (σ in kN) Heating time th in s 50 0.96 (0.034) 1200 60 16.47 (2.084) 1200 70 39.64 (0.053) 1200 80 62.57 (0.207) 1200 Fig. 10. Compression force displacement diagram for the actuator height ha = 7 mm. The maximum displacement is reached with a velocity of �w ¼ 0; 01 mm=s. 1 The specific heat of the paraffin usedwas determined by dynamic differential calorim- etry as proposed by Schimmelpfennig et al. [36]. The paraffins latent heat is measured by Schaerer et al. [37]. Table 4 Efficiency characteristic for the actuator with different temperatures θt Temperature θt in °C Positioning work W in Nm Minimum heat quantity E in J Actuators efficiency η in [−] 50 0.035 430.4 0.008% 60 0.81 1129.2 0.072% 70 2.43 1708.8 0.142% 80 4.43 1838.0 0.241% 7A. Mann et al. / Materials and Design 190 (2020) 108580 At the temperature θt = 50 °C, themelting temperature of paraffin has not been reached yet. Therefore, the volume increase is minimal, resulting in a slight force increase. For the target temperature θt=60 °C a significant increase of the axial force is detectable. In addition, the standard deviation is the highest of all tests. For the first three repetitions, the results are con- sistent. The fourth repetition showsa significant increase in the force (com- pare Fig. 9). The target temperature θt = 60 °C lies within the melting temperature of theparaffinwax. Obviously, the steady-state of the temper- ature distribution in the paraffin wax has not yet been set during the first three cycles using θt = 60 °C. An increasing temperature and a progressive melting of the paraffin wax explain the increase in force. θt = 70 °C and θt = 80 °C continue to display an increasing compression force F. At θt = 70 °C, the actuators show the same behavior as with lower temperatures but at a higher force level. At θt=80 °C, the actuator reaches themaximum compression force with nearly Fmax = 63 kN. All prior described tests are performed with a suppressed actuator dis- placement. In the final test, the effect of an actuator displacement is inves- tigated (Fig. 10). An actuator (ha=7mm) is heated for t=1200 s to reach the target temperature θt. The compression test machine initiates a dis- placement of w = 0.1 mm in direction of action of force with 0.01 mm/s followed by a change in direction returning to w = 0.0 mm using the same speed and recording the compression force F. Due to the optical dis- placement control an influence of a frame deflection can excluded. The ac- tuator is heated for the next t = 1200 s to reach the next target temperature θt and the described displacement procedure is repeated. For θt=50 °C, the displacement ofw=0.1mm leads to a complete unloading of the actuator with a resulting compression force F= 50 N after the dis- placement cycle is completed. For θt = 60 °C, a decreasing compression force for an increasing displacement is observed. The compression force shows a nearly linear decrease from F=11.8 kN to F=4.7 kN. The change in testing direction leads to an increasing compression force resulting in F=13.6 kN forw=0.0mm. The start and end displacements show a dif- ference in the compression force. Thus, the graph shows a hysteresis. The target temperatures θt = 70 °C and θt= 80 °C show similar results, whereby higher target temperatures lead to an increasing difference of the compression force at the start and end of the displacement. When the starting point of the displacement cycle is reached, the axial force de- creases at a constant temperatureuntil it is equal to the applied forcebefore thedisplacement. Furthermore, thegraph showsanonlinear slope. Thedis- placementmeasurementhas a standarddeviationofΔwdev=0.00044mm (Δwmin=−0.00150mm,Δwmax=0.00054mm),whereas an influence of acceleration and deceleration processes of the compression test machine cannot be excluded. The hysteresis may be induced by a temperature in- crease due to the compression, which cannot be measured with the given set-up. Therefore, the hysteresis is not necessarily linked to the actu- ator or paraffin wax behavior, whereas the literature suggests a tempera- ture dependent hysteresis behavior of paraffin wax [35]. Relaying on the work density defined by Krulevitch et al. it is possible to evaluate the efficiency of the actuator concept. For this purpose, the maxi- mum determined actuating force F at a displacementw of 0.1 mm is com- pared to the paraffin filling volume V of the actuator. The volume of the paraffin filling is determined by the geometric parameters of the actuator housing (see Fig. 3 and Table 1). The volume of the paraffin core of the actu- ator is V=2845mm3, with the deflection of d=0.1mm and a maximum force at this point of F=37.5 kN, theworkingdensity is calculated as follows. Wd ¼ Fmaxdmax 2 1 V ¼ 37500 N � 0:1 mm 2 1 2845mm3 ¼ 659 J=m3 ð2Þ In relation with the work densities presented by Ogden for known phase-change-material actuators, the presented actuator concept shows a considerable increase in the achievable work density. In com- parison to the strongest actuators mentioned (compare [34]) using the membrane principle, an increase by a factor of 7 was achieved. Even in comparison with the structurally stiffer actuators in piston de- sign (compare [30]), an increase by a factor of 2.5 can be achieved. Based on the force displacement characterization, an estimation of the actuators ideal efficiency is possible, as shown in Table 4. The work of the actuators is determined by integrating the compression force F over the displacement w. It is compared with the required amount of heat to warm the paraffin wax up to θt. The heat quantity re- sults from the filling volume, the paraffin's specific and latent heat,1 the housing's specific heat during the heating from θ0 = 27 ° C at θt. The as- sembly condition and thermal isolation determine the divergence to the estimated efficiency. In comparison with typical data of thermal actua- tors it is noticeable, that the actuators efficiency increases with higher temperatures and subsequently higher loads. Furthermore, the actua- tors efficiency reaches at θt = 80 ° C the upper limit of reported effi- ciency of thermal expansion actuators [38]. The efficiency further increases with higher temperatures θt. 4. Application of a thermal compensation element The applicability of the described actuator concept is demonstrated in this section. It illustrates the functionality of a paraffin actuator as a passive element, i.e. a compensating element without separate energy supply. Fig. 11. Test set-up for investigating the thermal compensation of pre-loads (a), application scenario for the actuator in a tool-clamping situation in manufacturing machines (b) 8 A. Mann et al. / Materials and Design 190 (2020) 108580 As mentioned before, tool clamping has a significant influence on the result of manufacturing processes [7]. In many cases, the perpetuation of the pre-stress of a clamping under variation of the thermal conditions is required (Fig. 11 b). In an experimental setup, a typical situation is emu- lated by a frame combining different materials having different coeffi- cients of thermal expansion (Fig. 11 a). The paraffin wax actuator is integrated to compensate this difference in thermal expansion of thema- terials and therefore to maintain the pre-stress of the system (Fig. 11 a). A pre-loaded screw introduces a pre-stressing between an aluminum frame and a steel bar. A load cell captures the generated pre-load Fp. As the thermal expansion coefficient of aluminum is approximately twice the thermal expansion coefficient of steel [39], a homogenous heating of the structure without an actuator results in a decreasing pre-load of the system (Fig. 12). Fig. 12. Pre-load compensation for an aluminum-steel combination placed in an oven in comparison to the same setup without an actuator Fig. 12 shows the results for a pre-load of Fp = 2.5 kN. In accordance with the discussion above, the set-up without an actuator shows a de- creasing pre-load with increasing temperature. At about 1500 s, the minimum pre-load for an oven temperature of θ = 50 °C is reached with Fp = 1.5 kN. In the second trial, a paraffin wax actuator is placed in the flux of force and is pre-loaded with Fp = 2.5 kN. As previously seen, the pre-load is decreasing until 1000 s in the oven at θ = 50 °C. However, after this point in time the paraffin wax inside the actuator starts tomelt and thus the pre-load is recovering to Fp= 2.5 kN. The de- crease of pre-load, initiated by the deviant thermal behavior of alumi- num and steel is compensated. The delayed thermal compensation is due to the thermal isolation by the guiding hull (Fig. 11 a). In a third trial the oven temperature is set to θ=60 °C, which leads to an increase in pre-load of nearly Fp = 2.8 kN in addition to the thermal compensa- tion. The results show that the properties of a passive paraffinwax actu- ator can be adapted to various scenarios. 5. Conclusion This paper presents a solution to tackle the special challenges in upscaling paraffin wax actuators. The actuator concept fulfills the re- quirements of a space and cost-efficient actuator that provides high stroke forces. A manufacturing process that serves the challenging join- ing process is presented; it provides a solid and dense, high-strength connection without melting the phase change material core of the actuator. The investigations carried out show that the presented compact actu- ator design enables to apply high forces typically acting inmanufacturing devices for clamping and adjustment tasks. Therefore, machinery prop- erty adjustments can be performed with paraffin wax actuators. The ex- hibited sheet metal structure consists of a cutting seal and a metallurgical welded joint. Forces of almost 63 kN can be achieved with an actuator with the diameter of 30 mm and the height ha = 7 mm. The temperature-dependent behavior of the actuator with blocked axial expansion was recorded. The highest tested temperature is θt = 80 °C. A further increase in temperature would probably result in a further in- crease in force. In turn, the actuator forces are limited by the rigidity and strength of the actuator housing. Furthermore, the compression force displacement diagram for the actuator of the height ha = 7 mm is recorded. Even if an actuator displacement of w = 0.1 mm is employed, a compression force of F=37.5 kN can still be reached, though a hyster- esis occurs. The comparison of the developed actuator concept with existing actuators in membrane design by means of the work density in- dicator shows an increase in power density by the factor 7. Finally, the actuator is integrated into an application scenario, which shows the potential of the new actuator concept. Data availability The raw/processed data required to reproduce these findings are available from the corresponding author upon reasonable request. Author contribution As being the authors of this research paper, we declare that thework is entirely original, any raw data exiting in the article can be provided. Besides, the work has never been submitted/published in any other journal. Authors' individual contributions 1. ArneMann: Execution and processing of the experimental investiga- tions and results 2. Thiemo Germann: Processing of the experimental results 3. Mats Ruiter: Execution and evaluation of welding tests 4. Prof. Dr.-Ing. Dipl.-Wirtsch.-Ing. Peter Groche: Supervision of the re- search, revision of the publication 9A. Mann et al. / Materials and Design 190 (2020) 108580 CRediT authorship contribution statement Arne Mann: Investigation. Thiemo Germann: Validation. Mats Ruiter: Investigation. Peter Groche: Supervision, Writing - review & editing. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influ- ence the work reported in this paper. Acknowledgements The authors want to express their gratitude to the DFG (Deutsche Forschungsgemeinschaft) for the support of the grand “Designmethods for novel, energy-efficient, closed phase change actuators with high ac- tion of force” (GR 1818/65-1). References [1] D.Y. Yang, M. Bambach, J. Cao, J.R. Duflou, P. Groche, T. Kuboki, A. Sterzing, A.E. Tekkaya, C.W. Lee, Flexibility in metal forming, CIRP Ann. 67 (2018) 743–765, https://doi.org/10.1016/j.cirp.2018.05.004. [2] Hohmann, J.; Schatz, T.; Groche, P. Intelligent wear identification based on sensory inline information for a stamping process. In Proceedings of 5th International Con- ference on Advanced Manufacturing Engineering and Technologies: NEWTECH 2017; Majstorovic, V., Jakovljevic, Z., Eds.; Springer International Publishing: Cham, s.l., 2017; pp 285–295, ISBN 978-3-319-56429-6. [3] W. Volk, P. Groche, A. Brosius, A. Ghiotti, B.L. Kinsey, M. Liewald, L. Madej, J. Min, J. Yanagimoto, Models and modelling for process limits in metal forming, CIRP Ann. (2019) https://doi.org/10.1016/j.cirp.2019.05.007. [4] J.M. Allwood, S.R. Duncan, J. Cao, P. Groche, G. Hirt, B. Kinsey, T. Kuboki, M. Liewald, A. Sterzing, A.E. Tekkaya, Closed-loop control of product properties in metal forming, CIRP Ann. 65 (2016) 573–596, https://doi.org/10.1016/j.cirp.2016.06.002. [5] P. Groche, F. Hoppe, T. Kessler, A. Kleemann, Industrial Working Environment 2025. NEBU 2018: Neuere Entwicklungen in der Blechumformung, 2018. [6] T. Bäume,W. Zorn, W.-G. Drossel, G. Rupp, Iterative process control and sensor eval- uation for deep drawing tools with integrated piezoelectric actuators, Manufactur- ing Rev 3 (2016) 3, https://doi.org/10.1051/mfreview/2016002. [7] D. Kraus, M. Lieberenz, P. Groche, Reduction of tool wear by systematic design of the tool clamping situation, J. Manuf. Process. 28 (2017) 449–456, https://doi.org/10. 1016/j.jmapro.2017.04.011. [8] K.Wegener, J. Mayr, M. Merklein, B.-A. Behrens, T. Aoyama, M. Sulitka, J. Fleischer, P. Groche, B. Kaftanoglu, N. Jochum, et al., Fluid elements in machine tools, CIRP Ann. 66 (2017) 611–634, https://doi.org/10.1016/j.cirp.2017.05.008. [9] P. Zoller, D.J. Walsh, Standard pressure-volume-temperature data for polymers, Technomic Publ, Lancaster, 1995, ISBN 1566763282. [10] S. Ogden, L. Klintberg, G. Thornell, K. Hjort, R. Bodén, Review on miniaturized paraf- fin phase change actuators, valves, and pumps, Microfluid. Nanofluid. 17 (2014) 53–71, https://doi.org/10.1007/s10404-013-1289-3. [11] A. Mann, C. Bürgel, P. Groche, A modeling strategy for predicting the properties of paraffin wax actuators, Actuators 7 (2018) 81, https://doi.org/10.3390/act7040081. [12] H. Mehling, L.F. Cabeza, Heat and Cold Storagewith PCM. An up to Date Introduction into Basics and Applications; with 28 Tables, Springer, Berlin, Heidelberg, 2008. [13] Y. Jin, Q. Wan, Y. Ding, PCMs heat transfer performance enhancement with ex- panded graphite and its thermal stability, Procedia Engineering 102 (2015) 1877–1884, https://doi.org/10.1016/j.proeng.2015.01.326. [14] C. Chen, Z. Huang, Y. Jiao, L.-A. Shi, Y. Zhang, J. Li, C. Li, X. Lv, S.Wu, Y. Hu, et al., In situ reversible control between sliding and pinning for diverse liquids under ultra-low voltage, ACS Nano 13 (2019) 5742–5752, https://doi.org/10.1021/acsnano.9b01180. [15] M.D. Lima, N. Li, M. Jung de Andrade, S. Fang, J. Oh, G.M. Spinks, M.E. Kozlov, C.S. Haines, D. Suh, J. Foroughi, et al., Electrically, chemically, and photonically powered torsional and tensile actuation of hybrid carbon nanotube yarnmuscles, Science 338 (2012) 928–932, https://doi.org/10.1126/science.1226762. [16] K.-Y. Chun, S. Hyeong Kim, M. Kyoon Shin, C. Hoon Kwon, J. Park, Y. Tae Kim, G.M. Spinks, M.D. Lima, C.S. Haines, R.H. Baughman, et al., Hybrid carbon nanotube yarn artificial muscle inspired by spider dragline silk, Nat. Commun. 5 (2014) 3322, https://doi.org/10.1038/ncomms4322. [17] S.M. Mirvakili, I.W. Hunter, Artificial muscles: mechanisms, applications, and chal- lenges, Adv. Mater. Weinheim. 30 (2018) https://doi.org/10.1002/adma.201704407. [18] D. Sangian, J. Foroughi, S. Farajikhah, S. Naficy, G.M. Spinks, A bladder-free, non- fluidic, conductive McKibben artificial muscle operated electro-thermally, Smart Mater. Struct. 26 (2017), 15011. https://doi.org/10.1088/1361-665X/26/1/015011. [19] . Vernet, S. Sealing Means. US2368181A, Jan 30, 1945. [20] J.F. Sherwood, Devices for Utilizing the Thermal Expansion of Wax. US2815642A, Dec 10, 1957. [21] S. Tibbitts, High output paraffin actuators: Utilization in aerospacemechanisms, The 22nd Aerospace Mechanisms Symposium, NASA. Langley Research Center 1988, pp. 13–28. [22] N. Kabei, M. Kosuda, H. Kagamibuchi, R. Tashiro, H. Mizuno, Y. Ueda, K. Tsuchiya, A thermal-expansion-type microactuator with paraffin as the expansive material. (basic performance of a prototype linear actuator): basic performance of a proto- type linear actuator, JSME Int. J. Ser. C Mech. Syst. Mach. Elem. Manuf. 40 (1997) 736–742, https://doi.org/10.1299/jsmec.40.736. [23] Behr Thermot-tronik GmbH, Dehnstoffarbeitselemente: Produktkatalog Behr Thermot-tronik, GmbH, 2014. [24] L. Klintberg, M. Karlsson, L. Stenmark, J.-Å. Schweitz, G. Thornell, A large stroke, high force paraffin phase transition actuator, Sensors Actuators A Phys. 96 (2002) 189–195, https://doi.org/10.1016/S0924-4247(01)00785-3. [25] L. Klintberg, M. Karlsson, L. Stenmark, G. Thornell, A thermally activated paraffin- based actuator for gas-flow control in a satellite electrical propulsion system, Sen- sors Actuators A Phys. 105 (2003) 237–246, https://doi.org/10.1016/S0924-4247 (03)00203-6. [26] E.T. Carlen, C.H. Mastrangelo, Electrothermally activated paraffin microactuators, J. Microelectromech. Syst. 11 (2002) 165–174, https://doi.org/10.1109/JMEMS.2002. 1007394. [27] R. Bodén,M. Lehto, U. Simu, G. Thornell, K. Hjort, J.-Å. Schweitz, A polymeric paraffin actuated high-pressure micropump, Sensors Actuators A Phys. 127 (2006) 88–93, https://doi.org/10.1016/j.sna.2005.11.068. [28] S. Svensson, G. Sharma, S. Ogden, K. Hjort, L. Klintberg, High-pressure peristaltic membrane micropump with temperature control, J. Microelectromech. Syst. 19 (2010) 1462–1469, https://doi.org/10.1109/JMEMS.2010.2076784. [29] M. Lehto, Paraffin Actuators in Microfluidic Systems, Acta Universitatis Upsaliensis Acta Universitatis Upsaliensis, Uppsala, 2007 (978-91-554-6942-9). [30] R. Bodén, Microactuators for Powerful Pumps, Acta Universitatis Upsaliensis, Upp- sala, 2008, ISBN 978-91-554-7352-5. [31] J.S. Lee, S. Lucyszyn, Bulk-Mikromachined hydraulic microactuator, in: C.-J. Sun (Ed.), Science and Technology of Hybrid Materials, ICMAT 2005, IUMRS-ICAM 2005: Proceedings of Symposium S Science and Technology of Hybrid Materials (ICMAT 2005), 3–8 July 2005 Singapore: The 3rd International Conference onMate- rials for Advanced Technologies: International Union of Materials Research Socie- ties’ 9th International Conference on Advanced Materials, Trans Tech Publications: Uetikon-Zuerich, United Kingdom, 2010. [32] P. Krulevitch, A.P. Lee, P.B. Ramsey, J.C. Trevino, J. Hamilton, M.A. Northrup, Thin film shape memory alloy microactuators, J. Microelectromech. Syst. 5 (1996) 270–282, https://doi.org/10.1109/84.546407. [33] J.S. Lee, S. Lucyszyn, Thermal analysis for bulk-micromachined electrothermal hy- draulic microactuators using a phase change material, Sensors Actuators A Phys. 135 (2007) 731–739, https://doi.org/10.1016/j.sna.2006.07.033. [34] S. Ogden, R. Bodén, K. Hjort, A Latchable valve for high-pressure microfluidics, J. Microelectromech. Syst. 19 (2010) 396–401, https://doi.org/10.1109/JMEMS.2010. 2041749. [35] B.-y. Chen, Y. Liu, X.-h. Zhang, C.-c. Sun, Study on Influence Factors of Thermal Hys- teresis in Paraffin Actuator, 2008 1145–1148, https://doi.org/10.1109/ICICTA.2008. 272. [36] M. Schimmelpfennig, K. Weber, F. Kalb, K.-H. Feller, T. Butz, M. Matthäi, Volumenausdehnung von Paraffinen aus Steigrohr-Messungen, 2007. [37] A.A. Schaerer, C.J. Busso, A.E. Smith, L.B. Skinner, Properties of pure normal alkanes in the C 17 to C 36 range, J. Am. Chem. Soc. 77 (1955) 2017–2019, https://doi.org/ 10.1021/ja01612a097. [38] J.E. Huber, N.A. Fleck, M.F. Ashby, The selection of mechanical actuators based on performance indices, Proc. R. Soc. Lond. A 453 (1997) 2185–2205, https://doi.org/ 10.1098/rspa.1997.0117. [39] VDI-Wärmeatlas, Mit 320 Tabellen; Gesellschaft Verfahrenstechnik und Chemieingenieurwesen, 11., bearb. und erw. Aufl, Springer Vieweg, Berlin, 2013, ISBN 978-3-642-19981-3. https://doi.org/10.1016/j.cirp.2018.05.004 https://doi.org/10.1016/j.cirp.2019.05.007 https://doi.org/10.1016/j.cirp.2016.06.002 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0020 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0020 https://doi.org/10.1051/mfreview/2016002 https://doi.org/10.1016/j.jmapro.2017.04.011 https://doi.org/10.1016/j.jmapro.2017.04.011 https://doi.org/10.1016/j.cirp.2017.05.008 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0040 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0040 https://doi.org/10.1007/s10404-013-1289-3 https://doi.org/10.3390/act7040081 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0055 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0055 https://doi.org/10.1016/j.proeng.2015.01.326 https://doi.org/10.1021/acsnano.9b01180 https://doi.org/10.1126/science.1226762 https://doi.org/10.1038/ncomms4322 https://doi.org/10.1002/adma.201704407 https://doi.org/10.1088/1361-665X/26/1/015011 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0090 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0090 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0095 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0095 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0095 https://doi.org/10.1299/jsmec.40.736 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0105 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0105 https://doi.org/10.1016/S0924-4247(01)00785-3 https://doi.org/10.1016/S0924-4247(03)00203-6 https://doi.org/10.1016/S0924-4247(03)00203-6 https://doi.org/10.1109/JMEMS.2002.1007394 https://doi.org/10.1109/JMEMS.2002.1007394 https://doi.org/10.1016/j.sna.2005.11.068 https://doi.org/10.1109/JMEMS.2010.2076784 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0135 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0135 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0140 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0140 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0145 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0145 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0145 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0145 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0145 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0145 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0145 https://doi.org/10.1109/84.546407 https://doi.org/10.1016/j.sna.2006.07.033 https://doi.org/10.1109/JMEMS.2010.2041749 https://doi.org/10.1109/JMEMS.2010.2041749 https://doi.org/10.1109/ICICTA.2008.272 https://doi.org/10.1109/ICICTA.2008.272 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0170 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0170 https://doi.org/10.1021/ja01612a097 https://doi.org/10.1021/ja01612a097 https://doi.org/10.1098/rspa.1997.0117 https://doi.org/10.1098/rspa.1997.0117 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0185 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0185 http://refhub.elsevier.com/S0264-1275(20)30114-3/rf0185 The challenge of upscaling paraffin wax actuators 1. Introduction 2. The closed laser-welded phase change material actuator 3. Experimental characterization of the phase change actuator 4. Application of a thermal compensation element 5. Conclusion Data availability Author contribution CRediT authorship contribution statement Declaration of competing interest Acknowledgements References