Industrially relevant Injection Moulding Apparatus for In-Situ Time-Resolving Small-Angle X-ray Scattering Measurements

This work is the first step in the development of a multi-scale digital twin of an injection moulding system. It presents the design and construction of an automated industrial injection moulding apparatus that can be used with a typical multi-user beamline such as the NCD-SWEET small-angle X-ray scattering at the 5 ALBA Synchrotron Light Source. This apparatus is focused on developing an understanding of how the mould temperature and the injection temperature affect both the orientation and morphology of the semi-crystalline polymer used to fabricate the injected parts. The system design follows current industrial practice and enables the collection of time-resolved X-ray scattering data at several points 10 within the mould cavity so we can understand the 4D morphology. In this work we show the effectiveness of the equipment using some results from the injection moulding of isotactic polypropylene and polyhydroxybutyrate.


Introduction
* Corresponding author: geoffrey.mitchell@ipleiria.pt;Tel.: +351-962426925 To analyse specific industrial processes, such as injection moulding, the polymer community has used the post-analysis of test samples including X-ray scattering to develop an understanding of the processes of crystallization and flow in that industrial process.However, over the last four decades, there has been a growing use of synchrotron-based small-angle X-Ray scattering (SAXS) allowing an enhancement of the knowledge and understanding of polymer crystallization processes using in-situ time resolving technologies [1].
Injection moulding is the most common process for the automated fabrication of plastic parts, partly as a consequence of the high throughput and the ability to obtain a wide range of product shapes.There is considerable active attention placed on the in-line process control and property measurements [2].The concept of injection moulding technology is rather simple [3].There is an injection of a volume of molten plastic into a mould cavity, usually metallic, with an empty space that corresponds to the negative of the desired component.To counteract the shrinkage effects, additional plastic material is supplied.As detailed by Speranza et al. (2019), once the cavity is filled, the material cools down in the mould, exposing the polymeric chains to "intensive shear and elongation flow fields" [4].
It is both shear and elongation fields that affect the solidification process and its morphology.Due to the difference in temperature between the mould and the plastic, the latter becomes solid either through reaching a glass transition, in the case of an amorphous polymer or undergoing a crystallization process The latter is the situation with semi-crystalline thermoplastics such as polyethylenes, polypropylenes and several bioplastics such as polyhydroxyalkanoates [5] and polybutylene succinate [6].These transitions lock in the shape of the part so that it is retained even when the part is removed from the mould.When the extraction temperature is reached, the part is then removed, and the final component is obtained [3].This manufacturing process is strongly dependent on the parameters of the injection moulding and the tool itself (the mould).Changes in temperature affect the end conditions of the plastic parts produced.As mentioned by Speranza et al. (2019), the changes in mould temperature highly influence the "molecular orientation and morphology formed in injection moulded samples" [4].
The employment of an ex-situ X-ray scattering approach, to study the complex behaviours of polymer injection moulding molecular orientations, has been a widely used technique to study phenomena that are of industrial interest [7].This comes with some disadvantages attached, since the ex-situ type of study "provides little insight into the complex time and spatially dependant processes that take place during mould filling" as described by Rendon et al (2009) [8].It is in this subject that new approaches can appear, and it opens a way of approaching research to industrial needs [8][9][10][11].
The scientific literature has relatively few reports of such systems relevant to injection moulding [12][13][14] in addition to our own work [15][16][17] and the two studies cited here represent quite different approaches.The impressive work of Yin et al [12] is based around the installation of an industrial scale injection moulding system on the BL10U1 beamline at the Shanghai Synchrotron Radiation Facility in China.They overcame the challenges of the precision positioning of a moulding machine with a weight of ~ 3.5 tons and used diamond windows to seal the mould cavity to provide a high transmission pathway for the X-rays through the mould.This system was used to compare the behaviour of polyethylene and polypropylene during injection moulding.Earlier work at the same facility involved a much smaller scale of equipment and led to the development of a microinjection moulding system, again using a metal mould with diamonds as the material for the X-ray transparent windows [14].The use of diamond windows is a well-established approach and widely used for the design of X-ray cells for high-pressure studies [18].The difference between a high-pressure stage and an injection moulding machine cavity is the scale of the active volume and the need to probe the time-resolution of the morphology of the polymer in the cavity throughout the mould.We have opted to follow a rather different approach which maintains the current mould industry design practice while introducing a level of X-ray transparency to enable useful X-ray scattering data to be obtained.The approach described here is more compatible with the standard multi-user beamlines available at synchrotron facilities in Europe and elsewhere.This approach is focused on maintaining the essentials of an industrial injection moulding system but in a more compact and lightweight system so that it can be mounted on the restricted space available on a multi-user beamline such as the NCD-SWEET beamline [19] at the ALBA Synchrotron Light Source [20] in Barcelona, Spain.
It is clear, that there is still much ground to cover in developing a thorough understanding polymer crystallisation and the development of morphology in a mould cavity, and perhaps more importantly how this can be controlled to produce high value products.There have been various attempts to evaluate crystallization following the imposition of a flow history, but these attempts although valuable in a science context have not been so relevant to industrial processing with its great heating and cooling rates and complex flow patterns.[8][9][10][11].
The aim of this work is to enable the time-evolution of the complex processes of flow and cooling within the mould to be observed separately, rather than superimposed as in the final product by designing, manufacturing, and testing a system that allows time-resolved X-ray scattering to be obtained during the injection process and subsequent cooling and crystallisation stages to build a clear and more comprehensive theory and practice of injection moulding.This paper is focused on the design, construction and testing of such a system.An important step in the design of any experimental system is to identify the mode and manner of data collection and how it relates to the problem in hand.X-ray scattering is a versatile technique and through use of the correct geometry differing types of information can be obtained.Figure 1 shows a schematic of the different regimes of quantitative data that can be obtained [21].Traditionally, laboratory based wide-angle X-ray scattering and small-angle X-ray scattering techniques utilise different instruments.However, beamlines at synchrotron light sources such as ALBA in Barcelona [20] have been constructed in which it is possible to measure both small-angle and wide-angle scattering simultaneously.
However, such possibilities come with drawbacks.Time-resolved data collection usually involves a fixed area detector.The small-angle X-ray scattering is measured without compromise.The scattering vector Q, the difference between the vectors representing the incident and scattered beams lies in the plane of the sample and the full azimuthal range is available between the scattering vector Q and the flow axis due to the small angles of scattering.The geometry for wideangle X-ray scattering is usually greatly compromised as the wide-angle X-ray scattering detector is typically placed above the line of the incident X-ray beam, so that only a restricted part of the azimuthal range is available in the active area of the detector.The Austrian SAXS beamline at Elettra [22] and the NCD-SWEET Beamline at ALBA [19] are unusual in this context as they allow the wide-angle X-ray scattering detector to be place in either the meridional or equatorial positions, the latter being a more obvious choice for polymers assuming the symmetry or flow axis is vertical [21].Nevertheless, even in such a configuration, the scattering vector is still tilted out of the sample plane containing the symmetry axis or flow axis and so azimuthal angles close to α=0°cannot be reached.This restriction becomes greater as the scattering angle increases, and so it varies across the length of the detector.
We can see from Figure 1, that small-angle X-ray scattering gives information on the chain folded lamellar crystals and these are predominant in determining the properties of the solid polymer.In contrast, wide-angle X-ray scattering gives access to the sharp Bragg peaks corresponding to the low order index planes of the crystals and their connection to the orientation of the chain folded lamellar crystals is not so straightforward.In the current context, there is a further consideration to make.Small-angle X-ray scattering, as its name suggests, is confined to relatively small-angles making the design of a sample stage such as the mould to both contain the sample and facilitate the passage of the incident and scattering X-ray paths straightforward.In contrast the provision of the same in a wide-angle context is much more demanding and, in some cases, may be impossible.Consideration of all these factors led us to conclude that for our first attempt at operando time-resolving X-ray scattering measurements during injection moulding we should restrict ourselves to recording the small-angle scattering patterns and the remainder of this manuscript is focused on just that.
The apparatus detailed in this work is designed to enable operando SAXS measurements during the injection moulding process, whilst maintaining relevance to industrial injection moulding practise.Sections 2 and 3 explain the design process and implementation of each identified solution.Some preliminary data is presented in section 7 to demonstrate the capabilities of the prototype to analyse the level of preferred orientation of the chain folded lamellar crystals which develops during the cooling stage for two commercial semi-crystalline polymers, isotactic polypropylene, and polyhydroxybutyrate.

Concept Development
First and foremost, to design a machine it is of utmost importance to understand what it should do, where it is going to be used, and who is going to use it.To comprehend this, it is noteworthy to first define its objectives which in this case lie on the necessity to record, in real time, quantitatively the small-angle X-ray scattering which results when hot fluid plastic is injected at high pressure in to a metallic mould with the option of the mould temperature maintained at a specific temperature.The data need to be recorded on a time scale commensurate with the time scale of the development of the morphology which forms as the fluid polymer cools to a solid sample.An upper bound to that time scale is estimated from previous studies to be ~1s.Quantitative data is essential to enable the process of analysis to evaluate the level of preferred orientation of the chain folded lamellar crystals, their thickness and the level of crystallinity.As described by Rendon et al. [8] the use of in-situ X-ray scattering allows "direct, real-time monitoring of molecular scale evolution in polymer materials undergoing complex processing operation".Through the understanding of the previous statement two points become clear.The first is that the device has to be able to enable the data during the "molecular scale evolution in polymer materials undergoing complex processing operation" [8], and the second is that the experience should be made and implemented in a facility able to use in-situ X-ray scattering technologies.In order to achieve the collection of quantitative data in the very early stages of the injection moulding cycle, where the sample volume will be particularly small, we need a bright X-ray source.
To first answer where to find the technology and facilities the team chose to develop this project in partnership with the NCD-SWEET SAXS-WAXS Beamline [19] at the ALBA Synchrotron Light Source in Barcelona, Spain [20], due to previous positive experience and the knowledge that already exists about the facilities and its reliability.Of particular importance is the fact that the beam intensity in this beamline has the ability to answer the second point identified earlier.Beamline NCD_SWEET uses a vacuum undulator as the source of the Xradiation with one of the intense harmonics being that corresponding to a wavelength of 1Å with 12.4KeV photons.The flux at the sample position is reported to be 1.5•10 12 ph/s at 12.4 keV @ 150 mA [23].Currently ALBA operates at a ring current of ~ 250mA, which will increase the figure by a factor of 1.7.The detector for small-angle X-ray scattering is a DECTRIS Pilatus3S 1 M detector system mounted up to 6.7m from the sample.This enables the SAXS data to be obtained over the |Q| range from 0.0017 Ᾰ -1 to 0.125 Ᾰ -1 where |Q| = 4πsin θ/ λ where 2θ is the scattering angle and λ is the incident wavelength.
With ALBA selected as the facility to host the injection moulding system, the first priority was to understand the constraints for mounting the machine at the beamline, such as geometric, as shown in Figure 2, and weight limitations.The length of the mounted system along the beamline is adjustable, but as in this region the beam passes through the air, a long length will lead to some scattering of the beam by the air, leading to a loss of flux and collimation.The weight of the injection moulding system is one of the major concerns, since a machine for the injection moulding process must be sturdy enough to handle the pressure, the weight of the components and the relative movement of each subassembly.The weight of a commercial machine can be several tons.Although the system is likely to be large and heavy, it needs to be precision mounted on the beamline to ensure the x-ray beam is centered in the high transmission path designed in the mould.The NCD-SWEET beamline provides a high precision positioning table equipped with linear movement in x, y and z together with angular movements about the centre of these stages and a yaw stage.The maximum weight which can be mounted on this assembly is 100Kg and clearly the weight needs to be balanced to minimise any off-axis stresses on the translation stages.With the constraints defined, it was possible to start defining each sub-assembly of the machine.The first task was to select an injection unit since it is a device that must be acquired from a manufacturer, restricting the rest of the apparatus to its connections and dimensions.Hence, after carefully studying the possibilities, the team opted for an industrial autonomous injection unit (Babyplast ® UAI I/25) from the Rambaldi Group [24].This injection system separates the injection unit from the power and control systems as shown in Figure 3.It also allows the operator to remotely control the process through an LCD touch display connected to the floor-mounted control unit.
An industrial injection moulding system contains four main units, the injection unit, the clamping unit, the control unit, and the power unit.With the selection of the autonomous injection unit, the first system and its control were defined.
Nevertheless, it is necessary to understand how the clamping unit, which holds the two parts of the mould closed, will operate for this apparatus to work effectively.

System-level design
Several sub-systems were developed to allow the installation and subsequent analysis of the liquid to solid transformation which accompany the injection moulding cycle.To improve the quality of the system designed (Figure 4), a system-level design approach had to be used.This means that the development had to be divided into three main focus areas: the mould and the part to inject, the clamping and support unit, and the clamping control unit.The other key functions required for these experiments connected with the time-resolved detection of the small-angle X-ray scattering patterns and their subsequent storage and analysis are already provided as part of the existing beamline [19].
Figure 4. Exploded view schematic of the apparatus.

Mould
The mould was designed for a study part developed from a conceptual model with objective of this experiment is to understand polymer crystallisation in injected moulded components, we identified the need for analysis throughout the length of the part, allowing an understanding of the full injection process to be developed in different locations.To reduce any influence of the change in flow direction and to allow for easy measurement during the injection process, the outcome required a flow entrance along the part, while allowing injection from above.Injection from above, allowed the injection unit to be mounted vertically above the mould in a central manner to maintain a balanced distribution of weight to avoid over stressing the translation stage on which the whole system is mounted.The injected part should have its length in the horizontal direction (Figure 5) to allow for the multiple points of analysis shown in Figure 5 with for easy and fast movement on the platform.The chosen solution was a component of 1 to 2mm in thickness and a length x height of 100x20 mm.However, this does not take away the fact that this apparatus is prepared for the analysis of different parts, it only means that it depends on the cavity machined onto the mould.To maintain equivalence with the mould industrial practice, both mould, platens, and frame were made from an aluminium alloy AW6082 [25].This alloy contains manganese to control the grain size to maximise its strength.As we have a relatively simple mould, one could assume that there is not much to add on a mould.Nevertheless, that is not the case, due to the necessity to prepare it for being evaluated through SAXS.We identified the advantages of focusing on the mould inserts to enable the transmission of the X-ray beam.Mould inserts are widely employed in the design and fabrication of moulds to allow complexity to be introduced in the mould cavity without the need to re-machine the complete mould cavity should any changes be required.This means that we should introduce two inserts one for the incident beam and one for the scattered x-rays leaving the mould, without compromise should anything need to be improved.
Hence, the conclusion was to make two inserts, one at the fixed half and another at the moving half, that would have "holes" where the X-rays could pass through.
Of course, these are not holes as that would prevent the development of a pressure in the mould cavity.These inserts, shown in Figure 6 possess directly machined precision windows of 0,08 mm in thickness, which ensures that the current mould practice is maintained whilst providing a level of X-ray transparency.To guarantee the evaluation at different locations, six holes were introduced into the inserts at 11 mm intervals.To allow the exit of the diffracted beam two sections appear in these holes, a first with 1,5 mm closest to the mould cavity to improve mechanical strength and the second of 5 mm to allow the beam to exit without interference from the mould.This approach maintained a continuous smooth surface of the same material in the mould cavity whilst providing an adequate transmission for the X-ray beam.
The CT image in Figure 6b confirmed the expected results and the mean thickness of the twelve windows were measured as 0.0832 mm with a standard deviation of 0.0015 mm.This gives an estimated attenuation < 30% for 12.4keV photons.The nature of the background scattering introduced by the mould insert is considered later.
For cooling the mould, conventional cooling was machined into the mould bringing the mould even closer to industrial practise.As injection moulding is a highly temperature dependant process and since the components were made of aluminium, two ceramic insulation platens were placed between the mould and its supports to reduce loss of heat through the components, Future developments will include conformal heating and cooling.To guarantee that the component could be extracted, pneumatically activated extractors were placed onto the mould.
At the end of the mould design process the system to analyse the part was set as, the X-ray would pass through the insert windows of the fixed half scattering along the part and exiting the moving part, with the mould closed, providing data that would allow for understanding the crystallisation of the polymer as shown in the rendered model (Figure 7.).
Figure 7.The mould and components developed for small-angle X-ray scattering and the initial expectation of the mechanism for analysis.
With both part and mould readily defined the next design step was to ensure the closing and opening of the mould through the clamping unit actions.

Clamping unit
The clamping unit allows for the movement of the mould, without this a mould This component is the structural assembly as it is to this that both clamping force, relative movement, and support of components must be connected to.Thus, the team subdivided the development of the clamping unit in three steps, the first was the support to guarantee that the dimensional constraints were met, the second was the movement, to allow opening and closing of the mould, and the third the locking device, to counterforce the injection pressure.To connect the device to the platform, a platen with screws corresponding to the ones hole pattern on the mounting platform present at ALBA synchrotron [23] were screwed to the support platens.This platen enabled the motion of the whole system to ensure the precision alignment with the beam through the motion systems present at the beamline.
To support the other components, the support system is made of two individual sub-assemblies, the first supports the injection unit above the mould, and the other supports both mould and movement accessories.For the injection unit support, a rig made of aluminium profiles connected by screws was machined with connections to the injection unit rods and to the support platens.In the case of the clamping unit itself, the apparatus consists of two support platens connected by four tie rods.
Design of the movement system was focused on simplifying the system.To that end, a ball screw mechanism mounted in the sides of the support platens fed by two stepper motors was the chosen solution, due to its precision while avoiding the mechanism to become asynchronous.With this solution, it was possible to reduce weight and greatly simplify the system.To allow the opening and closing of the mould the two ball screws were connected to the moving support platen through ball nuts making them move simultaneously.
To allow locking of the whole mechanism for the injection process, two motor brakes were connected to the ball screw's shaft (Figure 8b).This allowed the avoidance of the mould opening while injection was occurring.

Clamping unit automation
The system requires an automated control unit.This is necessary not only to allow for multiple injections in a small amount of time, but also since the whole system must be controlled remotely.On the NCD-SWEET beamline, the apparatus is mounted within a radiation proof enclosure and researchers are not allowed inside while the X-ray beam is present due to radiation safety.
Several components had to be incorporated to allow automation of the clamping unit.A Control Board as shown in Figure 8a, was developed and connected to the electrical motors.To control the movement of the motors, an USB-A port was placed onto the board and this was then connected to an external terminal which was then used to control the open/closing of the mould.

Finite Element Method Simulations
Since the injection moulding process involves both pressure and temperature, the variations in these two will induce deformation through pressure, expansion or contraction through temperature.Ansys ® software was used to understand the degree of changes produced in the inserts.The properties presented for Aluminium Aw 6082 T6 in Ansys for both Tensile Ultimate Strength, Yield Strength and Thermal conductivity are presented in Table 1.
To setup the simulation a fixed support was applied around the insert and to the surface in contact with the mould, the insert was set at a temperature of 50˚C.The pressure profile used for simulating the pressure flow inside the cavity was set as 20 MPa from the imported first rheological trials, these have shown a decrease in pressure along the length of the insert as expected.This effect is shown in the windows which are the critical zones due to the lower thickness, as shown in The results of the FEM (Finite Element Method) simulations showed that the effect of pressure will not affect the inserts as much as the changes in temperature.
The deformations presented in these show that there will exist a slight expansion on the inserts mainly in the area of the windows.When the hot plastic enters the mould cavity, the temperature of the mould will increase, especially in the area of the windows as the metal section is thinner.As a result there will be a differential expansion in the region of the thinned areas, which then protrude in to the mould cavity, leading to small depressions in the moulded part.There is no permanent change in the mould as the temperature rapidly equilibrates through the mould.
Figure 10a shows a false colour representation of these temporary distortions in the surface of the mould insert, the maximum deformation is 0.015mm.Figure 10b shows a photograph of the moulded part which shows a series of depressions, each ~0.63mm is diameter, along the centre line.The dimensions of the depression observed are similar to those predicted from the simulations.These temporary changes in the surface of the mould insert are sufficiently small that they will not affect the flow of the polymer in any significant manner, but they do provide a useful marker of the positions of the windows in the part which is most useful in any post-moulding characterisation.To understand if the thickness of the thinned areas can compromise the inserts or not, the Von-Mises stresses were evaluated and compared with the values in Table 1, which should be higher than the results in Figure 11.It is possible to conclude that, by observing the figure above that the design of the mould inserts is well within the specification for the material.

Rheological simulations
To understand some of the parameters to use in this apparatus and obtain the injection moulding processing conditions and other characteristics, namely of the moulding, during the injection cycle, some rheological simulations were performed in Moldex3D software (.Version 4.1).These studies had the objective of analysing some of the outputs such as, the clamping force value and its centroid, the variation of pressure and temperature in the locations of the respective sensors.
For the first trials an injection pressure of 20 MPa was applied.From the results of this simulation it was clear that the centroid of the clamping force was not centred with the mould, however, the offset was not significant.In this simulation only partial filling of the mould cavity was achieved.Experiments performed with the injection moulding system developed in the work showed similar levels of filling ~ 61%.
After understanding that both simulations and injected specimens have similar results, Moldex3D was used to evaluate the necessary parameters to guarantee filling of the part and respective values for sensor pressure and temperature.For ton/m 2 is required in order to guarantee a fully injected part.The complete pressure curve is shown in Figure 12.
Figure 12.Pressure results for whole injection process.

Experimental approaches
The experiments to be performed to follow the development of structure and morphology during and subsequent to the injection moulding process can be performed in a variety of ways.We can opt to focus on the behaviour of the polymer at a particular point in the mould using one of the predetermined window positions.In addition since the basis of injection moulding is that it is highly repeatable [17], we can combine several of these time sequences obtained under equivalent conditions but related to different positions, which allows us to collect 4D data.

Results and discussion
The complete injection moulding system weighs 90.86 kg, which means it is under the maximum weight allowed, allowing to successfully mount it on the NCD-SWEET Beamline as shown in Figure 13.There are some critical aspects for the moulding system present in this article.When using any experimental stage with windows, there will some scattering recorded at the detector which arises from the windows.The injection moulding system could be effectively aligned on the NCD_SWEET beamline using the available translational and rotational stages.Figure 14a shows the 2-D scattering pattern recorded with the injection mould closed but empty on the NCD-SWEET.We can observe an isotropic scattering pattern which decays with scattering angle quite rapidly.
Figure 14b shows the radial average of this scattering and we can see that it is devoid of any sharp features.We can associate this scattering with that arises from the aluminium alloy with precipitates within it.Rowolt et al have studied the microstructure of this specific alloy and their work shows that the precipitates are stable up to a temperature of ~300°C and clearly we do not expect the mould to reach such temperatures.As a consequence we can record this scattering at the start of each experiment and then subtract it from subsequent images so as to display the scattering from the polymer sample only.For higher temperatures we will need to record background measurements on a more regular basis.The corresponding time in each column is the same for each line.The first pattern on the left corresponds to an empty cavity.It shows the same well-defined isotropic pattern typical for an aluminium alloy which forms the window material as in Figure 14a.This scattering is easily subtracted from the subsequent experimental patterns when the mould contains the polymer.There are no specific features relating to the specular reflection of the X-ray beam within the mould insert and reflects the high quality of alignment which was obtained when the system was aligned on the beamline.
As we shift to the right, the SAXS pattern recorded 1 s after injection of the polypropylene at 190ºC.The mould temperature was held constant throughout these 50ºC.The SAXS patterns in Figure 16a exhibit a high level of anisotropy.The pattern is typical for a semi-crystalline polymer in which crystallisation has been nucleated by row nuclei and the chain folded lamellar crystals grow out normal to the row nuclei [21].The intensification of the peak onto the horizontal axis indicates that the lamella growth direction is vertical and so the row nuclei are organised parallel to the long axis of the mould and parallel to the fill direction [21].The intense peak can be used to evaluate the long period and the level of preferred orientation of the lamellar crystals.
We use the azimuthal variation of intensity as a function of α and a fixed value of |Q| corresponding to the peak maximum as shown in Figure 17. Figure 17a shows plots of I(α) at a fixed value of |Q|.The symmetry of the plot arises from the fact that every Xray Diffraction pattern for a low absorbance sample shows an inversion centre.In each of these plots we see intense peaks at α=180 and 0/360° which arise from the lamellar crystal stacks and much weaker peaks at 90 and 270°.We attribute the latter peaks to nanoscale fibrillar structures organised by the clarifier [27].The variation in intensity with α, can be used to quantitatively evaluate the level of preferred orientation of the lamellar crystals or the fibrils of the clarifier using rigorous mathematical procedures described in [28].The increase in intensity of the peak at α=180° arises in part from an increase in the level of preferred orientation and in part from the increase in the volume fraction of crystals.
Clearly, this data contains much information and in future work, we will exploit the availability of these data for the first time to development an understanding of the crystallisation processes which take place in the hot plastic injected in to the metal mould in order to understand how these are influenced by the injection temperature and pressure and by the temperature of the mould.In particular we will seek to identify the structure which develops in each 1s time slice of the time-resolved X-Ray patterns .
Typically injection moulding systems have little quantitative instrumentation to provide feedback on the process.In this work we are able to extract and record critical information on the development of the molecular organisation and morphology during the transformation of the molten plastics to the solid state throughout the injection moulding cycle.In the example discussed in this paper we only show the first few seconds.Digital twin technology has been developed for diverse industrial processes.A digital twin is a virtual equivalent of a real syste,.|A major difference between a digital twin and a standard simulation is the two way communications between the pair and the two transfer of data.The virtual systems is informed and driven by data from the real system.The availability of the structural information, as shown in this work, moves us closer to the realisation of a multi-scale digital twin of injection moulding.This will provide a powerful approach to the optimisation of mould design, the process parameters for new materials such as bioplastics .

Conclusions
This project has focused on designing, manufacturing, and testing an injection moulding system, which conforms to the current industrial mould making practices, to operate in the ALBA NCD-SWEET beamline for crystallisation analysis.The mould inserts windows lead to an attenuation of the primary X-ray beam, with 12.4keV photons of less than 30%.
The system used shows great promise and that is demonstrated by the data obtained following the injected samples of isotactic polypropylene, which shows progressive orientation of chain folded lamellar crystals throughout the process.Through these experiments it is possible to separate the different stages of crystallisation in the mould and their variation with injection temperature.In future work, the design of the mould cavity and the sampling points allow the morphology to be measured along the filling path, this data will also allow for the creation of a 3D model of the development of morphology under industrial conditions with industrial relevant polymers.

Figure 1 .
Figure 1.Scales of structure accessible in X-ray scattering experiments.Reproduced from [21] with permission.

Figure 2 .
Figure 2. Constraints at NCD-SWEET SAXS-WAXS Beamline for (a) beamline working volume dimensions and (b) working height dimensions (all dimensions in mm).
3 iterations, considering a part of circular geometry or a part of rectangular geometry in horizontal and vertical position.The main propose of the study part is to evaluate the effect of the temperature, namely mould and injection temperatures, during the injection moulding process in both the molecular orientation and morphology globally, of a semi-crystalline polymer.As the 150 mm

Figure 5 .
Figure 5. Selected design of the part to study and respective feeding and extraction.

Figure 6 (
Figure 6 (a) A photograph of the two mould inserts showing the mounting screws and the precision machined "holes".(b) An X-ray CT of one of the mould inserts used in this work.The insert shows an enlarged area demonstrating the thin "window" area at the base of the blind holes.
cannot work.It needs to open to allow the ejection of the moulded part and to be clamped shut prior to the injection stage.It is with this part that the team saw possible complications due to the constraints on weight.The constraint on weight was already compromised by the weight of the injection unit which is around 20 kg.

Figure 8 .
Figure 8.(a) schematic of the control board for automation, (b) photograph of mechanical limit switch for end of movement detection and mechanical brakes.

Figure 9 .
Figure 9. Pressure profile applied in the "windows" in the insert, decreasing from right to left.

Figure 10 (
Figure 10 (a, upper) Deformation results from the FEM simulation.X-direction, (b, lower) Photograph of a moulded part (length70mm x 20mm) showing the very shallow depressions along the centre line of the part as described in the text 420
an injection temperature of 240ºC and 50ºC of mould temperature it is expected to have 148ºC at the sensor point and 40.2 MPa of surface pressure after the injection phase.With these results we identified that a clamping force of19.6

Figure 13 .
Figure 13.The injection moulding system developed in this work mounted on the NCD-SWEET Beamline.The beam enters from the right and the vacuum path for the scattered x-rays can be seen on the left emerging from behind the support

Figure 14 (
Figure 14 (a) SAXS pattern of the empty mould cavity (b) the radial average of the data shown in Figure 14a.

Figure 15 (
Figure 15 (a) (Left) A photograph of a PHB moulded part, still attached to the cold runner, produced in the injection moulding system shown in Figure 14.6(b) (Right) A SAXS pattern obtained from the PHB part shown in Figure 15a

Figure 16 .Figure 16
Figure 16.shows a series of SAXS patterns recorded at the first sampling point in the mould cavity throughout the injection moulding cycle for a sample of REPSOL ISPLEN PR595 C2M, a random copolymer of propylene and ethylene with the ethylene component < 0.04 at the injection temperature of 190ºC and a mould temperature of 50ºC.Time varies from left to right.The horizontal direction on the page correspond to the flow direction within the mould cavity.This Figure contains three lines of plots.The first line shows the 2D SAXS patterns, the second line contains a 1D plots of the radial average of the data contained in the SAXS patterns as a function of the modulus of the scattering vector Q and the third line contains 1D plots of the azimuthal variation of the intensity as a function of alpha at a fixed value of |Q|.

Figure 17 (
Figure 17 (a left) Plots of the azimuthal variation in intensity I(α) at a fixed value of |Q| corresponding to peak in the lamellar crystal stack peak as a function of time t.(b, right) plots of I(α) versus time t for (a) α=180°diamond, (b) α=270° triangles