- Email: [email protected]
Combining cure depth and cure degree, a new way to fully characterize novel photopolymers
Accepted Manuscript Title: Combining cure depth and cure degree, a new way to fully characterize novel photopolymers Authors: C. Hofstetter, S. Orman, S. Baudis, J. Stampfl PII: DOI: Reference:
S2214-8604(18)30563-3 https://doi.org/10.1016/j.addma.2018.09.025 ADDMA 514
To appear in: Received date: Revised date: Accepted date:
1-8-2018 19-9-2018 20-9-2018
Please cite this article as: Hofstetter C, Orman S, Baudis S, Stampfl J, Combining cure depth and cure degree, a new way to fully characterize novel photopolymers, Additive Manufacturing (2018), https://doi.org/10.1016/j.addma.2018.09.025 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Combining cure depth and cure degree, a new way to fully characterize novel photopolymers C. Hofstetter,1 S. Orman,2 S. Baudis, 2 J. Stampfl 1,*
Institute of Materials Science and Technology, TU Wien, Getreidemarkt 9/308, 1060 Vienna, Austria
2
Institute for Applied Synthetic Chemistry, TU Wien, Getreidemarkt 9/163, 1060 Vienna, Austria
SC R
IP T
1
*Corresponding author. E-mail address: [email protected]
Abstract
U
Bottom-up stereolithography has become a common lithography-based additive manufacturing
N
technology (L-AMT) to fabricate parts with high feature resolution for biomedical applications. Novel vinyl ester based photopolymers, with their good biocompatibility and biodegradation behavior,
A
showed a promising capacity as bone replacement materials. Due to further tuning of the mechanical
M
properties, those biophotopolymers exhibit reduced curing speed in comparison to highly crosslinked resins e.g. acrylates. The slow structuring of the polymer network results in difficulties at the printing
ED
process. The Jacobs working curve characterizes the cure- and penetration depth of resins, but gives no information about the mechanical properties of the cured layer. The information of cure depth and the mechanical properties of the cured layer (cure degree) is desired. In this work, we simulated the
PT
conditions at L-AMT during the structuring process with a real-time near-infrared photorheometer to evaluate the cure degree of a cured layer at constant cure depth. Therefore, we investigated the curing behavior of mixtures with variable amount of photoinitiator (PI) and light absorber (LA) of vinyl ester
CC E
based biophotopolymers. We found, that a high amount of LA is crucial for good mechanical properties at constant cure depth. Moreover, we present a technique how to optimize a resin
A
formulation regarding the content of PI and LA.
Keywords: vat photopolymerization, photorheology, biophotopolymer, Jacobs working curve, mechanical properties of cured layer
1
Introduction
Digital light processing (DLP) is a lithography-based additive manufacturing technology (L-AMT) suitable to cure layers of photopolymer resin to produce three dimensional parts [1], [2]. The typical curing behavior of common resins is characterized by the fast building up of a polymer network (highreactivity), quick transition from liquid to solid (fast gelation) and a fast setting of the final mechanical properties (fast curing) [3]. This results in stiff and highly crosslinked polymers. Novel, vinyl ester based monomers are promising biodegradable photopolymers for biomedical
IP T
applications [4]. They are characterized by good biocompatibility and mechanical properties, but lack reactivity and show slow curing speed in comparison to common fast curing monomers e.g. acrylates (see Figure 1). For their use as bone replacement materials, the mechanical properties have to be
SC R
further tuned towards higher toughness and elongation at break [5]. This can be achieved, for example, by using monofunctional monomers, which reduce the polymer-network density. Unfortunately, this also reduces the speed of curing at the same time. This results in photopolymers which are difficult to process in L-AMT, since decreased crosslink- density leads to very soft structures, which might
U
dissolve in the surrounding resin or cannot withstand the mechanical forces acting during the building
N
process. The final mechanical properties of such polymers are reached in the following post-curing
PT
ED
M
A
procedure, not during the printing process [6].
Figure 1: Principle sketch of the curing behavior of fast curing systems (e.g. acrylates, dashed line) and slow curing systems (e.g. vinyl ester, solid line).
CC E
The bottom-up vat-polymerization technology is a layer based additive manufacturing process (see Figure 2). Therefore, the actual cure depth has to be larger than the layer thickness. This is necessary to ensure sufficient adhesion between the layers (interlaminar bonding). Due to the inferior material
A
properties of the only partially cured vinyl ester biophotopolymers during the printing process, the interlaminar bonding between the layers is often insufficient and the printed parts break off during the part-vat separation process [7]. A possibility to avoid delamination is to increase the exposure and consequently the cure depth [8]. Therefore, the mechanical properties of the solidified layer, and the interlaminar bonding between the layers would increase too. However, that would impair the fabrication accuracy for reasons of over-curing, particularly in the building direction [9], [10]. Consequently, increasing the cure depth is not an option, especially when printing complex scaffold structures or patient specific implants for biomedical applications.
IP T
Figure 2: Bottom-up vat-polymerization. Reprinted with permission from [11].
SC R
Creating the Jacobs working curve [12] has become the basic procedure for testing and characterizing of new photocurable resins. When analyzing photopolymers, the Jacobs working curve is essential,
because it gives information about the penetration depth (Dp), cure depth (Cd) and the cure speed of the resin and its influencing parameters [13]–[18]. Cure depth tests and the resulting working curve are
U
necessary to characterize photopolymers. The working curve gives the relationship between exposure
N
(Emax) in [mJ cm-2] and responding cure depth (Cd) in [µm]. Creating a logarithmic-linear plot with Cd in y-axis and Emax in x-axis results in a straight line, called the Jacobs working curve (see Figure 3).
A
The intersection of the Jacobs working curve with the x-axis is the critical exposure (Ec) in [mJ cm-2].
M
Ec is related to the gel point of the resin, the starting point for the transition from liquid to solid and the slope of the Jacobs working curve represents the penetration depth (Dp) in [µm] (see Eq. (1)). Thereby,
𝐶𝑑 = 𝐷𝑝 ∗ ln (
𝐸𝑚𝑎𝑥 𝐸𝑐
)
𝐸(𝑧) = 𝐸𝑚𝑎𝑥 ∗ exp (−
A
CC E
PT
(Eq. (2)) [18], [19].
ED
the light-absorption of the resin (dependent of depth (z) in [µm]) follows the Beer-Lambert equation
Figure 3: Example of Jacobs working curve.
Eq. (1)
𝑧 𝐷𝑝
)
Eq. (2)
However, the Jacobs working curve does not give any information about mechanical properties of the cured photopolymer layer. It does not differentiate between soft solids (low crosslink density) and rigid solids (high crosslink density). This may not be a problem for common, fast curing and highlycrosslinked materials, but when using a photopolymer with lower curing speed and reduced cross-link density, the polymer network and the resulting mechanical properties need time to build up. The information from the Jacobs working curve as depicted in Figure 3 is not sufficient for indicating the printability of such resins, since the Jacobs working curve gives no information about the kinetics of
IP T
the polymerization process. A possibility to characterize and further optimize the curing behavior in order to increase mechanical properties at constant Cd, is to vary the amount of light absorber (LA) and photoinitiator (PI)
SC R
[20], [21]. A common approach is to determine the Jacobs working curve followed by printing and testing specimens. However, the printing and further testing of a variety of specimens is time and material consuming. Moreover, this characterization is complex and the results often have large
U
variance.
In this work, we present an improved technique, where a real time-near infrared (RT-NIR)
N
photorheometer is utilized to determine the time-resolved curing behavior of a photopolymer during
A
exposure of a single layer [22]. Therefore, we investigated the influences of LA and PI on the mechanical properties and double bond conversion (DBC) during curing. Moreover, it is possible to
M
create an extension to the common Jacobs working curve, a dataset with information on mechanical properties and DBC during curing of a single layer. Hence, a way to find the optimal mixture of LA
Materials & Methods
2.1
Materials
PT
2
ED
and PI to enhance the printability of resins with a low curing speed at constant Cd is presented.
CC E
The basic components of the photopolymer were the monomers divinyl adipate (DVA, Sigma-Aldrich®) and vinyl decanoate (10MV, Sigma-Aldrich®), the chain transfer agent trimethylolpropane tris(3-mercaptopropionate) (Thiocure® TMPMP, Bruno Bock) (see Figure 4) with radical inhibitor pyrogallol (PYR, Sigma-Aldrich®) to ensure sufficient storage stability of the
A
formulations [23] . The variable components were the photoinitiator Ivocerin® (PI, Ivoclar Vivadent®) and light absorber Quinoline Yellow SS (LA, Sigma Aldrich®). Altogether, nine different batches, all based on the mixture of 60 wt% DVA, 30 wt% 10MV, 9.98 wt% TMPMP and 0.02 wt% PYR were prepared by magnetic stirring at 50 °C for 3 h. The additional amount of PI and LA into the basic mixture can be found in Table 1. The final mixtures were stirred again at 50 °C for 3 h.
10MV
TMPMP DVA
Figure 4: Monomers 10MV (monofunctional) and DVA (difunctional), and chain transfer agent TMPMP.
0.50
1.0
0.04
#1
#2
#3
0.08
#4
#5
#6
0.12
#7
#8
#9
L-AMT system
U
2.2
0.25
SC R
LA [wt%]
PI [wt%]
IP T
Table 1: Additional amount of PI and LA in basic mixture.
N
For this study, a digital light processing (DLP) L-AMT system with upside-down set-up and an
A
InVision® WUXGA 1080p light engine with 460nm LED was used. It was used with an irradiation
M
intensity of 19 mW cm-2, measured at the surface of the vat [24]. Cure depth measurements: Creating the Jacobs working curves for the different mixtures, a vat was filled 3 mm high with photopolymer formulation (vat filling height >> cure depth, to reduce oxygen
ED
inhibition). Square geometries (5x5 mm²) with a hole of 1 mm diameter in the center were cured. All squares were cured with the same intensity (19 mW cm-2), while the exposure times were varied. After
PT
the curing, non-cured material was removed from the vat and the surface of the cured squares was cleaned with solvent (ethyl lactate) and compressed air. Afterwards, the thickness of cured material was measured with a thickness gauge (Kroeplin C110T, Kroeplin with a contact size of Ø 5 mm). The
CC E
measurements were performed in triplicates. Based on the Jacobs working curves the exposure for Cd = 200 µm (ECd - 200 µm) was calculated for the different formulations.
A
2.3
RT-NIR-Photorheometer
A unique combination of an Anton Paar MCR302 WESP rheometer with a light curing system accessory, combined with an OmniCure® 460 nm LED and additionally, a coupled in Bruker Vertex 80 Fourier-transform infrared spectroscopy (FTIR) was used. This setup is capable of measuring the curing behavior of photopolymers, mechanical properties as well as double bond conversion (DBC), simultaneously (Figure 5) [22].
Due to the possibility to adjust the wavelength of the used light-source, intensity and the gap size between glass plate and the measuring system, it was possible to simulate the conditions during the structuring process in the L-AMT system and to obtain a time-resolved analysis of the polymerization
U
SC R
IP T
process.
N
exposure
2.3.1
Measurement procedure
M
A
Figure 5: Schematic illustration of RT-NIR-photorheology setup. Reprinted with permission from [22].
ED
The photorheometer was equipped with a special black-coated parallel aluminum plate measuring system (Ø 25 mm). The black coating was used, to reduce light reflection of the aluminum plate, which could cause multiple-exposure of the sample. Despite the coated measuring system the NIR-
PT
signal for the double bond detection was still sufficient. 75 µl of photopolymer were pipetted onto the center of the glass plate. Additionally, the aluminum plate was “coated” with a thin layer of fluid
CC E
photopolymer, to avoid bubbles in the liquid volume while approaching the glass plate. Before reaching the final gap height, the aluminum plate stopped and excessive material was removed. Afterward, the machine proceeded to measurement position. In order to simulate the curing process in an L-AMT system, a gap height of 100 µm was chosen (which is the common layer thickness in an L-
A
AMT system).
The material was measured under periodic oscillation with a frequency 1 Hz and a shear strain of 1 % [25]. Previously performed amplitude tests [26] confirmed linear elastic behavior using those test parameters. The test procedure was divided into two phases: The first 60 s without exposure, to equilibrate the sample and to analyze the uncured resin with near-infrared spectroscopy (NIR). In the second phase, the photopolymer was cured with a light intensity of 19 mW cm-2 for 200, 300 or 500 s
(which conforms 3800, 5700 and 9500 mJ cm-²), dependent on the amount of PI and LA in the formulation. The measurements were performed in triplicate. 3
Analyzing mechanical properties and double bond conversion
Storage modulus G’, loss modulus G’’ and the normal force resulting from polymerization shrinkage were measured with the rheological set up and the decreasing carbon double bond signal (characteristic wavenumber at 6256-6127 cm-1) to get the DBC with the NIR-set up of the
4
IP T
photorheometer during exposure [22]. Results & Discussion
The Jacobs working curve gives information about how much exposure is needed to cure a certain
SC R
layer, the critical exposure to material gelation and the penetration depth for tested resin. It is a
valuable tool and suitable to characterize highly crosslinked photopolymers with fast curing behavior. This behavior can be called a digital curing behavior. Effectively, there are only two states: not
U
crosslinked (fluid) and highly crosslinked (fully cured).
Unfortunately, Jacobs working curve is insufficient to characterize photopolymers with decreased cure
N
speed, where the polymer network and the resulting mechanical properties of the cured layer need
A
more time to build up. Without the information of the mechanical properties of the cured layer it is hard to process these resins with high accuracy, which is necessary for printing complex patient
M
specific implants. Therefore, we investigated photopolymer mixtures (DVA, 10MV and TMPMP) with different amount of LA (0.04, 0.08 and 0.12 wt%) and PI (0.25, 0.5 and 1.0 wt%) to analyze the
ED
differences during curing and to find the most suitable photopolymer mixture for L-AMT printing. Furthermore, we present a way to optimize the mixture regarding mechanical properties at constant
4.1
PT
cure depth. Working curves
We created nine working curves of different photopolymer-formulations (see Table 1). The
CC E
coefficients of determination (R²) of the logarithmic regression line was above 0.99 for all measured working curves. Furthermore, the standard deviation of the single measurement points was below 12 µm (see resin with PI 0.5 wt% and 0.04, 0.08 and 0.12 wt% LA as representative example in Figure
A
6 and the summarized data in Table 2). As expected, higher amount of PI leads to a decrease of Ec but has low influence on Dp. Moreover, higher amount of LA leads to a decrease of Dp and an increase of Ec. Hence, the amount of PI has less influence on Dp than the amount of LA. The calculated exposure for the cure depth of 200 µm (ECd - 200µm) lies within a range of 386 mJ cm-2 (for PI 1 wt% - LA 0.04 wt%) to 3426 mJ cm-² (for PI 0.25 wt% - LA 0.12 wt%).
IP T SC R
Figure 6: Working curves of photopolymers with varying amount of light absorber (LA) and 0.5 wt% photoinitiator (PI). Calculated exposure (ECd - 200 µm) for particular resin PI 0.5 wt% - LA 0.08 wt% at intersection with cure depth Cd = 200 µm (red dot).
Photorheology
U
4.2
N
We measured the curing properties of nine different photopolymer formulations (Table 1) via photorheology. We plotted the photorheometer results of resin PI 0.5 wt% - LA 0.08 wt% as
A
representative example, to show the typical G’ and DBC curve during curing and to explain the data
M
and results (see Figure 7).
Rheology- and NIR- plot: As the light exposure starts, the DBC increases (see Figure 7, dashed
ED
curve). First, a short but strong increase was visible due to the high reactivity of the chain transfer agent until 10 % DBC, which transfers into a lower slope of the co-polymerization of 10MV and DVA. After an increase to 58%, the DBC-gradient decreases steadily until the final curing stage is
PT
reached (at 90% DBC).
As expected, the G’ slope is delayed to the DBC curve. An irradiation dose of 577 mJ cm-2 is required
CC E
to observe an increase in G‘. At that exposure the photopolymer filled gap of 100 µm was solidified (ECd - 100µm). The ECd - 100µm was detected at the rise of the normal force (data not shown). From ECd - 100µm on, the G’ increases steadily, which has its reflection point at 1835 mJ cm-2 (E (∂G'/∂E) max, see Figure 10) and ends in a steady decrease of the slope, until no change of G’ is detectable
A
(G’max = 2.1 MPa).
The G’max values of the photopolymer mixtures slightly increase with the amount of PI (G’PI 0.25wt% = 2.05 MPa ± 0.05, G’PI 0.5wt% = 2.15 MPa ± 0.1, G’PI 1.0wt% = 2.3 MPa ± 0.1), but no expressive change could be observed with variable amount of LA. The DBCmax of the mixtures does not follow the trend of G’max for the variable amount of PI. Only the formulation with PI 0.25 % had lower DBC than the other resins (DBC PI 0.25wt% = 88.2 % ± 0.1, DBC PI 0.5wt% = 93 % ± 2.6, DBC PI 1.0wt% = 91.4 % ± 1.2, graphs not shown).
IP T SC R U
4.3
M
A
N
Figure 7: Photorheometer curing graph of double bond conversion (DBC, dashed curve, left axis) and storage modulus (G’, solid curve, right axis) of formulation PI 0.5 wt% - LA 0.08 wt% (summarized results in Table 2). The ECd - 200µm value was imported from the working curve (Figure 6). The values for DBCCd – 200µm and G’Cd – 200µm were calculated at the intersection of G’ and DBC curve with ECd - 200µm, respectively. The G’∂G'/∂E max value can be found at the G’ curve at maximum of the slope (see Figure 10).
Combination of cure depth and cure degree
ED
Combining the results from the Jacobs working curve (L-AMT system) with the mechanical properties and double bond conversion (DBC) (photorheometer) gives a unique insight into the material
PT
properties during curing. We present a method, where the cure depth parameter from the Jacobs working curve are combined with the analyzed mechanical properties and double bond conversion (DBC) during curing from RT-NIR-photorheology. Thereby, mechanical properties (G’) and double
CC E
bond conversion (DBC) of a cured layer (h) cured with defined exposure (ECd) represents the cure degree. With this approach, photopolymers with low cure speed can be fully characterized and further optimized for successful L-AMT printing.
A
To ensure a sufficient adhesion between the new cured layer to the previous one (interlaminar bonding) in an L-AMT process, the actual cure depth has to be larger than the layer thickness. Thereby, the Jacobs working curves give information about the required exposure to cure a layer of 200 µm (ECd - 200µm). The particular G’- and DBC curves (photorheometer) were combined with the exposure ECd - 200µm from the Jacobs working curves. This results in the cure degree (mechanical properties (G’Cd - 200µm) and double bond conversion (DBCCd - 200µm)) of a 100 µm cured layer, cured with exposure ECd - 200µm. The results of G’Cd - 200µm and DBCCd - 200µm can be found in Figure 7 dashed red lines.
4.3.1
Analyzing the influence of the amount LA and on the curing behavior
The results of G’Cd - 200µm and DBCCd – 200µm of the different mixtures are summarized in Figure 8 and Figure 9, respectively. The lowest values for G’Cd - 200µm were found at the resin with lowest amount of LA (0.04 wt%), followed by the mixture with 0.08 wt% of LA. The highest values for G’Cd - 200µm were found for resins with 0.12 wt% of LA (see Figure 8).
IP T
A similar trend is also visible for the DBCCd - 200µm results (see Figure 9), where lowest values were found for resins with 0.04 wt% of LA, followed by the mixtures with 0.08 wt% of LA. The highest values for DBCCd - 200µm were found for resins with 0.12 wt% of LA.
SC R
For the variable amount of PI, no clear progression is visible (see Figure 8). We assume, when the
ratio between the amount of LA and PI is constant, the influence of the amount of PI to G’Cd - 200µm is minor. However, as this ratio was changed to higher amount of PI (1 wt% of PI), the G’Cd - 200µm increased stronger with increasing amount of LA. Additionally, the G’max values of fully cured
U
formulations could also have influence of the partially cured state at G’Cd - 200µm (resin with 1 wt% of PI
N
has up to 20 % higher G’max as resin with 0.25 wt% of PI). Therefore, it was considered, that the increased amount of PI has influence on the G’Cd - 200µm, especially the closer the actual exposure gets
A
to G’max.
M
However, more important is the clear trend of G’ and DBC regarding the increasing amount of LA in different mixtures (see Figure 8 and Figure 9). Resins with high amount of LA have substantially
ED
higher G’Cd - 200µm and DBCCd - 200µm, than resins with lower amount of LA, independent of the amount of PI. Consequently, the higher the amount of LA in the photopolymer formulations, the higher the
A
CC E
PT
G’Cd - 200µm and DBCCd - 200µm values at constant layer thickness (see Table 2).
Figure 8: G’Cd - 200µm values from the combination of cure depth (working curve) and mechanical properties (photorheometer) of nine different mixtures. The highest G’Cd - 200µm values were found for mixtures with 0.12 wt% light absorber (LA), independent of the amount of photoinitiator (PI).
IP T
.
Cure speed (∂G'/∂E)
U
4.3.2
SC R
Figure 9: DBCCd – 200µm values from the combination of cure depth (working curve) and double bond conversion (DBC) (photorheometer) of nine different mixtures. The highest DBCCd – 200µm values were found for mixtures with 0.12 wt% light absorber (LA), independent of the amount of photoinitiator (PI).
Results of resin PI 0.5 wt% - LA 0.08 wt% as representative example are presented in Figure 10, to
N
explain the data of the typical ∂G'/∂E- curve. The summarized results for all mixtures can be found in
A
Table 2.
M
By formation of the first derivative of the storage modulus (from photorheometer data) with respect to the exposure (∂G'/∂E, interpreted as cure speed), it is possible to determine the exposure, where the G´
ED
has its maximum slope (exposure at highest cure speed E(∂G'/∂E) max) (see Figure 7) and at that point, the ratio of increasing G’ to required exposure reaches its highest value. Therefore, we define the E(∂G'/∂E) max as the optimal exposure of tested resin regarding G’ (see Figure 10). Certainly, it is possible
PT
to cure with higher exposure, but excessive more exposure is required to gain relatively little more G’. Curing with less than E(∂G'/∂E) max is less time consuming, but the mechanical properties of the cured
A
CC E
layer and the interlaminar bonding between the layers is weaker.
IP T SC R
Figure 10: Cure speed (∂G'/∂E) of formulation PI 0.5 wt% - LA 0.08 wt%. The red dot marks the maximum of ∂G'/∂E, at exposure E (∂G'/∂E) max. The ECd - 200µm value was taken from the Jacobs working curve (Figure 7).
Exposure index (Ei)
U
4.3.3
N
By dividing the actual exposure ECd - 200µm by the defined optimal exposure E (∂G'/∂E) max, it is possible to calculate the exposure index Ei and evaluate the actual exposure (see Eq. (3)). The aim is, to get the
𝐸(𝐶𝑑 ) 𝐸(𝜕𝐺`/𝜕𝐸 )𝑚𝑎𝑥 (ℎ)
= 𝐸𝑖
Eq. (3)
ED
M
A
actual exposure close to E (∂G'/∂E) max, which would result in Ei = 1.
PT
As expected, the higher the amount of PI and the lower the amount of LA, the lower the exposure to reach the maximum cure speed (E∂G'/∂E max) (Table 2). The same trend takes effect to reach G’Cd - 200µm. However, as we can see in values of calculated Ei in Table 2, the amount of PI is relevant, but the
CC E
important variable to increase the G’ at constant Cd of 200 µm is the amount of LA. Thus, the higher the amount of LA, the closer Ecd – 200µm comes to E(∂G'/∂E) max and therefore Ei gets closer to
A
optimum value 1.
Table 2: Summarized data (critical exposure (Ec), penetration depth (Dp), exposure to cure a 200 µm layer (ECd -200µm) of Jacobs working curves, analyzed data of particular E (∂G'/∂E) max from photorheology and calculated exposure index (Ei) by Eq. (3).
Ec Dp -2 [mJ cm ] [µm]
ECd - 200µm E (∂G'/∂E) max Ei [mJ cm-2] [mJ cm-2] (opt. = 1)
LA 0.08 wt%
LA 0.12 wt%
4.3.4
333 213 147
219 210 207
830 553 386
2054 1326 984
0.40 0.42 0.39
PI 0.25 wt% PI 0.50 wt%
456 286
131 146
2108 1119
2932 1841
0.70 0.61
PI 1.0 wt% PI 0.25 wt% PI 0.50 wt% PI 1.0 wt%
178 638 380 209
130 119 118 97
825 3426 2078 1656
1309 3844 2556 1775
0.63 0.89 0.81 0.93
IP T
LA 0.04 wt%
PI 0.25 wt% PI 0.50 wt% PI 1.0 wt%
Calculating the optimal amount of LA for given PI
SC R
Using the particular data of ECd - 200µm and E (∂G'/∂E) max of the different mixtures, it is possible to
calculate the optimal amount of LA (LAopt.) and the therefore needed exposure (ELA opt.) to reach Ei = 1, at constant Cd (diagram with 0.5 wt% PI is shown in Figure 11). The LAopt. derives from the intersection of the extrapolated trend lines (logarithmic for 1 and 0.5 wt% of PI, linear for 0.25 wt% of
N
determination (R²) of all created trend lines was above 0.99.
U
PI) of ECd - 200µm and E(∂G'/∂E) max at 0.04, 0.08 and 0.12 wt% LA, respectively. The coefficients of
A
The results for LAopt. (at ELA opt) are 0.125 wt% (1833 mJ cm-2), 0.143 wt% (3085 mJ cm-2) and
M
0.162 wt% (4769 mJ cm-2) for 1, 0.5 and 0.25 wt% PI, respectively. Unexpected is the linear trend of the values of E(∂G'/∂E) max and ECd - 200µm for increasing amount of LA (0.04, 0.08 and 0.12 wt%) for the formulation with 0.25 wt% PI (diagram is shown in the
ED
supplementary material). We assume, that the amount of PI is below a certain threshold, so that the absorbance of the resin is too low to show typical logarithmic behavior. Further experiments have to
PT
be done to investigate these untypical behavior. However, the values of E(∂G'/∂E) max and ECd - 200µm of resins with 1 and 0.5 wt% of PI follows the
CC E
expected logarithmic trend for increasing amount of LA (0.04, 0.08 and 0.12 wt%) (diagram with 0.5 wt% PI is shown in Figure 11). Consequently, it could be shown, that it is possible to calculate the optimum amount of LA by creating the cross section of the trend lines of E(∂G'/∂E) max (photorheology) and ECd - 200µm (Jacobs working
A
curve), if the amount of PI is sufficient.
IP T SC R
5
U
Figure 11: Calculating the optimal amount of LA (LA opt.) and the required exposure (ELA opt.), for resin with PI 0.5 wt%, at the intersection (red dot, Ei = 1) of extrapolated logarithmic trend lines of ECd – 200µm (from working curve) and- E (∂G'/∂E) max (from photorheometer curve) for Cd = 200 µm and layer thickness 100 µm.
Conclusion
N
The aim of this work was to establish the results from real-time near-infrared
A
(RT-NIR)-photorheology as a supplementary to the Jacobs working curve, to fully characterize (cure
M
depth and cure degree) photopolymers with low cross-link-density and cure speed. Moreover, through analyzing the data from Jacobs working curve and RT-NIR-photorheology of vinyl ester based resins with different amount of light absorber (LA) and photoinitiator (PI), the optimal mixture for
ED
lithography-based additive manufacturing (L-AMT) of those particular formulations was found. The following findings summarize the view regarding photocuring of functional photopolymers with
-
PT
reduced cross-link-density and speed of solidification: The RT-NIR-photorheology is a valuable tool to investigate different photopolymer mixtures.
CC E
With the combination of Jacobs working curve (cure depth) and the mechanical properties and double bond conversion (DBC) from the photorheometer (cure degree), it is possible to completely characterize the curing behavior of photopolymers.
A
-
At a constant cure depth and layer thickness the amount of PI in the photopolymer mixture has less influence than the amount LA on the cure degree (mechanical properties and double bond conversion) of the cured layer. Moreover, the higher the amount of LA, the higher the storage modulus (G’) and DBC of the cured layer. Hence, it is easier to print three-dimensional objects.
-
Complete curing during the printing process (up to G’max) is not reasonable for resins with low cure speed. For curing a layer in the printing process, we defined the optimal exposure at the highest cure speed E (∂G'/∂E) max. The final mechanical properties of the printed part should be reached in a separate post-curing step.
-
The Exposure index (Ei) can be used to evaluate a photopolymer, or to compare different mixtures of photopolymers at different exposure situations. The closer the formulation to Ei = 1, the better.
-
The optimal amount of LA (and the required exposure E LA opt.) to achieve Ei = 1 (at constant cure depth) can be found at the intersection of the regression lines of optimal exposure E (∂G'/∂E) max and needed exposure E Cd - 200µm, when the amount of PI is sufficient.
Acknowledgements
SC R
fabrication of different test parts (e.g. scaffold structures) using L-AMT.
IP T
The next steps in order to confirm und further investigate these promising findings should be the
The research presented in this work was conducted at the TU Wien, Institute of Materials Science and Technology. The author gratefully thanks for the financial support by the
A
CC E
PT
ED
M
A
N
U
Austrian Science Fund (FWF) under research program No. P27059.
References
[8] [9]
[10]
[11] [12] [13] [14]
IP T
CC E
[15]
SC R
[7]
U
[6]
N
[5]
A
[4]
M
[3]
ED
[2]
David Kazmer, 28 - Three-Dimensional Printing of Plastics - Kutz, Myer, in: Appl. Plast. Eng. Handb. Second Ed. William Andrew, 2017. Samuel Clark Ligon, Robert Liska, Jürgen Stampfl, Matthias Gurr, and Rolf Mülhaupt, “Polymers for 3D Printing and Customized Additive Manufacturing,” Chem. Rev., vol. 117, no. 5, pp. 10212–10290, 2017. Chunguang Li, Jue Cheng, Wenkai Chang, and Jun Nie, “Photopolymerization kinetics and properties of a trifunctional epoxy acrylate,” Des. Monomers Polym., vol. 16, no. 3, pp. 274–282, 2013. Christian Heller, Martin Schwentwenwein, Günter Russmüller, Franz Varga, and Jürgen Stampfl, “Vinyl Esters: Low Cytotoxicity Monomers for the Fabrication of Biocompatible 3D Scaffolds by Lithography Based Additive Manufacturing,” Polym. Sci. Part A, vol. 47, no. 24, pp. 6941–6954, 2009. Samuel Clark Ligon-Auer, Martin Schwentwenwein, Christian Gorsche, Jürgen Stampfl, and Robert Liska, “Toughening of photo-curable polymer networks: a review,” Polym. Chem., vol. 7, no. 2, pp. 257–286, 2015. Bernhard Steyrer, Philipp Neubauer, Robert Liska, and Jürgen Stampfl, “Visible Light Photoinitiator for 3D-Printing of Tough Methacrylate Resins,” materials, vol. 10, no. 12, 2017. Xiangquan Wu, Qin Lian, Dichen Li, and Zhongmin Jin, “Tilting separation analysis of bottomup mask projectionstereolithography based on cohesive zone model,” Mater. Process. Technol., vol. 243, no. 1, pp. 184–196, 2016. Jim H. Lee, Robert K. Prud’homme, and Ilhan A. Aksay, “Cure depth in photopolymerization: Experiments and theory,” Mater. Res. Soc., vol. 16, no. 12, pp. 3536–3544, 2001. Paul O’Neill, Nigel J. Kent, and Dermont Brabazon, “Mitigation and control of the overcuring effect in mask projection micro-stereolithography,” presented at the 20th International ESAFORM Conference on Material Forming, Dublin, 2017. Jae-Won Choi, Ryan B. Wicker, Seok-Hyon Cho, Chang‐Sik Ha, and Seok‐Hee Lee, “Cure depth control for complex 3D microstructure fabrication in dynamic mask projection microstereolithography,” Rapid Prototyp. J., vol. 15, no. 1, pp. 59–70, 2009. Stefan Baudis, “Knochen, Knorpel und Gefäße maßschneidern mit Licht,” Nachrichten Aus Chem., vol. 64, no. 4, pp. 406–410, 2016. Ian Gibson, David W. Rosen, and Brent Stucker, Additive Manufacturing Technologies, Rapid Prototyping to Direct Digital Manufacturing. Springer, 2010. Ewa Andrzejewska, “Photopolymerization kinetics of multifunctional monomers,” Prog. Polym. Sci., vol. 26, no. 4, pp. 605–665, 2001. Bernhard Steyrer, Bernhard Busetti, György Harakály, Robert Liska, and Jürgen Stampfl, “Hot Lithography vs. room temperature DLP 3D-printing of a dimethacrylate,” Addit. Manuf., vol. 21, pp. 209–214, 2018. Hongping Xiang et al., “Preparation, Characterization and Application of UV-Curable Flexible Hyperbranched Polyurethane Acrylate,” Polymers, vol. 9, no. 11, 2017. Li Feng and Byoung I. Suh, “Exposure Reciprocity Law in Photopolymerization of MultiFunctional Acrylates and Methacrylates,” Macromol. Chem. Phys., vol. 208, no. 3, pp. 295–306, 2007. Sara Dalle Vacche, Valérie Geiser, Yves Leterrier, and Jan-Anders E. Månson, “Time-intensity superposition for photoinitiated polymerization of fluorinated and hyperbranched acrylate nanocomposites,” Polymer, vol. 51, no. 2, pp. 334–341, 2010. Paul Francis Jacobs, Rapid Prototyping & Manufacturing: Fundamentals of StereoLithography. California: Society of Manufacturing Engineers; 1st edition, 1992. Andreas Gebhardt and Jan-Steffen Hötter, Additive Manufacturing, 3D Printing for Prototyping and Manufacturing. Munich: Hanser Publications, 2016. Tom Scherzer and Ulrich Scherzer, “Real-time FTIR-ATR spectroscopy to study the kinetics of ultrafast photopolymerization reactions induced by monochromatic UV light,” Vib. Spectrosc., vol. 19, no. 2, pp. 385–398, 1999. Mohamed Mohamed Zabti, “Effects of light absorber on micro stereolithography parts,” University of Birmingham, Birmingham, 2012.
PT
[1]
[16]
A
[17]
[18] [19] [20]
[21]
A
CC E
PT
ED
M
A
N
U
SC R
IP T
[22] Christian Gorsche et al., “Real Time-NIR/MIR-Photorheology: A Versatile Tool for the in Situ Characterization of Photopolymerization Reactions,” Anal. Chem., vol. 89, no. 9, pp. 4958– 4968, 2017. [23] Paricher Esfandiari, Samuel Clark Ligon, Jean Jacques Lagref, Richard Frantz, Zoubair Cherkaoui, and Robert Liska, “Efficient Stabilization of Thiol-ene Formulations in Radical Photopolymerization,” J. Polym. Sci. Part Polym. Chem., vol. 51, pp. 4261–4266, 2013. [24] Robert Gmeiner, Gerald Mitteramskogler, Jürgen Stampfl, and Aldo R. Boccaccini, “Stereolithographic Ceramic Manufacturing of High Strength Bioactive Glass,” Appl. Ceram. Technol., vol. 12, no. 1, pp. 38–45, 2014. [25] Nono Darsono, Hiroshi Mizunuma, and Hiromichi Obara, “Rheological Study of the Solidification of Photopolymer and Dispersed Nanotube Systems,” Appl. Rheol., vol. 21, no. 11, pp. 63566–63581, 2011. [26] Thomas G. Mezger, Applied Rheology - With Joe Flow on Rheology Road. Anton Paar GmbH.
S2214-8604(18)30563-3 https://doi.org/10.1016/j.addma.2018.09.025 ADDMA 514
To appear in: Received date: Revised date: Accepted date:
1-8-2018 19-9-2018 20-9-2018
Please cite this article as: Hofstetter C, Orman S, Baudis S, Stampfl J, Combining cure depth and cure degree, a new way to fully characterize novel photopolymers, Additive Manufacturing (2018), https://doi.org/10.1016/j.addma.2018.09.025 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Combining cure depth and cure degree, a new way to fully characterize novel photopolymers C. Hofstetter,1 S. Orman,2 S. Baudis, 2 J. Stampfl 1,*
Institute of Materials Science and Technology, TU Wien, Getreidemarkt 9/308, 1060 Vienna, Austria
2
Institute for Applied Synthetic Chemistry, TU Wien, Getreidemarkt 9/163, 1060 Vienna, Austria
SC R
IP T
1
*Corresponding author. E-mail address: [email protected]
Abstract
U
Bottom-up stereolithography has become a common lithography-based additive manufacturing
N
technology (L-AMT) to fabricate parts with high feature resolution for biomedical applications. Novel vinyl ester based photopolymers, with their good biocompatibility and biodegradation behavior,
A
showed a promising capacity as bone replacement materials. Due to further tuning of the mechanical
M
properties, those biophotopolymers exhibit reduced curing speed in comparison to highly crosslinked resins e.g. acrylates. The slow structuring of the polymer network results in difficulties at the printing
ED
process. The Jacobs working curve characterizes the cure- and penetration depth of resins, but gives no information about the mechanical properties of the cured layer. The information of cure depth and the mechanical properties of the cured layer (cure degree) is desired. In this work, we simulated the
PT
conditions at L-AMT during the structuring process with a real-time near-infrared photorheometer to evaluate the cure degree of a cured layer at constant cure depth. Therefore, we investigated the curing behavior of mixtures with variable amount of photoinitiator (PI) and light absorber (LA) of vinyl ester
CC E
based biophotopolymers. We found, that a high amount of LA is crucial for good mechanical properties at constant cure depth. Moreover, we present a technique how to optimize a resin
A
formulation regarding the content of PI and LA.
Keywords: vat photopolymerization, photorheology, biophotopolymer, Jacobs working curve, mechanical properties of cured layer
1
Introduction
Digital light processing (DLP) is a lithography-based additive manufacturing technology (L-AMT) suitable to cure layers of photopolymer resin to produce three dimensional parts [1], [2]. The typical curing behavior of common resins is characterized by the fast building up of a polymer network (highreactivity), quick transition from liquid to solid (fast gelation) and a fast setting of the final mechanical properties (fast curing) [3]. This results in stiff and highly crosslinked polymers. Novel, vinyl ester based monomers are promising biodegradable photopolymers for biomedical
IP T
applications [4]. They are characterized by good biocompatibility and mechanical properties, but lack reactivity and show slow curing speed in comparison to common fast curing monomers e.g. acrylates (see Figure 1). For their use as bone replacement materials, the mechanical properties have to be
SC R
further tuned towards higher toughness and elongation at break [5]. This can be achieved, for example, by using monofunctional monomers, which reduce the polymer-network density. Unfortunately, this also reduces the speed of curing at the same time. This results in photopolymers which are difficult to process in L-AMT, since decreased crosslink- density leads to very soft structures, which might
U
dissolve in the surrounding resin or cannot withstand the mechanical forces acting during the building
N
process. The final mechanical properties of such polymers are reached in the following post-curing
PT
ED
M
A
procedure, not during the printing process [6].
Figure 1: Principle sketch of the curing behavior of fast curing systems (e.g. acrylates, dashed line) and slow curing systems (e.g. vinyl ester, solid line).
CC E
The bottom-up vat-polymerization technology is a layer based additive manufacturing process (see Figure 2). Therefore, the actual cure depth has to be larger than the layer thickness. This is necessary to ensure sufficient adhesion between the layers (interlaminar bonding). Due to the inferior material
A
properties of the only partially cured vinyl ester biophotopolymers during the printing process, the interlaminar bonding between the layers is often insufficient and the printed parts break off during the part-vat separation process [7]. A possibility to avoid delamination is to increase the exposure and consequently the cure depth [8]. Therefore, the mechanical properties of the solidified layer, and the interlaminar bonding between the layers would increase too. However, that would impair the fabrication accuracy for reasons of over-curing, particularly in the building direction [9], [10]. Consequently, increasing the cure depth is not an option, especially when printing complex scaffold structures or patient specific implants for biomedical applications.
IP T
Figure 2: Bottom-up vat-polymerization. Reprinted with permission from [11].
SC R
Creating the Jacobs working curve [12] has become the basic procedure for testing and characterizing of new photocurable resins. When analyzing photopolymers, the Jacobs working curve is essential,
because it gives information about the penetration depth (Dp), cure depth (Cd) and the cure speed of the resin and its influencing parameters [13]–[18]. Cure depth tests and the resulting working curve are
U
necessary to characterize photopolymers. The working curve gives the relationship between exposure
N
(Emax) in [mJ cm-2] and responding cure depth (Cd) in [µm]. Creating a logarithmic-linear plot with Cd in y-axis and Emax in x-axis results in a straight line, called the Jacobs working curve (see Figure 3).
A
The intersection of the Jacobs working curve with the x-axis is the critical exposure (Ec) in [mJ cm-2].
M
Ec is related to the gel point of the resin, the starting point for the transition from liquid to solid and the slope of the Jacobs working curve represents the penetration depth (Dp) in [µm] (see Eq. (1)). Thereby,
𝐶𝑑 = 𝐷𝑝 ∗ ln (
𝐸𝑚𝑎𝑥 𝐸𝑐
)
𝐸(𝑧) = 𝐸𝑚𝑎𝑥 ∗ exp (−
A
CC E
PT
(Eq. (2)) [18], [19].
ED
the light-absorption of the resin (dependent of depth (z) in [µm]) follows the Beer-Lambert equation
Figure 3: Example of Jacobs working curve.
Eq. (1)
𝑧 𝐷𝑝
)
Eq. (2)
However, the Jacobs working curve does not give any information about mechanical properties of the cured photopolymer layer. It does not differentiate between soft solids (low crosslink density) and rigid solids (high crosslink density). This may not be a problem for common, fast curing and highlycrosslinked materials, but when using a photopolymer with lower curing speed and reduced cross-link density, the polymer network and the resulting mechanical properties need time to build up. The information from the Jacobs working curve as depicted in Figure 3 is not sufficient for indicating the printability of such resins, since the Jacobs working curve gives no information about the kinetics of
IP T
the polymerization process. A possibility to characterize and further optimize the curing behavior in order to increase mechanical properties at constant Cd, is to vary the amount of light absorber (LA) and photoinitiator (PI)
SC R
[20], [21]. A common approach is to determine the Jacobs working curve followed by printing and testing specimens. However, the printing and further testing of a variety of specimens is time and material consuming. Moreover, this characterization is complex and the results often have large
U
variance.
In this work, we present an improved technique, where a real time-near infrared (RT-NIR)
N
photorheometer is utilized to determine the time-resolved curing behavior of a photopolymer during
A
exposure of a single layer [22]. Therefore, we investigated the influences of LA and PI on the mechanical properties and double bond conversion (DBC) during curing. Moreover, it is possible to
M
create an extension to the common Jacobs working curve, a dataset with information on mechanical properties and DBC during curing of a single layer. Hence, a way to find the optimal mixture of LA
Materials & Methods
2.1
Materials
PT
2
ED
and PI to enhance the printability of resins with a low curing speed at constant Cd is presented.
CC E
The basic components of the photopolymer were the monomers divinyl adipate (DVA, Sigma-Aldrich®) and vinyl decanoate (10MV, Sigma-Aldrich®), the chain transfer agent trimethylolpropane tris(3-mercaptopropionate) (Thiocure® TMPMP, Bruno Bock) (see Figure 4) with radical inhibitor pyrogallol (PYR, Sigma-Aldrich®) to ensure sufficient storage stability of the
A
formulations [23] . The variable components were the photoinitiator Ivocerin® (PI, Ivoclar Vivadent®) and light absorber Quinoline Yellow SS (LA, Sigma Aldrich®). Altogether, nine different batches, all based on the mixture of 60 wt% DVA, 30 wt% 10MV, 9.98 wt% TMPMP and 0.02 wt% PYR were prepared by magnetic stirring at 50 °C for 3 h. The additional amount of PI and LA into the basic mixture can be found in Table 1. The final mixtures were stirred again at 50 °C for 3 h.
10MV
TMPMP DVA
Figure 4: Monomers 10MV (monofunctional) and DVA (difunctional), and chain transfer agent TMPMP.
0.50
1.0
0.04
#1
#2
#3
0.08
#4
#5
#6
0.12
#7
#8
#9
L-AMT system
U
2.2
0.25
SC R
LA [wt%]
PI [wt%]
IP T
Table 1: Additional amount of PI and LA in basic mixture.
N
For this study, a digital light processing (DLP) L-AMT system with upside-down set-up and an
A
InVision® WUXGA 1080p light engine with 460nm LED was used. It was used with an irradiation
M
intensity of 19 mW cm-2, measured at the surface of the vat [24]. Cure depth measurements: Creating the Jacobs working curves for the different mixtures, a vat was filled 3 mm high with photopolymer formulation (vat filling height >> cure depth, to reduce oxygen
ED
inhibition). Square geometries (5x5 mm²) with a hole of 1 mm diameter in the center were cured. All squares were cured with the same intensity (19 mW cm-2), while the exposure times were varied. After
PT
the curing, non-cured material was removed from the vat and the surface of the cured squares was cleaned with solvent (ethyl lactate) and compressed air. Afterwards, the thickness of cured material was measured with a thickness gauge (Kroeplin C110T, Kroeplin with a contact size of Ø 5 mm). The
CC E
measurements were performed in triplicates. Based on the Jacobs working curves the exposure for Cd = 200 µm (ECd - 200 µm) was calculated for the different formulations.
A
2.3
RT-NIR-Photorheometer
A unique combination of an Anton Paar MCR302 WESP rheometer with a light curing system accessory, combined with an OmniCure® 460 nm LED and additionally, a coupled in Bruker Vertex 80 Fourier-transform infrared spectroscopy (FTIR) was used. This setup is capable of measuring the curing behavior of photopolymers, mechanical properties as well as double bond conversion (DBC), simultaneously (Figure 5) [22].
Due to the possibility to adjust the wavelength of the used light-source, intensity and the gap size between glass plate and the measuring system, it was possible to simulate the conditions during the structuring process in the L-AMT system and to obtain a time-resolved analysis of the polymerization
U
SC R
IP T
process.
N
exposure
2.3.1
Measurement procedure
M
A
Figure 5: Schematic illustration of RT-NIR-photorheology setup. Reprinted with permission from [22].
ED
The photorheometer was equipped with a special black-coated parallel aluminum plate measuring system (Ø 25 mm). The black coating was used, to reduce light reflection of the aluminum plate, which could cause multiple-exposure of the sample. Despite the coated measuring system the NIR-
PT
signal for the double bond detection was still sufficient. 75 µl of photopolymer were pipetted onto the center of the glass plate. Additionally, the aluminum plate was “coated” with a thin layer of fluid
CC E
photopolymer, to avoid bubbles in the liquid volume while approaching the glass plate. Before reaching the final gap height, the aluminum plate stopped and excessive material was removed. Afterward, the machine proceeded to measurement position. In order to simulate the curing process in an L-AMT system, a gap height of 100 µm was chosen (which is the common layer thickness in an L-
A
AMT system).
The material was measured under periodic oscillation with a frequency 1 Hz and a shear strain of 1 % [25]. Previously performed amplitude tests [26] confirmed linear elastic behavior using those test parameters. The test procedure was divided into two phases: The first 60 s without exposure, to equilibrate the sample and to analyze the uncured resin with near-infrared spectroscopy (NIR). In the second phase, the photopolymer was cured with a light intensity of 19 mW cm-2 for 200, 300 or 500 s
(which conforms 3800, 5700 and 9500 mJ cm-²), dependent on the amount of PI and LA in the formulation. The measurements were performed in triplicate. 3
Analyzing mechanical properties and double bond conversion
Storage modulus G’, loss modulus G’’ and the normal force resulting from polymerization shrinkage were measured with the rheological set up and the decreasing carbon double bond signal (characteristic wavenumber at 6256-6127 cm-1) to get the DBC with the NIR-set up of the
4
IP T
photorheometer during exposure [22]. Results & Discussion
The Jacobs working curve gives information about how much exposure is needed to cure a certain
SC R
layer, the critical exposure to material gelation and the penetration depth for tested resin. It is a
valuable tool and suitable to characterize highly crosslinked photopolymers with fast curing behavior. This behavior can be called a digital curing behavior. Effectively, there are only two states: not
U
crosslinked (fluid) and highly crosslinked (fully cured).
Unfortunately, Jacobs working curve is insufficient to characterize photopolymers with decreased cure
N
speed, where the polymer network and the resulting mechanical properties of the cured layer need
A
more time to build up. Without the information of the mechanical properties of the cured layer it is hard to process these resins with high accuracy, which is necessary for printing complex patient
M
specific implants. Therefore, we investigated photopolymer mixtures (DVA, 10MV and TMPMP) with different amount of LA (0.04, 0.08 and 0.12 wt%) and PI (0.25, 0.5 and 1.0 wt%) to analyze the
ED
differences during curing and to find the most suitable photopolymer mixture for L-AMT printing. Furthermore, we present a way to optimize the mixture regarding mechanical properties at constant
4.1
PT
cure depth. Working curves
We created nine working curves of different photopolymer-formulations (see Table 1). The
CC E
coefficients of determination (R²) of the logarithmic regression line was above 0.99 for all measured working curves. Furthermore, the standard deviation of the single measurement points was below 12 µm (see resin with PI 0.5 wt% and 0.04, 0.08 and 0.12 wt% LA as representative example in Figure
A
6 and the summarized data in Table 2). As expected, higher amount of PI leads to a decrease of Ec but has low influence on Dp. Moreover, higher amount of LA leads to a decrease of Dp and an increase of Ec. Hence, the amount of PI has less influence on Dp than the amount of LA. The calculated exposure for the cure depth of 200 µm (ECd - 200µm) lies within a range of 386 mJ cm-2 (for PI 1 wt% - LA 0.04 wt%) to 3426 mJ cm-² (for PI 0.25 wt% - LA 0.12 wt%).
IP T SC R
Figure 6: Working curves of photopolymers with varying amount of light absorber (LA) and 0.5 wt% photoinitiator (PI). Calculated exposure (ECd - 200 µm) for particular resin PI 0.5 wt% - LA 0.08 wt% at intersection with cure depth Cd = 200 µm (red dot).
Photorheology
U
4.2
N
We measured the curing properties of nine different photopolymer formulations (Table 1) via photorheology. We plotted the photorheometer results of resin PI 0.5 wt% - LA 0.08 wt% as
A
representative example, to show the typical G’ and DBC curve during curing and to explain the data
M
and results (see Figure 7).
Rheology- and NIR- plot: As the light exposure starts, the DBC increases (see Figure 7, dashed
ED
curve). First, a short but strong increase was visible due to the high reactivity of the chain transfer agent until 10 % DBC, which transfers into a lower slope of the co-polymerization of 10MV and DVA. After an increase to 58%, the DBC-gradient decreases steadily until the final curing stage is
PT
reached (at 90% DBC).
As expected, the G’ slope is delayed to the DBC curve. An irradiation dose of 577 mJ cm-2 is required
CC E
to observe an increase in G‘. At that exposure the photopolymer filled gap of 100 µm was solidified (ECd - 100µm). The ECd - 100µm was detected at the rise of the normal force (data not shown). From ECd - 100µm on, the G’ increases steadily, which has its reflection point at 1835 mJ cm-2 (E (∂G'/∂E) max, see Figure 10) and ends in a steady decrease of the slope, until no change of G’ is detectable
A
(G’max = 2.1 MPa).
The G’max values of the photopolymer mixtures slightly increase with the amount of PI (G’PI 0.25wt% = 2.05 MPa ± 0.05, G’PI 0.5wt% = 2.15 MPa ± 0.1, G’PI 1.0wt% = 2.3 MPa ± 0.1), but no expressive change could be observed with variable amount of LA. The DBCmax of the mixtures does not follow the trend of G’max for the variable amount of PI. Only the formulation with PI 0.25 % had lower DBC than the other resins (DBC PI 0.25wt% = 88.2 % ± 0.1, DBC PI 0.5wt% = 93 % ± 2.6, DBC PI 1.0wt% = 91.4 % ± 1.2, graphs not shown).
IP T SC R U
4.3
M
A
N
Figure 7: Photorheometer curing graph of double bond conversion (DBC, dashed curve, left axis) and storage modulus (G’, solid curve, right axis) of formulation PI 0.5 wt% - LA 0.08 wt% (summarized results in Table 2). The ECd - 200µm value was imported from the working curve (Figure 6). The values for DBCCd – 200µm and G’Cd – 200µm were calculated at the intersection of G’ and DBC curve with ECd - 200µm, respectively. The G’∂G'/∂E max value can be found at the G’ curve at maximum of the slope (see Figure 10).
Combination of cure depth and cure degree
ED
Combining the results from the Jacobs working curve (L-AMT system) with the mechanical properties and double bond conversion (DBC) (photorheometer) gives a unique insight into the material
PT
properties during curing. We present a method, where the cure depth parameter from the Jacobs working curve are combined with the analyzed mechanical properties and double bond conversion (DBC) during curing from RT-NIR-photorheology. Thereby, mechanical properties (G’) and double
CC E
bond conversion (DBC) of a cured layer (h) cured with defined exposure (ECd) represents the cure degree. With this approach, photopolymers with low cure speed can be fully characterized and further optimized for successful L-AMT printing.
A
To ensure a sufficient adhesion between the new cured layer to the previous one (interlaminar bonding) in an L-AMT process, the actual cure depth has to be larger than the layer thickness. Thereby, the Jacobs working curves give information about the required exposure to cure a layer of 200 µm (ECd - 200µm). The particular G’- and DBC curves (photorheometer) were combined with the exposure ECd - 200µm from the Jacobs working curves. This results in the cure degree (mechanical properties (G’Cd - 200µm) and double bond conversion (DBCCd - 200µm)) of a 100 µm cured layer, cured with exposure ECd - 200µm. The results of G’Cd - 200µm and DBCCd - 200µm can be found in Figure 7 dashed red lines.
4.3.1
Analyzing the influence of the amount LA and on the curing behavior
The results of G’Cd - 200µm and DBCCd – 200µm of the different mixtures are summarized in Figure 8 and Figure 9, respectively. The lowest values for G’Cd - 200µm were found at the resin with lowest amount of LA (0.04 wt%), followed by the mixture with 0.08 wt% of LA. The highest values for G’Cd - 200µm were found for resins with 0.12 wt% of LA (see Figure 8).
IP T
A similar trend is also visible for the DBCCd - 200µm results (see Figure 9), where lowest values were found for resins with 0.04 wt% of LA, followed by the mixtures with 0.08 wt% of LA. The highest values for DBCCd - 200µm were found for resins with 0.12 wt% of LA.
SC R
For the variable amount of PI, no clear progression is visible (see Figure 8). We assume, when the
ratio between the amount of LA and PI is constant, the influence of the amount of PI to G’Cd - 200µm is minor. However, as this ratio was changed to higher amount of PI (1 wt% of PI), the G’Cd - 200µm increased stronger with increasing amount of LA. Additionally, the G’max values of fully cured
U
formulations could also have influence of the partially cured state at G’Cd - 200µm (resin with 1 wt% of PI
N
has up to 20 % higher G’max as resin with 0.25 wt% of PI). Therefore, it was considered, that the increased amount of PI has influence on the G’Cd - 200µm, especially the closer the actual exposure gets
A
to G’max.
M
However, more important is the clear trend of G’ and DBC regarding the increasing amount of LA in different mixtures (see Figure 8 and Figure 9). Resins with high amount of LA have substantially
ED
higher G’Cd - 200µm and DBCCd - 200µm, than resins with lower amount of LA, independent of the amount of PI. Consequently, the higher the amount of LA in the photopolymer formulations, the higher the
A
CC E
PT
G’Cd - 200µm and DBCCd - 200µm values at constant layer thickness (see Table 2).
Figure 8: G’Cd - 200µm values from the combination of cure depth (working curve) and mechanical properties (photorheometer) of nine different mixtures. The highest G’Cd - 200µm values were found for mixtures with 0.12 wt% light absorber (LA), independent of the amount of photoinitiator (PI).
IP T
.
Cure speed (∂G'/∂E)
U
4.3.2
SC R
Figure 9: DBCCd – 200µm values from the combination of cure depth (working curve) and double bond conversion (DBC) (photorheometer) of nine different mixtures. The highest DBCCd – 200µm values were found for mixtures with 0.12 wt% light absorber (LA), independent of the amount of photoinitiator (PI).
Results of resin PI 0.5 wt% - LA 0.08 wt% as representative example are presented in Figure 10, to
N
explain the data of the typical ∂G'/∂E- curve. The summarized results for all mixtures can be found in
A
Table 2.
M
By formation of the first derivative of the storage modulus (from photorheometer data) with respect to the exposure (∂G'/∂E, interpreted as cure speed), it is possible to determine the exposure, where the G´
ED
has its maximum slope (exposure at highest cure speed E(∂G'/∂E) max) (see Figure 7) and at that point, the ratio of increasing G’ to required exposure reaches its highest value. Therefore, we define the E(∂G'/∂E) max as the optimal exposure of tested resin regarding G’ (see Figure 10). Certainly, it is possible
PT
to cure with higher exposure, but excessive more exposure is required to gain relatively little more G’. Curing with less than E(∂G'/∂E) max is less time consuming, but the mechanical properties of the cured
A
CC E
layer and the interlaminar bonding between the layers is weaker.
IP T SC R
Figure 10: Cure speed (∂G'/∂E) of formulation PI 0.5 wt% - LA 0.08 wt%. The red dot marks the maximum of ∂G'/∂E, at exposure E (∂G'/∂E) max. The ECd - 200µm value was taken from the Jacobs working curve (Figure 7).
Exposure index (Ei)
U
4.3.3
N
By dividing the actual exposure ECd - 200µm by the defined optimal exposure E (∂G'/∂E) max, it is possible to calculate the exposure index Ei and evaluate the actual exposure (see Eq. (3)). The aim is, to get the
𝐸(𝐶𝑑 ) 𝐸(𝜕𝐺`/𝜕𝐸 )𝑚𝑎𝑥 (ℎ)
= 𝐸𝑖
Eq. (3)
ED
M
A
actual exposure close to E (∂G'/∂E) max, which would result in Ei = 1.
PT
As expected, the higher the amount of PI and the lower the amount of LA, the lower the exposure to reach the maximum cure speed (E∂G'/∂E max) (Table 2). The same trend takes effect to reach G’Cd - 200µm. However, as we can see in values of calculated Ei in Table 2, the amount of PI is relevant, but the
CC E
important variable to increase the G’ at constant Cd of 200 µm is the amount of LA. Thus, the higher the amount of LA, the closer Ecd – 200µm comes to E(∂G'/∂E) max and therefore Ei gets closer to
A
optimum value 1.
Table 2: Summarized data (critical exposure (Ec), penetration depth (Dp), exposure to cure a 200 µm layer (ECd -200µm) of Jacobs working curves, analyzed data of particular E (∂G'/∂E) max from photorheology and calculated exposure index (Ei) by Eq. (3).
Ec Dp -2 [mJ cm ] [µm]
ECd - 200µm E (∂G'/∂E) max Ei [mJ cm-2] [mJ cm-2] (opt. = 1)
LA 0.08 wt%
LA 0.12 wt%
4.3.4
333 213 147
219 210 207
830 553 386
2054 1326 984
0.40 0.42 0.39
PI 0.25 wt% PI 0.50 wt%
456 286
131 146
2108 1119
2932 1841
0.70 0.61
PI 1.0 wt% PI 0.25 wt% PI 0.50 wt% PI 1.0 wt%
178 638 380 209
130 119 118 97
825 3426 2078 1656
1309 3844 2556 1775
0.63 0.89 0.81 0.93
IP T
LA 0.04 wt%
PI 0.25 wt% PI 0.50 wt% PI 1.0 wt%
Calculating the optimal amount of LA for given PI
SC R
Using the particular data of ECd - 200µm and E (∂G'/∂E) max of the different mixtures, it is possible to
calculate the optimal amount of LA (LAopt.) and the therefore needed exposure (ELA opt.) to reach Ei = 1, at constant Cd (diagram with 0.5 wt% PI is shown in Figure 11). The LAopt. derives from the intersection of the extrapolated trend lines (logarithmic for 1 and 0.5 wt% of PI, linear for 0.25 wt% of
N
determination (R²) of all created trend lines was above 0.99.
U
PI) of ECd - 200µm and E(∂G'/∂E) max at 0.04, 0.08 and 0.12 wt% LA, respectively. The coefficients of
A
The results for LAopt. (at ELA opt) are 0.125 wt% (1833 mJ cm-2), 0.143 wt% (3085 mJ cm-2) and
M
0.162 wt% (4769 mJ cm-2) for 1, 0.5 and 0.25 wt% PI, respectively. Unexpected is the linear trend of the values of E(∂G'/∂E) max and ECd - 200µm for increasing amount of LA (0.04, 0.08 and 0.12 wt%) for the formulation with 0.25 wt% PI (diagram is shown in the
ED
supplementary material). We assume, that the amount of PI is below a certain threshold, so that the absorbance of the resin is too low to show typical logarithmic behavior. Further experiments have to
PT
be done to investigate these untypical behavior. However, the values of E(∂G'/∂E) max and ECd - 200µm of resins with 1 and 0.5 wt% of PI follows the
CC E
expected logarithmic trend for increasing amount of LA (0.04, 0.08 and 0.12 wt%) (diagram with 0.5 wt% PI is shown in Figure 11). Consequently, it could be shown, that it is possible to calculate the optimum amount of LA by creating the cross section of the trend lines of E(∂G'/∂E) max (photorheology) and ECd - 200µm (Jacobs working
A
curve), if the amount of PI is sufficient.
IP T SC R
5
U
Figure 11: Calculating the optimal amount of LA (LA opt.) and the required exposure (ELA opt.), for resin with PI 0.5 wt%, at the intersection (red dot, Ei = 1) of extrapolated logarithmic trend lines of ECd – 200µm (from working curve) and- E (∂G'/∂E) max (from photorheometer curve) for Cd = 200 µm and layer thickness 100 µm.
Conclusion
N
The aim of this work was to establish the results from real-time near-infrared
A
(RT-NIR)-photorheology as a supplementary to the Jacobs working curve, to fully characterize (cure
M
depth and cure degree) photopolymers with low cross-link-density and cure speed. Moreover, through analyzing the data from Jacobs working curve and RT-NIR-photorheology of vinyl ester based resins with different amount of light absorber (LA) and photoinitiator (PI), the optimal mixture for
ED
lithography-based additive manufacturing (L-AMT) of those particular formulations was found. The following findings summarize the view regarding photocuring of functional photopolymers with
-
PT
reduced cross-link-density and speed of solidification: The RT-NIR-photorheology is a valuable tool to investigate different photopolymer mixtures.
CC E
With the combination of Jacobs working curve (cure depth) and the mechanical properties and double bond conversion (DBC) from the photorheometer (cure degree), it is possible to completely characterize the curing behavior of photopolymers.
A
-
At a constant cure depth and layer thickness the amount of PI in the photopolymer mixture has less influence than the amount LA on the cure degree (mechanical properties and double bond conversion) of the cured layer. Moreover, the higher the amount of LA, the higher the storage modulus (G’) and DBC of the cured layer. Hence, it is easier to print three-dimensional objects.
-
Complete curing during the printing process (up to G’max) is not reasonable for resins with low cure speed. For curing a layer in the printing process, we defined the optimal exposure at the highest cure speed E (∂G'/∂E) max. The final mechanical properties of the printed part should be reached in a separate post-curing step.
-
The Exposure index (Ei) can be used to evaluate a photopolymer, or to compare different mixtures of photopolymers at different exposure situations. The closer the formulation to Ei = 1, the better.
-
The optimal amount of LA (and the required exposure E LA opt.) to achieve Ei = 1 (at constant cure depth) can be found at the intersection of the regression lines of optimal exposure E (∂G'/∂E) max and needed exposure E Cd - 200µm, when the amount of PI is sufficient.
Acknowledgements
SC R
fabrication of different test parts (e.g. scaffold structures) using L-AMT.
IP T
The next steps in order to confirm und further investigate these promising findings should be the
The research presented in this work was conducted at the TU Wien, Institute of Materials Science and Technology. The author gratefully thanks for the financial support by the
A
CC E
PT
ED
M
A
N
U
Austrian Science Fund (FWF) under research program No. P27059.
References
[8] [9]
[10]
[11] [12] [13] [14]
IP T
CC E
[15]
SC R
[7]
U
[6]
N
[5]
A
[4]
M
[3]
ED
[2]
David Kazmer, 28 - Three-Dimensional Printing of Plastics - Kutz, Myer, in: Appl. Plast. Eng. Handb. Second Ed. William Andrew, 2017. Samuel Clark Ligon, Robert Liska, Jürgen Stampfl, Matthias Gurr, and Rolf Mülhaupt, “Polymers for 3D Printing and Customized Additive Manufacturing,” Chem. Rev., vol. 117, no. 5, pp. 10212–10290, 2017. Chunguang Li, Jue Cheng, Wenkai Chang, and Jun Nie, “Photopolymerization kinetics and properties of a trifunctional epoxy acrylate,” Des. Monomers Polym., vol. 16, no. 3, pp. 274–282, 2013. Christian Heller, Martin Schwentwenwein, Günter Russmüller, Franz Varga, and Jürgen Stampfl, “Vinyl Esters: Low Cytotoxicity Monomers for the Fabrication of Biocompatible 3D Scaffolds by Lithography Based Additive Manufacturing,” Polym. Sci. Part A, vol. 47, no. 24, pp. 6941–6954, 2009. Samuel Clark Ligon-Auer, Martin Schwentwenwein, Christian Gorsche, Jürgen Stampfl, and Robert Liska, “Toughening of photo-curable polymer networks: a review,” Polym. Chem., vol. 7, no. 2, pp. 257–286, 2015. Bernhard Steyrer, Philipp Neubauer, Robert Liska, and Jürgen Stampfl, “Visible Light Photoinitiator for 3D-Printing of Tough Methacrylate Resins,” materials, vol. 10, no. 12, 2017. Xiangquan Wu, Qin Lian, Dichen Li, and Zhongmin Jin, “Tilting separation analysis of bottomup mask projectionstereolithography based on cohesive zone model,” Mater. Process. Technol., vol. 243, no. 1, pp. 184–196, 2016. Jim H. Lee, Robert K. Prud’homme, and Ilhan A. Aksay, “Cure depth in photopolymerization: Experiments and theory,” Mater. Res. Soc., vol. 16, no. 12, pp. 3536–3544, 2001. Paul O’Neill, Nigel J. Kent, and Dermont Brabazon, “Mitigation and control of the overcuring effect in mask projection micro-stereolithography,” presented at the 20th International ESAFORM Conference on Material Forming, Dublin, 2017. Jae-Won Choi, Ryan B. Wicker, Seok-Hyon Cho, Chang‐Sik Ha, and Seok‐Hee Lee, “Cure depth control for complex 3D microstructure fabrication in dynamic mask projection microstereolithography,” Rapid Prototyp. J., vol. 15, no. 1, pp. 59–70, 2009. Stefan Baudis, “Knochen, Knorpel und Gefäße maßschneidern mit Licht,” Nachrichten Aus Chem., vol. 64, no. 4, pp. 406–410, 2016. Ian Gibson, David W. Rosen, and Brent Stucker, Additive Manufacturing Technologies, Rapid Prototyping to Direct Digital Manufacturing. Springer, 2010. Ewa Andrzejewska, “Photopolymerization kinetics of multifunctional monomers,” Prog. Polym. Sci., vol. 26, no. 4, pp. 605–665, 2001. Bernhard Steyrer, Bernhard Busetti, György Harakály, Robert Liska, and Jürgen Stampfl, “Hot Lithography vs. room temperature DLP 3D-printing of a dimethacrylate,” Addit. Manuf., vol. 21, pp. 209–214, 2018. Hongping Xiang et al., “Preparation, Characterization and Application of UV-Curable Flexible Hyperbranched Polyurethane Acrylate,” Polymers, vol. 9, no. 11, 2017. Li Feng and Byoung I. Suh, “Exposure Reciprocity Law in Photopolymerization of MultiFunctional Acrylates and Methacrylates,” Macromol. Chem. Phys., vol. 208, no. 3, pp. 295–306, 2007. Sara Dalle Vacche, Valérie Geiser, Yves Leterrier, and Jan-Anders E. Månson, “Time-intensity superposition for photoinitiated polymerization of fluorinated and hyperbranched acrylate nanocomposites,” Polymer, vol. 51, no. 2, pp. 334–341, 2010. Paul Francis Jacobs, Rapid Prototyping & Manufacturing: Fundamentals of StereoLithography. California: Society of Manufacturing Engineers; 1st edition, 1992. Andreas Gebhardt and Jan-Steffen Hötter, Additive Manufacturing, 3D Printing for Prototyping and Manufacturing. Munich: Hanser Publications, 2016. Tom Scherzer and Ulrich Scherzer, “Real-time FTIR-ATR spectroscopy to study the kinetics of ultrafast photopolymerization reactions induced by monochromatic UV light,” Vib. Spectrosc., vol. 19, no. 2, pp. 385–398, 1999. Mohamed Mohamed Zabti, “Effects of light absorber on micro stereolithography parts,” University of Birmingham, Birmingham, 2012.
PT
[1]
[16]
A
[17]
[18] [19] [20]
[21]
A
CC E
PT
ED
M
A
N
U
SC R
IP T
[22] Christian Gorsche et al., “Real Time-NIR/MIR-Photorheology: A Versatile Tool for the in Situ Characterization of Photopolymerization Reactions,” Anal. Chem., vol. 89, no. 9, pp. 4958– 4968, 2017. [23] Paricher Esfandiari, Samuel Clark Ligon, Jean Jacques Lagref, Richard Frantz, Zoubair Cherkaoui, and Robert Liska, “Efficient Stabilization of Thiol-ene Formulations in Radical Photopolymerization,” J. Polym. Sci. Part Polym. Chem., vol. 51, pp. 4261–4266, 2013. [24] Robert Gmeiner, Gerald Mitteramskogler, Jürgen Stampfl, and Aldo R. Boccaccini, “Stereolithographic Ceramic Manufacturing of High Strength Bioactive Glass,” Appl. Ceram. Technol., vol. 12, no. 1, pp. 38–45, 2014. [25] Nono Darsono, Hiroshi Mizunuma, and Hiromichi Obara, “Rheological Study of the Solidification of Photopolymer and Dispersed Nanotube Systems,” Appl. Rheol., vol. 21, no. 11, pp. 63566–63581, 2011. [26] Thomas G. Mezger, Applied Rheology - With Joe Flow on Rheology Road. Anton Paar GmbH.