Printer Friendly

Characterization of the auto-curing behavior of rapid prototyping materials for three-dimensional printing using dielectric analysis.

INTRODUCTION

During the last 20 years, printing technologies were implemented in the field of rapid prototyping (RP) because of their high efficiency proved on ink jet office printers [1-4]. In the Voxeljet three-dimensional printing (TDP) process, a curable resin called binder is printed on a flat powder bed. As the powder particles have diameters of 30-150 [micro]m, the prototypes exhibit a layer structure corresponding to a thickness of 80-150 [micro]m [5].

Prototypes with good mechanical performance are achieved with curable monomers--either by printing a UV-curable acrylate resin on a flat surface [3, 6] or a liquid binder containing an initiator on polymer powders containing a curing agent [7].

The two technologies differ significantly with respect to the curing time of the binder. The UV-curable binder is cured immediately after application by UV-light. The other binder is printed on the powder which wets and starts to swell the powder particles. After a certain induction time, the curing process starts as the binder has to mobilize and activate the initiator molecules which start to cure the liquid resin. The curing process has to be sufficiently slow to give the resin time for stress relaxation, but simultaneously high efficiency of the process must be attained.

Prototypes with optimal performance are achieved if the curing behavior of the powder-binder systems is well adjusted [8-10]. This requires precise knowledge of the curing kinetics and the effects of each binder component on the curing process. To investigate the curing kinetics of reactive polymer systems, usually methods such as Fourier Transform infrared spectroscopy (FTIR) or differential scanning calorimetry (DSC) are used. In FTIR, the time-dependent intensity of a band being characteristic for the curing reaction is evaluated to calculate the degree of cure (DC) [11-13] using the equation

[DC.sub.FTIR(t)] = (1 - [I.sub.characteristic band](t)/[I.sub.characteristic band](t=0)) (1)

with time-dependent degree of cure [DC.sub.FTIR] (t) determined by FTIR and intensity of characteristic 1R band [I.sub.characteristic band). Whereas in DSC, the time-dependent progress of exothermal reaction peak is evaluated and normalized to the curing heat [14-17]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

with time-dependent degree of cure [DC.sub.DSC](t) determined by DSC, heat flow [??](t), and curing heat [DELTA][Q.sub.curing].

Dielectric analysis (DEA) has been applied to investigate the curing behavior of polyester and epoxy resins in aircraft and automotive industry [18-24], and recently in dental composites [25-29]. The relevant quantity is the time-dependent ion viscosity [[eta].sup.ion] (t) which is sensitive to changes of the ion mobility. In DEA experiments, usually the complex dielectric constant [[epsilon].sup.*] is measured:

[[epsilon].sup.*] ([omega], T) = [[epsilon]' ([omega], T) - i[[epsilon]" ([omega], T) (3)

with dielectric constant [epsilon]', dielectric loss [epsilon]", frequency [omega], and temperature T. In a curing process, the quantity of interest is the time-dependent dielectric loss [epsilon]" (T, [omega], T) as it may change by two or more orders of magnitude if a resin reacts from the liquid state to the glassy state [27, 29, 30]. The dielectric loss consists of a conductivity part and a dipole part

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (4)

with ion conductivity [[sigma].sub.ion] dielectric permittivity [[omega].sub.0], and dielectric loss of the dipoles [[epsilon]".sub.dipole].

Therefore, one defines the ion viscosity [[eta].sup.ion] as an appropriate quantity to monitor the curing process of a polymer resin (corresponding to the rheological viscosity) whereas the dipole part can be neglected as long as the resin is in the liquid phase:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)

Zahouily ct al. [30] found that the ion viscosity depends on ion mobility [[mu].sub.ion], ion charge [q.sub.ion] and ion concentration [c.sub.ion] in the resin:

[[eta].sup.ion] (t, [omega], T) = 1/[[mu].sub.ion](t, [omega], T) * [c.sub.ion](t) * [q.sub.ion](t) (6)

Thus, the ion viscosity changes either if ions gain more mobility because of swelling and solution processes or the concentration of ions is changed by chemical reactions or addition of binder containing ions. The changes of ion concentration and ion mobility in the PMMA particles during swelling are not known. Thus, Eq. 5 is a simplified approach to the issue by evaluating the sample as a bulk. A mathematically description of the ion viscosity seeing the sample as a complex two-phase system governed by curing and swelling processes as well as temperature change will be an ambitious task for further studies.

The aim of this article is to show that the DEA can provide relevant information about the different stages of the curing process of 3DP materials if a measuring method allowing the direct monitoring of the ongoing curing process of the binder printed on a PMMA powder is developed. Furthermore, it is demonstrated how the DEA can be implemented in the rapid prototype printer system of the plant to control the curing process under the production conditions.

EXPERIMENTAL

Materials

The rapid prototyping printing materials used in this study are two component polymer systems consisting of PMMA powder and a liquid binder.

PMMA powder has an average particle size of 55 [micro]m and contains peroxide based cross-linking agent, Fig. 1 (left). The liquid binder VXP1 was supplied by voxeljet AG and consists of 2-hydroxyethyl methacrylate (HEMA) and styrene in the ratio 1:2. It was tested at two stages--new (VXP1-new) and 6 months old (VXP1-old).

The binder containing the activator swells the PMMA particles, radicalizes the cross-linking agent, and starts the curing process of the main binder components HEMA and styrene. As the liquid binder is solidified, the particles are fused together, Fig. 1 (right). To separate reaction heats generated by the binder curing from possible heats caused by swelling of the powder due to a binder addition, binders containing solely HEMA and styrene were also investigated, Table 1.

Methods and Sample Preparations, Data Evaluation

Equipments used in the study (Netzsch Geratebau, Germany) are summarized in Table 2.

DEA. For the lab experiments, the DEA was performed using 1DEX sensors having sensor geometry of (26 X 13) mm and electrode distance of 115 [micro]m. Well reproducible ion viscosity curves were achieved using the experimental setup shown in Fig. 2. To measure the temperature change during the curing reaction, a thermocouple was installed under the sensor. This arrangement avoided interference with the DEA measurement. Calibration measurements showed that the temperature difference between the thermocouple and the reacting powder is 1 K.

The DEA measuring procedure was as follows: 0.66 g of PMMA powder was put on the sensor and cautiously brought to a uniform height of approximately 1.5 mm. Then, the DEA measurement was started to obtain the ion viscosity of the PMMA powder. After 10 min, 0.33 g of binder were evenly dropped over the PMMA powder being on sensor surface with an Eppendorf pipette.

The DEA measurements were performed for 120 min at a temperature of 26[degrees]C with frequencies of 10, 1000, and 10,000 Hz. The frequency of 100 Hz was not applied as it often showed interference with the 50 Hz net frequency. Every 20 s, a data point was collected.

The DEA measurements in the production plant were performed with the above used IDEX sensor (disposable sensor) and the TMS sensor (tool mount sensor) having a sensor diameter of 18 mm and an electrode distance of 500 [micro]m (Table 2). Both sensors were mounted on a holder to affirm that the sensor surfaces are in the same plain. This affirmed that the same process step was measured. The most important difference of the two sensor types is the electrode distance as it corresponds directly to the sensing depth [28-30], DEA sensors are described in detail in Ref. [31].

The log [[eta].sup.ion] (t)-curves were evaluated with respect to the minimum ion viscosity, the onset and the endset of the curing process using the tangent method shown in Fig. 3. This method provides the time to minimum ion viscosity [t.sub.min], the start time [t.sub.start], and the end time [t.sub.end], which characterize the kinetics of the curing process together with the corresponding ion viscosities [[eta].sup.ion.sub.min], [[eta].sup.ion.sub.start], and [[eta].sup.ion.sub.end] and the initial ion viscosity [[eta].sup.ion.sub.0] of the PMMA powder. The temperature change during the DEA experiments is evaluated with respect to the times of peak onset, peak, and peak endset.

DSC Measurements. To determine the curing enthalpy, the DSC measurements were performed in the isothermal mode at a temperature of 26[degrees]C for 90 min. The sampling rate was 2 data points per minute. Both sample and reference crucibles were used without lids.

The measuring procedure was as follows: A sample crucible was put into the DSC cell containing 24 mg of PMMA powder being slightly compressed. Then, the DSC measurement was started. After 10 min, 12 mg of the binder was evenly dropped over the PMMA powder in the sample crucible with an Eppendorf pipette.

As the new binder VXP1 shows immediately after binder addition a small exothermal response, the DSC trace of VXP1 containing no activator was used as a baseline to determine the curing heat. The curing heat of aged VXP1 and HEMA could be determined using a horizontal line as baseline.

Compatible solvents act as softeners in polymers. Possible shifts of the glass transition temperature [T.sub.g] of PMMA to lower temperatures were investigated by dynamic DSC measurements. The DSC traces of solid PMMA, PMMA powder, and swollen PMMA powder were determined in the temperature range -40[degrees]C to + 150[degrees]C at a heating rate of 10 K/min using crucibles with unpinned lids. The sample of swollen PMMA powder was prepared by putting a certain amount of PMMA powder into the crucible, adding a corresponding amount of binder without activator, lock it with the lid. Then the sample was stored for 24 h in a refrigerator at +6[degrees]C to allow for swell and reduce binder evaporation.

RESULTS

The ion viscosity of the PMMA powder measured with a frequency of 10 Hz drops from high values in the dry state ([[eta].sup.ion.sub.0] [approximately equal to] [10.sup.12] W* cm) by three to five decades in the moment of the addition of a binder. In the case of VXP1 binder, it passes through a minimum and increases slowly for 60 min to values being half a decade below the initial ion viscosity. A long-term measurement showed that the increase still continues slowly even after 120 min. The ion viscosity curves show a good reproducibility, Fig. 4, especially if one takes into account that the temperatures within these four measurements differ between 26 and 30[degrees]C as in the real processing.

Three phases in the curing process can be distinguished in the ion viscosity curves after binder addition. First initiation, second primary curing, and third post-curing, Fig. 3. The two steps "initiation" and "primary curing" are better separated for 1,000 and 10,000 Hz. Furthermore, the time-dependent ion viscosities depend strongly on the chosen frequency, Fig. 5, as the ion viscosities before binder addition and at [t.sub.min], [t.sub.start], and [t.sub.end] exhibit significantly different frequency dependency, Table 3. For frequencies of 10, 1000, and 10,000 Hz (change by a factor 1000), the ratios of the ion viscosities (normalized to the 10,000 Hz value) change in a different manner: The initial ion viscosity [[eta].sup.ion.sub.0] decreases in the steps of 235:5:1 (10:1000:10,000 Hz). The minimum ion viscosity [[eta].sup.ion.sub.min] decreases in steps of 5:1:1. The start ion viscosity [[eta].sup.ion.sub.start] decreases in steps of 1.7:1.5:1 and the end ion viscosity (at 100 min) [[eta].sup.ion.sub.end] decrease in steps of 299:6:1.

The time to minimum ion viscosity [t.sub.min] is shifted toward the moment of binder addition with increasing frequency, Table 3.

To understand the curing process of VXP1 better, the effects of its constituents HEMA and styrene on the time and frequency dependent ion viscosity were investigated as well as aging of VXP1 during storage in a refrigerator for 6 months, Fig. 6. For all three frequencies, the ion viscosity curves of new VXP1 show a different time dependency. After the sharp drop due to a binder addition, it reaches quickly a minimum. Then, it increases continuously until reaching the end ion viscosity [[eta].sup.ion.sub.end]. The curing process with new VXP1-binder lasts typically 15 min longer than for HEMA and VXP1-old, Table 4. Furthermore, the ion viscosities at 10 Hz are at least one decade below that of HEMA and old VXP1-binder, Fig. 6-10 Hz, whereas they are at least a decade above for 1000 Hz and 10,000 Hz, Fig. 6-1000 and 10,000 Hz.

The HEMA and VXP1-old binders show a sharp drop with a subsequent plateau during the initiation phase before the primary curing leads to a sharp increase. The ion viscosity curves at 10 Hz differ in two features significantly (Fig. 6). First, after remaining constantly on the plateau level for 15 min, the curve shows a further decrease before suddenly the primary curing starts. Second, at the end of this sudden increase, the slope of the log [[eta].sup.ion] (t)-curve decreases significantly and remains constant for another 10 min. Therefore, two end times are provided for the ion viscosity curves at 10 Hz in Table 4.

The styrene binder shows a small drop after binder addition compared with the other binders and almost no further change of ion viscosity, Fig. 6. In the ion viscosity curves at 10 and 1000 Hz, a little step may be identified in the time range of 45-50 min after binder addition, however, even after 120 min, the samples were soft and uncured.

Temperature was recorded simultaneously to the ion viscosity measurements, Fig. 7. The four binders differ in their curing performance which leads to different temperature profiles, Table 5. The styrene binder shows no temperature change during the measurement. This means that no curing reaction takes place although the sample gets a soft consistency. The HEMA binder starts first curing reactions after 17 min and generates a sharp temperature peak having a half width of 5 min. The complete curing reaction lasts about 25 min. This means that pure HEMA binder cures rather quickly after the initiation phase. The VXP1-new binder generates a different temperature curve which shows a small temperature increase already after 5 min. This increase remains constant for another 20 min before a wide temperature peak occurs. The whole curing reaction lasts for approximately 70 min. The VXP1-old binder generates a temperature profile somewhat between VXP1-new and HEMA binders. The temperature peak is more pronounced and shifted to shorter times. The phase of a constant temperature increase is hardly to be seen.

The times to maximum temperature [t.sub.max] depicted in Table 5 coincide well with the end times [t.sub.end] of the ion viscosity measurements (Table 4). This means that the end of the primary curing is reached at the moment of maximum temperature increase.

The differences in the curing behavior of HEMA and both VXP1 binders are also seen in the corresponding DSC traces, Fig. 8. VXP1-new binder generates, after a short onset time [t.sub.onset] of 4 to 5 min, a constant heat flow over the next 10 to 15 min before primary curing generates a wide not pronounced heat flow peak after 27 min. Obviously, the aging process has three effects on the VXP1-new binder: shift of the heat flow peak to shorter times ([t.sub.max] = 19 min) toward the HEMA peak ([t.sub.max] = 16-7 min), heat flow peak becomes more pronounced and HEMA shaped, and vanishing of the initial heat flow plateau. Interestingly, the start times [t.sub.start] of all binders (Table 5) seem to be similar (22-25 min), but not the further reaction kinetics. The rate of heat release is significantly smaller as the reaction starts already after 5 min and ends after 70 min.

The time to the heat flow peak [t.sub.max] of VXP1-new binder is shifted by 2.5 min to shorter times if the sample mass is reduced to the half, Fig. 8. This affects directly the sample thickness in the crucibles, and therefore, the temperature increases during primary curing. This may explain the differences between the [t.sub.max] values determined by DSC and DEA. Furthermore, the VXP1-new sample showed a smaller peak area [DELTA][Q.sub.curing] in the DSC-curve indicating a lower curing enthalpy and thus, a lower degree of cure (DC). It is a known effect in DSC measurements that higher sample mass of curing material might be responsible for a higher temperature increase in the sample due to heat conduction effects. Higher reaction temperature might also cause a higher DC.

The onset temperatures of the glass transition of solid PMMA, PMMA powder, and swollen PMMA powder differ significantly, Table 6. Especially, the glass temperature of the swollen PMMA powder shows that VXP1-new binder acts as a softener.

The measurement of ion viscosity allows also for the tracing of the curing process under processing conditions, Fig. 9. The drops of ion viscosity are seen after each binder addition taking place typically after every 1.5 min. They are less pronounced as the amount of binder is less than half that of the lab tests. However, the depth sensitivity is limited to the electrode distances meaning that the IDEX sensor sees one or two processing steps, whereas the TMS sensor sees five to six processing steps. Thus, the process control of a whole RP part is not possible, but the DEA can be used to compare the measured ion viscosities of the process with a reference curve to document changes.

DISCUSSION

Time-dependent ion viscosity [[eta].sup.ion] (t, [omega], T) is a suitable quantity to trace curing and polymerization processes because of the occurring viscosity changes. These viscosity changes have a direct impact on the mobility of available ions. In the case of TDP the ion viscosity drops by three to five decades in the moment of binder addition. Then, it increases in a way which is characteristic for a given powder-binder combination, Fig. 6. If there is a curing reaction, the three phases--initiation, primary curing, and post-curing--can clearly be distinguished for all binders except of pure styrene binder showing no curing reaction at all. However, the ion viscosity shows that the processes taking place in the initiation and primary curing phases differ significantly in their time dependency, and thus in the underlying reaction mechanisms as well.

During the initiation, the binder swells the PMMA particles and mobilizes the initiator molecules trapped in the PMMA particles. Initiator-activator reactions generate radicals starting the copolymerization of styrene and HEMA. The initiation lasts a certain time until the ion viscosity has reached the starting ion viscosity [[eta].sup.ion.sub.start] which was found to be 5 X [10.sup.8] [[OMEGA]cm for the investigated PMMA-VXP1 system and reached after typically 34 min. It is obvious that VXP1-new binder generates different initiation processes than VXP1-old and HEMA binders exhibiting the ion viscosity plateau for about 20 min. The DSC curve of VXP1, Fig. 8, shows a constant exothermal heat flow already after an onset time of 4 min until primary curing starts. This means that chemical reactions take place already after 4 min. The corresponding ion viscosity increases slowly and continuously during this initiation phase. This increase can be explained by both reduction of monomer mobility because of the swelling of the PMMA particles combined with ion diffusion into the particles or a slow and immediately starting polymerization process.

Both processes increase the overall viscosity, and thus the corresponding ion viscosity. It should be noted that the change of the ion viscosity of VXP1 during the initiation phase--although being 2 orders of magnitude--contributes only less than 10% to the total change.

Interestingly, the slopes of the log(ion viscosity)-curves coincide well independently of frequency within the linear range of the initiation phase, Fig. 5. This fact together with the frequency independent starting ion viscosity [[eta].sup.ion.sub.start] indicate that the underlying initiation processes must be of a special nature. Possibly this behavior is linked to the heterogeneity of these TDP systems as both pure and fully cured PMMA powder systems exhibit the pronounced and almost reciprocal frequency dependency as expected according to Eq. 5.

After reaching the start time [t.sub.start] the reaction rate increases significantly as the system transfers to the phase of primary curing. This sudden change can be explained by the Trommsdorf effect [36, 37]. The polymer chains have reached a length which reduces their mobility significantly. The rate of radical annihilation reactions goes to zero and the polymerization is only determined by the monomer mobility. Because of the polymerization the viscosity of the binder system has reached a level --shown by the relatively high starting ion viscosity [[eta].sup.ion.sub.start]--keeping the polymerization rate moderate during the primary curing phase compared to VXP1-old and HEMA. This benefits twice to the performance of the final rapid prototyping parts as thermal stresses are kept low and enough relaxation time is given to them.

The curing behavior generated by the pure HEMA binder differs significantly. The ion viscosity remains on a constant but strongly frequency dependent level for 15 min. Then the ion viscosity measured at 10 Hz drops by more than one decade, whereas only small ion viscosity changes are observed at 1000 and 10,000 Hz. If one associates the ion viscosity at 10 Hz with the long range mobility of ions and the ion viscosity at 1000 and 10,000 Hz with short range mobility this means that a minimum long range mobility of the monomers is required to start polymerization. The decrease of ion viscosity of HEMA (approx. 15 min after binder addition) and VXP1-old (approx. 17 min after binder addition) at 10 Hz in Fig. 6 in the initiation phase coincide exactly with the increasing temperature (Fig. 7) due to the polymerization reaction. This phenomena can be associated with the increasing ion mobility due to the temperature increase. The ion viscosity curve minimum and the subsequent strong increase can be attributed to the viscosity increase due to polymerization. The fact that the ion viscosity is increased by almost three decades within less than 2 min after passing the minimum at 10 Hz shows that the polymer chain radicals have to be immobilized. This leads immediately to Trommsdorf-like polymerization behavior which is decelerated when the glass temperature of the HEMA styrene copolymer exceeds ambient temperature. Furthermore, it seems that the short range mobility does not depend on frequency as the ion viscosities at 1000 and 10,000 Hz are similar during the initiation phase.

Further, the behavior of VXP1-old binder can be explained because of the higher volatility of styrene indicated by lower melting and boiling temperatures, lower viscosity, and higher vapor pressure, Table 6. During storage the VXP1 binder lost styrene due to evaporation and permeation. This leads to an increasing HEMA concentration in the binder containers. As the number of ions in the binder remains constant it causes an increasing ion concentration, and thus decreasing the ion viscosity. The HEMA/styrene ratio of VXP1-old and VXP1-new binder was investigated with gas chromatography/mass spectroscopy (GC/MS) and compared qualitatively. The ratio of the HEMA and styrene peaks differed significantly showing that only a small content of styrene was left in VXPl-old. Small rests of styrene in the binder reduce its viscosity as styrene has a significant lower viscosity than HEMA decreasing the ion viscosity further. These two effects explain the lower level of the ion viscosity of VXP1-old during the initiation phase. Pure HEMA and VXP1-old binders cure rapidly within 2 min, whereas the VXP1-new binder requires almost 15 min, Table 4.

However, binder systems containing small amount of styrene obviously behave differently to that containing a large amount. Styrene and PMMA have similar solubility parameters, Table 6. Therefore, mainly styrene is responsible for swelling of the PMMA particles, and the release of initiator molecules. As styrene acts as a softener in PMMA, the glass transition temperature is decreased especially in areas of high styrene concentrations. In the areas where the glass transition temperature approaches ambient temperature, the release of initiator molecules becomes efficient and the curing starts. However, the degree of swelling, and thus the release of initiator molecules depend on the partial concentration of styrene in the binder. If the styrene concentration is low only small amounts of initiator molecules can be released and activated. The result is a rather low rate of curing which may lead to the limited increase of the ion viscosity at 1000 and 10,000 Hz during the initiation phase. During this phase, the HEMA-styrene copolymers have grown to a length that immobilizes their radical ends and primary curing starts in a Trommsdorf-like manner.

The ion viscosity curves of styrene binder show only a small drop with binder addition and no curing reaction. According to Eq. 6, the ion viscosity depends reciprocally on the ion mobility and the concentration of ions. As styrene has a significantly lower viscosity, and thus higher mobility than HEMA, it must be a liquid containing almost no ions. With respect to the behavior of the aged VXP1-old binder this means that the lower ion viscosity results from a reduction of the viscosity because of the remaining styrene and as well as an increase of the ion concentration because of the styrene evaporation. HEMA is essential to start the curing process of these binders.

The differences in the curing behavior of both VXP1 and HEMA binders provide a clear picture of the steps occurring within the curing process of this TDP rapid prototyping system consisting of initiator containing PMMA powder and activator containing binder. First, in the moment of addition, the binder wets the PMMA particles. Second, then mainly styrene starts to swell the PMMA particles from the surface and mobilizes the initiator molecules which can now react with the activator molecules of the binder to radicals. This effect is pronounced for high styrene concentrations in the binder as it leads to a significant decrease of the glass transition temperature of the PMMA due to softening. Third, at the beginning, only few radicals are generated and the curing rate is low. With ongoing radical generation the curing rate slowly increases until the HEMA-styrene chains are immobilized. Fourth, at this moment, the primary curing starts in a Trommsdorf-like manner. The curing rate increases further, and the chains grow until the mixture of copolymer and binder reaches its glass transition temperature. Fifth, now, the curing rate decreases significantly, and the system transfers into the post-curing phase.

CONCLUSION

The measuring method using the ion viscosity determined by DEA equipment with flat sensors allows precise characterization of the curing process of TDP materials and its reaction kinetics. A concept was developed which describes the curing steps until the binder comes into the post-curing phase. The styrene concentration in the binder determines the curing kinetics in a crucial way. High styrene concentrations swell remarkable parts of the PMMA particles, lead to a significant softening of the PMMA with a continuous release of initiator molecules. Thus, the radical concentration also continuously increases as well as the number of polymerizing chains which diffuses into the liquid binder. The polymerization-induced viscosity increase is reflected in the increase of the ion viscosity during the initiation phase. If the radical ends of the chains are immobilized, Trommsdorf-like polymerization starts, but the curing rate is moderated by already high viscosity in the case of VXP1-new binder. Low styrene concentration as in the case of aged VXP1 binder only swell the surface of the PMMA particles and mobilize a certain number of initiator molecules. The radical concentration as well as the number of polymerizing chains is stable. Polymerization processes only take place close to the particle surfaces. This does not strongly affect the viscosity of the liquid binder, and therefore, the ion viscosity remains on a plateau for a certain time before Trommsdorf-like polymerization starts in an uncontrolled way. Finally, it has been shown that the approach presented can be an important tool in the development and tailoring of powder-binder systems, especially as it is applicable under plant conditions.

REFERENCES

[1.] L.M. Sherman, Plast. Technol., 6, 8 (2004).

[2.] H. Lipson and M. Kurman, Fabricated: The New World of 3D Printing, John Wiley & Sons, 12-13 (2013).

[3.] I. Gibson, D.W. Rosen and B. Stucker, Additive Manufacturing Technologies--Rapid Prototyping to Direct Digital Manufacturing, Springer, 171-172 (2010).

[4.] A. Azari and S. Nikzad, Rapid. Prototyping. J., 15, 3 (2009).

[5.] Available at http://www.voxeljet.de/en/systems/material. Accessed on September 23, 2014.

[6.] SYS Systems Ltd Information Brochure. 3D Printing for Professionals. Available at http://www.sys-uk.com/documents/3d printing-for-professionals.pdf. Accessed on September 23, 2014.

[7.] Available at http://www.voxeljet.de/en/systems. Accessed on September 23, 2014.

[8.] B.M. Tymrak, M. Kreiger, and J.M. Pearce, Mater. Des., 58, 6 (2014).

[9.] R.A. Giordano, B.M. Wu, S.W. Borland, L.G. Cima, E.M. Sachs, and M.J. Cima, J. Biomater. Sci. Polym. E., 8, 1 (1997).

[10.] D. Dimitrov, K. Schreve, and N. De Beer, J. New Generation Sci., 4, 1 (2006).

[11.] K.C. Wu and J.W. Halloran, J. Mater. Sci., 40, 1 (2005).

[12.] M. Atai and D.C. Watts, Dent. Mater., 22, 8 (2006).

[13.] M. Guvendiren, B. Purcell, and J.A. Burdick, "Photopolymerizable Systems," in Polymer Science: A Comprehensive Reference, vol. 9, Polymers in Biology and Medicine, R.S. Langer, D.A. Tirrell, Eds., Elsevier B.V. 415, Chapter 22. (2012).

[14.] J.M. Barton, Adv. Polym. Sci., 72, 1 (1985).

[15.] J. Vaidyanathan, T.K. Vaidyanathan, Y. Wang, and T. Viswanadhan, J. Oral. Rehahii, 19, 1 (1992).

[16.] A. Maffezzoli, A. Della Pietra, S. Rengo, L. Nicolais, and G. Valetta, Biomaterials, 15, 15 (1994).

[17.] M. Rosentritt, M. Behr, A. Leibrock, and G. Handel, J. Mater. Sci., 10. 2 (1999).

[18.] N.T. Smith and D.D. Shepard. Sensors, 12, 10 (1995).

[19.] S.H. Nahm, J. Coat. Technol. Res., 3, 4 (2006).

[20.] A. Mcilhagger, D. Brown, and B. Hill, Compos. Part A Appl. Sci., 31, 12 (2000).

[21.] J.H. Choi, I.Y. Kim, and D.G. Lee, J. Mater. Process. Technol., 132, 1 (2003).

[22.] H.G. Kim and D.G. Lee, Compos. Struct., 57, 1 (2002).

[23.] K. Hongkyeong and C. Kookheon, Bull. Korean Client. Soc., 20, 11 (1999).

[24.] D. Kranbuehl, S. Delos, E. Yi, J. Mayer, and T. Jarvie, Polym. Eng. Sci., 26, 5 (1986).

[25.] M. Rosentritt and M. Behr, J. Mater. Sci., 41, 10 (2006).

[26.] M. Rosentritt, A.C. Shortall, and W.M. Palin, Dent. Mater., 26, 6 (2010).

[27.] J. Steinhaus, B. Hausnerova, T. Haenel, M. Grossgarten, and B. Moeginger, Dent. Mater., 30, 3 (2014).

[28.] J. Steinhaus, B. Moeginger, M. Grossgarten, M. Rosentritt, and B. Hausnerova, Dent. Mater., 30, 6 (2014).

[29.] J. Steinhaus, B. Moeginger, M. Grossgarten, and B. Hausnerova, Mater. Eng., 18, 1 (2011).

[30.] K. Zahouily, C. Decker, E. Kaisersberger, and M. Gruener, Eur. Coat. J., 11 (2003).

[31.] Netzsch product brochure: Dielectric Cure Monitoring; Method, Technique, Applications, 2013, 9, (2013_02_13) Available at http://www.netzsch-thermal-analysis.com/en/literature-digital-media/literature/brochures.html. Accessed on September 30, 2014.

[32.] Netzsch poster: Thermal Properties of Polymers, Version 2.1, Available at http://www.netzsch-thermal-analysis.com/en/litera ture-digital-media/literature/posters.html. Accessed on September 30, 2014.

[33.] GESTIS chemicals database of the German, Institut fur Arbeitsschutz (IFA) for CAS-Nr. 868-77-9 and CAS-Nr. 100-42-5. Accessed on June 14, 2014.

[34.] M. Belmares, M. Blanco, W.A. Goddard, R.B. Ross, G. Caldwell, S.H. Chou, J. Pham, P.M. Olofson, and C. Thomas, J. Comput. Chem., 25, 15 (2004).

[35.] S. Mandal and V.G. Pangarkar, J. Memhr. Sci., 209, 1 (2002).

[36.] D.C. Watts, Dent. Mater., 21, 1 (2005).

[37.] L.G. Lovell, K.A. Berchtold, J.E. Elliott, H. Lu, and C.N. Bowman, Polym. Adv. Technol., 12, 6 (2001).

Johannes Steinhaus, (1,2) Berenika Hausnerova, (2,3) Bernhard Moeginger, (1) Mohamed Harrach, (1) Daniel Guenther, (4) Florian Moegele (4)

(1) Department of Natural Sciences, Bonn-Rhein-Sieg University of Applied Science, von-Liebig-Str. 20, 53359 Rheinbach, Germany

(2) Faculty of Technology, Tomas Bata University in Zlin, namesti T. G. Masaryka 5555, 760 01 Zlin, Czech Republic

(3) Centre of Polymer Systems, Tomas Bata University in Zlin, Nad Ovcirnou 3685, 760 01 Zlin, Czech Republic

(4) Voxel jet AG, Paut-Lenz-Strasse 1a, 86316 Fried berg, Germany

Correspondence to: J. Steinhaus; e-mail: johannes.steinhaus@h-brs.de Contract grant sponsor: Netzsch Geratebau GmbH and Operational Program Research and Development for Innovations (European Regional Development Fund [ERDF] and national budget of Czech Republic; contract grant sponsor: Centre of Polymer Systems; contract grant number: CZ. 1.05/2.1.00/03.0111; B.H.).

DOI 10.1002/pen.24100

TABLE 1. Investigated binder types--binders contain
additives to start the curing process.

Binder type       HEMA:styrene ratio

Commercial VXPI      Approx. 1:2
HEMA                    01:00
Styrene                 00:01

TABLE 2. Thermo-analytical equipment.

Method                Measured quantity      Measuring device

Dielectric            Complex dielectric     Netzsch DEA 230/1 Epsilon
  analysis (DEA)      constant [epsilon] *
                      (t,[omega])
Differential                                 Netzsch 200 PC Phox DSC
  scanning
  calorimetry (DSC)
DEA oven                                     Netzsch Gefran DEA oven
DEA-IDEX sensor
Electrode distance:
  115 [micro]m
                      Ion viscosity
                        .sub.ion]
                        (t,[omega])
                      Heat How difference
                      Glass temperature
                        [T.sub.g]
                      Curing heat [DELTA]
                        [Q.sub.curing]
Electrode distance:   Temperature
  500 [micro]m

TABLE 3. Frequency-dependent ion viscosities and
characteristic times of VXP1-new binder determined by DEA.

                                                         Frequency

                                                          1,000  10,000
Parameter                              Unit        10 Hz   Hz      Hz

[[eta].sup.ion.sub.0]             G[OMEGA] (a) cm  1330   28.6    5.7
[DELTA][[eta].sup.ion.sub.0]                        363     2     2.1
[[eta].sup.ion.sub.min]           M[OMEGA] (a) cm   32     6.5    6.3
[DELTA][[eta].sup.ion.sub.min]                       1     1.8    1.8
[t.sub.min] (a)                   min               4.5    0.6    0.4
[DELTA][t.sub.min]                                  1.7    0.3    0.3
[[eta].sup.ion.sub.start]         M[OMEGA] (a) cm   784    689    457
[DELTA][[eta].sup.ion.sub.start]                    339    326    25.9
[t.sub.start] (a)                 min               23    24.2    24.9
[DELTA][t.sub.start]                                4.2    4.1    4.6
[[eta].sup.ion.sub.100]           G[OMEGA] (a) cm   386    7.2    1.3
[DELTA][[eta].sup.ion.sub.100]                      57     0.7    0.5
[t.sub.end] (a)                   min              35.5   38.8    39.1
[DELTA][t.sub.end]                                  4.2    4.4    4.2

(a) Characteristic times are corrected with respect to the time
of binder addition at t = 10 min.

TABLE 4. Frequency-dependent times of cure of the binders HEMA,
VXPl-old, and VXPI-new determined by DEA.

                       Start time [t.sub.start]

              10 Hz         1,000 Hz      10,000 Hz
Binder        [min]          [min]          [min]

HEMA       23.8           23.6           22.8
           [+ or -] 1.1   [+ or -] 0.3   [+ or -] 0.8

Styrene    34.5           34.3           n.e.
           [+ or -] 0.6   [+ or -] 0.5

VXP1-old   26.2           24.3           25

VXP1-new   26.8           27.6           29.6
           [+ or -] 3.4   [+ or -] 3.1   [+ or -] 2.9

                       End time [t.sub.end]

              10 Hz         1,000 Hz      10,000 Hz
Binder        [min]          [min]          [min]

HEMA       25.3/35.5      25.7           26.5
           [+ or -] 0.9   [+ or -] 0.9   [+ or -] 0.8

Styrene    48.1           45.6           n.e.
           [+ or -] 0.9   [+ or -] 0.9

VXP1-old   28.8/36.2      26.2           27.5

VXP1-new   38.9           42.4           43
           [+ or -] 3.7   [+ or -] 4.1   [+ or -] 3.3

n.e. = impossible to evaluate

The VXP1-old measurement was an additional measurement 6 months
later. After ageing there was not enough binder left for a multiple
determination. Comment to the evaluation: [r.sub.start] and
[t.sub.end] were determined according to Fig. 3 taking into account
that they refer to the time of binder addition after approximately
10 to 11 min after the start of the DEA equipment. As the 10-Hz
curves of HEMA and VXP1-old differ significantly, [t.sub.start] was
determined by the minimum and [t.sub.end] as the intercept of the
slope of the sharp ion viscosity increase with the slope of the
long-term ion viscosity.

TABLE 5. Frequency-dependent times of curing of the binders HEMA,
VXP1-old, and VXPi-new determined from the DEA temperature measurement;
curing heat AQcllrillg obtained from DSC.

           [t.sub.onset]   [t.sub.start]   [t.sub.max]
Binder     [min]           [min]           [min]

HEMA       17.1            22.2            24.9
           [+ or -] 0.5    [+ or -] 1.4    [+ or -] 0.7

VXP1-old   16.2            22.1            25.8
           --              --              --

VXP1-new   5.2             24.8            37.5
           [+ or -] 0.2    [+ or -] 1.3    [+ or -] 3.3

           [t.sub.end]    [DELTA][t.sub.max]   [DELTA][Q.sub.curing]
Binder     [min]          [K]                  [W/g]

HEMA       28.6           10.5                 95.7
           [+ or -] 0.7   [+ or -] 1.8         --

VXP1-old   31.8           8.9                  97.7
           --             --                   --

VXP1-new   47.8           2.4                  88.5
           [+ or -] 1.0   [+ or -] 0.4         --

The VXP1-old measurement was an additional measurement 6 months
later. After aging, there was not enough binder left for a multiple
determination.

TABLE 6. Physical properties of PMMA, HEMA, and styrene, [32-35].

Substance                 Unit                        PMMA

Chemical structure        --                      [FORMULA NOT
                                             REPRODUCIBLE IN ASCII]

Glass temperature         [degrees]C      [approximately equal to] 105
  [T.sub.g] [32]
[T.sub.g] onset
  temperatures
  (measured)
PMMA solid second run     [degrees]C      [approximately equal to] 84
PMMA powder                               [approximately equal to] 53
PMMA with HEMA-styrene                    [approximately equal to] 40
Melting temperature       [degrees]C                   --
  [T.sub.m] [33]
Boiling temperature       [degrees]C                   --
  [T.sub.b] [33]
Vapor pressure            hPa                          --
  [p.sub.v] at 21
  [degrees]C [33]
Density [rho] [33]        g/[cm.sup.3]                1.19
Viscosity [eta] at        mPa * s                    Solid
  20[degrees]C [33]
Solubility parameter      M[Pa.sup.0.5]                19
  [delta] [34, 35]

Substance                     HEMA         Styrene

Chemical structure        [FORMULA NOT   [FORMULA NOT
                          REPRODUCIBLE   REPRODUCIBLE
                           IN ASCII]      IN ASCII]
Glass temperature              --             --
  [T.sub.g] [32]
[T.sub.g] onset                --             --
  temperatures
  (measured)
PMMA solid second run
PMMA powder
PMMA with HEMA-styrene
Melting temperature          -12.0          -30.6
  [T.sub.m] [33]
Boiling temperature           251            145
  [T.sub.b] [33]
Vapor pressure                0.1            7.14
  [p.sub.v] at 21
  [degrees]C [33]
Density [rho] [33]            1.07           0.91
Viscosity [eta] at           0.696          5.784
  20[degrees]C [33]
Solubility parameter          51.1            19
  [delta] [34, 35]
COPYRIGHT 2015 Society of Plastics Engineers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2015 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Steinhaus, Johannes; Hausnerova, Berenika; Moeginger, Bernhard; Harrach, Mohamed; Guenther, Daniel;
Publication:Polymer Engineering and Science
Article Type:Report
Date:Jul 1, 2015
Words:6497
Previous Article:Block and random copolyamides of poly(m-xylylene adipamide) and poly(hexamethylene isophthalamide-co-terephthalamide): methods of preparation and...
Next Article:Processing of composite polystyrene foam with a honeycomb structure.
Topics:

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters