Design and application of discontinuous resindistribution patterns for semi-pregs
Sarah G. K. Schechter, Lessa K. Grunenfelder & Steven R. Nutt
To cite this article: Sarah G. K. Schechter, Lessa K. Grunenfelder & Steven R. Nutt (2020)Design and application of discontinuous resin distribution patterns for semi-pregs, AdvancedManufacturing: Polymer & Composites Science, 6:2, 72-85, DOI: 10.1080/20550340.2020.1736864
To link to this article: https://doi.org/10.1080/20550340.2020.1736864
Design and application of discontinuous resin distribution patternsfor semi-pregs
Sarah G. K. Schechter , Lessa K. Grunenfelder and Steven R. Nutt
Department of Chemical Engineering and Materials Science, University of Southern California, Los Angeles, California, USA
ABSTRACTVacuum-bag-only (VBO) prepregs fabricated with discontinuous resin (semi-pregs) on a uni-directional fiber bed reportedly exhibit high through-thickness permeability and yield high-quality laminates, even under adverse process conditions, such as poor vacuum or long out-times. In this work, semi-pregs were fabricated using fiber beds of various weaves, fibertypes, and areal weights (200–670 GSM). Flat and curved laminates were produced and char-acterized, confirming that porosity-free parts can be manufactured from a range of constitu-ent materials using VBO semi-pregs. Contoured laminates were produced with negligibleporosity, although a slight increase in bulk factor of prepreg plies was observed (D�0.1). Inaddition, design considerations and limitations for the fabrication of semi-pregs were pre-sented. The findings demonstrate that polymer film dewetting can be used effectively toproduce semi-pregs that yield porosity-free laminates via VBO processing, imparting robust-ness to out-of-autoclave cure of prepreg laminates.
GRAPHICAL ABSTRACT
KEYWORDSPrepreg; polymer matrixcomposites; porosity;defects; out of autoclave;vacuum bag only
1. Introduction
The application of discontinuous resin on differentfibers beds was evaluated for use in commonlyencountered manufacturing conditions for out-of-autoclave (OoA)/vacuum-bag-only (VBO) processingof composite prepregs, and design considerations weredetermined for the fabrication of such prepregs. Thestudy was motivated by limitations of current methods
for producing composite parts for aerospace, whichare commonly cured in high-pressure autoclaves(heated pressure vessels). These methods consistentlyyield defect-free parts, but autoclave processing hasdrawbacks, including high capital and operational cost,size restrictions, limited throughput (production bot-tlenecks), and high resource use (e.g. energy, nitrogengas). Alternative manufacturing methods are sought to
CONTACT Sarah G. K. Schechter [email protected] Department of Chemical Engineering and Materials Science, University of SouthernCalifornia, Los Angeles, CA, USA� 2020 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permitsunrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
reduce costs, overcome limitations, and accelerate pro-duction rates, particularly in aerospace [1].
VBO processing of composites offers a viablealternative to conventional autoclave cure methods[2, 3]. VBO processing is an OoA method in whichprepreg is vacuum bagged and cured in an oven.VBO processed parts can match the quality of partsmanufactured in an autoclave under favorable con-ditions. However, adverse process conditions,including poor vacuum, incomplete air evacuation,and/or high humidity, often yield defects, particu-larly porosity, which degrade mechanical perform-ance [4]. For every percent of void content in acomposite, up to a total void content of �4%, theinterlaminar shear strength decreases �7%, regard-less of the resin, fiber type, or fiber surface treat-ment [5]. Because VBO processing is limited to aconsolidation pressure of 0.1MPa (1 atm), air andother gases can be trapped during layup. Once theresin gels, trapped bubbles remain, resulting inpotentially unacceptable porosity levels (>1%). ForVBO processing to gain wider acceptance in high-performance applications, the process must consist-ently yield low porosity levels.
Background. Previous work has shown that use ofOoA/VBO prepregs with a discontinuous resin pat-tern (semi-pregs) virtually eliminates porositycaused by entrapped air or moisture, without anautoclave cure [6]. For example, Grunenfelder [7–9]showed that even under adverse process conditions,parts cured using semi-preg contained near-zerointernal porosity and no surface defects. In contrast,conventional VBO prepreg (with continuous resinfilms) yielded unacceptable levels of porosity(3–8%). Tavares [10] measured the through-thick-ness air permeability of a commercial semi-preg(‘Zpreg’) and an equivalent unidirectional (UD) pre-preg constructed with continuous resin. The perme-ability of the semi-preg material was three orders ofmagnitude greater than that of the continuous filmprepreg before and during the cure cycle, a resultattributed to a network of dry interconnected porespaces in the semi-preg design. As these studiesdemonstrate, prepregs featuring discontinuous resinfilms increase the capacity for air evacuation in thez-direction (transverse) by creating efficient egresspathways. This effect is attributed to the muchshorter air evacuation distances in the through-thickness direction (on the order of millimeters) ascompared to the in-plane direction (on the order ofmeters). Discontinuous resin not only increases theefficiency of air evacuation but also entraps less airbetween plies. VBO prepregs with discontinuousresin have the potential to mitigate process-induceddefects and impart robustness to the manufacture ofaerospace composites via OoA methods.
Semi-pregs can be produced by different meth-ods. We have previously reported one such method,which yields finely tuned resin distribution patternsand high through-thickness permeable prepregs[11–13]. The method relies on polymer film dewet-ting on a low surface energy substrate [14]. Usingthis method, resin film is first perforated on a sili-cone-coated backing paper (substrate) to create anarray of nucleation sites. The film is then heated,causing resin to recede from the nucleation sites.The resin recession is driven by the surface tensionbetween the low surface energy substrate (the sili-cone-coated backing paper) and the resin [14–17].The dewetted resin is transferred onto a fiber bedby briefly pressing the constituents in a hydraulicpress. This technique can be used to create a varietyof resin patterns, including stripes, islands, andgrids. In addition, the resin patterns created withthis technique can in principle be applied to any dryfiber bed.
Objectives. The goal of this study is to assess thedesign and application of the semi-preg format invarious common but challenging manufacturingconditions. Previous work has demonstrated thebenefits of discontinuous resin application in pre-preg processing. What has yet to be developed is abroad range of product forms for semi-pregs. Thekey developments described in this work are: (1) thefabrication of semi-preg produced from fiber bedswith various fiber types, weave types, and arealweights, and (2) the production of complex partsusing a woven semi-preg format. In this context,design considerations and limitations of semi-pregfabrication are also discussed.
Results reveal that low porosity laminates can befabricated using semi-pregs with various wovenfiber bed architectures. The use of discontinuousresin increases overall resin thickness, whichslightly increases the bulk factor. To address con-cerns regarding the increase in bulk factor, curvedlaminates with concave and convex corners wereproduced. Ultimately, part quality did not decreasewith the use of semi-pregs in the production ofcomplex shapes. Finally, initial design guidelinesare offered that outline the dimensional limitationsof resin patterns, degree-of-impregnation (DOI) ofsemi-preg, placement of constituents consideringthe general direction of the fiber bed and the resinpattern, and the use of the polymer film dewettingtechnique on some of the most widely used resinsystems. This work provides the first guidelines forthe design of semi-pregs – an important modifica-tion of OoA prepreg format that has the potentialto enhance the robustness of VBO processing ofprepreg laminates.
Prepreg fabrication. A total of seven fiber beds wereevaluated. Three fiber areal weights (200, 370, and670 GSM) were selected. Unless otherwise noted,fabrics were carbon fiber. At 200 GSM, four fabrictypes were evaluated: plain weave, glass fiber plainweave (Fibre Glast Development Corporation, Ohio,USA), 3K twill (Fibre Glast), and spread tow(Textreme, Sweden). The production of spread towfiber beds involves the spreading of a tow into a thinand flat UD tape and then weaving into a fabric. For370 GSM, a 5-harness satin (Fibre Glast) and a 6Ktwill (Fibre Glast) fiber bed were evaluated. Finally, at670 GSM, a 12K twill fiber bed (Fibre Glast) was eval-uated. A twill weave was selected for each areal weightto enable direct comparisons. The designations 3K,6K and 12K refer to the numbers of fibers withineach tow (i.e. 3000, 6000 or 12,000 fibers per tow).
Semi-preg and conventional prepreg plies werefabricated with epoxy resin (PMT-F4, Patz Materials& Technology, California, USA) at a resin contentof 35–36%. Semi-pregs were produced via selectivedewetting of resin film [11, 12]. With the neat resinfilm on the silicone-coated backing paper (a lowsurface energy substrate), nucleation sites wereintroduced using either a hand-held spike roller(HR-2, Robert A. Main & Sons, Inc., New Jersey,USA) or a box cutter. The spike roller pins werespaced at 6.35mm, and the roller was passed overthe entire film in straight passes. Using a box cutter,the resin on the silicone-coated backing paper was
scored in straight passes. The resin film was thenplaced in an air-circulating oven (Blue M Oven,Thermal Product Solutions, Pennsylvania, USA) todewet and grow the openings at the nucleation sites.The resin film was heated for 2min at 104 �C.Subsequently, the resin film was attached to thefiber beds by pressing the constituents briefly in anunheated hydraulic press (G30H-18-BCX, WabashMPI, Indiana, USA). Images of discontinuous resinpatterns produced using a spike roller, after applica-tion to various fiber beds, are presented in Figure 1.
To produce flat laminates, 16 plies of prepreg werestacked in a [0/90]4s sequence, where applicable.Initially, each ply was cut to 150mm � 150mm, andafter stacking, the edges of the stack were trimmed,resulting in dimensions of 140mm � 140mm. Thelaminates were vacuum bagged using standard con-sumables. Rather than utilizing edge breathing (com-mon practice with commercial VBO prepregs), theperimeter of each laminate was sealed with vacuumtape to restrict air evacuation solely to the through-thickness direction. Sealed edges were used to approxi-mate process conditions that limit or prevent in-planeair evacuation (e.g. large or complex parts, parts withply drops or corners). Laminates were cured accordingto the recommended cure cycle: a ramp of 1.5 �C permin to 121 �C followed by a 2-h dwell.
Commercial prepregs are produced from variousresin systems, including cyanate ester and bismalei-mide (BMI). To assess the compatibility of semi-preg fabrication via resin dewetting with differentresin systems, the technique was employed using a
Figure 1. Semi-pregs prior to cure, showing discontinuous resin distribution (fabricated via polymer film dewetting) on eachof the fiber bed types evaluated.emi-pregs
74 S. G. K. SCHECHTER ET AL.
cyanate ester resin film (54 GSM, PMT-F27, PatzMaterials & Technology, California, USA), a BMIresin film (71.5 GSM, RS-8HT, Toray, California,USA), and a toughened epoxy resin film (71.5 GSM,CYCOM 5320-1, Solvay, USA).
Porosity measurements. To evaluate the surfaceporosity of cured laminates, images of regions38mm � 38mm were recorded using a digitalmicroscope (VHX-5000, Keyence Corporation ofAmerica, California, USA). These images wererecorded at three locations across the laminate sur-face. For bulk void content, mutually orthogonalsections were prepared from the center of eachlaminate. Cross-sections were polished, and regions5mm � 20mm were imaged.
Images were analyzed using software (ImageJ) todetermine void contents. A series of steps was per-formed manually to produce binary porosity images.Using the software, images were converted to blackpixels for voids and white pixels for the rest of thelaminate. The percent porosity was calculated fromthe number of black and white pixels:
Porosity %ð Þ ¼ pblackpwhite þ pblack
� 100 % (1)
where p is the number of pixels.Bulk factor. Bulk factor is defined as the ratio of
the initial thickness, ti, of the prepreg stack (prior tocure) to the final thickness, tf, of the laminate (aftercure):
Bulk Factor ¼ titf
(2)
To calculate the bulk factor, the thickness of alaminate was measured prior to cure using a caliperat four locations around the perimeter. After cure,the laminate was sectioned, and a caliper was usedto measure the thickness of the laminate at fourlocations throughout the cross-section.
Complex shapes. To fabricate curved laminates, acustom test fixture [18] was utilized, featuring a 60�
concave and a 60� convex corner (with a roundedradius of 9.5mm), yielding parts with common
geometric complexities (Figure 2a). Tooling wasmachined from a single billet of aluminum to avoidleaks, and tool surfaces were fine polished. 8-plylaminates with dimensions of 75mm � 130mmwere produced using a 370 GSM 5-harness satinfiber bed.
An accepted metric for corner quality is thicknessvariability between the flanges and the curved por-tion. Parts produced at a concave corner areexpected to have corner thickening from pooling ofresin. On the other hand, parts made at a convexcorner are expected to have corner thinning. Todetermine the thickness variability, nine locationswere measured at the flanges and corner, as illus-trated in Figure 2b. The average thickness, x̄ , wascalculated using the following equation:
x ¼P
xin
(3)
where xi is the thickness at each individual locationand n is the number of measurement locations. Thecoefficient of variation (CoV) was then calculatedusing the following equation:
where the variables are defined identically as theprevious equation.
In situ monitoring. To monitor the resin flowfront during processing, the surfaces of the semi-pregs were tracked in situ using the techniquedescribed by Hu et al. [19, 20]. Four-ply semi-pregstacks using a 5-harness satin fiber bed were fabri-cated using either a grid or a striped pattern. Thestacks were laid against a glass window of an ovenand vacuum-bagged with standard consumables,sealing edges to prevent gas egress. Resin flow wasmonitored and recorded during a standard curecycle using a digital microscope and time-lapsevideo acquisition (Dino-Lite US, Dunwell Tech,Torrance, CA, USA). Temperature was monitored
Figure 2. (a) A custom test fixture, which allowed simultaneous production of a 60� concave and a 60� convex corner lamin-ate (with a rounded radius of 9.5mm). (b) Locations of measurement along the flanges and corners to calculate the coefficientof variation (CoV).
with a USB thermocouple data logger (LascarElectronics EasyLog EL-USB-TC-LCD), which wasattached adjacent to the laminate and directly onthe glass tool plate.
3. Results
3.1. Quality analysis of semi-preg formats
Key differences exist between UD and woven fiberbeds. Woven fabric exhibits crimp, or fiber wavi-ness, whereas crimp in UD fiber beds is negligible.Crimp varies based on weave type and can rangefrom minimal (i.e. spread tow) to large (i.e. plain
weave). High crimp generally reduces laminatestrength [21]. Secondly, woven fiber beds containpinhole openings between crossing tows, whichenhance transverse air evacuation, while UD fiberbeds do not. Finally, flow of resin in a woven fiberbed is multidimensional. Resin flows more rapidlythrough pinholes and in the depressions at tow inter-sections. Resin therefore flows through pinholes anddepressions (macroflow) before saturating tows(microflow). This dual-scale flow does not occur inUD fiber beds. These differences are expected to affectboth gas and resin flow in semi-preg materials.
Both conventional prepregs and semi-pregs werefabricated with a variety of fiber types, weave types,
Figure 3. Surfaces and cross-sections of each of the semi-preg formats evaluated. The fiber bed types are indicated as: PW(plain weave), GF (glass fiber plain weave), 3KT (3 K twill), ST (spread tow), 5HS (5-harness satin), 6KT (6 K twill), and 12KT(12 K twill).
76 S. G. K. SCHECHTER ET AL.
and fiber bed areal weights, after which cured laminatequality was assessed. As described in Section 2, lami-nates were fabricated from three areal weights (200,370, and 670 GSM), four weave types (plain weave,twill, satin weave, and spread tow), and two fibertypes (carbon and glass fibers). Micrographs of surfaceand internal porosity of all samples are presented inFigure 3, with the seven fiber bed types indicated asPW (plain weave), GF (glass fiber plain weave), 3KT(3K twill), ST (spread tow), 5HS (5-harness satin),6KT (6K twill), and 12KT (12K twill).
The first and second columns of Figure 3 depictthe surface and internal porosity, respectively, of lami-nates produced with conventional OoA prepreg for-mats (continuous film) using the seven different fiberbeds. All seven conventional laminates exhibited highlevels of surface and internal porosity, both of whichwere apparent via visual inspection. In all seven con-ventional laminates, porosity was concentrated at loca-tions corresponding to fabric pinholes and thedepressions at tow intersections. During cure, thesefeatures create points of low pressure toward whichresin and air migrate. Specifically, in the conventionalspread tow laminate (fourth row, first and second col-umns), in addition to porosity at the pinholes anddepressions, intra-tow porosity was also observed dueto the flatness of this fabric. In these prepregs pro-duced with spread tow fabric, the surface porosityobserved resembled defects observed with UD fiberbeds, as shown in previous work [11, 12].
The third and fourth columns of Figure 3 depictthe surface and internal porosity, respectively, of lami-nates produced with semi-preg using the seven differ-ent fiber beds. Each of the four semi-preg laminatesfabricated with 200 GSM fiber beds displayed near-zero surface and internal porosity, as depicted in thefirst through fourth rows of the third and fourth col-umns. Notably, in the laminate made from spread towsemi-preg (fourth row, third column), surface porositywas not detected, but vestiges of the applied discon-tinuous resin pattern were observed on the cured part
surface. Similar surface features have been reported inprevious studies of semi-preg materials made fromUD fiber beds [11, 12]. Semi-preg laminates producedwith 370 and 670 GSM fiber beds exhibited both sur-face and internal porosity after curing, as depicted inthe fifth through seventh rows of the third throughfourth columns. The surface and internal porosity wasattributed to incomplete saturation of the fiber bedduring cure, indicating that resin flow distances weretoo long to fully infiltrate. This intra-tow (flow-induced) porosity was greater for 670 GSM than for370 GSM fabrics, indicating a potential upper limit onfabric thickness for the present format of semi-pregprocessing. No inter-ply porosity (gas induced) wasobserved in any semi-preg samples.
A graphic summary of the surface and internalporosity for all samples is shown in Figure 4, withthe fiber bed types indicated as in Figure 3. The sur-face porosity (Figure 4a) of all conventional prepregsamples was greater than 1%, except for the 3K twillsample at 0.4%. All semi-preg samples, in contrast,exhibited less than 1% surface porosity, with mostsamples displaying no surface voids. While surfaceporosity was visible for the fiber beds with largerareal weights (�370 GSM), the measured surfaceporosity was �0.1%. Internal porosity (Figure 4b)for all samples produced using prepreg with con-tinuous film was 1% or more, with porosity levelsincreasing with fiber areal weight. Except for thepanel produced with 12K twill, the semi-preg sam-ples showed less than 1% porosity. Interestingly, the12K twill laminate produced from semi-pregresulted in more porosity (12.0%) than the samplefabricated from prepreg with continuous resin(7.1%). These findings indicate that an upper limiton fabric weight may exist for successful semi-pregprocessing, or that the cure process or semi-pregformat might require modification.
The potential benefits of semi-preg materialshave been previously reported, but a versatile rangeof product forms has not been developed. Here, we
Figure 4. (a) Surface porosity and (b) internal porosity of each of the prepreg and semi-preg formats evaluated. The fiber bedtypes are indicated as in Figure 3.
have demonstrated that semi-preg can be fabricatedfrom a range of fiber types, weave types, and fiberbed areal weights. The laminates produced fromsemi-preg with low and intermediate areal weightfabrics showed lower porosity (< 1%) than compar-able laminates produced with conventional formatprepreg. However, in laminates produced from highareal weight fabrics, porosity increased markedly(12.0%), because fiber bundles were too thick tofully impregnate during cure. To resolve this issue,further work must be undertaken, such as employ-ing a higher degree of impregnation (DOI) toreduce the flow distance (which will be discussed inSection 3.4) or utilizing a tailored resin system withlonger flow times at low viscosity.
3.2. Bulk factor
Bulk factor also was calculated for each laminatedescribed in Section 3.1. Bulk factor is relevant forfabrication of contoured parts because a large bulkfactor can cause wrinkling and/or bridging of plies.In principle, a bulk factor of 1.0 represents nochange in thickness during cure, a characteristicgenerally preferred. However, prepregs that incorp-orate dry regions for air evacuation intrinsically pos-sess bulk factors > 1. Bulk factor is furtherincreased when discontinuous resin distributions areused [11, 12].
For the semi-preg and prepreg formats evaluated(Figure 5), bulk factors varied with weave typebecause of the inherent differences in fiber bedcharacteristics (i.e. crimp). In addition, the bulk fac-tor generally increased as areal weight increased.The average difference in bulk factor between con-ventional prepreg and semi-preg with the same fibertype was �0.1. The fiber bed architecture that pro-duced the largest overall bulk factor was 5-harnesssatin (5HS). This fiber bed was selected for further
study via fabrication of panels with concave andconvex corner geometries (discussed in Section 3.3).
3.3. Complex shapes
Laminates with complex geometries were evaluatedto determine if and how the bulk factor increaseassociated with a semi-preg design affected partquality. Laminates with concave and convex cornerswere cured using both semi-pregs and conventionalformat prepregs, for a total of four curved lami-nates. Prepregs with a higher bulk factor undergogreater compaction during cure and often will notconform readily to a curved surface. When the bulkfactor is high, prepreg and consumable materialscan bridge over concave molds or wrinkle over con-vex tooling [22–29], introducing in-plane stressesand reducing the compaction pressure at corners.
Images of the four sample cross-sections are pre-sented in Figure 6. As the four images show, neitherthe semi-preg nor the conventional format prepregproduced wrinkling or extensive bridging at lamin-ate corners. However, a key distinction between thetwo prepreg formats was the levels of porosity inthe cured laminates. The conventional prepreg sam-ple exhibited both intra-tow and inter-ply porosity,as shown in the first column. In contrast, the semi-preg sample exhibited negligible intra-tow porosity,as shown in the second column. These results wereconsistent with those discussed in Section 3.1.
The CoV (Equation (4)) is a metric of cornerpart quality which describes the variability in thick-ness between the flanges and the curved surface of acomplex part. The calculated CoV of both convexand concave corners made with conventional pre-preg was 0.08. For the corners made with semi-preg,the CoV was 0.09, indicating that discontinuous filmmay slightly increase the variability in corner thick-ness. This increase can be attributed to either inherent
Figure 5. The measured bulk factor of each of the prepreg and semi-preg formats evaluated.
78 S. G. K. SCHECHTER ET AL.
experimental variation or to the higher bulk factor ofthe semi-preg.
Design considerations and limitations for the fabri-cation of semi-pregs were explored in the context ofprocessing. Certain aspects of design and fabricationof semi-pregs are critical to achieving high-qualityparts via OoA/VBO processing. Here, we examine:(1) dimensional limitations of the discontinuouspattern due to resin thickness and to uniformity; (2)DOI of the resin in the fiber bed; (3) placement ofdiscontinuous patterns and fiber beds that are eitherisotropic or anisotropic; and (4) polymer film dew-etting of different resin systems.
Feature dimensions. Polymer film dewetting typicallyoccurs at the edge of a dry region (i.e. perforations,tears, or deep depressions). When producing resinpatterns for semi-preg by dewetting, nucleation sitesare introduced by scoring or piercing the resinwhere openings are desired. However, a regular pat-tern is more difficult to achieve with resin that isnonuniform (high discontinuity), because uninten-tional nucleation sites will exist. Nonuniformity inresin films typically consists of uneven thickness,gaps, and dry space, which often arises during film-ing. These defects are more prevalent in thinnerresin films and have a greater effect.
Examples of striped resin patterns with differentspacings of imposed nucleation sites are presentedin Figure 7. The first row depicts the resulting dis-continuous resin patterns at 26 GSM when scoredat 5-mm and 1-mm intervals. As shown in the leftmicrograph, the resin did not dewet uniformlywhen scored at 5mm, and dewetting nucleated
spontaneously between the scorings. As depicted inthe right micrograph, the resulting pattern becamemore uniform when the scoring was reduced to 1-mm intervals, although some dewetting stilloccurred along the resin stripes.
The second row depicts the resulting discontinuousresin patterns at 54 GSM when scored at 10-mm and5-mm intervals. As shown in the right micrograph, bydoubling the resin areal weight from 26 GSM to 54GSM, uniform stripes were achieved with dewettingafter scoring at 5-mm intervals. However, as depictedin the left micrograph, when scoring spacing wasincreased to 10mm, feature precision was lost, andresin stripes were irregular in shape. The third rowdepicts a resulting discontinuous resin pattern at 188GSM. Here, the patterning was uniform, even withscoring at 20-mm intervals.
Uniform patterns can be achieved with thin resinfilms if nucleation sites are closely spaced. However,uniform patterns are difficult to achieve for thinand nonuniform resin films when nucleation sitesare spaced further apart, as depicted in the leftmicrographs of the first and second rows of Figure7. On the other hand, uniform patterns are readilyachieved with thick resin films, as shown in thethird row. To achieve uniform patterns across allresin thicknesses, the resin filming process must becarefully controlled to restrict random dewetting.
Degree of impregnation. The method described forfabricating semi-pregs via polymer film dewettingresults in a DOI near zero. Low DOI values requiretransverse microflow to saturate fiber tows prior togelation. Results in Section 3.1 revealed that lowporosity laminates were not achieved when semi-preg was fabricated with the thickest fiber bed(12K, 670 GSM). Presumably, the porosity resultedfrom the longer flow distances required, which
Figure 6. Part quality obtained from processing 5-harness satin prepregs and semi-pregs at a concave and convex corner.
prevented full saturation of fiber tows. Oneapproach to this problem is to increase the DOI toreduce the flow distances required to achieve satur-ation during cure. However, attempts to increaseDOI can potentially eliminate resin discontinuitiesand compromise the semi-preg format, asdescribed below.
The original intention of creating conventionalOoA prepregs with partial DOI (versus full satur-ation of the fiber bed) was to retain dry space at thecenter of fiber tows, which was a technologicaladvancement in prepreg manufacturing [30]. Inconventional OoA prepregs, these in-plane dryregions allow air to evacuate the prepreg via edgebreathing. DOI is defined as:
Degree of Impregnation DOIð Þ ¼ 1� Adry
Atow(5)
where Adry is the dry tow area and Atow is the entiretow area (dry and saturated) [31]. Commercial OoAprepregs exhibit levels of DOI which range from0.05 to 0.5. However, the prepreg fabricationmethod presented here involves briefly pressingfibers and resin films (continuous or discontinuous)using an unheated hydraulic press. This method
results in a negligible DOI into the fiber bed, andthe resin adheres only to the outer fibers of tows.
Images of the DOI of the resin into a 5-harnesssatin fiber bed produced in this manner is presentedin Figure 8(a,b). The micrograph of the surface isdepicted in Figure 8(a), and the micrograph of thecross-section is depicted in Figure 8(b). The cross-sectional image (Figure 8(b)) shows, in addition toexhibiting near-zero DOI, extensive macroflowbetween tows, as well as entrapped air bubbles. Alarger DOI would reduce the resin flow distancesrequired, as well as the bulk factor. Increasing theDOI was attempted by applying a pressure of0.09MPa to the semi-preg at 50 �C for 15min usinga hydraulic press. Images of the surface and cross-section of the semi-preg after this treatment are dis-played in Figure 8(c,d). This treatment altered anddisrupted the discontinuous resin pattern, as illus-trated by comparing Figure 8(a) and 8(c). Whenpressure and temperature were applied, resin flowedpreferentially along the length of the fibers beforeinfiltrating the fiber tows. In Figure 8(d), anexample of an entire tow bundle is outlined in pur-ple, and the remaining dry fiber region is outlined
Figure 7. Resin with various areal weights that were cut at different distances to demonstrate the limitations of featuredimensions for uniform patterning.
80 S. G. K. SCHECHTER ET AL.
in red. Using Equation (5), the volumes of thesetwo outlined regions were used to calculate theDOI. Here, the DOI increased to 0.25 þ/� 0.10.
A negligible DOI in semi-preg materials withthick fiber beds inevitably will result in long flowdistances. However, conventional means of increas-ing the DOI (i.e. high temperature and applied pres-sure) alter discontinuous resin patterns. Furtherdevelopment in prepreg production methods will berequired to achieve semi-pregs with controlled resinpatterning and increased DOI (and reducedbulk factors).
Isotropy & anisotropy. Some fiber beds (i.e. UD) andsome discontinuous resin patterns (i.e. stripes) areanisotropic. Other fiber beds (i.e. satin weaves) arenot anisotropic, but have adjacent, unidirectionallyaligned segments of tows (AUDAST). An exampleof an AUDAST fiber bed is 5-harness satin fabric.On one side of this fabric, segments of adjacenttows are aligned globally in one direction. On theobverse side, segments of adjacent tows are alsoaligned globally in one direction, but perpendicularto the segments of the first side. The segments oneach side of the fabric are woven together, whichresults in each side of the fabric containing
perpendicular tow warps. Thus, the fabric surface isnot entirely isotropic. This locally anisotropic struc-ture of the AUDAST fabric surface affects resin flowin semi-pregs, particularly during initial stages, andcan potentially disrupt, distort, and even eliminatepatterns of discontinuous resin.
When imposing resin patterns on such fiber beds(anisotropic and AUDAST), the placement of theresin pattern in relation to the fiber orientation atthe fabric surface can affect the ability to saturatethe fiber bed. During cure, resin flow occurs preferen-tially along fiber directions, rather than transversely[11, 12, 32]. Thus, if an anisotropic resin pattern isplaced parallel to an anisotropic or AUDAST fiberbed (e.g. striped resin pattern on 5-harness satin fiberbed), fiber bed saturation may not be achieved. Suchcare need not be required with isotropic fabrics (i.e.plain, twill weaves) and/or an isotropic resin patterns(i.e. uniform grid, islands).
To determine the influence of resin patternanisotropy and fiber orientation on fiber bedimpregnation, in situ observations were performedusing a transparent tool plate [19, 20]. A typicalcure cycle was employed using vacuum baggedsemi-pregs with select resin distributions applied toa 5-harness satin fiber bed. Figure 9 shows the
Figure 8. (a) The surface of semi-preg with a DOI of zero. (b) The cross-section of a semi-preg with a DOI of zero. (c) Surfaceof semi-preg that underwent elevated temperature and pressure to increase the DOI. (d) Cross-section of semi-preg with anincreased DOI.
progression of resin flow at the tool/part interface atdifferent temperatures during the cure of each ofthe semi-preg formats analyzed.
The first row of Figure 9 depicts the flow of resinduring the cure cycle of a 5-harness satin semi-pregwith an isotropic grid pattern. Resin quickly filledpinholes and depressions created by tow crimp ofthe 5-harness satin fabric (macroflow), as shown inthe 50 �C micrograph. Subsequently, microporespaces within the fiber tow bundles were slowlyfilled by microflow, as shown in the 85 �C micro-graph. By the end of the cure cycle, the fiber bedwas saturated, and exhibited only minor surfaceporosity, as shown in the ‘End of Cure’ micrograph.
The second row of Figure 9 depicts the flow ofresin during the cure cycle of a 5-harness satin semi-preg with an anisotropic striped pattern applied paral-lel to the primary fiber direction. Resin initially filledadjacent pinholes and the depressions of adjacenttows, as shown in the 50 �C micrograph. Thereafter,resin flowed preferentially into fiber tows that wereperpendicular to the resin stripes, as shown in the85 �C micrograph. This process was much slower thanthe infiltration times required for the grid pattern andresulted in regions of dry fiber tows (�9.2% of theentire surface area). The third row of Figure 9 depictsthe flow of resin during the cure cycle of a 5-harnesssatin semi-preg with an anisotropic striped patternapplied perpendicular to the primary fiber direction.Resin flow saturated this fiber bed much like the gridpattern depicted in the first row.
The results described above show that anisotropicresin patterns (e.g. stripes), when aligned parallel toan anisotropic or AUDAST fiber bed, yield onlypartial saturation. On the other hand, isotropic resinpatterns, such as grids, are more robust andimmune to adverse effects of fiber orientation. Thus,when designing semi-preg formats, isotropic pat-terns can be advantageous, as they effectively miti-gate nonuniform flow issues and achieve full fiberbed saturation regardless of fiber orientation.Anisotropic patterns such as stripes must be intelli-gently applied to ensure proper placement with ref-erence to the fiber orientation. However, stripesmay be simpler and more convenient to producerelative to other pattern designs, such as islands orgrid, depending on the production method of thediscontinuous resin distribution [6, 7, 10].
Alternative resin systems. Prepregs are produced with awide variety of thermoset resin types and formula-tions. To explore the versatility of semi-preg fabrica-tion via polymer film dewetting, the technique wasapplied to select resin systems: cyanate ester, BMI, anda standard commercial OoA epoxy (CYCOM 5320-1,Solvay, USA).
Images of the discontinuous resin patternspressed onto a UD carbon fiber bed and onto aplain weave glass fiber bed are shown in Figure 10.A grid pattern of cyanate ester was created andapplied to glass fibers (cyanate ester is often rein-forced with fibers other than carbon), as shown in
Figure 9. Resin flow for various pattern types and pattern placement with respect to the fiber bed during the cure cycle.
82 S. G. K. SCHECHTER ET AL.
the left micrograph. Similarly, grid patterns of BMIresin and a commercial epoxy resin (CYCOM 5320-1)were created and applied to UD carbon fibers, asshown in center and right micrographs. To achievethe same feature dimensions as the epoxy resin pat-terns studied in previous sections, the cyanate esterand BMI resin were heated to 104 �C for only10� 15 s. If heated for longer times, the resin filmdewetted extensively, forming small droplets.However, achieving the same feature dimensions inthe commercial epoxy required heating to 104 �Cfor more than 12min. While a dewetting processcan potentially be applied to any resin system, dew-etting conditions require adaptation to the specificresin system and substrate. The BMI and cyanateester films used here required less time (10� 15 s)to achieve feature dimensions equivalent to the ori-ginal epoxy resin system (2min), a result of thehigher interfacial tension with the silicone-coatedbacking paper. The commercial epoxy film had alower interfacial tension with the backing paper andthus required more time (> 12min) to achieve simi-lar dimensions.
The results shown in Figure 10 and describedabove demonstrate that polymer film dewetting can beused to produce discontinuous resin patterns from arange of resin systems. However, a resin can only beutilized for the fabrication of OoA parts if resin curekinetics are amenable to low pressure processing [4].VBO prepreg systems are designed to remain relativelyviscous in the early stages of cure to limit infiltration,prevent resin bleed, and retain sufficient dry areas forair evacuation [2]. The overall viscosity profile, how-ever, must also permit sufficient flow during the curecycle to fully saturate the fiber bed [3]. The rheologicalevolution of VBO resins must, in essence, balance theneed to prevent voids caused by entrapped gases (airand cure induced volatiles) as well as voids caused byinsufficient flow. Provided these conditions can bemet, a semi-preg fabricated with dewetted resin can beexpected to consistently yield high quality laminates
with OoA/VBO cure, even under adverse pro-cess conditions.
4. Conclusions
In general, laminates fabricated using semi-preg with avariety of fiber bed architectures exhibited near-zeroporosity. In addition, semi-preg formats increased thebulk factor of VBO-fabricated laminates by �0.1.Nevertheless, production of complex shapes usingsemi-preg revealed that part quality (i.e. wrinkling,bridging, and thickness variation) was similar to thequality achieved with conventional prepreg. By incor-porating multiple design considerations, the dewettingtechnique was applied to other resin systems, such ascyanate ester and BMI, to create uniform and repeat-able discontinuous resin patterns.
The use of semi-pregs to produce parts via VBOprocessing shows promise and may indeed impartrobustness. Nevertheless, some key limitations of theapproach were identified, and these may require fur-ther study. Current fabrication methods to createsemi-pregs result in larger bulk factors than conven-tional OoA prepregs (D�0.1), due primarily to thenear-zero DOI, and secondarily to thicker resinfrom the dewetting process. In addition to longerflow distances, the low DOI will limit the ability torework plies (lift and reposition) when required.Only practical experience will determine how ser-ious these drawbacks will be [33]. However, theadvantages imparted by the short breathe-out dis-tances in semi-pregs (enhancing robustness to theVBO manufacturing process) are expected to out-weigh these drawbacks. In particular, the use ofsemi-pregs may help to restore the process robust-ness sacrificed by abandoning autoclave curing forOoA processing. Nonetheless, one crucial challengethat remains is to demonstrate the scalability of thedewetting process. Current work has demonstratedonly a lab-scale batch process and a continuous pro-cess has not yet been shown. However, the process
Figure 10. Bismaleimide (BMI), cyanate ester, and commercial epoxy (CYCOM 5320-1) films that were created into discontinu-ous distributions by initiating nucleation sites by a spike roller and subsequently heating the resins at 104 �C.
may in fact be backwards-compatible with hot-meltprepregging, since in principle, imprint/de-wet stepscan be incorporated into existing prepreg produc-tion lines.
The development of semi-preg materials is apotential route toward robust OoA composite man-ufacturing. This work established that a large rangeof semi-preg product forms can be fabricated andmanufactured into high-quality laminates for bothflat and complex parts. The inherent manufacturingrobustness imparted by the methods presented herecan potentially expand the applicable uses of VBOprepregs within aerospace manufacturing and toother nonaerospace applications.
Acknowledgements
The authors are grateful to Claire Carlton for her assist-ance and for material donations from AirtechInternational (Cole Standish), Textreme (Jim Glaser),Toray Advanced Composites (Steve Smith), and Solvay.
Disclosure statement
No potential conflict of interest was reported by the authors.
Funding
This project was supported by the M.C. GillComposites Center.
ORCID
Sarah G. K. Schechter http://orcid.org/0000-0001-9000-3093Lessa K. Grunenfelder http://orcid.org/0000-0001-6561-401XSteven R. Nutt http://orcid.org/0000-0001-9877-1978
References
[1] Hueber C, Horejsi K, Schledjewski R. Review ofcost estimation: methods and models for aerospacecomposite manufacturing. Adv Manuf PolymCompos Sci. 2016;2:1–13.
[2] Ridgard C. Out of autoclave composite technologyfor aerospace, defense and space structures.Proceedings of the SAMPE 2009 Conference,Baltimore, MD; 2009.
[3] Steele M, Corden T, Gibbs A. The development ofout-of-autoclave composite prepreg technology foraerospace applications. SAMPE Tech Conference,Long Beach, CA; 2001.
[4] Centea T, Grunenfelder LK, Nutt SR. A review ofout-of-autoclave prepregs: material properties, pro-cess phenomena, and manufacturing considera-tions. Compos Part A Appl Sci Manuf. 2015;70:132–154.
[5] Pethrick RA. Bond inspection in composite struc-tures. Compr. Compos. Mater. 2000;5:359–392.
[6] Roman M, Howard SJ. Boyd JD. Curable prepregswith surface openings. US patent 2014/0174641A1, 2014.
[7] Grunenfelder LK, Dills A, Centea T, et al. Effect ofprepreg format on defect control in out-of-auto-clave processing. Compos Part A Appl Sci Manuf.2017;93:88–99.
[8] Nutt S, Grunenfelder L, Centea T. High permeabil-ity composite prepreg constructions and methodsfor making the same. US patent PCT/US2018/012665, 2018.
[9] Grunenfelder LK, Katz S, Centea T, et al.Through-thickness permeable prepreg for robustvacuum bag only processing. SAMPE J. 2018.
[10] Tavares SS, Michaud V, Månson J. Assessment ofsemi-impregnated fabrics in honeycomb sandwichstructures. Compos Part A Appl Sci Manuf. 2010;41:8–15.
[11] Schechter SGK, Centea T, Nutt SR. Polymer filmdewetting for fabrication of out-of-autoclave pre-preg with high through-thickness permeability.Compos Part A Appl Sci Manuf. 2018;114:86–96.
[12] Schechter SGK, Centea T, Nutt S. Fabrication ofout-of-autoclave prepreg with high through- thick-ness permeability by polymer film dewetting.Sample Tech Conference, Long Beach, CA; 2018;p. 2018.
[13] Schechter SGK, Centea T, Nutt S. Effects of resindistribution patterns on through-thickness airremoval in vacuum-bag-only prepregs. ComposPart A Appl Sci Manuf. 2020;130: 105723.
[14] Kohl JG, Singer IL. Pull-off behavior of epoxybonded to silicone duplex coatings. Prog OrgCoat. 1999;36:15–20.
[15] Kheshgi HS, Scriven LE. Dewetting: Nucleationand growth of dry regions. Chem Eng Sci. 1991;46:519–526.
[16] Roy S, Bandyopadhyay D, Karim A, et al.Interplay of substrate surface energy and nanopar-ticle concentration in suppressing polymer thinfilm dewetting. Macromolecules. 2015;48:373–382.
[17] Coussy O. Surface energy and capillarity. Mech.Phys. Porous Solids. Hoboken, NJ: John Wiley &Sons, Ltd; 2010. p. 107–147.
[18] Ma Y, Centea T, Nutt SR. Defect reduction strat-egies for the manufacture of contoured laminatesusing vacuum BAG-only prepregs. Polym Compos.2017;38:2016–2025.
[19] Hu W, Grunenfelder LK, Nutt SR. In-situ observa-tion of void transport during vacuum bag-onlycure. SAMPE Tech Conference, Long Beach, CA;2016.
[20] Hu W, Grunenfelder LK, Centea T, et al. In-situmonitoring of void evolution in unidirectional pre-preg. J Compos Mater. 2017;52:2847–2858.
[21] H€orrmann S, Adumitroaie A, Viechtbauer C, et al.The effect of fiber waviness on the fatigue life ofCFRP materials. Int J Fatigue. 2016;90:139–147.
[22] Albert C, Fernlund G. Spring-in and warpage ofangled composite laminates. Compos Sci Technol.2002;62:1895–1912.
[23] Hoa SV. Factors affecting the properties of compo-sites made by 4D printing (moldless compositesmanufacturing). Adv Manuf Polym Compos Sci.2017;3:101–109.
[24] Kar KK, Sharma SD, Mohanty A, et al. Short-termeffect of distilled water, seawater and temperature
84 S. G. K. SCHECHTER ET AL.
on the crushed and interlaminar shear strength offiber reinforced plastic composites made by thenewly proposed rubber pressure molding tech-nique. Polym Compos. 2008;37:915–924.
[25] Kar KK, Sharma SD, Kumar P. Effect of rubberhardness on the properties of fiber reinforced plas-tic composites made by the newly proposed rubberpressure molding technique. Polym Compos. 2007;37:915–924.
[26] Kar KK, Sharma SD, Sah TK, et al. Developmentof rubber pressure molding technique using butylrubber to fabricate fiber reinforced plastic compo-nents based on glass fiber and polyester resin. JReinf Plast Compos. 2007;26:269–283.
[27] Kar KK, Sharma SD, Behera SK, et al.Development of rubber pressure molding tech-nique using silicone rubber to fabricate fiber-rein-forced plastic components based on glass fiber andepoxy resin. J Elastom Plast. 2007;39:117–131.
[28] Kar KK, Sharma SD, Behera SK, et al.Development of rubber pressure molding tech-nique using butyl rubber to fabricate fiber
reinforced plastic components based on glass fiberand epoxy resin. J Appl Polym Sci. 2006;101:1095–1102.
[29] Kar KK, Sharma SD, Kumar P, et al. Pressure dis-tribution analysis of fiber reinforced plastic com-ponents made by rubber pressure mouldingtechnique. J Appl Polym Sci. 2007;105:3333–3354.
[30] Repecka L, Boyd J. Vacuum-bag-only-curable pre-pregs that produce void-free parts. SAMPE TechConference, Long Beach, CA; 2002.
[31] Palardy-Sim M, Hubert P. Characterization of thedegree of impregnation of out-of-autoclave pre-preg. 20th International Conference on CompositeMaterials, Copenhagen, July 19–24th 2015.
[32] Martinez P, Jin B, Nutt S. Effect of fiber bed archi-tecture on single resin droplet spread for prepregmanufacturing. CAMx Conference Proceedings,Anaheim, CA; 2019.
[33] Elkington M, Bloom D, Ward C, et al. Handlayup: understanding the manual process. AdvManuf Polym Compos Sci. 2015;1:138–151.
