burnout effects on cellular ceramics obtained from gelatine gelcasted emulsified suspensions

J

A

Dctac©

K

1

tbfcccsg

mct

nj

h0

ARTICLE IN PRESS+ModelECS-9872; No. of Pages 9

Available online at www.sciencedirect.com

ScienceDirect

Journal of the European Ceramic Society xxx (2014) xxx–xxx

Burnout effects on cellular ceramics obtained from gelatine gelcastedemulsified suspensions

C. Freitas b, N. Vitorino a,∗, J.C.C. Abrantes a,b, J.R. Frade a

a Department of Materials and Ceramic Engineering, CICECO, University of Aveiro, 3810 Aveiro, Portugalb UIDM, ESTG, Polytechnic Institute of Viana do Castelo, 4900-348 Viana do Castelo, Portugal

Received 6 August 2014; received in revised form 15 October 2014; accepted 16 October 2014

bstract

ifferential thermal analysis, thermogravimetry and dilatometry were used to study burnout changes occurring during the earliest stages of firing ofellular ceramics obtained by gelcasting of emulsified suspensions. These analyses provide information on microstructural rearrangements duringhe initial burnout stages, and for the effects of heating rate, isothermal plateau and corresponding time on cell size distributions, average cell sizend other microstructural features of resulting cellular ceramics. Compressive strength and its scattering were also related to those microstructuralhanges induced by burnout conditions.

2014 Published by Elsevier Ltd.

eywords: Cellular ceramics; Taguchi planning; Emulsification; Gelcasting; Heat treatment

tdHbtasttag

. Introduction

Gelcasting has raised increasing interest as a versatile methodo process ceramics with different shapes,1,2 and to yield strongodies for green machining. Combination of gelcasting andoaming of ceramic suspensions3,4 has also been used to pro-ess porous ceramics with wide range of relevant microstructuralharacteristics, often by co-addition of pore formers. Classi-al pore formers may be polymeric sponges,5 natural productsuch as starch,6 synthetic microspheres of suitable polymers,7

raphite particles, etc.Emulsification of ceramic suspensions is also a versatile

Please cite this article in press as: Freitas C, et al. Burnout effects on cellulJ Eur Ceram Soc (2014), http://dx.doi.org/10.1016/j.jeurceramsoc.2014.1

ethod to process highly porous ceramics, with emphasis onellular ceramics.8 The use of highly volatile alkanes is attrac-ive for its prospects to allow recovery and reutilization of

∗ Corresponding author. Tel.: +351 258 819 700; fax: +351 258 827 636.E-mail addresses: [email protected] (C. Freitas),

[email protected] (N. Vitorino), [email protected] (J.C.C. Abrantes),[email protected] (J.R. Frade).

sdttercmc

ttp://dx.doi.org/10.1016/j.jeurceramsoc.2014.10.021955-2219/© 2014 Published by Elsevier Ltd.

hese organics by low temperature volatilization and upon con-ensation, before drying the resulting green cellular ceramics.owever, highly volatile alkanes are ill suited for gelcastingecause gelling additives often require previous treatment atemperatures approaching 100 ◦C. Therefore, one proposed anlternative method based on emulsification of ceramic suspen-ions with molten paraffin, to allow subsequent gelcasting atemperatures below the melting point of this fugitive phase9;his allows solidification of paraffin droplets before gelcastingnd drying, without undue coarsening, yielding relatively strongreen bodies.

Burnout conditions must be controlled to avoid overpres-ure of burnout gases resulting from pore former (paraffinroplets), gelling additive (dried gelatine), tensioactive addi-ives, etc. Otherwise, overpressure of burnout gases may leado failure of very fragile green ceramics. In addition, the pres-nce of molten paraffin is more likely to allow microstructuralearrangements, compared to solid pore formers. Thus, burnout

ar ceramics obtained from gelatine gelcasted emulsified suspensions.0.021

onditions may play significant effects on shrinkage andicrostructural features of resulting cellular ceramics, with

orresponding impact on microstructure dependent properties,

dx.doi.org/10.1016/j.jeurceramsoc.2014.10.021

http://www.sciencedirect.com/science/journal/09552219


mailto:[email protected]





ARTICLE IN PRESS+ModelJECS-9872; No. of Pages 9

2 C. Freitas et al. / Journal of the European

2 ºC min-1

5 ºC min-1

- 5 ºC min-1

Tiso

tisoββ

0

400

800

1200

1600

201612840

Tem

pera

ture

/ ºC

Time / h

Fs

sepswd

2

mD(fip(w6ad

mdbtrrphie

aaermopt

eaabsti((

ovmafm(cu

3

3

mem(ladopa

gmTfidtems

(wtaitrange (300–400 C), ascribed to combustion of paraffin, and the

ig. 1. Representative thermal cycle showing the burnout stage and subsequentintering for a representative sample, E3 in Table 1.

uch as mechanical strength,10,11 permeability,12,13 thermal orlectrical conductivity,14,15 etc. This is the main purpose of theresent study. Cellular ceramics obtained by gelcasting of emul-ified suspensions may be a model for other cellular ceramicshen pore formers undergo melting before oxidation or thermalecomposition.

. Experimental conditions

Alcoa CT3000 powder was used to prepare aqueous alu-ina suspension with a solid load of 50 vol%, with addition ofolapix PC-67 for dispersion and stabilization. This suspension

prepared at room temperature) was emulsified with paraf-n (Merck 1.07337.2500) above its melting temperature, witharaffin:suspension volume ratio = 1.5, by stirring at 1000 rpmBrinkmann Heidolph Mechanical Overhead Stirrer RZR1) andith additions of sodium lauryl sulphate (Sigma–Aldrich L-026) as tensioactive, and collagen (Oxoid LP0008) as a gellingdditive. Further details about emulsification and subsequentrying can be found elsewhere.9,16

The initial stage of heat treatment was adjusted based on ther-al analyses to avoid excessively fast weight losses, and risks of

isruption by evolution of large volumes of gases resulting fromurnout of paraffin and other additives. This stage is shown ashick lines in Fig. 1, and was adjusted by changing the heatingate (β), the isothermal temperature (Tiso) and time (tiso). Theemaining steps of the firing cycle were identical for every sam-le, with heating rate of 2 ◦C min−1 until 500 ◦C, followed byeating at 5 ◦C min−1 at temperatures above 500 ◦C and sinter-ng at 1550 ◦C for 2 h. The cooling rate was −5 ◦C min−1 forvery case.

Thermogravimetry (Netzsch STA409EP) was used to identify relevant temperature range for relatively slow weight losses,nd differential thermal analysis allowed one to distinguishndothermic phase changes from exothermic decompositioneactions. Dilatometry (Netzsch DIL402EP) was also used toonitor shrinkage or expansion during the initial burnout stages


f heat treatment. These results were used to establish a Taguchilan to study combined effects of initial heating rate, tempera-ure and duration of the isothermal plateau, with three levels for

aad

Ceramic Society xxx (2014) xxx–xxx

ach factor (Table 1). The relative impacts of a given factor on relevant microstructural parameters (average cell size, etc.),nd a related property (compressive strength) were assessedy averaging the results for experiments with this factor at theelected level. For example, the impact of isothermal plateauemperature on cell size was taken as the average of exper-ments (dc,1 + dc,4 + dc,7)/3 for the lowest level Tiso = 200 ◦C,dc,2 + dc,5 + dc,8)/3 for the intermediate level Tiso = 250 ◦C, anddc,3 + dc,6 + dc,9)/3 for the highest level Tiso = 300 ◦C.

The resulting cellular ceramics were used to assess the impactf burnout conditions on microstructural characteristics and rele-ant properties such as compressive strength. Scanning electronicroscopy (SEM – Hitachi SU1510) was used to determine

verage cell sizes, size distributions and other microstructuraleatures, using a suitable stereology code.17 The Archimedesethod was used to obtain information on total (x) and closed

xc) porosity. A Lloyd LR30K equipment was used to measureompressive strength under crosshead speed of 15 mm min−1,sing nine prismatic samples with length:width ratio = 2:1.

. Results and discussion

.1. Thermochemical changes during the burnout stage

Weight losses of samples processed by emulsification witholten paraffin and gelcasting (Fig. 2) are mainly determined by

limination of pore former (paraffin). Thus, one performed ther-ogravimetry of a paraffin sample (p) at identical heating rate

5 K min−1); this shows gradual increase in the rate of weightosses, reaching a peak at temperatures close to 300 ◦C, andpproaching complete elimination (≈98%) at 400 ◦C. Thermalegradation of gelatine was also assessed by thermal analysesf gelatine gelcasted alumina samples (a + g in Fig. 2), showingeaks of weight losses at temperatures in the order of 100 ◦C andlso in the order 300 ◦C, in agreement with literature results.18,19

Fig. 2 also shows the combined contributions of paraffin andelatine (g + p). This sample was prepared by emulsification ofolten paraffin in liquid gelatine, in the volume ratio 1.5:1.his sample was cooled to room temperature, causing solidi-cation of the paraffin droplets and inducing gelling, and thenried at about 50 ◦C; this sample shows an enhanced shoulder atemperatures in the range 200–300 ◦C, suggesting that burnoutffects of paraffin and gelatine in previously emulsified samplesay include a significant contribution which is not revealed for

eparate additives.Fig. 3 shows differential thermal analysis of samples

a + g + p) obtained by emulsification of alumina suspensionsith paraffin as pore former and gelatine as gelcasting addi-

ive. The relevant changes include the endothermic contributionscribed to melting of paraffin droplets, the exothermic shouldern the range 200–300 ◦C, ascribed to thermal degradation of gela-ine, a double exothermic peak in the intermediate temperature

◦


dditional exothermic peak at temperatures above 400 ◦C, alsoscribed to elimination of paraffin. Note that paraffin burnoutoes not show the initial shoulder. The DTA results for the



C. Freitas et al. / Journal of the European Ceramic Society xxx (2014) xxx–xxx 3

Table 1Taguchi plan to assess effects of heating rate β, isothermal plateau temperature Tiso, and time tiso on average cell size (dc), total porosity (x), closed porosity (xc),compressive strength (σc) and Weibull modulus (mW).

Sample β (K min−1) Tiso (◦C) tiso (h) dc (�m) x (%) xc (%) σc (MPa) mW

E1 1 200 1 23 65 6.0 41 3.0E2 1 250 2 20 66 3.8 42 2.2E3 1 300 4 21 65 2.9 58 5.7E4 2 200 2 24 65 4.4 36 1.6E5 2 250 4 15 66 4.0 54 3.6E6 2 300 1 14 66 4.3 35 2.7E7 4 200 4 24 65 2.6 36 4.1E8 4 250 1 19 66 3.0 50 5.3E9 4 300 2 19 67 6.0 65 7.5

f

→-3

-2

-1

0

0 200 40 0 60 0-100

-80

-60

-40

-20

0

Temper atu re /ºC

Los

ses (

wt%

)

a+g+pg+ppa+g

f

→

-0.0 3

-0.0 2

-0.0 1

0

0 20 0 400 60 0-1

-0.5

0

Temperat ure / ºC

DT

G /(

%K

-1)

a+g+ppg+pa+g

F −1 lcastp y em(

ap

bati

Fcpe(

gmim

ig. 2. Thermogravimetry performed at 5 K min for paraffin (p), alumina gearaffin in gelatin solution (g + p), and alumina + paraffin + gelatin obtained ba + g + p).

lternative (g + p) sample show that the intermediate temperatureeaks merge in a single broader peak.

Samples (g + p and a + g + p) also show an exothermic contri-ution at temperatures in the order of 220 ◦C for sample (g + p)

◦


nd circa 250 C for (a + g + p); this is even more pronouncedhat the DTG shoulder (Fig. 2). Note also that this contributions absent in the cases of samples containing only paraffin or

ig. 3. Differential thermal analyses at 5 K min−1 for paraffin (p), alumina gel-asted with gelatin (a + g), paraffin + gelatin obtained by emulsification of moltenaraffin in gelatin solution (g + p), and alumina + paraffin + gelatin obtained bymulsification of molten paraffin in alumina suspensions with gelatin additiona + g + p).

dda

pbt

F1

ed with gelatin (a + g), paraffin + gelatin obtained by emulsification of moltenulsification of molten paraffin in alumina suspensions with gelatin addition

elatine combined with alumina. Thus, the additional exother-ic contribution in samples (g + p) and (a + g + p) suggests

nteraction between gelatine and molten paraffin. However, thisay also be due to interaction between gelatine and sodium

odecyl sulphate, co-existing in (a + g + p) and also (g + p). Evi-ence of such interactions between gelatine and surfactants canlso be found in the literature.20,21

Fig. 4 shows differential thermal analyses of (a + g + p) sam-


les at different heating rates. These DTA results were dividedy the heating rate to account for corresponding differences inime scale, which drops with increase in heating rate. This allows

2

1K/min

5

10

-10

10

30

50

70

0 100 200 300 40 0 500

(DTA

/ββ)/[

mV

/(K/m

in)]

Temperature / ºC

ig. 4. Differential thermal analyses of (a + g + p) samples at 1, 2, 5 and0 K min−1.




125ºC

-0.6

-0.4

-0.2

0

0 10 0 20 0 300 400 500

DT

G/ (

%K

-1)

Temperature / ºC

1K/min

2K/min

5K/min

10K/min

Fig. 5. Differential thermogravimetry for (a + g + p) samples at 1, 2, 5 and1 −1

otiiiok

(actrcfd

cssoem(ad

arbitftlrpsnt

ba

carophpttfd

3

rmT(wEaEpalnbsIsl(u

daotitldmoptdmthan for E8, whereas average cell sizes show the reverse order(Table 1). The size distribution of sample E5 is also closer to

0 K min .

ne to compare relative changes in peak areas. The shoulder inhe range 200–300 ◦C shows significant dependence on heat-ng rate and splits into two separate contributions, whereas thentermediate double peaks cannot be de-convoluted for low heat-ng rates. The high temperature contribution is also dependentn heating rate, as expected for thermally activated solid stateinetics.

Thermogravimetry and differential thermogravimetryFig. 5) also show the effects of heating rate on the shouldert temperatures in the range 200–300 ◦C and on the lastontribution at temperatures above 400 ◦C. However, this failso de-convolute contributions in the intermediate temperatureange (300–400 ◦C). In addition, one easily observes a smallontribution with a peak rate at about 125 ◦C, accountingor about 0.5 wt% losses, and probably corresponding toehydration of gelatin.19

Dilatometry (Fig. 6) shows evidence that the burnout pro-esses are responsible for a significant contribution to totalhrinkage of the resulting cellular ceramics. The first step maytart at temperatures below 100 ◦C and may be ascribed to lossesf residual humidity and dehydration of gelatine.18 Yet, the rel-vant temperatures of the earliest shrinkage contributions onlyatch those of DTG (Fig. 5) at relatively low heating rates

1–2 K min−1). Therefore, one cannot exclude other factors suchs a slight initial deformation under loading exerted by theilatometer.

The main shrinkage contribution in Fig. 6 occurs at temper-tures in the range 300–400 ◦C and is fairly consistent with theange of highest weight losses (Fig. 5). Thus, shrinkage shoulde due to partial relaxation of the green ceramic wall during elim-nation of paraffin droplets; this may depend on the degree ofhermal degradation of gelatine. Note that shrinkage is enhancedor the highest heating rates, and occurs in narrower tempera-ure range; this corresponds to much shorter scale and is moreikely to shift denaturation of gelatine to higher temperature, asevealed by suppression of the DTA and DTG shoulders at tem-eratures in the range 200–300 ◦C (Figs. 4 and 5). Otherwise,lower heating should promote greater shrinkage upon elimi-


ation of molten paraffin. Slight expansion at temperatures inhe range 500–1000 ◦C may be ascribed to thermal expansion

np


ecause massive weight losses occur at lower temperatures andre unlikely to cause swelling.

The combined information extracted from thermal analysesonsistently suggests that burnout conditions should be adjustedt temperatures in the range 200–300 ◦C, to avoid risks of dis-uption or undue microstructural changes with negative impactn properties. These effects were, thus, the basis for the selectedarameters of the Taguchi plan (Table 1). The relevant levels ofeating rate in the initial stage and time scale of the isothermallateau were adjusted as βi+1 = 2βl and ti+1 = 2ti, to account forhe relevant ranges in DTA, DTG and dilatometry. Note alsohat solid state kinetics often shows decaying rate, as expectedor prospective relevant processes such as coarsening of paraffinroplet sizes, which is expected to vary as D2 ∝ t.22

.2. Microstructural characterization

Relations between early shrinkage and weight losses occur-ing in burnout conditions (Figs. 2–4) suggest significanticrostructural effects on the resulting cellular ceramics.his is, indeed, revealed by scanning electron microstructures

Figs. 7 and 8). Some samples show larger average cell sizes,ith emphasis on samples with a plateau at 200 ◦C (E1, E4 and7) and possibly also for samples with low heating rate (E1, E2nd E3). The highest average size is, indeed, found for sample1 which combines slowest heating and the lowest isothermallateau temperature. In addition, size distributions (Fig. 9) devi-te from the normal distribution, mainly for experiments withowest isothermal plateau temperature (200 ◦C). Onset of a sig-ificant fraction of relatively large cell sizes is probably causedy coarsening of molten paraffin droplets during the earliertages of heat treatment, before thermal degradation of gelatine.ndirect evidence of this is provided by cells with non-sphericalhape (e.g., E1 in Fig. 7). Note that thermal degradation of col-agen is expected to occur at temperatures above about 200 ◦CFigs. 2–5). Paraffin burnout requires even higher temperatures,sually above 300 ◦C (Fig. 5).

Coarsening of paraffin droplets may be halted by thermalegradation of gelatine, thus explaining the decrease in finalverage cell size for samples with isothermal plateau at 250 ◦Cr above, (Table 1). One also observes closer matching betweenhe average cell size and the peak of the distribution function onncreasing the isothermal plateau temperature (Fig. 9). In addi-ion, size distributions show smaller fractions of cell sizes mucharger than the average size. Thus, coarsening and abnormal sizeistributions are observed mainly for samples with the isother-al plateau at 200 ◦C. Still, this may also depend on the rate

f heating of the initial stage and/or duration of the isothermallateau, as found on comparing the size normalized distribu-ions for E5 and E8. These differences cannot be ascribed toifferences in total time scale, including the ramp and isother-al plateau. Note that this time scale is much longer for E5


ormal (Fig. 9). Thus, these results suggest that the heating ratelays a prevailing role.




1 K min-1

5 K min-1

2 K min-1

10 K min-1

-3%

-2%

-1%

0%

0 200 400 60 0 800 100 0 12 00

ΔL/L0

Temper ature / ºC

1 K min-1

5 K min-1

2 K min-1

10 K min-1

-0.3

-0.2

-0.1

0.0

0 200 40 0 60 0

103 (

dlnL

/dT

) / K

-1

Temperatu re / ºC

F gelag

apa

w

ig. 6. Dilatometry at 1, 2, 5 and 10 K min−1 obtained for alumina + paraffin +elatin.

Fig. 10 confirms that the isothermal plateau temperature plays


prevailing effect on average cell size, mainly for the lowestlateau temperature. The impact of heating rate and partiallylso duration of the isothermal plateau are less clear. In fact, one

std

Fig. 7. Scanning electron microstructures of cellular ceram

tin obtained by emulsification of molten paraffin in alumina suspensions with

ould expect more extensive coarsening with increasing time


cale, i.e., increasing time and also decrease in the heating rate;his is consistent only with differences for lowest and interme-iate values of time (tiso) and heating rate (β). These trends are

ics obtained under the conditions shown in Table 1.



6 C. Freitas et al. / Journal of the European Ceramic Society xxx (2014) xxx–xxx

ar cer

rcfsha(cw

cdt

sa

Fig. 8. SEM microstructures of cellul

everted for the highest values of these parameters, mainly if oneonsiders the impact of highest heating rate. The reverted trendor the highest heating rate must, thus, be ascribed to other causesuch as displacement of thermal degradation of dried gelatine toigher temperatures, as discussed above and revealed by DTAnd DTG shoulders at temperatures in the range 200–300 ◦C


Figs. 4 and 5). Delayed thermal degradation of gelatine mayontribute to retain the ability of the gelcasted green ceramicalls to undergo microstructural rearrangements involved in

Itt

amics showing closer views of struts.

oarsening of molten paraffin droplets. Delayed thermal degra-ation of gelatine may also explain the enhanced shrinkage athe onset of paraffin burnout in the range 250–300 ◦C (Fig. 6).

The effects of time may also involve a combination of oppo-ite trends corresponding to increase in paraffin droplet sizend more complete weight losses during the isothermal plateau.


ncomplete weight losses raise the risks of expansive effects ashe residual organic contents are burned on heating at higheremperatures.




0

0.4

0.8

1.2

0 1 2 3

(Dav

/Nt).

(δδN

/δ δD

)

Di/Dav

E1

E6

0

0.4

0.8

1.2

0 1 2 3

(Dav

/Nt).

(δδN

/δδD

)

Di/Da v

E5E8

Fig. 9. Normalized cell size distributions for E1 (triangles), E6 (circles), E5 (diamonds), and E8 (squares).

low middle high16

18

20

22

24

d c/μμm

Level

ratetimeTempe ratur e

F(

mtiftbtisouW

wm

pmsmwc

5.7

1.67.5

-2.2

-1.2

-0.2

0.8

43.532.5

ln[-l

n(1-

N/N

t)]

ln(σσc / MPa)

E1 E2E3 E4E5 E6E7 E8E9

Fi

fts(

Significant effects of time are expected by taking into accountthat longer time allows more complete evolution of burnoutgases, prior to next steps of the heating cycle, above the

low middle hig h35

40

45

50

55

σσ c/ M

Pa

ratetimeTemperatur e

ig. 10. Relative impact of heating rate, isothermal temperature (Tiso) and timetiso) on average cell size.

Mechanical properties of cellular materials show strongicrostructural dependence, with well-established evidence

hat flexural and compressive strength depend on porosity,ts distribution, average cell size and other microstructuraleatures.10,23,24 Though compressive strength is usually higherhan tensile strength,23 one expects reasonable correlationsetween those results. Compressive testing is by far simplero screen processing–microstructure–property relationships,ncluding effects of average parameters, such as average cellizes, and probably also other factors responsible for scatteringf compressive strength; this is emphasized by the Weibull mod-lus, i.e., the slope of a suitable representation (Fig. 11) of theeibull distribution:

N

Nt

= 1 − exp

[−

(σc

σo

)mW]

(1)

here N is the cumulative number of samples, mW is the Weibullodulus and σо is a fitting parameter.Figs. 12 and 13 also show the impact of different burnout

arameters on the average compressive strength and Weibullodulus. The average compressive strength (Table 1 and Fig. 12)

hows prevailing dependence on isothermal plateau temperature,


ainly for samples with the lowest isothermal plateau (200 ◦C)hen coarsening of paraffin droplets yields increase in average

ell size and also abnormal size distributions with a significantFc

ig. 11. Weibull plots of compressive strength for samples with heat treatmentsn Table 1.

raction of cells well above the average. These abnormal dis-ributions also explain the increased scattering of compressivetrength results, indicated by decrease in Weibull modulusFig. 13).


Level

ig. 12. Relative impact of heating rate, isothermal temperature and time onompressive strength.




low midd le high2

3

4

5

6W

eibu

ll m

odul

lus

Level

ratetimeTempera ture

FW

icomtctc

lrEmshatgrdmg(ttio

4

bsetbiwti

hmy

A

aP9

Q0

R

1

1

1

1

1

1

1

11

mal analysis and atomic force microscopy. Biophys J 2011;101(1):228–36.

ig. 13. Relative impact of heating rate, isothermal temperature and time oneibull modulus.

sothermal plateau as shown in Fig. 1. Longer time may alsoontribute to development of stronger struts. Still, the impactsf time on average compressive strength (Fig. 12) and Weibullodulus (Fig. 13) are less important than for the isothermal

emperature, probably because the resulting highly porous greeneramics allows ready evolution of burnout gases. Note also thathe impact on Weibull modulus is even smaller than on averageompressive strength.

The effects of heating rate confirm different trends for theowest and highest levels, as emphasized also by comparingesults for samples E3, E6 and E9 (Table 1). Note that sample6 combines low results of compressive strength and Weibullodulus, with smallest cell sizes. On the other way, sample E9

hows the best results of mechanical strength, possibly becauseigh heating rate prevents early thermal degradation of gelatine,s emphasized by displacement of the DTA and DTG shoulder inhe range 200–300 ◦C (Figs. 4 and 5), thus retaining the ability ofelcasted green ceramic walls to accommodate microstructuralearrangements; this should facilitate escape of burnout gasesuring the isothermal plateau. Microstructural rearrangementsay also promote thicker cell walls or stronger struts, as sug-

ested in Fig. 8 for sample E9, with best compressive strengthTable 1). Electron microstructures of sample E6 show muchhinner cell walls (Fig. 8), due to the expected dependence of wallhickness on cell size, and probably also because the heating rates lower than required to sustain microstructural rearrangementsn reaching the isothermal plateau.

. Conclusions

Microstructural rearrangements occur during the initialurnout stages upon firing cellular ceramics processed by emul-ification of suspensions with molten paraffin. The most criticalffects are related to coarsening of paraffin droplets for slow ini-ial heating and/or during an isothermal plateau at temperaturesefore thermal degradation of gelatine. This coarsening yieldsncrease in average cell size and abnormal size distributions,


ith negative impact on compressive strength, and scattering ofhese results, described by decrease in Weibull modulus. Highnitial heating rate displaces thermal degradation of gelatine to

1


igher temperatures, enhancing the ability to undergo furthericrostructural rearrangements, facilitating escape of gases and

ielding thicker walls or stronger struts.

cknowledgements

This work was financially supported by FEDER-COMPETEnd FCT, Portugal, within Projects PEst-C/CTM/LA0011/2013,TDC/CTM/ENE/2073/2012, and a PosDoc grant SFRH/BPD/9367/2013.

SEM facilities were funded by FEDER Funds throughREN – Aviso SAIECT-IEC/2/2010, Operacão NORTE-07-162-FEDER-000050.

eferences

1. Omatete OO, Janney MA, Strehlow RA. Gelcasting – a new ceramic formingprocess. Am Ceram Soc Bull 1991;70(10):1641.

2. Chen YL, Xie ZP, Yang JL, Huang Y. Alumina casting based on gelation ofgelatine. J Eur Ceram Soc 1999;19(2):271–5.

3. Sepulveda P. Gelcasting foams for porous ceramics. Am Ceram Soc Bull1997;76(10):61–5.

4. Sepulveda P, Binner JGP. Processing of cellular ceramics by foamingand in situ polymerisation of organic monomers. J Eur Ceram Soc1999;19(12):2059–66.

5. Ramay HR, Zhang MQ. Preparation of porous hydroxyapatite scaffolds bycombination of the gel-casting and polymer sponge methods. Biomaterials2003;24(19):3293–302.

6. Lyckfeldt O, Ferreira JMF. Processing of porous ceramics by ‘starch con-solidation’. J Eur Ceram Soc 1998;18(2):131–40.

7. Ganesh I, Jana DC, Shaik S, Thiyagarajan N. An aqueous gelcast-ing process for sintered silicon carbide ceramics. J Am Ceram Soc2006;89(10):3056–64.

8. Barg S, de Moraes EG, Koch D, Grathwohl G. New cellular ceramicsfrom high alkane phase emulsified suspensions (HAPES). J Eur CeramSoc 2009;29(12):2439–46.

9. Vitorino N, Abrantes JCC, Frade JR. Cellular ceramics processed byparaffin emulsified suspensions with collagen consolidation. Mater Lett2013;98:120–3.

0. Brezny R, Green DJ. The effect of cell-size on the mechanical-behavior ofcellular materials. Acta Metall Mater 1990;38(12):2517–26.

1. Chen F, Yang Y, Shen Q, Zhang L. Macro/micro structure dependence ofmechanical strength of low temperature sintered silicon carbide ceramicfoams. Ceram Int 2012;38(6):5223–9.

2. Innocentini MDM, Sepulveda P, Salvini VR, Pandolfelli VC, Coury JR.Permeability and structure of cellular ceramics: a comparison between twopreparation techniques. J Am Ceram Soc 1998;81(12):3349–52.

3. Moreira EA, Innocentini MDM, Coury JR. Permeability of ceramicfoams to compressible and incompressible flow. J Eur Ceram Soc2004;24(10–11):3209–18.

4. Petrasch J, Meier F, Friess H, Steinfeld A. Tomography based determina-tion of permeability, Dupuit–Forchheimer coefficient, and interfacial heattransfer coefficient in reticulate porous ceramics. Int J Heat Fluid Flow2008;29(1):315–26.

5. Rice RW. Evaluation and extension of physical property-porosity modelsbased on minimum solid area. J Mater Sci 1996;31(1):102–18.

6. Sanches MF, Vitorino N, Abrantes JCC, Frade JR, Rodrigues Neto JB, HotzaD. Effects of processing parameters on cellular ceramics obtained by paraffinemulsified suspensions. Ceram Int 2014;40:9045–53.

7. Abrantes JCC. Estereologia. Software package. UIDM/IPVC; 2001.8. Bozec L, Odlyha M. Thermal denaturation studies of collagen by microther-


9. Cucos A, Budrugeac P. Simultaneous TG/DTG-DSC-FTIR characteriza-tion of collagen in inert and oxidative atmospheres. J Therm Anal Calorim2014;115(3):2079–87.


http://refhub.elsevier.com/S0955-2219(14)00568-8/sbref0005






























































































































































































































































































































































































































































































































































pean

2

2

2

2

C. Freitas et al. / Journal of the Euro

0. Cooke DJ, Dong CC, Thomas RK, Howe AM, Simister EA, Pen-fold J. Interaction between gelatin and sodium dodecyl sulfate at theair/water interface: a neutron reflection study. Langmuir 2000;16(16):


6546–54.1. Sovilj V, Dokic P, Sovilj M, Erdeljan A. Influence of surfactant–gelatin

interaction on microcapsule characteristics. Trends Colloid Interface Sci X1996;100:235–40.

2

Ceramic Society xxx (2014) xxx–xxx 9

2. Wagner C. Theorie der alterung von niederschlagen durch umlosen(Ostwald-Reifung). Z Elektrochem 1961;65(7–8):581–91.

3. Brezny R, Green DJ. Uniaxial strength behavior of brittle cellular materials.


J Am Ceram Soc 1993;76(9):2185–92.4. Presser V, Schultheiss S, Kohler C, Berthold C, Nickel KG, Vohrer A, et al.

Lessons from nature for the construction of ceramic cellular materials forsuperior energy absorption. Adv Eng Mater 2011;13(11):1042–9.






































































































































































burnout effects on cellular ceramics obtained from gelatine gelcasted emulsified suspensions

Documents