Sei sulla pagina 1di 15

ARTICLE IN PRESS

Biomaterials 27 (2006) 13621376 www.elsevier.com/locate/biomaterials

Review

A comparison of micro CT with other techniques used in the characterization of scaffolds


Saey Tuan Hoa, Dietmar W. Hutmachera,b,
b

Faculty of Engineering, Division of Bioengineering, National University of Singapore, 10 Kent Ridge Crescent, Singapore 119280, Singapore Faculty of Medicine, Department of Orthopaedic Surgery, National University of Singapore, 10 Kent Ridge Crescent, Singapore 119280, Singapore Received 13 July 2005; accepted 18 August 2005 Available online 19 September 2005

Abstract The structure and architecture of scaffolds are crucial factors in scaffold-based tissue engineering as they affect the functionality of the tissue engineered constructs and the eventual application in health care. Therefore, effective scaffold assessment techniques are required right at the initial stages of research and development so as to select or design scaffolds with suitable properties. Various techniques have been developed in evaluating these important features and the outcome of the assessment is the eventual improvement on the subsequent design of the scaffold. An effective evaluation approach should be fast, accurate and non-destructive, while providing a comprehensive overview of the various morphological and architectural characteristics. Current assessment techniques would include theoretical calculation, scanning electron microscopy (SEM), mercury and ow porosimetry, gas pycnometry, gas adsorption and micro computed tomography (CT). Micro CT is a more recent method of examining the characteristics of scaffolds and this review aims to highlight this current approach while comparing it with other techniques. r 2005 Elsevier Ltd. All rights reserved.
Keywords: Scaffold characterization; Micro CT

Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . 2. Architectural and structural parameters . 3. Theoretical method and SEM analysis . . 4. Mercury porosimetry . . . . . . . . . . . . . . 5. Gas pycnometry . . . . . . . . . . . . . . . . . 6. Gas adsorption . . . . . . . . . . . . . . . . . . 7. Flow porosimetry . . . . . . . . . . . . . . . . 8. Micro CT . . . . . . . . . . . . . . . . . . . . . . 9. A micro CT study . . . . . . . . . . . . . . . . 10. Conclusion . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1363 1363 1364 1367 1368 1368 1369 1370 1371 1373 1374

Corresponding author. Faculty of Engineering, Division of Bioengineering, National University of Singapore, 10 Kent Ridge Crescent, Singapore 119280, Singapore. Tel.: +65 94309696. E-mail addresses: biedwh@nus.edu.sg, g0201956@nus.edu.sg (D.W. Hutmacher).

0142-9612/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2005.08.035

ARTICLE IN PRESS
S.T. Ho, D.W. Hutmacher / Biomaterials 27 (2006) 13621376 1363

1. Introduction In scaffold-based tissue engineering, scaffold architectural features are studied after the design and fabrication process; hence, allowing an assessment of the feasibility and the precision level of the process. These architectural characteristics would include porosity, pore size, surface area to volume ratio, interconnectivity, anisotrophy, strut thickness, cross sectional area and permeability [1]. The mechanical strength and biological functionality of the scaffold are inuenced by these characteristics. Various techniques can be used to evaluate them and they would include theoretical assessment, scanning electron microscopy (SEM) analysis, ow and mercury porosimetry, gas pycnometry and adsorption. Researchers have only recently employed micro computed tomography (CT) in the study of scaffolds. The authors research group has performed theoretical calculations, SEM, mercury porosimetry and micro CT evaluation of scaffolds fabricated via various techniques [25]. Gas adsorption and ow porosimetry have potential applications in the assessment of scaffold architecture; however, there is yet to be any report of their usage in the current literature. As one goes about selecting a suitable technique in characterizing scaffolds, the associated virtues and pitfalls of each technique should be scrutinized. Sometimes a combination of techniques is required so as to achieve an in depth study of the scaffold properties. However, the most attractive option is a single technique which is nondestructive, yet capable of providing a comprehensive set of data. It appears that micro CT can potentially fulll this role. When it was rst developed, micro CT was used extensively in the study of the trabecular structure [6] and important structural data can be derived after scanning the specimens, but it is not without its limitations. This review compares micro CT and other scaffold evaluation techniques and the nal section dwells on a micro CT study that was conducted on novel scaffolds. 2. Architectural and structural parameters Molecular transport in tissue engineered constructs is dependent on vasculature growth and diffusion. At the implantation sites, vasculature provides the main mode of transport, while for in vitro culturing, diffusion would be the sole approach. In order to ensure the survivability of the cells, the pore network has to be optimized so as to facilitate molecular transport [7]. Molecular transport would include the exchange of oxygen, nutrient, metabolic wastes and molecular signaling. These biochemical exchanges are essential for cell migration and proliferation. When molecular transport is hampered due to poor diffusion, cellscaffold constructs exhibit peripheral cellular growth while the interior of the construct undergoes necrosis [8]. In the light of this, scaffolds are evaluated against a series of properties (Table 1), so as to assess the ease of diffusion within the scaffold. Moreover, researchers

are able to gauge the mechanical properties and cellular response of the scaffold from these characteristics. One of the key properties that is widely discussed in the literature is porosity [3,9,10]. Porosity would determine cell seeding efciency, diffusion and the mechanical strength of the scaffold. High porosity and high surface area to volume ratio are required for uniform cell delivery, cellular attachment and neo-tissue in growth [11,12]. In a landmark study by Langer et al., a porosity of 90% was recommended for optimum diffusive transport within a cellscaffold construct under in vitro conditions [13]. However, scaffolds with such a high porosity would possess very low mechanical strength. In creating scaffolds for load bearing tissues such as bone and cartilage, mechanical integrity is a concern, hence there must be a compromise between the porosity and mechanical properties of the scaffold [14]. Studies on scaffold design have revealed that besides porosity, other factors such as pore interconnectivity and permeability affect molecular transport. A highly porous scaffold may have non-interconnected pores, thus lowering the diffusion efciency. Suh et al. cultured chondrocytes on scaffolds of equal porosity but of varying degrees of interconnectivity. He observed superior cell attachment and proliferation in scaffolds with highly interconnected pores [15]. However, one should be mindful that decient molecular transport may occur in the interconnected pores due to poor permeability. The scaffold permeability describes the size of the pore interconnections or fenestrations and this inuences the ease at which a uid ows through the pore network. Besides fenestrations, permeability is also dependent on porosity, interconnectivity, pore orientation, pore size and distribution [16]. A permeation experiment can be conducted so as to measure the permeability of the scaffold and permeability is calculated using Darcys law [17,18]. In this investigation, the rate of uid ow is measured at a known hydrostatic pressure. When viscous uid ow is assumed, Eq. (1) is used: PC mFL=hA, (1)

where F is the ow rate, L the linear dimension of the scaffold in the ow direction, h the pressure drop across scaffold, A the cross-sectional area of scaffold and m the uid viscosity. Besides permeation measurements, Scheidegger noted that permeability could be derived from known porosity, tortuosity and average specic surface area (Eq. (2)). In his calculations, he simplied the pore network to a series of parallel channels. Tortuosity is dened as the ratio of the length of the twist ow path connecting 2 points to the straight-line distance between the points. Average specic surface area is dened as the ratio of the pore area to the scaffold volume [19]: k P3 =T 2 S2 , (2)

where P is the porosity, T the tortuosity and S the average specic surface area.

ARTICLE IN PRESS
1364 S.T. Ho, D.W. Hutmacher / Biomaterials 27 (2006) 13621376 Table 1 Important architectural and structural characteristics of scaffolds Characteristics 1. 2. 3. 4. 5. 6. 7. 8. Porosity Surface area to volume ratio Interconnectivity Pore size Strut/wall thickness Anisotropy Cross-section area Permeability Formulae/elaboration Porosity 100% volume of pores/sum of volume of pores and scaffold material Surface area to volume ratio surface area of scaffold struts/volume of scaffold material Interconnectivity 100% volume of interconnected pores/volume sum of interconnected and closed pores Average diameter of the pores Average diameter of scaffold struts or average thickness of the scaffold walls A measure of the non uniformity in the alignment of the scaffold struts A measure of the area in a specied sectional plane of the scaffold A measure of the ease at which uid passes through the scaffold pores

Other parameters which are used to assess the functionality of the scaffold would include surface area and pore size. A large surface to volume ratio would assist cellular adhesion. Moreover, pore sizes should be sufciently large so as to encourage cellular in growth. Depending on the tissue type (soft or hard tissues), some of the tissue engineered constructs must possess sufcient mechanical strength so as to withstand loading, thus strut/wall thickness, anisotropy and the cross sectional area of the scaffold would be of interest. In the creation of scaffolds, 2 methodologies are observed. Researchers would determine the design of the scaffolds rst and a selection of a fabrication technique which satises the requirements follows. In the second approach, a fabrication method is chosen and from that platform, scaffold designs are optimized. The determination of architecture properties is crucial in both methodologies, thus it is important to have an understanding of the efcacies associated with each scaffold evaluation method so as to select the suitable assessment technique. It should be noted that the only some of the parameters are of emphasis and this would usually depend on the scaffold design. Scaffold design can be divided into two broad categories. The rst design incorporates a precise geometrical layout and scaffolds that fall into this category include regular honeycombed scaffolds (Fig. 1A) and woven textile meshes (Fig. 1B). Honeycombed scaffolds are fabricated using fused deposition modeling (FDM) [3], 3D printing [20] and stereolithography [21]. Woven Textile meshes are fabricated via precise weaving techniques. In this highly regulated design, scaffold evaluation would focus on porosity, interconnectivity, surface area to volume ratio and cross-sectional area, while the evaluation of the other characteristics would serve as a study on the precision level of the fabrication process. The second category of scaffold design involves the formation and deposition of scaffold struts or walls in a non-precise manner. Scaffolds which belong to this category include foams (Fig. 1C) and nanober meshes (Fig. 1D). The scaffold walls in foams are deposited via porogen decomposition [1], gas forming [22] and salt leaching [23]. Nanober meshes are fabricated using

electrospinning [24] and phase separation [25] methods. Since consistency and control are limited, numerous properties inclusive of those in Table 1 would be of interest. 3. Theoretical method and SEM analysis The theoretical method is commonly used in the assessment of scaffolds and there are various ways in doing so but most of these approaches are only capable of estimating scaffold porosity. The 2 main theoretical approaches are unit cube analysis and mass technique. In the unit cube method (Eq. (3)), the volume sum of the scaffold material and the associated pore spaces is calculated by taking linear measurements of the scaffold cube. The volume of the scaffold material is calculated from the known deposition pattern (Eq. (3)). This approach is commonly adopted for regular honeycombed scaffolds, fabricated via rapid prototyping [3,26]. The calculation assumes that the scaffold struts are uniform and the layers that they form do not fuse into each other. In this assumption, strut diameter and layer spacing are considered to be equal and consistent [3,26]. However, when scaffolds are fabricated via extrusion techniques like in the case of FDM [3] and 3D deposition [26], the ow dynamics of the extruded semi-solid struts would be of concern. Firstly, the struts are of inconsistent diameter, and fusion occurs between the semi-solid bers, these features affect the accuracy of the calculated porosity (Fig. 2). It should be noted that Eq. (3) needs to be modied when scaffolds of non 901 deposition angles are encountered: Porosity 1 V f =V A 100%,
2

(3)

where Vf is the scaffold material volume ( PLd n1n2/4), VA the apparent scaffold cube volume ( Lwh), d the strut diameter, L the strut length, w the scaffold width, h the scaffold height, n1 the number of struts per layer and n2 the number of layers per scaffold. In the mass technique, the volume of the scaffold material is derived by dividing the mass of the scaffold cube by its material density. The apparent scaffold cube volume is calculated from its linear measurements

ARTICLE IN PRESS
S.T. Ho, D.W. Hutmacher / Biomaterials 27 (2006) 13621376 1365

Fig. 1. SEM pictures of the different types of matrices. Polycaprolactonetri calcium phosphate (80%:20%) regular honeycombed scaffold (A) with a laydown pattern of 601/1201/1801 is created using FDM. The open arrow indicates a scaffold strut. A PLGA woven textile mesh is shown in (B). Using salt leaching, a polycaprolactone foam is fabricated (C). The close arrow indicates a scaffold wall. Electrospun nanobers of polycaprolactone collagen blend (75%:25%) is shown in (D). The magnications of (A), (B), (C) and (D) are 20, 20, 100 and 2000 times, respectively.

and scaffold porosity is derived using Eq. (4) [26]. This method is applicable for scaffolds with controlled and uncontrolled geometries, but it is dependent on the accuracy of the linear measurements (L, w and h) of the cube. Hence edge effects introduced during the process of excising a scaffold cube would be a concern as rough edges are created and inaccurate linear dimensions are taken: Porosity 1 V g =V A 100%, (4)

Porosity Dp DApp =DP , Dp density of polymer 1=f1 X C =DA X C =DC g, 5 where XC is the degree of crystallinity as determined via differential scanning calorimetry, DA the density of amorphous polymer, DC density of crystalline polymer and DApp the apparent density ( scaffold mass/apparent scaffold cube volume). Besides the unit cube analysis and mass technique, there are other theoretical calculations that are based on similar concepts but with slight variations, these would include the Archimedes method [16,27] and the liquid displacement technique [28,29]. In the Archimedes method (Fig. 3), the dry mass of the scaffold is recorded. Following that, the scaffold is prewet with ethanol and soaked in water under negative pressure. When the scaffold is submerged under water, the submerged mass is recorded. It is then taken out of the water and its wet mass is measured, and the measured values are used to compute the porosity value in Eq. (6): Porosity M wet M dry =M wet M submerged , (6)

where Vg is the volume of scaffold material ( mass of scaffold/density of scaffold material), VA the apparent scaffold cube volume ( Lwh). For polymer scaffolds, the polymer density is dependent on the amorphous and crystalline states, thus Ma et al. has taken this into consideration and has modied the mass technique (Eq. (5)). The density of the polymer is calculated by using the degree of crystallinity, densities of the amorphous and crystalline polymer. The calculation which Ma et al. uses, requires the measurement of the linear dimensions and the mass of the scaffold so as to determine the apparent density as shown in Eq. (5):

ARTICLE IN PRESS
1366 S.T. Ho, D.W. Hutmacher / Biomaterials 27 (2006) 13621376

where Mwet is the wet mass, Mdry the dry mass and Msub the submerged mass. The liquid displacement method (Fig. 4) uses an ethanol bath of known volume. The scaffold is rst placed in the bath and the entrapped air is removed under vacuum. The

volume sum of the ethanol and the scaffold is measured. The scaffold is removed from this bath and the remaining volume of ethanol in the bath is measured again. Using Eq. (7), the scaffold porosity is derived [28,29]. The Archimedes method may not be appropriate for hydrophobic polymer scaffolds as water does not penetrate easily into the pores moreover water induces shrinkage or swelling for the polymeric scaffolds [29]. The liquid displacement technique resolves this issue by using ethanol. Porosity V 1 V 3 =V 2 V 3 , (7)

Fig. 2. SEM pictures of scaffolds created via FDM: (A) shows inconsistent ber diameter and layer fusion, while (B) reveals edge effects introduced during the sectioning. The pressure applied during sectioning results in compression thus merging the struts at the edge (circled).

where V1 is the an initial known volume of ethanol, V2 the volume sum of ethanol and submerged scaffold and V3 the remaining volume of ethanol in the bath after scaffold removal. SEM analysis complements these theoretical calculations of porosity [3,30,31] and it allows direct measurements of pore size and strut/wall thickness. A visual estimation of interconnectivity, cross-section area and anisotropy can also be achieved. However to examine the scaffold interior, physical sectioning is needed and this would introduced unnecessary compression and edge effects to the scaffold architecture, thereby compromising the results. Owing to layer fusion, inconsistent strut diameter and complex architectures, surface area to volume ratio cannot be determined quantitatively via the theoretical and SEM approach. In view of the various biomedical applications and requirements, different scaffolds architectures have been experimented with. Besides FDM and 3D deposition scaffolds, investigators have also studied textiles [32], nanober meshes [33] and foams [10]. Textiles have regular geometric layout while nanober meshes and foams have a lack of it. As textiles and nanober meshes are exible and at, it is difcult to measure the dimensions of a unit cube, hence theoretical calculations would only yield an approximation for the porosity value. On the other hand when evaluating the porosity of rigid foams, the mass technique would sufce. SEM is capable of providing a qualitative analysis of these 3 types of scaffold architectures, however one should be mindful that only the supercial regions are assessed.

Dry mass

Wet mass

Submerged mass

Mdry

Mwet

Msub

Scaffold material
Fig. 3. Archimedes method.

Water

ARTICLE IN PRESS
S.T. Ho, D.W. Hutmacher / Biomaterials 27 (2006) 13621376 1367

V1

V2

V3

Scaffold material

Ethanol

Fig. 4. Liquid displacement method.

4. Mercury porosimetry Mercury porosimetry is a well known and established method which is often used to study porous materials. This technique employs mercury which is a non-wetting liquid that does not intrude into pore spaces except under sufcient pressure [34]. In 1921, Washburn related this pressure to the size of the intruded pore space which is shown in the Washburn equation as follows [35]: DP 4g cos y, where D is the pore diameter, g the surface tension of mercury, 480 dyne cm1, P the applied pressure and y the contact angle between pore wall and mercury, 1401. In this technique, the sample is introduced into a penetrometer and is subjected to gas evacuation while allowing mercury in ow. Ambient pressure is applied and the sample is enveloped by mercury. From the mercury volume measurements, the bulk volume of the sample is derived (Fig. 5A). Bulk volume is equivalent to apparent volume and it consists of scaffold material and pore spaces. Maximum pressure is attained through incremental steps so as to promote mercury intrusion. At maximum pressure, the total volume of the intruded mercury is measured, and this allows the derivation of total open pore volume and porosity (Fig. 5C) [35]. As mercury intrusion occurs, the cumulative volume of the intruded mercury at each pressure step is recorded in an intrusion curve. Using the Washburn equation, pore sizes and pore volume distribution by pore size are calculated. After reaching maximum pressure, incremental pressure reduction occurs and an extrusion prole is obtained. Hysteresis loops are noted as the intrusion and extrusion curves usually do not follow the same path. Pore cavity to pore throat ratio and the distribution of pore cavity sizes associated with pore throat sizes can be calculated from these loops [35]. From the intrusion and extrusion curves, parameters such as cumulative surface area to pore size, pore tortuosity, sample compressibility, permeability and densities are derived [35]. However, this technique does not inform the user of the scaffolds interconnectivity, strut/wall thickness, anisotropy and cross-sectional area.

(A)
2 1

(B)
2 1

Mercury

Mercury

(C)
2 1

Mercury

Fig. 5. Mercury intrusion with incremental increase in mercury pressure. At ambient pressure, the sample is enveloped by mercury (A). This allows the derivation of apparent scaffold volume. As mercury pressure increases, the large pores (1) are lled with mercury rst (B). Pore sizes and pore volume distribution by pore size are calculated as the mercury pressure increases. At higher pressures (C), mercury intrudes into the ne pores (2) and when the pressure reaches a maximum, total open pore volume and porosity are calculated. The closed pores (3) are neglected in this assay as they are not intruded with mercury.

Scaffolds with pore sizes ranging from 0.0018 to 400 mm can be studied using mercury porosimetry [3537]. Pore sizes smaller than 0.0018 mm are not intruded with mercury and this is a source of error for porosity calculations [35]. Furthermore, mercury porosimetry does not account for closed pores (Fig. 6) as mercury does not intrude into closed pores [36]. As the porosimeter is capable of applying pressures up to 414 MPa, sample collapse and compression is a possibility, hence the entire scaffold architecture is disrupted, resulting in erroneous measurements. Hence, mercury porosimetry is not suitable for fragile compressible scaffolds [36] such as exible foams (with porosities higher than 90%), textiles and nanober sheets. As mercury reacts with metals such as gold and aluminum to form amalgams, scaffolds containing these metals have to be evaluated with other techniques [36]. Other concerns

ARTICLE IN PRESS
1368 S.T. Ho, D.W. Hutmacher / Biomaterials 27 (2006) 13621376

would include the fact that the Washburn equation assumes prefect cylindrical pores, which is not the case in reality. After the mercury intrusion test, sample decontamination at specialized facilities is required as the highly toxic mercury is trapped within the scaffolds. Therefore this dangerous and destructive assay can only be conducted in well-equipped labs.

5. Gas pycnometry In porosity studies, there is a need to measure the scaffold material volume and gas pycnometry is capable of doing so [3,4]. This technique operates by detecting the change in pressure when the scaffold specimen displaces gas. A specimen of unknown volume Vx is placed in the sample chamber of known volume Vc (Fig. 7). Upon sealing, the chamber pressure Pc is measured. A reference chamber of volume Vr with a pressure Pr is separated from the sample chamber by a closed valve and Pr is higher than Pc. Both systems are allowed to equilibrate to a pressure of PE when the valve is opened. Using gas law, Vx is determined from Eq. (8) [38]. Once the scaffold material volume is determined, the derivation of scaffold porosity using the unit cube approach follows (Eq. (3)): V x PE V c PE V r Pc V c Pr V r =PE Pc , (8)

The accuracy of this volumetric measurement is dependent on the absence of moisture and volatile substances, as they would result in partial pressures and system instability. Hence, the analysis gas is either a pure gas or dry air and pretreatment in a vacuum oven for volatile substance removal is a prerequisite for samples [38]. Helium is a common analysis gas because of its inert nature and it has the smallest molecules that are capable of occupying all the open pores [39]. There are 2 sources of error in this volumetric measurement, rstly it does not account for the closed pores which the gas molecules cannot enter. Secondly, the reliability of this method is hampered by the difculty in taking accurate linear dimensions (L, w and h) of the unit scaffold cube. This is a concern for exible, at textiles and nanober scaffolds. Furthermore gas pycnometry does not inform the user of the other scaffold parameters such as surface area to volume ratio, interconnectivity, permeability and pore size. Thus it has to be coupled with other techniques such as micro CT in order to achieve a more conclusive assessment. 6. Gas adsorption In solids, electrical forces of attraction hold down the atoms in their xed positions. As the outermost atoms have lesser neighbors than those beneath them, an imbalance of attractive forces occurs [40]. To counter this, surface atoms attract surrounding gas molecules via Van der Waals and electrical forces [41]. This physical phenomenon is known as gas adsorption and it can be applied in the study of porous materials. The gas adsorption procedure begins with the placement of the specimen in an evacuated chamber where a small amount of absorbate gas is introduced. Absorbates would include nitrogen, benzene vapor, argon and krypton [42,43]. Adsorption isotherms are derived from the pressure and volume measurements of the chamber. Absorbate molecules form an initial thin layer on the available surfaces (Fig. 8B). At this stage, the surface area can be calculated using Brunauer, Emmett and Teller (BET) theory [44] and the Langmuir model [4547]. As gas adsorption continues, multi adsorption layers form and capillary condensation occurs. At this stage, pore sizes can be derived using the Barnett, Joyner and Halenda (BJH) method. When the pores are completely saturated with absorbate molecules (Fig. 8D), the total pore volume of the scaffold is calculated. After that, the gas is withdrawn and

where Vx is the specimen volume or scaffold ber volume, Vc the Chamber volume, Pc the initial chamber pressure, Vr the reference chamber volume, Pr the reference chamber pressure and PE the pressure at equilibrium.

B A

Fig. 6. Different kinds of pores within the scaffold. There are 2 broad categories of pores, closed pore (red) and open pores. Open pores can be subdivided into through pores (blue) and blind pores (green). (A) denotes pore cavity, while (B) denotes pore throat. Diagram is modied from [52].

Pc

Pr

PE

PE

Vc Vx

Vr

Vc Vx

Vr

Fig. 7. Schematic diagram of gas pycnometry.

ARTICLE IN PRESS
S.T. Ho, D.W. Hutmacher / Biomaterials 27 (2006) 13621376 1369

(A)

(B)

(C)

(D)

Fig. 8. Gas adsorption process. Gas molecules are attracted to the pore surfaces (A) and they will initially form a monolayer of molecules covering it (B). The formation of this monolayer allows the calculation of the surface area. As the adsorption process continues, multiple layer lling occurs (C) and pore sizes can be calculated at this stage. Total pore volume is obtained when the pores are completely with the molecules (D).

desorption isotherms are generated. Hysteresis is noted as adsorption and desorption isotherms fail to overlay completely, and from these isotherms and hysteresis loops, several key information are elucidated. This would include pore shapes, cumulative pore size distributions, pore volume and areas of mesopores and micropores. Mesopores are pores with diameters that measure 250 nm, while that of micropores is lesser than 2 nm [48]. This technique is relevant to the study of nano-featured and nano-modied scaffolds [49] as it is capable of examining the micro-porosity of scaffolds. Gas adsorption setups are capable of assessing scaffolds with pore sizes ranging from 0.35 to 400 nm [50] or 3.5 to 2000 mm [51]. Setups capable of small pore size measurements can be used to evaluate the architectural parameters of nanobers, dense foams and textiles. However, gas adsorption is not a suitable approach in studying scaffolds with low specic surface area, as erroneous surface area measurements would be encountered for values lesser than 0.01 m2 g1 [52]. The gas adsorption analysis does not measure scaffold material volume, hence it has to be complemented with the mass technique so that the surface area to volume ratio can be derived. Moreover, this technique does not evaluate other important architectural parameters such as interconnectivity, strut thickness, anisotropy, permeability and cross sectional area. Closed pores do not allow the entry of absorbate molecules, thus gas adsorption fails to account for the presence of these pores. Depending on the type of specimen and tests

required, the gas adsorption study would require between 2.5 h to 2 days [51,53]. This technique have been employed in petrochemical, ceramic and munitions industries [40]; however, there is yet to be a literature report of its application on biomedical scaffolds. 7. Flow porosimetry Flow porosimetry is a non-destructive approach that has been used to measure the pore sizes of porous materials such as ltration media and paper products [54,55]. Potential biomedical applications of this technique should not be ignored even though there are no publications in the biomaterials and tissue engineering elds of its use in examining scaffolds. In this technique, a fully wetted scaffold sample is sealed in a chamber and gas is allowed to pass through it. At the bubble point, there is sufcient gas pressure to overcome the uid capillary action in the largest pore [56]. The pressure is increased while the ow rate is measured till all the scaffold pores are empty and dry. Pore size distributions and mean pore size can be derived once the ow rate and the applied pressure are known [56]. Flow porosimetry is capable of measuring pore sizes within the range of 0.013500 mm [57]. Furthermore, compressive stresses can be applied onto the sample so as to study the effects of compression on pore size and distribution [58]. Flow porosimetry can be coupled with other scaffold evaluation techniques so as to determine the porosity and cross

ARTICLE IN PRESS
1370 S.T. Ho, D.W. Hutmacher / Biomaterials 27 (2006) 13621376

sectional area of the scaffold, moreover using Eq. (9), the calculation of the surface area of through pores can be accomplished (Fig. 6) [52]: QL=Pr a P3 =mkS 2 1 Pr 2 zpP2 =S1 P2prPave 1=2 , 9

where Q is the volume ow rate, L the thickness of porous medium, Pr the pressure drop across porous medium, a the cross sectional area, P the porosity, S the through pore surface area per unit volume of scaffold, m the viscosity of gas, r the density of gas at average pressure, Pave the mean pressure in porous medium and k and z are constants. Even though ow porosimetry is a relatively user friendly technique, it is unable to determine several important scaffold properties such as porosity, interconnectivity, strut thickness, anisotropy, permeability and cross sectional area. Since the scaffold material volume cannot be derived from ow porosimetry, surface area to volume ratio cannot be calculated. As the gas only passes through the through pores of the scaffold, properties of these pores are measured while that of closed and blind pores are omitted (Fig. 6). 8. Micro CT Feldkamp et al. pioneered micro CT when he developed an X-ray-based microtomographic system to analyze trabecular samples at a spatial resolution of 50 mm [59]. Since then, micro CT had been used extensively in the study of trabecular architecture [60] and there are increasing applications of it in other areas. Its popularity can be attributed to its ability to provide precise quantitative and qualitative information on the 3D morphology of the specimen. The interior of the specimen can be studied in great detail without resorting to physical sectioning nor using toxic chemicals. Moreover, after scanning, the intact samples can be subjected to other tests, therefore resolving the problem of sample scarcity. As researchers began to recognize the potential of this radiographic technique, various biomedical applications are being explored which would include the assessment of scaffolds, regenerated tissue [61] and vasculature networks [6264]. In micro CT scanning, the specimen is divided into a series of 2D slices which are irradiated from the edges with X-rays. Upon transversing through the slice, the X-rays are attenuated and the emergent X-rays with reduced intensities are captured by the detector array. From the detector measurements, the X-ray paths are calculated and the attenuation coefcients are derived. A 2D pixel map is created from these computations and each pixel is denoted by a threshold value which corresponds to the attenuation coefcient measured at a similar location within the specimen. As the attenuation coefcient correlates to the material density, the resultant 2D maps reveal the material phases within the specimen. The quality of the 2D maps is

dependent on the scanning resolution which ranges from 1 to 50 mm [59,65]. At high resolutions, intricate details are imaged, however more time is required for high resolution scanning and the resultant large data set poses a challenge for data storage and processing. 3D modeling programs such as Mimics (Materialize, Belgium), Velocity and Anatomics stack the 2D maps to create 3D models, without these programs, visualization and analysis would have been impossible. As computation is inherent in this technique, the selection of software and hardware facilities would inuence the efciency and effectiveness of this radiographical assessment. The recent use of micro CT in scaffold research enabled morphological studies to be carried out, yielding comprehensive data sets. Guldberg et al. conducted micro CT studies on porous poly (L-lactide-co-DL-lactide) scaffolds formed using porogens [1]. These scaffolds were not of precise geometrical layout and the structural parameters derived using micro CT were correlated to the percentage of porogen used. This in turn led to the optimization of the scaffold fabrication process. On the other hand, Wang et al. employed micro CT in the characterization of the internal geometry of poly-e-caprolactone (PCL) scaffolds that were fabricated via precision extruding deposition [66]. From the scan data, measurements of the scaffold material volume, surface area, strut width and pore sizes were taken. Using threshold inversion, the pore network was visualized and the pore interconnectivity was studied [66]. Mechanical tests are conducted so as to ascertain the mechanical properties of the scaffold, however most of these tests are destructive; hence, a non-destructive method is sought. Researchers have considered nite element modeling (FEM) as an alternative to mechanical testing as simulations can be carried out via computations. FEM requires the input of precise 3D structural and architectural information of scaffolds, which can be obtained from micro CT scans [67]. Similarly, diffusion patterns in scaffolds can also be investigated through uid ow studies which are simulated via FEM. Scaffolds with intricate interior structures can be scrutinized using micro CT, as any spatial location of the architecture can be digitally isolated out. This is crucial for scaffolds that exhibit different geometric layouts at different spatial locations. Micro CT possesses this key advantage over other techniques such as mercury and ow porosimetry. Within the digitally excised scaffold cube, scaffold material volume and surface area are measured, thus allowing the calculation of porosity and surface area to volume ratio. Three-dimensional imaging allows a close up view of any specic location, thus the observation of pore shape and the measurement of pore size and strut/wall thickness can be conducted in these close ups. Scaffold anisotropy is evaluated via algorithms [1] and the cross-sectional area can be measured from the 2D slice images. By inverting the threshold, a negative image is created which captures only the scaffold pores. By measuring the total and the interconnected pore volumes,

ARTICLE IN PRESS
S.T. Ho, D.W. Hutmacher / Biomaterials 27 (2006) 13621376 1371

interconnectivity is derived [66]. In studying scaffold permeability, a suitable visualization program needs to be selected. As micro CT employs penetrative X-rays, closed pores can be imaged. The exibility of micro CT analysis allows the evaluation of foams, textiles and nanober scaffolds. There are associated concerns despite of the numerous advantages of using micro CT. Image thresholding is a crucial step that has to be executed prior to 3D modeling and it affects the subsequent analysis and visualization [10,68]. In the conventional approach, the thresholding range is selected via histographics and visual estimation and the problem arises when the scaffold composes of multiple materials whose thresholding ranges overlap and this renders the digital separation of these materials a difcult task. Moreover, as polychromatic X-ray beams are used in micro CT, the lower energy rays would be readily attenuated by the sample resulting in a high exposure at the center of the scaffold. This effect is known as beam hardening and as a result thresholding is no longer dependent solely on radiodensity but also on the specimen size [61,69]. Micro CT analysis is not suitable for scaffolds containing metals as X-rays are heavily attenuated by these metals. The presence of metals results in dark and bright grainy artifacts which obscure important details in the scan images [70,71]. As micro CT is a relatively new technology, improved algorithms and setups are anticipated, thus resolving such imaging errors. 9. A micro CT study The potential of micro CT scanning is demonstrated in the following 2 applications. The rst application involves the evaluation of polymer scaffolds that were produced via rapid prototyping, while the second study evaluates the feasibility of using micro CT to analyze ne exible scaffold meshes. Two rapid prototyped scaffolds of different geometric layouts (Table 2) were scanned at a resolution of 35 mm using a Skyscan 1076 in vivo microtomograph. The scaffolds were made of a copolymer comprising of poly ethylene glycol (PEG), PCL and poly lactic acid (PLA). An averaging of 5 and a 1 mm aluminum lter were used

during scanning. The rotational step was 0.81 over an angle of 1801. The scan les were reconstructed at a step size of 4 using a modied Feldkamp algorithm as provided by Skyscan. The reconstructed data was loaded onto Mimics and visualized at a thresholding range of 793 to 83 Hounseld units (HU). The determinants of the scaffold architecture would include the selected nozzle diameter and ll gap. Strut thickness is affected by the nozzle diameter, while ll gap affects the distance between the adjacent struts of a common layer. This study seeks to correlate these parameters to the actual measurements taken in the fabricated scaffolds. The micro CT results indicated that the rapid prototyping approach was precise as the measured pore sizes matched that of the ll gap (Fig. 9). However, due to the ow dynamics of the extruded laments, strut thickness deviated from the nozzle diameter and strut fusion was observed (Fig. 9H). Porosity, surface area to volume ratio and interconnectivity were calculated in the micro CT analysis (Table 2). Hence this study demonstrated the versatility of micro CT in evaluating scaffolds as it is a single technique which is capable of characterizing the scaffolds in multiple aspects. Pycnometer measurements of porosity were recorded and these values were in general agreement with that of micro CT. Flexible, ne and sparse textile meshes pose a problem in scaffold evaluation. The porosity of compressible and sparse meshes cannot be derived via theoretical calculations, gas pycnometry and adsorption as all of them require a sectioned cube of specic dimensions, which is a difcult procedure for these meshes. Being compressible, erroneous results would be encountered if these meshes were subjected to mercury porosimetry. Using SEM, pore sizes and strut thickness can be measured but SEM only offers qualitative information on porosity and interconnectivity. To get around these limitations, micro CT was used and the sample in this second study was a poly D,L-lactide-coglycolide (PLGA) mesh that was woven out of 160 mm thick laments. The 2 layered, braided architecture of this scaffold was clearly observed at a resolution of 18 mm in Fig. 10. With digital analysis, porosity (94.83%), surface area to volume ratio (35.05 mm2 mm3), interconnectivity (100%) and the average cross-sectional area (0.57 mm2) were calculated. Therefore, this demonstrates the feasibility

Table 2 Characteristics of the 2 rapid prototyped scaffolds No Material Angle Nozzle size (mm) Fill gap (mm) Porositypycnometer (%) 72.0570.41 70.6070.67 Porositymimics (%) Surface area/volumemimics (mm2/mm3) Interconnectivity (%)

Copolymer of PEG, PCL and PLA Copolymer of PEG, PCL and PLA

01/901/ 1801 01/601/ 1201

0.5

1.5

74.99

8.65

100

0.5

1.5

75.45

9.07

100

ARTICLE IN PRESS
1372 S.T. Ho, D.W. Hutmacher / Biomaterials 27 (2006) 13621376

Fig. 9. (AD) are the 3D models of scaffold 1 (refer to Table 2), while (EH) belongs to scaffold 2. (A) and (E) are the isometric views, while (B) and (F) are top views. The measured pore sizes correlates well with the ll gap selected as shown in (C) and (G). (D) and (H) show the diameters of the struts and in the case of scaffold 2, a deviation from the nozzle diameter and strut fusion were observed.

ARTICLE IN PRESS
S.T. Ho, D.W. Hutmacher / Biomaterials 27 (2006) 13621376 1373

Fig. 10. A PLGA mesh shown in (A) was subjected to micro CT scanning. The 3D models are as shown in (BD). The isometric (B), side (C) and top (D) views revealed the braided interlocking architecture of the bilayer scaffold. Cross-sectional area was calculated as shown in (E). The rectangle encloses the individual laments of the scaffold.

of using micro CT in the study of ne compressible meshes, which cannot be analyzed using other techniques. 10. Conclusion In conclusion, the evaluation of scaffold architecture parameters is necessary as it facilitates the improvement of the scaffold design process and its eventual biomedical applications. Hence, an appropriate technique needs to be selected in analyzing the scaffold and an understanding of the potentials and concerns of the technique is crucial.

Table 3 summarizes each technique as discussed in this review. Micro CT is a rather new technique but it has demonstrated various key advantages. Being a computational method, numerous parameters can be calculated and this depends on the computational capability of the software and hardware. Furthermore, it is non-destructive, hence samples remain intact for further assays. However, it is not without its drawbacks that include beam hardening and the difculty of thresholding. Since it is a new technique, future advancements would be anticipated so as to address such problems.

ARTICLE IN PRESS
1374 S.T. Ho, D.W. Hutmacher / Biomaterials 27 (2006) 13621376 Table 3 A summary of the properties that can be derived (denoted by +) by the individual techniques and those that cannot be (denoted by ) Theoretical method Porosity Surface area/ volume Pore size Strut/wall thickness Anisotropy Cross section area Permeability Other parameters that can be derived + SEM Mercury porosimetry Gas pycnometry Gas adsorption Flow porosimetry + Micro CT

Qualitative

+ + Interconnectivity

+ Nil

+ Qualitative +

+ + + + + + Dependent on software  Observation of pore shape and morphology.

Nil

+ + + Qualitative Qualitative Nil

+ +

 Pore volume vs.     


pore size. Pore cavity to pore throat ratio. Pore cavity sizes vs. pore throat sizes. Cumulative surface area vs pore size. Pore tortuosity. Sample compressibility

 Pore shapes  Cumulative 


pore size distributions. Pore volume and areas of mesopores and micropores

 Effects of
compressive stresses can be studied. Pore size distributions

Other drawbacks

 Not precise
due to the inaccurate measurements of the linear dimensions of the scaffold cube.

 Physical  
sectioning required. Destructive assay. Only supercial regions are analyzed.

 Does not account    


for closed pores. Sample compression. Mercury reacts with metals. Assumption of prefect cylindrical pores. Toxic and destructive assay.

 Not precise
due to inaccurate measurements of the linear dimensions of the scaffold cube. Does not account for closed pores.

 Erroneous
measurements are encountered for low surface area samples. Can be time consuming. Does not account for closed pores.

 Porosity and
cross sectional area values are prerequisites for the calculation of the surface area of through pores. Does not account for closed and blind pores.

 Thresholding
difculty

 Beam
hardening

 

Scaffolds with regular geometry Foams Textile scaffolds Nanober scaffolds

Qualitative

Y N N

Qualitative Qualitative Qualitative

Y N N

Y N N

Y Y Y

Y Y Y

Y Y Y

Other useful characteristics and drawbacks are also shown. The feasibility of applying these techniques on foams, textile and nanober scaffolds are as indicated. Y means suitable while N means not suitable.

References
[1] Lin ASP, Barrows TH, Cartmell SH, Guldberg RE. Microarchitecture and mechanical characterization of oriented porous polymer scaffolds. Biomaterials 2003;24:4819.

[2] Schantz JT, Teoh SH, Lim TC, Endres M, Lam CXF, Hutmacher DW. Repair of calvarial defects with customized tissue-engineered bone grafts I. Evaluation of osteogenesis in a three-dimensional culture system. Tissue Eng 2003;9(S1):q 11326.

ARTICLE IN PRESS
S.T. Ho, D.W. Hutmacher / Biomaterials 27 (2006) 13621376 [3] Zein I, Hutmacher DW, Tan KC, Teoh SH. Fused deposition modeling of novel scaffold architectures for tissue engineering applications. Biomaterials 2002;23:116985. [4] Lam CXF, Teoh SH, Hutmacher DW. In vitro degradation studies of customized PCL scaffolds fabricated via FDM. International Conference on Biological and Medical Engineering, 2002. [5] Ang TH, Sultana FSA, Hutmacher DW, Wong YS, Fuh JYH, Mo XM, et al. Fabrication of 3D chitosan-hydroxyapatite scaffolds using a robotic dispensing system. Mater Sci Eng 2002;20: 3542. [6] Muller R, Hildebrand T, Ruegsegger P. Non-invasive bone biopsy: a new method to analyse and display the three-dimensional structure of trabecular bone. Phys Med Biol 1994;39:14564. [7] Karande TS, Ong JL, Agrawal CM. Diffusion in musculoskeletal tissue engineering scaffolds: design issues related to porosity, permeability, architecture, and nutrient mixing. Ann Biomed Eng 2004;32(12):172843. [8] Ishaug-Riley S. Bone formation by three-dimensional stromal osteoblast culture in biodegradable polymer scaffolds. J Biomed Mater Res 1997;36(1):1728. [9] Schantz JT, Hutmacher DW, Chim H, Ng KW, Lim TC, Teoh SH. Induction of ectopic bone formation by using human periosteal cells in combination with a novel scaffold technology. Cell Transplant 2002;11:12538. [10] Meinel L, Karageorgiou V, Fajardo R, Snyder B, Shinde-Patil V, Zichner L, et al. Bone tissue engineering using human mesenchymal stem cells: effects of scaffold material and medium ow. Ann Biomed Eng 2004;32(1):11222. [11] Kim BS, Mooney DJ. Development of biocompatible synthetic extracellular matrices for tissue engineering. Trends Biotechnol 1998;16(5):22430. [12] Mooney DJ, McNamara K, Hern D, Vacanti JP, Langer R. Stabilized polyglycolic acid bre-based tubes for tissue engineering. Biomaterials 1996;17(2):11524. [13] Freed LE, Vunjak-Novakovic G, Biron RJ, Eagles DB, Lesnoy DC, Barlow SK, et al. Biodegradable polymer scaffolds for tissue engineering. Biotechnology (NY) 1994;12(7):68993. [14] Hollister SJ, Maddox RD, Taboas JM. Optimal design and fabrication of scaffolds to mimic tissue properties and satisfy biological constraints. Biomaterials 2002;23(20):4095103. [15] Suh SW, Shin JY, Kim J, Beak CH, Kim DI, Kim H, et al. Effects of different particles on cell proliferation in polymer scaffolds using a solventcasting and particulate leaching technique. Am Soc Artif Int Organs J 2002;48:4604. [16] Ma PX, Zhang R. Microtubular architecture of biodegradable polymer scaffolds. J Biomed Mater Res 2001;56:46977. [17] Agrawal CM, McKinney JS, Huang D, Athanasiou KA. The use of the vibrating particle technique to fabricate highly porous and permeable biodegradable scaffolds. In: Agrawal CM, Parr JE, Lin ST, editors. Synthetic bioabsorbable polymers for implants, ASTM STP 1396. West Conshohocken, PA: American Society for Testing and Materials; 2000. p. 99114. [18] Agrawal CM, Mckinney JS, Lanctot D, Athanasiou KA. Effects of uid ow on the in vitro degradation kinetics of biodegradable scaffolds for tissue engineering. Biomaterials 2000;21(23): 244352. [19] Scheidegger AE. The physics of ow through porous media. New York: Macmillan; 1957. p. 236. [20] Landers R, Hubner U, Schmelzeisen R, Mulhaupt R. Rapid prototyping of scaffolds derived from thermoreversible hydrogels and tailored for applications in tissue engineering. Biomaterials 2002;23:443747. [21] Sachlos E, Reis N, Ainsley C, Derby B, Czernuszka JT. Novel collagen scaffolds with predened internal morphology made by solid freeform fabrication. Biomaterials 2003;24:148797. [22] Shea LD, Wang D, Franceschi RT, Mooney DJ. Engineered bone development from a pre-osteoblast cell line on three-dimensional scaffolds. Tissue Eng 2000;6(6):60517. 1375 [23] Laurencin CT, Ambrosio AMA, Borden MD, Cooper JA. Tissue engineering: orthopedic applications. Ann Rev Biomed Eng 1999;1:1946. [24] Boland ED, Matthews JA, Pawlowski KJ, Simpson DG, Wnek GE, Bowlin GL. Electrospinning collagen and elastin: preliminary vascular tissue engineering. Frontiers Biosci 2004;1(9):142232. [25] Ma PX, Zhang R. Synthetic nano-scale brous extracellular matrix. J Biomed Mater Res 1999;46(1):6072. [26] Woodeld TBF, Malda J, Wijin JD, Peters F, Riesle J, Blitterswijk CAV. Design of porous scaffolds for cartilage tissue engineering using a three-dimensional ber-deposition technique. Biomaterials 2004;25(18):414961. [27] Chu TMG, Orton DG, Hollister SJ, Feinberg SE, Halloran JW. Mechanical and in vivo performance of hydroxyapatite implants with controlled architectures. Biomaterials 2002;23:128393. [28] Ramay HR, Zhang M. Preparation of porous hydroxyapatite scaffolds by combination of the gel-casting and polymer sponge methods. Biomaterials 2003;24:3293302. [29] Zhang R, Ma PX. Poly(a-hydroxyl acids)/hydroxyapatite porous composites for bone-tissue engineering. I. Preparation and morphology. J Biomed Mater Res 1999;44(4):44655. [30] Kellomaki M, Niiranen H, Puumanen K, Ashammakhi N, Waris T, Tormala P. Bioabsorbable scaffolds for guided bone regeneration and generation. Biomaterials 2000;21:2495505. [31] Walsh D, Furuzono T, Tanaka J. Preparation of porous composite implant materials by in situ polymerization of porous apatite containing e-caprolactone or methyl methacrylate. Biomaterials 2001;22:120512. [32] Principles of mercury porosimetry and gas sorption. www. quantachrome.com [33] Yoshimoto H, Shin YM, Terai H, Vacanti JP. A biodegradable nanober scaffold by electrospinning and its potential for bone tissue engineering. Biomaterials 2003;24:207782. [34] Principles of mercury porosimetry and gas sorption. www. quantachrome.com [35] Webb PA. An introduction to the physical characterization of materials by mercury intrusion porosimetry with emphasis on reduction and presentation of experimental data. Micromertics Instrument Corp., 2001. [36] Higgins C. New thermo style. www.pragolab.cz [37] www.thermoquest.com. [38] Paul AW. Volume and density determinations for particle technologists. Micromeritics Instrument Corp., 2001. p. 116. [39] Bomberg M, Plagge R. Joint In situ protocols. p. 18. [40] Autosorb-1 series. www.quantachrome.com [41] What is adsorption? http://ias.vub.ac.be/Genneral/Adsorption.html #anchor848631 [42] Adsorption in mesopores. Marczewski A. http://isotherm.adsorption. org/ [43] Gas adsorption. http://www.beckman.com/products/instrument/ partchar/technology/gasadsorption.asp [44] Brunauer, Emmett, Teller. J Am Chem Soc 1938;60:309. [45] Langmuir I. J Am Chem Soc 1918;40:1361. [46] Langmuir I. J Am Chem Soc 1932;54:2798. [47] Langmuir I. Nobel lecture. 1932. [48] Adsorption glossary. Marczewski A. http://isotherm.adsorption.org/ [49] Ma PX. Scaffolds for tissue fabrication. Mater Today 2004;7(5): 3040. [50] Autosorb-6B. www.quantachrome.com [51] BET liquisorb. http://www.pmiapp.com.com/products/bet_liquisorb. html [52] Jena A, Gupta K. A novel technique for surface area and particle size determination o components of fuel cells and batteries. www.pmi.com [53] BET sorptometer. http://www.pmiapp.com/products/bet_sorptometer. html [54] Cyclic compression porometer product resources. http://www. pmiapp.com/

ARTICLE IN PRESS
1376 S.T. Ho, D.W. Hutmacher / Biomaterials 27 (2006) 13621376 [64] Ortiz MC, Garcia-Sanz A, Bentley MD, Fortepiani LA, GarciaEstan J, Ritman EL, et al. Microcomputed tomography of kidneys following chronic bile duct ligation. Kidney Int 2000;58:163240. [65] Weiss P, Obadia L, Magne D, Bourges X, Rau C, Weitkamp T, et al. Synchrotron X-ray microtomography (on a micro scale) provides three-dimensional imaging representation of bone ingrowth in calcium phosphate biomaterials. Biomaterials 2003;24:4591601. [66] Wang F, Shor L, Darling A, Khalil S, Sun W, Guceri S, et al. Precision extruding deposition and characterization of cellular poly e-caprolactone tissue scaffolds. Rapid Prototyping J 2004;10(1):429. [67] Jaecques SVN, Van Oosterwyck H, Muraru L, Van Cleynenbreugel T, De Smet E, Wevers M, et al. Individualised, micro CT-based nite element modeling as a tool for biomechanical analysis related to tissue engineering of bone. Biomaterials 2004;25:168396. [68] Duvall CL, Taylor WR, Weiss D, Guldberg RE. Quantitative microcomputed tomography analysis of collateral vessel development after ischemic injury. Am J Physiol Heart Circulatory Physiol 2004;287:H30210. [69] Bonse U, Busch F. X-ray computed microtomography (mCT) using synchrotron radiation (SR). Progress Biophys Molec Biol 1996;65(1/ 2):13369. [70] Wang G. Computed tomography principles. http://dolphin.radiology. uiowa.edu/ge/teaching.html [71] Wang G, Vannier MW. Computerized tomography. Encyclopedia Electr Electron Eng 1998:122. [55] In-plane porometer product resources. http://www.pmiapp.com/ products/in_plane_porometer.html [56] Clamp-on porometer product resources. http://www.pmiapp.com/ products/clamp_on_porometer.html [57] Bubble point tester resources. http://www.pmiapp.com/products/ bubble_point_tester.html [58] Compression porometer product resources.http://www.pmiapp.com/ products/compression_porometer.html [59] Feldkamp LA, Goldstein SA, Partt AM, Jesion G, Kleerekope M. The direct examination of three-dimensional bone architecture in vitro by computed tomography. J Bone Min Res 1989;4:311. [60] Ru egsegger P, Koller B, Mu ller R. A microtomographic system for the nondestructive evaluation of bone architecture. Calcif Tissue Int 1996;58:249. [61] Verna C, Bosch C, Dalstra M, Wikesjo UME, Trombelli L. Healing patterns in calvarial bone defects following guided bone regeneration in rats. J Clin Periodontol 2002;29:86570. [62] Bentley MD, Ortiz MC, Ritman EL, Romero JC. The use of microcomputed tomography to study microvasculature in small rodents. Am J PhysiolRegulatory, Integrative Comparative Physiol 2002;282:126779. [63] Jorgensen SM, Demirkaya O, Ritman EL. Three-dimensional imaging of vasculature and parenchyma in intact rodent organs with X-ray micro-CT. Am J Physiol Heart Circulatory Physiol 1998;44: H110314.

Potrebbero piacerti anche