Annals of Biomedical Engineering

, Volume 34, Issue 1, pp 89–101

Methods for Fabrication of Nanoscale Topography for Tissue Engineering Scaffolds


    • Department of Biomedical EngineeringBoston University
  • Tejal A. Desai
    • Department of Biomedical EngineeringBoston University
    • Department of Physiology and Division of BioengineeringUniversity of California

DOI: 10.1007/s10439-005-9005-4

Cite this article as:
Norman, J.J. & Desai, T.A. Ann Biomed Eng (2006) 34: 89. doi:10.1007/s10439-005-9005-4


Observations of how controlling the microenvironment of cell cultures can lead to changes in a variety of parameters has lead investigators to begin studying how the nanoenvironment of a culture can affects cells. Cells have many structures at the nanoscale such as filipodia and cytoskeletal and membrane proteins that interact with the environment surrounding them. By using techniques that can control the nanoenvironment presented to a cell, investigators are beginning to be able to mimic the nanoscale topographical features presented to cells by extracellular matrix proteins such as collagen, which has precise and repeating nanotopography. The belief is that these nanoscale surface features are important to creating more natural cell growth and function. A number of techniques are currently being used to create nanoscale topographies for cell scaffolding. These techniques fall into two main categories: techniques that create ordered topographies and those that create unordered topographies. Electron Beam lithography and photolithograpghy are two standard techniques for creating ordered features. Polymer demixing, phase separation, colloidal lithography and chemical etching are most typically used for creating unordered surface patterns. This review will give an overview of these techniques and cite observations from experiments carried out using them.


Electron beam lithographyPhotolithographyPolymer demixingColloidal lithographyPhase separationElectrospinningChemical etchingNanotechnologyNanoscale features


The goal of tissue engineering is to create a living construct that will mimic the complexities of human tissue function. Developments over the last decade have moved the field of tissue engineering from the simple cultures of cells to precisely engineered constructs that closely resemble native tissues in both appearance and function. Any engineered tissue generally has two main parts: The cells and the scaffold.75 The cells provide the biological functionality while the scaffold provides a platform for cell growth. A variety of fabrication techniques can be used to create complex two- and three-dimensional structures.19 In even the simplest of constructs, however, there are a host of cell–material interactions that will affect both the biological functionality as well as the overall appearance of the tissue.

The properties of the scaffold and how they affect the cell population has become a key area of study in understanding how to better build artificial tissues. One of the most widely studied areas over the last 10–15 years has been how topographical cues provided by the scaffold, to the cells, affect growth parameters such as cell adhesion, morphology, viability, apoptosis, genetic regulation, and motility, as well as numerous other parameters of interest to specific tissues and cell types.

A tissue engineer needs to consider how the cells and the scaffold material will interact in an engineered construct. It is now well known that scaffold features at the scales of individual cells (1–100 μm) can influence cell growth and function.19 While all of the mechanisms are still not completely understood, the observations have led to the engineering of scaffolds with micron-size features that individual cells can directly interact with. This has given researchers the ability to create scaffolds that organize cells into prescribed patterns. With proper design, these patterned scaffolds, in conjunction with the appropriate cells, can recreate the fine structure and functionality of native tissues.52

All size scales of a scaffold need to be considered when engineering an artificial tissue: The superstructure, the overall shape of the scaffold; the microstructure, the cellular level structure of the surface; and the nanostructure, the subcellular level structure of the surface. The first two levels of this hierarchy have been very well studied over the years and methods for controlling them are well established. The nanoscale structure, however, is less well understood and less controllable, yet just as important as the levels above it.

The extracellular matrix (ECM), the substratum in which cells live in vivo, has considerable topographic detail down to the nanometer-scale (e.g., the 66 nm repeat banding of collagen fibers). Cells are in intimate contact with the extracellular matrix forming adhesions to the fibers via integrins, transmembrane proteins that interact with specific amino acid sequences found within the proteins that make up the ECM (typically the RGD sequence in fibronectin57). Researchers are now looking to see if and how these domains of nanometer-size variations in topography of the ECM affect the cells growing on them. It has been very well established that micrometer-size features on a culture surface can affect parameters such as migration, adhesion, and morphology.63 Controlling the nanotopography of biomaterials will allow researchers to systematically study the effects of nanotopography on live cultures.

When cells are cultured outside the body on a polymeric scaffold, or any one of a variety of materials used as substrates for a biomedical device or implant, often, only the macro- and micro-scale features are considered. However, the nanoscale features of all of these materials may greatly affect, for better or worse, the culture in question. The inherent nanotopography encountered by the cells on such materials is most likely different from that found in their native environment and may provide physical cues leading to less than desirable growth and function. Gaining understanding and control of the nano-environment will give researchers another method to control cell cultures in order to recreate more physiologically accurate tissues.

Natural tissues have a hierarchical structure. All tissues have properties at the macro-scale (>1 mm), micro-scale (1 μm–1 mm), and nanoscale (<1 μm) that greatly affect their function. For example, skeletal muscle has features at each of these scales that must be present for proper contractility. At the macro-scale, skeletal muscle is composed of bundles of muscle cells. At the micro-scale each muscle cell contains a dense parallel arrangement of myofibrils. At the nanoscale, each of these myofibrils is comprised of interdigitating myofilaments. There are two types of myofilaments, thin filaments that are 5–8 nm in diameter and 1 μm in length and thick filaments that are 10 nm in diameter and 1.6 μm in length. The precise arrangement of the thick and thin filaments in relation to each other and the cell is the fundamental mechanism by which a single muscle cell contracts. The higher level arrangement of the myofibrils and ultimately the cells themselves is what allows for the proper contractility of a muscle group. Similar hierarchical structure can be found in nearly every tissue in the human body.

Individual cells also have a hierarchical structure. Typically cells have lengths or diameters in the tens of microns. When cells migrate, they extend psuedopods (filipodia) from the main body of the cell. Gustafson and Wolpert have made striking observations of psuedopods on the order of 500 nm extending from the body of a cell (sea urchin mesenchyme cells) to explore and probe the surrounding environment.32 This exploration by the cells appears to happen randomly with psuedopods sweeping the surface until a point of stable contact is made. Cell movement then occurs in the direction of the contact through retraction of the attached psuedopod. This apparent exploration of the surrounding by cells is what has led researchers to develop scaffolds that provide cues to the cells as they migrate across its surface. Finding the cues that will cause the cells to make a preferential adhesion to the surface has been one of the major endeavors for tissue engineers over the last decade. Chemical and physical cues are the two most widely studied modifications that are made to the surface of a material. One of the most difficult tasks is determining the role played by each of these treatments and finding the proper balance between the two types of treatments for optimal cell attachment and growth. Physical cues, artificially created on the surface of a material alone have been shown to significantly affect organization of cells in culture through the phenomenon of contact guidance.23 This has been demonstrated in both two-dimensional and three-dimensional cultures with artificial tissues being created that are strikingly similar in both appearance and function to natural tissues.22,53

Nanotechnology allows for the construction of devices that interact at the subcellular level. Application of nanotechnology to tissue engineering is a new and quickly expanding field. In this review, only the most common nanofabrication techniques used for tissue engineering will be addressed. Also, only those techniques that create physical nanoscale structures are covered. Engineering of surface chemistries for enhanced cell adhesion and proliferation, while a widely studied area, and well within the realm of nanotechnology, will not be addressed. More esoteric techniques will be left for brief discussion at the end of this review. Observations by various research groups using the described techniques will be highlighted. However, it must be noted that the number of techniques and biological systems being studied is highly varied. New techniques pushing the lower size limits of nanotechnology are continuously being developed and improved upon.


The success with using micro-scale surface features to study a variety of cellular phenomena has led researchers to want to study how nanoscale cellular extensions (e.g., filipodia) interact with their environment to effect cell growth, proliferation, and expression. This change in scale (several orders of magnitude lower) is motivated by the various nanoscale structures that comprise the extracellular matrix. For example, the basement membrane of the corneal epithelium of the rhesus macaque was found to be composed of a varying nanotopography consisting of elevations ranging from 76 to 379 nm and pores ranging from 22 to 216 nm with a mean interpore distance of 87 nm.1 This rich three-dimensional topography at the nanoscale causes an increase in surface area of the basement membrane of up to 400%. This increased area over which cell–surface interactions can take place may have a number of important consequences to regulating tissue growth. It is these types of questions that nanofabrication can play a role in answering through the controlled fabrication of model substrates that will allow for a systematic study of surface topographies and their effects on a variety of growth parameters.

The standard for fabricating micro-scale topographies has been photolithography performed in the near-UV. This process involves exposing a silicon wafer coated with a photoresist to a near-UV light source (typically 365 or 405 nm) through a mask that has the desired pattern on it. This mask selectively allows light through to the wafer thereby recreating the pattern in the photoresist. Development of the wafer brings out the desired features in the photoresist. However, due to the inherent lower limit of this method (a function of the wavelength of light used) it is incompatible with creating nanometer-size features. New methods for producing surfaces with nanometer topographies are needed to take the place of near-UV photolithography. There have been many different techniques employed to create nanoscale topographical features. Some of these, such as electron beam lithography (EBL), have been successfully employed in the micrometer-size domain for creating textured surfaces, while others are unique to creating nanometer-scale features. Several of the more common techniques are described further below.
Table 1.

Techniques for creating nanotopographies




Polymer demixing

Electron beam lithography

Phase separation




Colloidal lithography


Chemical etching



aElectrospinning can be use to create a simple ordered topography of aligned fiber bundles

The types of nanotopographical features that can be created on materials falls into two main categories: Unordered topographies and ordered topographies. Unordered topographies are typically those that spontaneously occur during processing. Examples of such topographies can be made using techniques such as polymer demixing, colloidal lithography, and chemical etching. Surfaces patterned in this manner tend to give features that are random in orientation and organization with imprecise or no control over feature geometry. These techniques, however, are usually simpler, quicker, and less costly than the more complex equipment and processes needed to create ordered topographies. Ordered topographies are those that can be created with techniques like photolithography and electron beam lithography. These methods allow the creation of prescribed patterns that are well ordered and geometrically precise. However, they usually require very expensive equipment and a high level of expertise. See Table 1.

The value of an ordered topography over an unordered topography was investigated by Curtis et al.8 In their experiment, surfaces of various nanoscale ordered patterns were created using electron beam lithography and surfaces with nanoscale unnordered patterns were created using colloidal lithography. Rat epitonone fibroblasts showed higher level of adhesion to the surface with an unordered pattern than to that of both planar surfaces and ordered surfaces. The surfaces with the ordered patterns had even lower levels of adhesion than flat surfaces. Not only must size and shape of the features be considered by the bioengineer when designing a scaffold, but also the technology used must be carefully considered to create a surface appropriate to the desired results.

What follows is a description of the main fabrication methods that have been used for creating nanoscale features for tissue engineering scaffolds and substrates. Examples of studies done with various cell types are cited as demonstration of the types of results obtained by various research groups. Table 2 lists the discussed fabrication methods and shows the size scale of features that can be created as well as some of the main advantages and disadvantages to each method.
Table 2.

Nanoscale fabrication methods

Fabrication method

Feature sizes



Electron beam lithography

For positive results: ≥3–5 nm for a single feature. ≥30–40 nm for arrays of features

Precise geometries and patterns can be created. No mask needed. Patterning is compouter-controlled

Expensive equipment, time consuming, small surface coverage. Increasing area, lowers resolution. Negative resists have even lower resolution

Polymer demixing

Vertical: ≥13 nm

Simple, fast and inexpensive

Only sample features can be created (e.g., pits, islands, ribbons)

Colloidal lithography

≥20 nm

Easier to pattern larger areas than with EBL

Specific feature geometries not possible

Chemical etching

Dependant on etchant used and time as well as many other factors

Simple, fast, no special equipment needed

Specific feature geometries not possible

Nanofiber scaffolds

Allows for 3-D scaffolds to be created


≥3 nm

Can create aligned fibrous meshes. Can be used with biological polymers such as collagen

Can only create fibers

 Phase separation

Pore sizes ≥1 nm

Good for creating porous scaffolds. Simple, no special equipment necessary. Porosity is easily controlled

No organized pattern possible


Tailored by molecule design

Molecules undergo spontaneous self-assembly into higher order structures

Requires engineering of molecules that will self-assemble

Nanoporous membranes

≥1 nm


 As scaffolds


Relatively easy fabrication, inexpensive, allows for precise control over pore size and distribution. Alumina membranes are bioinert

Strength of materials can be insufficient to withstand physiological loads

 For cell encapsulation


Can provide an immune barrier for forcing cells. Pore size can be precisely controlled with very small deviations

Difffusion in and out of the membrane may be impeded resulting in cell death

Photolithography (near-UV)

≥0.5 nm

Can create precise geometries and patterns

Expensive equipment, feature size is the largest of reviewed methods

Electron Beam Lithography

Electron beam lithography has been used to create surface topographies at the nanometer scale for studying cellular growth and behavior on these surfaces.8,68 Electron beam lithography involves the use of high-energy electrons to expose an electron-sensitive resist.45 Both positive and negative type resists are available; however, negative resists result in lower feature resolution. Unlike photolithography, a physical mask is not needed to pattern the surface with the beam of electrons. The pattern is programmed into the unit in order to precisely control how the beam travels over the surface (see Fig. 1). The resolution of this technique takes many factors into consideration including electron scattering in the resist and backscattering from the substrate.45

Schematic of electron beam lithography process. The basic componenets of importance are detailed in the top left figure. Moving from top left to bottom right, a computer-controlled electron gun scans an electron beam across the resist-coated surface creating the preprogrammed pattern. Note that no mask is needed to define the pattern area on the substrate.

With optimization of the process, EBL has the ability to create single surface features down to about 3–5 nm.68 However, for applications of studying cellular growth and behavior, single features are not of practical use. Typically, high-density arrays of a single feature such as posts or channels are used in such experiments. This allows for enough surface area to be covered with the pattern to observe a cell population's response to the surface features. The desire to create a relatively large surface area of features drops the limit of EBL to approximately 30–40 nm as demonstrated by Vieu et al.68 At these resolutions, development of the resist was seen to play an important part in bringing out the pattern. Ultrasonic agitation is needed to overcome intermolecular forces that prevent the resist from dissolving. Also, using pure isopropyl alcohol (IPA) for development assists in the development of very high-density arrays.

Electron beam lithography allows for surfaces with regular patterns of nanotopographical features to be created. The ability to create an ordered array of a single feature allows for investigators to observe a large population of cells in order to gain a significant understanding of how the cells respond to a single feature. That is, the response of a cell to a nanoscale post on a surface would be difficult and statistically insignificant if only one post and one cell were observed. By creating tens of thousands of posts on a surface and seeing how a population of cells reacts can give insight into the mechanisms of single-cell surface interactions. Dalby et al. investigated fibroblast response to arrays of nanopits created using EBL.11 Specific filipodia–pit interactions were evident on surfaces with pits of 75 and 120 nm. The fibroblast filipodia were seen to extend to and end on pits of these diameters (Fig. 2). Recognition of the pits by the fibroblasts seems obvious. Smaller pits with diameters of 35 nm, however, failed to show specific interactions, although it was evident that fibroblasts were still reacting to the smaller features due to the observation that cells with their body on flat parts of the surface had many filipodia extending into the 35 nm pitted domain.

Fibroblast filipodia interacting with a 120 nm pit. Pits were created in resist-coated on silicon wafers using electron beam lithography and then replicated in poly(caprolactone). Permission for figure provided by Dalby et al.11.

Fabrication using EBL can be time consuming and costly. To overcome this, nanometer patterns can be replicated in polymeric materials making the ability to mass reproduce the patterns created a much faster and inexpensive process. For cell studies, many samples are often needed to obtain statistically significant results. When dealing with microstructured surfaces, researchers often replicate the master pattern created on a silicon wafer by using some type of curable polymer. Polydimethylsiloxane (PDMS) has been by far the most common material used for replicating patterns created on the surface of silicon wafers.27 A variety of polymers have been used for this, however, with the choice depending on the application. Nanometer patterns created using EBL can be replicated in the same manner, making the ability to mass reproduce the patterns created a much faster and inexpensive process. Replication has been demonstrated in polycaprolactone (PCL) using hot embossing11,30 and solvent casting.8 This can dramatically cut the time and costs associated with the creation model substratum using EBL.

Polymer Demixing

A unique method for creating nanoscale topographical features for use as cell substrates is through the use of polymer demixing. Polymer demixing involves the spontaneous phase separation of polymer blends [e.g., polystyrene/poly(4-bromostyrene) blends], which occurs under conditions such as spin casting onto silicon wafers.2 This process is a quick and inexpensive method for creating cell culture surfaces with nanoscale features. Polymer demixing can be used to create topographies similar to those commonly used to study cell growth on nanostructured surfaces. Nanoscale pits, islands, and ribbons can be created through the careful adjustment of polymer ratio and concentration. Changing the polymer ratio changes the shape of the features while changing the polymer concentration changes the feature size.3,4

Polymer demixing does not lend itself to creating ordered arrays of structures or structures with precise geometries. The features created using polymer demixing are somewhat uncontrollable in the horizontal direction, yet very precise control can be achieved in the vertical direction. Features appear randomly placed over the surface (Fig. 3). Because of this lack of control over pattern organization, polymer demixing may not be ideal for creating model surfaces to study cellular interactions with nanoscale features. Despite this limitation, nanotextured surfaces created by polymer demixing have been used to study the interactions of various cell types to nanotopographies of various heights.7,10,12,13,14,17,56 These studies have shown fibroblast filipodia interacting specifically with 10 nm high islands on the surface.14 Fibroblast interaction with nano-islands was also shown to have upregulated activity in many genes including those associated with cell shape, movement, and ECM production.17 While fibroblasts are seen to interact quickly and strongly with nano-islands created by polymer demixing, they seem to grow less well on these nanopatterned surfaces.12 It is suggested that this is likely because the cells remain motile on the surface instead of settling down to make strong contacts with the material as they do on a flat surface.

Example of nano-islands created using polymer demixing. Average island height is 27 nm. Permission for figure provided by Dalby et al.12.

Similar studies with endothelial cells showed that polymer demixed nano-islands played an important role in cell spreading, proliferation, and morphology.13 Cells were seen to be more spread on surfaces containing nano-islands than on control polystyrene and poly(4-bomostyrene) surfaces. On surface with features as small as 13 nm, cells were seen to form multilayers after 1 week. A common morphological feature found to the epithelial cells on the nano-islands was an acruate morphology similar to that found in endothelial cells lining the vasculature. This result suggests that the nanotopography alone can induce proper morphology by providing similar physical cues to native ECM.

Colloidal Lithography

Colloidal lithography provides another inexpensive method for creating nanoscale topographies. This technique allows for the production of surfaces with controlled heights and diameters. Colloidal lithography involves the use of nanocolloids as an etch mask. These nanocolloids are dispersed as a monolayer and are electrostatically self-assembled over a surface. Directed reactive ion beam bombardment or film evaporation can then be used to etch way the area surrounding the nanocolloids, as well as the nanocolloids themselves (Fig. 4).18,33,34 Common biologically relevant patterns such as nanocolumns and nanopits can be created with this method. Directed reactive ion beam bombardment can be used to produce nanocolumns while film evaporation will leave nanopits after removal of the colloid. To vary the surface structure, the coverage of the monolayer of colloid and the size of the colloid can be varied. The spacing between particles can be controlled by changing the ionic strength of the colloid solution. Patterning with particle sizes of 20 nm has been demonstrated.71 With this technique, large surface areas can be patterned (∼cm2), making it a suitable technique for creating functional biomaterials for cell studies.33

Shematic of colloidal lithography process. Nanocolloids are electrostatically dispersed over the surface and are then etched away along with the surface itself leaving a patterned substrate.

This method was used by Dalby et al. to produce nanocolumns in order to investigate fibroblast sensing of nanotopography.15 An increased density of filipodia was observed on nanocolumns of approximately 160 nm height and 100 nm width, suggesting that filipodia (50–70 nm wide) play an important part in gathering topographical information from the cells environment. Also, they suggest that it is the direct transmittal of forces encountered by the cell, as it spreads across the nanotopography, to the nucleus via the cytoskeleton that account for changes in cell morphology, adhesion, and proliferation.16 A very interesting observation made was that some fibroblasts cultured on nanocolumns 160 nm in height and 100 nm in diameter showed macrophage-like morphology with actin- and tubulin-associated cavities forming around nanocolumns. This observation, along with observed localization of clathrin at the cell peripheries, and increased levels of dynamin (two proteins associated with the process of endocytosis), suggest that the fibroblasts are attempting to endocytose the nanocolumns.9

The morphology of epithelial cells was also studied using surfaces created by colloidal lithography.5 These tests showed increased cell spreading of rat pancreatic epithelial cells on hemispherical nano-pillars with varying diameters. Progressively larger diameters resulted in increased percentages of cells showing blebs protruding in various directions from the main body of the cell. The fact that the cells showed protrusions and were nonrounded (as compared to flat surfaces) indicates that the cells may be differentiating into (neuro)endocrine cells based on the nanotopographical cues.

Chemical Etching

Chemical etching is a means of producing nanoscale features on the surface of a material by soaking it in an etchant. Typical etchants are hydrofluoric acid (HF) and sodium hydroxide (NaOH). As the material is etched away, the surface is roughened creating pits and protrusions at the nanometer scale. This process is essentially a surface treatment and cannot create structures with any prescribed geometry or organization. It can, however, provide a very quick, easy, and inexpensive means of creating a nanostructured surface by changing the scale of the roughness on the material surface. The change in the scale of roughness of the surface has been shown to significantly affect many cellular growth parameters. Thapa et al. produced poly(lactic-co-glycolide) (PLGA) and poly(ether urethane) (PU) films with feature dimensions of 50–100 nm.66 The films were roughened by soaking in 1N NaOH for 1 h. Bladder smooth muscle cells grown on these nano-roughened surfaces were seen to have increased levels of viability and function. Levels of viability and function were also seen to increase with decreased feature size. This provides further evidence that the cells need certain features at the nanoscale in the topography of the environment in which they grow.

Silicon wafers can be etched using HF to create nanometer-scale roughness on the surface of the wafer. Etch time can control the roughness with longer etch times leading to rougher surfaces (i.e., smaller surface pits and protrusions). Surfaces prepared this way have been used to study neural cell adhesion and viability.24,25 Interestingly, adhesion was seen to improve in only a certain range of roughnesses (20–50 nm) while adhesion was negatively affected above and below this range.25

The creation of nanostructured surface has also moved from those created on flat, planar surface to three-dimensional scaffolds with nanoscale features. Pattison et al. created three-dimensional PLGA scaffolds with nanometer-size features.54 Chemical etching using NaOH was done to create nanometer-scale roughness on every surface of the scaffold. Roughness features smaller than 100 nm were created in this way. Also, porosity of the scaffolds was increased two-fold over conventional PLGA scaffolds using this method. Bladder smooth muscle cells were grown on these scaffolds and shown to have increased adhesion and proliferation. Many of the scaffold properties changed through the process of making it nanostructured. All of these changes in tandem may play an important part in these improved results. Not only may the nanoroughened surface increase adhesion, through its own mechanisms, but the increased surface area provides more room for cell adhesion and more room for the population to grow into. Also, increased porosity allows for greater infiltration of the scaffold by cells as well as increased nutrient and waste diffusion.


The goal of all scaffolding techniques has been to recreate an environment for cell growth that in some way mimics the natural environment of the cells. Collagen is an extacellular matrix component found in nearly every tissue in the human body. Types I, II, and III collagens form a fibrous network that among other things, provide structural strength to a tissue and support cell growth. All of these collagen types form triple helical structures that pack together to form fibrils. The typical ECM protein fiber, type I collagen, has a diameter ranging from 50 to 500 nm.62 Many researchers are attempting to mimic this fibrous network in the hope that by recreating the ECM at the same scale as the cells in it would normally be exposed to, the cells will “feel more at home” and grow in their normal manner. To recreate the ECM, scaffolds made of nanofibers (i.e., fibers with diameters in the nanometer range) are created and seeded with cells. The desire is to have the diameter of the artificial fibers approach that of natural collagen fiber bundles.

There are three main methods for making these nanofibrous scaffolds: Electrospinning, phase separation, and self-assembly.64


Electrospinning can be used to create ultra-fine fiber down to the nanoscale.28 The process of creating nanofibers is explained in detail in the review by Huang et al.37 Basically, it involves electrically charging a suspended droplet of polymer melt or solution. To create nanofibers, the polymer is suspended from a capillary. An electric field is applied to the end of the capillary tube. When the electrostatic charge overcomes the surface tension of the droplet a polymer jet is formed. The jet elongates and thins due to an instability process. As the solvent evaporates from the jet, an electrically charged polymer is left behind. These solidified fibers are then collected on a grounded surface (Fig. 5). The cited review details 50 different polymers that have been electropsun with diameters ranging from <3 nm to over 1 μm. The electrospinning process can even be applied to natural biopolymers such as collagen.36,47,60 Matthews et al. were able to create fibrous mats of Type I collagen with 100 nm fibers that exhibited the 67 nm banding pattern characteristic of native collagen.47

Electrospining schematic. The polymer jet is collected on a grounded surface. After collection, it can be post-processed to fit the desired application.

This process of electrospinning creates an unordered mesh of fibers (Fig. 6). These electrospun fiber meshes have a very high surface-to-volume ratio with very high porosities. This increases surface area available for growth and mass transport of nutrients and wastes to and away from the cells. These factors are of primary concern in any tissue-engineered scaffold.

Randonly oriented electruspun PLGA nanofiber scaffold. Permission for figure provided by Smith et al.64.

Electropun fibrous scaffolds have been used with a variety of cells to their response to the nanostructured scaffold. Electrouspun silk scaffolds have been used successfully to support bone marrow stromal cell growth.38 Polycaprolactone nanofibrous meshes have been used to support colonies of cardiac myocytes that began spontaneously contracting after 3 days of culture.61 Various electrospun fiber scaffolds have also been investigated using fibroblasts, keratinocytes, osteoblasts, and smooth muscle cells and have been shown to support adhesion and proliferation.6,29,30,39,48,49 Type II collagen nanofibrous mats were created for culturing chondrocytes by Shields et al. as possible replacement for cartilage.60 These scaffolds showed good infiltration of chondrocytes into the interior.

Although electrospinning typically falls in the realm of creating an unordered scaffold matrix, the advantage of scaffolds with directional cues to many tissues, including neural cells, has been demonstrated in the past.31 Novel work by Yang et al. has shown the creation of aligned electrospun fiber scaffolds for use as three-dimensional scaffolds for neural stem cells (NSC) from mice (Fig. 7).74 The fibers were made of poly(l-lactic acid) and were aligned by collecting the fiber jets on a rotating, as opposed to stationary, plate. Average fiber diameter was 250 nm. NSCs cultured in these aligned fiber scaffolds showed elongation and neurite extension parallel to the fibers. Neurite length was also observed to be longer on the nanometer-scale fibers as opposed to similar scaffolds created with micrometer diameter fibers.

Aligned PLLA nanofibers created using electrospinning. Permission for figure provided by Yang et al.73.

This same technique for creating aligned nanofiber scaffolds has been used to create biodegradable scaffolds for smooth muscle cells.72 SMCs aligned along the axis of the fibers displayed a spindle-like contractile phenotype. This type of organization is typical of SMCs in the vasculature.

Another method for creating aligned nanofibrous meshes is to do some post-processing on the collected fibers. Zhong et al. used the standard electrospinning technique to create an unorganized collection of fibers.76 The collected fibers were then uniaxially stretched causing alignment of the fibers. Cardiac myocytes cultured on these scaffolds organized and grew to follow the fiber direction.

Aligned bundles of cells are a fairly common motif throughout the human body. In tissue engineering, the standard technique for aligning cells has been to use microchannels. Novel methods of electrospinning, such as described here, may provide a vast improvement over current micro-scale alignment methods by moving tissue engineering in to the nanoscale domain. The use of aligned electrospun fiber networks gives the ability to create large scale three-dimensional scaffolds that provide alignment cues to the cells, something that is difficult to do using microfabrication techniques.

Methods of electrospinning to make more complex spatial designs and using multiple polymers have been developed.59 This essentially involves layer-by-layer collection of different polymers on the same collection plate for creating multilayer scaffolds or moving the collection plate to mix two polymers while they are simultaneously collected.

Phase Separation

Phase separation relies on no special equipment and allows for three-dimensional scaffolds to be created with fibers in the sub-micron range. Ma et al. have used a thermally induced phase separation process to create nanofibrous foams (Fig. 8). This process occurs in five steps: Polymer dissolution, phase separation and gelation, solvent extraction from the gel, freezing followed by freeze drying in water.64 This process forms a continuous fiber network that can be tailored to any application. Pore structure and fiber diameter are easily controlled and batch-to-batch consistency is high. Pore structure is easily varied by changing the solvent used in the process.69 This change in solvent can change the phase separation process from liquid–liquid to solid–liquid depending on the solvent mixtures used thus changing the final pore morphology. Initial polymer concentration plays an important part in porosity of the final scaffold and fiber diameter. High initial concentration of polymer leads to thicker fibers and less porous networks.73 This process can form fibrous networks with fiber diameters in the same range as type I collagen fibers. The fibers created with this technique are randomly oriented and unlike with electrospinning, no work has been reported of organized nanofibers scaffolds created via phase separation.

PLLA scaffold created using phase separation. Fibers range from 50 to 500 nm. Permission for figure provided by Smith et al.64.

Liquid–liqiud phase separation was used by Yang et al. to produce nanofibrous PLLA scaffolds for nerve tissue engineering.73 Nerve stem cells cultured on these scaffolds showed in-growth by the cells and differentiated morphology with neurites extending to twice the length of the cell body after several days of incubation.

The ease with which these scaffolds can be created and the demonstrated ability of these highly porous networks to support cell growth makes them an attractive solution for creating nanofibrous matrices for use as tissue engineering scaffolds.


Self-assembled nanofibrous scaffolds involves the use of engineered polymers that undergo self-assembly to form scaffolds composed of a matrix of nanofibers. The key to the self-assembly of these scaffolds is the engineering of the molecules themselves. A common technique for creating these self-assembling scaffolds is to engineer amphiphilic peptide sequences.35 The process and physics of this method are detailed in the reviews by Tu and Tirrell67 and MacPhee and Woolfoson.44 Self-assembly is largely driven by the hydrophobic moiety. In water, the hydrophobic portions of the peptide sequences will be driven away from water and towards each other. Proper engineering of the peptide amhiphiles allows control over the self-assembled matrix.

Fields and Tirrell et al. have created peptide amphiphiles that mimic the triple helix super secondary structure of collagen.26 These peptide amphiphiles contain a type IV collagen peptide sequence head group and a lipophilic tail composed of long-chain mono- or di-alkyl esters. The triple helical structure is formed by the collagen peptide head while hydrophobic interactions between the tail groups induce and stabilize the self-assembled three-dimensional structure of the scaffold. To study the bioactivity of these scaffolds, melanoma cells were cultured on them. Adhesion was examined on the head and the tails individually as well as on the entire peptide amphiphile. Adhesion was significantly enhanced on the triple helical structure of the scaffold over that of the individual elements.


One of the most widely studied systems in tissue engineering is the vascular system. Due to their extremely small size (5–10 μm in diameter) and complexity, the ex vivo engineering of capillary networks for later implantation has proven difficult. Studies have shown that the local topography plays an important role in the development of capillaries in vivo.51 It is the hypothesis of Moldovan et al. that preformed tunnels found in cardiac tissue of transgenic mice are thought to form into new capillaries when they connect to existing vessels. Micro- and nano-electromechanical systems (MEMS and NEMS) techniques lend themselves to developing scaffolds to replicate tissues that display growth and organization based on spatial cues. Nanoporous membranes with pores in the range of 10–20 nm, developed for use as nanofilter based drug-delivery capsules have been suggested as a platform for growing micro-vascular networks in vitro.20 The suggested method is to fill nanofilter-based capsules with angiogenic or angiostatic substances and seed the surface of the membrane with endothelial cells. By micromachining capillary-sized grooves into the surface of the membrane, the endothelial cells could assemble into micro-vascular like networks that could then be implanted and the engineered network would then, in theory, connect to the existing vasculature.50

Another widely studied system in which nanoporous membranes have been the scaffold of choice is bone. Nanoporous alumina has received much attention for its use in bone tissue engineering. The use of nanoporous alumina stems from the idea that the feature size matches that of the inorganic particles in bone. Swan et al. have produced porous alumina membranes with pore sizes ranging from 30 to 80 nm created using a two-step anodization process.42 These alumina nanoporous scaffolds were found to increase osteoblast adhesion and matrix production in longer term cultures. Adhesion and proliferation were increased for up to 4 days of culture,55 while matrix production was increased for 4 weeks of culture [Figs. 9(top) and 9(bottom)]. Futher modification of the porous membranes was achieved by immobilizing RGDC moieties onto the alumina.41 This addition of this peptide sequence did not clog the pores and increased osteoblast adhesion as well as matrix production 2 days into culture.

Osteoblast growing on nanoporous alumina membrane. (Top) Cell processes are seen extending into nanopores. (Bottom) Cells producing increased levels of extracellular matrix (fibrous bundles).


The success of a tissue-engineered construct when transplanted into a living being will depend foremost on the response it elicits from the host's immune system. Immune rejection of the construct is due to antibody recognition of foreign antigens present on either the cells or the scaffold material (if biological materials are used). Researchers have attempted to overcome this recognition of foreign antigens by creating immunoisolation devices to house the engineered tissue construct. These devices block antibodies from getting at the transplanted cells by creating a physical barrier to diffusion into the scaffold. The key to these immunoisolation devices is to block antibodies from entering, but allow nutrients in and wastes and cellular products out. To do this, porous materials (e.g., polymers) have been investigated as size-selective barriers. The problem with these materials, however, is that they usually exhibit a range of pore sizes that do not adequately block all antibodies. For immunoprotection, a well-controlled pore size is needed.

Nanoporous capsules with uniform pore dimensions and pores sizes down to 7 nm were developed for islet cell transplantation using bulk and surface micromachining techniques43 (Fig. 10). Studies with these capsules have shown that cell viability and functionality were not compromised by the encapsulation.21 The porous membranes of these capsules were shown to provide sufficient insulin and glucose diffusion for nutrient exchange for the encapsulated cells. These membranes also showed an almost complete deselection of IgG over extended periods of time. In vivo studies done on rats with encapsulated insulinoma cells showed a reversal in diabetes and normal blood glucose levels over 2 weeks (total encapsulation time was 14 days).

Nanoporous membrane created using bulk and surface micromachining techniques. Pore size is 24.5 nm.


The exploration of the nano-environment that cells are exposed to in vivo is leading researchers to try to mimic that environment with special materials and processes. In using microtechnology, the field of tissue engineering began with the systematic study of how the scale and geometry of various features affected cells seeded on them. As better understanding and control of the micro-environment of the cells was gained, these techniques have begun to be employed to create constructs that mimic native tissues. The use of nanotechnology for tissue engineering is now going through the initial stages where techniques are being refined and simplified model systems are being used to gain insight into cellular responses to their nano-environment. Many of the techniques described in this review are being continuously improved upon to gain better control of the size and organization of the environment created.

As with microtechnology for tissue engineering, most of the new techniques stem from advances in the microprocessor industry. The desire to continue fabricating microchips in silicon with increasingly higher numbers of transistors has pushed scientists and engineers to explore new methods of nanofabrication. Shorter wavelength lasers can allow the exposure of ever smaller features using photolithography. Wavelengths in the deep UV (157 nm) and ‘soft’ X-ray regions (2–50 nm) may allow pushing the current lower limits of photolithography to the natural limits of CMOS-based devices.46,58

The use of the atomic force microscope (AFM) and the scanning tunneling microscope (STM) to image surface in the nanometer and below domain has led researchers to take active control of the stylus on these instruments and use them for writing nanoscale patterns as opposed to reading them. Scanning probe lithography (SPL) uses the tip of the AFM or STM to write nanometer-scale features on a surface as opposed to reading them. SPL has been used to etch sub-100 nm patterns using photoresist or self-assembled monolayers (e.g., alkanethiols) as the resist.40,70 SAMS are deposited on a silicon or metal surface and are then etched away using the nano-probe tip to scribe the desired pattern (Fig. 11). Chemical etching is then used to create the pattern in the material surface.65

Schematic of scanning probe lithography (SPL) that uses an STM or AFM tip to etch away SAMS (resist) deposited on a silicon wafer or metal substrate.


The authors would like to acknowledge the financial support of National Heart, Lung, and Blood Institute Grant NIH (64956) and Johnson & Johnson.

Copyright information

© Biomedical Engineering Society 2006