Journal of Micromechanics and Microengineering

PAPER

Micro-patterning of porous alumina layers withaligned nanoporesTo cite this article: Michael D Canonica et al 2015 J. Micromech. Microeng. 25 015017

 

View the article online for updates and enhancements.

Related contentTechnical NoteLauri Sainiemi, Jukka Viheriälä, TiinaSikanen et al.

-

A nanoporous silicon nitride membraneusing a two-step lift-off pattern transferwith thermal nanoimprint lithographyBhargav P Nabar, Zeynep Çelik-Butler,Brian H Dennis et al.

-

Fabrication of polymer nanowires viamaskless O2 plasma etchingKe Du, Ishan Wathuthanthri, Yuyang Liu etal.

-

Recent citationsNanoengineered thrusters for the nextgiant leap in space explorationPaulo C. Lozano et al

-

This content was downloaded from IP address 18.34.5.38 on 30/03/2018 at 00:03

1 © 2015 IOP Publishing Ltd Printed in the UK

1. Introduction

Porous alumina is a well-known nanostructured material that can be made in a variety of ways, including via anodization of a polished aluminum foil in acid or basic solutions, to create porous anodic alumina (PAA) that is comprised of vertically aligned capillaries having a diameter of several nanometers and distributed in a random manner [1, 2]. By adjusting the parameters of the anodization, diameters of the capillaries or pores can be tuned from few to hundreds of nanometers [3–5]. In addition, highly ordered porous alumina with straight capil-laries can be obtained by etching the porous alumina produced during the first anodization and anodizing the aluminum layer a second time [6, 7].

Due to its tunable porosity, its electrical insulation and its chemical resistance, PAA has been used in many applications such as; dyeing of aluminum layers [1], humidity sensors [8], microcantilevers [9], antireflection structures [10], photonics crystals [11], insulating layers for electronic devices [12, 13] and hermetic packaging for microsystems [13–15]. Moreover, due to its capillaries, high aspect ratio structures of alumina were fabricated [9, 16, 17] and due to its nanometer-range porosity, PAA was employed, as a template or as a mask for the fabrication of nanostructured materials [18], nanotubes [19, 20], nanowires [21, 22], nanoholes [23] and nanodots [24].

In this paper, nanoporous alumina microfeatures intended for applications such as electrospray propulsion [25], drug

delivery, nanoprinting, etc, were fabricated by wet etching of commercially available PAA layers. Prior to any processing, the morphology and the size of the pores of the PAA layers were studied using electron microscopy and image processing tools. Next, chemical composition of these layers was obtained using energy-dispersive x-ray spectroscopy (EDS) and, wet etching of PAA for different temperatures and times was intensively studied. Based on the wet etch results, microfeatures were fab-ricated using a reliable fabrication process and characterized.

2. Experimental

2.1. Properties

Commercially available PAA discs with pore size of 200 nm and 20 nm, usually used as filters, were chosen as substrates for the fabrication of microfeatures (Anopore Inorganic Membranes, GE Healthcare Life Sciences). These substrates were selected because their thickness of about 60 µm, allows making struc-tures of several micrometers high, their diameters that range from 13 to 47 mm allow patterning many structures on the same disc and their capillaries are opened on both ends and neither aluminum, nor porous alumina barrier layers had to be etched.

The morphologies of the randomly distributed nanopores of the two types of disc were studied by observation with a scanning electron microscope (SEM) (Zeiss Supra 40). These observations of non-tilted and tilted samples showed that

Journal of Micromechanics and Microengineering

Micro-patterning of porous alumina layers with aligned nanopores

Michael D Canonica, Brian L Wardle and Paulo C Lozano

Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139, USA

E-mail: plozano@mit.edu

Received 10 July 2014, revised 24 September 2014Accepted for publication 21 October 2014Published 30 December 2014

AbstractArrays of nanoporous microfeatures of alumina are demonstrated via isotropic chemical etching of commercially available nanoporous alumina substrates. Etch rates are established for four morphologies of aligned nanopores having pore sizes in the range 20–220 nm and porosities from 30–60%. Parametrically varying etch conditions yields a process for creating microfeatures over large areas which is used to create a 384 porous tip array at a pitch of 70 µm with the 200 nm porosity.

Keywords: alumina, porous, anodic, AAO, wet, etch, nanopore

(Some figures may appear in colour only in the online journal)

M D Canonica et al

Micro-patterning of porous alumina layers with aligned nanopores

Printed in the UK

015017

jMM

© 2015 IOP Publishing Ltd

2015

25

j. Micromech. Microeng.

jMM

0960-1317

10.1088/0960-1317/25/1/015017

Paper

1

journal of Micromechanics and Microengineering

IOP

0960-1317/15/015017+10$33.00

doi:10.1088/0960-1317/25/1/015017J. Micromech. Microeng. 25 (2015) 015017 (10pp)

M D Canonica et al

2

these discs are composed of two layers with different thick-nesses, pore shapes and sizes (see figure 1) [26]. One layer has circular pores with a diameter of about 200 nm and a thick-ness of about 60 µm. The other layer has a thickness of less than 1 µm and pores formed by branches and having the size defined by the vendor. In addition, cross-sections of a broken substrate showed that capillaries are not straight and contain some local defects generated during the anodization process.

The size of the pores for both faces of each disc was meas-ured by processing SEM images using the software Matlab. The processing of the images had for purpose to obtain the position, the surface, the radius of each pore on the image and the porosity of the layer, (defined as total surface of pores of an image divided by the total surface of the images). Pictures of the faces of the discs with a resolution of 1536 × 2048 pixels were taken at a magnification of 20 KX, which means that a pixel of the image corresponds to 7.5 nm. These images were cropped to remove the scale bar and converted to black and white using a threshold defined by the method of Otsu [27]. Then, to detect the walls between the pores, the region smaller than 30 pixels were removed and morphological closing of the images was performed using a disc of diameter 2 pixels as the structuring element. Finally, the pores were detected, their surfaces and porosities were calculated and assuming that each pore was circular, their diameters were calculated.

For the layer with pore diameter of about 20 nm, due to the small size of pores, the magnification of the image was 166 KX (1 pixel represents 0.88 nm), the regions smaller than 30 pixels were not removed and the threshold for the conver-sion to black and white was determined manually. Moreover, due to the low contrast of the SEM image, not all the pores were detected. For each layer, the median and standard devia-tion of the distribution of pore diameter as well as the porosity were reported in table 1. Each layer was then named according to the median pore diameter and the porosity.

These measurements and calculations show quantitatively that both types of disc have a thick layer, layers 220 nm/60% and 200 nm/40%, with a pore diameter of about 200 nm and a standard deviation of 108 nm and 61 nm. These two layers are used as handle and support for the thin layers, layers 140 nm/40% and 20 nm/30%, which act as a filter and have a standard deviation of 40  nm and 17  nm. Based on pore diameters and porosities, we found that the four layers are all structurally different.

The x-rays spectrum for each layer was obtained with an EDS detector coupled to a SEM (Philips XL-30) at a voltage of 5 kV. For these four spectra, two major peaks were observed at 1.49 keV and 0.52 keV, which corresponds to the K alpha emission lines of aluminum and oxygen respectively. Also, two small peaks were observed at 0.28 keV and 2 keV,

Figure 1. PAA layers were named in the following manner: median of the distribution of pore diameter/porosity (a) Layer 140 nm/40% of the 200 nm disc. (b) Layer 20 nm/30% of the 20 nm disc. (c) Layer 220 nm/60% of the 200 nm disc. (d) Layer 200 nm/40% of the 20 nm disc.

J. Micromech. Microeng. 25 (2015) 015017

M D Canonica et al

3

which are the emission lines of carbon and phosphorous. The presence of carbon was probably due to contamination of the substrates and phosphorous was probably used in the fabrica-tion process of the discs, for example, for pore widening or the etching of the barrier layer.

Based on the measured spectra, the chemical composition in atomic percentage of aluminum and oxygen was calculated and reported in table 1. In addition, since substrates are only 60 µm-thick and porous, no major difference of composition was observed between the two layers of a substrate because the electrons penetrated the disc over several micrometers.

2.2. Layout and fabrication

As an example of microfeatures, single tips and arrays of tips were patterned on porous alumina layers using a layout that consisted of circles of diameter, d, placed in a triangular packing of pitch, p (see figure 2(a)). The number of tips was defined by the number of layer, N, around the central circle and can be expressed by equation (1). The size of the region with the tips was limited by a circle of diameter, D, 1.5 mm. Different designs with different values for these parameters were processed; the diameter of the circles ranged from 20 to 100 µm, the pitch ranged from 60 to 120 µm and the number of layers from 0 for a single tip to 29.

    = · · + +N NNumber of tips 6

( 1)

21 (1)

Prior to any processing, the layers of the discs were identi-fied by observing the pore shapes using a SEM. A piece of polyimide tape was placed on the disc to indicate the layer to process but also to ease the handling of the 60 µm-thick and brittle discs. A 500 nm-thick layer of parylene C was depos-ited on both layers of the discs as a mask for the etching of alumina. SEM observation showed that the parylene did not enter into the capillaries but formed an homogenous layer on both sides of the substrate. Parylene was selected as a mask for its chemical resistance during etching of PAA and because it can be removed by oxygen plasma. Next, a 200 nm-thick aluminum layer was deposited in an electron beam evapo-rator on the parylene layer to process. This aluminum layer was patterned by photolithography and wet etching. The PAA discs were covered by photoresist OCG 825 and spun for 40 s at 2000RPM. The substrates were baked at 100 °C for 1 min on a hotplate and exposed to UV light using a mask aligner for 6 s (lamp power: 10 mW cm−2). The photoresist

was developed for 1 min in developer OCG 934 (3 : 2). The development time was precisely respected to avoid etching of the aluminum layer by the developer. Finally, the photore-sist was baked at 90 °C for 2 h and the aluminum layer was etched in a solution of aluminum etchant type A at room tem-perature for 5 min. The patterned aluminum layer was then used as a mask to pattern the parylene layer by reactive ion etcher (RIE). A directional oxygen plasma (150 W, 95 mTorr, 95  sccm) was used to pattern the parylene for 5  min and remove the photoresist.

To pattern the PAA discs by wet etching (see figure 2(b)), a solution of phosphoric acid 85 wt% (without chromic acid) was selected [28]. The discs were etched in a beaker containing 200 cm3 of etchant and placed on a hotplate used to regulate the temperature of the solution. To ensure etching uniformity, each disc was placed in the beaker in the same position and tempera-ture was measured next to the sample. Prior to the fabrication

Table 1. Statistical results of the pore size of PAA layers obtained by image processing of SEM images and chemical composition obtained by x-ray analysis. For layer 20 nm/30%, the number of detected pores can not be compared with others because the scale of the image was not the same.

Name of layerNumber of pores [-]

Pore diameter

Porosity [%]

Atomic Conc.

Median [nm] Std [nm] AlK [%] OK [%]

200 nm disc 140 nm/40% 2878 137 40 41.9 40.00 60.00200 nm disc 220 nm/60% 1112 221 108 57.8 40.00 60.0020 nm disc 20 nm/30% 779 22 17 28.4 48.57 51.4320 nm disc 200 nm/40% 1453 203 61 42.3 46.69 53.31

Note: Std stands for standard deviation.

Figure 2. (a) Layout that was used to pattern PAA discs and obtain single tip or arrays of tips. The array was defined by three parameters, the diameter of the circles, d, the pitch between the circles, p and the number of layer, N, around the central circle. The dashed line represents the position of the cross-section b. (b) Patterned PAA discs were etched using a parylene mask, which was patterned by RIE using an aluminum mask.

Porous alumina disc Parylene C Aluminum

(b)

60µm

(a)

p

d

0 1 N

Underetching

Dep

th

J. Micromech. Microeng. 25 (2015) 015017

M D Canonica et al

4

of arrays of tips, to determine the optimum conditions, the wet etching of PAA discs was studied for different parameters.

After wet etching of porous alumina structures, although most of the aluminum layer was etched in phosphoric acid, the remains of aluminum were etched in an aluminum etchant for 5 min. Then, the parylene layers on both sides of the discs were etched by oxygen plasma (150 W, 95 m Torr, 95 sccm) for 10 min for the non-etched side and 30 min for the etched side because parylene remains were attached to the etched structures. Finally, the discs were diced in square of 5 × 5 mm2 using a conventional wafer dicer.

Using a SEM, morphology of etched regions was studied and the lateral etching or underetching was measured (see figure 2(b)). Then, the depth of the etched cavity were charac-terized using a profilometer (Veeco Dektak 150) and a white light interferometer (Veeco Wyko NT5100).

3. Results and discussion

3.1. Wet etch results

To determine the optimum etching conditions for the fabri-cation of tip arrays in phosphoric acid (85 wt%), one set of

similar samples from the manufacturer was patterned and characterized by varying several parameters: layer patterned, temperature of the etchant (50 °C and 70 °C) and etching time.

3.1.1. Morphologies. Due to the hexagonal shape of the porous alumina cell and the round shape of the pores, wet etching of this material usually generates alumina nanowires [26, 29]. During wet etching, the etchant propagates in the capillaries over a certain distance and begins etching the walls between the pores, when the thinnest walls between the pores are etched, a piece of wall placed between the hexagonal cells remains and forms a nanowire. However, if the nanowires are etched before the production of new parts of nanowire, nanotips remain on top of the capillaries. These two different behaviors were observed in the etched regions of the porous alumina layers.

The morphology of etched regions that were obtained by etching layer 220 nm/60% in a solution of phosphoric acid (85 wt%) at 50 °C was reported for different etching times in figure 3. In the first 10 min of etching, several micrometers long nanowires were formed in the etched region and formed braids with their neighbors (see figure 3(a)). Then, an equilib-rium between the etching of the nanowires and the production

Figure 3. Formation of nanowires by etching layer 220 nm/60% in phosphoric acid 85 wt% at 50 °C. Top view is on the left of each image and tilted view on the right. (a) Etching time: 10 min. (b) Etching time: 20 min. (c) Etching time: 30 min. (d) Etching time: 40 min.

J. Micromech. Microeng. 25 (2015) 015017

M D Canonica et al

5

of new part of nanowires was reached and nanowires with a length of about a micrometer were observed (see figures 3(b)–(c)). After etching most of the disc, capillaries with sharp nanotips but no nanowire remained (see figure 3(d)) because as revealed by SEM images, in this part of layer 220 nm/60%, the walls between the pores are larger than those on the surface of the layer, thereby, the nanowires were completely etched before the production of new nanowires. The same morpholo-gies were observed when this layer was etched in a solution at 70 °C, however, the change of morphology was two times faster than at 50 °C.

When layer 140  nm/40%, was etched at 50  °C, nanow-ires several micrometers long were observed and remained throughout the etching process. However, at 70 °C, nanowires were quickly etched and nanotips were observed.

For both layers of the 20  nm discs, 20  nm/30% and 200 nm/40% and for both temperatures, analysis of the mor-phology of etched regions showed nanotips on top of the capillaries but no nanowires as depicted in figure 4.

3.1.2. Etching. For each etching time and for each layer, a different and dedicated disc was used and the profile of an

etched cross used as dicing mark and having a total surface of 60  ×  300  µm2 was measured. Indeed, first etching tests, demonstrated that etch rates of PAA are dependent on the sur-face of the etched region, smaller surfaces were etched faster, therefore, all measurements were performed in the etched dic-ing marks. The etched region being not completely planar, the average depth and the standard deviation were calculated from the measured profile of the region and plotted as function of etching time and for both temperatures in figure 5. The etch rates were then obtained by calculating the slope of the linear regression of data and reported in table 2. Due to limited num-ber of available samples, not all the points were obtained for layers 20 nm/30% and 200 nm/40%.

The similarities of etch profiles between the two layers of the 20 nm discs, layers 20 nm/30% and 200 nm/40%, which correspond to each face of the disc show that the thin layer (less than 1  µm), layer 20  nm/30% does not influence the etch profile because it is quickly etched and the back of the thick layer (layer 200 nm/40%), is etched. Also, the linearity of these etch profiles tends to confirm as observed by SEM that the alumina walls between the pores of layer 200 nm/40% have a constant thickness.

Figure 4. Formation of nanotips by etching layer 20 nm/30% in phosphoric acid 85 wt% at 50 °C. Top view is on the left of each image and tilted view on the right. (a) Etching time: 10 min. (b) Etching time: 20 min. (c) Etching time: 40 min.

J. Micromech. Microeng. 25 (2015) 015017

M D Canonica et al

6

On the other hand, for layers of the 200 nm discs (layer 140 nm/40%, layer 220 nm/60%), etch profiles show a decrease of etch rate after etching 40  µm of layer 220  nm/60% and when the first 10 µm of layer 140 nm/40% were etched. These decreases confirm as observed that the alumina walls of layer 220 nm/60% have a thickness that increases from the top to the bottom of the layer, thereby, at the bottom more material has to be etched and etch rate decreases.

These results show also that for both temperatures, layers 20 nm/30% and 200 nm/40% of the 20 nm discs were etched

at a lower etch rate than layers 140 nm/40% and 220 nm/60% of the 200  nm discs. This is in agreement with the porosi-ties of the 60  µm-thick layers, layer 200  nm/40% has a smaller porosity than layer 220  nm/60%, thereby, for layer 200  nm/40% more material were etched and its etch rates were smaller than layer 220 nm/60%.

3.1.3. Underetching To design precisely porous alumina microfeatures, underetching was studied by measuring with a SEM, the difference between the dimensions of the etched structures and the lithography mask (see figure 2(b)). Mea-surements of patterned structures performed before alumina etching showed that imprecision due to previous fabrication steps (photolithography, aluminum and parylene etching) was of +/ − 2 µm, therefore, for the calculation of alumina underetching, this imprecision was not taken into account. However, to compensate for local differences, for each sam-ple, six measurements were performed and the results with error bars and linear regressions were plotted as a function of etching time in figure  6. Based on the slopes of these linear regressions, the underetch rates were calculated and reported in table 2.

Table 2. Etch and underetch rates calculated using the slopes of the linear regressions of figures 5 and 6. For layer 220 nm/60%, etch rate at 50 °C was not determined because etch profile was not linear.

Etch rate [µm min−1]

Underetch rate [µm min−1]

50 °C 70 °C 50 °C 70 °C

140 nm/40% 1.84 2.48 0.51 1.14220 nm/60% NA 1.94 0.48 0.7920 nm/30% 0.82 1.35 0.42 0.90200 nm/40% 0.81 1.26 0.39 0.95

Figure 5. Etch depth as a function of time of wet etched PAA layers in phosphoric acid 85 wt%. (a) Etchant at 50 °C. (b) Etchant at 70 °C. Error bars show the standard deviation of the samples. LR means linear regression and FC means fitted curve.

0 5 10 15 20 25 30 35 400

10

20

30

40

50

60

Etching time [min]

Dep

th [

µm]

140nm/40% LR220nm/60%FC20nm/30% LR200nm/40% LR

0 5 10 15 20 25 30 35 400

10

20

30

40

50

60

Etching time [min]

Dep

th [

µm]

140nm/40%LR220nm/60% LR20nm/30% LR200nm/40% LR

(a)

(b)

Etchant temperature: 50°C

Etchant temperature: 70°C

J. Micromech. Microeng. 25 (2015) 015017

M D Canonica et al

7

At the temperature of 50  °C, no major difference of underetch rates was observed between layers and the alu-mina under the parylene mask was etched at a rate of about 0.45 µm min−1. When the temperature was increased to 70 °C,

the underetch rates increased of about 2.3 times, to about 0.95 µm min−1 for all layers except layer 220 nm/60%, which increased only of 1.6 times because the sample measured after 30 min of etching at 70 °C was not etched uniformly,

Figure 6. Lateral etching or underetching as a function of time of PAA layers wet etched in phosphoric acid 85 wt%. Linear regression based on the measurements were plotted. (a) Etchant at 50 °C. (b) Etchant at 70 °C. Error bars show the standard deviation of the samples. LR means linear regression.

0 5 10 15 20 25 30 35 400

5

10

15

20

25

30

Etching time [min]

Dis

tanc

e [µ

m]

140nm/40%LR220nm/60%LR20nm/30%LR200nm/40%LR

0 5 10 15 20 25 30 35 400

10

20

30

40

50

60

Etching time [min]

Dis

tanc

e [µ

m]

140nm/40% LR220nm/60%LR20nm/30%LR200nm/40% LR

Etchant temperature: 50°C

Etchant temperature: 70°C(a)

(b)

Figure 7. Layers 20 nm/30% and 200 nm/40% were etched in a solution of phosphoric acid 85 wt% for 30 min at 50 °C and 10 min at 70 °C. (a) When layer 200 nm/40% was patterned, nanowires were formed on the sides of the structures (tilted view). (b) When layer 20 nm/30% was patterned, microstructures without nanowires on the sides were obtained (tilted view).

J. Micromech. Microeng. 25 (2015) 015017

M D Canonica et al

8

decreasing the underetch value and decreasing the slope of the linear regression.

3.2. Array of tips

Analysis of the morphologies of the etched regions, showed that etched surfaces with and without nanowires were pro-duced. For the fabrication of porous tips, it was decided to

process layers of 20 nm discs (20 nm/30% and 200 nm/40%) in order to obtain etched regions without nanowires, but also because these discs have the lowest etch rate, allowing a better control of the depth of the patterned structures.

Based on wet etching results, microstructures were obtained by etching the discs in two steps: 30 min at 50 °C and 10 min at 70 °C in phosphoric acid 85 wt%. Although, it could have been possible to etch the discs at 50 °C only, the second step

Figure 8. Micro-patterned PAA discs (tilted views). (a) Array of 384 tips, the diameter of the array is 1.5 mm. (b) Close-up view of multiple tips. (c) Single tip with a top plateau having a diameter of about 20 µm and a height of 50 µm. (d) Side of the top of single tip, pore size varies due to lateral etching. Top grey surface has pores with diameters of 20 nm. (e) Sharp tips were also obtained using this process and a different lithography mask. (f) Close-up view of a sharp tip, formed by the etching of a single capillary.

J. Micromech. Microeng. 25 (2015) 015017

M D Canonica et al

9

at higher temperature cleaned the etched regions. The time for these two steps was defined to obtain nanoporous tips with a top diameter and a height of about 40 to 50 µm. Knowing the etching process and underetching of the layer, the diameter, d, of the layout was calculated to be 45 µm for making sharp tips, if structures with plateaux are required, then, this diameter has to be increased.

First fabrication tests demonstrated that when layer 200 nm/40% was patterned, nanowires were produced contin-uously on the side of the structures due to lateral etching (see figure 7(a)). For the application, nanowires on the side of the microfeatures are not desired, therefore, since, this behavior was not observed for layer 20 nm/30% (see figure 7(b)), the structures were processed by etching this layer.

Different kinds of arrays were patterned using the fabrica-tion process described above. Figures 8(a)–(d) show an array of 384 nanoporous tips with a top diameter of 20 µm and a height of 50  µm. This array was obtained using the layout described in figure 2(a) with the following parameters, a diam-eter, d of 60 µm, a pitch, p, of 70 µm and a number of layer, N, of 12 to completely fill the 1.5 mm-diameter circle with struc-tures. By decreasing the diameter, d, of the layout to 40 µm, sharp tips were also obtained as shown in figures 8(e)–(f).

For most applications, uniformity of the structures is required, therefore, the height and the top diameter of the plateau of the structures of the chip from figures  8(a)–(d) were measured precisely by white light interferometry (Veeco Wyko NT5100). Out of the 384 tips, 48 were meas-ured: a mean of 49.0 µm with a standard deviation of 1.39 µm was measured for the height and a mean of 18.0 µm with a standard deviation of 1.30 µm was measured for the top diam-eters. The low values of the standard deviations demonstrate that uniform structures can be obtained by wet etching of randomly distributed porous alumina. However, if the PAA substrate is composed of defects generated during the ano-dization process, non-uniformities will appear locally in the etched regions and the uniformity will be affected.

Based on etch rates reported in table 2, the predicted depth of the etched cavities was calculated to be 38.1 µm. Thereby, a difference of 10.9 µm was obtained between calculated and measured depths of the structures. This difference could be explained because the surface of the etched region around the tips was not the same as the one of the tests, thereby, modi-fying the estimated underetch rates.

Also, the underetching for these structures can be calcu-lated; since a diameter of the circles, d, on the lithography mask was 60.0 µm and since the measured top diameter was 18  µm, the underetching was 21.0  µm. Based on underetch rates presented in table 2, the expected underetching for the two etch steps was 21.6 µm. A comparison between these two values shows that the underetch rates of table 2 are accurate for the design of porous alumina structures.

4. Conclusions

Commercially available PAA discs usually used as filters, with pore diameter of 200 nm and 20 nm, were characterized

and patterned by wet etching to make microfeatures com-posed of vertically aligned nanopores. Since each substrate was composed of two layers, the morphology of all four layers were characterized using electron microscopy and image processing tools and these layers demonstrated pores with a diameter ranging from 20 to 200  nm and porosities from 30–60%.

These porous alumina layers were patterned by wet etching in phosphoric acid using a mask of parylene. Etch and underetch rates were investigated for each layer and at two temperatures. Depending on the etching parameters, nanow-ires or nanotips were produced in the etched regions.

Based on wet etch results and by carefully selecting etching conditions, arrays of porous tips were designed and fabricated over large area of PAA substrates. Sharp tips of a few nanometers were processed as well as an array of 384 porous tips with height of 48 µm and pitch of 70 µm. All of these tips are composed of capillaries with a diameter of 200 nm, which can transport fluid from the back of the disc to the surface of the tips, something of great interest for electrospray, microfluidics applications, etc. Moreover, the low value of standard deviation for the height of these structures, shows that uniform porous alumina microfeatures can be obtained by this process. Finally, integration of these alumina microstructures with their nanopores on microsystems will contribute to devices with innovative function.

Acknowledgments

The authors would like to acknowledge the staff at the Micro-systems Technologies Laboratory (MTL) and at the Center for Materials Science and Engineering (CMSE) at MIT for their support. This work was supported by Lockheed Martin Cor-poration, through an MIT-Energy Initiative grant.

References

[1] Diggle J W, Downie T C and Goulding C W 1969 Anodic oxide films on aluminum Chem. Rev. 69 365–405

[2] Li A, Müller F, Birner A A, Nielsch K and Gösele U 1998 Hexagonal pore arrays with a 50–420 nm interpore distance formed by self-organization in anodic alumina J. Appl. Phys. 84 6023–6

[3] Lee W, Ji R, Ross C, Gösele U and Nielsch K 2006 Wafer-scale Ni imprint stamps for porous alumina membranes based on interference lithography Small 2 978–82

[4] Masuda H, Asoh H, Watanabe M, Nishio K, Nakao M and Tamamura T 2001 Square and triangular nanohole array architectures in anodic alumina Adv. Mater. 13 189–92

[5] Masuda H, Abe A, Nakao M, Yokoo A, Tamamura T and Nishio K 2003 Ordered mosaic nanocomposites in anodic porous alumina Adv. Mater. 15 161–4

[6] Masuda H and Fukuda K 1995 Ordered metal nanohole arrays made by a two-step replication of honeycomb structures of anodic alumina Science 268 1466–8

[7] Masuda H, Yamada H, Satoh M, Asoh H, Nakao M and Tamamura T 1997 Highly ordered nanochannelarray architecture in anodic alumina Appl. Phys. Lett. 71 2770–2

J. Micromech. Microeng. 25 (2015) 015017

M D Canonica et al

10

[8] Kim Y, Jung B, Lee H, Kim H, Lee K and Park H 2009 Capacitive humidity sensor design based on anodic aluminum oxide Sensors Actuators B 141 441–6

[9] Lee P S et al 2008 Microcantilevers with nanochannels Adv. Mater. 20 1732–7

[10] Kanamori Y, Hane K, Sai H and Yugami H 2001 100 nm period silicon antire ection structures fabricated using a porous alumina membrane mask Appl. Phys. Lett. 78 142–3

[11] Müller F et al 2002 High aspect ratio microstructures based on anisotropic porous materials Microsyst. Technol. 8 7–9

[12] Yook J M, Kim K M and Kwon Y S 2009 Air-cavity transmission lines on anodized aluminum for highperformance RF modules IEEE Microw. Wirel. Compon. Lett. 19 623–5

[13] Yeo S K, Kim K M, Chun J H and Kwon Y S 2009 Quasi-coaxial vertical via transitions for 3D packages using anodized aluminum substrates IEEE Microw. Wirel. Compon. Lett. 19 365–7

[14] He R and Kim C J 2006 A low temperature vacuum package utilizing porous alumina thin film encapsulation IEEE Int. Conf. on Micro Electro Mechanical Systems (Istanbul, Turkey, 12–15 January 2006) pp 126–9

[15] Hellin Rico R, Du Bois B, Witvrouw A, Van Hoof C and Celis J P 2007 Fabrication of porous membranes for MEMS packaging by one-step anodization in sulfuric acid J. Electrochem. Soc. 154 K74

[16] Songsheng T, Reed M, Han H and Boudreau R 1995 High aspect ratio microstructures on porous anodic aluminum oxide IEEE Int. Conf. on Micro Electro Mechanical Systems (Stockholm, Sweden, 25–29 June 1995) pp 267–72

[17] Jee S, Lee P, Yoon B J, Jeong S H and Lee K H 2005 Fabrication of microstructures by wet etching of anodic aluminum oxide substrates Chem. Mater. 17 4049–52

[18] Sulka G D 2008 Highly Ordered Anodic Porous Alumina Formation by Self-Organized Anodizing (Weinheim: Wiley) pp 1–1

[19] Shingubara S 2003 Fabrication of nanomaterials using porous alumina templates J. Nanopart. Res. 5 17–30

[20] Dickey M, Weiss E, Smythe E, Chiechi R, Capasso F and Whitesides G 2008 Fabrication of arrays of metal and metal oxide nanotubes by shadow evaporation ACS Nano 2 800–8

[21] Peng X and Chen A 2004 Electrochemical fabrication of novel nanostructures based on anodic alumina Nanotechnology 15 743–8

[22] Lombardi I, Magagnin L, Cavallotti P, Carraro C and Maboudian R 2006 Electrochemical fabrication of supported Ni nanostructures through transferred porous anodic alumina mask Electrochem. Solid-State Lett. 9 D13–6

[23] Yanagishita T, Nishio K and Masuda H 2005 Fabrication of metal nanohole arrays with high aspect ratios using two-step replication of anodic porous alumina Adv. Mater. 17 2241–3

[24] Chen P L, Kuo C T, Tsai T G, Wu B W, Hsu C C and Pan F M 2003 Self-organized titanium oxide nanodot arrays by electrochemical anodization Appl. Phys. Lett. 82 2796–8

[25] Courtney D, Li H and Lozano P 2013 Electrochemical micromachining on porous nickel for arrays of electrospray ion emitters J. Microelectromech. Syst. 22 471–82

[26] Xiao Z et al 2002 Fabrication of alumina nanotubes and nanowires by etching porous alumina membranes Nano Lett. 2 1293–7

[27] Otsu N 1979 A threshold selection method from gray-level histograms IEEE Trans. Syst. Man Cybern. 9 62–6

[28] Tsukamoto T and Ogino T 2012 Fabrication of three-dimensional porous alumina microstructures using imprinting method J. Electrochem. Soc. 159 C155–9

[29] Hu G, Zhang H, Di W and Zhao T 2009 Study on wet etching of AAO template Appl. Phys. Res. 1 78

J. Micromech. Microeng. 25 (2015) 015017