PAPER www.rsc.org/materials | Journal of Materials Chemistry

Publ

ishe

d on

24

Sept

embe

r 20

09. D

ownl

oade

d by

Uni

vers

ity o

f Il

linoi

s at

Chi

cago

on

31/1

0/20

14 0

2:58

:04.

View Article Online / Journal Homepage / Table of Contents for this issue

Two dimensional anisotropic etching in tracked glass

Pradeep Ramiah Rajasekaran,a Justin Wolff,a Chuanhong Zhou,a Mary Kinsel,a Christina Trautmann,c

Samir Aouadib and Punit Kohli*a

Received 3rd July 2009, Accepted 19th August 2009

First published as an Advance Article on the web 24th September 2009

DOI: 10.1039/b913151e

We describe in this paper the creation of a two-dimensional pore gradient using hydrofluoric acid (HF)

chemical etching of tracked glasses (TGs). The first gradient was along the plane of TGs where the pore

diameters of the conical pores were modulated, and a second gradient was formed in pores along the

axis of each pore. The 2-D pore gradient in TGs was characterized with optical and electron

microscopies. We demonstrate that the pore gradient was formed only when the TGs had a thin layer of

polydimethylsiloxane (PDMS) on their surface and when the etching solution was not stirred. The 2-D

pore gradient was also found to be dependent upon the TG orientation with respect to the etching

solution. Following the reaction between HF and PDMS, the resulting insoluble precipitate was

deposited, due to gravity, at the bottom of the vertical TG’s etching surface, accruing at the mouths of

the pores. This precipitate deposition at pore mouths appeared to hinder the diffusion of HF to the pore

surface. This build up also caused the retention of the by-products inside the pores which further

suppressed the etching of the glass. The etching process was inhibited more at the bottom of the chips

than that at the top presumably due to the formation of a precipitate gradient on TG surface. Using

a 2-D pore gradient containing TGs, many different experiments can be performed simultaneously

which will improve the throughput rate and aid analysis in many potential applications in materials and

life sciences.

Introduction

Track etching1 is a well-known technique for making pores in

various materials such as polyesters,2 nylons,3 polycarbonate

membrane4,5 and glass.6,7 Tracking is usually performed by

irradiating substrates with heavy ions accelerated to �15% the

speed of light using a linear particle accelerator. Other irradia-

tion methods include nuclear reactors, radioactive sources and

scanning ion microbeams. These treatments allow the construc-

tion of damaged tracks in materials.1 The interactions between

accelerated heavy ions and substrates result in the damage of

chemical bonds in the material along the direction of ion

movement. Following chemical etching, the nanometre dimen-

sional tracks in materials can be transformed into pores of

various shapes and sizes.8 Due to the presence of damaged tracks

in materials, the chemical etching along the tracks is much faster

than the etching in the lateral direction. This results in pore

formation without significant reduction in the substrate’s thick-

ness. Depending upon the etching conditions and properties of

the material being etched, the pores can be conical,9 cylin-

drical,10,11 diamond,12 or cigar in shape.13 In general, the isotropic

etching of tracks in non-crystalline materials will result in

cylindrical pores whereas conical pores are formed using aniso-

tropic etching of damaged tracks.1 The etching of damaged

aThe Nanoscience Exploration Research and Development Group (NERDGroup), Departments of Chemistry and Biochemistry, Southern IllinoisUniversity, Carbondale, IL, 62901, USA. E-mail: pkohli@chem.siu.edubDepartment of Physics, Southern Illinois University, Carbondale, IL,62901, USAcGSI Helmholtzzentrum f€ur Schwerionenforschung GmbH, Planckstr,Darmstadt, 64291, Germany

8142 | J. Mater. Chem., 2009, 19, 8142–8149

tracks in crystalline materials can result in pore shapes other than

those with a circular cross-section.14 Commercially, the track

etch technology has been used in the fabrication of nano- and

micron-size filters15 and in the synthesis of nanomaterials using

porous membranes as templates.16–18

One-dimensional anisotropic etching for conical pore forma-

tion with relatively uniform diameters in polymeric materials

using track-etch technology is well-studied by the groups of

Siwy,19–26 Martin,9,12,19,20,26–29 and Azzaroni.30,31 We report here

the formation of a two-dimensional pore gradient in TG chips. In

this paper, the diameters of the conical pores were varied along

the y-axis in the x–y plane, and a second pore gradient was

formed along the length of the conical pores (z-axis) (Fig. 1).

Although there are reports of one-dimensional conical pore

formation in TGs, to the best of our knowledge, a 2-D pore

gradient in track-etch glass is not reported in the literature. We

also propose a mechanism that explains the 2-D pore gradient

Fig. 1 The schematic of a 2-D pore gradient formed in tracked etched

glass after HF chemical etching. The diameter of the conical pores is

gradually changed along the y-axis in the x–y plane and a second pore

gradient was formed along the length of conical pores (z-axis).

This journal is ª The Royal Society of Chemistry 2009

Fig. 3 The schematic of the etching apparatus showing various

components. TG was held between two polycarbonate blocks with

O-rings, and the whole apparatus is fastened with a clamp (not shown) to

hold TG in place. One side of TG was in contact with Ca(NO3)2 whereas

its second side was in contact with HF. Two Pt electrodes were connected

to a digital multimeter. The inset on the top right shows the whole

assembled device.

Publ

ishe

d on

24

Sept

embe

r 20

09. D

ownl

oade

d by

Uni

vers

ity o

f Il

linoi

s at

Chi

cago

on

31/1

0/20

14 0

2:58

:04.

View Article Online

formation in the TG. Through the use of 2-D gradient containing

chips many different experiments can be performed simulta-

neously using one chip. This will improve analysis efficiency and

throughput in many potential biological and materials sciences

applications.

Experimental

All the glass substrates used in our experiments were �70 mm

thick and were purchased from SCHOTT (catalogue # D263).

The glasses were tracked with an ion fluence of 104–105 ions cm�2

at GSI (Darmstadt, Germany).

Chemical etching of tracked glass (TG) chips

A 24.5% HF solution was used for etching the TG in all cases

unless otherwise stated. Fig. 2 shows the assembly used

for etching the TG in our experiments. The TG was adhered

to a microscope glass slide (MG, 25 mm by 25 mm, thickness

� 1 � 0.1 mm, drilled hole �5 mm in diameter) with a poly-

dimethylsiloxane (PDMS) film �0.2–0.5 mm thick, to provide

structural support for the delicate TG. We refer to this assembly

as TG-PDMS-MG in our discussion. PDMS was also used as

a cushion to absorb stresses when the assembly was positioned

within a two U-half-cell apparatus. The etching was performed

on the open circular area (diameter � 5 � 0.5) that was perme-

able to the HF solution. Fig. 3 shows the etching apparatus

which consists of the two U-half cells. TG-PDMS-MG was

placed in between two solutions, a 24.5% HF in 0.5 M NaCl

etching solution in one half cell and a saturated Ca(NO3)2

blocking solution in the other half U-cell. There were two

configurations of the U-cell apparatus utilized in this experiment.

The configuration illustrated in Fig. 3, we call the vertical

configuration. Some experiments were also performed in which

the chip along with the apparatus was rotated 90� from the

vertical configuration. This arrangement we refer to as the

horizontal configuration. In the horizontal configuration, HF

and Ca(NO3)2 were poured into the top and bottom cells, above

and below the TG respectively. NaCl served as an electrolyte to

increase the ionic conductivity of the HF solution. Two Pt wires,

(Alfa-Aesar) functioning as electrodes, were inserted in the cells,

one in each cell, situated on both sides of the TG. The measured

Fig. 2 The assembly used for the fabrication of 2-D pore gradient in the

tracked glass (TG) substrates. TGs were adhered to a thin PDMS layer

with microscope glass (MG). TG was exposed to HF via a hole drilled

through MG of (5 � 0.5) mm diameter.

This journal is ª The Royal Society of Chemistry 2009

increase in the ionic current or decrease in the ionic resistance was

used to assess the progress of the etching process. These

measurements were obtained using either a Fluke 8840A Multi-

meter or a Keithley 6487 picoammeter/voltage source. Etching

was stopped when a spike in ionic current or a decrease in ionic

resistance was observed. These observed spikes indicated the

point in time when a continuous ionic path was detected between

the two electrodes; i.e., it is the time when the TG’s width was fully

penetrated across the TG by HF along the tracks. Once etching

was complete, HF was neutralized by washing the TG chips in

copious amounts of Ca(NO3)2. The TG was then rinsed thor-

oughly with deionized water and ethanol. Under these experi-

mental conditions, the etching time for our chips was�700–900 s

corresponding to an etching rate of �75–100 nm s�1. The char-

acterization of pore formation after chemical etching was per-

formed with scanning electron microscopy (SEM) using a Hitachi

570 instrument and an optical microscope (Lieca DMIRB)

equipped with a QImage (Cooled Mono 12-bit) CCD camera.

Water contact angle measurements were performed on a goni-

ometer (CAM-Micro, Tantec Inc.). Our reported values of the

contact angle were the averages of at least three measurements.

Simulation and experiment

Although the chips were made of a glass with high optical

transparency, when utilizing an inverted optical microscope in

bright field mode, the pores appeared as dark circles with smaller

brighter circles at the center. The larger darker and smaller

brighter circles in the optical micrographs correspond to the

larger and smaller diameters of the pore. Using the 2-D finite

difference time domain (FDTD) method, we performed simula-

tions to gain a better understanding of the observed high-

contrast optical micrographs of the conical pores.32

The propagation of the light through the glass chip satisfies

Maxwell’s equations which are expressed by:

V � E ¼ �vB/v t (1)

V � H ¼ 3vE/v t (2)

J. Mater. Chem., 2009, 19, 8142–8149 | 8143

Publ

ishe

d on

24

Sept

embe

r 20

09. D

ownl

oade

d by

Uni

vers

ity o

f Il

linoi

s at

Chi

cago

on

31/1

0/20

14 0

2:58

:04.

View Article Online

V � ¼ (iv/vx + jv/vy + kv/vz)x is curl operator; i, j and k are

orientation unit vectors; E and H are the electric and magnetic

fields of the light respectively; and B ¼ mH. m and 3 are

permeability and permittivity of glass with respect to the media

(air in our case). To make it easier to compute, we assume that

transverse magnetic (TM) polarized monochromatic light

(l ¼ 600 nm) is categorically incident on the larger open side of

the glass chip. Also in the simulation, we assumed a CCD

detector was fixed at the smaller pore side of the chip.

LDI-MS

A thin layer of the precipitate formed by the reaction between

TG, PDMS and HF, was deposited on a stainless steel target.

The target was inserted into the ion source of a Bruker Daltonics

Microflex LR time-of-flight mass spectrometer (Billerica, MA).

Laser desorption ionization (LDI) mass spectra were then

acquired by firing a pulsed nitrogen laser (337 nm, 20 Hz repe-

tition rate). The resultant positive ions were then recorded.

Results and discussion

The conical pores formation in the TG using anisotropic chem-

ical etching is well-established in literature.9,21,25 In the aniso-

tropic etching process, an enhanced etching rate was achieved on

one side of the TG (the etching side). On the contrary, a reduced

rate was obtained on the other side of the TG (the blocking side)

due to the neutralizing reaction between the etching and blocking

solutions (eqn (1)). In this paper, we report the preparation and

characterization of a 2-D anisotropic gradient in TG-PDMS-

MG. First, the gradient was established in the z-axis direction

where the pore diameter was uniformly modulated along the

thickness of the glass chip, and a second pore gradient was

created along the y-axis (Fig. 1). The smaller and larger diame-

ters of the conical pores can be controlled by means of altering

the etching conditions and by utilizing different tracked materials

with varying properties. We chose glass as our substrate because

it is a chemically, thermally and mechanically stable material.

Only a few chemicals such as HF and strong bases, for example,

Fig. 4 (A) shows the scanning electron micrograph (SEM) of a 2-D pore grad

area of TG exposed to HF (corresponding to the area of the drilled hole in th

bottom portions respectively of (A), zoomed-in to show the 2-D pore gradien

8144 | J. Mater. Chem., 2009, 19, 8142–8149

NaOH and KOH are known to chemically attack glass.33 In

particular, the reaction between glass (SiO2) and HF is extremely

fast due to attack by active species such as HF2� and HF on

glass34 (see eqn (2)).

The conical pore formation in the TG was easily inspected

with optical and electron microscopies. The gradient in the pore

diameters along the y-axis is shown in the SEM (Fig. 4) and

optical micrographs (Fig. 5). The top micrograph corresponds

to a top etched portion of the glass with a dl � (43 � 1) mm

whereas the bottom micrograph has a dl � (15 � 1) mm. Fig. 5B

shows the theoretical predictions of q of the pores formed in our

studies. The cone angle (q) that we calculated turned out to be

17.81�. We estimated q based on simple trigonometric calcula-

tions. For these calculations, it is assumed that (1) the HF

concentration and the etching rate are constant, (2) the larger

diameter dimension was assumed to be 45 mm; (3) the thickness

of the glass chip was 70 mm. These simple calculations provide

some guidance about cone angles of pores. q is highly dependent

upon the etching conditions (such as concentration and

temperature of the etching solution, humidity, etching time, and

the presence of impurities in the etching solution) and material

properties of the etching surface. We also note that q can be

accurately estimated only for break-through pores (i.e. both

smaller and larger pores are open). This is because the height of

the conical pores that are closed (at smaller openings) is hard to

determine using electron or optical micrographs under our

experimental conditions.

Fig. 6A and 6B show the optical micrograph and its surface

intensity plot respectively for a typical conical pore. The larger

pore appears as a dark circle whereas the smaller pore as

a smaller bright circle. At first, we were surprised to see a large

color contrast between the smaller and larger pores in the

optical micrographs because the whole chip is composed of

transparent glass, and it should be fully transparent to visible

light.

To gain more understanding regarding the optical micro-

graphs of the conical pores, we performed simulations by

solving Maxwell’s electromagnetic equations (Fig. 7). The

brighter and darker circles in the optical micrographs

ient containing tracked-etched glass. The black dotted line represents the

e MG). Figures (B), (C) and (D) are SEM images of the top, middle, and

t in TG.

This journal is ª The Royal Society of Chemistry 2009

Fig. 5 (A) The optical micrographs of conical pores fabricated using

2-D pore gradient technique. The Figure on the left shows the schematic

of an area of TG exposed to HF. The images on the right side are

a collection of optical micrographs from 5 areas of TG corresponding to

an area shown in the schematic in the left figure. The 50 micron scale bar

is common to all the micrographs. (B) Theoretical calculations for cone

angle with assumptions that the larger pore diameter is 45 mm, the small

side a couple 100 nm (just after ‘‘breakthrough’’ of the pores), and the

thickness of the TG did not change significantly after etching. The slant

height and the cone angles can be calculated using simple trigonometry

formulas. The cone angle (q), in general, is highly dependent on the ratio

of larger to smaller pore diameters.

Fig. 6 (A) shows is an optical micrograph of a conical pore open at both

ends using an inverted optical microscope in bright field mode. (B) shows

a 3-D optical surface intensity plot of the pore in (A).

Fig. 7 Simulation of light propagation through a hollow conical pore

using a 2-dimensional finite difference time domain (FDTD) method.

(A) shows the interference of the reflected light (J0K0 interface) from the

inside of the cone with incident light. (C) depicts the refracted light

(at J0K0 interface) and reflected light (KK0). (G) Incident light penetrating

through the pore (corresponds to the bright area in Fig. 6A). (D) Inter-

ference of light from area B (JJ0) and C (J0K0 and K0K interfaces).

(F) light from C is refracted by the KK0 plane, it corresponds to dark area

in the Fig. 6A.

This journal is ª The Royal Society of Chemistry 2009

Publ

ishe

d on

24

Sept

embe

r 20

09. D

ownl

oade

d by

Uni

vers

ity o

f Il

linoi

s at

Chi

cago

on

31/1

0/20

14 0

2:58

:04.

View Article Online

correspond to the smaller and larger pore diameters respectively

in the TG. Our simulations provided us detailed information on

this interesting phenomenon. Most of the incident light that

impinged on the pore was reflected or refracted by the inner wall

of the conical pore. The reflective light then interfered with the

incident light, which is shown in region A. The refractive light

was redirected into the glass (region C) with the refractive angle

qt satisfying Snell’s law: nisin(qi) ¼ ntsin(qt). ni and ni are

refractive indices of the incident and transmission medium

respectively, and qi is the incident angle. We also observed that

some of the incident light directly penetrated into the smaller

pore opening. In region G, a distinct diffraction pattern was

observed which corresponds to the smaller bright circles in our

experiments. In our simulation, we also see that some interfer-

ence patterns developed near the edges of the glass (i.e. near

points J and J0) due to the scattering caused by their sharp edges

and to the given boundary conditions used in the simulation.

This phenomenon was not seen in the optical micrographs

because experimentally there were no latent boundary condi-

tions such as sharp edges and planes, except at the chip corners.

In region D, the light transmitted normally, interfered with the

refractive light from region C, and this resulted in brighter

interference strips. Region D sparks special interest because its

simulated intensity is 2–3 times larger than the intensity in

region F and is consistent with our experimental results

regarding intensity (red circle at the edge of larger in Fig. 6B). F

is the dark circle region where the direction of the penetrating

light from region C was refracted away due to a ‘‘prism-like’’

effect of the conical pore. This ‘‘prism-like’’ effect led to

a decrease in the number of photons that reached the detector.

Ultimately, the high contrast between the dark and light areas

in the optical micrographs is a result of the interference due to

J. Mater. Chem., 2009, 19, 8142–8149 | 8145

Fig. 8 Comparison of pore diameter (dl)-versus-distance (D) from

bottom of TG after HF etching under different experimental conditions.

A significant 2-D pore gradient was observed only for TG-PDMS-MG

without stirring HF in vertical configuration. Each pore diameter is an

average of three measurements.

Publ

ishe

d on

24

Sept

embe

r 20

09. D

ownl

oade

d by

Uni

vers

ity o

f Il

linoi

s at

Chi

cago

on

31/1

0/20

14 0

2:58

:04.

View Article Online

the conical shape of the pore. It presents an easier, faster and

less expensive way for characterizing the conical pores in glass.

In our experiments, a reproducible 2-D anisotropic gradient

was only observed when the following two conditions were met:

(a) the presence of a thin PDMS film in between the TG and MG;

and (b) an unstirred HF solution during etching when the TG

was in vertical orientation. Fig. 8 shows four graphs comparing

pore diameter-versus-distance (D) starting from the bottom of

the TG etched area under four different experimental conditions:

(1) the etching of TG-PDMS-MG without stirring in vertical

configuration (blue squares); (2) the etching of TG-PDMS-MG

while stirring in the vertical configuration (red triangles); (3) the

etching of TG-PDMS-MG without stirring in the horizontal

configuration (green triangles); and (4) the etching of TG-3M

super 77-MG without stirring in the vertical configuration

(orange triangles).

The effect of stirring on the track-etched pore formation and

on the etching rate is previously described.35–37 In the present

study, the etching rate of the tracked substrate was found to

increase with stirring which is attributable to the removal of the

HF-PDMS reaction’s precipitate from the etching surface at

the HF–glass interface due to the solution’s circulation within the

U-half cell chamber. The pore gradient along the y-axis for

TG-PDMS-MG in (1) was�7.7 mm per 1 mm of TG. In contrast,

Fig. 9 Proposed mechanism of 2-D

8146 | J. Mater. Chem., 2009, 19, 8142–8149

we did not observe any pore gradient along the y-axis of the chip

for experiment (2) in which the etching solution was stirred.

These experiments suggest that a 2-D pore gradient in

TG-PDMS-MG in vertical configuration only occurs when the

etching solution was not stirred.

Similarly, the results of experiments (1) and (3) suggest that the

vertical configuration is necessary for the modulation of the

diameters of conical pores in the y-axis of the TG. Finally, to

evaluate the effect of adhesive chemistry on the etching rate of

the damaged tracks in glass substrates, we fabricated a TG glass

assembly with 3M Super 77 multi-purpose adhesive in place of

PDMS for sticking the TG to the MG (experiments (1) and (4)).

We chose this adhesive because its chemistry is very different

from PDMS. Since it does not contain siloxane polymer, it is

more resistant to HF attack than PDMS. HF is a weak acid, and

it reacts vigorously with Si containing materials due to the strong

affinity between Si and F38 (eqn (2)). The HF etching of the 3M

Super 77 containing TG did not show a uniform 2-D pore

gradient in the substrates. The pores were uniform throughout

the etching area (dl ¼ (13 � 1) mm) except at the adhesive–HF

interface where the pore diameter was somewhat smaller

(dl ¼ (8 � 1) mm). The dimensions of all the pores were smaller

because the etching was performed for only �250 s (about 1/3rd

of the total etching time). The reduced pore diameter at the

adhesive–HF interface was probably due to the hydrophobic

nature of the adhesive where the HF concentration was lowered.

More etching experiments were conducted in which the TG was

coated with a thin coating of wax. In these experiments, no

conical pore gradient was observed in the TG along the y-axis.

These experiments clearly indicate that the hydrophobicity of the

TG surface played a minor role in the formation of the 2-D pore

gradient.

We propose a mechanism to explain the formation of the 2-D

pore gradient in the PDMS containing TG after it was etched

with HF (Fig. 9). We propose that the TG surface exposed to HF

contained a thin film of PDMS. In addition, PDMS was also

present along the circular edge and in-between the TG and MG.

We argue that following the reaction between PDMS and HF

some of PDMS was etched away in the HF solution, and the

hydrophobic by-products of this reaction precipitated on TG.

Due to gravity, the precipitate was deposited at the mouth

openings of the pores and was built up against the HF exposed

TG. The precipitate blocked the diffusion of HF into the pores

and inhibited the PDMS-HF by-products from exiting out of the

pores. We further propose that a precipitate gradient formed

pore formation in tracked glass.

This journal is ª The Royal Society of Chemistry 2009

Table 1 Structural assignments of ion signals observed for the LDI ofthe PDMS-HF reaction by-products.

m/z Chemical structure

101.6

106.4

171.6

330.7

766.6

826.6

Publ

ishe

d on

24

Sept

embe

r 20

09. D

ownl

oade

d by

Uni

vers

ity o

f Il

linoi

s at

Chi

cago

on

31/1

0/20

14 0

2:58

:04.

View Article Online

along the y-axis in the opposite direction of the pore gradient.

Since the pore formation is a surface chemical reaction, it

appears that there is a increased restriction of HF to the pore

sites at the lower portion of the HF exposed TG than to the sites

present in the upper portion. The pore blockage and the degree of

blockage of HF resulted in the incrementally decreased chemical

etching rate in the lower portion of the exposed etching area of

the TG, and this, we suggest, contributed to the creation of the

pore gradient along the y-axis.

To further test our hypothesis that the pore gradient was

formed due to the presence of PDMS film on glass, we treated

TG-PDMS-MG first with O2 plasma39 (for 10 min, 100 W power,

and flow rate of O2 40 sccm using a FA2000 Reactive Ion Etcher

from BSETEQ Inc.) followed by five H2SO4 washings (18 M

H2SO4 and 10 s each). Since both of these treatments are known

to chemically attack polysiloxane,40 we expected a decrease or

elimination of a pore gradient along the y-axis as a result of these

treatments. Indeed, we observed a relatively uniform pore

distribution (dl � (45 � 10) mm and ds � (5 � 1) mm) over the

whole etching area of O2-plasma/H2SO4 treated TG. These

observations reinforce our hypothesis that the PDMS coating

contributed to the formation of 2-D pore formation in TGs.

Additional evidence regarding the contribution of PDMS to

the 2-D pore gradient formation came from the experiments

involving different PDMS film thicknesses in the TG-PDMS-

MG assemblies. Thicker PDMS adhesive films were expected to

generate a larger 2-D pore gradient in the TG than for thinner

PDMS adhesive films because of the increased amount of

precipitate formation with thicker PDMS films. We fabricated

two different TG-PDMS-MG assemblies that contained PDMS

thicknesses of (0.5 � 0.1) mm and (0.2 � 0.1) mm. As expected,

a larger pore gradient formed when utilizing thicker PDMS

layered samples as compared to a thinner PDMS layer on the TG

(Fig. 10).

Finally, we also tested the chemical composition of the

precipitate formed from the reaction between TG-PDMS-MG

and HF. For this, we synthesized the precipitate by reacting an

assembly made up of TG, PDMS and microscopic cover slide in

Fig. 10 Comparison of 2-D pore gradients formed in TG with PDMS of 2 different thicknesses. A larger pore gradient in y-axis was seen for TGs with

thicker PDMS (0.5 � 0.05 mm) than for a TG with thinner PDMS film (0.2 � 0.05 mm). Each pore diameter was an average of three measurements.

This journal is ª The Royal Society of Chemistry 2009 J. Mater. Chem., 2009, 19, 8142–8149 | 8147

Publ

ishe

d on

24

Sept

embe

r 20

09. D

ownl

oade

d by

Uni

vers

ity o

f Il

linoi

s at

Chi

cago

on

31/1

0/20

14 0

2:58

:04.

View Article Online

a mass ratio of 1 : 0.7 : 0.3 with a 24.5% HF solution overnight.

The major product of the reaction between glass and HF is

H2SiF6 which is water soluble. Thus, it did not affect the etching

of the TG. Table 1 shows the composition of the precipitate

formed by the reaction between the assembly and HF using mass

spectrometric analysis (also see Fig. 11). The major component

of the precipitate was composed of silicon, methyl and fluorine

containing oligomers. The mass spectrometry and water solu-

bility analysis of the precipitate suggested that the precipitate was

hydrophobic in nature. Fig. 11 shows a representative LDI mass

spectrum in the mass-to-charge (m/z) range of 100 to 2000

formed by the summation of 1000 laser shots. The ion formation

efficiency in LDI rapidly decreases as analyte molecular masses

increase above 2000 amu. This behavior was also observed in this

experiment, where no ion signals were observed at m/z > 2000.

Much larger molecular mass oligomers may be present in the

precipitate but would not be desorbed and ionized using this

technique. Table 1 summarizes the major ions observed in the

low and high mass-to-charge regions, 100 < m/z < 600 and

600 < m/z < 1500, respectively. The ion signals appearing at

m/z 101.6, 106.4, 171.5 and 330.7 were assigned to small sodiated

(or potassiated) oligomers composed of silicon, methyl,

hydrogen, oxygen and fluorine. A short series of ion signals was

observed between m/z 766.5 and 1035.5, and was assigned to

sodium- and potassium-containing linear poly(dimethyl, meth-

ylhydrogen)siloxane (n ¼ 5 to 7). The most abundant ion signal

appearing at m/z 766.0 belonged to the sodiated linear series

member with n ¼ 5. A longer series of ion signals was observed

between m/z 692.6 and 1244.9. These ion signals were assigned to

sodium and potassium cyclic poly(dimethyl, methylhy-

drogen)siloxane (n¼ 5 to 9), the most abundant ion signal in this

series at m/z 826.5 belonging to the sodiated cyclic series member

with n ¼ 6.

The water-insoluble precipitate from the TG-PDMS-MG and

HF reaction was also filtered out and was deposited over an

�2 cm2 area of a TG. The etching of this precipitate-covered TG

showed a 2-D pore formation only in the area that was not

Fig. 11 A representative LDI mass spectrum with mass-to-charge (m/z) ran

observed between m/z � 766.6 and 1035.6 with an average repeat unit of m/z

8148 | J. Mater. Chem., 2009, 19, 8142–8149

exposed to the precipitate. The etching on the area covered with

precipitate was retarded as evidenced by the comparison of the

pore diameters in the two regions of the TG. Furthermore,

a uniform 2-D gradient was not observed in this precipitate

devoid area. These results conclusively indicated that a PDMS

coating is needed for the 2-D pore gradient formation in

damaged tracked glass. We also note that glass used in the

experiments also contained small quantities of metal oxides like

BaO, CaO, ZnO etc. These metal-oxides form sparingly soluble

metal fluorides upon reaction with HF. We cannot rule out the

minor contributions these metal fluorides might have made to the

formation of the 2-D pore gradient in the TG.

Now the question is, how siloxane polymer film is formed on

the TG surface? Sylgard 184 (Dow Corning) is a two-part silicone

elastomer, and it contains a number of low-molecular weight

oligomers and high vapor pressure monomers. We believe that

some of siloxane monomers and oligomers formed a thin film

during TG-PDMS-MG fabrication, especially during curing

time when the temperature was �90–100 �C. At these tempera-

tures, it is possible that some monomers and oligomers diffused

across the TG surface resulting in the formation of a thin film of

PDMS. Our water contact angle on the TG was (65� 15)0. These

measurements suggested a hydrophobic TG which is consistent

with our argument for the presence of a PDMS film on the TG.

Finally, 2-D pore gradient substrates can be used in many

potential applications where a conical pore gradient is needed.

This paper provides a fast and simple fabrication method for

making cones of different dimensions in the tracked glass,

a capability which was previously not available to the scientific

community. The conical pores have been used in a wide range of

applications including bioseparation, biosensing, nanofluidic

diodes and rectifiers, and artificial ion channels and ion

pumps.22–29,41,42 In all of these experiments, either a single pore or

pores of uniform diameters were used. In principle, with the glass

chips fabricated in the present study, many different experiments

can be performed simultaneously using one single chip. Thus,

these chips are expected to increase the throughput rate of

ging between 100 and 2000. The inset shows a short series of ion signals

� 134.0 �0.2.

This journal is ª The Royal Society of Chemistry 2009

Publ

ishe

d on

24

Sept

embe

r 20

09. D

ownl

oade

d by

Uni

vers

ity o

f Il

linoi

s at

Chi

cago

on

31/1

0/20

14 0

2:58

:04.

View Article Online

analysis and may also help in reducing the cost of analysis in

biological and materials sciences.

In conclusion, we demonstrated the fabrication of glass

substrates with pore gradients in two directions. The 2-D pore

gradient was characterized with optical and electron micros-

copies. Our experiments indicated that a 2-dimensional pore

gradient was formed due to the blocking of the pore mouths in

the tracked-glass substrates by a precipitate formed by a reaction

between PDMS and HF. We proposed that a precipitate gradient

was formed in the TG which led to a pore gradient direction

along the y-axis.

Acknowledgements

We acknowledge financial support from the National Institute of

Health and the National Science Foundation. We also

acknowledge the Southern Illinois University Carbondale Mass

Spectrometry Facility. Instrumentation in the Mass Spectrom-

etry Facility was obtained with support of the National Science

Foundation Grant No. 0405819 and the State of Illinois. We

thank Rashid Zakeri (Indiana University) for his assistance in

chemical etching of glass and Drs. John Bozzola and Steve

Schmitt and Ms. Hillary Gates (IMAGE center at SIUC) for

electron microscopy analysis.

References

1 Nuclear Tracks in Solids. Principles and Applications, ed. R. L.Fleischer, P. B. Price and R. M. Walker, University of CaliforniaPress, Berkeley, 1975.

2 P. C. Kalsi, P. D. Sawant, A. Ramaswami and V. K. Manchanda,J. Radioanal. Nucl. Chem., 2007, 273, 473–477.

3 R. Kumar, R. Prasad, Y. K. Vijay, N. K. Acharya andV. K. C. Udyan, Nucl. Instrum. Methods Phys. Res., Sect. B, 2003,212, 221–227.

4 P. Makphon, W. Ratanatongchai, S. Chongkum, S. Tantayanon andP. Supaphol, J. Appl. Polym. Sci., 2006, 101, 982–990.

5 J.-F. Allaeys, B. Marcilhac and J.-C. Mage, J. Phys. D: Appl. Phys.,2007, 40, 3714–3717.

6 A. Dallanora, T. L. Marcondes, G. G. Bermudez, P. F. P. Fichtner,C. Trautmann, M. Toulemonde and R. M. Papal�eo, J. Appl. Phys.,2008, 104, 024307.

7 G. S. Randhawa and H. S. Virk, Appl. Radiat. Isot., 1996, 47(3), 351–354.

8 R. L. Fleischer, P. B. Price and R. T. Woods, Phys. Rev., 1969, 188,563.

9 P. Scopece, B. A. Lane, P. Ugo and C. R. Martin, Nanotechnology,2006, 17, 3951–3956.

10 K.-L. Tung, Y.-L. Chang and C.-J. Chuang, Tamkang J. Sci. Eng.,2001, 4(2), 127–132.

11 P. Djardin, E. N. Vasina, V. V. Berezkin, V. D. Sobolev andV. I. Volkov, Langmuir, 2005, 21, 4680–4685.

This journal is ª The Royal Society of Chemistry 2009

12 S. Lee, Y. Zhang, H. S. White, C. C. Harrell and C. R. Martin, Anal.Chem., 2004, 76, 6108–6115.

13 P. Y. Apel, I. V. Blonskaya, A. Y. Didyk, S. N. Dmitriev,O. L. Orelovitch, D. Root, L. I. Samoilova and V. A. Vutsadakis,Nucl. Instrum. Methods Phys. Res., Sect. B, 2001, 179, 55–62.

14 Xu Fan, John E. Wharton and Charles R. Martin, Small, 2007, 3,1718–1722.

15 http://www.2spi.com/catalog/spec_prep/filter3.html (accessed on 06/26/090).

16 (a) C. R. Martin, Science, 1994, 266, 1961–1966; (b) C. R. Martin andP. Kohli, Nat. Rev. Drug Discovery, 2003, 2, 29–27; (c) E. Gatebe,H. Herron, R. Zakeri, R. R. Pradeep, S. Aouadi and P. Kohli,Langmuir, 2008, 24, 11947–11954.

17 J.-R. Ku, S.-M. Lai, N. Ileri, P. Ramı́rez, S. Mafe and P. Stroeve,J. Phys. Chem. C, 2007, 111, 2965–2973.

18 K.-Y. Chun, S. Mafe, P. Ramırez and P. Stroeve, Chem. Phys. Lett.,2006, 418, 561–564.

19 Z. Siwy, L. Troffin, P. Kohli, L. A. Baker, C. Trautmann andC. R. Martin, J. Am. Chem. Soc., 2005, 127, 5000–5001.

20 C. R. Martin and Z. S. Siwy, Science, 2007, 317, 331.21 S. Howorka and Z. S. Siwy, Chem. Soc. Rev., 2009, 38, 2360–2384.22 Z. S. Siwy, M. R. Powell, A. Petrov, E. Kalman, C. Trautmann and

R. S. Eisenberg, Nano Lett., 2006, 6, 1729.23 C. C. Harrell, P. Choi, L. P. Horne, L. A. Baker, Z. S. Siwy and

C. R. Martin, Langmuir, 2006, 22, 10837–10843.24 I. Vlassiouk and Z. S. Siwy, Nano Lett., 2007, 7, 552.25 Z. S. Siwy and A. Fulin, Phys. Rev. Lett., 2002, 89, 198103.26 E. A. Heins, Z. S. Siwy, L. A. Baker and C. R. Martin, Nano Lett.,

2005, 5(9), 1824–1829.27 P. Kohli, C. C. Harrell, Z. Cao, R. Gasparac, W. Tan and

C. R. Martin, Science, 2004, 305, 984.28 J. E. Wharton, P. Jin, L. T. Sexton, L. P. Horne, S. A. Sherrill,

W. K. Mino and C. R. Martin, Small, 2007, 3(8), 1424–1430.29 L. T. Sexton, L. P. Horne, S. A. Sherrill, B. W. Gregory, L. A. Baker

and C. R. Martin, J. Am. Chem. Soc., 2007, 129, 13144–13152.30 M. Ali, B. Yameen, R. Neumann, W. Ensinger, W. Knoll and

O. Azzaroni, J. Am. Chem. Soc., 2008, 130, 16351–16357.31 B. Yameen, M. Ali, R. Neumann, W. Ensinger, W. Knoll and

O. Azzaroni, J. Am. Chem. Soc., 2009, 131, 2070–2071.32 Computational Electrodynamics: The Finite-Difference Time-Domain

Method, A. Taflove, Artech House, Boston, MA, 1995.33 K. Kim, S. K. Dhungel, S. Jung, D. Mangalaraj and J. Yi, Sol. Energy

Mater. Sol. Cells, 2008, 92, 960–968.34 J. S. Judge, J. Electrochem. Soc., 1971, 118, 1772.35 E. Manea; I. Cernica; M. Lupu; C. Podaru., Semiconductor

Conference 28 Sept.–2 Oct. 2003, CAS 2003, vol. 1, 204, 2003.36 S. Ndoni, E. Martin, M. E. Vigild and R. H. Berg, J. Am. Chem. Soc.,

2003, 125, 13366–13367.37 J. P. Y. Ho, C. W. Y. Yip, D. Nikezic and K. N. Yu, Radiat. Meas.,

2003, 36, 141–143.38 G. A. C. M. Spierings, J. Mater. Sci., 1993, 28, 6261–6273.39 D. Bodas and C. Khan Malek, Microelectron. Eng., 2005, 83, 4–9.40 J. N. Lee, C. Park and G. M. Whitesides, Anal. Chem., 2003, 75, 6544–

6554.41 N. Li, S. Yu, C. C. Harrell and C. R. Martin, Anal. Chem., 2004, 76,

2025–2030.42 T. G. Knudstrup, I. A. Bitsanis and G. B. Westermann-Clark,

Langmuir, 1995, 11, 893.

J. Mater. Chem., 2009, 19, 8142–8149 | 8149