Proton beam micromachining on PMMA, Foturan and CR-39 materials

Nuclear Instruments and Methods in Physics Research B 210 (2003) 260–265

www.elsevier.com/locate/nimb

Proton beam micromachining on PMMA, Foturanand CR-39 materials

I. Rajta a, I. G�oomez-Morilla b, M.H. Abraham b, �AA.Z. Kiss a,*

a Institute of Nuclear Research of the Hungarian Academy of Sciences, P.O. Box 51, Debrecen H-4001, Hungaryb Department of Materials, University of Oxford, Parks Road, Oxford OX1 3PH, UK

Abstract

In this paper we investigate further the potential of proton beam micromachining (PBM) on three different materials:

the polymers PMMA and CR-39, and the photowritable glass Foturan. A focused beam of 2 MeV protons delivered by

the nuclear microprobe of ATOMKI was used to pattern these materials. The parameters of PBM and the obtained

structures are presented.

� 2003 Elsevier B.V. All rights reserved.

PACS: 87.80.Mj

Keywords: Proton beam micromachining; PMMA; Foturan; CR-39

1. Introduction

Proton beam micromachining (PBM) is a noveldirect-write process for the production of 3D mi-

crostructures, which has been pioneered in Oxford

[1] and Singapore [2], and it is gradually spreading

into other nuclear microprobe facilities [3] in-

cluding Debrecen. At Debrecen we have per-

formed PBM irradiations on three types of

materials: PMMA, Foturan, and CR-39. PMMA,

a high density polymethylmethacrylate, has al-ready been used for various experiments in Oxford

[4] and Singapore [5]. Its irradiation parameters

were reported by Kan et al. [6] for the first time.

Foturan [7] is a photosensitive glass (composition:

* Corresponding author. Tel.: +36-52-417266; fax: +36-52-

416181.

E-mail address: [email protected] (A.Z. Kiss).

0168-583X/$ - see front matter � 2003 Elsevier B.V. All rights reser

doi:10.1016/S0168-583X(03)01025-5

B2O3, CeO2, Sb2O3, Ag2O<1%, K2O 1–20%,

SiO2 > 80%, Al2O3, Na2O, ZnO, Li2O 1–10%;

q ¼ 2:37 g/cm3) which has been applied to createstructures by UV induced lithography [8] or direct

write laser fabrication [9]. The potential of this

glass is enhanced by the fact that it can be applied

in corrosive and high temperature environments

which is of major importance for applications in

chemistry and biology. Its temperature stability

and chemical resistance are notably higher than

those parameters for plastics. The use of Foturanas a material for proton beam micromachining has

been developed at Oxford [10–12] and a systematic

investigation of its properties is in progress. CR-39

is a thermoset polymer (allyl-diglycol-carbonate,

C12H18O7, q ¼ 1:31 g/cm3), and it is widely used in

applied nuclear physics as the basic material for

Solid State Nuclear Track Detectors (SSNTD).

CR-39 also seemed to be promising material forPBM as has been demonstrated in our preliminary

ved.

I. Rajta et al. / Nucl. Instr. and Meth. in Phys. Res. B 210 (2003) 260–265 261

report [13], and its easy availability and cheapness

could be interesting for further applications. All of

these materials are positive resists.

The aim of this work is to compare the abovethree materials in equal PBM conditions, and at

the same time to investigate the performance and

capabilities of our facility in proton induced mi-

cromachining technique.

2. Experimental

All irradiations in this paper have been per-

formed on the nuclear microprobe facility at

ATOMKI, Debrecen, Hungary [14]. The proton

energy was 2 MeV. Beam currents of 5–40 pA were

focused down to 1–2 lm spotsize. The scan size

was varied between 250 and 1000 lm. The re-

quired dose for PMMA, CR-39 and Foturan are

100 nC/mm2 (¼ 6.3� 1013 p/cm2) [2], 1–3� 1014 p/cm2 [13] and 1 nC/mm2 [10], respectively.

The delivered dose was measured using the

PIXE signals from the samples. Since all the irra-

diated samples consist of light elements, an Ultra

Thin Windowed (UTW) Si(Li) detector was used,

and we integrated over the whole spectrum which

consisted of mainly carbon and oxygen K lines. In

order to do this, first we calibrated the UTW de-tector for charge. A spectrum was then collected

during each irradiation, and the total counts were

compared to the calibrated beam chopper current

integrator [15]. The beam chopper itself, which is

generally used for monitoring the beam in other

applications of our scanning nuclear microprobe,

could not be used here because its beam cutting

frequency matched the scanning frequency in sucha way that it produced regularly spaced inhomo-

Table 1

Typical beam and scanning parameters

PMMA CR- 3

Pixel resolution 1024 1024 1024 1024

Beam current (pA) 10 10 10 10

Scan size (mm2) 1 0.25 0.0625 1

Dose (nC/mm2) 100 100 100 300

Time (s) 10,000 2500 625 30,000

Sweeps 10 10 3 10

Pixel dwell time (ms) 0.95 0.24 0.20 2.86

geneity in our early PBM products. Furthermore,

the RBS yield from the wings of the chopper was

not high enough. The collected PIXE total counts

were in the range of 1000–10,000 counts, thismeans 1–3% statistical error, which is better than

the original charge calibration of the chopper. All

of these counts came from the sample, the noise

was first eliminated by a lower level discriminator,

and we did not receive any counts with the beam in

the chamber in the absence of a sample.

The beam scanning was done using a National

Instruments (NI) card (model 6711), and the newC++ version of the program IonScan, developed

specifically for PBM applications called IonScan

2.0 [16]. Each sample has been scanned over sev-

eral times to eliminate the effect of possible insta-

bilities in the beam intensity. A home made

electrostatic beam blanking system was used to

eliminate unwanted beam exposure. The IonScan

software produces the blanking signal via the NIcard and then this blanking signal drives the home

made amplifier that rises and eliminates the volt-

age on the plates. Both the rise time and fall time

of the voltage on the plates are less than 1 ls,which is negligible compared to the pixel dwell

time (see Table 1). The dwell time was in the range

of 100 ls and 1 ms, it is adjustable in IonScan.

The scanning of the beam was typically doneusing the parameters shown in Table 1. Our

scanned structures were usually about 3% of the

total number of pixels, and we scanned at the same

speed, therefore the irradiation times were about

5 min each. At smaller scan sizes we had to reduce

the total accumulated charge to deliver the same

dose. In order to do this we did not want to reduce

the pixel dwell time for two reasons: the dwell timewould be comparable to the blanking rise and fall

9 Foturan

1024 1024 1024 1024 1024

10 10 5 5 5

0.25 0.0625 1 0.25 0.0625

300 300 1 1 1

7500 1875 200 50 12.5

10 10 3 3 1

0.72 0.18 0.06 0.016 0.012

Fig. 1. Optical microscope pictures before development (a) CR-

39, (b) PMMA and (c) Foturan.Fig. 2. Optical microscope pictures after development (a) CR-

39, (b) PMMA and (c) Foturan.

262 I. Rajta et al. / Nucl. Instr. and Meth. in Phys. Res. B 210 (2003) 260–265

time, and we would have introduced hysteresis.

Thus we tried to reduce the beam current but it

became more unstable. Finally, we decided to run

fewer sweeps at a more stable beam intensity (three

sweeps would still eliminate small fluctuations).

However, in a few cases during the Foturan irra-

diations it was necessary to reduce the dwell time

to the 10 ls range, and reduce the number of

sweeps even further.

I. Rajta et al. / Nucl. Instr. and Meth. in Phys. Res. B 210 (2003) 260–265 263

After the irradiations Differential Interference

Contrast (DIC) microscopy was performed, which

is sensitive for the refractive index change in the

samples. Using this technique we were able to seethe irradiation tracks before etching. In Fig. 1

some structures are seen after irradiation but be-

fore etching, while in Fig. 2 the same structures are

shown after development.

Exposed PMMA samples were etched using the

procedure and recipe described in [2]. The proce-

dure used to develop (anneal) the Foturan glass

was the same one used in UV lithography [8], it wasthen followed by the standard etch protocol. The

CR-39 etch was done using the method that has

been routinely applied over the last 20 years of ex-

perience in the nuclear track counting technique [17].

Scanning electron microscope (SEM) images

were taken on our microstructures applying 5 kV

acceleration voltage, and secondary electron sig-

nals to generate the images. The samples were goldcoated to avoid charging.

Fig. 3. Electron microscope pictures after chemical develop-

ment (a) PMMA and (b) Foturan.

3. Results

Fig. 1(a) shows a resolution test on CR-39. The

irradiation pattern was designed to see how many

lines we can separate after irradiation, and whetherthese lines are still separated after etching. We con-

cluded that the etching process did not ruin the

structures, they were equally good quality before

and after chemical etching for CR-39 and PMMA

as well.

Fig. 1(b) shows a detail of a structure similar to

a ring resonator. The width of the gap between the

straight guide and the oval is crucial for such ap-plications. The sizes are given below (see the

comments after Fig. 2(b)).

Fig. 1(c) shows the same structure in Foturan.

The structure was not visible immediately after the

irradiation, as was the case on other samples, but

the first step of the annealing had to be done. No

chemical was used, the structure has not been

etched yet, it only changed the effective refractiveindex by forming a crystalline ceramic state at the

exposed areas, which was visible in DIC micro-

scopy. This ceramic is etched more rapidly than

the gas phase etching in hydrofluoric acid.

The pictures on Fig. 2 were taken after thechemical development on the same structures that

were shown on Fig. 1. Concerning the resolution

no differences are seen between the structures

before and after development. However there is a

difference between the resist materials PMMA

and Foturan in the case of ‘‘ring resonator’’

structure. In PMMA the straight and curved lines

are separated both before and after development.(Figs. 1(b) and 2(b)) while in Foturan they are

overlapping in both cases (Figs. 1(c) and 2(c)).

Both ‘‘ring resonator’’ structures are the same

size, but the patterns used to make them are

slightly different – the gap is 4 pixels (3.9 lmapproximately) wide in the structure made in

264 I. Rajta et al. / Nucl. Instr. and Meth. in Phys. Res. B 210 (2003) 260–265

Foturan and 10 pixels (9.8 lm approximately)

wide in the one made in PMMA, while the width

of both the straight and curved lines are the same

in both cases (4 pixels). With a smaller beam spotsize it might be possible to separate these lines in

both patterns.

The irradiations on Figs. 1(a) and 2(a) were

done using the OMDAQ software [18] for scan-

ning the beam, which has its drawbacks of rela-

tively low scan resolution (256� 256 pixels), and

raster scanning type pattern generation. The fast

scan direction was vertical on these samples, andthe unevenness on the two horizontal lines shows

the problems due to hysteresis. However, the rest

of the samples were scanned as described above

using the NI card and IonScan 2.0 with vector

scanning and up to 4096� 4096 resolution (we

used 1024� 1024 pixels).

Some of the typical SEM images are presented

on Fig. 3. (Note, the rough surface on Fig. 3(a) canbe eliminated using a different type of PMMA.)

The ranges in each material were calculated using

the TRIM code [19], and the SEM images showed

the same depth after etching. The depth of the

resulted patterns in each material corresponds to

the calculated ranges 63, 40 and 59 lm for

PMMA, Foturan and CR-39, respectively.

4. Conclusions

Proton beam micromachining was demon-

strated at the Institute of Nuclear Research of the

Hungarian Academy of Sciences using three

different types of resists: PMMA, Foturan and

CR-39 type Solid State Nuclear Track Detectormaterial. These materials show refractive index

change which may indicate that they can be used in

microphotonics without further development. In

the case of Foturan the issue of light absorption

and scatter in the polycrystaline material still needs

exploration. Conversely they can be chemically

developed for other applications (e.g. microelec-

tromechanical systems). The chemical develop-ment is relatively simple for CR-39 and PMMA,

however Foturan requires hydrofluoric acid as a

developer, which makes it a rather dangerous

process. It is interesting to note that UTW PIXE

was used for dose normalization, this gave us more

yield than a large RBS detector would have done,

because the polymer samples consist of light ele-

ments for which UTW PIXE has large sensitivity,and the RBS yield is rather low.

Acknowledgements

The authors are indebted to Dr. I. Hunyadi for

assistance in CR-39 development, Mrs. M. Mo-

gyor�oosi for assistance in PMMA development andDr. Cs. Cserh�aati for assistance in DIC microscopy.

This project was supported by the Hungarian

National Research Foundation OTKA (grant nos.

A080 and partly by T032264 and F042474). One of

the authors (I.R.) is a Bolyai fellow.

References

[1] M.B.H. Breese, G.W. Grime, F. Watt, Nucl. Instr. and

Meth. 77 (1993) 169.

[2] S.V. Springham, T. Osipowicz, J.L. Sanchez, L.H. Gan, F.

Watt, Nucl. Instr. and Meth. 130 (1997) 155.

[3] P. Polesello, C. Manfredotti, F. Fizzotti, R. Lu, E. Vittone,

G. Lerondel, A.M. Rossi, G. Amato, L. Boarino, S.

Galassini, M. Jaksic, Z. Pastuovic, Nucl. Instr. and Meth.

B 158 (1999) 173.

[4] D.G. de Kerckhove, M.B.H. Breese, M.A. Marsh,

G.W. Grime, Nucl. Instr. and Meth. 136–138 (1998)

379.

[5] F. Watt, Nucl. Instr. and Meth. 158 (1999) 165.

[6] J.A. van Kan, J.L. Sanchez, B. Xu, T. Osipowicz, F. Watt,

Nucl. Instr. and Meth. B 158 (1999) 179.

[7] MGT Mikroglas Technik AG, Mainz, Germany.

[8] T.R. Dietrich et al., Microelectron. Eng. 30 (1996)

497.

[9] M. Abraham et al., MRS 2002 Fall Meeting Symposium

H: Three-Dimensional Nanoengineered Assemblies.

[10] I. G�oomez-Morilla, M.H. Abraham, G.W. Grime, in

preparation.

[11] W. Hanesen, P. Fuqua, F. Livingston, A. Huang, M.

Abraham, D. Taylor, S. Janson, H. Helvajian, The

Industrial Physicist (June/July) (2002) 18.

[12] I. G�oomez-Morilla, D.G. de Kerckhove, M.H. Abraham,

G.W. Grime, 3rd Conf. on Postgrad. Res. in Electronics,

Photonics, Communications and Software, 9–11 April

2001, Keele, UK.

[13] I. Rajta, L. Bartha, I. Uzonyi, I. Hunyadi, Sz. Szilasi, �AA.

Z. Kiss, Proc. 16th European Conf. on Solid-State

Transducers, 15–18 September 2002, Prague, Czech

Republic.

I. Rajta et al. / Nucl. Instr. and Meth. in Phys. Res. B 210 (2003) 260–265 265

[14] I. Rajta, I. Borb�eely-Kiss, Gy. M�oorik, L. Bartha, E. Koltay,�AA.Z. Kiss, Nucl. Instr. and Meth. B 109–110 (1996) 148.

[15] I. Uzonyi, L. Bartha, Nucl. Instr. andMeth. 161 (2000) 339.

[16] A.A. Bettiol, J.A. van Kan, T.C. Sum, F. Watt, Nucl. Instr.

and Meth. 181 (2001) 49.

[17] Gy. Somogyi, I. Hunyadi, R. Ili�cc, A. Loose, Zs. Varga,

Nucl. Tracks Radiat. Meas. 8 (1984) 61.

[18] G.W. Grime, M. Dawson, Nucl. Instr. and Meth. B 104

(1995) 107.

[19] J.F. Ziegler, TRIM2000 private communication.