Microelectronics Journal 37 (2006) 519–525
`
`www.elsevier.com/locate/mejo
`
`Etch characteristics of KOH, TMAH
`and dual doped TMAH for bulk micromachining of silicon
`
`K. Biswas*, S. Kal
`
`Microelectronics Laboratory, Advanced Technology Centre/E & ECE Department, Indian Institute of Technology, Kharagpur 721302, India
`
`Received 20 May 2005; received in revised form 25 July 2005; accepted 27 July 2005
`Available online 2 November 2005
`
`Abstract
`
`High precision bulk micromachining of silicon is a key process step to shape spatial structures for fabricating different type of
`microsensors and microactuators. A series of etching experiments have been carried out using KOH, TMAH and dual doped TMAH at
`different etchant concentrations and temperatures wherein silicon, silicon dioxide and aluminum etch rates together with !100O silicon
`surface morphology and !111O/!100O etch rate ratio have been investigated in each etchant. A comparative study of the etch rates and
`etched silicon surface roughness at different etching ambient is also presented.
`From the experimental studies, it is found that etch rates vary with variation of etching ambient. The concentrations that maximize silicon
`etch rate is 3% for TMAH and 22 wt.% for KOH. Aluminum etch rate is high in KOH and undoped TMAH but negligible in dual doped
`TMAH. Silicon dioxide etch rate is higher in KOH than in TMAH and dual doped TMAH solutions. The !111O/!100O etch rate ratio is
`highest in TMAH compared to the other two etchants whereas smoothest etched silicon surface is achieved using dual doped TMAH. The
`study reveals that dual doped TMAH solution is a very attractive CMOS compatible silicon etchant for commercial MEMS fabrication which
`has superior characteristics compared to other silicon etchants.
`q 2005 Elsevier Ltd. All rights reserved.
`
`Keywords: MEMS; Bulk micromachining; KOH; TMAH; Dual doped TMAH; Surface roughness; !111O/!100O Etch rate ratio
`
`1. Introduction
`
`Bulk silicon micromachining is an essential process step
`for the fabrication of MEMS devices. Anisotropic wet
`chemical etching of silicon is frequently used for shaping
`quite intricate three-dimensional structures such as proof
`masses, cantilevers, diaphragms, trenches and nozzles on
`silicon substrate [1]. Presently, dry etching techniques (RIE
`and DRIE) are employed for high aspect ratio silicon
`micromachining but wet chemical etching still dominates
`over dry etching due to its low process cost, simple etch
`setup, higher etch rate, better surface smoothness, high
`degree of anisotropy and lower environmental pollution.
`The study of silicon, silicon dioxide, aluminum etch rates
`along with topology of the etched silicon surface for various
`silicon etchants at different temperatures is necessary to
`
`* Corresponding author. Tel.: C91 3222 281479; fax: C91 3222 255303.
`E-mail address: kanishka@ece.iitkgp.ernet.in (K. Biswas).
`
`0026-2692/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.
`doi:10.1016/j.mejo.2005.07.012
`
`develop a CMOS compatible commercial process for
`fabricating micromechanical devices. Compatibility of
`MEMS fabrication process with commercial IC fabrication
`technology allows for integration of microtransducers with
`integrated circuits which provides on-chip signal condition-
`ing, interface control and remote signal transmission. The
`commonly used silicon etchants are classified into three
`main groups:
`(i) Alkali metal hydroxides [1–3]
`(ii)
`Diamines based [4] and (iii) Quaternary ammonium
`hydroxides [5,6]. KOH (potassium hydroxide) is a non-
`toxic, economical and commonly used alkali metal
`hydroxide silicon etchant which requires simple etch setup
`and provides high silicon etch rate, high degree of
`anisotropy, moderate Si/SiO2 etch rate ratio and low etched
`surface roughness [2,3]. But KOH damages exposed
`aluminum metal
`lines very quickly and is not CMOS
`compatible due to the presence of alkali metal ions in it.
`EDP (ethylenediamene pyrocatecol) is a diamine based
`silicon etchant which has moderate silicon etch rate, high
`Si/SiO2 etch rate ratio, low degree of anisotropy and is
`partly CMOS compatible [4]. But EDP ages quickly,
`
`Raytheon2035-0001
`
`Sony Corp. v. Raytheon Co.
`IPR2015-01201
`
`

`
`520
`
`K. Biswas, S. Kal / Microelectronics Journal 37 (2006) 519–525
`
`requires a complex etching apparatus and careful handling
`as it produces reaction gases which are health hazardous and
`so require special safety measures. TMAH (tetramethyl
`ammonium hydroxide) is the most preferred quaternary
`ammonium hydroxide based silicon etchant [5,6]. TMAH is
`gaining popularity despite its high cost and complex etch
`setup because it is a non-toxic, CMOS compatible organic
`solution which has moderately high silicon etch rate and
`high selectivity to masking layers. Ordinary TMAH yields
`rough etched silicon surfaces but when TMAH is doped
`with suitable amounts of silicic acid and AP (ammonium
`the mixture provides
`(NH4)2S2O8),
`peroxodisulphate,
`complete aluminum passivation along with smooth etched
`surfaces [7–9].
`In the present investigation, studies have been made
`on (i) silicon etch rate, (ii) !111O/!100O etch rate
`ratio and (iii) etched !100O silicon surface
`(ER)
`roughness using KOH, TMAH and dual doped TMAH
`solutions. The etching experiments were performed using
`different KOH (10, 22, 33 and 44 wt.%) and TMAH (3,
`8, 12 and 20%) solution concentrations and at different
`bath temperatures (50 to 80 8C). The dual doped TMAH
`etching experiments were carried out using 2 and 5%
`TMAH solution. Our earlier study [9] reported that 2%
`TMAH doped with 30 gm/l silicic acid and 5 gm/l of AP
`whereas 5% TMAH doped with 38 gm/l silicic acid and
`7 gm/l AP provides high silicon etch rate, smooth etched
`silicon surface and almost complete passivation of the
`exposed overlaying aluminum metal
`interconnection
`lines. This paper reports the experimental results and a
`comparison of the silicon etch rates, !111O/!100O
`ER ratio and etched silicon surface roughness
`for
`different Si-etchants, namely KOH, TMAH and dual
`doped TMAH solution.
`
`2. Experimental
`
`Both p-type (resistivity 10–40 U-cm) and n-type
`single crystal !100O silicon
`(resistivity 4–6 U-cm)
`substrates of diameter 4-inch and thickness 525 mm,
`were used for studying anisotropic etching of silicon in
`different etchants. Initially, a silicon dioxide layer of
`thickness around 1 mm was thermally grown by cyclic
`oxidation process. Rectangular and square oxide windows
`of suitable dimensions were opened at the front side of
`the wafer by photolithography process. Thereafter a
`0.8 mm thick aluminum layer was deposited by thermal
`evaporation technique and photolithographically pat-
`terned. Before insertion of
`the samples
`in etching
`solution, native oxide was removed without damaging
`the aluminum metal patterns using a special native oxide
`etchant followed by rinsing in de-ionized water. Each set
`of experiment was repeated twice using two samples held
`vertically in the etching solution during etching. Etching
`at a particular ambient was carried out for a span of
`
`30 min wherein the second sample was inserted 15 min
`after the first sample. The silicon and aluminum etch
`rates and etched surface roughness were measured by
`Dektak3 surface profilometer by averaging the readings
`obtained from the x- and y-scans. The silicon dioxide
`thickness was measured using an ellipsometer. The !
`111O/!100O ER ratio were determined using an
`optical microscope and SEM.
`
`2.1. Experiments using KOH-water solution
`
`The KOH-water solution was prepared by diluting
`commercially available KOH pellets
`(84% pure,
`E.Merck, India). All experiments were carried out in a
`closed glass beaker with a constant temperature bath. No
`external mechanical stirring or reflex condenser was used
`in etch bath of KOH during experiments. KOH etching
`was performed using both p-type and n-type silicon
`substrates. The KOH concentration was varied from 10
`to 44 wt.% and temperature of the etch bath from 50 to
`80 8C in steps of 10 8C. The silicon dioxide etch rate was
`determined after 60 min of etching in KOH.
`
`2.2. Experiments using ordinary TMAH-water solution
`
`The TMAH-water solution was prepared by diluting
`commercially
`available TMAH (25 wt.%, Merck,
`Germany). All experiments were carried out in a closed
`glass vessel with a constant temperature bath. A water
`cooled reflex condenser was used to prevent changes in
`etchant concentration during etching. A magnetic stirrer
`rotating with speed of 100 rpm was used constantly
`during etching to facilitate uniform etching. The TMAH
`etching experiments were carried out using 3 to 20%
`TMAH at different bath temperatures (50 to 80 8C). An
`etching time of 120 min was used to determine silicon
`dioxide etch rate in TMAH.
`
`2.3. Experiments using dual doped TMAH solution
`
`For this study, 2 and 5% TMAH doped with silicic
`acid and AP was used at 60, 70 and 80 8C. The 2%
`TMAH solution was mixed with 30 gm/l silicic acid and
`5 gm/l of AP whereas the 5% TMAH solution contained
`38 gm/l of silicic acid and 7 gm/l AP [9]. During etching
`mechanical agitation was provided continuously with the
`help of a magnetic stirrer rotating at 100 rpm and a water
`cooled reflex condenser prevented any change of etchant
`concentration. The variation of aluminum, silicon and
`silicon dioxide etch rates were measured in each case.
`The silicon surface roughness measurements were also
`carried out.
`
`Raytheon2035-0002
`
`

`
`K. Biswas, S. Kal / Microelectronics Journal 37 (2006) 519–525
`
`521
`
`3. Results and discussions
`
`3.1. KOH solution
`
`Fig. 1 shows the variation of the etch rates of n-type and
`p-type silicon for four different KOH concentrations at
`different bath temperatures. In the present study, 22 wt.%
`KOH solution provides maximum silicon etch rates of 89.2
`and 88.1 mm/h for n-type and p-type silicon respectively at
`80 8C. The dopant type of silicon substrate has little effect
`on the etch rate of silicon although n-type etches slightly
`faster than p-type silicon. Fig. 2 shows the variation of
`SiO2 and Al etch rates at different temperatures and KOH
`concentrations. The silicon dioxide etch rate increases
`continuously with increase in temperature irrespective of
`KOH solution concentration. The maximum silicon dioxide
`etch rate is 450 nm/h at 80 8C using 33 wt.% KOH. The Al
`etch rate is appreciable in all KOH concentrations with
`maximum etch rate of 3.0 mm/min. The etched silicon
`surface smoothens with both increase in KOH concen-
`tration and bath temperature. Silicon surface roughness
`degrades with increase in etch duration due to the masking
`of hydrogen bubbles evolved during etching which
`significantly contributes to surface roughness [10–13].
`Fig. 3(a) shows the SEM micrograph of p-type silicon
`surface etched in 22 wt.% KOH at 70 8C. The maximum
`silicon roughness value is 0.83 mm resulting from 10 wt.%
`
`Fig. 2. Variation of the silicon dioxide and aluminum etch rates with KOH
`concentration.
`
`KOH at 50 8C and the smoothest silicon surface has
`roughness value of 0.12 mm rough which is obtained using
`33 wt.% KOH.
`
`3.2. TMAH solution
`
`Fig. 4 shows the variation of the silicon etch rates with
`TMAH solution. The maximum silicon etch rate obtained is
`60.2 mm/h using 3% TMAH at 80 8C. The silicon etch rate
`decreases with increase in TMAH concentration. Similar to
`KOH, dopant type of silicon substrate type has negligible
`effect on silicon etch rate. The silicon dioxide and Al etch
`rate for various TMAH concentrations is shown in Fig. 5.
`The silicon dioxide etch rate increases with both increase in
`temperature and decrease in TMAH concentration. The
`maximum silicon dioxide etch rate value is almost half in
`TMAH compared to KOH which is an attractive feature of
`TMAH. Undoped TMAH solution attacks aluminum film
`but to a less extent than KOH. The maximum aluminum
`etch rate is 1.44 mm/min in 20% TMAH at 80 8C. The
`etched silicon surface smoothness increases with both
`increase in TMAH concentration and temperature. The
`random pyramidal hillocks formed during etching signifi-
`cantly contributes to the surface roughness than other
`factors [11,14–16]. Fig. 3(b) shows the SEM micrographs of
`the etched silicon surface at TMAH concentration of 5% at
`70 8C. The inset SEM photograph of Fig. 3(b) shows the
`random hillock formation on the etched silicon surface after
`2 hours of etching in 3% TMAH at 50 8C. The maximum
`and minimum silicon roughness values are 3.7 and 0.04 mm
`which are obtained using 3% TMAH and 20% TMAH,
`respectively.
`
`3.3. Dual doped TMAH solution
`
`Fig. 1. Variation of the silicon etch rate with KOH concentration.
`
`The variation of the etch rates of silicon, aluminum and
`silicon dioxide is presented in Fig. 6. The etch rate of silicon
`
`Raytheon2035-0003
`
`

`
`522
`
`K. Biswas, S. Kal / Microelectronics Journal 37 (2006) 519–525
`
`Fig. 3. SEM microphotograph of silicon!100O after etching in (a) 22 wt.% KOH at 70 8C (b) 8% TMAH at 70 8C (Inset: Hillock formation at low
`concentrations of TMAH) and (c) 5% dual doped TMAH at 80 8C [9].
`
`in dual doped TMAH is almost comparable to that of
`KOH and TMAH but the masking silicon dioxide and
`aluminum etch rates are quite small which is desirable for
`MEMS fabrication technology. Dual doped TMAH also
`improves surface smoothness of the etched surfaces due to
`the suppression of
`random hillock formation by the
`oxidizing agent AP. The silicon surface roughness is
`three orders less in dual doped TMAH solutions in
`comparison to KOH and TMAH. Fig. 3(c) shows the SEM
`micrograph of the n-type silicon surface etched in 5% dual
`doped TMAH at 80 8C [9].
`
`3.4. Comparative study
`
`A comparative study of the etching characteristics of
`silicon,
`silicon dioxide and aluminum using three
`anisotropic silicon etchants
`(KOH, TMAH and dual
`doped TMAH) has been made. Fig. 7 silicon shows the
`comparison of the silicon etch rates for the three etchants
`whereas Fig. 8 presents the comparison of aluminum
`and silicon dioxide etch rates. The etched silicon !
`100O surface roughness values
`for all
`the etchant
`concentrations are compared in Fig. 9. An analysis of
`
`Fig. 4. Variation of the silicon etch rates with TMAH concentration.
`
`Raytheon2035-0004
`
`

`
`K. Biswas, S. Kal / Microelectronics Journal 37 (2006) 519–525
`
`523
`
`Fig. 5. Variation of the silicon dioxide and aluminum etch rates with TMAH concentration.
`
`the comparison of the etch characteristics (Figs. 7–9)
`reveals that (a) the highest silicon etch rate is obtained
`using KOH solution at 808C whereas dual doped TMAH
`solution provides complete aluminium passivation, (b) the
`silicon dioxide etch rate is approximately two times higher in
`KOH solutions compared to ordinary TMAH and three
`orders higher than dual doped TMAH solutions, (c) surface
`roughness of silicon etched in KOH solution is one order less
`compared to surfaces etched in undoped TMAH solution and
`
`(d) the etched silicon surface roughness is almost three orders
`less in dual doped TMAH compared to the other two
`etchants. The method used for determining the ER ratios of
`the etchants is shown in Fig. 10. The !111O etch rate was
`determined using SEM microphotographs. Fig. 11 presents
`of !111O/!100O ER ratios
`the
`comparison
`of the etchants. The !111O/!100O ER ratio was
`1.68% for KOH, 3.44% for TMAH, 2.48 and 2.83% for
`2 and 5% dual doped TMAH, respectively. The results
`
`Fig. 6. Variation of the silicon, aluminum and silicon dioxide etch rates with dual doped TMAH.
`
`Raytheon2035-0005
`
`

`
`524
`
`K. Biswas, S. Kal / Microelectronics Journal 37 (2006) 519–525
`
`Fig. 7. Comparison of the variation of the silicon etch rates with KOH, TMAH and dual doped TMAH.
`
`indicate that !100O etch selectivity is better in dual doped
`TMAH compared to TMAH. However, !100O etch
`selectivity is much better in KOH compared to either
`TMAH and dual doped TMAH.
`
`4. Conclusions
`
`An experimental study of !100Osilicon, silicon
`dioxide, aluminum etch rates and etched surface mor-
`phology has been carried out in three different anisotropic
`wet chemical etchants at a variety of etching conditions.
`Silicon etch rates and !100O etch selectivity is highest in
`KOH solution but KOH etches the masking silicon dioxide
`layer and aluminum metallization much faster than TMAH
`
`Fig. 8. Comparison of the variation of the silicon dioxide and aluminum
`etch rates with KOH, TMAH and dual doped TMAH.
`
`Fig. 9. Comparison of the variation of the !100O silicon surface
`roughness with KOH, TMAH and dual doped TMAH.
`
`Raytheon2035-0006
`
`

`
`K. Biswas, S. Kal / Microelectronics Journal 37 (2006) 519–525
`
`525
`
`maximum in undoped TMAH. N-type silicon has higher
`etch rate and smoother etched surface compared to p-type
`silicon in all etchants. The !111O/!100O ER ratio is
`highest in TMAH and is found to be independent of etchant
`concentration and etching temperature for all three etchants.
`The 5% dual doped etchant has been successfully used in the
`author’s laboratory to fabricate a variety of MEMS devices
`and microstructures.
`
`References
`
`[1] H. Seidel, L. Csepregi, A. Heuberger, H. Baumgartel, Anisotropic
`etching of crystalline silicon in alkaline solutions, J. Electrochem.
`Soc. 137 (11) (1990) 3612–3625.
`[2] I. Barycka, I. Zubel, Silicon anisotropic etching in alkali solutions: I,
`Sens. Actuators A70 (1998) 250–259.
`[3] K.E. Petersen, Silicon as a mechanical material, Proc. IEEE 70 (5)
`(1982) 420–457.
`[4] R.M. Finne, D.L. Klein, A water-amine complexing agent system for
`etching in silicon, J. Electrochem. Soc. 114 (9) (1967) 965–970.
`[5] O. Tabata, R. Asahi, H. Funabashi, K. Shimaoka, S. Sugiyama,
`Anisotropic etching of silicon in TMAH solutions, Sens. Actuators
`A34 (1) (1992) 51–57.
`[6] P-. Hei, H-. Yah Peng, C-. Ming Hsieh, M.K. Chyu, The characteristic
`behavior of TMAH water solution for anisotropic etching on both
`silicon substrate and SiO2 layer, Sens. Actuators A93 (2001) 132–137.
`[7] K. Lian, S. Smith, A.J. Walton, A.M. Gundlach, J.T.M Stevenson,
`Characterisation of aluminium passivation for TMAH based
`anisotropic etching for MEMS applications, in: Proceedings of the
`IEEE International Conference on Microelectronic Test Structures
`(ICTMS00), Monterey, CA, USA (2000) 210–214.
`[8] P.M. Sarro, D. Brida, W.v.d. Vlist, S. Brida, Effect of surfactant on
`surface quality of
`silicon microstructures etched in saturated
`TMAHW solutions, Sens. Actuators A85 (2000) 340–345.
`[9] K. Biswas, S. Das, D.K. Maurya, S. Kal, S.K. Lahiri, Bulk
`micromachining of silicon in TMAH-based etchants for aluminum
`passivation and smooth surface, Microelectron. J., in press.
`[10] K. Sato, K. Sato, M. Shikida, T. Yamashiro, M. Tsunekawa, S. Ito,
`Roughening of single-crystal silicon surface etched by KOH water,
`Sens. Actuators A73 (1999) 122–130.
`[11] M. Shikida, K. Sato, K. Tokoro, D. Uchikawa, Differences in
`anisotropic etching properties of KOH and TMAH solutions, Sens.
`Actuators A80 (2000) 179–188.
`[12] E. Palik, O. Glembocki, Etching roughness for (100) silicon surfaces
`in aqueous KOH, J. Appl. Phys. 70 (1991) 3291–3300.
`[13] P.M.M.C. Bressers, J.J. Kelly, J.G.E. Gardeniers, M. Elwenspoek,
`Surface morphology of p-type (100) silicon etched in aqueous alkaline
`solution, J. Electrochem. Soc. 143 (1996) 1744–1750.
`[14] J.T.L. Thong, W.K. Choi, C.W. Chong, TMAH etching of silicon and the
`interaction of etching parameters, Sens. Actuators A63 (1997) 243–249.
`[15] Osamu Tabata, pH-controlled TMAH etchants for silicon micro-
`machining, Sens. Actuators A53 (1996) 335–339.
`[16] J.T.L. Thong, P. Luo, W.K. Choi, S.C. Tan, Evolution of hillocks during
`silicon etching in TMAH, J. Micromech. Microeng. 11 (2001) 61–69.
`
`Fig. 10. The method used for determining the !111O/!100O etch rate
`ratios.
`
`Fig. 11. Comparison of the variation of the !111O/!100O etch rate
`ratios (in %) with KOH, TMAH and dual doped TMAH.
`
`and dual doped TMAH. In TMAH, the silicon dioxide etch
`rate is low whereas the silicon etch rate is comparable to
`KOH. Low concentration dual doped TMAH provides
`complete aluminum passivation during silicon microma-
`chining. The etched silicon surface roughness is lowest in
`dual doped TMAH solution, optimally high in KOH and
`
`Raytheon2035-0007