Gabriel2009. ALD of optical coatings inside microchannels

SAVE OFFLINE

WB5 12:00 PM – 12:15 PM

Atomic layer deposition of optical coatings inside microchannels Nicholas T. Gabriel and Joseph J. Talghader Department of Electrical and Computer Engineering University of Minnesota Minneapolis, Minnesota 55455 Email: [email protected]

I. I NTRODUCTION Very high-aspect ratio channels may be coated using atomic layer deposition (ALD) due to the unique self-limiting nature of the process. Reactive ion-etched trenches with aspect ratios near 100:1 have been coated [1]–[3]. For optical and microfluidic applications, most channels are centimeters deep with diameters of tens to hundreds of micrometers. This results in a similarly high aspect ratio as etched trenches but the larger area creates more difficult problems of temperature and gas flow uniformity. To quantitatively explore the behavior, we create an air wedge between 2 square wafers of silicon approximately 7 cm on a side, with the air gap varying in thickness linearly from 0–1650 µm over 7 cm. A Fabry-Perot optical cavity composed of an ALD Al2 O3 /HfO2 multilayer was deposited inside the wedge, and the measured resonant wavelength only shifted by −10.3% over a distance of 6 cm along the gas flow direction and over a 118–1533 µm range of gap thickness. Our first experiments to test optical coatings inside microchannels were performed using 12 cm long glass capillary tubes with an interior dimension defined by a square 500 µm opening, a 240:1 aspect ratio. Fig. 1 shows cross-sectional SEM images at the midpoint of a capillary coated with 1000 cycles of ALD Al2 O3 . The inset indicates that the interior coating is near the expected 110 nm thickness for this 0.11 nm/cycle process, but it is difficult to further quantify the properties of transparent films deposited on the inside of narrow transparent tubes.

Fig. 1. Cross-sectional SEM images at the midpoint of 12 cm capillary tube after coating with 1000 cycles of Al2 O3 . The 500 µm inner dimension can be seen at 100x, and the inset at 35000x magnification indicates the presence of a roughly 100 nm film. The white box indicates the location of the inset.

Thickness and refractive index at 633 nm was measured on single films deposited in this manner using a variable-angle spectroscopic ellipsometer (VASE) from J.A. Woollam with x-y translation, yielding maps of these parameters as a function of position in the wedge. Measurements were performed on the bottom piece over a 36 cm2 square area at the center each sample, thus leaving about 0.5 cm near each edge unmeasured. Subsequently, a 6-layer optical cavity was deposited in a similar wedge structure consisting of layers L-H-2L-H-L-H, from bottom-to-top, where “L” represents a low-index quarterwave layer of Al2 O3 and “H” is a high-index quarter-wave layer of HfO2 . This yields a resonance at the design wavelength of 500 nm. Translated to ALD cycles, the “L” layers were 714 cycles of Al2 O3 and “H” were 690 cycles of HfO2 . The reflectance of the resulting multilayer was measured from 425–675 nm at several locations along the bottom piece of this wedge structure, this time with fixed s-polarization and a 15◦ angle of incidence from normal. We chose to measure only along the diagonal line y = −x based on the HfO2 thickness results (as shown in Fig. 2) because nearly the full range of observed thickness variation occurs along that diagonal.

II. E XPERIMENTAL To create the air wedge structures, standard (100) silicon wafers were cleaved into nearly square pieces approximately 7 cm on a side. Two of these were centered in the ALD reactor (a Savannah S200 from Cambridge NanoTech) with the polished surfaces facing one-another, in contact along one edge and open at the opposite edge except for 3 small 1650 µm spacers. The process gas flow was perpendicular to this ∼1.35◦ incline, presenting a variable crosssection to the flow. Al2 O3 was deposited from trimethylaluminum (TMA) kept at room temperature [4], and HfO2 from tetrakis(dimethylamido)hafnium kept at 75◦ C [3]; both processes used water as the complementary precursor.

978-1-4244-2382-8/09/$25.00 ©2009 IEEE

III. R ESULTS As expected from the capillary result, the Al2 O3 process shows remarkable ability to coat within the wedge struc-

79

3

55.2

0.8

2

55

0.7

1

54.8

0

54.6

−1

54.4

(−3.0 (−2.5 (−2.0 (−1.5 (−1.0 (−0.5 (+0.0 (+0.5 (+1.0 (+1.5 (+2.0 (+2.5 (+3.0

Thickness (nm)

0.5 R

y (cm)

0.6

+3.0 +2.5 +2.0 +1.5 +1.0 +0.5 −0.0 −0.5 −1.0 −1.5 −2.0 −2.5 −3.0

cm) cm) cm) cm) cm) cm) cm) cm) cm) cm) cm) cm) cm)

0.3 0.2

54.2

−2

0.4

cm, cm, cm, cm, cm, cm, cm, cm, cm, cm, cm, cm, cm,

0.1 −3 −3

54 −2

−1

0 x (cm)

1

2

0

3

460

480

(a)

500 λ (nm)

520

(a)

3

45

520

Prediction (from Fig. 2 data) Measured (Fig. 3a)

44 2

510 42 41

0 40 −1

39

Thickness (nm)

y (cm)

1

λ at minimum R (nm)

43

500

490

480

38 −2

470

37 −3 −3

−2

−1

0 x (cm)

1

2

3

460

(b)

−3

−2

−1

0 x (cm)

1

2

3

(b)

Fig. 2. Measured (a) Al2 O3 and (b) HfO2 thickness profiles in wedge; dots represent actual measurement locations with linear interpolation in between. Flow is in the +x direction and wedge gap decreases in the +y direction from 1533 µm at y = −3 cm to 118 µm at y = 3 cm.

Fig. 3. Measured optical cavity performance along the y = −x diagonal. A subset of the measured spectra is shown in (a), and comparison of resonance location to simulated prediction in (b).

number of cycles. The distinguishable peaks and agreement with simulation indicates that the use of a cavity is a viable characterization technique and the microchannel ALD growth behavior is quite repeatable.

ture, with only 1.2 nm thickness variation across the 36 cm2 measured area shown in Fig. 2a after 500 ALD cycles. The thickness profile in this case seems insensitive to the flow direction, instead showing slightly decreasing thickness with increasing distance from any opening of the wedge. The overall thickness range of 54.0–55.2 nm represents a maximum variation from the peak of only −2.2%. 500 ALD cycles of HfO2 were deposited in a wedge structure yielding the thickness profile shown in Fig. 2b. There is significantly more variation than was seen in Al2 O3 . Thickness varies by −20% from a peak of 45.0 nm, thinnest at (x, y) = (3, −3), which is the corner furthest from the gas inlet and near the wedge opening. The refractive index data for each film is not shown, but it was found to range between 1.640–1.650 for Al2 O3 and 2.050–2.072 for HfO2 at 633 nm. Reflected intensity versus wavelength was measured for the 6-layer optical cavity along the y = −x diagonal, yielding distinct resonance shown in Fig. 3a. Comparison of the reflection minima to simulated predictions based on the single-film results are shown in Fig. 3b. The simulations were performed using the single-film thickness and refractive index results, with expected thickness scaled to the appropriate

ACKNOWLEDGMENTS The authors would like to acknowledge funding from the Joint Technology Office and the Air Force Office of Scientific Research under grant FA9550-05-1-0399. We also thank Dr. Radek Uberna of Lockheed Martin for providing capillary tube samples. R EFERENCES [1] D. M. Hausmann, J. Becker, S. Wang, and R. G. Gordon, “Rapid vapor deposition of highly conformal silica nanolaminates,” Science, vol. 298, no. 5592, pp. 402–6, 2002. [2] J. Niinist¨o, K. Kukli, M. Kariniemi, M. Ritala, M. Leskel¨a, N. Blasco, A. Pinchart, C. Lachaud, N. Laaroussi, Z. Wang, and C. Dussarrat, “Novel mixed alkylamido-cyclopentadienyl precursors for ALD of ZrO2 thin films,” Journal of Materials Chemistry, vol. 18, no. 43, pp. 5243–7, 2008. [3] D. M. Hausmann, E. Kim, J. Becker, and R. G. Gordon, “Atomic layer deposition of hafnium and zirconium oxides using metal amide precursors,” Chem. Mater., vol. 14, no. 10, pp. 4350–8, 2002. [4] G. S. Higashi and C. G. Fleming, “Sequential surface chemical reaction limited growth of high quality Al2 O3 dielectrics,” Appl. Phys. Lett., vol. 55, no. 19, pp. 1963–5, Nov. 1989.

80