Direct bond transfer layers for manufacturable sealing of microfluidic chips

Abstract

Techniques for use of wafer bonding techniques for sealing of microfluidic chips are provided. In one aspect, a wafer bonding sealing method includes the steps of: forming a first oxide layer coating surfaces of a first wafer, the first wafer having at least one fluidic chip; forming a second oxide layer on a second wafer; and bonding the first wafer to the second wafer via an oxide-to-oxide bond between the first oxide layer and the second oxide layer to form a bonded wafer pair, wherein the second oxide layer seals the at least one fluidic chip on the first wafer. The second wafer can be at least partially removed after performing the bonding, and fluidic ports may be formed in the second oxide layer. A fluidic chip device is also provided.

Claims

1. A device, comprising: a first oxide layer coating surfaces of a first wafer, the first wafer comprising at least one fluidic chip that includes fluidic channels joined by nanochannel structures, wherein each of the nanochannel structures comprises an array of pillars for particle sorting, and wherein the first oxide layer is a conformal oxide layer covering top and sidewall surfaces of each of the pillars; and a second oxide layer bonded to the first oxide layer via an oxide-to-oxide bond, wherein the second oxide layer seals the at least one fluidic chip on the first wafer.

2. The device of claim 1, wherein the pillars are patterned in the first wafer.

3. The device of claim 1, wherein the second oxide layer is bonded to the array of the pillars via the conformal oxide layer.

4. The device of claim 1, wherein the second oxide layer is bonded to the conformal oxide layer at the top surfaces of the pillars.

5. The device of claim 1, wherein the fluidic channels comprise microchannels.

6. The device of claim 5, further comprising: fluidic ports formed in the second oxide layer at opposite ends of each of the microchannels.

7. The device of claim 5, wherein the microchannels have a width of from about 10 m and about 50 m, and ranges therebetween.

8. The device of claim 1, wherein the first oxide layer is thinner than the second oxide layer.

9. The device of claim 1, wherein the first oxide layer has a thickness of from about 5 nm to about 50 nm, and ranges therebetween.

10. The device of claim 1, wherein the second oxide layer has a thickness of from about 0.5 m to about 2 m, and ranges therebetween.

11. The device of claim 1, further comprising: fluidic ports formed in the second oxide layer.

12. The device of claim 11, wherein the fluidic ports are partially formed through the second oxide layer, such that a portion of the second oxide layer remains separating the fluid ports from the fluid channels.

13. The device of claim 11, wherein the portion of the second oxide layer that remains separating the fluid ports from the fluid channels has a thickness of from about 50 nm to about 300 nm, and ranges therebetween.

14. The device of claim 1, wherein at least one of the first oxide layer and the second oxide layer comprises a thermal oxide.

15. The device of claim 14, wherein the thermal oxide comprises thermal SiO.sub.2.

16. The device of claim 1, wherein the first wafer comprises multiple fluidic chips.

17. The device of claim 1, further comprising: an adhesive tape on portions of the second oxide layer at locations for fluidic ports to be formed in the second oxide layer.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 is a diagram illustrating an exemplary (first) wafer (100) on which one or more fluidic chips (102) have been printed according to an embodiment of the present invention;

(2) FIG. 2 is a diagram illustrating an enlarged view of one of the fluidic chips (102) which includes one or more microchannels (202) joined by nanochannel structures (204) according to an embodiment of the present invention;

(3) FIG. 3 is a diagram illustrating an enlarged view of one of the nanochannel structures (204) according to an embodiment of the present invention;

(4) FIG. 4 is a diagram illustrating an enlarged view of one of the pillar sorting arrays (402) according to an embodiment of the present invention;

(5) FIG. 5 is a cross-sectional diagram illustrating one of the pillar sorting arrays (402) according to an embodiment of the present invention;

(6) FIG. 6 is a cross-sectional diagram illustrating a (first) thin oxide layer (602) having been formed on the surfaces of the pillar array (402) to help facilitate wafer bonding according to an embodiment of the present invention;

(7) FIG. 7 is a cross-sectional diagram illustrating a (second) wafer (700) having a (second) oxide layer (702) formed thereon according to an embodiment of the present invention;

(8) FIG. 8 is a cross-sectional diagram illustrating the first wafer (100) and the second wafer (700) having been bonded together by way of an oxide-to-oxide bond between the first (602) and second (702) oxide layers according to an embodiment of the present invention;

(9) FIG. 9 is a cross-sectional diagram illustrating the (second) wafer (700) having been removed leaving behind only the (second) oxide layer (702) over/sealing the channels according to an embodiment of the present invention;

(10) FIG. 10 is a cross-sectional diagram illustrating fluidic ports having been formed in the second oxide layer (702) according to an embodiment of the present invention;

(11) FIG. 11 is a cross-sectional diagram illustrating a microfluidic jig (1102) being used to open a partially formed fluidic port just prior to use according to an embodiment of the present invention;

(12) FIG. 12 is a cross-sectional diagram illustrating an alternative technique for forming fluidic ports using an adhesive tape according to an embodiment of the present invention;

(13) FIG. 13 is an image of a silicon (Si) pillar array sample prepared by the present techniques to have 50 nanometers (nm) of thermally grown silicon dioxide (SiO.sub.2) on the surface according to an embodiment of the present invention; and

(14) FIG. 14 is an image of the Si pillar array of FIG. 12 having been sealed using the present direct wafer bond transfer process according to an embodiment of the present invention.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

(15) Provided herein are techniques for sealing micro/nanofluidic channels and features using direct wafer bonding. Advantageously, the present techniques for capping fluidic chips are fully cleanroom compatible to enable high-volume production and three-dimensional integration. As will be described in detail below, a thin thermal oxide is formed on the walls and surface of a first wafer containing the (micro/nano) fluidic features, which provides a good bonding surface for a second wafer consisting of bulk silicon (Si) covered with a thick uniform silicon dioxide (SiO.sub.2) layer that becomes the ceiling for the fluidic features and provides a good oxide-to-oxide bond between the two wafers. The two wafers are then tacked together in a wafer bonding tool with the two oxide surfaces in contact and subsequently annealed at a high temperature to fortify the bond. Finally, the silicon of the second wafer is removed. The final microfluidic features can therefore also consist of a high quality SiO.sub.2 covering all surfaces, which is easy to modify chemically.

(16) The present wafer bonding approach has several notable advantages over other bonding processes, such as anodic bonding, eutectic bonding, glass fritting, and polymer adhesive bonding. For example, wafer bonding has the ability to survive the high temperatures required to make stacked structures; wafer bonding has a high bond strength; and wafer bonding is compatible with standard microelectronic processing (see for example, Moriceau et al., Overview of recent direct wafer bonding advances and applications, Advances in Natural Sciences: Nanoscience and Nanotechnology 1(4):043004 (2010) (Published February 2011), the contents of which are incorporated by reference as if fully set forth herein).

(17) The process begins, as shown in FIGS. 1-4, and shown by way of example with a wafer 100 on which one or more fluidic chips 102 have been printed. In the example shown in FIG. 1, multiple fluidic chips 102 printed on the wafer. This is however only an example, and it is to be understood that the present techniques may be performed in the same manner described for implementations involving a single chip. As will be described in detail below, the present techniques involve a bonding process between a pair of wafers. Wafer 100 is the first wafer in the pair, and thus may also be referred to herein as the first wafer. FIG. 2 provides an enlarged view of one of the fluidic chips 102. As shown in FIG. 2, each of the fluidic chips 102 includes pairs of microchannels 202 joined by nanochannel structures 204. The terms micro and nano are used herein to denote the relative sizes of features. Regarding the micro/nanochannels for instance, as shown in FIG. 2, the microchannels 202 feed into the relatively smaller nanochannel structures 204. By way of example only, the microchannels can have a width of from 10 micrometers (m) to about 50 m, and ranges there between, whereas the nanochannels can have a width of from 30 nanometers (nm) to about 500 nm, and ranges there between.

(18) An enlarged view of one of the nanochannel structures 204 is shown in FIG. 3. Each of the nanochannel structures 204 contains a pillar arrangement for particle sorting (also referred to herein as a pillar sorting array). The pillar sorting arrays can operate on the principle of deterministic lateral displacement to sort, separate, and enrich microscale and nanoscale particles. See, for example, Huang et al., Continuous Particle Separation Through Deterministic Lateral Displacement, Science, vol. 304 (May 2004), the contents of which are incorporated by reference as if fully set forth herein.

(19) An enlarged view of one of the pillar sorting arrays is provided in FIG. 4. As shown in FIG. 4, the pillar array includes a plurality of (e.g., silicon) pillars 402 patterned in the wafer. FIG. 4 provides a top-down view of the pillar array. In order to better illustrate the present process, reference will now shift to cross-sectional views of the pillar array (e.g., based on a cut along line A-Asee FIG. 4). See, for example, FIG. 5. As shown in FIG. 5, the pillars 402 define a plurality of fluid channels. It is via these fluid channels that liquid samples will pass through the array. While the following description details the present process from the perspective of sealing fluidic channels in the pillar array, it is to be understood that the instant techniques are employed (in the same manner described) to seal any/all of the channels or features present in this any other silicon and oxide based fluidic chip design. Thus, the channels shown in the figures and described below might generally represent any of the micro/nano fluidic channels present on the chip.

(20) As highlighted above, the present techniques employ novel wafer bonding processes for sealing the above-described fluidic chips. To help facilitate bonding, a thin, conformal oxide layer 602 is formed on the surfaces of the pillar array. See FIG. 6. This oxide layer 602 may also be referred to herein as a (first) bonding oxide layer so as to differentiate it from the (second) oxide layer that will be formed on a second wafer (see below). According to an exemplary embodiment, the oxide layer 602 is formed on the top and sidewall surfaces of the pillars 402 (as shown in FIG. 6) having a thickness of from about 5 nm to about 50 nm, and ranges therebetween. By way of example only, oxide layer 602 can be formed in this manner using a thermal oxidation process. In that case, oxide layer 602 may also be referred to herein as a thermal oxide. In order to form a thin oxide layer 602, it may be preferable to use a dry thermal oxidation process, i.e., wherein molecular oxygen is the oxidant.

(21) As highlighted above, the present wafer bonding process will involve providing a second wafer 700, also having a bonding oxide layer, and bonding the first wafer 100 to the second wafer 700 via an oxide-to-oxide bond between the respective bonding oxide layers. Thus, as shown in FIG. 7, a (second) wafer 700 is provided, and a (second) oxide layer 702 is formed on the wafer 700. Since the (second) wafer 700 will be used to provide the physical ceiling oxide used to seal the fluid channels, wafer 700 may also be referred to herein as the ceiling wafer, so as to distinguish it from wafer 100 containing the fluid channels.

(22) According to an exemplary embodiment, wafer 700 is a bulk silicon (Si) wafer having a thickness of from about 550 micrometers (m) to about 750 m, and ranges therebetween. A thermal oxidation process may be used here as well to form the oxide layer 702 on the wafer 700. However, the oxide layer 702 is formed thicker than oxide layer 602 (i.e., a thicker oxide layer 702 can enhance the oxide-to-oxide bond and is necessary to provide a robust seal that can withstand pressure applied to the channels to drive fluidics, while one runs the risk of occluding trenches in the first wafer with a thicker oxide layer 602). By way of example only, oxide layer 702 can be formed on the wafer 700 to a thickness of from about 0.5 m to about 2 m, and ranges therebetween. A wet thermal oxidation process (e.g., using water vapor as the oxidant) is well suited for growing thicker oxide layers. As will be apparent from the description that follows, the oxide layer 702 on wafer 700 will serve as the ceiling for the fluidic channels (on wafer 100).

(23) Prior to forming the oxide layer 702 on wafer 700, it is preferable to clean the wafer 700 to remove any potential contaminants. By way of example only, a standard RCA clean may be performed to remove organic and other contaminants from the surface of the wafer 700. Further, prior to bonding the wafers together, it is preferable to clean the oxide bonding surfaces. By way of example only, a distilled water megasonic cleaning, and subsequent infrared (IR) drying of the wafers 100 and 700 can be performed post formation of the oxide layers 602 and 702 before tacking the wafers together for bonding.

(24) Wafers 100 and 700 are then bonded together by way of a direct oxide-to-oxide bond between oxide layers 602 and 702, forming a bonded wafer pair. See FIG. 8. According to an exemplary embodiment, the bonding process is performed by pressing the (bonding) oxide layers 602 and 702 together (using a bonding tool) at room temperature with a force initiating from a center contacting area. The contacting area will expand from the center outward across the layers. Thereafter, a thermal anneal at from about 800 C. to about 1,000 C., and ranges therebetween, for a duration of from about 40 minutes to about 60 minutes, and ranges therebetween in an inert gas (e.g., nitrogen (N.sub.2)) ambient is used to enforce the bonding quality. By way of example only, a suitable bonding tool is the CL-200 substrate bonder available from SUSS MicroTec AG, Garching, Germany. As shown in FIG. 8, the oxide layer 702 (originating on the second wafer 700) now serves as the ceiling of the fluid channels.

(25) Following wafer bonding, according to an exemplary embodiment, the (second) wafer 700 is removed from the structure leaving behind only the (second) oxide layer 702 over and sealing the channels. See FIG. 9. For instance, the thickness of wafer 700 is first reduced (e.g., to about 50 m) using a coarse followed by a thin wafer polish process via chemical mechanical polishing (CMP). Alternatively, a Smart Cut process may be employed in the same manner to partially remove the wafer 700. The Smart Cut process is described, for example, in U.S. Pat. No. 5,374,564 issued to Bruel, entitled Process for the Production of Thin Semiconductor Material Films, the contents of which are incorporated by reference as if fully set forth herein. The remaining portion of wafer 700 can then be removed using a tetramethylammonium hydroxide (TMAH) etch process to leave only the (from about 0.5 m to about 2 m, and ranges therebetweensee above) thick oxide layer 702 ceiling bonded directly to the pillar array of wafer 100. It is notable that the thin oxide layer 602 serves to prevent any etching of wafer 100 during the TMAH etch. As an alternative to TMAH, potassium hydroxide (KOH) wet etching or xenon difluoride (XeFl.sub.2) etching may also be employed to remove the remaining portions of wafer 100.

(26) Advantageously, the present ceiling techniques enable further processing not possible with conventional configurations (such as those using glass to seal the channels). For instance, as shown in FIG. 10, fluidic ports can be formed in the oxide layer 702 to access the channels. It is notable that the fluidic ports are preferably formed at the ends of the microchannel regions (i.e., the fluidic ports reside in the regions represented by open squares around the perimeter of each chip in FIG. 1). The reason for this is that the channels are much wider in these regions than in the pillar regions, making the oxide much simpler to puncture. Thus, the views in FIGS. 10 and 11 depict the highlighted regions in FIG. 1 (the regions at the ends of the microchannel regions). The fluidic ports can be formed using a standard lithography and etching process (such as reactive ion etching or RIE). The fluidic ports can be formed completely through the (ceiling) oxide layer 702 or, as shown in FIG. 10, the fluidic ports can be formed extending only partway through the (ceiling) oxide layer 702. By only partially forming the fluidic ports, a part of the oxide layer 702 will remain over the channels preventing any fluid from entering the channels prior to use via wicking or capillary action. The thin remaining oxide can then be punctured by a needle embedded within a microfluidic jig, such that the oxide is penetrated and the microfluidic channels opened just prior to use. According to an exemplary embodiment, the fluidic ports are partially formed through the oxide layer 702 (such as shown in FIG. 10) leaving a portion of the oxide layer having a thickness of from about 50 nm to about 300 nm, and ranges therebetween, separating the fluidic ports and the channels.

(27) FIG. 11 illustrates the optional use of a microfluidic jig 1102 to fully open up the (partially formed) fluidic ports prior to use. As shown in FIG. 11, the microfluidic jig 1102 contains a needle(s) that can pierce the thin remaining portion of the (ceiling) oxide layer 702 to gain access to the channels just prior to use.

(28) Alternatively, an adhesive tape may be applied in the desired area of the fluidic port regions, where ports have not been preetched to form any fluidic ports. Then, just prior to use, the adhesive tape can be peeled off (and with it local portions of the ceiling oxide layer 702) which will locally remove oxide in these regions forming the fluidic ports.

(29) The present techniques are further illustrated by way of reference to the following non-limiting examples. FIG. 13 illustrates an uncapped Si pillar array having 50 nm of thermally grown SiO.sub.2 on the surface prepared using the present techniques. FIG. 14 illustrates a processed sealed array using the direct wafer bond transfer process described above. In this example, the ceiling oxide (provided via the second wafer) was a 2 m thick layer of SiO.sub.2.

(30) Lithography and anisotropic RIE of the ceiling SiO.sub.2 in select regions was then used to establish fluidic ports (not shown in the images) to access the microfluidic channels and a protective layer, such as a spin applied photoresist when fluidic ports have not be RIE etched completely through the ceiling oxide (702), or bonding tape for dicing when holes are etched completely through to the microchannels, was then added for wafer dicing to separate the individual chips. These chips ran fluidics containing sub-micron particles demonstrating fully functional sorting capabilities at high pressure and withstood over 20 Bars of fluidic pressure without leakage or damage to the oxide during stress tests, which is far more than is typically employed in microfluidic applications.

(31) Although illustrative embodiments of the present invention have been described herein, it is to be understood that the invention is not limited to those precise embodiments, and that various other changes and modifications may be made by one skilled in the art without departing from the scope of the invention.