Information
-
Patent Application
-
20020042193
-
Publication Number
20020042193
-
Date Filed
September 28, 200122 years ago
-
Date Published
April 11, 200222 years ago
-
CPC
-
US Classifications
-
International Classifications
Abstract
The copper interconnect formed by the use of a damascene technique is improved in dielectric breakdown strength (reliability). During post-CMP cleaning, alkali cleaning, deoxidizing process due to hydrogen anneal or the like and acid cleaning are carried out in the order. After the post-CMP cleaning and before forming an insulation film for a cap film, hydrogen plasma and ammonia plasma processes are carried out on the semiconductor substrate. In this manner, a copper-based buried interconnect is formed in an interlayer insulation film structured of an insulation material having a low dielectric constant.
Description
BACKGROUND OF THE INVENTION
[0001] This invention relates to a fabrication method of a semiconductor integrated circuit device and, more particularly, to a technique applicable effectively for a fabrication method of a semiconductor integrated circuit device having buried interconnects having copper in the main conductor layers thereof.
[0002] In the semiconductor integrated circuit devices, electronic devices, etc., there is a technique established to form interconnects as an interconnect-forming technology, wherein a conductor film, e.g., aluminum or tungsten, is deposited on an insulation film and then patterned by the usual photolithography and dry-etch technique, thereby being formed into an interconnect.
[0003] In the above interconnect-forming technique, however, there is conspicuous increase of interconnect resistance due to the scale-down in the devices and interconnects forming a semiconductor integrated circuit devices or the like, resulting in occurrence of interconnect delay. Thus, there is a limitation in further improving the performance of a semiconductor integrated circuit device or the like. For this reason, in recent years studies have been made on the interconnect-forming technology called damascene, for example. The damascene technology roughly divided with single-damascene technique and dual-damascene technique.
[0004] The single damascene technique is a method that, after forming an interconnect trench in an insulation film, a main conductor layer for interconnect formation is deposited in the interconnect trench and further the main conductor layer is polished to be left only at the inside of the interconnect trench by a chemical mechanical polishing (CMP) for example, thereby forming a buried interconnect at the inside of the interconnect trench.
[0005] Meanwhile, the dual damascene technique is a method that, after forming an interconnect trench and a hole for connection to the lower-leveled interconnect in an insulation film, a main conductor layer for interconnect formation is deposited on the insulation film and in the interconnect trench and hole, and further the main conductor layer is polished to be left only at the inside of the interconnect trench and hole by CMP or the like, thereby forming a buried interconnect at the inside of the interconnect trench and hole.
[0006] In any of the techniques, the interconnect main conductor material uses, for example, a low-resistance material, such as copper, from a viewpoint of improving the performance of the semiconductor integrated circuit device or the like. Copper, having a merit of lower resistance than aluminum and greater allowable current in reliability by two orders of magnitude than that, requires a smaller film thickness in providing the same interconnect resistance and hence reduces the capacitance of between adjacent interconnects.
[0007] However, copper is ready to diffuse into an insulation film as compared to the other metals, such as aluminum and tungsten. For this reason, it is considered that, where using copper as an interconnect material, there is a need to form a copper-diffusion-preventing thin conductive barrier film on a surface of a copper main-conductor layer (bottom and side surfaces), i.e. on an inner wall of the interconnect trench. Meanwhile, there is an art that a cap film, for example, of silicon nitride is deposited over the entire upper surface of an insulation film formed with an interconnect trench in a manner covering the upper surface of a buried interconnect thereby preventing copper in the buried interconnect from diffusing into the insulation film through the upper surface of the buried interconnect.
[0008] The buried interconnect technology having a copper main-conductor layer is described, for example, in JP-A-11-330246. This describes a technique that, after forming a copper interconnect in an interconnect opening formed in a dielectric layer, a barrier layer is formed and further a plasma process is carried out using only ammonia as a source gas, thereby improving the adhesion between the copper interconnect and the copper barrier layer. Meanwhile, JP-A-11-16912, for example, discloses a technique to vanish an oxide layer formed in an interconnect part exposed at a bottom of a connection hole by carrying out heat plasma or ultraviolet-ray illumination process in a deoxidizing atmosphere.
[0009] Meanwhile, the post-CMP cleaning technique is described, for example, in “Monthly Semiconductor World, published October 1998” Sep. 20, 1998, by Press Journal, pp 62-72.
SUMMARY OF THE INVENTION
[0010] In the meanwhile, according to the result of study by the present inventors it has been found that there exist the following problems in the buried interconnect technique using copper in a main conductor layer thereof.
[0011] Firstly, there exists a problem that, where using copper as an interconnect material, the TDDB (Time Dependence on Dielectric Breakdown) life is conspicuously short as compared to other metal materials (e,g. aluminum, tungsten). Moreover, in addition to the scale-down in interconnect pitch and the trend of increasing effective electric field intensity, there is a recent tendency to use, as an interlevel insulation film, an insulation material having a lower dielectric constant than that of silicon dioxide in view of decreasing interconnect capacitance. However, because the low-dielectric-constant dielectric material generally is low in dielectric strength, there is still more difficulty in securing a TDDB life.
[0012] Incidentally, TDDB test is one of acceleration test methods to presume an interlevel dielectric breakdown strength, i.e., a test method for presuming an dielectric breakdown time (life) in the usual-use environment from a time to a dielectric breakdown under a high electric field at a predetermined temperature higher than that in the usual-use environment. The TDDB life is a life to be presumed from such a TDDB test. The TDDB life will be referred later.
[0013] Secondly, if a silicon nitride film is used as a cap film on the buried interconnect having a copper main-conductor layer, silicide or copper oxide is formed at the interface of copper and silicon nitride film during forming a cap film, causing a problem of increasing the resistance of the buried interconnect. The experiment by the present inventors has first found that such silicide or copper oxide is one of the major causes of copper diffusion as referred later.
[0014] Thirdly, there is a problem that strip-off occurs between the interconnect layer of the buried interconnect and an insulation film (e.g. the cap film) formed in the upper level thereof.
[0015] Fourthly, there is a problem that interconnect resistance increases due to the cleaning process after a CMP process for forming a buried interconnect (hereinafter, referred also to as post-CMP cleaning). It has been found that this problem is particularly conspicuous where a buried interconnect is formed by a plating method.
[0016] It is an object of the invention to provide a technique capable of improving the dielectric breakdown strength between the interconnects having copper in the main conductor layers thereof.
[0017] Another object of the invention is to provide a technique capable of preventing against resistance increase in the interconnect having copper main-conductor layer.
[0018] Another object of the invention is to provide a technique improve the adhesion between an interconnect having copper main conductor layer and a cap film.
[0019] The foregoing and other objects and novel features of the invention will be made apparent from the description of the present Specification and the appended drawings.
[0020] The inventions as disclosed in the present Application, if briefly explained on the summary of the representative ones thereof, include the followings.
[0021] Namely, the present invention has a step of carrying out a plasma process on a semiconductor substrate in a hydrogen gas atmosphere and a step of carrying out a plasma process on the semiconductor substrate in an ammonia gas atmosphere, after forming by polishing technique an interconnect including copper as a main conductor layer in an interconnect opening formed in a first insulation film on the semiconductor substrate but before depositing a second insulation film on the first insulation film and interconnect.
[0022] Also, the invention has a step of carrying out a heating process in a hydrogen gas atmosphere, a step of carrying out on a semiconductor substrate a plasma process in an ammonia gas atmosphere and a step of carrying out a plasma process in a hydrogen gas atmosphere, after forming by polishing technique an interconnect including copper as a main conductor layer in an interconnect opening formed in a first insulation film on the semiconductor substrate but before depositing a second insulation film on the first insulation film and interconnect.
[0023] Also, the invention has a step of carrying out on a semiconductor substrate a cleaning process using an alkali solution, a deoxidizing process and a cleaning process using the acid solution and a step of depositing a second insulation film on the first insulation film and interconnect, after forming by polishing technique an interconnect including copper as a main conductor layer in an interconnect opening formed in a first insulation film on the semiconductor substrate.
[0024] Also, the invention has a step of carrying out a cleaning process using an alkali solution, a deoxidizing process and a cleaning process using the acid solution and a step of carrying out a plasma process in a hydrogen gas atmosphere to the substrate, a step of carrying out a plasma process on the semiconductor substrate in an ammonia gas atmosphere and a step of depositing a second insulation film on the first insulation film and interconnect, after forming by polishing technique an interconnect including copper as a main conductor layer in an interconnect opening formed in a first insulation film on the semiconductor substrate.
BRIEF DESCRIPTION OF THE DRAWINGS
[0025]
FIG. 1 shows a sample used in TDDB life measurement in the present Application, wherein
[0026]
FIG. 1A is a plan view and FIGS. 1B and 1C are explanatory views showing respectively a section on line B-B′ and a section on line C-C′ in FIG. 1A;
[0027]
FIG. 2 is a concept view showing the outline of measurement;
[0028]
FIG. 3 is one example of current-voltage measurement result;
[0029]
FIG. 4 is an explanatory view of coverage in an interconnect trench or connection hole in a conductive barrier film;
[0030]
FIGS. 5A and 5B are sectional views of a buried interconnect for explaining the technical problem studied by the present inventors;
[0031]
FIG. 6 is a sectional view of a buried interconnect for explaining the technical problem studied by the present inventors;
[0032]
FIG. 7 is a graph showing the data measured on the TDDB characteristics of a copper interconnect, an aluminum interconnect and a tungsten interconnect;
[0033]
FIG. 8 is a graph showing a TDDB life;
[0034]
FIG. 9 is a graph showing a TDDB life;
[0035] FIGS. 10A-10D are graphs showing XPS data;
[0036] FIGS. 11A-11D are graphs showing XPS data;
[0037] FIGS. 12A-12D are graphs showing XPS data;
[0038] FIGS. 13A-13D are graphs showing XPS data;
[0039] FIGS. 14A-14D are graphs showing mass analysis results;
[0040] FIGS. 15A-15D are graphs showing mass analysis results;
[0041]
FIG. 16 is a graph showing an interconnect resistance;
[0042]
FIG. 17A is a sectional view due to tracing of a TEM picture showing an interconnect part in the case of no processing and
[0043]
FIG. 17B is a sectional view due to tracing of a TEM picture showing an interconnect part in the case of carrying out an ammonia plasma process;
[0044] FIGS. 18A-18C sectional view due to tracing of a TEM picture showing for comparison;
[0045]
FIG. 19A and 19B are explanatory views showing a mechanism of TDDB deterioration;
[0046]
FIG. 20A and 20B are explanatory views showing a mechanism of TDDB improvement;
[0047]
FIG. 21 is a graph showing a TDDB life;
[0048]
FIG. 22 is a flowchart showing a fabrication method of a semiconductor integrated circuit device as one embodiment of the invention;
[0049]
FIG. 23A is a plan view of an essential part in a fabrication process of the semiconductor integrated circuit device as one embodiment of the invention and
[0050]
FIG. 23B is a sectional view on line X-X in FIG. 23A;
[0051]
FIG. 24A is a plan view of an essential part in a fabrication process of the semiconductor integrated circuit device following FIG. 23 and
[0052]
FIG. 24B is a sectional view on line X-X in FIG. 23A;
[0053]
FIG. 25 is a sectional view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 24;
[0054]
FIG. 26 is a sectional view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 25;
[0055]
FIG. 27 is a sectional view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 26;
[0056]
FIG. 28 is a sectional view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 27;
[0057]
FIG. 29 is a graph showing a result of TDDB characteristic;
[0058]
FIG. 30 is a graph showing a dependency of an interconnect resistance on hydrogen anneal;
[0059]
FIG. 31 is a sectional view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 28;
[0060]
FIG. 32 is a sectional view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 31;
[0061]
FIG. 33 is a sectional view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 32;
[0062]
FIG. 34 is a graph showing a TDDB characteristic at a time that a hydrogen plasma process and an ammonia plasma process have been carried out in combination;
[0063]
FIG. 35 is an essential part sectional view of one example of a semiconductor integrated circuit device;
[0064]
FIG. 36 is an explanatory view of an overall structural example of a CMP apparatus used in one embodiment of the invention;
[0065]
FIG. 37 is an explanatory view of a structural example of a polish processing section in the CMP apparatus of FIG. 36;
[0066]
FIG. 38 is an explanatory view of a structural example of a post-cleaning processing section in the CMP apparatus of FIG. 36;
[0067]
FIG. 39 is an explanatory view of a brush of the post-cleaning processing section of FIG. 38;
[0068]
FIGS. 40A and 40B are explanatory views of another post-cleaning processing scheme;
[0069]
FIGS. 41A and 41B are explanatory views of still another post-cleaning processing scheme;
[0070]
FIGS. 42A and 42B are explanatory views of an overall structural example of a plasma processing apparatus used in one embodiment of the invention;
[0071]
FIG. 43 is a flowchart of a fabrication method of a semiconductor integrated circuit device of another embodiment of the invention;
[0072]
FIG. 44 is a flowchart of a fabrication method of a semiconductor integrated circuit device of another embodiment of the invention;
[0073]
FIG. 45 is a flowchart of a part of a fabrication method of a semiconductor integrated circuit device of another embodiment of the invention;
[0074]
FIG. 46 is a flowchart of a part of a fabrication method of a semiconductor integrated circuit device of another embodiment of the invention;
[0075]
FIG. 47 is a flowchart of a part of a fabrication method of a semiconductor integrated circuit device of another embodiment of the invention;
[0076]
FIG. 48 is a flowchart of a part of a fabrication method of a semiconductor integrated circuit device of another embodiment of the invention;
[0077]
FIG. 49 is a flowchart of a part of a fabrication method of a semiconductor integrated circuit device of another embodiment of the invention;
[0078]
FIG. 50 is a flowchart of a part of a fabrication method of a semiconductor integrated circuit device of another embodiment of the invention;
[0079]
FIG. 51 is a flowchart of a part of a fabrication method of a semiconductor integrated circuit device of still another embodiment of the invention;
[0080]
FIG. 52 is an explanatory view showing one example of an overall structure of a CMP apparatus to be used in a fabrication method of a semiconductor integrated circuit device of another embodiment of the invention;
[0081]
FIG. 53 is an explanatory view showing one part of a CMP apparatus to be used in forming a buried interconnect;
[0082]
FIG. 54 is an explanatory view of the CMP apparatus showing a state of polishing a copper main-conductor film;
[0083]
FIG. 55 is an essential-part sectional view of a semiconductor substrate in a fabrication process of a semiconductor integrated circuit device in another embodiment of the invention;
[0084]
FIG. 56A is an essential-part plan view of the semiconductor substrate showing a fabrication method of the semiconductor integrated circuit device following FIG. 55 and
[0085]
FIG. 56B is an essential-part sectional view in FIG. 56A;
[0086]
FIG. 57 is an essential-part sectional view of the semiconductor substrate in the fabrication process of the semiconductor integrated circuit device following FIG. 56;
[0087]
FIG. 58A is an essential-part plan view of the semiconductor substrate showing the fabrication method of the semiconductor integrated circuit device following FIG. 57 and
[0088]
FIG. 58B is an essential-part sectional view in FIG. 58A;
[0089]
FIG. 59 is an essential-part sectional view of the semiconductor substrate in the fabrication process of the semiconductor integrated circuit device following FIG. 58;
[0090]
FIG. 60A is an essential-part plan view of the semiconductor substrate showing the fabrication method of the semiconductor integrated circuit device following FIG. 59 and
[0091]
FIG. 60B is an essential-part sectional view in FIG. 60A;
[0092]
FIG. 61 is a flowchart showing the fabrication method of the semiconductor integrated circuit device explained in FIG. 52 to FIG. 60;
[0093]
FIG. 62 is a graph showing a TDDB life;
[0094]
FIG. 63 is a graph showing a TDDB life;
[0095]
FIG. 64 is an essential-part sectional view of a semiconductor substrate in a fabrication process of a semiconductor integrated circuit device in another embodiment of the present invention;
[0096]
FIG. 65A is an essential-part plan view of the semiconductor substrate showing the fabrication method of the semiconductor integrated circuit device following FIG. 64 and
[0097]
FIG. 65B is an essential-part sectional view;
[0098]
FIG. 66 is an essential-part sectional view of the semiconductor substrate in the fabrication process of the semiconductor integrated circuit device following FIG. 65;
[0099]
FIG. 67 is a graph showing an amount of silicon contained in a copper interconnect upon carrying out each process;
[0100]
FIG. 68 is a graph showing a conductive-barrier film thickness dependency in the resistance of a buried copper interconnect;
[0101]
FIG. 69 is a graph showing a dependency of a TDDB characteristic on conductive-barrier film thickness;
[0102]
FIG. 70 is a graph showing a TDDB characteristic after anneal in the case of the absence of a conductive barrier film and of a thickness of smaller than 10 nm;
[0103]
FIGS. 71A and 71B are essential-part sectional views of a copper buried interconnect layer of a semiconductor integrated circuit device of another embodiment of the invention;
[0104]
FIG. 72A is a plan view of an essential part in a fabrication process of a semiconductor integrated circuit device of an embodiment of the invention and
[0105]
FIG. 72B is a sectional view on line A-A in FIG. 72A;
[0106]
FIG. 73A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 72 and
[0107]
FIG. 73B is a sectional view on line A-A in FIG. 73A;
[0108]
FIG. 74A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 73 and
[0109]
FIG. 74B is a sectional view on line A-A in FIG. 74A;
[0110]
FIG. 75A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 74 and
[0111]
FIG. 75B is a sectional view on line A-A in FIG. 75A;
[0112]
FIG. 76A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 75 and
[0113]
FIG. 76B is a sectional view on line A-A in FIG. 76A;
[0114]
FIG. 77A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 76 and
[0115]
FIG. 77B is a sectional view on line A-A in FIG. 77A;
[0116]
FIG. 78A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 77 and
[0117]
FIG. 78B is a sectional view on line A-A in FIG. 78A;
[0118]
FIG. 79A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 78 and
[0119]
FIG. 79B is a sectional view on line A-A in FIG. 79A;
[0120]
FIG. 80A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 79 and
[0121]
FIG. 80B is a sectional view on line A-A in FIG. 80A;
[0122]
FIG. 81A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 80 and
[0123]
FIG. 81B is a sectional view on line A-A in FIG. 81A;
[0124]
FIG. 82A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 81 and
[0125]
FIG. 82B is a sectional view on line A-A in FIG. 82A;
[0126]
FIG. 83A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 82 and
[0127]
FIG. 83B is a sectional view on line A-A in FIG. 83A;
[0128]
FIG. 84A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 83 and
[0129]
FIG. 84B is a sectional view on line A-A in FIG. 84A;
[0130]
FIG. 85A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 84 and
[0131]
FIG. 85B is a sectional view on line A-A in FIG. 85A;
[0132]
FIG. 86A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 85 and
[0133]
FIG. 86B is a sectional view on line A-A in FIG. 86A;
[0134]
FIG. 87A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 86 and
[0135]
FIG. 87B is a sectional view on line A-A in FIG. 87A;
[0136]
FIG. 88A is a plan view of the essential part in the fabrication process of the semiconductor integrated circuit device following FIG. 87 and
[0137]
FIG. 88B is a sectional view on line A-A in FIG. 88A; and
[0138]
FIGS. 89A and 89B are essential- part sectional views of a copper buried interconnect layer of a semiconductor integrated circuit device of still another embodiment of the invention.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0139] For explaining the embodiments of the present invention, the terms in this Specification if explained have their meanings in the followings.
[0140] 1. TDDB (Time Dependence on Dielectric Breakdown) life is a time (life) determined by extrapolation to an electric field intensity in actual use (e.g. 0.2 MV/cm) from a graph prepared by applying a comparatively high voltage to between electrodes under a measuring condition at a predetermined temperature (e.g. 140° C.) to plot a time from voltage application up to dielectric breakdown against an applied electric field. FIG. 1 shows a sample used in a TDDB-life measurement in the present Application, wherein FIG. 1A shows a plan view and FIGS. 1B and 1C show, respectively, a section on B-B′ line and a section on C-C′ line in FIG. 1A. The sample in actual can be formed in a TEG (Test Equipment Group) region on a semiconductor wafer (hereinafter referred merely to as wafer). As shown in the figure, a pair of comb-shaped interconnects L are formed in a second interconnect level M2 and connected respectively to the pads P1, P2 in the uppermost level. An electric field is applied to between the comb-shaped interconnects L to measure a current. The pads P1, P2 are measuring terminals. The comb-shaped interconnect L is 0.5 μm in any of interconnect width, interconnect spacing and interconnect thickness. Also, the interconnect opposite length was provided 1.58×105 μm. FIG. 2 is a concept figure showing the outline of measurement. The sample is held on a measuring stage S, and a current-voltage measuring instrument (I/V measuring instrument) is connected between the pads P1, P2. The sample stage S is heated by a heater H in order to adjust a sample temperature at 140° C. FIG. 3 is one example of a current-voltage measurement result. Exemplification is on the case of a sample temperature 140° C. and an electric field intensity of 5 MV/cm. Although the TDDB life measurement includes a constant-voltage stress scheme and a low-current stress scheme, the present Application uses a constant-voltage stress scheme that the mean electric field applied to an insulation film is constant. After voltage application, current density decreases with lapse of time, and thereafter an abrupt current increase (dielectric breakdown) is observed. Herein, the time that the leak current density reaches 1 RA/cm2 is given a TDDB life (TDDB life at 5 Mv/cm). Incidentally, although the TDDB life refers to a breakdown time (life) at 0.2 MV/cm unless otherwise referred to, there is a case that the term of TDDB life is used in broad meaning as a time to a breakdown while mentioning to a predetermined electric field intensity. Also, unless otherwise mentioned, the TDDB life is at a sample temperature 140° C. Incidentally, although the TDDB life refers to a measurement on a comb-shape interconnect L as in the foregoing, it is needless to say that it reflects an actual interlevel breakdown life.
[0141] 2. Plasma process refers to a process that a plasma chemical or mechanical (bombardment) action is applied to an exposed surface of a substrate or, a member when an insulation film, metal film or the like is formed on a substrate, in a plasma-state environment. Generally, although plasma is caused by ionizing a gas through an action of a high-frequency electric field or the like while supplementing, as required, a process gas to a reaction chamber having a particular gas (step gas) replaced, perfect replacement to a process gas is practically impossible to achieve. Accordingly, in the present Application, ammonia plasma when referred to, for example, is not intended to mean perfect ammonia plasma and accordingly the existence of an impurity gas in the plasma (nitrogen, oxygen, carbon dioxide, water vapor, etc.) is not to be excluded. Similarly, the inclusion of other dilution gas or additive gas in plasma is not excluded as a matter of course.
[0142] Reduction-atmosphere plasma refers to a plasma environment that there predominantly exist reaction species, such as radicals, ions, atoms and molecules, having a deoxidizing action or action to withdraw oxygen, wherein radical and ion include atomic or molecular-formed radicals or ions. Also, it is satisfactory that, besides a single species, species in a plurality of kinds may be included in the environment. For example, environment may have hydrogen radical and NH2 radical existing simultaneously.
[0143] 3. If in the present Application there is an expression, for example, of comprising copper, it is intended to use copper as a major component. Namely, it is natural that, generally, copper even if high purity yet contains impurities. The additives or impurities included in a member comprising copper are not to be excluded. In the Specification, the expression of comprising copper in high purity is intended to mean copper comprising a material high in purity (e.g. 4N (99.99%)) or higher, wherein it is premised to contain an arbitrary impurity in an amount of nearly 0.01%. This is not limited to copper but is true for other metals (titanium nitride, etc.).
[0144] 4. The gas concentration, in the Specification, refers to a flow rate ratio in mass flow rate. Namely, the concentration of a gas A of 5% in a gas mixture of gases A and B refers to Fa/(Fa+Fb)=0.05 provided that the mass flow rate of the gas A is Fa and the mass flow rate of the gas B is Fb.
[0145] 5. Chemical Mechanical Polish (CMP) generally refers to polishing of a polish surface by relative movement in a direction of the surface while supplying a slurry in a state relatively contacted with a polishing pad formed of a sheet material in a soft cloth form. Besides this, the present Application includes CML (Chemical Mechanical Lapping) for polishing the polish surface while being relatively moved to a surface of a hard grindstone, those using fixed grits and grit-free CMP without using grits.
[0146] 6. Grit-free chemical mechanical polish generally refers to chemical mechanical polish using a slurry having a grit weight concentration of 0.5% by weight or less. Grit-used chemical mechanical polish refers to chemical mechanical polish using a slurry having a grit weight concentration of 0.5% by weight or higher. However, these are relative in nature. Where the first-step polish is by grit-free chemical mechanical polish and the following second-step polish is by grit-used chemical mechanical polish or where the first-step polish concentration is smaller by one order of magnitude, preferably by two orders of magnitude, than the second-step polish concentration, there are cases that the first-step polish is referred to as grit-free chemical mechanical polish. In this Specification, grit-free chemical mechanical polish includes not only the case that unit planarizing process overall for a subjective metal film is made by grit-free chemical mechanical polish but also the case the major process is made by grit-free chemical mechanical polish and the minor process is by grit-used chemical mechanical polish.
[0147] 7. Polishing fluid (slurry), generally, refers to a suspension having polish grits mixed in a chemical etchant, and the present Application includes those not mixed with polishing grits in the nature of the invention.
[0148] 8. Grits (slurry particles) generally refers to a powder, such as alumina and silica, contained in slurry.
[0149] 9. Anticorrosive refers to an agent which inhibits or suppresses CMP polish from proceeding by forming a protection film having a corrosive or hydrophobic nature or the both on the surface of a metal. Generally, benzotriazole (BTA) or the like is used (for detail, see JP-A-8-64594).
[0150] 10. The conductive barrier film is generally a conductive film having a diffusion barrier property formed comparatively thin on a side surface or bottom of a buried interconnect in order to prevent copper from diffusing into an interlayer insulation film or the lower-leveled layer. Generally, used is a refractory metal or nitride thereof, such as titanium nitride (TiN), tantalum (Ta) and tantalum nitride (TaN).
[0151] 11. The buried interconnect or buried metal interconnect generally refers to an interconnect patterned by an interconnect-forming technique for removing an unwanted portion of a conductive film on an insulation film after burying a conductive film inside a trench or hole formed in the insulation film, as in single damascene or dual damascene. Meanwhile, generally, single damascene refers to an interconnect burying process for burying a plug metal and a interconnect metal in two stages. Similarly, dual damascene generally refers to an interconnect burying process for burying a plug metal and an interconnect metal at one time. Generally, there are often cases that a copper-buried interconnect is used in a multi-level structure.
[0152] 12. Selective removal, selective polish, selective etch or selective chemical mechanical polish, when referred to, in any has a selective ratio of 5 or greater.
[0153] 13. Concerning selective ratio, when referred to a selective ratio “of A to B” (or “of B to A”) is X, it means that, where taking an example of a polish rate, the calculation of a polish rate for A with reference to a polish rate for B provides X.
[0154] 14. The semiconductor integrated circuit device when referred to in the present Application includes not only those particularly fabricated on single-crystal silicon substrates but, in excepting the case explicit not so, also those fabricated on other substrates such as SOI substrates or TFT (Thin Film Transistor) liquid crystal manufacturing substrates. Also, the wafer refers to a single-crystal silicon substrate (generally in a disk form) used in the manufacture of a semiconductor integrated circuit device, an SOS (Silicon On Sapphire) substrate, a glass substrate, other insulating, semi-insulating or semiconductor substrates or a combined substrate thereof.
[0155] 15. The wafer (semiconductor integrated circuit substrate or semiconductor substrate) refers to silicon or other semiconductor single-crystal substrate used in the manufacture of semiconductor integrated circuits (generally, nearly in a flat disk form), sapphire substrate, glass substrate, other insulation or semi-insulation semiconductor substrate or a combined substrate thereof. Incidentally, the substrate in its part or entirety or the gate electrode in its entirety or part may be formed of another semiconductor, e.g. SiGe.
[0156] 16. The semiconductor integrated circuit chip (semiconductor integrated circuit substrate) or semiconductor chip (semiconductor substrate) refers to those that a wafer completed of wafer process is divided into unit-circuit groups.
[0157] 17. Silicon nitride or a silicon nitride film, when referred to, includes not only Si3N4 but also a silicon-nitride insulation film having the similar composition.
[0158] 18. The cap film is an insulation film having high insulation and diffusion barrier properties formed in a position above a buried interconnect excepting an electrical connection, and generally formed of a material other than that of the major portion of the interlayer insulation film, e.g. silicon nitride film.
[0159] 19. The wafer process is called also a first-half process, which is a process of from a mirror-polished wafer (mirror wafer) state through a device-and-interconnect forming process to a state that a surface protection film is formed to finally enable electric tests with a probe.
[0160] 20. The long throw sputtering technique refers to a sputtering technique, one method for improving bottom coverage, for allowinb stable discharge at reduced pressure by putting the target-to-substrate spacing distant in order to allow only the vertical component of sputtering particles to reach the substrate.
[0161] 21. The collimate sputtering technique refers to a sputtering technique having a mechanism for forcibly enhancing the vertical component by inserting a grit-formed plate between a target and a substrate in order to obtain a sufficient film thickness even at a bottom when forming a film in an interconnect opening, such as an interconnect trench, contact hole or through-hole, great in aspect ratio.
[0162] 22. The coverage in a conductive-barrier-film interconnect opening (interconnect trench, contact hole, through-hole or the like) has side coverage and bottom coverage. FIG. 4 shows typically a state that a barrier film 62 is deposited by a sputtering technique, on an upper surface of an insulation film 60 as well as in an interconnect trench 61 formed in the insulation film 60. The depo-film thickness of a barrier film, when referred to, generally refers to a film thickness D1 of a barrier film 62 on a top surface of the insulation film 60. Side coverage refers to coverage of the barrier film 62 on a sidewall (including the intersecting corner at between the side surface and the bottom surface) of the interconnect trench 61, wherein the film thickness D2 is minimum at that point. Meanwhile, bottom coverage refers to a coverage of the barrier film 62 in the bottom of the interconnect trench 61, wherein the film thickness D3 is the greatest next to the depo-film thickness. For example, according to a result of experiment by the present inventors, where a barrier film was deposited in an interconnect trench having an aspect ratio, for example, of 1 by the usual sputtering technique without especial consideration to directivity, the barrier film had a depo-film thickness of 100 nm, a side coverage of nearly 30 nm and a bottom coverage of nearly 50 nm. Meanwhile, where a barrier film was deposited by the long sputtering technique, the barrier film had a depo-film thickness of 100 nm, a side coverage of nearly 20 nm and a bottom coverage of nearly 90 nm.
[0163] The low-dielectric-strength insulation film or insulation material is to be defined as an insulation film having a dielectric constant lower than a dielectric constant of a silicon oxide film (e.g. TEOS (Tetraethoxysilane) oxide film) included in a protection film formed as a passivation film. Generally, the insulation film having a dielectric constant nearly equal to or lower than a dielectric constant of ε=4.1-4.2 of a TEOS oxide film is referred to as a low-dielectric-constant insulation film.
[0164] In the below embodiments, explanations will be made by dividing with a plurality of sections or embodiments. It is however noted that, unless otherwise explicitly noted, they are in a relationship that one is a modification, detail, supplementary explanation to a part or the entire of another, instead of not related to one another.
[0165] Also, in the below embodiment, where reference is made to the number of elements or the like (including number, numeral, quantity, range and the like), the particular number is not limited to, i.e. the number may be greater than or smaller than the particular number, unless otherwise explicitly noted or clearly limited to a particular number on the principle.
[0166] Furthermore, in the below embodiments, it is needless to say that the constituent elements thereof (including element steps) are not necessarily essential except for the case of especially explicitly noted or clearly essential on the principle.
[0167] Similarly, in the below embodiment, when reference is made to a shape, positional relationship or the like of the constituent elements or the like, those substantially approximated or similar to the shape or the like are included except for the case of especially explicitly noted or clearly considered not so on the principle. This is true for the numeral and range.
[0168] Also, those having the same function are attached with the identical reference numerals throughout the drawings for explaining the embodiments, to omit repeated explanations.
[0169] Meanwhile, in the embodiments, the MISFET (Metal Insulator Semiconductor Field Effect Transistor) representative of a field-effect transistor is abbreviated merely as MIS, the p-channel MISFET is as pMIS, and the n-channel MISFET is as nMIS.
[0170] (Embodiment 1)
[0171] Explanation is now made on the problem in a post-CMP cleaning technique studied by the present inventors and the solving means therefor.
[0172] The post-CMP cleaning technique studied by the present inventors is to carry out alkali cleaning and acid cleaning in the order on a semiconductor substrate of after CMP processing. Alkali cleaning has an object to remove foreign matter, such as the slurry used upon CMP processing. In alkali cleaning, used was a cleaning liquid containing, for example, ammonia (NH4OH). Also, acid cleaning has an object to improve TDDP characteristic or remove the remaining metal. In acid cleaning, used was a cleaning liquid containing, for example, hydrogen fluoride (HF).
[0173]
FIGS. 5A and 5B shows respective examples of sections in the buried interconnects 65a, 65b that a cleaning process is carried out on the buried interconnects that are different in interconnect dimensions. The buried interconnect 65a, 65b are formed in a state of buried in an interconnect trench 67 formed in an insulation film 66, each of which is structured as a main conductor film having a conductor film based on copper (Cu).
[0174] As shown in FIG. 5, according to a result of experiment by the present inventors, the etch action on copper is intense in the cleaning process, particularly in the acid cleaning. The removal of interconnect is almost negligible in the dimensions of the buried interconnect 65a exemplified in FIG. 5A. However, scale-down was made as in the dimensions of the buried interconnect 65b exemplified in FIG. 55B, there were found problems with interconnect-resistance increase, variation, step occurrence and the like particularly in a solitary buried interconnect pattern.
[0175] Also, in the post-CMP cleaning, as shown in FIG. 6 there was found poor appearance alike corrosion (hereinafter, etch corrosion) EC in an edge of a dense buried-interconnect 65 group. This occurs also in coarse buried interconnection or solicit buried interconnection. Also, according to a result of study by the present inventors, it was found that the above various problems were especially conspicuous where the buried interconnect was formed by a plating method. The reason of etch-corrosion occurrence can to be considered that the buried interconnect formed by a plating technique is readily corroded as compared to a buried interconnect formed by a sputter technique and, in addition, the friction by the pad upon CMP processing concentrates on an edge of a dense buried interconnect, coarse buried interconnect or solitary buried interconnect to thereby promote the formation of a brittle oxide film (CuO), or the dissolution of the buried interconnect is accelerated due to the non-linear diffusion of an agent used upon post-CMP cleaning. The occurrence of etch corrosion incurs increase or variation in interconnect resistance.
[0176] For this reason, the resent embodiment carries out a deoxidizing process after CMP before post-CMP cleaning or after alkali cleaning before acid cleaning. Due to this, because the copper surface oxidized by CMP can be deoxidized, the entire copper interconnect can be improved in resistance to a cleaning liquid. Due to this, it is possible to suppress or prevent the occurrence of steps in an upper portion of a copper interconnect or etch corrosion as above. Therefore, it is possible to suppress or prevent a copper-based buried interconnect from increasing or varying in interconnect resistance.
[0177] Next, explanation is made on the problem of TDDB life deterioration and solving means therefor.
[0178] As in the foregoing, it is generally known that, in the case of using copper as an interconnect material, the TDDB life is conspicuously shortened than that of other materials (e.g. aluminum, tungsten). FIG. 7 is a graph showing the data measured on the TDDB characteristics of a copper interconnect, an aluminum interconnect and a tungsten interconnect. TDDB life is assined on the vertical axis while electric-field intensity is on the horizontal axis. If extrapolating an aluminum-interconnect characteristic (data A) and tungsten-interconnect characteristic (data B), a TDDB life at an electric field intensity of 0.2 MV/cm (normal use state) is by far in excess of 3×108 sec (10 years) as a development goal of the present inventors. On the other hand, if extrapolating a copper-interconnect characteristic (data C), it is to be known that almost no margin is available for the 10-year development goal.
[0179] In this experiment, the aluminum interconnect was formed by patterning using film deposition and photolithography while the tungsten interconnect was formed by a damascene technique similarly to the copper interconnect. Namely, the copper interconnect and the tungsten interconnect are different only in material without having any difference in structure. Nevertheless, there is conspicuous difference in TDDB characteristic, suggesting a cause of different interconnect materials. Incidentally, the TDDB characteristic herein shows the data taken at a temperature of 140° C.
[0180] It is generally considered that the cause of such TDDB life deterioration is due to diffusion of the copper applied as an interconnect material to the surrounding thereby causing decrease in the interlevel breakdown strength. However, as a result of experiment and consideration on copper diffusion phenomenon by the present inventors, the following mechanism of copper difusion phenomenon has been found for the first time.
[0181] Namely, the copper in an interconnect is more predominant than the copper in an atomic form, in respect of the factor that the ionized copper supplied from copper oxide or copper silicide drifts and diffuses on an interlevel potential. Also, the diffusion route of copper is predominantly at an interface between the insulation layer formed with a copper interconnect and the cap film. Namely, copper oxide or copper silicide is formed on a surface of a copper interconnect. Copper ions are formed from such a copper compound so that the ionized copper drifts and diffuses along the interface of an interconnect-forming insulation film and the cap film due to the interlevel electric field. The diffused copper atoms form a cause to increase leak current. The increase of leak current increases thermal stress, and dielectric breakdown ultimately occurs on a leak path resulting in TDDB life. Accordingly, the inventors could have greatly improved the TDDB life by carrying out a deoxidizing process such as an ammonia (NH3) plasma process on the semiconductor substrate after forming a buried interconnect by CMP before forming a cap film. As concerned with this, there is a description in Japanese Patent Application No. 226876/1999 or Japanese Patent Application No. 104015/2000. Incidentally, the result obtained in Japanese Patent Application No. 226876/1999 or Japanese Patent Application No. 104015/2000 if explained is as the following.
[0182]
FIG. 8 is a graph showing a TDDB life of a TEG sample formed in the same level as the second-leveled interconnect, wherein the data taken upon performing the ammonia plasma process is shown on a line A. For comparison, also shown is the TDDB life data taken upon not performing an ammonia plasma process (line Ref). As apparent from the figure, there is observed an improvement of life by six orders of magnitude in the case of performing the ammonia plasma process as compared to the comparative data.
[0183]
FIG. 9 shows the data (line B) taken in the case that the silicon oxide film formed with an interconnect trench for a buried interconnect is replaced to a silicon nitride film greater in density and firmness. Even where the insulation film is replaced to a silicon nitride, there is no difference from the case of replacing the insulation film to a silicon nitride film unless an ammonia plasma process is carried out (line Ref). On the other hand, if a silicon nitride film is applied for the insulation film to carry out an ammonia plasma process thereon, the TDDB life is improved furthermore. However, it can be understood that the rate of improvement is less and the factor due to the ammonia plasma process performed is more predominant. This demonstrates that the factor governing the TDDB life is more predominant at an interface of an insulation film than in the bulk thereof.
[0184] Herein, the inventors has conducted surface analysis on copper and a silicon oxide film in order to analyze the mechanism of improving the TDDB life due to an ammonia plasma process. The result of analysis and the like are explained below.
[0185]
FIG. 10 to FIG. 12 are graphs showing results of XPS (X-ray Photoelectron Spectroscopy) analysis on a copper interconnect surface. In each figure, A and C show Cu2p spectral results while B and D N1s spectral results.
[0186]
FIGS. 10A and 10B are a result of analysis of a copper film surface in an as-depo state. From the fact that a peak of Cu2p is observed and a peak of N1s is in a noise level, it is understood that no nitrogen exists in the as-depo-state copper film. FIGS. 10C and 10D are a result of analysis on a copper interconnect surface of immediately after carrying out only CMP on the copper film. There is observed a peak of N1s together with a peak of Cu2p. Because there is a case that slurry contains benzotriazole (BTA), it is to be presumed that the observation is on the nitrogen in BTA remained on the copper surface. FIGS. 11A and 11B are a result of analysis on a copper interconnect surface in a state of having done to post cleaning after CMP. No change is seen in a Cu2p peak but a Nls peak is lowered. It can be considered that BTA has been removed by the cleaning. FIGS. 11C and 11D are a result of analysis on a copper interconnect surface in a state having been allowed to stand for 24 hours in an air atmosphere after post cleaning. A peak is observed in copper oxide (CuO), together with a peak in Cu2p. In the N1s peak, a change due to allowing to stand is not observed. It can be seen that the copper surface is oxidized due to allowing to stand to produce copper oxide.
[0187]
FIGS. 12A and 12B are a result of analysis on a copper interconnect surface in a state an ammonia plasma process is carried out on the copper interconnect thus oxidized. The peak in the copper oxide nearly goes out. On the other hand, an N1s peak occurs strongly. It can be considered that the copper surface is deoxidized and oxygen is drawn out and further the surface is nitrided. For comparison, analysis was made on a copper interconnect surface in a state a hydrogen heating process at 350° C. has been carried out on an oxidized copper interconnect. The result is given in FIGS. 12C and 12D. Comparing the Cu2p peaks in between FIG. 12C and FIG. 12A, it can be considered that deoxidizability is stronger in the hydrogen heating process because of the fact closer to the as-depo state (FIG. 10A). On the other hand, the copper surface is merely deoxidized in the hydrogen heating process in view of almost no N1s peak observed.
[0188] From the above results, it can be understood that the copper interconnect surface is deoxidized by the ammonia plasma process and further a nitrogen layer is formed. It can be considered that this nitrogen layer acts to prevent copper from reacting with the silane contained in a material gas used upon depositing a silicon nitride film after the ammonia plasma process and to suppress copper from forming a silicide. The prevention of silicide formation has an effect to suppress interconnect resistance from increasing.
[0189]
FIG. 13 is graphs showing a result of XPS analysis conducted on a silicon oxide film surface. FIG. 14 and FIG. 15 are graphs showing a result of mass spectroscopy (TDS-APIMS) conducted on a silicon oxide film. In the analysis of the silicon oxide film, analyses were made on a state of having done up to cleaning after CMP (profile C), a state a hydrogen plasma process has been made after post-CMP cleaning (profile D), a state an ammonia plasma process has been made after post-CMP cleaning (profile E), and a state a nitrogen plasma process has been made after post-CMP cleaning (profile F). Incidentally, the deviation of the profile C by approximately 1 eV toward the higher energy is due to the affection of charge-up.
[0190]
FIGS. 13A and 13B are both on the data that an Si2p spectrum has been observed, wherein FIG. 13A is on the analysis in a depth of approximately 10 nm while FIG. 13B is in a depth of approximately 2 nm. FIGS. 13C, 13D and 13E are on the data observing N1s, O1s and C1s spectrums, respectively. From FIG. 13B, a broad peak is seen at a lower energy side (at around 102 eV) of a hydrogen plasma process (profile D). This can be considered due to an existence of Si—H bonds, i.e. Si—H is presumably formed on a silicon oxide film surface by the hydrogen plasma process. In FIG. 13A, the ammonia plasma process (profile E) and the nitrogen plasma process (profile F) have their 105-eV peaks each asymmetrically broadened toward the lower energy side. The peak in the asymmetric portion (103.5 eV) is considered due to Si—O—N bonds. It is presumed that the silicon oxide film surface is nitrided by the ammonia plasma process and nitrogen plasma process. Meanwhile, from comparison between FIGS. 13A and 13B, nitriding is considered stronger in the surface. The nitriding due to the ammonia plasma process and nitrogen plasma process can be confirmed also in FIG. 13C. In FIG. 13E, almost no carbon is detected in the hydrogen plasma process (profile D). It is understood that the organic matter in the surface has been removed by the hydrogen plasma process. Also, the 289-eV peak after the CMP (profile C) is considered due to C—O bonds. It is considered that slurry remains after the CMP. FIG. 13F shows an N-amount presumed value by determining an existence ratio of them from an Si peak and an N peak. It is considered that nearly equivalent nitriding has been done in the ammonia plasma process and the nitrogen plasma process.
[0191]
FIGS. 14A, 14B, 14C and 14D are, respectively, graphs of measurements on a mass number of 41 (Ar—H), a mass number of 27 (C2 H3), a mass number of 57 (C4 H9) and a mass number of 59 (C3 H7 O). Also, FIGS. 15A, 15B, 15C and 15D are, respectively, graphs of measurements on a mass number of 28 (Si, C2 H4), a mass number of 44 (SiO, C3 H6), a mass number of 29 (SiH, C2 H5) and a mass number of 31 (SiH3).
[0192] It is understood from FIG. 14A that there is almost no difference in desorption amount of hydrogen but the desorbing temperature in the hydrogen plasma process (profile D) is 520° C. that is lower as compared to the other cases (560° C.). In FIGS. 14A, 14B and 14C, desorption of organic matter is seen in each process.
[0193] On the other hand, in FIGS. 15A-15D there are seen the existences of peaks of other than desorption of organic matter. That is, the peaks at 300° C.-400° C. seem to be Si, SiO, SiH and SiH3. Comparing the figures, SiO desorption is seen in each of hydrogen, ammonia and nitrogen plasmas but SiH/SiH3 desorption is not obserbed in the ammonia plasma process. Namely, in the ammonia plasma process, Si—O—N is formed which is to be desorbed at a comparatively low energy. Also, the energy required for desorption is the highest in the nitrogen plasma process, and almost the same in the hydrogen plasma process and the ammonia plasma process.
[0194] From these results, Si—OH and Si—O—, forming a cause of dangling bond in the silicon oxide film surface, is considered to terminate at a weak bond Si—O—N by the ammonia plasma process. In the formation of a silicon nitride film after the ammonia plasma process, Si—O—N immediately close to the surface desorbs so that the Si—O bond in the bulk and the Si—N in the silicon nitride film firmly bond together to form a continuous interface. This is considered a mechanism to improve the adhesion at the interface. On the other hand, where an ammonia plasma process is not carried out, it can be considered that condensation reaction takes place between the surface of a silicon oxide film abundant in Si—OH bond and ammonia as a material gas for a silicon nitride film thereby causing a number of Si—O-bonds as a cause of dangling bonds. If there exist a number of dangling bonds at the interface of the silicon oxide film and the silicon nitride film, it can be considered that a leak path is formed there forming a cause of leak current between interconnects and eventually of dielectric breakdown.
[0195] From the above analysis results, it can be considered that the oxidized copper interconnect surface is deoxidized and converted into cu single element in a state electrically stable than ionized copper and the interface of the silicon oxide film/silicon nitride film becomes a continuous firm film thereby decreasing the leak current and greatly improving the TDDB life.
[0196] The inventors have taken TEM pictures of the interface of the interconnect layer and the silicon nitride film (cap film) in the cases of an ammonia plasma process has been carried out and not been carried out. As a result, in the case of the present embodiment having carried out an ammonia plasma process, a thin film existence at the interface could have been confirmed. It can be considered that the thin film is a nitriding layer mentioned before. On the other hand, where an ammonia plasma process is not carried out, such a film cannot be confirmed.
[0197] Also, the present embodiment can reduce the resistance of the Cu interconnect. FIG. 16 is a result of measurement of interconnect resistances in the case of carrying out various processes. In the cases of no process (not carrying out a plasma process) and carrying out a plasma process, the values are significantly low as compared to those of the other cases (hydrogen plasma process, hydrogen anneal, nitrogen plasma process). FIG. 17 and FIG. 18 are trace figures of the TEM pictures observing the interface of a Cu interconnect and a cap film (silicon nitride film) in the case of carrying out each of these processes.
[0198] In the cases of no process and an ammonia plasma process (FIG. 17), no differential one is seen at the interface. However, in the case of hydrogen anneal and nitrogen plasma processes (FIG. 18), a copper silicide (CuSi) layer is formed at the interface. This silicide layer seems to be a cause of resistance increase. Such a silicide layer is formed in the reaction with a silane gas upon forming a silicon nitride film (cao film). However, where an ammonia plasma process has been carried out, an extremely thin silicon nitride film is formed on a copper surface. This silicon nitride film is considered to serve as a blocking layer against silicide formation. On the other hand, it can be considered that, by only deoxidizing the copper surface, an active copper surface is exposed and promoted for reaction with silicon thereby readily forming a silicide layer. Meanwhile, in the case of a hydrogen plasma process (FIG. 18C), any formation is seen in the interface. However, there are many cases of not forming such a formation. It can be considered that silicidation degree is low in the case of a hydrogen plasma process.
[0199] It has been found for the first time by a study result of the present inventors that the following model can be considered from the foregoing analysis results. FIG. 19A shows a schematic illustration of a TDDB deterioration mechanism, and FIG. 19B shows an energy band of the same. Namely, where the ammonia plasma process is not carried out, the surface of the copper interconnect is influenced by the subsequent surface processes to form thereon copper oxide (CuO), and copper silicide (Cu compound) upon forming a cap film (silicon nitride film) 68. Such copper oxide or copper silicide is ready to ionize as compared to pure copper. Such ionized copper is drifted by an interlevel electric field and diffused into an interlevel insulation film.
[0200] Meanwhile, where the ammonia plasma process is not carried out, the interface of between an insulation film (silicon oxide film) 66 forming a copper-based buried interconnect 65 and a cap film (silicon nitride film) 68 is formed with a number of CMP damages, organic matters or dangling bonds, thus being non-continuous and insufficient in adhesion. The presence of such dangling bonds have an action to promote the diffusion of copper ions so that the copper ions are drifted and diffused along the interface. Namely, a leak path is formed at the interface between the interconnects. The leak current flowing through the leak path is added with an long-time leak action and the thermal stresses due to current, and thereafter acceleratedly increased in current value, resulting in dielectric breakdown (lowered TDDB life).
[0201] Contrary to this, FIGS. 20A and 20B show a schematic figure of a TDDB-improvement mechanism and an energy band of the same where the ammonia plasma process has been carried out. Because of carrying out an ammonia plasma process on a surface of a copper-based buried interconnect 65, an oxide layer on the surface of the buried interconnect 65 is deoxidized to vanish. Because a thin nitride layer is formed on the surface of the buried interconnect 65, a copper silicide is not formed upon forming a silicon nitride film 68. Due to this, it is possible not to cause a source substance predominantly supplying copper ions forming a cause of current leak and dielectric breakdown. Also, because the ammonia plasma process is carried out onto the surface of an insulation film (silicon oxide film) 66, it is possible to make the connection to the cap film (silicon nitride film) 68 continuous and decrease the density of dangling bonds thus suppressing the formation of a leak path. Also, the surface of the insulation film 66 can be made in a clean state. Accordingly, it is possible to suppress the occurrence of copper ions forming a cause of lowering the TDDB life and to form such a junction interface of the insulation film 66 and the cap film 68 as can suppress the diffusion of copper. This can improve the TDDB life.
[0202] Also, it can be considered from the foregoing analysis that the TDDB life is to be improved also by a hydrogen plasma process. Namely, a hydrogen plasma process deoxidizes the copper surface thereby terminating, at Si—H, the dangling bonds, such as Si—O—, and Si—OH as a cause thereof. When forming a cap film (silicon nitride film), the weak-bonded Si—H in the surface is desorbed and replaced with Si—N. This forms a continuous interface of insulation film (silicon oxide film) 66 and a cap film (silicon nitride film) 68. However, the interconnect resistance increases as in the foregoing.
[0203]
FIG. 21 is a graph showing the data of TDDB life where a hydrogen plasma process is carried out. For reference, a line Ref (no-processing) and a line A (ammonia plasma process) were shown. With a hydrogen plasma process (line C), it is understood that the TDDB life is improved greatly. In the case of a hydrogen plasma, it is expected that plasma damage be relieved. Accordingly, this is extremely effective when such a material can be applied that is another material replacing, as a cap film, a silicon nitride film and not to cause a reaction product with Cu. Incidentally, in a nitrogen plasma process (line D), the TDDB life is rather lowered. As understood from FIGS. 14 and 15, this seems to be caused by rather an increase of organic-matter deposition due to the nitrogen plasma process. Furthermore, because the adhesion between the buried interconnect 65 and insulation film 66 and the cap film 68 is improved, there is an effect to increase the strip-resisting strength at the interface and increase the margin.
[0204] In the ammonia plasma process, a single gas of ammonia or hydrogen is not limited to but processing may be due to a mixture-gas plasma with an inert gas of nitrogen, argon, helium or the like, i.e., a mixture gas of ammonia and hydrogen, nitrogen, argon or helium or a mixture gas of hydrogen and ammonia, nitrogen, argon or helium. Furthermore, it is satisfactory to use a mixture gas in a hypercomplex system of a ternary or quarternary system as selected from those gases. In this case, hydrogen, ammonia or hydrogen plus ammonia in sum is required mixed by 5% or more with respect to a total flow rate (mass flow rate).
[0205] The present inventors have qualitatively analyzed the TDDB deterioration mechanism as above and established a reliable process (Japanese Patent Application No. 226876/1999 or Japanese Patent Application No. 104015/2000). However, because the material of an interlevel insulation film have recently used a low-dielectric-constant insulation material from a viewpoint of reducing the interlevel capacitance or the like, the interlevel dielectric strength is naturally lower than the case of using a silicon oxide film. Moreover, the effective electric field intensity increases due to the scaling down of line pitch (e.g. scaling down of from 1.0-μm pitch to 0.5-μm or 0.44-μm pitch). These result in further difficulty in securing reliability, i.e. TDDB life. Namely, in addition to the foregoing ammonia plasma surface treatment, it is essentially required to develop a process for further improve the TDDB life.
[0206] Accordingly, the present embodiment, after CMP and post-CMP washing, carries out hydrogen plasma or ammonia plasma processes and thereafter forms a cap film in a continuous fashion without release to the air. This can improve the TDDB life furthermore and secure sufficient reliability even in the combination of a low-dielectric-constant insulation film and a copper-based buried interconnect. Also, by combining with the foregoing post-CMP cleaning technique of the present embodiment, it is possible to further improve interconnect electric characteristics and TDDB life.
[0207] Next, explanation is made on the case that the technical idea of the invention is applied to a fabrication method for a CMIS (Complimentary MIS)- LSI (Large Scale Integrated Circuit), with reference to a FIG. 22 flowchart and FIGS. 23 to 35. Incidentally, FIG. 23 and FIG. 24A are essential-part plan views, wherein B in each figure is a sectional view on line X-X in A of the relevant figure.
[0208] In FIG. 23, a semiconductor substrate 1 forming a wafer (hereinafter referred merely to as substrate) comprises p-type single-crystal silicon having a specific resistance of nearly 1-10 Ωcm, for example. The semiconductor substrate 1 has a device isolation trench 2 formed in a main surface (device-forming surface) thereof. In the device isolation trench 2 is formed a trench-formed device isolation region (trench isolation) buried, for example, with a silicon oxide film. Also, a p-type well 4 and an n-type well 5 are formed in the semiconductor substrate 1 at a main-surface side thereof. Boron, for example, is introduced in the p-type well 4 and phosphorus, for example, is introduced in the n-type well 5. In the region forming the p-type well 4 and n-type well 5 encompassed by the device isolation region, an nMISQn and a pMISQp are formed.
[0209] The nMISQn and the pMISQn each have a gate dielectric film 6 formed by a silicon oxide film having a thickness, e.g. of nearly 6 nm. The film thickness of the gate dielectric film 6, referred herein, signifies a silicon-oxide-reduced film thickness (hereinafter referred merely to as reduced film thickness), which in some cases might not coincide with the actual film thickness.
[0210] The gate dielectric film 6 may be structured by a silicon oxide nitride film in place of the silicon oxide film. The silicon oxide nitride film, having a high effect to suppress the occurrence of interfacial levels in the film as compared to a silicon oxide film and reduce electron traps, can improve the resistance to hot carriers and reduce dielectric strength. To form a silicon oxide nitride film, the semiconductor substrate 1 may be thermally processed in a nitrogen-contained gas atmosphere such as NO, NO2 or NH3. Meanwhile, the similar effect to the above can be obtained by forming a silicon-oxide gate electrode film 6 on a surface of the p-type well 4 and n-type well 5 and thereafter the semiconductor substrate 1 is thermally processed in the foregoing nitrogen-contained gas atmosphere thereby segregating nitrogen to the interface of the gate dielectric film 6 and the semiconductor substrate 1.
[0211] Meanwhile, the gate dielectric film 6 may be formed by a composite insulation film, for example, of a silicon nitride film or silicon oxide film and a silicon nitride film. If the silicon-oxide gate dielectric film 6 is reduced in silicon-dioxide-reduced film thickness down to 5 nm or less, particularly 3 nm or less, the lowering of dielectric strength is directly actualized due to the occurrence of tunnel current or stress-caused hot carriers. Because the silicon nitride film has a dielectric constant higher than that of the silicon oxide film, the silicon-dioxide-reduced film thickness thereof is smaller than the actual film thickness. Namely, where having a silicon nitride film even if physically thin, it is possible to obtain a capacitance equivalent to that of a relatively thin silicon oxide film. Consequently, by structuring the date dielectric film 6 by a single silicon nitride film or a composite film of silicon oxide together therewith, its effective film thickness can be increased greater than a gate dielectric film structured by a silicon oxide film thereby improving the occurrence of tunnel leak current or the lowering in dielectric strength due to hot carriers. Also, the silicon oxide nitride film is immune to penetrate impurity as compared to the silicon oxide film. Accordingly, by structuring a gate oxide film 6 by a silicon oxide nitride film, it is possible to suppress from varying the threshold voltage resulting from the diffusion of the impurity in a gate electrode material into the semiconductor substrate.
[0212] Herein, the reduced film thickness dr of a single insulation film or composite insulation film is a film thickness defined by the following formula provided that the specific dielectric constant of a insulation film being considered is εi, the film thickness thereof is di and the specific dielectric constant of silicon dioxide is εS.
1
[0213] For example, the dielectric constant of silicon oxide (SiO2) is 4-4.2 while the dielectric constant of silicon nitride (Si3N4)is 8. Accordingly, computing with the dielectric constant of silicon nitride taken twice the dielectric constant of silicon oxide, a silicon nitride film having a film thickness of 6 nm, for example, has a silicon-dioxide-reduced film thickness of 3 nm. Namely, a gate dielectric film formed by a silicon nitride film having a film thickness of 6 nm is equivalent in capacitance to a gate dielectric film formed by a silicon oxide film having a film thickness of 3 nm. Also, a gate dielectric film formed by a composite film of a silicon oxide film having a film thickness of 2 nm and a silicon nitride film having a film thickness of 2 nm (reduced film thickness=1 nm) is equal in capacitance to a gate dielectric film formed by a single silicon oxide film having a film thickness of 3 nm.
[0214] The gate electrode 7 of the nMISQn and pMISQp is formed by a lamination film having, for example, a low-resistance polycrystal silicon film, a WN (tungsten nitride) film and a W (tungsten) film. It is noted that the gate electrode 7 may be formed using a lamination film having a tungsten silicide film or cobalt (Co) silicide film deposited on a low-resistance polycrystal silicon film. Also, the material of the gate electrode 7 may use an alloy of polycrystal or single-crystal silicon (Si) and germanium (Ge). On the gate electrode 7 is formed a gate cap film 8 formed of silicon oxide, for example. Also, sidewalls 9 are formed, for example, of silicon oxide on the side surfaces of the gate electrode 7.
[0215] The nMISQn has n− semiconductor region 10a and n+ semiconductor region 10b as source/drain semiconductor regions for the nMISQn both introduced, for example, with phosphorus or arsenic. The pMISQp has p− semiconductor region 11a and p+ semiconductor region 11b as source/drain semiconductor regions for the pMISQp both introduced, for example, with boron. Meanwhile, a silicide layer 12, for example, of titanium silicide, cobalt silicide or the like is formed in a surface of the n+ semiconductor region 10b and p+ semiconductor region 11b.
[0216] An insulation film 13 is deposited on the substrate 1. This insulation film 13 is formed by a film high in reflow property to fill a narrow space around the gate electrode, e.g. a BPSG (Boron-doped Phospho Silicate Glass) film. This may be structured by an SOG (Spin On Glass) film to be formed due to a spin application technique. The insulation film 13 is formed with contact holes 14a-14c. In a bottom of the contact hole 14a, 14b, part of a top surface of the silicide layer 12 is exposed. Also, in the bottom of the contact hole 14c, part of a top surface of the gate electrode 7 is exposed. Within the contact hole 14a-14c, a plug 15 is formed. The plug 15 is formed, for example, by depositing a titanium nitride (TiN) film and tungsten (W) film over the insulation film 13 including the inside of the contact hole 14a-14c by a CVD technique or the like and removing an unwanted portion of the titanium nitride film and tungsten film on the insulation film 13 by a CMP technique or etch-back technique to leave these films only inside the contact hole 14a-14c.
[0217] A first-level interconnect M1 is formed on the insulation film 13. The first-level interconnect M1 is formed, for example, of tungsten and electrically connected to the source/drain or gate electrode 7 of the nMISQn or pMISQp through the plug 15. Also, an insulation film 16a and an insulation film 16b are deposited on the insulation film 13 in the order of from the lower level in a manner covering the first-level interconnect Ml. The insulation film 16a is formed by an insulation film having a low dielectric constant, e.g. organic polymer, while the insulation film 16b is formed, for example, silicon oxide serving to secure a mechanical strength of the interlayer insulation film.
[0218] The organic polymer structuring an insulation film 16a includes, for example, SiLK as a material based on polyallylether (PAE) (by U.S. The Dow Chemical Co.; specific dielectric constant=2.7, heat-resisting temperature=490° C. or higher, dielectric strength=4.0-5.0 MV/Vm) or FLARE (by U.S. Honeywell Electronic Materials; specific dielectric constant=2.8, heat-resisting temperature=40° C. or higher) and so on. The PAE-bsed material has a feature of high basic properties, excellent mechanical strength, thermal stability and low cost.
[0219] Also, the material of an insulation film 16a can use a SiOC-based material, a SiOF-based material, an HSQ (hydrogen silsequioxane)-based material, an MSQ (methyl silsesquioxane)-based material, a porous HSQ-based material, a porous MSQ-based material or a porous organic material, in place of the PAE-based material.
[0220] The SiOC-based material includes, for example, BlackDiamond (by U.S. Applied Materials; specific dielectric constant=3.0-2.4, heat-resisting temperature=450° C.), CORAL (by U.S. Novellus Systems; specific dielectric constant=2.7-2.4, heat-resisting temperature=500° C.), Aurora2.7 (by ASM Japan; specific dielectric constant=2.7, heat resisting temperature=450° C.) and p-MTES (by Hitachi Kaihatsu, specific dielectric constant=3.2).
[0221] The HSQ-based material includes, for example, OCD T-12 (by Tokyo Ohka Kogyo; specific dielectric constant=3.4 -2.9, heat-resisting temperature 450° C.), FOx (by U.S. Dow Corning; specific dielectric constant=2.9) and OCL T-32 (by Tokyo Ohka Kogyo; specific dielectric constant=2.5, heat-resisting temperature=450° C.).
[0222] The MSQ-based material includes, for example, HSG-R7 (by Hitachi Chemical, specific dielectric constant=2.8, heat-resisting temperature=650° C), OCD T-9 (by Tokyo Ohka Kogyo; specific dielectric constant=2.7, heat-resisting temperature=600° C.), LKD-T200 (by JSR; specific dielectric constant=2.7-2.5, heat-resisting temperature=450° C.), HOSP (by U.S. Honeywell Electronic Materials; specific dielectric constant=2.5, heat-resisting temperature=550° C.), HSG-RZ25 (by Hitachi Chemical, specific dielectric constant=2.5, heat-resisting temperature=650° C.), OCL T-31 (by Tokyo Ohka Kogyo; specific dielectric constant=2.3, heat-resisting temperature=500° C.) and LKD-T400 (by JSR; specific dielectric constant=2.2-2, heat-resisting temperature=450° C.).
[0223] The porous HSQ-based material includes, for example, HLK (by U.S. Dow Corning; specific dielectric constant=2.5-2), OCL T-72 (by Tokyo Ohka Kogyo; specific dielectric constant=2.2-1.9, heat-resisting temperature=450° C.), Nanoglass (by U.S. Honeywell Electronic Materials; specific dielectric constant=2.2-1.8, heat-resisting temperature=500° C. or higher) and MesoELK (by U.S. Air Productsand Chemicals; specific dielectric constant=2 or lower).
[0224] The porous MSQ-based material includes, for example, HSG-6211X (by Hitachi Chemical; specific dielectric constant=2.4, heat-resisting temperature=650° C.), ALCAP-S (by Asahi Kasei Kogyo; specific dielectric constant=2.3-1.8, heat-resisting temperature=450° C.), OCL T-77 (by Tokyo Ohka Kogyo; specific dielectric constant=2.2-1.9, heat-resisting temperature=600° C.), HSG-6210X (by Hitachi Chemical; specific dielectric constant=2.1, heat-resisting temperature=650° C.), and silica aerogel (Kobe Steel; specific dielectric constant 1.4-1.1).
[0225] The porous organic material includes, for example, PolyELK (by U.S. Air Productsand Chemicals; specific dielectric constant=2 or lower, heat resisting temperature=490° C.).
[0226] In the insulation film 16a, 16b, a through-hole 17 is opened to expose part of the first-level interconnect M1. A plug 18, for example, of tungsten is formed in the through-hole 17.
[0227] The SiOC-based material and SiOF-based material for forming an insulation film 16a and the insulation film 16b are formed by the CVD technique. In the case of Black Diamond in the above, for example, a mixture gas of trimethylsilane and oxygen is used as a material gas. For P-MTES in the above, the material gas uses, for example, a mixture gas of methyltriethoxysilane and N2O. The other insulation materials having low dielectric constant are formed by an application technique.
[0228] First, in the present embodiment, an insulation film 19a of a silicon nitride film having a film thickness, for example, of 50 nm is deposited on the foregoing semiconductor substrate 1 by the plasma CVD technique or the like, as shown in FIGS. 24A and 24B. The insulation film 19a may use a silicon carbide (SiC) or silicon oxide film formed by the plasma CVD technique, in place of the silicon nitride film. The silicon carbide-based material to be formed by the plasma CVD technique includes, for example, BLOk (by AMAT; specific dielectric constant=4.3), wherein a mixture gas of trimethylsilane and helium, for example, is used upon formation thereof. Also, the silicon oxide-based material to be formed by the plasma CVD technique includes, for example, PE-TMS (by Canon; specific dielectric constant=3.9), wherein a mixture gas of trimethoxysilane and nitride oxide (N2O), for example, is used upon formation thereof. These if used can reduce the dielectric constant greater than that of the silicon nitride film and hence decrease the interconnect capacitance or the like, making possible to improve the operation speed of the semiconductor integrated circuit device.
[0229] Subsequently, insulation films 16c, 16d are deposited in the order from the lower level on the insulation film 19a (step 100 in FIG. 22). The insulation film 16c comprises a material selected from the low dielectric constant of insulation films as the same in the insulation films 16a. Also, the insulation film 16d is formed of the same material as that of the insulation film 16b. Thereafter, the insulation films 16d, 16c, 19a are preferentially removed by the dry etching using a photoresist film as a mask thereby forming an interconnect trench (interconnect opening) 20a (step 101 in FIG. 22). To form an interconnect trench 20a, the insulation film 16d, 16c is preferentially etched using the insulation film 19a as an etch stopper, and thereafter the insulation film 19a is etched. In this manner, a thin insulation film 19a is previously formed in a layer underlying the insulation film 16c, 16d where an interconnect trench 20a is to be formed. Etching is once stopped at the surface of the insulation film 19a and thereafter the insulation film 19a is etched thereby improving the depth precision of the interconnect trench 20a. Thus, the interconnect trench 20a can be formed without encountering overetching.
[0230] Next, a buried interconnect for a second-level interconnect is formed in the interconnect trench 20a.
[0231] First, as shown in FIG. 25 a thin conductive barrier film 21a, for example, of titanium nitride (TiN) is formed over the insulation film 16c, 16d including the inside of the interconnect trench 20a by the sputtering technique or the like (step 102 in FIG. 22). The conductive barrier film 21a has the actions to prevent the diffusion of copper for forming a main conductor film, hereinafter referred, to improve the adhesion of the main conductor film to the insulation film 16c, 16d and to improve the wetness of copper upon reflowing the main conductor film. The film having such actions preferably uses, in place of titanium nitride, a refractory-metal nitride such as tungsten nitride (WN) and tantalum nitride (TaN), less reactive with copper. Meanwhile, it is possible to use, in place of the titanium nitride, a material of a refractory-metal nitride added with silicon (Si) or a refractory metal, such as tantalum (Ta), titanium (Ti), tungsten (W), titanium tungsten (TiW) or the like, less reactive with copper.
[0232] The present embodiment exemplifies a conductive barrier film 21a in the case having the maximum thickness of 50 nm. However, according to the result of consideration by the present inventors, it has been revealed that the conductive barrier film 21a can be further reduced in thickness or eliminated. This will be hereinafter referred.
[0233] Subsequently, a main conductor film 22a, for example, of copper is deposited onto the conductive barrier film 21a (step 103 in FIG. 22). In the present embodiment, the main conductor film 22a has been formed by a plating technique. The use of a plating technique makes it possible to form a main conductor film 22a favorable in film quality with preferred burying and at low cost. In this case, a thin conductor film of copper was first deposited on the conductive barrier film 21a by the sputtering technique, and thereafter a comparatively thick conductor film of copper was grown thereon by an electrolytic plating technique or electroless plating technique using a plating solution based, for example, on copper sulfate, thereby depositing a main conductor film 22a.
[0234] It is noted that the main conductor film 22a can be formed by a sputtering technique. The sputtering technique for forming such a conductive barrier film 21a and main conductor film 22a may be the usual sputtering technique but preferably uses a sputtering technique having a high directivity, e.g. a long-throw sputtering technique or a collimate sputtering technique, in view of improving burying property and film quality. Also, the main conductor film 22a can be formed by a CVD technique.
[0235] Subsequently, the substrate 1 is thermally processed in a nonoxidizing atmosphere (e.g. hydrogen atmosphere) at nearly 475° C., for example, thereby reflowing the main conductor film 22a to bury copper in the interconnect trench 20a without giving gaps.
[0236] Next, the main conductor film 22a and conductive barrier film 21a is polished by a CMP technique (step 104 in FIG. 22), and thereafter a buried interconnect 23a based on copper is formed in the interconnect trench 20a through a cleaning process with pure water or the like, as shown in FIG. 26. The detail of the CMP process and CMP apparatus will be hereinafter referred.
[0237] Subsequently, a post-CMP cleaning process is entered in a state the surface of the substrate 1 is kept in a wet state. First, an alkali cleaning process is carried out for the substrate 1 (step 105 in FIG. 22). Herein, the aim is to remove the foreign matter, such as slurry, used in the CMP process, wherein the acid slurry adhered to the substrate during CMP is neutralized to arrange the zeta potential of the substrate 1, foreign matter and cleaning brush. In order to eliminate an absorbing force acted between them, scrub cleaning (or brush cleaning) is conducted to the surface of the substrate 1 while supplying thereto a weak alkali agent solution, for example, of nearly PH 8 or higher. As an alkali agent solution was used Diluted Amino Ethanol (DAE having a composition of 2-aminoethanol and CH2CH2OH in an concentration of approximately 0.001%-0.1%, preferably 0.1%). This agent solution is less in etch action to copper and has a cleaning capability equivalent to NH4OH. The detail of the alikali cleaning process will be hereinafter explained.
[0238] Subsequently, a deoxidizing process is carried out to the substrate 1 (step 106 in FIG. 22). Herein, as shown in FIG. 27 a thermal process (hydrogen (H2) anneal) was carried out to the substrate 1 in a hydrogen atmosphere at, for example, 200° C.-475° C., preferably 300° C., for 0.5-5 minutes, preferably approximately 2 minutes. This makes it possible to deoxidize the copper oxide caused in the surface of the buried interconnect 23a during CMP and hence prevent the buried interconnect 23a from being etched due to the following acid cleaning. Due to this, it is possible to suppress and prevent simultaneously the variation in interconnect resistance and the occurrence of steps and further the occurrence of etch corrosion. Also, where no deoxidizing process is carried out, there is a case that the organic matter of BTA or the like adhered to the surface of the substrate 1 during the CMP process acts as a mask during a cleaning process making impossible to favorably removing a surface layer of the insulation film 16d. However, by conducting a deoxidizing process as in the present embodiment, because the organic matter of BTA or the like adhered during CMP can be removed, the surface layer of the insulation film 16d can be removed sufficiently and evenly. These makes it possible to greatly improve the TDDB life for the semiconductor integrated circuit device.
[0239] Subsequently, an acid cleaning process is carried out to the substrate 1 shown in FIG. 28 (step 107 in FIG. 22). Herein, the aim is to improve the TDDB characteristic, remove the remaining metal, decrease the dangling bonds on the surface of the insulation film 16d, remove the roughening on the surface of the insulation film 16d and so on. A hydrogen fluoride solution is fed to the surface of the substrate 1 to remove foreign particles due to etching. The mere insertion of hydrogen-fluoride cleaning can improve the TDDB characteristic. This is to be considered due to the removal of the surface damaged layer by an acid process thereby improving the adhesion at the interface. The hydrogen fluoride (HF) cleaning uses, for example, brush scrub cleaning wherein selected are the conditions of an HF concentration of 0.5% and a cleaning time of 20 seconds.
[0240]
FIG. 29 shows a result of the TDDB characteristic according to the present embodiment. As compared to the TDDB characteristic in the successive sequence of alkali cleaning and acid cleaning in the figure, it is to be understood that the TDDB characteristic in the sequence of alkali cleaning, hydrogen anneal and acid cleaning is improved by two orders of magnitude. If considering the reliability of the buried copper interconnect structure using a low-dielectric-constant insulation material in the interlayer insulation film, the improvement of TDDB life by the two orders of magnitude provides a quite effective process. The reason of improvement in TDDB life by insertion of ydrogen anneal between alkali cleaning and acid cleaning is considered because of the removal of the organic matter such as BTA adhered during CMP. It is to be presumed that, if acid cleaning is done while being adhered with organic matter, the cleaning on the adjacent insulation film surface (lift-off) determinative for the TDDB life cannot be made sufficiently. On the other hand, the present embodiment, because of carrying out a cleaning process after performing a hydrogen anneal process, can lift off the insulation-film surface sufficiently and evenly thereby making possible to improve the TDDB life.
[0241] Meanwhile, as in the foregoing, the occurrence of interconnect resistance proceeds in the order of the formation and promotion of an oxide film due to CMP, the removal of a copper oxide film by an acid solution such as a hydrogen-fluoride cleaning, the increase (variation) in interconnect resistance and the occurrence of steps. Accordingly, by performing a hydrogen anneal process at a time that alkali cleaning is ended, the copper oxide film caused upon CMP can be deoxidized into copper thus making possible to suppress or prevent the copper interconnect from being etched in the following acid cleaning. This can suppress or prevent simultaneously the increase of interconnect resistance and the occurrence of variation and steps, and further the occurrence of etch corrosion. FIG. 30 is a graph showing a dependency of interconnect resistance upon hydrogen anneal in the case the present embodiment is applied. The interconnect resistance could have been reduced by approximately 6% by carrying out alkali cleaning, hydrogen anneal and acid cleaning in the order as compared to that by continuously performing alkali cleaning and acid cleaning. Also, the variation of resistance could have been decreased from 6.4% to 5.9%.
[0242] Although the above example explained on the case of carrying out hydrogen anneal as a deoxidizing process, this is not limited to but, for example, hydrogen plasma or ammonia plasma may be carried out. In this case, in addition to the foregoing effects, the deoxidizing process time can be shortened thereby providing an effect of promoting the improvement of throughput. The advantage of hydrogen anneal as compared to hydrogen plasma or ammonia plasma lies in an actual experience of broad use in the device process and capability of performing a process comparatively easily because of absence of requirement to form a vacuum state and the like.
[0243] Also, prior to or in concurrent with the post-CMP cleaning process, the substrate 1 at its surface may be subjected to pure-water scrub cleaning, superpure-water ultrasonic cleaning, pure-water feed water cleaning or pure-water spin cleaning, or the substrate 1 at its backside be subjected to pure-water scrub cleaning.
[0244] Next, after the cleaning process, a cap-insulation-film forming process (step 111 in FIG. 22) is entered by way of, for example, a pure-water rinse process, a drying process such as spin drying and IPA (IsoPropyl Alcohol) vapor drying (step 108 in FIG. 22) and post process (steps 109, 110 in FIG. 22).
[0245] Herein, as shown in FIG. 31 a hydrogen plasma process is first carried out on the surface of the substrate 1 (the surface the buried interconnect 23a is exposed). The condition for the hydrogen plasma process, provided that the diameter of the substrate 1 is 8 inches (=approximately 20 cm) for example, can be given a processing pressure of 5.0 Torr (=6.6661×102 Pa), an RF power of 600 W, a substrate temperature of 400° C., a hydrogen gas flow rate of 500 cm3/min and a processing time of 10-30 seconds. The electrode-to-electrode distance was given 600 mils (15.24 mm).
[0246] Subsequently, after the hydrogen plasma process 109, an ammonia (NH3) plasma process is continuously carried out, without release to the air, to the surface of the substrate 1 (the surface the buried interconnect 23a is exposed), as shown in FIG. 32. The condition for the ammonia plasma process is the same as the hydrogen plasma process condition in process 109, excepting that the ammonia flow rate has been given approximately 200 cm3/min for example.
[0247] It is noted that the plasma process condition is, of course, not limited to the exemplified conditions. According to the considerations by the present inventors, plasma damage can be decreased with increase in the pressure while the TDDB life can be decreased in in-substrate variation and increased in life with increase in the substrate temperature. Also, there have obtained the finding that hillocks are caused more readily on the surface of Cu as RF power and process time increase. Considering these findings and conditional variation due to device structure, setting is possible within the ranges, e.g. a processing pressure of 0.5-6 Torr (=0.66661×102−7.99932×102 Pa), an RF power of 300-600 W, a substrate temperature of 350° C.-450° C., a hydrogen gas flow rate of 50-1000 cm3/min, an ammonia gas flow rate of 20-500 cm3/min, an electrode-to-electrode distance of 150 -1000 mils (3.81-25.4 mm).
[0248] Subsequently, after the ammonia plasma process 110, an insulation film 19b (cap film) is continuously deposited, without release to the air, on a surface of the buried interconnect 23a and insulation film 16d by the CVD technique or the like, as shown in FIG. 33. The insulation film 19b is, for example, of the same material and in the same thickness as the insulation film 19a, and has a similar modification to the insulation film 19a.
[0249] In this manner, the present embodiment carries out a hydrogen plasma process and ammonia plasma process to the substrate 1 prior to the deposition of an insulation film 19b for a cap film.
[0250] In ammonia plasma, the copper oxide (CuO, CuO2) in a copper interconnect surface oxidized in CMP is deoxidized into copper (Cu). Also, a copper nitride (CuN) layer, for preventing copper silicidation upon set flow, is formed in a surface (in an extremely shallow region) of the buried interconnect 23a. In an upper surface of the insulation film 16d between the interconnects (in an extremely shallow region), the formation of SiN or SiH proceeds to compensate for the dangling bonds in the surface of the insulation film 16d, which can also improve the adhesion of between the cap film (silicon nitride film) and the buried interconnect 23a and insulation film 16d to reduce the leak current through the interface. This effect improves the TDDB life.
[0251] On the other hand, hydrogen plasma is extremely high in capability to remove organic-based matters as compared to the ammonia plasma process or the like, as has been described in Japanese Patent Application No. 226876/1999. Accordingly, this can completely remove the BTA contained in the CMP slurry, slurry ingredients, post-CMP-cleaning organic acid and the residual organic matter produced in the process, thereby decreasing the leak current through the interface. As a result, the TDDB life can be improved furthermore.
[0252] Consequently, by carrying out a hydrogen plasma process and an ammonia plasma process in the order, it is possible to obtain the deoxidization in the surface of the copper-based buried interconnect 23a and the formation of a silicide-resisting barrier layer, the cleaning and SiH effect/SiN effect at the insulation-film interface thus realizing further improvement. FIG. 34 shows a TDDB characteristic in actual combination of a hydrogen plasma process and an ammonia plasma process. CMP conditions and post-CMP cleaning conditions are all the same. It has been revealed that, where the interlayer insulation film is structured by depositing a silicon nitride film formed by a plasma CVD technique on a silicon oxide film formed by a plasma CVD technique using, for example, a TEOS (Tetraethoxysilane) gas, a sample processed by a combination of hydrogen plasma and ammonia plasma is improved in TDDB life by the two orders of magnitude as compared to the case of a single amonia plasma process.
[0253] Meanwhile, FIG. 34 shows an approximated evaluation characteristic where the interlayer insulation film is structured of such a low-dielectric-coefficient material (e.g. SILK in the foregoing) as was explained in the present embodiment, from the fact of a SiLK dielectric strength of nearly 4.0-5.0 MV/cm and the TDDB characteristic evaluation experiences on an organic SOG (Spin On Glass) interlayer structure and the like. With only an ammonia plasma process, there is a case that is insufficient in an operating environment, for example, of approximately 0.13-0.17 MV/cm for 10 years. On the contrary, the present embodiment using hydrogen plasma and ammonia plasma can secure sufficient reliability for that operating environment.
[0254]
FIG. 35 shows one example of a CMIS-LSI formed up to a seventh-level interconnect in the manner as the foregoing. Although not limited, the dimensions of each part will be described in the below.
[0255] The film thickness and the interconnect pitch (center-to-center distance of between adjacent interconnects) in the first-level interconnect M1 are, for example, nearly 0.4 μm or nearly 0.25 μm. Also, the second-level interconnect M2 to the fifth-level interconnect M5 are made by the foregoing Cu-interconnect-forming method. In the second-level interconnect M2 and third-level interconnect M3, the conductive barrier film has a thickness, for example, of nearly 0.05 μm, and the main conductor films has a thickness, for example, of nearly 0.35 μm, the interconnect width and the interconnect pitch are, for example, nearly 0.5 μm or nearly 0.25 μm. In the fourth-level interconnect M4 and fifth-level interconnect M5, the conductive barrier film has a thickness, for example, of nearly 0.05 μm, and the main conductor film has a thickness, for example, of nearly 0.95 μm, the interconnect width and the interconnect pitch are, for example, nearly 1.0 μm or nearly 0.25 μm. Also, the sixth-level interconnect M6 is in a three-layer structure having, for example, a tungsten film, an aluminum film and a tungsten film. Also, the seventh-level interconnect M7 is, for example, of an aluminum film. The seventh-level interconnect M7 has a pad to be formed with a bump electrode or connected with a bonding wire, although not shown. Incidentally, the one reason of structuring the seventh-level interconnect (M7) by the lamination film of aluminum and tungsten is that the lamination film is generally used in an uppermost layer of the usual semiconductor integrated circuit device not adopting a damascene interconnect structure wherein the capability of securing the reliability in connection to a bump electrode or bonding wire has been empirically verified. The through-hole connecting between the first-level interconnect M1 and the second-level interconnect M2 has a diameter, for example, of nearly 0.45 μm or nearly 0.25 μm. The through-hole connecting between the second-level interconnect M2 and the third-level interconnect M3 has a diameter, for example, of nearly 0.5 μm or nearly 0.25 μm. The through-hole connecting between the third-level interconnect M3 and the fourth-level interconnect M4 has a diameter, for example, of nearly 0.5 μm or nearly 0.25 μm. The through-hole connecting between the fourth-level interconnect M4 and the fifth-level interconnect M5 has a diameter, for example, of nearly 1.0 μm or nearly 0.25 μm. The through-hole connecting between the fifth-level interconnect M5 and the sixth-level interconnect M6 has a diameter, for example, of nearly 0.5 μm or nearly 0.25 μm.
[0256] Next, explanation is made on one example of the CMP apparatus used in the present embodiment.
[0257]
FIG. 36 shows one example of an overall structure of the CMP apparatus used in the present embodiment.
[0258] The CMP apparatus is a sheet-processing type CMP apparatus used for polishing a conductive barrier film 21a and main conductor film 22a, and comprises a loader 25a for accommodating a plurality of substrates each formed with a main conductor film 22a on the surface, a polish processing section 25b for polishing and planarizing a main conductor film 22a, a polish processing section 25c for polishing and planarizing a conductive barrier film 21a, an ani-corrosion processing section 25d for carrying out an anti-corrosion process to a surface of a polished substrate 1, an immersion processing section 25e for keeping a substrate 1 completed of anti-corrosion process not to dry its surface before post-cleaning, a post-cleaning processing section 25f for post-cleaning a substrate 1 completed of anti-corrosion process and an unloader 25g for accommodating a plurality of substrates 1 completed of post-cleaning. The post-clean processing section 25f has a first post-clean processing section 25f1 for alkali cleaning, a deoxidization processing section 25f2 for deoxidization and a second post-clean processing section 25f3 for acid cleaning. The post-clean processing section 25f may be an independent device separated from the CMP apparatus.
[0259] As shown in FIG. 37, the polish processing sections 25b, 25c of the CMP apparatus has a housing 26a opened in an upper part thereof. A rotary spindle 26b attached on the housing 26a has, at a top end, a polishing platen 26d to be rotatively driven by a motor 26c. The polishing platen 26d has, on its surface, a polishing pad 26e formed by evenly bonding a synthetic resin having a multiplicity of pores.
[0260] Also, the polish processing section 25b, 25c has a wafer carrier 26f for holding a substrate 1. The driving spindle 26g having the wafer carrier 26f is rotatively driven in unison with the wafer carrier 26f by a motor (not shown) and vertically movable in the above of the polishing platen 26d.
[0261] A substrate 1 is held with its main or polish surface faced down on the wafer carrier by a vacuum suction mechanism (not shown) provided in the wafer carrier 26f. The wafer carrier 26f is formed, at its lower end, with a recess 26f1 to receive a substrate 1. When a substrate 1 is received in the recess 26f1, its polish surface is in a state flush with or slightly projecting from the lower end surface of the wafer carrier 26.
[0262] A slurry feed pipe 26h is provided in the above of the polishing platen 26d, to feed polishing slurry S to between a surface of the polishing pad 26e and a polish surface of the substrate 1, so that the polish surface of the substrate 1 is chemically, mechanically polished by the polishing slurry S fed through an lower end of the same. Polishing slurry S uses a dispersion or dissolution based on polishing grits, for example, of alumina and an oxidizer of hydrogen peroxide water or ferric nitrate solution.
[0263] Also, the polish processing section 25b, 25c has a dresser 26i as a tool for shaping (dressing) the surface of the polishing pad 26e. The dresser 26i is attached at the lower end of the driving spindle 26j vertically movable in the above of the polishing platen 26d, to be rotatively driven by a motor (not shown).
[0264] The polished substrate 1 at its surface is subjected to a anti-corrosion process in the anti-corrosion processing section 25d. The anti-corrosion processing section 25d is in a similar structure to the structure of the polish processing section 25b, 25c. Herein, the substrate 1 at its main surface is pressed onto the polishing pad provided on the surface of the polishing platen and mechanically removed of polishing slurry. Thereafter, an agent containing an anti-corrosive, for example, of benzotriazole (BTA) is fed to the main surface of the substrate 1 thereby forming a hydrophobic protection film in a surface of the copper interconnect formed on the main surface of the substrate 1.
[0265] The substrate 1 completed of the anti-corrosion process is temporarily stored in the immersion processing section 25 in order to prevent its surface from being dried. The immersion processing section 25e is to maintain the substrate 1 completed of the anti-corrosion process not to be dried at its surface until post-cleaning of it, and structured to store, for example, a predetermined number of substrates 1 by immersing them in an immersion tank (stocker) having pure water allowed to overflow. At this time, by supplying the pure water cooled down to such a temperature that electrochemical corroding reaction on the buried interconnect 23a does not substantially proceed, the buried interconnect 23a can be prevented more positively from corroding. The prevention of the substrate 1 against drying may be conducted by a method other than the storage in the immersion tank provided that the substrate 1 at its surface can be at least held in a wet state, e.g. pure-water shower supply.
[0266]
FIG. 38 shows a scrub cleaning device 27 as one example of the first and second post-clean processing sections 25f1, 25f3. This scrub cleaning device 27 has a stage 27b to be rotatively driven by a motor 27a so that a substrate 1 held on the stage 27b rotates at a desired speed in a horizontal plane. Also, cylindrical brushes 27c are pressed onto the upper and lower surfaces of the substrate 1 rotating on the stage 27b. These brushes 27c are rotated at a desired speed in a vertical plane relative to the main surface of the substrate 1 by a rotation drive mechanism. Also, the brushes 27c are supplied with cleaning liquid through piping. The cleaning liquid supplied to the rush 27c exudes, little by little, from an interior of the brush 27c structured by a porous member of a synthetic resin, for example, of polyvinylalcohol (PVA) to a surface (contact surface with the substrate 1), thereby wetting the surface of the substrate 1 contacted with the brush 27c.
[0267] In order to remove the slurry or the like adhered on the substrate 1, the brushes 27c are pressed onto the upper and lower surface of the substrate 1, and then the brushes 27c are rotated in a vertical plane relative to the main surface of the substrate 1 while rotating the substrate in the horizontal plane. At this time, the both brushes 27c are pressed onto the substrate at such a pressure that a plurality of projections, hereinafter referred, provided in the surface thereof at their tips are slightly recessed (e.g. nearly 1 mm). The rotation velocity of the substrate is, for example, nearly 20 rpm while the rotation velocity of the brush 27c is, for example, 120 rpm. In this manner, the opposite surfaces of the substrate 1 are scrub-cleaned by the brushes 27c for a predetermined period of time, thereby removing the slurry or the like adhered on the surface.
[0268]
FIG. 39 is a magnifying perspective view of the brush 27c. As shown in the figure, a multiplicity of precise cylindrical projections 27c1 are provided with spacing on a surface of the brush 27c, i.e. contact surface with a substrate 1. The brush 27c is structured by a porous member of a synthetic resin, for example, of PVA. In the present embodiment, the number of the projections 27c1 increases in a direction of from a central portion of the brush 27c toward the both ends. Namely, the projections 27c1 are arranged such that the number becomes the minimum in the central part of the brush 27c and the maximum at the both ends.
[0269] Where using such a brush 27c, the contact time with the brush 27c is long at a central part of a substrate contacted with the central part of the brush 27c as compared at the peripheral part of the substrate 1 contacted with the ends of the brush 27c. On the contrary, however, because the projections 27c contacting the substrate 1 is minimum in the number at the central part of the brush 27c and maximum in the number at the ends thereof, the contact area between the brush 27c and the substrate 1 is greater at the peripheral part than in the central part of the substrate 1. This can make equivalent the value of time x area that the brush 27c is contacted with the substrate 1 throughout the surfaces of the substrate. Due to this, the amount of cut (etch amount) in the surface of the buried interconnect 23a can be made equivalent throughout the surfaces of the substrate 1.
[0270] According to an experiment by the present inventors, the amount of cut in the surface of the buried interconnect 23a has been made nearly equivalent throughout the surface of a substrate 1, by optimizing the number of the projections 27c1 under the following condition. Incidentally, the diameter of the substrate 1 used herein is, for example, nearly 125 mm, the rotation velocity of the substrate is, for example, nearly 22 rpm, and the diameter of the brush 27c is, for example, nearly 55 mm. The brush 27c was pressed onto the substrate at a pressure to recess by 1 mm the contact surface with the substrate 1.
[0271] Where the substrate was equally divided into ten from the center to the outermost periphery wherein the number of the projections to contact each region has gradually increased along the center of the brush to the end, for a brush rotation velocity of 120 rpm the optimal number of the projections was 1, 1, 2, 3, 3, 4, 5, 6, 7 and 8 (in the number) in the order of from the substrate center. Also, for a brush rotation velocity of 30 rpm, the optimal number of projections was 3, 3, 8, 8, 8, 8, 8, 8, 8 and 8 (in the number) in the order of from the substrate center.
[0272] The other means for make equivalent the time x area that the brush 27c is contacted with the substrate 1 throughout the surface of the substrate 1 can adopt various means for changing the shape, the number and the like of the projections 27c1 such that the contact area between the brush 27c and the substrate 1 decreases as the center of the substrate 1 is neared and increases as the peripheral part of the substrate is neared, e.g. the diameter of the projection 27c1 is gradually increased in the direction of from the center of the brush 27c toward the both ends (or the diameter of the projection 27c1 is gradually decreased in the direction of from the both ends of the brush 27c toward the center).
[0273] Also, in place of the above means for decreasing the contact area between the brush 27c and the substrate 1 in the direction of from the peripheral part of the substrate toward the center, the pressure of the brush 27c pressed to the surface of the substrate 1 may be decreased in the direction of from the peripheral part of the substrate 1 toward the center (or increased in the direction of from the center toward the peripheral part). In this case, even if the contact area between the brush 27c and the substrate 1 is nearly the same in every part of the substrate 1, the effect similar to the foregoing can be obtained.
[0274] In order to decrease the pressure of the brush 27c pressed to the surface of the substrate 1 as the center is neared from the peripheral part of the substrate 1, for example the projections 27c1 may be decreased in height in the direction of from the both ends of the brush 27c toward the center or the brush 27c may be decreased in diameter.
[0275] Meanwhile, by optimizing the ratio of substrate rotation velocity/brush rotation velocity (hereinafter referred to as W/B ratio), the in-plane evenness on the substrate 1 can be realized without changing the number of projections 27c1 between the center and the both ends of the brush 27c. According to a result of experiment by the present inventors, the effective cleaning condition in post-CMP cleaning was at a W/B ratio=1.2 or greater for the roll-type cleaning apparatus and at a W/B ratio=2.0 or greater for the disk-type cleaning apparatus.
[0276] Incidentally, the art for making equivalent the amount of cut on the surface of the substrate 1 upon post-CMP cleaning as in the above is described in Japanese Patent Application No. 176769/2000 by the present inventors.
[0277] Although in the post-cleaning process the roll-type cleaning scheme was explained, this is not limited to but various modification is possible, e.g. a disk-type cleaning scheme can be adopted in alkali cleaning. Also, in acid cleaning, a disk-type cleaning scheme or pen-type cleaning scheme can be adopted. FIG. 40 shows one example of a disk-type cleaning scheme. FIG. 40A is a plan view and FIG. 40B is a sectional view thereof. In the state that the disk brushes 28 circular in plane are placed on the upper and lower surfaces of a substrate 1, the substrate 1 at its surface is cleaned while rotating the substrate 1 and disk brushes 28 in a horizontal plane. Meanwhile, FIG. 41 shows one example of a pen-type cleaning scheme. FIG. 41A shows a plan view and FIG. 41B a sectional view thereof. In the state that a pen brush 29 is placed on a surface (surface formed with the buried interconnect) of a substrate 1, the pen brush 29 is rotated and swung to clean a main surface and end surface of the substrate 1 while rotating the substrate 1 in a horizontal plane. The pen brush 29 is suited for removing the substance having not been removed in the precedent-staged cleaning process (e.g. roll-type cleaning or disk-type cleaning). The brush material, etc. of the disk brush 28 and pen brush 29 are the same as the foregoing.
[0278] The substrate 1 completed of the post-cleaning process, after pure-water rinse and spin dry, is accommodated in a dry state to the unloader 25g (see FIG. 36) and transferred collectively on a plurality-unit basis to the next process.
[0279] Incidentally, it is possible to make the immersion processing section (substrate storing section) 25e for preventing against drying in the surface of the substrate 1 completed of the anti-corrosion process in a shade structure so that illumination light or the like cannot be illuminated to the surface of the substrate 1 in storage. This can prevent short current from occurring due to the photovoltaic effect. To make the immersion processing section 25e in a shade structure, concretely the immersion tank (stocker) at its surrounding is covered by a shading sheet or the like thereby reducing the interior illuminance at least to 500 lux or below, preferably 300 lux or below, more preferably 100 lux or below.
[0280] Meanwhile, immediately after the polishing process, i.e. before starting of an electrochemical corrosive reaction due to an oxidizer in the polishing slurry remained on the surface, transfer may be immediately made to the dry processing section to remove the water content from the polishing slurry by forcible drying. In this case, the substrate 1 subjected to a polishing process in the polish processing section 25b, immediately after a polishing process, i.e. before starting of an electrochemical corrosive reaction due to an oxidizer in the polishing slurry remained on the surface, is immediately transferred to the dry processing section to be forcibly removed of the water content in the polishing slurry by forcible drying. Thereafter, the substrate 1 while being maintained in the dry state is transferred to the post-clean processing section 25f where it is subjected to a post-cleaning process and thereafter accommodated to the unloader 25g through pure-water rinse and spin dry. In this case, because the substrate 1 at its surface is kept in the dry state in the duration of immediately after the polishing process to a start of post cleaning, an electrochemical corrosive reaction is suppressed from starting. This can effectively prevent the Cu interconnect from corroding.
[0281] Next, explanation is made on one example of the plasma processing apparatus used in the processes of from post-processing to cap-film formation FIG. 42A is a sectional view showing one example of a plasma processing apparatus, and FIG. 42B is a plan view thereof. It is noted that this plasma processing apparatus is described in Japanese Patent Application No. 226876/1999 by the present inventors.
[0282] This plasma processing apparatus 30 uses, for example, P5000 by AMAT. The plasma processing apparatus 30 has a load lock chamber 30a mounted with two process chambers 30b1, 30b2 and a cassette interface 30c. The load lock chamber 30a has therein a robot 30d to transfer a substrate 1. A gate valve 30e is provided between the load lock chamber 30a and the process chamber 30b1, 30b2 to keep a high vacuum state within the load lock chamber 30a even during processing.
[0283] The process chamber 30b1, 30b2 has therein a susceptor 30f to hold a substrate 1, a baffle plate 30g to regulate gas flow, a support member 30h to support the susceptor 30f, an electrode 30i in a mesh form arranged opposed to the susceptor 30f, and an insulation plate 30j arranged nearly opposed to the baffle plate 30g. The insulation plate 30j serves to suppress parasitic discharge in an unnecessary region other than the region between the suceptor 30f and the electrode 30i. On a back side of the suceptor 30f, lamps 30m are arranged in a reflection unit 30k so that the infrared ray 30n emitted from the lamp 30m is passed through a quartz window 3p and illuminate to the susceptor 30f and substrate 1. This heats up the substrate 1. It is noted that the substrate 1 is placed face up on the susceptor 30f (in a state the main surface, or buried interconnect formed surface, positioned up).
[0284] The process chamber 301, 30b2 can be evacuated in its interior to high vacuum, into which a process gas and radio-frequency power is supplied through a gas port 30q. The process gas is passed through the mesh electrode 30i and supplied to a vicinity of the substrate 1. The process gas is exited through a vacuum manifold 30r. By controlling the supply flow rate and exit velocity of the process gas, pressure is put under control. High-frequency power is applied to the electrode 30i to generate plasma at between the susceptor 30f and the electrode 30i. High-frequency power uses a frequency, for example, of 13.56 MHz.
[0285] Within the process chamber 30b1 is made, for example, hydrogen plasma process and ammonia plasma process as in the foregoing. It is noted that such hydrogen plasma process and ammonia plasma process may be respectively carried out in separate plasma process chambers. Meanwhile, cap-film (silicon nitride film) deposition is carried out within the process chamber 30b2. Because the process chamber 30b1 and the process chamber 30b2 are mechanically connected together through the load lock chamber 30a, the substrate 1 after hydrogen plasma processing and ammonia plasma processing can be transferred to the process chamber 30b2 without encountering vacuum fracture. Thus, plasma processing (post-processing) and cap-film formation can be carried out in a continuous fashion.
[0286] In plasma processing (post processing), for example the following operations are carried out. A substrate 1 is transferred from a cassette interface 30c to the load lock chamber 30a by the robot 30d. The load lock chamber 30a is evacuate fully to a pressure-reduced state, and the substrate 1 is transferred into the process chamber 30b1 by the use of the robot 30d. The gate valve 30e of the process chamber 30b1 is closed to evacuate the process chamber 30b1 to a full vacuum. Then, a hydrogen gas or ammonia gas is introduced to the process chamber 30b1 where it is adjusted in pressure and kept at a predetermined pressure. Thereafter, an electric field is applied from the radio-frequency power source to the electrode 30i to plasma-process the surface of the substrate 1. After a lapse of a predetermined time, the high-frequency electric field is shut down to cease the plasma. Thereafter, the process chamber 30b1 is evacuated and the gate valve 30e is opened to transfer the substrate 1 out to the load lock chamber 30a by the robot 30d. It is noted that the substrate 1 at its surface will not be exposed to the air ambient because the load lock chamber 30a is kept in a high vacuum state.
[0287] Subsequently, in cap-film formation, for example the following operations are carried out. First, using the robot 30d a substrate 1 is transferred into the process chamber 30b2. The gate valve 30e of the process chamber 30b2 is closed to evacuate the process chamber 30b2 fully to a vacuum, and thereafter a mixture gas of silane (SiH4), ammonia and nitrogen is introduced into the process chamber 30b2 where it is adjusted in pressure and kept at a predetermined pressure. Thereafter, an electric field is applied from the radio-frequency power source to the electrode 30i to cause plasma, thereby depositing an insulation film for a cap film (see FIG. 33). After a lapse of a predetermined time, the high-frequency electric field is shut down to cease the plasma. Thereafter, the process chamber 30b2 is evacuated and the gate valve 30e is opened, to transfer the substrate 1 out to the load lock chamber 30a. Furthermore, using the robot 30d the substrate 1 is exited to the cassette interface 30c.
[0288] (Embodiment 2)
[0289] The present embodiment explains a modification to the foregoing post-CMP cleaning process.
[0290]
FIG. 43 shows a flowchart for the modification. Exemplification is herein made on the case of carrying out the foregoing deoxiding process (step 106) after the CMP process (step 104) and thereafter post-cleaning process in the order of alkali cleaning (step 105) and acid cleaning (step 107). The other points than the above are similar to those of the Embodiment 1. In also this case, obtained is the similar effect to the post-CMP cleaning process. In this case, a deoxidization processing section may be provided in the CMP apparatus.
[0291] Meanwhile, FIG. 44 shows another example of a flowchart. Exemplification is herein made on the case that the invention is applied where acid cleaning only is carried out without performing alkali cleaning. The others than the above are similar to those of the Embodiment 1. The TDDB characteristic is improved by merely carrying our acid cleaning. This can be considered owing to interfacial characteristic improvement by the removal of damaged layers. In also this case, a deoxidization processing section may be provided in the CMP apparatus.
[0292] (Embodiment 3)
[0293] The present embodiment explains a modification to the foregoing post process.
[0294]
FIG. 43 shows part of the flowchart. Exemplification is herein made on the case that, after an ammonia plasma process, a hydrogen plasma process is continuously made while maintaining a vacuum. The others than the above are similar to those of the Embodiments 1 and 2. Meanwhile, FIG. 44 exemplifies a case that an ammonia plasma process only is carried out in a post process. The others than the above are similar to those of the Embodiments 1 and 2. In these cases, the TDDB characteristic could have been improved despite the interconnect resistance has decreased lower than that of the Embodiments 1 and 2.
[0295] (Embodiment 4)
[0296] The present embodiment explains a modification to the foregoing post process.
[0297]
FIG. 47 shows part of the flowchart. Exemplification is herein made on the case that a hydrogen anneal process (step 200) is carried after a drying process (step 108) but before a hydrogen plasma process (step 109). Namely, exemplification is made on a case that the processes are carried out in the order of a drying process (step 108), a hydrogen anneal process (step 200), a hydrogen plasma process (step 109), an ammonia plasma process (step 110) and a cap film forming process (step 111).
[0298] Meanwhile FIG. 48 exemplifies a case that the post process of FIG. 47 is changed in sequence. Namely, exemplification is made on the case that hydrogen anneal (step 200), an ammonia plasma (step 110) and hydrogen plasma (step 109) are carried out in the order.
[0299] Furthermore, FIG. 49 exemplifies a case that, after hydrogen anneal (step 200), ammonia plasma (step 110) only is carried out as a post process. Namely, exemplification is made on the case that hydrogen anneal (step 200), ammonia plasma (step 110) and cap-film forming process (step 111) are carried out in the order.
[0300] In any of the above, the conditions for hydrogen anneal was given as a processing temperature, for example, of 200-475° C., preferably approximately 300° C., and a processing time, for example, of 0.5-5 minutes, preferably approximately 2 minutes.
[0301] This method is suited particularly for the case to form a main conductor film comprising copper for a buried interconnect by a plating technique. Also, this is suited for the case not to carry out hydrogen anneal during post-cleaning process or the immediately before deoxidizing process. In this manner, by carrying out a hydrogen anneal process (step 200), the copper formed by the plating technique can be recrystallized thus making possible to reduce interconnect resistance. Meanwhile, if a cap film (insulation film 19b) is deposited without performing the hydrogen anneal, there is a case that the cap film be stripped off due to thermal stress. However, a hydrogen anneal process if carried out can suppress or prevent this.
[0302] It is noted that, in the post process of the Embodiment 3, the means for carrying out hydrogen plasma and ammonia plasma processes and the means of the Embodiment 4, if combined with the case of not carrying out a deoxidizing process during the post-CMP cleaning process or immediately before that as shown in FIG. 50 or FIG. 51, can obtain the effect of each means. FIG. 50 exemplifies a case that alkali cleaning (step 105) and acid cleaning (step 107) are carried out in the order during post-CMP cleaning while FIG. 51 exemplifies a case that acid cleaning (step 107) only is carried out during post-CMP cleaning.
[0303] (Embodiment 5)
[0304] A fabrication method of a semiconductor integrated circuit device of the present embodiment will be explained, with reference to FIG. 52 to FIG. 61.
[0305]
FIG. 52 is a schematic view showing one example of the overall structure of a CMP apparatus to be used for forming copper-based buried interconnects.
[0306] As shown in the figure, the CMP apparatus 31 has a polish processing section 31a and a post-cleaning section 31b provided in a rear stage thereof. The polish processing section 31a is provided with two platens (first platen 31c1, second platen 31c2) to perform a polishing process on a substrate 1, a clean station 31d to preliminarily clean the substrate 1 completed of polishing and subjecting an anti-corrosion process to a surface thereof, a rotary arm 31g to move the substrate 1 between the loader 31e, the first platen 31c1, the second platen 31c2, the clean station 31d, the unloader 31f and so on.
[0307] In a rear stage of the polish processing section 31a, a post-cleaning section 31b is provided to perform scrub cleaning on the surface of a substrate 1 completed of preliminary cleaning. The post-cleaning section 31b is provided with a loader 31h, a first cleaning section 31i1, a second cleaning section 31i2, a spin drier 31j, an unloader 31k and so on. Meanwhile, the post-cleaning section 31b, in the entire, is surrounded with a shade wall 31m in order to prevent light from illuminating onto the surface of a substrate 1 under cleaning. The interior is made in a darkroom state at a 180 lux or less, preferably 100 lux or less. This is because, if light is illuminated to a substrate 1 adhered with a polishing solution in a wet state, a short current flows through a pm junction due to silicon photovoltaic power to dissociate Cu ions from the surface of a Cu interconnect connected to a p-side (+ side) of the pn junction thereby causing interconnect corrosion.
[0308] As shown in FIG. 53, the first platen 31c1 is rotatively driven in a horizontal plane by a drive mechanism 31n provided under the same. Also, the first platen 31c1 has, on its upper surface, a polishing pad 31p formed by evenly bonding a synthetic resin such as polyurethane having a multiplicity of pores. In the above of the first platen 31c1 is provided a wafer carrier 31r to be vertically moved and rotatively driven in a horizontal plane by a drive mechanism 31q. A substrate 1 is held with its main surface (polish surface) directed down by a wafer chuck 31s and retainer ring 31t provided in a lower end of the wafer carrier 31r, to be pressed at a predetermined load onto the polishing pad 31p. Slurry (abrasive liquid) S is fed through a slurry feed pipe 31u to between a surface of the polishing pad 31p and a polish surface of the substrate 1, to chemically mechanically polish the polish surface of the substrate 1. Meanwhile, in the above of the first platen 31c1 is provided a dresser 31w to be vertically moved and rotatively driven in a horizontal plane by a drive mechanism 31v. In a lower end of the dresser 31w is attached a base member electro-deposited with diamond particles. The polishing pad 31p in its surface is periodically ground by the substrate material in order to prevent loading due to polishing grits. It is noted that the second platen 31c2 is nearly in the same structure as the first platen 31c1 except in that two slurry feed pipes are provided.
[0309] To form a copper-based buried interconnect by using the CMP apparatus 31, the substrate 1 received by the loader 31e is transferred to the polish processing section 31a by the use of the rotary arm 31g. This substrate 1 has passed the same process as that explained in FIGS. 23 to 25 mentioned before. As shown in FIG. 54, the substrate 1 is chemically mechanically polished using slurry free of grits (grit-free chemical mechanical polish) on the first platen 31c1 (first step CMP) thereby being removed of the copper-based main conductor film 22a in an outside of the interconnect trench 20a, as shown in FIG. 55.
[0310] Herein, grit-free chemical mechanical polish means chemical mechanical polish using an abrasive liquid (slurry) having grits comprising a powder of alumina, silica or the like in a content of 0.5% by weight or less. The abrasive liquid preferably has a grit content of 0.1% by weight or less, more preferably 0.05% by weight or less or 0.01% by weight or less.
[0311] Also, the abrasive liquid to be used is adjusted in its pH to be fallen in a copper corrosive range and in its composition to have a polish selective ratio of a main conductor film 22a to a conductive barrier film 21a of at least 5 or greater. The abrasive liquid alike this is exemplified by slurry containing oxidizer and organic acid. The oxidizer is exemplified by hydrogen peroxide, ammonium hydroxide, ammonium nitrate and ammonium chloride. The organic acid is exemplified by citric acid, malonic acid, boletic acid, malic acid, adipic acid, benzoic acid, phthalic acid, tartaric acid, lactic acid, succinic acid and oxalic acid. Among these, hydrogen peroxide not containing metal ingredients and not a strong acid is a suitable oxidizer for use in an abrasive liquid. Meanwhile, citric acid, generally used as food additives, is low in toxicity, low in waste-fluid harm, less odor and high in water dissolution and hence is a suitable organic acid for use in an abrasive liquid. The present embodiment uses an abrasive liquid having pure water added with hydroxide peroxide in an amount of 5% by volume and citric acid in an amount of 3% by weight to have a grit content of less than 0.01% by weight.
[0312] If chemical mechanical polish is made using an abrasive liquid as above, oxidation occurs in a surface of copper to form a thin oxide layer in the surface. When an oxide-solubilizing substance is supplied, the oxide layer turns into a solution and liquates out thereby decreasing the thickness of the oxide layer. The portion thinned in the oxide layer is again exposed to a oxidizing substance to increase the thickness of the oxide layer. Repeating the reaction, chemical mechanical polish proceeds. It is noted that the chemical mechanical polish using such a grit-free abrasive liquid is concretely described in Japanese Patent Application No. 299937/1997 and Japanese Patent Application No. 317233/1998.
[0313] The polish conditions, as one example, are under a load of 250 g/cm2, a wafer-carrier rotation velocity of 3.0 rpm, a platen rotation velocity of 25 rpm and a slurry flow rate of 150 cc/min, wherein the polish pad uses a hard pad (IC1400) by U.S. Rodel. The end point of polish is at a time that the main conductor film 22a is removed to expose the underlying conductive barrier film 21a. The detection of an end point is due to detecting a platen or wafer-carrier rotation torque signal intensity changed when a subject of polish is changed from a main conductor film 22a to a conductive barrier film 21a. Meanwhile, it is satisfactory to open a part of the polishing pad in order to detect an end point depending upon a spectral change in light reflection from the substrate 1 surface, or detect an end point depending upon an optical spectral change of slurry.
[0314] As shown in FIG. 55, by carrying out the grit-free chemical mechanical polish, the main conductor film in a portion outside the interconnect trench 20a is almost removed to expose the underlying conductive barrier film 21a. However, as shown with magnification in FIGS. 56A and 56B, an unremoved portion of the main conductor film 22a in the present polish remains in a cavity (shown at the arrow) of the conductive barrier film 21a caused due to the underlying steps.
[0315] Next, in order to remove the conductive barrier film 21a outside the interconnect trench 20a and the main conductor film 22a locally left on an upper surface thereof, the substrate 1 is transferred from the first platen 31c1, shown in FIGS. 52 to 54, to the second platen 31c2 to carry out chemical mechanical polish using abrasive liquid containing grits (slurry) (grit-used chemical mechanical polish) (second-step CMP). Herein, grit-used chemical mechanical polish means chemical mechanical polish using an abrasive liquid having grits of aluminum, silica or the like in a content of greater than 0.5% by weight. The present embodiment uses an abrasive liquid of pure water added with hydrogen peroxide in an amount of 5% by volume, citric acid in an amount of 0.03% by weight and grits in an amount of 0.5% by weight, but not limited to this. The abrasive liquid is fed to a polishing pad 31p of the second platen 31c2 through the slurry feed pipe 31u.
[0316] Meanwhile, in the grit-used chemical mechanical polish, the conductive barrier film 21a outside the interconnect trench 20a is removed following the removal of the locally left main conductor film 22a on the upper surface of the conductive barrier film 21a. For this reason, polish is made under a condition that the polish selective ratio of the conductor barrier film 21a to the main conductor film 22a is lower than that of the foregoing grit-free chemical mechanical polish, to suppress the surface of the main conductor film 22a at an inside of the interconnect trench 20a from being polished.
[0317] The polish conditions, as one example, are under a load of 120 g/cm2, a wafer-carrier rotation velocity of 30 rpm, a platen rotation velocity of 25 rpm and a slurry flow rate of 150 cc/min, wherein the polish pad uses IC1400 by Rodel. The polish amount corresponds to a film thickness of the conductive barrier film 21a. The end point of polish is put under control by a time computed from a film thickness of and polish rate on the conductive barrier film 21a.
[0318] As shown in FIG. 57, the grit-used chemical mechanical polish removes almost the conductive barrier film 21a in an outside of the interconnect trench 20a to expose the underlying insulation film 16d. However, as shown with magnification in FIGS. 58A and 58B, an unremoved portion of the conductive barrier film 21a in the polish remains in a cavity of the insulation film 16d caused due to an underlying step (shown at the arrow) or the like.
[0319] Next, carried out is selective chemical mechanical polish to remove the locally left conductive barrier film 21a on the insulation film 16d outside the interconnect trench 20a while suppressing the polish of the main conductor film 22a at an inside of the interconnect trench to a possible less extent (third-step CMP). This selective chemical mechanical polish is made under a condition that the polish selective ratio of the conductive barrier film 21a to the main conductor film 22a is at least 5 or greater. Also, this chemical mechanical polish is under a condition that the ratio of a polish rate on the insulation film 16d to a polish rate on the main conductor film 22a is greater than 1.
[0320] The selective chemical mechanical polish generally uses an addition of an anti-corrosive agent to an abrasive liquid containing grits in an amount of greater than 0.5% by weight as was used in the grit-used chemical mechanical polish. The anti-corrosive agent refers to an agent which forms an anti-corrosive protection film over a surface of the main conductor film 22a thereby preventing or suppressing polish from proceeding, and uses Benzotriazole (BTA), BTA derivatives such as BTA carbonic acid, dodecylmercaptan, triazole, tolyltriazole or the like. Particularly, the use of BTA forms a stable protection film.
[0321] Where using BTA as an anti-corrosive agent, although the concentration thereof is dependent upon the kind of slurry, a sufficient effect is available by an addition of usually 0.001-1% by weight, more preferably 0.01-1% by weight, and further preferably 0.1-1% by weight (three stages). The present embodiment uses a mixture of the abrasive liquid as was used in the grit-used chemical mechanical polish in the second step, as an abrasive liquid, added with BTA as an anti-corrosive agent in an amount of 0.1% by weight. This however is not limited to. Meanwhile, in order to avoid the lower in polish rate due to the addition of the anti-corrosive agent, polyacrylic acid, polymethacrylic acid, ammonium salt of these, ethylenediaminetetraacetic acid (EDTA) or the like may be added as required. It is noted that the chemical mechanical polish using slurry containing such an anti-corrosive agent is concretely described in Japanese Patent Application No. 209857/1998, Japanese Patent Application No. 299937/1997 and Japanese Patent Application No. 317233/1998.
[0322] This selective chemical mechanical polish (third-step CMP) is continuously carried out on the second platen 31c2 of the CMP apparatus shown in FIGS. 52 to 54 after completing the grit-used chemical mechanical polish (second-step CMP). The abrasive liquid added with the anti-corrosive agent is fed to a surface of the polishing pad 31p through the slurry feed pipe 31u. The polish condition, as one example, is under a load of 120 g/cm2, a wafer-carrier rotation velocity of 30 rpm, a platen rotation velocity of 25 rpm and a slurry flow rate of 190 cc/min.
[0323] As shown in FIG. 59 and FIGS. 60A and 60B, the selective chemical mechanical polish removes all the conductive barrier film 21a at the outside of the interconnect trench 20a to thereby form a buried interconnect 23a at an inside of the interconnect trench 20a.
[0324] On a surface of the substrate 1 completed of buried interconnect 23a formation is adhered slurry residue containing particles of grits or the like or metal particles such as Cu oxide. Consequently, in order to remove the slurry residue, first the substrate 1 is cleaned with pure water containing BTA in the clean station 31d shown in FIG. 52. In this case, mega-sonic cleaning may be used together that liberates the slurry residue from the substrate 1 surface by applying a high-frequency vibration of 800 kHz or higher to the cleaning liquid. Next, in order to prevent drying in the surface, the substrate 1 held in a wet state is transferred from the polish processing section 31a to the post-cleaning section 31b. Scrub cleaning is made using a cleaning liquid containing NH4OH in an amount of 0.1% by weight in the first cleaning section 31i1, and subsequently scrub cleaning is made using pure water in the second cleaning section 31i2. As noted before, the post-cleaning section 31b in its entirety is covered with a shade wall 31m in order to prevent against the occurrence of corrosion in the buried interconnect 23a due to illumination of light onto the substrate 1 under cleaning.
[0325] The substrate 1 completed of the scrub cleaning (post cleaning) is dried by the spin drier 31j and then transferred to the next process. The following processes are similar to those of the Embodiment 1. FIG. 61 is a flowchart showing a part of the buried-interconnect forming process. The other processes are the same as the means of Embodiments 1 to 4.
[0326] According to the present embodiment, the TDDB life can be improved greater than that of the Embodiment 1. FIG. 62 is a graph showing a TDDB life in the present embodiment. The data in this embodiment is shown at a line E. For reference, shown at the same time are the data of no-processing (line Ref) and of grit-used chemical mechanical polish (Embodiment 1) (line A). It is noted that the TDDB characteristic is improved as shown at a line F by carrying out only grit-free chemical mechanical polish without performing an ammonia plasma process. It can be considered that the reason of improving the TDDB life in a grit-free case in this manner is because the damage to a silicon oxide film can be reduced. In a grit-use case, slurry contains grits (alumina, etc.) having a particle size of 2 μm-3 μm (secondary particle size). The grits causes micro-scratches to give damage to a surface of the silicon oxide film (insulation film 16d). In the grit-free case, however, the slurry contains no grits or, if contained, a little bit of grits, which greatly relieves the damage. It can be considered that the TDDB characteristic is improved due to the above.
[0327] (Embodiment 6)
[0328] The present embodiment uses, as an agent liquid, the foregoing organic acid or a mixture agent liquid of hydrogen fluoride and organic acid during an acid cleaning process (step 107 in FIG. 22) in the post-CMP cleaning process of the Embodiment 1. The others than the above are the same as those in Embodiments 1 to 5. In the case of using citric acid as the organic acid, brush scrub cleaning for example is used wherein the conditions of a citric acid concentration of 5% and a cleaning time of 45 seconds are to be selected.
[0329] By thus using organic acid cleaning, the damaged layer in the surface caused by CMP or the like can be removed. This improves the TDDB life. FIG. 63 is a graph showing a TDDB life in the present embodiment case. The data that citric acid is applied in the present embodiment case is shown at a line H while the data that HF cleaning is applied is shown at a line I. For reference, shown at the same time are the data of no-processing (line Ref) and the data in the present embodiment (line A). Meanwhile, where using organic acid, there is also an effect to remove only metal ions without causing damage to the underlying layer. Namely, impurities such as Fe, K and Ca can be selectively removed.
[0330] (Embodiment 7)
[0331]
FIG. 64 to FIG. 66 is a plan view and sectional view showing a fabrication method of a semiconductor integrated circuit device in an embodiment of the invention. An interconnect part only is shown in FIG. 64 to FIG. 66.
[0332] As shown in FIG. 64, a buried interconnect 23a and an insulation film 19b for a cap film are formed similarly to the Embodiment 1, followed by formation thereon of an insulation film 16e having a low dielectric constant comprising a silicon oxide film or the like and an insulation film 16f comprising a silicon oxide film or the like by a plasma CVD technique using TEOS as a material gas.
[0333] The insulation film 16e with a low dielectric constant is structured by a silicon-oxide-based insulation film having a specific dielectric coefficient (ε) of 3.0 or lower, e.g. an inorganic SOG film made of hydrogen silses quioxane as a material, a application-type insulation film such as an organic SOG film made as a material of tetra alkoxy silane+alkyl alkoxy silane, or a fluorocarbon polymer film formed by a plasma CVD technique. The use of such a low dielectric coefficient of silicon oxide film reduces an interlevel parasitic capacitance thereby promoting to improve the operation speed of the semiconductor integrated circuit device.
[0334] Next, with using a pattern as shown in FIG. 65A, through-holes 33 are opened as shown in FIG. 65B. Photolithography and etching are used for opening through-holes 33. In the meanwhile, the insulation film 16e low in dielectric constant, having a film structure having a coarse surface, possesses many Si—OH bonds. Due to this, it has been empirically revealed that the film formed in the upper layer thereof is not favorable in film quality and interfacial state. It has also been revealed that, where a conductive barrier film (titanium nitride or the like) is formed directly without processing, the TDDB characteristic is not favorable. For this reason, an ammonia plasma process as was explained in the embodiment is carried out to an exposed portion of the insulation film 16 in an inside of the through-hole 33. Due to this, the Si—OH bonds in the surface are improved in nature and turned into Si—O—N bonds as was explained in the Embodiment 1.
[0335] Next, as shown in FIG. 66 a plug 34 is formed, for example, of titanium nitride and tungsten in the through-hole 33. During deposition of titanium nitride, Si—O—N bonds are desorbed similarly to Embodiment 1 thereby improving the interface of the titanium nitride and the low-dielectric-coefficient insulation film 16e and hence enhancing adhesion. It is of course that such an in-through-hole plasma process is also applicable to an interconnect trench.
[0336] Meanwhile, in place of the ammonia plasma process it is satisfactory to use a hydrogen plasma process or a plasma process mixed with nitrogen, argon, helium and the like. Of course, both of ammonia plasma process and hydrogen plasma process as in the foregoing may be applied. This can improve the effect furthermore. The other points than this are the same as those of the Embodiments 1-6.
[0337] Incidentally, in an ashing process for removing a photoresist film after opening a through-hole 33, oxidation in some cases occurs in the surface of the buried interconnect 23a at a bottom of the through-hole 33. There is an art described in JPA-11-16912 as a technique for removing such an oxide layer.
[0338] (Embodiment 8)
[0339] As in the foregoing, it is considered that a barrier film for preventing copper diffusion is essentially required in putting the copper interconnect into practical use. However, as the interconnect scales down, interconnect resistance increases due to increase of a sectional area of a high-resistance barrier film occupying in an interconnect sectional area, with a result that a problem arises the merit of application of copper as an interconnect material weakens.
[0340] For this reason, the present inventors have again conducted an experiments and studied on the copper diffusion phenomenon. As a result, the present inventors have first found a substantial mechanism on a copper diffusion phenomenon as in the foregoing. FIG. 67 is a graph showing a silicon content in a copper interconnect upon each kind of a surface processing (ammonia plasma process, hydrogen plasma process, hydrogen anneal process, nitrogen plasma process) and not-processing. The inspection result has been prepared by the inspections conducted after the copper interconnect (including TiN film (barrier film)) forming process, the cleaning process (not including the deoxidizing process), the surface processing process (only each single plasma process) in various kind, the cap-film forming process and the interlayer-insulation-film forming process. Also, it is considered that similar results be obtained for other impurities, such as oxygen and sulfur.
[0341] The copper silicide in various surface processes is mainly due to set flow upon forming a cap film (silicon nitride or the like). In a hydrogen anneal process or nitrogen plasma process at this inspection time point, the content of Si is high in or around the surface (d=approx. 10 nm-60 nm) and in an interior (d=approx. 90 nm-300 nm) of a copper interconnect as compared to that of the ammonia plasma process or hydrogen plasma process. Particularly, it can be understood that it is extremely high in or around the surface. In these processes, the TDDB characteristic was poor.
[0342] On the other hand, in an ammonia plasma process or hydrogen plasma process at the inspection time point, the content of silicon is low in or around the surface (d=approx. 10 nm-60 nm) and in an interior of a copper interconnect as compared to that of the hydrogen plasma process or nitrogen plasma process. Particularly, the silicon content in or around the surface is extremely low. Namely, in these processes, there is less content of impurities in the copper interconnect and cleanness is high in the surface of the insulation film where the interconnect trench, and further there are less dangling bonds or the like in the surface of the insulation film where the interconnect trench is formed. Accordingly, the TDDB characteristic was favorable as shown in FIG. 21. In this manner, where there exists a TiN film (conductive barrier film), the TDDB characteristic is determined only by the affection of an interface.
[0343] From such a new point of view, the present inventors have first found that, by forming not-ionized neutral Cu in a side wall and bottom of an interconnect trench (increasing copper purity), carrying out the ammonia plasma process or hydrogen plasma process or combining these with the CMP process or cleaning process or the like, even if the film thickness of a barrier film is reduced in thickness to smaller than 10 nm or the barrier film itself is eliminated, the TDDB life of a semiconductor integrated circuit device having a Cu interconnect can be improved.
[0344] Herein, FIG. 68 shows a dependency of an interconnect resistance (ratio of TiN·xnm/TiN·50 nm) on a TiN film (barrier film) thickness. The figure shows an actual measurement value and a theoretical value (calculated value) for a trench form having a interconnect width, for example, of approximately 0.4 μm and approximately 1.0 μm and a interconnect-trench depth, for example, of approximately 0.4 μm. It is noted that the thickness of the TiN film is given by a film thickness at a bottom of the interconnect trench.
[0345] It is understood from FIG. 68 that, as the thickness of the TiN (barrier film) decreases, the interconnect resistance also decreases whereby a calculated value and an actual measurement value comes into coincidence. Consequently, as compared to an interconnect resistance for a TiN film thickness of 50 nm, where there is no TiN film, the interconnect resistance can be greatly decreased by nearly 19% for an interconnect width of nearly 0.4 μm and by nearly 15% for an interconnect width of nearly 1.0 μm. It is also understood that, even where the film thickness of the TiN film is nearly 10 nm, the interconnect resistance can be decreased by nearly 16% for an interconnect width of nearly 0.4 μm and by nearly 12% for an interconnect width of nearly 1.0 μm.
[0346] Meanwhile, FIG. 69 shows a TiN film dependency of a TDDB characteristic where a copper interconnect is formed by a long-throw sputtering technique. It is understood from the figure that a TDDB characteristic for a TiN film thickness of 10 nm-50 nm is equivalent to the foregoing. On the other hand, it is understood that a TDDB characteristic of a sample in the absence of a TiN film is moderate in inclination as compared to a TDDB characteristic for a TiN film of nearly 10-50 nm but fully in excess of a new system target (e.g. 0.2 MV/cm, 110 ° C., 10 yrs.=3×108 sec.).
[0347] Meanwhile, FIG. 70 shows a TDDB characteristic on each copper interconnect due to a presence or absence of heating process in the absence of a TiN film is absent and a TiN film is formed in a thickness of nearly 10 nm. It is understood from the figure that, even on a sample in the absence of a TiN film, the TDDB characteristic is not deteriorated by the heating process, for example, at 400° C. for 3 hours. From the evaluation result on FIG. 69 and FIG. 70, it has been first revealed by the inventor's experiment that, even where a TiN film is absent, i.e. an interconnect is formed by copper only, it is possible to achieve sufficient reliability and form a practical Cu interconnect. It is noted that this is described in Japanese Patent Application No. 104015/2000 by the present inventors.
[0348] Next, a concrete example of an interconnect structure of the semiconductor integrated circuit device of the present embodiment is shown in FIG. 71. FIG. 71 is a sectional view showing, by taking out, an interconnect part of the semiconductor integrated circuit device, wherein FIG. 71A shows a part formed by a single damascene technique and FIG. 71B a part formed by a dual damascene technique.
[0349] It is noted that, in FIG. 71, references 16g, 16i comprise the same material selected from the low-dielectric-constant insulation films as that of an insulation film 16a while references 16h, 16j are the same material as that of the insulation film 16b. Also, FIG. 71B exemplifies a case that, in insulation films 16c, 16d is formed such a through-hole 35 as exposing a part of an upper surface of a first-level interconnect M1. A reference 19c is formed, for example, of silicon nitride.
[0350] The interconnect width (width of a interconnect trench 20a) and adjacent interconnect spacing (distance between the side surfaces of adjacent interconnects) is, for example, 0.4 μm or smaller. The interconnect width and adjacent interconnect spacing being considered by the present inventors is on a semiconductor integrated circuit device having an interconnect structure, for example, of 0.25 μm or smaller, or 0.2 μm or smaller. The aspect ratio of an interconnect trench 42 is, for example, 1.
[0351] Meanwhile, the thickness of a conductive barrier film 21a is, for example, smaller than 10 nm, preferably nearly 6-7 nm. In the present embodiment, even where the film thickness was, for example, 5 nm or the smaller 3 nm, or the further smaller nearly 2 nm, the TDDB characteristic could have been improved. The thickness of the conductive barrier film 21a, referred herein, signifies a surface point to be deposited in the smallest thickness. Namely, because herein the conductive barrier film 21a in the interconnect opening (interconnect trench 20a or through hole 35) is deposited the smallest in film thickness at the sidewall, referred to is the thickness of the conductive barrier film 21a at the sidewall. Furthermore, in such a case, there are the following two structures, for example. One is that, on the sidewall in the interconnect opening (including bottom corners of the trench or hole), the conductive barrier film 21a has the smallest thickness equal to the above thickness (e.g. smaller than 10 nm, preferably nearly 6 nm-7 nm, 5 nm or smaller, 3 nm or smaller or nearly 2 nm). Another is that, on the sidewall in the interconnect opening, the conductive barrier film 21a has the greatest thickness equal to the above thickness (e.g. smaller than 10 nm, preferably nearly 6 nm-7 nm, 5 nm or smaller, 3 nm or smaller or nearly 2 nm).
[0352] By forming the conductive barrier film 21a having a thickness of smaller than 10 nm, the conductive barrier film 21a is superior, in adhesion to the insulation film 16b, 16c, 16d, 16g, 16h, to the copper film, thus preventing the main conductor film 22a from stripping off during CMP processing. Also, a reliable copper interconnect structure can be realized despite the interconnect resistance increases as compared to the case of not providing a conductive barrier film 21a (explained in the hereinafter embodiment). Also, the TDDB characteristic can be improved as compared to the case of not providing a conductive barrier film 21a. It can be considered that, where there is no conductive barrier film 21a at all, copper bombards against the sidewall of the interconnect trench 20a to react with SiO2 during forming a main conductor film 22a with a result that a slight amount of copper occurs. From no deterioration in the TDDB characteristic even after heating process, it can be considered that the above is due to the affection of a copper ion layer at its slight Cu/SiO2 interface. Consequently, according to the present embodiment, it can be considered that the barrier film 21a even smaller than slight 10 nm serves as a barrier against ionized copper thereby improving the TDDB characteristic.
[0353] Meanwhile, the concentration of the other ingredient than copper of the main conductor film is 0.8 atomic % or less or 0.2 atomic % or less. Also, according to the result of measurement by the present inventors, the concentration of the other ingredient than copper could be reduced, for example, to 0.08 atomic % or less or the lower 0.05 atomic %, or the further lower 0.02 atomic % or less. The concentration value of the other ingredient than copper is a value of at a time that the semiconductor chip is completed, i.e. at a time that semiconductor chips are cut out of the semiconductor wafer through a wafer process, and a value calculated on the assumption that the other ingredient than copper is diffused into the interconnect by the heat of upon the film-forming process for an insulation film or metal film after forming the copper interconnect (e.g. for tungsten or the like, heat at about 450° C. is applied during forming a film). In an actual copper interconnect, it can be considered that the other ingredient than copper distributes in a state the concentration is high in an upper layer portion of the copper interconnect (portion the cap film contacts) but gradually decreases toward the center of the copper interconnect. The other ingredient than copper includes, for example, silicon, oxygen, sulfur (sulfur is considered in the case the copper interconnect is formed by a plating technique) or an arbitrary combination of them.
[0354] It is noted that the material of an interlayer insulation film can use a silicon oxide film in place of a low-dielectric-coefficient material such as of the insulation film 16a. Also, a PSG film if used serves to prevent Cu from diffusing thus further improving the TDDB life. Consequently, the reliability of the semiconductor integrated circuit can be improved furthermore.
[0355] Next, one example of a copper interconnect structure forming method by such a single damascene will be described with reference to FIGS. 72 to 76. Incidentally, FIGS. 72A to 76A each show an essential-part plan view in a fabrication process of a semiconductor integrated circuit device while FIGS. 72B to 76B each show a sectional view on line A-A in the corresponding figure. Also, FIGS. 72A-75A are plan views, wherein hatching is provided to a metal film for easy understanding.
[0356] As shown in FIG. 72, first an interconnect trench is formed. A top surface of a plug 18 is exposed at a bottom of the interconnect trench 20a. Subsequently, as shown in FIG. 73, a conductive barrier film 21a, for example, of tantalum is deposited in a depo-film thickness, for example, of nearly 30 nm by a sputtering technique or the like. On this occasion, a conductive barrier film 21a is formed in a film thickness, for example, of smaller than 10 nm, nearly 6 nm-7 nm, at a greatest thickness point or a smallest thickness point on a sidewall of the interconnect trench 20a. It is noted that, the conductive barrier film is herein of tantalum, it may be of titanum nitride (TiN) or by other exemplified film.
[0357] Thereafter, a main conductor film 422a of copper is deposited in a depo-film thickness, for example, of nearly 300 nm on the conductive barrier film 21a by a sputtering technique. The conditions in this case, for example, is as follows. Pressure is 0.02 Pa for example, direct-current (DC) power is 10 kW for example, distance between a target and a substrate 1 is 300 nm-400 nm for example, and temperature is room temperature for example.
[0358] In this manner, the present embodiment deposits a main conductor film 22a by a sputtering technique thereby making possible to reduce the generation of compounds as compared to the CVD technique or plating technique. Also, the target in that case used oxygen-free Cu having a high purity, for example, of 99.999% (5N) or higher, preferably 99.9999% (6N) or higher. This can make the concentration of copper in the main conductor film 22a being formed, for example, to 99.999% or higher, preferably 99.9999% or higher. Consequently, copper can be deposited with further higher purity.
[0359] In depositing a conductive barrier film 21a and main conductor film 22a, the usual sputtering technique is satisfactory. However, a high-directivity sputtering technique may be used, including a long-throw sputter technique and a collimate sputtering technique. It is possible, in such a case, to improve coverage of a metal film to an interconnect trench 20a.
[0360] Then, a hydrogen anneal process is carried out. This favorably buries the main conductor film 22a in the interconnect trench 20a. The conditions on that occasion is, for example, nearly, 475° C., 3 minutes, 26.6644×102 Pa, and 500 cm3/min.
[0361] Subsequently, the main conductor film 22a and conductive barrier film 21a is removed of extra portions by a similar CMP technique to those explained in the Embodiments 1-7, as shown in FIG. 74, thereby forming a main conductor film 22a. Subsequently, an anti-corrosion process and post-cleaning process is carried out that is similar to that of the foregoing embodiment. Thereafter, as shown by the meshed hatching in FIG. 75, a hydrogen plasma process and ammonia plasma process is carried out on a surface of the insulation film 16d and main conductor film 22a, as explained in the Embodiments 1-4.
[0362] The ammonia plasma process forms SiH bonds and SiN bonds in a surface region of the insulation film 16d and the like, with a result that the surface region of the insulation film 16d can be improved in film quality, cleanness and electrical stability thus making possible to improve copper-diffusion preventive property. Also, as was explained in the Embodiment 1, the adhesion can be improved to the cap film. Also, copper nitride (CuN) is formed in a surface region of the main conductor film 22a. The copper oxide acts to prohibit silicon and oxygen from bonding to resultingly prevent the formation of copper silicide and copper oxide, making possible to improve copper purity. it is therefore possible to prevent copper diffusion and improve the TDDB life. Also, because of high purity of copper, the resistance of a copper interconnect in a state completed as a semiconductor chip can be decreased according to a purpose thereof. Due to this, the semiconductor integrated circuit device can be improved in performance.
[0363] Meanwhile, as a result of SiH formation in a surface region of the insulation film 16d by carrying out the hydrogen plasma process, almost the same effect has been obtained as in the case of the ammonia plasma process. Also, according to a result of experiment by the present inventors, although, in the hydrogen plasma process, copper reacts with nearly several percent of silicon in a later cap-film forming process, leak current could be greatly reduced as compared to the case of hydrogen anneal process, nitrogen plasma process or no processing, thus making possible to improve the TDDB life. Also, the resistance of the copper interconnect could be reduced as compared to the case of a hydrogen anneal or nitrogen anneal process.
[0364] Thereafter, as shown in FIG. 76, an insulation film 19b for a cap film is deposited similarly to the Embodiment 1 or the like. In this manner, an interconnect structure shown in FIG. 71A is obtained.
[0365] Next, one example of a copper interconnect structure forming method by a dual damascene technique will be explained, with reference to FIG. 77 to FIG. 85. It is noted that A in each of FIG. 77 to FIG. 85 shows a plan view of an essential part in a fabrication process of a semiconductor integrated circuit device, and B shows a sectional view on line A-A in A of each figure. Also, A in each of FIG. 81 to FIG. 84 is a plan view, wherein hatching is provided to a metal film for easy understanding.
[0366] First, as shown in FIG. 77, a reflection preventing film 40 is applied onto an insulation film 16h, on which a resist pattern 41 is formed. The photoresist pattern 41, a mask pattern for example for forming planar circular hole, is formed by the usual photolithography technique. Subsequently, as shown in FIG. 78, using the photoresist pattern as a mask, the reflection preventing film 40 in a portion exposed therefrom is etch-removed by a dry etching technique. Furthermore, using an insulation film 19a as an etching stopper, the underlying insulation films 16h, 16g, 16d, 16c are etch-removed by a dry etching technique thereby forming a through-hole 35. In this stage, the insulation film 19a is exposed at the bottom of the through-hole 35.
[0367] Next, after the photoresist pattern 41 and reflection preventing film 40 is removed as shown in FIG. 79 by an ashing technique, a reflection preventing film 42 is again applied over the entire surface of the insulation film in a manner being buried in the through-hole 35, as shown in FIG. 80. Subsequently, as shown in FIG. 81, a photoresist pattern 43 is formed on the reflection preventing film 42. The photoresist pattern 43 is, for example, a mask pattern for forming a interconnect trench in a planar strip form, which is formed by the usual photolithography technique. Thereafter, as shown in FIG. 82, using the photoresist pattern as a mask, the reflection preventing film 42 in a portion exposed therefrom is etch-removed by a dry etching technique. Furthermore, using an insulation film 19c of silicon nitride or the like as an etching stopper, the insulation films 16h, 16g are etch-removed by a dry etching technique thereby forming an interconnect trench 20a. In this stage, the insulation film 19c is exposed at a bottom of the interconnect trench 20a. In this stage, the insulation film 19c is exposed at the bottom of the interconnect trench 20a.
[0368] Then, the photoresist pattern 43 and reflection preventing film 42 is removed as shown in FIG. 83 by an ashing technique or the like, and thereafter the insulation film 19a, 19c in a portion exposed at the bottom of the interconnect trench 20a and through-hole 35 is selectively removed. In this etching process, the etching selective ratio of the silicon oxide film and silicon nitride film is taken great to perform etching under a condition that the silicon nitride film is more readily etch-removed than the silicon oxide film. This exposes a part of insulation film 16d and first-level interconnect M1 at a bottom of the interconnect trench 20a and through hole 35. This is for providing an electrical connection between the first-level interconnect M1 and upper-leveled buried interconnect. This is also for reducing interconnect capacitance by decreasing the silicon nitride film (insulation films 19a, 19c) higher in dielectric constant than the silicon oxide film to a possible extent. In this manner, an interconnect trench 20a and through-hole 35 is formed.
[0369] Next, as shown in FIG. 85, a conductive barrier film 21a, for example, of tantalum (Ta) is deposited under the same condition as the foregoing by a sputtering technique. On this occasion, a conductive barrier film 21a is formed, e.g. to smaller than 10 nm, nearly 6 nm-7 nm, in a greatest thickness point or a smallest thickness point, on a sidewall of an interconnect trench 20a and through-hole 35. It is noted that, although the conductive barrier film herein is of Ta, the film may be of TiN or other exemplified films as in the foregoing.
[0370] Subsequently, a main conductor film 22a is deposited to a depo-fim thickness of nearly 50 nm on the conductor barrier film 21a by a sputtering technique. The target in that case used oxygen-free copper having a high purity, for example, of 99.999% (5N) or higher, preferably 99.9999% (6N) or higher. This can make the concentration of copper in the main conductor film 22a being formed, for example, to 99.999% or higher, preferably 99.9999% or higher. Consequently, copper with high purity can be deposited on a bottom and sidewall of the copper interconnect.
[0371] Subsequently, furthermore a main conductor film 22a of copper is formed by electroplating technique or the like thereby obtaining a design film thickness of the main conductor film 22a. The condition under which copper is buried in the through-hole 35 by the electroplating technique is, for example, a current density of 0.5-1.0 A/dm2, for nearly 40 seconds. Also, the condition under which copper is buried in the interconnect trench 20a is, for example, a current density of 1.0-2.0 A/dm2, nearly 140 seconds.
[0372] Then, a hydrogen anneal process is carried out similarly to the case of the single damascene technique. This process, in some cases, may be omitted. Subsequently, the main conductor film 22a and conductive barrier film 21a is polished by a CMP technique similar to the above as shown in FIG. 86 thereby being removed of extra portions, and thereafter an anti-corrosion process and post-cleaning process similarly to the above is carried out. Thereafter, as shown by meshed hatching in FIG. 87, an ammonia plasma process and hydrogen plasma process as explained in the Embodiments 1-4 is carried out onto a surface of the insulation film 16h and buried interconnect 23a. This makes it possible to obtain the similar effect to the case explained in the single damascene technique.
[0373] Thereafter, as shown in FIG. 88 an insulation film 19b is deposited similarly to the Embodiment 1 or the like thus obtaining an interconnect structure shown in FIG. 71B.
[0374] With the present embodiment as above, it is possible to obtain, in addition to the effects obtained in the present embodiment, the same effects as were explained in the Embodiments 1-7 as concerned with the same structural parts as those of the Embodiments 1-7.
[0375] (Embodiment 9)
[0376] A concrete example of an interconnect structure for a semiconductor integrated circuit device of the present embodiment will be shown in FIG. 89. FIG. 89 is a sectional view shown by taking out an interconnect part of the semiconductor integrated circuit device, wherein FIG. 89A shows a point formed by a single damascene technique and FIG. 89B a point formed by a dual damascene technique.
[0377] In the present embodiment, a conductive barrier film is not formed. Namely, copper only is buried in the interconnect trench 20a or through-hole 35. Consequently, the buried interconnect 23a at its sidewall and bottom is in a state nearly in direct contact with the insulation films 16b -16d. Meanwhile, the concentration and distribution of other ingredient than copper of the buried interconnect 23a is the same as that explained in the Embodiment 8. Also, the dimensions, such as interconnect width (width of the interconnect trench 20a) and adjacent interconnect spacing (distance between the opposed side surfaces of adjacent interconnects), are the same as those explained with FIG. 71 in the Embodiment 8.
[0378] In the present embodiment alike this, the TDDB life could be improved as was explained in the Embodiment 8. Accordingly, it is possible to improve the yield and reliability of the semiconductor integrated circuit device. Also, in the present embodiment, because a conductive barrier film is not provided but the main conductor film 22a of copper only is buried in the interconnect trench and through-hole 35, making possible to greatly improve interconnect resistance. Meanwhile, because direct connection is provided between different-leveled interconnections without through a conductive barrier film (Ta, TiN or the like) (although shown herein was the structure directly connecting between the buried interconnect 23a and the first-level interconnect M1, there are cases that direct connection is made between different-leveled copper interconnects), the contact resistance between the different-leveled interconnects can be greatly decreased thus making possible to decrease the resistance through a precise through-hole. Accordingly, even where the interconnect trench 20a and through-hole 35 be scaled down, the semiconductor integrated circuit device can be improved in performance.
[0379] The forming method for such a copper interconnect structure is similar to that explained in the Embodiment 8. In this case, it is preferred to carry out a hydrogen plasma process, an ammonia plasma process, a plasma process mixed with nitrogen, argon, helium and the like, or a hydrogen plasma process and ammonia plasma process, after the forming process of an interconnect opening (interconnect trench 22a and through-hole 35) in FIG. 72 and FIG. 84 and before the main conductor film deposition process.
[0380] As in the foregoing, where an ammonia plasma process is carried out, SiH bonds and SiN bonds are formed (e.g. thin silicon nitride film in a thickness of smaller than 10 nm is formed) in the surface of the insulation film 16b, 16c, 16d, 16g, 16h (including the inside of the interconnect trench 22a and through-hole 35). Also, where a hydrogen plasma process is carried out, SiH bonds are formed in the surface of the insulation film 16b, 16c, 16d, 16g, 16h (including the inside of the interconnect trench 22a and through-hole 35). This can improve film quality, cleanness and electrical stability in the upper surface of the insulation film 16d, 16h and surface region of the insulation film 16d, 16c, 16h, 16g in an inside of the interconnect trench 20a and through-hole 35 as was explained in the Embodiment 7, thus improving the diffusion preventing property of copper. Meanwhile, similarly to the explanation in the Embodiment 1, it is also possible to improve the adhesion between the insulation film 16d, 16h and the cap film (insulation film 19b). Incidentally, as was explained in the Embodiment 7, the nitride film (in this case WN film) formed in the first-level interconnect M1 may be removed by carrying out an ammonia plasma process and then lightly carrying out a dry etching process. In the case of adopting such a method, the buried interconnect 23a at its sidewall and bottom is made in a state in direct contact with a thin silicon nitride film formed in the interconnect trench 20a and through-hole 35.
[0381] In the preset embodiment alike this, it is possible to obtain the following effects, in addition to the effects obtained in the structures of the Embodiments 1-8. Namely, due to the absence of a conductive barrier film, it is possible to greatly decrease the resistance of the copper buried interconnect 23a and the contact resistance at a connect point between different-leveled interconnections. Therefore, the semiconductor integrated circuit device can be improved in performance.
[0382] In the above, although the invention by the present inventors was concretely explained by way of the embodiments thereof, it is needless to say that the invention is not limited to the embodiments but may be modified in various ways within a range of not departed from the gist thereof.
[0383] In the embodiment, explanation was made on the case that continuous formation is made without breaking the vacuum after carrying out post-processing (plasma-processing) the cap film. However, a cap film may be formed after once breaking the vacuum after the post-process. Although the effect of the invention is available greater without breaking the vacuum, a thin nitride layer is formed by an ammonia plasma process in the post process so that the formation of an oxide layer can also be suppressed by vacuum breakage for exposure to the air atmosphere. Accordingly, even where vacuum breakage is made, it is possible to provide the effect of the embodiment to a certain extent.
[0384] The above explanations were made on the case that the invention made by the present inventors is applied to the CMIS-LSI art as a background utilization field thereof. However, the invention is not limited to that art but can be applied also, for example, to a semiconductor integrated circuit device having a memory circuit, such as a DRAM (Dynamic Random Access Memory), an SRAM (Static Random Access Memory), a flash memory (EEPROM; Electrically Erasable Programmable Read Only Memory) and an FRAM (Ferroelectric Random Access Memory), a semiconductor integrated circuit device having a logic circuit such as a microprocessor, or a hybrid semiconductor integrated circuit device having a memory circuit and a logic circuit on the common semiconductor substrate. The invention is applicable to a fabrication method for a semiconductor integrated circuit device, semiconductor device, electronic circuit device, electronic device or micro-machine having at least a microscopic copper interconnect structure.
[0385] The effects obtained by the typical ones of the inventions disclosed in the present Application, if briefly explained, include the followings.
[0386] (1) The present invention has a step of carrying out a plasma process on a semiconductor substrate in a hydrogen gas atmosphere and a step of carrying out a plasma process on the semiconductor substrate in an ammonia gas atmosphere, after forming, by polishing technique, an interconnect including copper as a main conductor layer in an interconnect opening formed in a first insulation film on the semiconductor substrate but before depositing a second insulation film on the first insulation film and interconnect. This makes it possible to improve the dielectric breakdown strength between the interconnects having copper in their main conductor layers thereof.
[0387] (2) The present invention has a step of carrying out a plasma process on a semiconductor substrate in a hydrogen gas atmosphere and a step of carrying out a plasma process on the semiconductor substrate in an ammonia gas atmosphere, after forming, by polishing technique, an interconnect including copper as a main conductor layer in an interconnect opening formed in a first insulation film on the semiconductor substrate but before depositing a second insulation film on the first insulation film and interconnect. This makes it possible to prevent or suppress silicide or copper oxide from being formed on the interconnect. Accordingly, it is possible to prevent against increase of the resistance of the interconnect having copper in a main conductive layer thereof.
[0388] (3) The present invention has a step of carrying out a plasma process on a semiconductor substrate in a hydrogen gas atmosphere and a step of carrying out a plasma process on the semiconductor substrate in an ammonia gas atmosphere, after forming, by polishing technique, an interconnect including copper as a main conductor layer in an interconnect opening formed in a first insulation film on the semiconductor substrate but before depositing a second insulation film on the first insulation film and interconnect. This makes it possible to improve the adhesion between the interconnect having copper in a main conductor layer thereof and the cap film.
[0389] (4) The present invention has a step of carrying out on a semiconductor substrate a cleaning process using an alkali solution, a deoxidizing process and a cleaning process using the acid solution in the order and a step of depositing a second insulation film on the first insulation film and interconnect, after forming, by polishing technique, an interconnect including copper as a main conductor layer in an interconnect opening formed in a first insulation film on the semiconductor substrate. This makes it possible to deoxidize, before acid cleaning, the copper oxide formed in an interconnect-exposed region during a polish process or the like. Accordingly, it is possible to prevent increase of the resistance of an interconnect having copper in a main conductor layer thereof.
[0390]
FIG. 1
[0391] 1. M2 LAYER
[0392] 2. TH1 LAYER
[0393]
FIG. 2
[0394] 3. SAMPLE
[0395] 4. I/V MEASURING INSTRUMENT
[0396]
FIG. 3
[0397] 1. CURRENT DENSITY (A/cm2)
[0398] 2. STRESS TIME (SEC.)
[0399] 3. TESTING TEMPERATURE 140° C.
[0400] 4. ELECTRIC FIELD INTENSITY 5 MV/cm
[0401] 5. TDDB LIFE
[0402]
FIG. 7
[0403] 1. TDDB LIFE (SEC.)
[0404] 2. ELECTRIC FIELD INTENSITY (MV/cm)
[0405] 3. 10 YRS.
[0406] 4. Cu INTERCONNECT Al INTERCONNECT w INTERCONNECT
[0407] 5. TESTING TEMPERATURE 140° C.
[0408]
FIG. 8
[0409] 1. TDDB LIFE (SEC.)
[0410] 2. ELECTRIC FIELD INTENSITY (MV/cm)
[0411] 3. NH3 PLASMA EFFECT
[0412] 4. TDDB LIFE IMPROVED BY 6 ORDERS
[0413] 5. NO NH3 PLASMA
[0414] 6. NH3 PLASMA PRESENT
[0415] 7. TESTING TEMPERATURE 140° C.
[0416]
FIG. 9
[0417] 8. TDDB LIFE (SEC.)
[0418] 9. ELECTRIC FIELD INTENSITY (MV/cm)
[0419] 10. TESTING TEMPERATURE 140° C.
[0420] 11. NH3 PLASMA EFFECT
[0421] 12. TDDB LIFE IMPROVED BY 6-8 ORDERS
[0422] 13. TO P-TEOS LAYER/NO NH3 PLASMA
[0423] 14. TO P-SiN LAYER/NO NH3 PLASMA
[0424] 15. TO P-TEOS LAYER/NH3 PLASMA PRESENT
[0425] 16. TO P-SiN LAYER/NH3 PLASMA PRESENT
[0426]
FIG. 10
[0427] 1. IMMEDIATELY AFTER Cu/AS DEPO
[0428] 2. NORMALIZED STRENGTH (a.u.)
[0429] 3. IMMEDIATELY AFTER Cu/AS DEPO
[0430] 4. Cu/CMP ONLY
[0431] 5. Cu/CMP ONLY
[0432]
FIG. 11
[0433] 1. IMMEDIATELY AFTER Cu/CMP+POST CLEANING
[0434] 2. IMMEDIATELY AFTER Cu/CMP+POST CLEANING NORMALIZED STRENGTH (a.u.)
[0435] 3. 24H ELAPSED AFTER Cu/POST CLEANING
[0436] 4. 24H ELAPSED AFTER Cu/POST CLEANING
[0437]
FIG. 12
[0438] 1. ALLOWED TO STAND AFTER Cu/POST CLEANING+NH3 PLASMA PROCESS
[0439] 2. ALLOWED TO STAND AFTER Cu/POST CLEANING+NH3 PLASMA PROCESS
[0440] 3. NORMALIZED STRENGTH (a.u.)
[0441]
FIG. 13
[0442] 1. CHARGE UP
[0443] 2. CONDITION
[0444]
FIG. 14
[0445] 1. THERMAL DESORPTION OF HYDROGEN
[0446] 2. ORGANIC MATTER
[0447] 4. THERMAL DESORPTION OF ORGANIC MATTER
[0448]
FIG. 15
[0449] 1. THERMAL DESORPTION OF Si & ORGANIC MATTER
[0450] 2. ORGANIC MATTER
[0451] 1. THERMAL DESORPTION OF SiO & ORGANIC MATTER
[0452] 4. THERMAL DESORPTION OF SiH & ORGANIC MATTER
[0453] 5. THERMAL DESORPTION OF SiH3 & ORGANIC MATTER
[0454]
FIG. 16
[0455] 1. Cu INTERCONNECT RESISTANCE
[0456] 2. SURFACE-PROCESS NAME
[0457] 3. NO PROCESSING
[0458]
FIG. 17
[0459] (a) NO PROCESSING
[0460] (b) NH3 PLASMA
[0461]
FIG. 18
[0462] (a) H2 ANNEAL
[0463] (b) N2 PLASMA
[0464] (c) H2 PLASMA
[0465]
FIG. 19
[0466] 1. Cu COMPUND
[0467] 2. INTERFACIAL LEVEL
[0468] 3. CMP-DAMAGED LAYER (LEVEL, INCLUDING REMAINING Cu)
[0469] 4. DRIFT CURRENT INCREASE
[0470] 5. DIELECTRIC BREAKDOWN
[0471] 6. ELECTRIC FIELD E
[0472] 7. DRIFT CURRENT INCREASE
[0473] 8. LEVEL INCREASE
[0474] 9. TDDB BREAKDOWN
[0475] 10. INTERLAYER FILM/CAP FILM
[0476]
FIG. 20
[0477] 1. NH3 PLASMA PROCESS→Cu2+ GENERATION PREVENTION BY DEOXIDATION & BARRIER LAYER
[0478] 2. NH3, H2 PLASMA PROCESS→INTERFACIAL LEVEL DECREASE
[0479] 3. GRIT-FREE CMP, HYDROGEN FLUORIDE CLEANING→LOW CMP-DAMAGED LAYER
[0480] 4. DRIFT CURRENT DECREASE
[0481] 5. TDDB LIFE IMPROVEMENT
[0482] 6. ELECTRIC FIELD E
[0483] 7. Cu2+ GENERATION PREVENTION
[0484] 8. LEVEL DECREASE
[0485] 9. DRIFT CURRENT DECREASE
[0486] 10. TDDB LIFE IMPROVEMENT
[0487] 11. INTERLAYER FILM/CAP FILM
[0488]
FIG. 21
[0489] 1. TDDB LIFE (SEC.)
[0490] 2. ELECTRIC FIELD INTENSITY (MV/cm)
[0491] 3. TESTING TEMPERATURE 140° C.
[0492] 4. H2-Pl. EFFECT
[0493] 5. HF/POST-CMP CLEANING P-SiN FILM FORMING 400° C.
[0494] 6. NO PROCESSING
[0495] 7 (360 C, NO HF)
[0496]
FIG. 22
[0497]
100
INSULATION FILM DEPOSITION
[0498]
101
INTERCONNECT OPENING FORMING
[0499]
102
CONDUCTIVE BARRIER FILM DEPOSITION
[0500]
103
MAIN CONDUCTOR FILM DEPOSITION
[0501]
105
ALKALI CLEANING
[0502]
106
DEOXIDIZING PROCESS
[0503]
107
ACID CLEANING
[0504]
108
DRYING PROCESS
[0505]
109
H2 PLASMA
[0506]
110
NH3 PLASMA
[0507]
111
CAP FILM FORMING
[0508] 1. INTERCONNECT-FORMING CONDUCTOR FILM DEPOSITION
[0509] 2. POST CLEANING PROCESS
[0510] 3 POST PROCESS
[0511] 4 NOT RELEASED TO THE AIR
[0512]
FIG. 27
[0513] 1. H2 ANNEAL
[0514]
FIG. 29
[0515] 1. TDDB LIFE (SEC.)
[0516] 2 ELECTRIC FIELD INTENSITY (MV/cm)
[0517] 3. (LINE & SPACE) 4 10 YRS.
[0518] 5. DAE+HF CLEANING
[0519] 6. DAE→H2 ANNEAL→HF CLEANING
[0520] 7. APPROX. 2 ORDER UP
[0521]
FIG. 30
[0522] 1. CUMULATIVE RATE (%)
[0523] 2. INTERCONNECT RESISTANCE (MΩ/)
[0524] 3. DAE+HF CLEANING
[0525] 4. DAE→H2 ANNEAL→HF CLEANING
[0526]
FIG. 31
[0527] 1. HYDROGEN PLASMA
[0528]
FIG. 32 AMMONIA PLASMA
[0529]
FIG. 33
[0530]
1
: SEMICONDUCTOR SUBSTRATE
[0531]
16
c
,16d: INSULATION FILM
[0532]
21
a
: CONDUCTIVE BARRIER FILM
[0533]
22
a
: MAIN CONDUCTOR FILM
[0534]
23
a
: BURIED INTERCONNECT
[0535]
FIG. 34
[0536] 1. TDDB LIFE (SEC.)
[0537] 2. ELECTRIC FIELD INTENSITY (MV/cm)
[0538] 3. 10 YRS.
[0539] 4. SILK STRUCTURE ESTIMATION CHARACTERISTIC
[0540] 5. SILK DIELECTRIC STRENGTH 3.0-3.5V
[0541] 6. NH3 PLASMA
[0542]
7
. 2 ORDER UP
[0543] 8. NH3 PLASMA
[0544] 9. H2 PLASMA+NH3 PLASMA
[0545] 10. H2 PLASMA +NH3 PLASMA
[0546] 11. pTEOS/pSiN STRUCTURE
[0547] 12. H2 PLASMA
[0548]
FIG. 36
[0549] 1. ANTI-CORROSION PROCESSING SECTION
[0550] 2. POLISH PROCESSING SECTION 2 (BARRIER FILM POLISH)
[0551] 3. POLISH PROCESSING SECTION 1 (Cu FILM POLISH)
[0552] 4. LOADER
[0553] 5. UNLOADER
[0554] 6. POST-CLEAN PROCESSING SECTION 2 (HEAVY METAL REMOVAL & RELIABILITY-IMPROVE PROCESSING SECTION)
[0555] 7. DEOXIDIZATION PROCESSING SECTION
[0556] 8. POST-CLEAN PROCESSING SECTION 1 (SLURRY REMOVAL PROCESSING SECTION)
[0557] 9. IMMERSION PROCESSING SECTION
[0558] 10. POST-CLEAN PROCESSING SECTION
[0559]
FIG. 43
[0560]
100
INSULATION FILM DEPOSITION
[0561]
101
INTERCONNECT OPENING FORMING
[0562]
102
CONDUCTIVE BARRIER FILM DEPOSITION
[0563]
103
MAIN CONDUCTOR FILM DEPOSITION
[0564]
106
DEOXIDIZING PROCESS
[0565]
105
ALKALI CLEANING
[0566]
107
ACID CLEANING
[0567]
108
DRYING PROCESS
[0568]
109
H2 PLASMA
[0569]
110
NH3 PLASMA
[0570]
111
CAP FILM FORMING
[0571] 1. INTERCONNECT-FORMING CONDUCTOR FILM DEPOSITION
[0572] 2. POST-CLEANING PROCESS
[0573] 3 POST PROCESS
[0574] 4 NOT RELEASED TO THE AIR
[0575]
FIG. 44
[0576]
100
INSULATION FILM DEPOSITION
[0577]
101
INTERCONNECT OPENING FORMING
[0578]
102
CONDUCTIVE BARRIER FILM DEPOSITION
[0579]
103
MAIN CONDUCTOR FILM DEPOSITION
[0580]
106
DEOXIDIZING PROCESS
[0581]
107
ACID CLEANING
[0582]
108
DRYING PROCESS
[0583]
109
H2 PLASMA
[0584]
110
NH3 PLASMA
[0585]
111
CAP FILM FORMING
[0586] 1. INTERCONNECT-FORMING CONDUCTOR FILM DEPOSITION
[0587] 2. POST-CLEANING PROCESS
[0588] 3 POST PROCESS
[0589] 4 NOT RELEASED TO THE AIR
[0590]
FIG. 45
[0591]
108
DRYING PROCESS
[0592]
110
NH3 PLASMA
[0593]
109
H2 PLASMA
[0594]
111
CAP FILM FORMING
[0595] 2. NOT RELEASED TO THE AIR
[0596]
FIG. 46
[0597]
108
DRYING PROCESS
[0598]
110
NH3 PLASMA
[0599]
111
CAP FILM FORMING
[0600] 3. POST PROCESS
[0601] 4. NOT RELEASED TO THE AIR
[0602]
FIG. 47
[0603]
108
DRYING PROCESS
[0604]
200
H2 ANNEAL
[0605]
109
H2 PLASMA
[0606]
110
NH3 PLASMA
[0607]
111
CAP FILM FORMING
[0608] 1. POST PROCESS
[0609] 2. NOT RELEASED TO THE AIR
[0610]
FIG. 48
[0611]
108
DRYING PROCESS
[0612]
200
H2 ANNEAL
[0613]
110
NH3 PLASMA
[0614]
109
H2 PLASMA
[0615]
111
CAP FILM FORMING
[0616] 3. POST PROCESS
[0617] 4. NOT RELEASED TO THE AIR
[0618]
FIG. 49
[0619]
108
DRYING PROCESS
[0620]
200
H2 ANNEAL
[0621]
110
NH3 PLASMA
[0622]
111
CAP FILM FORMING
[0623] 1. POST PROCESS
[0624] 2. NOT RELEASED TO THE AIR
[0625]
FIG. 50
[0626]
105
ALKALI CLEANING
[0627]
107
ACID CLEANING
[0628]
108
DRYING PROCESS
[0629] 3. POST-CLEANING PROCESS
[0630]
FIG. 51
[0631]
107
ACID CLEANING
[0632]
108
DRYING PROCESS
[0633] 1. POST-CLEANING PROCESS
[0634]
FIG. 61
[0635]
103
MAIN CONDUCTOR FILM DEPOSITION
[0636]
104
[0637] GRIT-FREE CHEMICAL MECHANICAL POLISH
[0638] GRIT-USED CHEMICAL MECHANICAL POLISH
[0639] SELECTIVE CHEMICAL MECHANICAL POLISH
[0640] POST-CLEANING PROCESS
[0641]
FIG. 62
[0642] 1. TDDB LIFE (SEC.)
[0643] 2. ELECTRIC FIELD INTENSITY (MV/cm)
[0644] 3. TESTING TEMPERATURE 140° C
[0645] 4. GRIT-FREE EFFECT
[0646] 5. GRIT-FREE+NH3-P1 EFFECT
[0647] 6. NO QCTT→NH3-Pl.
[0648] 7. QCTT→NH3-Pl. PRESENT
[0649]
[0650]
8
. NO HS4000→NH3-Pl.
[0651] 9. HS4000→NH3-Pl. PRESENT
[0652] 10. HS4000→NH3-Pl. PRESENT (HF CLEANING)
[0653]
FIG. 63
[0654] 1. TDDB LIFE (SEC.)
[0655] 2. ELECTRIC FIELD INTENSITY (MV/cm)
[0656] 3. TESTING TEMPERATURE 140° C.
[0657] 4. HF CLEANING EFFECT
[0658] 5. NH3-Pl. EFFECT
[0659]
6
. CITRIC ACID CLEANING+NH3-Pl. EFFECT
[0660]
7
. HF CLEANING+NH3-Pl. EFFECT
[0661]
8
. NO NH4OH CLEANING→NH3-Pl.
[0662]
9
. NO HF CLEANING→NH3-Pl.
[0663]
10
. NH4OH CLEANING→NH3-Pl. PRESENT
[0664]
11
. CITRIC ACID CLEANING→NH3-Pl. PRESENT
[0665]
12
. HF CLEANING→NH3-Pl. PRESENT
[0666]
FIG. 67
[0667] 1. Si CONTENT (ATOMIC%)
[0668] 2. SURFACE PROCESS NAME
[0669] 3. DEPTH FROM SURFACE: d
[0670] 4. EDX ANALYSIS POINT
[0671] 5. Cu INTERCONNECT SECTIONAL VIEW
[0672] 6. IN OR AROUND SURFACE
[0673] 7. IN Cu FILM
[0674] 8. NO PROCESSING
[0675]
FIG. 68
[0676] TiNx nm/TiN 50 nm<ref>RATIO (%)
[0677] 1. TiN FILM THICKNESS
[0678] 2. TiN FILM THICKNESS=TRENCH-BOTTOM FILM THICKNESS
[0679] 3. (MEASUREMENT VALUE)
[0680] 4. (CALCULATED VALUE)
[0681]
FIG. 69
[0682] 1. TDDB LIFE (SEC.)
[0683] 2. ELECTRIC FIELD INTENSITY (MV/cm)
[0684] 3. TESTING TEMPERATURE 140° C.
[0685]
FIG. 70
[0686] 1. TDDB LIFE (SEC.)
[0687] 2. ELECTRIC FIELD INTENSITY (MV/cm)
[0688] 3. TESTING TEMPERATURE 140° C.
Claims
- 1. A fabrication method of a semiconductor integrated circuit device, comprising the steps of:
(a) a step of forming a first insulation film on a main surface of a semiconductor substrate; (b) a step of forming an interconnect opening in said first insulation film; (c) a step of depositing a first conductor film for interconnect formation and a second conductor film based on copper for interconnect formation in the order on said first insulation film including an inside of said interconnect opening; (d) a polishing process of polishing said first and second conductor films for interconnect formation thereby forming an interconnect in said interconnect opening; (e) a process of carrying out a cleaning process on the semiconductor substrate of after said polishing process; (f) a process of carrying out a plasma process in a hydrogen atmosphere on the substrate of after said cleaning process; (g) a process of carrying out a plasma process in an ammonia atmosphere on the substrate of after said cleaning process; and (h) a process of depositing a second insulation film on said first insulation film and interconnect after said hydrogen gas plasma and ammonia gas plasma processes.
- 2. A fabrication method of a semiconductor integrated circuit device according to claim 1, wherein said cleaning process has a cleaning process using an alkali solution and a cleaning process using an acid solution.
- 3. A fabrication method of a semiconductor integrated circuit device according to claim 2, wherein said alkali solution includes aminoethanol.
- 4. A fabrication method of a semiconductor integrated circuit device according to claim 2, wherein said cleaning process has a step to carry out a cleaning process using said alkali solution, a deoxidizing process and a cleaning process using said acid solution in the order.
- 5. A fabrication method of a semiconductor integrated circuit device according to claim 2, having a step to carry out a deoxidizing process after said polishing process before said cleaning process.
- 6. A fabrication method of a semiconductor integrated circuit device according to claim 2, having a step to carry out a deoxidizing process after said polishing process before a cleaning process using said acid solution.
- 7. A fabrication method of a semiconductor integrated circuit device according to claim 6, wherein said deoxidizing process is to carry out a heating process on the semiconductor substrate in a hydrogen gas atmosphere.
- 8. A fabrication method of a semiconductor integrated circuit device according to claim 6, wherein said deoxidizing process is to carry out a hydrogen gas plasma process on the semiconductor substrate.
- 9. A fabrication method of a semiconductor integrated circuit device according to claim 6, wherein said dexidizing process is to carry out an ammonia gas plasma process on the semiconductor substrate.
- 10. A fabrication method of a semiconductor integrated circuit device according to claim 1, wherein said cleaning process has a step to carry out a cleaning process using an acid solution.
- 11. A fabrication method of a semiconductor integrated circuit device according to claim 10, having a step to carry out a deoxidizing process after said polishing step before said cleaning process using an acid solution.
- 12. A fabrication method of a semiconductor integrated circuit device according to claim 1, wherein a heating process is carried out on the semiconductor substrate in a deoxidizing gas atmosphere after said cleaning process and before said hydrogen gas plasma and ammonia gas plasma processes.
- 13. A fabrication method of a semiconductor integrated circuit device according to claim 1, wherein said second conductor film based on copper for interconnect formation is deposited by a plating technique.
- 14. A fabrication method of a semiconductor integrated circuit device according to claim 1, wherein said first insulation film is formed by an insulation film having a low dielectric constant.
- 15. A fabrication method of a semiconductor integrated circuit device according to claim 1, wherein said second insulation film is formed on said first insulation film and interconnect continuously without releasing to the air after said hydrogen gas plasma and ammonia gas plasma processes.
- 16. A fabrication method of a semiconductor integrated circuit device according to claim 1, wherein said second insulation film is formed by a silicon nitride film or silicon carbide film.
- 17. A fabrication method of a semiconductor integrated circuit device according to claim 1, wherein a process of said polishing step has a grit-free chemical mechanical polish.
- 18. A fabrication method of a semiconductor integrated circuit device, comprising the steps of:
(a) a step of forming a first insulation film on a main surface of a semiconductor substrate; (b) a step of forming an interconnect opening in said first insulation film; (c) a step of depositing a conductor film based on copper for interconnect formation on said first insulation film including an inside of said interconnect opening; (d) a polishing process of polishing said conductor film based on copper for interconnect formation thereby forming an interconnect in said interconnect opening; (e) a process of carrying out a cleaning process on the semiconductor substrate of after said polishing process; (f) a process of carrying out a plasma process in a hydrogen atmosphere on the substrate of after said cleaning process; (g) a process of carrying out a plasma process in an ammonia atmosphere on the substrate of after said cleaning process; and (h) a process of depositing a second insulation film on said first insulation film and interconnect after said hydrogen gas plasma and ammonia gas plasma processes.
- 19. A fabrication method of a semiconductor integrated circuit device, comprising the steps of:
(a) a step of forming a first insulation film on a main surface of a semiconductor substrate; (b) a step of forming an interconnect opening in said first insulation film; (c) a step of depositing a first conductor film for interconnect formation and a second conductor film based on copper for interconnect formation in the order on said first insulation film including an inside of said interconnect opening; (d) a polishing process of polishing said first and second conductor films for interconnect formation thereby forming an interconnect in said interconnect opening; (e) a process of carrying out a cleaning process using an alkali solution, a deoxidizing process and a cleaning process using said acid solution in the order on the semiconductor substrate of after said polishing process; and (f) a process of depositing a second insulation film on said first insulation film.
- 20. A fabrication method of a semiconductor integrated circuit device according to claim 19, wherein said alkali solution includes aminoethanol.
- 21. A fabrication method of a semiconductor integrated circuit device according to claim 19, wherein said deoxidizing process is to carry out a heating process on the semiconductor substrate in a hydrogen gas atmosphere.
- 22. A fabrication method of a semiconductor integrated circuit device according to claim 19, wherein said deoxidizing process is to carry out a hydrogen gas plasma process on the semiconductor substrate.
- 23. A fabrication method of a semiconductor integrated circuit device according to claim 19, wherein said deoxidizing process is to carry out an ammonia gas plasma process on the semiconductor substrate.
- 24. A fabrication method of a semiconductor integrated circuit device according to claim 19, wherein said second conductor film based on copper for interconnect formation is deposited by a plating technique.
- 25. A fabrication method of a semiconductor integrated circuit device according to claim 19, wherein said first insulation film is formed by an insulation film having a low dielectric constant.
- 26. A fabrication method of a semiconductor integrated circuit device according to claim 19, having a step to carry out an ammonia gas plasma process on the semiconductor substrate after said step (e) before the step of forming said second insulation film.
- 27. A fabrication method of a semiconductor integrated circuit device according to claim 19, wherein said second insulation film is formed by an SiN film, an SiC film or a SiOc film.
- 28. A fabrication method of a semiconductor integrated circuit device according to claim 19, wherein a process of said polishing step has a grit-free chemical mechanical polish.
- 29. A fabrication method of a semiconductor integrated circuit device, comprising the steps of:
(a) a step of forming a first insulation film on a main surface of a semiconductor substrate; (b) a step of forming an interconnect opening in said first insulation film; (c) a step of depositing a first conductor film for interconnect formation and a second conductor film based on copper for interconnect formation in the order on said first insulation film including an inside of said interconnect opening; (d) a polishing process of polishing said first and second conductor films for interconnect formation thereby forming an interconnect in said interconnect opening; (e) a process of carrying out a deoxidizing process and a process of using said acid solution in the order on the semiconductor substrate of after said polishing process; and (f) a process of depositing a second insulation film on said first insulation film and interconnect.
- 30. A fabrication method of a semiconductor integrated circuit device according to claim 29, wherein said deoxidizing process is to carry out a heating process on the semiconductor substrate in a hydrogen gas atmosphere.
- 31. A fabrication method of a semiconductor integrated circuit device according to claim 29, wherein said deoxidizing process is to carry out a hydrogen gas plasma process on the semiconductor substrate.
- 32. A fabrication method of a semiconductor integrated circuit device according to claim 29, wherein said deoxidizing process is to carry out an ammonia gas plasma process on the semiconductor substrate.
- 33. A fabrication method of a semiconductor integrated circuit device according to claim 29, wherein said second conductor film based on copper for interconnect formation is deposited by a plating technique.
- 34. A fabrication method of a semiconductor integrated circuit device according to claim 29, wherein said first insulation film is formed by an insulation film having a low dielectric constant.
- 35. A fabrication method of a semiconductor integrated circuit device according to claim 29, having a step to carry out an ammonia gas plasma process on the semiconductor substrate after the step (e) before a step of forming said second insulation film.
- 36. A fabrication method of a semiconductor integrated circuit device according to claim 29, wherein said second insulation film is formed by an SiN film, an SiC film or a SiOc film.
- 37. A fabrication method of a semiconductor integrated circuit device according to claim 29, wherein a process of said polishing step has a grit-free chemical mechanical polish.
- 38. A fabrication method of a semiconductor integrated circuit device, comprising the steps of:
(a) a step of forming a first insulation film on a main surface of a semiconductor substrate; (b) a step of forming an interconnect opening in said first insulation film; (c) a step of depositing a conductor film based on copper for interconnect formation on said first insulation film including an inside of said interconnect opening; (d) a process of polishing said conductor film based on copper for interconnect formation thereby forming an interconnect in said interconnect opening; (e) a process of carrying out a cleaning process using an alkali solution, a deoxidizing process and a cleaning process using said acid solution in the order on the semiconductor substrate of after said polishing process; and (f) a process of depositing a second insulation film on said first insulation film.
- 38. A fabrication method of a semiconductor integrated circuit device, comprising the steps of:
(a) a step of forming a first insulation film on a main surface of a semiconductor substrate; (b) a step of forming an interconnect opening in said first insulation film; (c) a step of depositing a conductor film based on copper for interconnect formation on said first insulation film including an inside of said interconnect opening; (d) a process of polishing said conductor film based on copper for interconnect formation thereby forming an interconnect in said interconnect opening; (e) a process of carrying out a deoxidizing process and a cleaning process using said acid solution on the semiconductor substrate of after said polishing process; and (f) a process of depositing a second insulation film on said first insulation film.
Priority Claims (1)
Number |
Date |
Country |
Kind |
2000-300853 |
Sep 2000 |
JP |
|