INTEGRATED CIRCUIT WITH MULTI-THRESHOLD BULK FINFETS

Abstract

A method for manufacturing a FinFET having a fm that has a fm body includes selecting a desired electrical performance parameter, selecting a base dimension of the fin, identifying a combination of fin-body doping and fin-geometry that causes the FinFET to have the desired electrical performance parameter, doping the fin body according to the identified fin-body doping, and fabricating the fm according to the fin-geometry.

Claims

1-41. (canceled)

42. A manufacture comprising a plurality of bulk FinFETs integrated into an integrated circuit, at least two of said FinFETs having different fin-body doping levels.

43. The manufacture of claim 42, further comprising a first fin having a first fin-body doping level and a first fin-geometry, and a second fin having a second fin-body doping level and a second fin-geometry, said first and second fins being constituents of corresponding first and second bulk FinFETs from said plurality of FinFETs, wherein said first fin-body doping level and said first fin-geometry are selected to achieve a first electrical-performance parameter for said first FinFET, and wherein said second fin-body doping level and said second fin-geometry are selected to achieve a second electrical-performance parameter for said second FinFET, said second electrical-performance parameter differing from said first electrical-performance parameter.

44. The manufacture of claim 43, wherein said first and second fin-geometries are characterized by sidewall angles that differ by a value that is within a range between twenty degrees and an integer multiple of ten degrees.

45. The manufacture of claim 43, wherein said first and second fin-geometries are characterized by different sidewall angles such that a difference between said different sidewall angles is within a sidewall-angle difference range having a lower bound and an upper bound, wherein said lower bound is defined by a product of a differential and a first integer, wherein said upper bound is defined by a product of a differential and a second integer, wherein said second integer is said first integer incremented by one, wherein said differential is ten degrees and said first integer is zero.

46. The manufacture of claim 43, wherein said first and second fin-geometries are characterized by different sidewall angles.

47. The manufacture of claim 43, wherein said first electrical-performance parameter is a parameter that is derived from a combination of leakage current and drive current.

48. The manufacture of claim 43, wherein said first fin has a cross section having a first side, a second side, and a third side, wherein said first and third sides slope towards each other.

49. The manufacture of claim 43, wherein said first and second fin-body doping levels differ from each other.

50. The manufacture of claim 43, wherein said first fin has a base dimension that is equal to a base dimension of said second fin.

51. The manufacture of claim 43, wherein said first electrical-performance parameter is drive current.

52. The manufacture of claim 43, wherein said first fin has a triangular cross-section.

53. The manufacture of claim 43, wherein said first fin has a cross-section having a first side and a second side, both of which have a first length, and a third side having a second length, wherein a first rounded vertex connects said third side to said first side, and wherein a second rounded vertex connects said second side to said third side.

54. The manufacture of claim 43, wherein said first fin has a transverse cross-section having first, second, and third sides, said first and second sides being joined at a vertex and said second and third sides being joined at a vertex.

55. The manufacture of claim 43, wherein said first electrical-performance parameter is indicative of a rate of charge leakage.

56. The manufacture of claim 43, wherein said first fin has a fin axis that extends along said integrated circuit, and wherein, when cut by a plane transverse to said fin axis, said first fin displays a rectangular cross-section.

57. The manufacture of claim 43, wherein a difference between sidewall angles in said first and second fin geometries is within a range having a lower bound and an upper bound, wherein said lower bound is defined by a product of a differential and a first integer, wherein said upper bound is defined by a product of said differential and a second integer, wherein said second integer is said first integer incremented by one, wherein said differential is ten degrees and said first integer is one.

58. The manufacture of claim 43, wherein said first fin has a transverse cross-section having a first side and a second side, both of which have a first length, and a third side, which have a second length, wherein a first rounded vertex connects said third side to said first side, and wherein a second rounded vertex connects said second side to said third side.

59. The manufacture of claim 43, wherein said first fin-geometry differs from said second fin-geometry.

60. The manufacture of claim 43, wherein said first fin has a cross-section that has two sides connected by a rounded vertex.

61. The manufacture of claim 43, wherein two sides connected by a vertex are constituents of a transverse cross-section of said first fin.

62. The manufacture of claim 43, wherein said first fin has a trapezoidal cross-section.

63. A method comprising manufacturing an integrated circuit having a plurality of bulk FinFETs, at least two of said bulk FinFETs having different fin parameters, wherein said fin parameters comprise a fin-body doping level and a fin geometry, wherein manufacturing said integrated circuit comprises selecting a desired electrical performance parameter for a bulk FinFET having a fin, identifying a combination of fin parameters that causes said FinFET to have said desired electrical performance parameter, doping said fin body according to said identified fin-body doping, and fabricating said fin according to said fin-geometry.

64. The method of claim 63, wherein identifying said combination of fin parameters comprises identifying a fin-geometry having a rectangular cross-section.

65. The method of claim 63, wherein identifying said combination of fin parameters comprises identifying a fin geometry in which a rounded vertex joins first and second sides that slope away from said vertex in opposite directions.

66. The method of claim 63, wherein identifying a combination of fin parameters comprises identifying a fin-geometry having a selected sidewall angle, wherein identifying a fin-geometry having a selected sidewall angle comprises selecting said selected sidewall angle to be in a range having a lower bound defined by a product of a range breadth and a first integer and an upper bound defined by a product of said range breadth and a second integer, wherein said second integer is equal to said first integer incremented by one, wherein said range breadth is five degrees, and wherein said first integer is three.

67. The method of claim 63, wherein identifying said combination of fin parameters comprises identifying a fin geometry having a first side, a second side, and a third side, wherein said first side and said second side meet at a right angle, and wherein said second side and third side meet at a right angle.

68. The method of claim 63, wherein identifying said combination of fin parameters comprises concurrently identifying a fin-geometry and a fin-body doping level.

69. The method of claim 63, wherein identifying said fin-geometry comprises identifying a fin-geometry having first and second sides that slope toward each other, and a third side having a first end that meets an end of said first side and a second end that meets an end of said second side.

70. The method of claim 63, wherein identifying said combination of fin parameters comprises identifying a fin-body doping level, and, after having identified said fin-body doping level, identifying a fin-geometry.

71. The method of claim 63, wherein identifying a combination of fin parameters comprises identifying a fin-geometry having a selected sidewall angle.

72. The method of claim 63, wherein identifying said combination of fin parameters comprises identifying a fin geometry having first and second sides that slope towards each other and meet at a vertex.

73. The method of claim 63, wherein identifying said fin-geometry comprises identifying a fin-geometry having a trapezoidal cross-section.

74. The method of claim 63, further comprising selecting said electrical parameter to be drive current.

75. The method of claim 63, wherein identifying a fin-geometry having a rectangular cross-section comprises identifying a fin-geometry comprising a first side and a second side, both of which have a first length, and a third side having a second length, wherein a first rounded vertex connects said third side to said first side, wherein a second rounded vertex connects said second side to said third side.

76. The method of claim 63, further comprising selecting said electrical parameter to be derived from a combination of leakage current and drive current.

77. The method of claim 63, further comprising selecting said electrical parameter to be leakage current.

78. The method of claim 63, wherein identifying a combination of fin parameters comprises identifying a fin-geometry having a selected sidewall angle, wherein identifying a fin-geometry having a selected sidewall angle comprises selecting said selected sidewall angle to be in a range having a lower bound defined by a product of a range breadth and a first integer and an upper bound defined by a product of said range breadth and a second integer, wherein said second integer is equal to said first integer incremented by one, wherein said range breadth is five degrees, and wherein said first integer is two.

79. The method of claim 63, wherein identifying a combination of fin parameters comprises identifying a fin-geometry having a selected sidewall angle, wherein identifying a fin-geometry having a selected sidewall angle comprises selecting said selected sidewall angle to be in a range having a lower bound defined by a product of a range breadth and a first integer and an upper bound defined by a product of said range breadth and a second integer, wherein said second integer is equal to said first integer incremented by one, wherein said range breadth is five degrees, and wherein said first integer is one.

80. The method of claim 63, wherein identifying said combination of fin parameters comprises identifying a fin geometry having a triangular cross-section.

81. The method of claim 63, wherein identifying a combination of fin parameters comprises identifying a fin-geometry having a selected sidewall angle, wherein identifying a fin-geometry having a selected sidewall angle comprises selecting said selected sidewall angle to be in a range having a lower bound defined by a product of a range breadth and a first integer and an upper bound defined by a product of said range breadth and a second integer, wherein said second integer is equal to said first integer incremented by one, wherein said range breadth is five degrees, and wherein said first integer is four.

82. The method of claim 63, wherein identifying said fin-geometry comprises identifying a fin-geometry having a transverse cross-section that includes a first rounded vertex, a first side, a second rounded vertex, and a second side, said first and second sides having a first length, and a third side that has a second length, wherein said first rounded vertex connects said third side to said first side and wherein said second rounded vertex connects said second side to said third side.

83. The method of claim 63, wherein identifying said combination of fin parameters comprises identifying a fin-geometry before identifying a fin-body doping level.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0041] FIG. 1 is an isometric view of a typical FinFET;

[0042] FIG. 2 is a cross-section of the FinFET of FIG. 1;

[0043] FIG. 3 shows another cross-section of a FinFET;

[0044] FIG. 4 shows finFETs with rounded corners; and

[0045] FIG. 5 shows leakage current as a function of body doping concentration.

DETAILED DESCRIPTION

[0046] Referring to FIG. 1, a typical bulk FinFET 10 features a fin 12 that protrudes through an oxide layer 14 from a substrate 16. The fin 12 has a first end doped as a source 18 and a second end doped as a drain 20. Between the source and drain is a channel, which is not shown because the sides and top of the fin 12 are covered by a gate electrode 22.

[0047] As shown in the cross-section of FIG. 2, the fin 12 in FIG. 1 has an active fin portion 24 that protrudes above the substrate 16 and a fin body 26 that penetrates an the oxide layer 14, which thus functions as an isolation trench towards the substrate 16. As shown in FIG. 3, the active fin portion 24 is characterized by a height H, a base width, W.sub.bottom and a top width, W.sub.top. In many embodiments, the corners of the fin's cross-section are rounded, as shown in FIG. 4.

[0048] The fin 12 can have a variety of cross-sections, including triangular, and quadrilateral, the latter including square, rectangle, and trapezoid. For convenience, the fin will be referred to by its cross-section, with the understanding that the actual fin 12 is a solid prism formed by extending its cross-section along a line.

[0049] In one embodiment, the fin height is 35 nanometers, the base width is 15 nanometers, and the top width is also 15 nanometers. Such a fin would therefore be a rectangular fin. In the case of triangular or trapezoidal fins, the base width would be greater than the top width.

[0050] The fin can also be characterized by either one of two complementary angles made between the wall of the active fin portion 14 and the substrate 16. The complementary angle that is inside the fin 12 shall be referred to herein as the sidewall angle. A rectangular fin will have a ninety-degree sidewall angle. Triangular and trapezoidal fins will have a sidewall angle that is smaller than ninety degrees.

[0051] In general, the active fin portion 24 will be intrinsic to maximize carrier mobility, but the fin body 26 will be doped. The fin body 26 should be doped heavily enough to avoid leakage under the fin 12 that arises from the gate's inability to control short channel effects below the isolation trench 28, but not so heavily as to promote gate induced drain leakage resulting from band-to-band tunneling (BTBT). Thus, if one were to plot leakage current as a function of fin-body doping, the result would have a minimum, as shown in FIG. 5.

[0052] To cause the transistor to have a desired electrical parameter, it is useful to first optimize the doping of the fin body 26 and to then select an appropriate fin-geometry. Optimization of doping is generally carried out experimentally, the simplest way being by creating a computer model and sweeping across a range of doping levels. Optimization over fin-geometry can be also be carried out by similar computer modeling techniques. A suitable computer model is implemented by software sold under the name TCAD by Synopsys.

[0053] It is also possible to select a fin-geometry and to then choose a doping level that achieves a selected electrical performance parameter for the transistor.

[0054] Suitable electrical parameters whose values are selected can include leakage current, drive current, a ratio formed by the foregoing currents, a threshold, or functions thereof.

[0055] The methods described herein can be used to conveniently create multi-threshold integrated circuits. In such integrated circuits, different FinFETs have different thresholds. This makes it possible to have low leakage transistors for stand by functions and high drive current transistors for operational functions all integrated on the same chip. This advantage arises because all the FinFETs have the same base dimension, but different cross-sections and/or fin-body doping levels.

[0056] Contrary to the teachings of the prior art, which disclosed that fin-geometry does not affect leakage current, the invention described herein exploits the fact that if the fin base is suitably doped, it is in fact possible to manipulate geometry to control leakage current and other electrical performance characteristics of a FinFET.

[0057] The procedure described herein relies on the recognition that achieving a desired FinFET electrical performance involves identifying a point in a two dimensional optimization space that achieves a particular electrical performance parameter defined along an axis perpendicular to that space. The variables in the two dimensional optimization space would be fin-body doping level and some geometric parameter, a suitable one being the effective width W.sub.eff, which is the component of the fin cross section perimeter adjacent to the gate oxide calculated simply using the Pythagorean Theorem for the fin sides and the area of the semi-circle with corner radius set to W.sub.top/2 for the fin top:

[00001]$W_{\mathrm{eff}}=2.Math.\sqrt{{\left(\frac{W_{\mathrm{bottom}}-W_{\mathrm{top}}}{2}\right)}^2+{\left(H-\frac{W_{\mathrm{top}}}{2}\right)}^2}+.Math.\frac{W_{\mathrm{top}}}{2}$