WO2010115994A1

WO2010115994A1 - Determining active doping profiles in semiconductor structures

Info

Publication number: WO2010115994A1
Application number: PCT/EP2010/054736
Authority: WO
Inventors: Fabian Dortu
Original assignee: Imec; Multitel; Katholieke Universiteit Leuven K.U.Leuven R&D
Priority date: 2009-04-10
Filing date: 2010-04-10
Publication date: 2010-10-14

Abstract

An optical measurement method and system is disclosed to determine an active doping profile of a semiconductor substrate, the method comprising directing at least one pump laser pulse of a pump laser beam on the semiconductor substrate, with at least one pump laser pulse having a pump laser pulse power generating an amount of excess carrier concentration in the semiconductor substrate which is at least comparable to the peak concentration of the active doping profile; directing a probe laser beam on the semiconductor substrate such that the probe laser beam is at least partially reflected by the generated amount of excess carrier concentration; measuring a reflection signal of the probe laser beam, the reflection signal being induced by the generated amount of excess carrier concentration, as a function of a predetermined parameter and determining at least part of the active doping profile from the reflection signal as function of the predetermined parameter.

Description

DETERMINING ACTIVE DOPING PROFILES IN SEMICONDUCTOR

STRUCTURES

Field of the invention

The present invention relates to non-destructive optical measurement techniques, data processing techniques, apparatus and/or systems for determining the active doping profile in semiconductor layers. In particular it relates to using optical energy to create charge carriers in these semiconductor layers and to probe changes in reflectivity created by these charge carriers as function of the depth in the semiconductor layer where these carriers agitate and/or methods for data processing based thereon. More particularly, the present invention relates to methods, apparatus and/or systems for extracting the active doping profile in ultra shallow junctions in a semiconductor substrate.

Background of the invention

In semiconductor processing, methods are required for the determination of properties of semiconductor materials, such as Si, SiGe, GaAs, ..., and their dependence on processing conditions. Introducing species into a semiconductor material by, for example, ion implantation can change the properties of the bulk material. Other methods that can change the properties of the bulk material are manufacturing of the substrate, annealing such as for example rapid thermal processing (RTP) or rapid thermal annealing (RTA), etc. In CMOS (Complementary Metal Oxide Silicon) devices for example, it is important to be able to determine the junction depth and profile of the source and drain regions formed in the semiconductor substrate. For advanced high-performance CMOS technologies, it is, for example, crucial to be able to quickly and reliably characterize ultra shallow junctions (USJ) . Especially, as CMOS structures, such as for example transistors, become increasingly smaller, the doping profiles, in particular the carrier profiles shrink accordingly. Advanced CMOS structures will have gate lengths less than 50 nm and junction depths less than 70 nm. The exact determination of these profiles becomes more difficult and at the same time more critical. Also higher doping concentrations are used for such USJ profiles, more specifically, doping concentrations higher than le20 /cm Process conditions need to be optimized in order to obtain the desired junction depth and profile and, hence, to yield the required device characteristics. One of the many crucial issues in fabricating state-of-the-art CMOS transistors is the precise control over the positioning and electrical characteristics of source/drain and extension regions. Besides the currently used low energy ion implantation and fast annealing techniques, much effort is placed in new techniques such as laser annealing (LTA) and low temperature Solid Phase Epitaxial Regrowth (SPER) to achieve higher concentration levels (above solubility) and steeper profiles (smaller thermal budget) .

Various methods exist to investigate the properties of the semiconductor carrier profile. Some of these techniques, however, are destructive. Presently, for doping characterization people typically use a combination of one- dimensional techniques such as Secondary-Ion-Mass- Spectroscopy (SIMS) for the total profile, Spreading- Resistance-Profile (SRP) for the electrically carrier profile and Four-Point-Probe (FPP) measurement for sheet resistance. SIMS and SRP have the disadvantage that they are off-line techniques, applicable only on small pieces of material. In case of SRP the semiconductor substrate to be characterized is cleaved along a diagonal cleavage line and a two-point electrical measurement is then performed at subsequent positions along this cleavage line. For SIMS the material from the substrate under examination will be locally removed and subjected to further analysis. Furthermore a measurement on one specific position on a wafer takes about a day, taking into account the sawing, preparation of the sample, measurement, calculation, etc. Conventional FPP can quickly measure whole wafers, but does not give any profile information and still requires rather large analysis areas, typically larger than 1 mm². Furthermore, probe penetration leads to unreliable results on ultra-shallow profiles, particular when less than 30 nm deep. Recently some new promising techniques have emerged. Examples of such new techniques are two-dimensional carrier imaging techniques such as Scanning-Capacitance-Measurement (SCM) or Scanning-Spreading-Resistance-Microscopy (SSRM), but one still needs small pieces for the measurements, a complicated and critical sample preparation is required and the depth resolution still needs improvement (5-10 nm) . Furthermore these two-dimensional techniques depend critically on the availability of more reliable one- dimensional calibration profiles.

Other known techniques are non-destructive such as Photo Modulated Optical Reflectance (PMOR) techniques, for example, the Carrier Illumination (CI) technique, as disclosed in US 6,049,220 and US 6,323,951, and the Therma Probe (TP) technique, also called Thermawave technique or thermal wave technique as disclosed in "Non-destructive analysis of ultra shallow junctions using thermal wave technology" by Lena Nicolaides et al . in Review of Scientific Instruments, volume 74, number 1, January 2003. The PMOR technique, also often called Photo Modulated Thermo Reflectance (PMTR) , is a pump-probe technique using light intensity as modulation variable. A pump laser intensity is modulated at a given frequency, which modifies the sample reflectance through light-matter interaction processes, i.e through the generation of excess carriers in the substrate. A probe laser measures the change in reflectance at the same modulation frequency (or higher harmonics) through a lock-in based detection technique. The conversion of the (differential) PMOR reflectance signal to doping profile (i.e. carrier concentration and electrical junction depth) may for example be done through extensive correlation of PMOR measurements with SRP and/or SIMS measurements on a wide range of profiles (i.e. calibration curves) .

For in-line monitoring of the pre- and postanneal process steps, PMOR techniques have established themselves as a fast, noncontact, nondestructive tool with wafer mapping capability. For these process monitoring applications, the exact quantitative interpretation of the PMOR signal is less important as long as high repeatability and sensitivity for a particular profile or process parameter can be demonstrated.

The PMOR signal strongly depends on the carrier profile, the kind of dopant, the peak dopant concentration, ... When the carrier profile is unknown it is - using current state of the art methods - impossible to extract the electrical junction depth and doping concentration (with high precision) .

However, as stated in the International Technology Roadmap for Semiconductors (ITRS) 2008 update edition, electrical characterization of the source and drain (S/D) extension regions is a key challenge for future CMOS and non-CMOS devices. According to ITRS 2008 by 2015 source/drain extension junction depths should be less than 10 nm, peak doping concentrations higher than 4e20 /cm and lateral abruptness of the profile less than 1.5 nm/decade. According to this roadmap metrology methods should be able to measure the doping concentration with a precision of 2%. There is a need for non-destructive and non-contact optical methods for characterization of ultra-shallow junctions. There is thus a need for optical methods for qualitatively and quantitatively reconstructing one-dimensional in depth active doping profiles with high precision.

There is a need for methods to derive from the measured optical signals the full carrier concentration profile on any arbitrary carrier profile. There is a need for solving the inverse problem, i.e. reconstructing the active doping profile from measured power curves, without the need for using calibration curves.

Summary of the invention

It is an object of the present invention to provide good methods, systems and devices for determining an active doping profile (also referred to as (equilibrium) carrier profile) of a semiconductor substrate, even if the sample has a high concentration of active dopants. It is an advantage of embodiments according to the present invention that a good, non-destructive technique for determining an active doping profile of a semiconductor substrate is provided, even if the sample has a high concentration of active dopants.

According to a first inventive aspect an optical measurement method to determine an active doping profile of a semiconductor substrate is disclosed, the method comprising directing at least one pump laser pulse of a pump laser beam on the semiconductor substrate. The at least one pump laser pulse thereby has a pump laser pulse power generating an amount of excess carrier concentration in the semiconductor substrate which is comparable to or higher than the peak concentration of the active dopant profile. The method also includes directing a probe laser beam on the semiconductor substrate such that the probe laser beam is at least partially reflected by the generated amount of excess carrier concentration and measuring the reflection signal of the probe laser beam as a function of a predetermined parameter. The method furthermore comprises determining at least part of the active doping profile from the reflection signal as function of the predetermined parameter. The latter may be performed using a predetermined algorithm, as will be described further. According to embodiments of the present invention generating an amount of excess carrier concentration in the semiconductor substrate which is comparable to or higher than the peak concentration of the active dopant profile comprises generating an amount of excess carrier concentration in the semiconductor substrate inducing a layer-plasma reflectivity component coefficient which is at least comparable to a substrate-plasma reflectivity component coefficient. The layer-plasma reflectivity component coefficient defines the reflectivity induced by the generated amount of excess carriers in the doped layer whereas the substrate-plasma reflectivity component coefficient defines the inherent reflectivity induced in the semiconductor substrate.

According to embodiments of the present invention the generated amount of excess carrier concentration is comparable to or higher than a predetermined percentage of the peak carrier concentration of the carrier profile. According to embodiments of the present invention comparable to or higher than comprises equal to or higher than the peak concentration of the active doping profile divided by 10. According to embodiments of the present invention comparable to or higher than comprises equal to or higher than the peak concentration of the active doping profile .

According to embodiments of the present invention directing at least one pump laser pulse comprises directing a series of pump laser pulses. The pump laser power of each of the pump laser pulses of the series of pump laser pulses may be identical .

The peak carrier concentration of the carrier profile may be at least as high as the values from the ITRS roadmap for the used impurity species (e.g. B, As, P, ..) and semiconductor substrate (e.g. Si, Ge, SiGe, GaAs, InGaAs, ...) . According to ITRS Update 2008 the peak concentration of the carrier profile may be at least 4e20 /c^m3. According to embodiments of the present invention the at least one pump laser pulse, and thus the generated amount of excess carrier concentration is chosen such that the increase in local temperature in the semiconductor substrate is lower than 30 percent of the melting point of the semiconductor material, preferably lower than 20 percent of the melting point of the semiconductor material, more preferably lower than 10 percent of the melting point of the semiconductor material.

According to embodiments of the present invention each pump laser pulse may be characterized by an average pump laser pulse irradiance I_puise. The average pump laser pulse irradiance I_puise is defined by the ratio of the pump laser pulse peak power P_puise to the focal spot area of the pulse and is expressed in W/cm². The average pump laser pulse irradiance I_puise during the pulse duration is preferably higher than Ie7 W/cm². The pump laser pulse duration is preferably smaller than the recombination time of the semiconductor substrate material under optical injection, i.e. the recombination time for the generated amount of excess carriers. Preferably the pump laser pulse duration is smaller than 10 picoseconds. The average pump laser pulse irradiance I_puise during the pulse duration D is preferably higher than Ie7 W/cm².

According to embodiments of the present invention each pump laser may be characterized by an average pump laser irradiance I. The average pump laser intensity or average pump laser irradiance I is defined by the ratio of the pump laser average power Pw to the focal spot area of the laser beam . The average power Pw of the pump laser being defined by the pump laser pulse energy to the pump laser pulse period P (being the total time between two pulses, i.e. on and off state of the pump laser) .

The pump laser pulse period P is preferably larger than Ie-

5 seconds .

The average pump laser irradiance I is preferably smaller than leβ W/cm²

According to embodiments of the present invention a pulsed pump laser being able to generate short laser pulses is used. With short laser pulses is meant laser pulses in the range femto- to nanoseconds laser pulses.

According to embodiments of the present invention the predetermined parameter is any of the pump pulse power, the pump-probe time delay or the probe wavelength.

When the predetermined parameter comprises the pump laser pulse power, a power curve may be formed. The power curve may be a relation expressing the reflection signal as function of the pump laser pulse power. When the predetermined parameter comprises the pump-probe time delay, a pump-probe time delay curve may be formed. The pump-probe time delay curve may be a relation expressing the reflection signal as function of the pump- probe time delay.

When the predetermined parameter comprises the probe wavelength, a probe wavelength curve may be formed. The probe wavelength curve may be a relation expressing the reflection signal as function of the probe wavelength. According to embodiments of the present invention directing a probe laser beam comprises directing at least one probe laser pulse, wherein directing the at least one probe laser pulse is performed after a predetermined pump-probe time delay ΔT with respect to directing the at least one pump laser pulse. The at least one probe laser pulse may be directed after the at least one pump laser pulse. The probe laser pulse may be defined by a probe laser pulse duration. According to embodiments of the present invention the pump laser pulse duration and the probe laser pulse duration are substantially identical.

According to embodiments of the present invention the pump- probe time delay ΔT is smaller than the time wherein recombination effects of the excess carriers start to dominate .

The pump laser pulse duration is preferably equal or smaller than the pump-probe time delay ΔT .

According to embodiments the (series of) pump laser pulse (s) and the (series of) probe laser pulse (s) may be generated from a common source laser The common laser may comprise a pulsed laser oscillator, generate a pulsed laser beam which may be split into a high power pulsed pump laser beam and into a low power pulsed probe beam. According to embodiments the (series of) pump laser pulse (s) and the probe beam pulse may be generated from different source lasers. The pulsed pump laser beam may be generated by a pump laser and the pulsed probe laser beam may be generated by a probe laser which is different from the pump laser.

According to embodiments of the present invention the probe laser beam comprises a plurality of different wavelengths. The probe laser beam may have a broad range spectrum or may have multiple discrete lines or wavelength ranges which can be individually selected. According to embodiments of the present invention the probe wavelengths may range from deep UV (100 nm) to near infrared (1000 nm) . The probe wavelength may also range up to 2000 nm.

According to a second inventive aspect a system for measuring a bulk property of a semiconductor substrate is disclosed. The system comprises a pulsed pump laser for generating an amount of excess carriers in the semiconductor substrate, a probe laser for impinging a laser beam on the semiconductor substrate, the probe laser beam at least partially reflected by the excess carriers, thus generating a reflection signal, a means for measuring the reflection signal due to the generated amount of excess carriers, e.g. a detector in function of a predetermined parameter, and a means for determining at least part of the carrier profile based on the measured reflection signal. The means for measuring the reflection signal may be an optical detector. The means for determining at least part of the carrier profile may be one or more processors, programmed for determining the carrier concentration and/or the junction depth based on a predetermined algorithm or on a neural network. According to embodiments of the second inventive aspect the probe laser beam may be a pulsed probe laser beam. The pulsed pump laser beam and the pulsed probe laser beam may be generated from a common laser.

According to embodiments of the second inventive aspect the pulsed probe laser may have a time delay with respect to the pulsed pump laser. The pump-probe time delay is preferably smaller than the time wherein diffusive and/or recombination effects of the excess carriers start to dominate. Advantageously, the system also may comprise a controller for controlling the pulsed pump laser beam, the probe laser beam and the means for measuring the reflection. Such controlling may comprise synchronizing the pulsed pump laser and the probe laser beam and/or the means for measuring the reflection. For silicon the pump-probe time delay is preferably about 30 picoseconds or less. According to embodiments of the second inventive aspect the means for generating a probe laser beam is a means for generating a plurality of wavelengths for the generated probe laser beam.

It is an advantage of certain embodiments of the present invention that a complete active doping profile can be measured for samples having a high doping concentration without substantially destroying the samples. It is an advantage of certain embodiments of the present invention that carrier profiles in ultra shallow junctions can be determined non-destructively, i.e. without sample preparation .

It is an advantage of certain embodiments of the present invention that doping incorporation may be monitored at key points in the process flow and thus leading to an enhanced product quality. It is an advantage of certain embodiments of the present invention that a method for determining the active dopant profile may be applied in-line, i.e. in the production process environment.

It is an advantage of certain embodiments of the present invention that a user friendly and easy to operate method may be applied for determining the active dopant profile of a semiconductor substrate in a short measurement time. It is also an object of some embodiments of the present invention to provide good methods, systems and devices for processing optical data for determining a carrier profile of a semiconductor substrate, even if the sample has a high concentration of active dopants. It is an advantage of some embodiments of the present invention that an efficient and accurate method for processing optical data is obtained resulting in an active doping profile.

It is an advantage of certain embodiments of the present invention that a complete active doping profile can be determined or reconstructed from an optical measurement on the doping profile. The active doping profile may be any arbitrary doping profile.

It is an advantage of embodiments of the present invention that a unique solution may be determined for the active doping profile based on an optical measurement of the active doping profile.

It is an advantage of embodiments of the present invention that an unknown arbitrary doping profile may be reconstructed from an optical measurement in a fast and flexible way while no prior assumptions need to be made about the initial shape of the profile.

Brief description of the figures Figure 1 illustrates the basics of Carrier Illumination, an example of a state-of-the-art photo modulated optical reflectance technique.

Figure 2 illustrates (top) the generated excess carrier concentration profile for a carrier profile and (bottom) an example of a photo modulated optical reflectance power curve .

Figure 3 illustrates schematically the acquisition of a photo modulated optical reflectance power curve using a continuous wave pump laser.

Figure 4 illustrates the relationship between the carrier profile, the excess carrier concentration profile and the derivative of the excess carrier concentration profile.

Figure 5 illustrates the relationship between the differential temperature and differential excess carrier concentration in function of the applied pump laser power.

Figure 6 gives a schematic representation according to embodiments of the present invention.

Figure 7 gives a schematic representation of an optical measurement method for determining a carrier profile according to embodiments of the present invention.

Figure 8 gives an overview of a series of pump laser pulses and its characteristics according to embodiments of the present invention.

Figure 9 gives a schematic overview of the timescales for various electron and lattice processes in laser-excited solids .

Figure 10 shows the differential reflectance ΔR or measured reflectance change in function of the pump-probe time delay according to embodiments of the present invention.

Figure 11 gives a schematic representation of a system according to embodiments of the present invention. Figure 12 shows schematically a general setup according to embodiments of the present invention.

Figure 13A to 13D illustrate probing of a doping profile using a PMOR technique, such as for example Carrier

Illumination, as function of the carrier injection, as can be used in embodiments of the present invention.

Figure 14A and Figure 14B illustrate examples of a method for determining a carrier profile of a semiconductor substrate based on an optical measurement according to an embodiment of the present invention.

Detailed description of the invention

The present invention will be described with respect to particular embodiments and with reference to certain drawings but the invention is not limited thereto but only by the claims. The drawings described are only schematic and are non-limiting. In the drawings, the size of some of the elements may be exaggerated and not drawn on scale for illustrative purposes. The dimensions and the relative dimensions do not correspond to actual reductions to practice of the invention.

It is to be noticed that the term "comprising", used in the claims, should not be interpreted as being restricted to the means listed thereafter; it does not exclude other elements or steps. It is thus to be interpreted as specifying the presence of the stated features, integers, steps or components as referred to, but does not preclude the presence or addition of one or more other features, integers, steps or components, or groups thereof. Thus, the scope of the expression "a device comprising means A and B" should not be limited to devices consisting only of components A and B. It means that with respect to the present invention, the only relevant components of the device are A and B.

The invention will now be described by a detailed description of several embodiments of the invention. It is clear that other embodiments according to the invention can be configured according to the knowledge of persons skilled in the art without departing from the true spirit or technical teaching of the invention, the invention being limited only by the terms of the appended claims.

Where in embodiments of the present invention reference is made to a high carrier concentration, reference is made to a carrier concentration of at least Iel8 /cm³, more preferably at least Iel9 /cm³, even more preferably at least le20 /cm³.

Where in embodiments of the present invention reference is made to intensity or irradiance I of a laser beam, this is defined by the ratio of the power P of the laser beam to the focal spot area A of the laser beam:

Intensity or Irradiance I (W/cm²) = power P (W) / focal spot area A (cm^Λ2)

Where in embodiments of the present invention reference is made to power P of a pulsed laser beam, this is defined by the ratio of the energy of the laser beam to the duration D of the laser beam (exposure) :

Power P (W) = energy E (J) / duration D (sec)

Where in embodiments of the present invention reference is made to fluence F of a laser beam, this is defined by the ratio of the energy E of the laser beam to the focal spot area A of the laser beam:

Fluence F (J/cm²) = energy E (J) / focal spot area A (cm²)

= Irradiance I (W/cm²) * duration D (sec)

For the purpose of teaching the invention, embodiments of the present invention not being limited thereto, the description discusses the application of the invention for the conventional Carrier Illumination™ technique. In the further description, when using the term CI power curves, a representation is meant of the variation of the Cl-signal (y-axis) versus the pump laser power (x-axis) . However, the invention can also be applied to any other non-destructive measurement technique using a pump laser to create excess carriers in a semiconductor substrate and at least one probe laser whose reflection in the semiconductor substrate is correlated to the presence of the excess carriers. Instead of the Cl-signal, the value of a parameter representative or characteristic for the respective nondestructive technique should then be determined and plotted on the y-axis, yielding a curve representing the variation of that parameter as a function of the pump laser power, the probe laser wavelength, the time delay between a pump laser pulse and a probe laser pulse, etc.

The methods according to the invention may be used for determining the (active) carrier concentration level and electrical junction depth in a semiconductor substrate having a doped structure.

The doped structure may be an arbitrary profile. The doped structure may be monotonic or non-monotonic . The doped structure may be a highly-lowly doped or lowly-highly doped structure .

The doped structure may be a highly-lowly doped structure also often referred to as non-retrograde doped structure. A (non-) retrograde doping profile is defined by a profile wherein the concentration of dopants (decreases) increases from the surface of the semiconductor substrate to the areas located deeper in the semiconductor substrate. With highly-lowly doped structure is meant a structure having an active doping profile (also referred to as carrier profile) with a maximum near the surface and decreasing towards the bulk of the substrate. Examples of highly-lowly doped structures may be structures formed and/or doped by CVD (chemical vapour deposition) , ion implantation, diffusion, ....

The semiconductor substrate 100 may be formed in any suitable way, e.g. by depositing an in-situ doped layer 100b on top of layer 100a, yielding a uniform doping profile or e.g. by implanting dopants into the substrate 100, yielding a doped region 100b and an undoped region 100a. By using e.g. ion implantation for implanting dopants into the substrate 100, any kind of doping profile can be obtained depending on the choice of implant species, the energy and implantation dose used. Layer 100b can be doped with a dopant of the same or the opposite type of dopant used to dope the underlying layer 100a.

In the following the present invention will be described with reference to a silicon substrate but it should be understood that the invention applies equally well to other semiconductor substrates. In embodiments, the "substrate" may include a semiconductor substrate such as e.g. a silicon, a germanium (Ge), a silicon germanium (SiGe), a gallium arsenide (GaAs), a gallium arsenide phosphide (GaAsP) or an indium phosphide (InP) substrate. The "substrate" may include for example, an insulating layer such as a Siθ2 or a Si3N₄ layer in addition to a semiconductor substrate portion. Thus, the term substrate also includes silicon-on-glass, silicon-on sapphire substrates. The term "substrate" is thus used to define generally the elements for layers that underlie a layer or portions of interest. Also, the "substrate" may be any other base on which a layer is formed, for example a glass or metal layer. Accordingly a substrate may be a wafer such as a blanket wafer or may be a layer applied to another base material, e.g. an epitaxial layer grown onto a lower layer. The term "crystalline substrate" includes various forms of crystalline material such as monocrystalline or microcrystalline .

By way of explanation, in Fig. 1, first the basic idea or methodology of Carrier Illumination is illustrated. An example of a Carrier Illumination tool is the Boxer Cross Carrier Illumination CI BX-IO. Again, it has to be mentioned that this is only meant for explanation of the invention, embodiments of the invention not being limited thereto. The invention may also be applied to other nondestructive techniques such as, for example, the thermal wave technique. As already described, Carrier Illumination™ uses two lasers, i.e. a pump laser generating a pump laser beam 201, e.g. a red laser beam and a probe laser generating a probe laser beam 202a, e.g. an infrared (IR) laser beam, both laser beams 201, 202a impinging on a semiconductor substrate 100, e.g. a silicon substrate, comprising a doped layer 100a and an undoped (or lower doped) layer 100b. In between the doped 100a and undoped (or lower doped) 100b region an electrical junction (EJ) 101 is present. The incident pump laser beam 201 excites an excess charge carrier distribution in the substrate 100. The excess carrier profile N(z) as function of depth z into the substrate 100 is shown, indicated by graph 500. The impinging probe laser beam 202a reflects, and thus generates a reflected probe laser beam 202b, due to index of refraction changes which are due to gradients of the charge carrier distribution N(z) .

The pump laser injects charges and generates an excess carrier profile, which is a function of the active doping concentration. By changing a parameter, for instance the amplitude of the modulated pump power, a spectrum of the differential reflectance (ΔR) is obtained, from which the active doping profile may then be reconstructed. Typically, prior-art CI-measurements involve the monitoring of the CI- signal as the generation power of the pump laser 201 is swept from low to high (full) power. The resulting curves, plotting Cl-related signals as a function of the pump laser power, are referred to as power curves as for example shown in Fig. 2 in the bottom graph 600.

A power curve 600 measures both ac and dc components of the reflected signal, the dc component being proportional to the reflectance R and the ac component being proportional to the differential reflectance ΔR. A ΔR/R signal is thus measured as function of the modulation amplitude of the pump laser power. Typically in prior art the pump laser power is swept from low to full power in an amount of steps S and from these S measurements a power curve 600 is generated.

Figure 3 schematically shows the principle for generating a power curve by modulating a continuous wave pump laser beam. The measurement starts with a first modulation of the pump laser beam 701a, 701b with a first modulation amplitude of the pump power. Typically each step starts with a settling time 701a and is followed by an acquisition time 701b. The measurement (i.e. detecting the reflectance signal from the reflected probe laser beam) starts during the acquisition time after the settling time. During the settling time, the first modulation of the pump power generates an amount of excess carriers in the semiconductor substrate. Due to the drift-diffusion mechanism of these excess carriers, the measurement of the reflected probe laser beam signal (which is proportional to the ΔR/R signal 601) occurs after a settling time during the acquisition time. As the level of excess carriers will increase proportionally to the applied generation pump power, one might, in a simplistic view, expect a proportional increase of the CI-signal.

In a next step the modulation power of the pump laser beam is increased. A second modulation of the pump laser beam 702a, 702b generates a second amount of excess carriers in the semiconductor substrate, yielding a second reflectance signal which is detected by the reflected probe laser beam generating another point 602 of the power curve 600. This is repeated several times 703a, 703b until the highest modulation power 704a, 704b is applied by the pump laser beam generating the highest and last value 604 of the power curve 600. The (continuous wave) probe beam and (continuous wave) pump beam are both illuminating the semiconductor substrate simultaneously.

As an example, for the CI BX-10 system, the modulation amplitude of the pump irradiance may for example be increased from 0 to 3.79e5 W/cm² (which represents a pump power of 56.7 mW for a pump laser spot area of about 3.7 μm) stepwise, in one embodiment for example in 22 steps of 1.72e4 W/cm² (which represents 2.58 mW) and the signals are recorded at each step. The measurement starts after a settling time and the signals are integrated during an acquisition time of typically 100 ms (corresponding to 200 modulation periods of 2 kHz) . The settling time is typically about 100 ms seconds.

As the level of excess carriers will increase proportionally to the applied generation pump power, one might, in a simplistic view, expect a proportional increase of the CI-signal. Figure 4 illustrates this mechanism. In this figure, the carrier concentration (curve 400), the excess carrier profile (curve 401) and the derivative of the excess carrier profile (curve 402) are plotted as a function of the depth into the semiconductor substrate 100. As the charge carrier injection level increases with increasing power of the pump signal, as indicated by arrow 403 in figure 4, one also probes different concentration levels along the slope of the carrier profile, and thus probes different profile depths, leading to a further change in the CI-signal. As such, the power curve will also contain information on the profile abruptness. For abrupt profiles, dN(z)/dz (curve 800) peaks where the profile reaches the injection level, as illustrated by the vertical dashed line 404 in figure 4.

A PMOR measurement is based on the generation of an amount of excess carriers in the semiconductor substrate which can be translated to a reflectance signal which is proportional to the active doping concentration. However for resolving high doping concentrations, e.g. higher than Iel8 /cm³, e.g. in the range of Iel8 to 3e20 /cm³, in the doped layer 100a of the semiconductor substrate 100, a high substrate excess carrier concentration is required.

AR

A PMOR reflection signal on a box-like active doping profile may be defined by a three-component signal in the following analytical equation:

(1 )

wherein

- N_act is the peak carrier concentration of the carrier profile

- ΔN_sub is the generated amount of excess carrier concentration in the substrate

- m_e and m_h are the electron and hole mass respectively - β is the Drude coefficient

- n₀ is the refractive index at the semiconductor surface

- X_j is the electrical junction position

- λp_robe is the wavelength of the probe beam

- δ is the thermorefractive index

- ΔT_surface is the excess temperature at the surface due to the generated amount of excess carrier concentration .

The first component is the layer plasma component, which is the product of the layer plasma component coefficient and a (l-cos()) term, related both to the active doping concentration in the doped layer 100a and to the excess carrier concentration in the substrate 100b. The second component is the substrate plasma component, which is the product of the substrate plasma component coefficient and a cos () term, linked to the excess carrier concentration in the substrate 100b. The third and final component is the thermal component which is the same as the thermal component on a homogeneous silicon substrate. Although Equation (1) is related to a box-like carrier profile, an analogue analytical equation may be used for arbitrary carrier profiles. For an arbitrary carrier profile an analogous analytical equation may be used. With a continuous wave pump laser ΔN_sub is limited to values lower than Iel8 /cm and much smaller than N_act . As a

consequence

is much lower than ΔN_sub and thus the

N_act layer-plasma component becomes negligible to the substrate- plasma component. With a continuous wave pump laser one is therefore not sensitive to high carrier concentration. In order to be sensitive to high carrier concetration (i.e. high N_act ) , ΔN_sub should be of the order of N_act . I f ΔN_sub i s

comparable to N_act, then also IΔΛU² is comparable to N_act,

N_act and as such the layer-plasma component has the same order of magnitude as the substrate-plasma component. Therefore, as a rule of thumb, as can be extracted from the equation (1), the amount of excess carrier concentration should at least be larger than the active doping concentration divided by 10 and ideally should be of the same order. Ultra shallow junctions in the coming years will feature doping concentrations as high as 5e20 /cm3, so that substrate excess carrier concentrations of at least 5el9 /cm3 will be required. For example for a doping profile with a concentration in the range Iel8 to 3e20 /cm , this means that pump powers in the range of a few milliwatts to hundreds of watts are required, leading to damage or even melting of the studied sample.

Also for arbitrary profiles the amount of excess carrier concentration should at least be larger than the active doping concentration divided by 10 and ideally should be of the same order.

Figure 5 shows the differential excess carrier concentration 502 and the differential temperature 501 in the semiconductor substrate when applying high pump illumination power from 0 to 250 W (or irradiance W/cm²) . For example in order to generate an amount of excess carriers equal to about 1.3e20 /cm³ a pump energy of about 80 W is necessary, leading to a temperature in the substrate of about 1000 Kelvin which is also the temperature around which silicon starts to melt and turn from solid into liquid silicon.

For a continuous wave pump laser, as used in state of the art PMOR techniques, a maximum pump energy of about 60 mW may be achieved, which corresponds to a maximum amount of excess carrier concentration of about 4el8 /cm³ (Fig. 5) . According to a first inventive aspect of the present invention a pulsed pump laser is used for generating an amount of excess carriers in the semiconductor substrate. In a pulsed mode of operation, the output of the pump laser varies with respect of time, typically taking the form of alternating 'on' (i.e. with a power) and 'off (i.e. with no power) periods. As much energy as possible is created in an as short time as possible (in the 'on' mode) . According to a first inventive aspect an optical measurement method 10 to determine a carrier profile of a semiconductor substrate is disclosed, the method comprising directing 11 at least one pump laser pulse of a pump laser beam on the semiconductor substrate. The at least one pump laser pulse having a pump laser pulse power generating an amount of excess carrier concentration in the semiconductor substrate which is comparable to or higher than the peak concentration of the active dopant profile. The method also includes directing 12 a probe laser beam on the semiconductor substrate such that the probe laser beam is at least partially reflected by the generated amount of excess carrier concentration and measuring 13 the reflection signal of the probe laser beam as a function of a predetermined parameter. The method furthermore comprises determining 14 at least part of the carrier profile from the reflection signal as function of the predetermined parameter. The latter may be performed using a predetermined algorithm, as will be described further. Figure 7 gives a schematic presentation of the optical measurement method according to embodiments of the present invention. A series of pump laser pulses 71, 72, 73, 74 is focused on the semiconductor substrate, each pump laser pulse having a pump laser pulse power generating an amount of excess carrier concentration, the series comprising at least one pump laser pulse 74 generating an amount of excess carrier concentration in the semiconductor substrate comparable to or higher than at least the peak concentration of the carrier profile. According to embodiments of the present invention comparable to or higher than comprises equal to or higher than the peak concentration of the carrier profile divided by 10. According to embodiments of the present invention comparable to or higher than comprises equal to or higher than the peak concentration of the carrier profile. The relationship between the necessary amount of excess carrier concentration - generated by the pulsed pump laser - and the peak carrier concentration of the carrier profile may also be defined using the layer-plasma component and the substrate-plasma component from for example equation (1) for a box-profile. An analogue definition may be done for an arbitrary profile.

According to embodiments of the present invention generating an amount of excess carrier concentration in the semiconductor substrate which is comparable to or higher than the peak concentration of the active dopant profile comprises generating an amount of excess carrier concentration in the semiconductor substrate inducing a layer-plasma reflectivity component coefficient according to for example in equation (1) which is at least comparable to the substrate-plasma reflectivity component coefficient according to for example in equation (1) .

Each pump laser pulse 71, 72, 73, 74 comprises a first time period 71a, 72a, 73a, 74a where a pulse power ΔP is applied during a first pulse duration (from to to ti, from t2 to t3, from tk-i to t_k) and a second time period 71b, 72b, 73b, 74b where no pulse power is applied during a second duration (from ti to t2) After the first time period 71a, 72a, 73a, 74a, i.e. after the pump laser pulse duration the reflected signal is measured using a probe laser beam. Thus during the second time period 71b, 72b, 73b, 74b, when no pulse power is applied, the reflected signal is measured. According to embodiments of the present invention directing at least one pump laser pulse comprises directing a series of pump laser pulses, the series comprising the at least one pump laser pulse and each pump laser pulse having the same pump laser pump power as the at least one pump laser pulse. Thus for each pump laser pulse power, the pump laser pulse cycle of the at least one pump laser pulse may be repeated a few times before applying the next pump laser pulse power. A plurality of series of pump laser pulses 701, 702, 703, 70k may be applied; each series of pump laser pulses having another pump laser pulse power, but each pump laser pulse of the series of pump laser pulses may have the same pump laser pulse power. The more pump laser pulses in a series of pump laser pulses, the higher the signal-to-noise ratio, and thus the better the precision of the reflection signal will be.

At least one pump laser pulse 74 has a pump laser pulse power which is able to generate a maximum amount of excess carriers comparable to or higher than at least the peak concentration of the carrier profile, preferably equal to or higher than the peak concentration of the carrier profile divided by 10, even more preferably equal to or higher than the peak concentration of the carrier profile. If for example the active doping profile has a peak concentration of 4e20 /cm³, which is a typical value for source/drain extensions, the at least one pump laser pulse should be able to generate an amount of excess carriers equal to or higher than le20 /cm , more preferably equal to or higher than 4e20 /cm³.

According to embodiments of the present invention the at least one pump laser pulse, and thus the generated amount of excess carrier concentration is chosen such that the increase in local temperature in the semiconductor substrate is lower than 30 percent of the melting point of the semiconductor material, preferably lower than 20 percent of the melting point of the semiconductor material, more preferably lower than 10 percent of the melting point of the semiconductor material.

According to embodiments of the present invention each pump laser pulse may be characterized by a pump laser pulse power 810, by a pump laser pulse duration D 811, by a pump laser pulse period P 813. The pump laser pulse period P 813 is the sum of the pump laser pulse duration D 811 and time duration 812 (Fig. 8) .

From a general formula the pump laser power P may be defined by the following formula:

With P(t) being the (instantaneous) power of the pump laser in function of time and Δt being a time period. When the integral is made over a period Δt = [813], then P = Pw is the average pump laser power. When the integral is made over the pulse duration Δt = [811], then P=P_pui_se is the average pulse pump laser power.

According to embodiments of the present invention each pump laser pulse may be characterized by an average pump laser pulse intensity or average pump laser pulse irradiance Ipuise- The average pump laser pulse intensity or average pump laser pulse irradiance I_puise is defined by the ratio of the pump laser pulse peak power P_puise (810) to the focal spot area of the pulse and is expressed in W/cm²: Average pump laser pulse Intensity or Irradiance I (W/cm²) = pulse peak power (W) / focal spot area (cm^Λ2) . The pulse peak power P_puise of the pump laser being defined by the pump laser pulse energy 810 to the pump laser pulse duration D 811.

The pump laser pulse duration D 811 is preferably smaller than the recombination time of the semiconductor substrate material under optical injection, i.e. the recombination time for the generated amount of excess carriers. For example, for silicon, the pump laser pulse duration is smaller than 10 picoseconds.

The average pump laser pulse irradiance I_pui_se during the pulse duration D is preferably higher than Ie7 W/cm². According to embodiments of the present invention each pump laser may be characterized by an average pump laser intensity or average pump laser irradiance I. The average pump laser intensity or average pump laser irradiance I is defined by the ratio of the average pump laser power Pw to the focal spot area of the pulse and is expressed in W/cm²: Average pump laser Intensity or Irradiance I (W/cm²) = average power (W) / focal spot area (cm^Λ2) . The average power Pw of the pump laser being defined by the pump laser pulse energy 810x811 to the pump laser pulse period P 813 (being the total time between two pulses, i.e. on and off state of the pump laser) .

The pump laser pulse period P 813 is preferably larger than le-5 seconds.

The average pump laser irradiance I is preferably smaller than leβ W/cm².

The power of the pump laser pulse is proportional with the amount of excess carries which can be generated in the semiconductor substrate. The higher the amount of excess carriers to be generated, the higher the pump laser pulse power should be.

According to embodiments of the present invention a pulsed pump laser being able to generate short laser pulses is used. With short laser pulses is meant laser pulses in the range of le-15 to le-8 seconds, i.e. femto- to nanoseconds laser pulses. Preferably, the laser pulse should be long enough to let the generated excess carriers to thermalize and to organize themselves in the electrical field of the carrier profile. Practically, the pump laser pulse duration (811) should be substantially larger than the Debye time εε t_DB= —— , which is typically le-14 seconds or even less for qNμ carrier concentrations above Iel7 /cm3. Furthermore, if the laser pulse duration or/and pump-probe time delay are smaller than the carrier recombination time, surface recombination and bulk defect issues are also eliminated. The use of very short laser pulses according to embodiments of the present invention allows to achieve high excess carrier concentrations and low excess temperatures at the same time. Otherwise said by using short laser pulses with a low enough repetition rate (1 to 100 kHz), thermal effects can be dramatically reduced, if not completely eliminated.

By using a pulsed picosecond (or femtosecond) pump beam and according to embodiments a variable pump-probe time delay in the pico- to millisecond range also the excess carrier dynamics the picosecond to millisecond scale may be determined, i.e. processes from carrier generation and thermalization to carrier recombination and diffusion. This is shown in region A from figure 9, which gives a schematic overview of the timescales for various electron and lattice processes in laser-excited solids.

If the pulse duration is short enough so that both the carrier recombination and diffusion can be neglected, the typical pump pulse energy needed to achieve the required injection levels can easily be calculated.

The fluence F (J/cm²) is given by,

F=IA,= ^lιυβA>

(l-R)αr where I is the irradiance (W/cm²) , hv = 1.5 eV is the photon energy, R = 0.35 is the sample reflectivity and α = 600/cm is the absorption coefficient (the given values correspond to a pump laser wavelength λ=830 nm) , G is the carrier generation rate (cm^s^"1) and Δt is the pulse duration (s) . The excess carrier concentration generated at the end of the pulse is given by GΔt and should reach 5e20 /cm3 in order to probe high enough doping concentrations such as those encountered in future ultra shallow junctions .

One can calculate the required fluence F = 0.29 J/cm2 for a pump laser at λ=830 nm. This corresponds to a pulse energy of 36 nJ considering a spot radius of 2 μm. If the pulse duration is le-12 seconds, the corresponding average pump pulse irradiance is I = 2.9ell W/cm . This is well below the damage threshold for semiconductors which is around Iel3 W/cm , depending on material and wavelength used) . Lasers with characteristics according to embodiments of the present invention are commercially available at a reasonable cost. Further, if the differential reflectance (ΔR/R in the range le-5 and Ie-I) can be measured with enough accuracy during the pulse duration, a method for deconvoluting a complete dopant profile, i.e. for all doping concentrations can be performed. According to embodiments of the present invention, the pulsed laser may for example be a neodymium-doped yttrium aluminium garnet (Nd:YAG) laser, an Argon (Ar) laser or an aluminum-gallium arsenide (AlGaAs) laser, a fibre laser. Also a probe laser beam is directed 12 on the semiconductor substrate such that the probe laser beam is at least partially reflected by the generated amount of excess carriers .

The reflected signal of the probe laser beam is measured 13 as function of a predetermined parameter.

According to embodiments the predetermined parameter is the pump pulse power. Thus the reflected signal is measured as function of the pump pulse power as such creating a power curve .

According to embodiments the predetermined parameter is a pump-probe time delay between the pump pulse and the probe beam. Thus the reflected signal is measured as function of the pump-probe time delay.

According to embodiments directing a probe laser beam comprises directing at least one probe laser pulse, wherein directing the at least one probe laser pulse is done with a pump-probe time delay ΔT with respect to directing the at least one pump laser pulse. Directing the at least one probe laser pulse is done after directing the at least one pump laser pulse. The pump-probe time delay ΔT should be smaller than the time wherein recombination effects of the excess carriers start to dominate. The pump laser pulse duration is equal or smaller than the pump-probe time delay ΔT.

The probe laser pulse may be defined by a probe laser pulse duration. The pump laser pulse duration and the probe laser pulse duration are preferable identical. The reflected signal comprises the differential reflectance ΔR from the semiconductor substrate. Figure 10 shows the time-dependent reflectivity change of silicon on a 2 picosecond time scale. The differential reflectance ΔR or measured reflectance change strongly varies (see Figure 10) while the excess carriers are being generated during the pump pulse with a coherent spike A and a pulse-width limited drop B. It subsequently plateaus for a few picoseconds C and eventually tends to recover its initial state trough surface or Auger recombination D. The variation in reflectivity is strongly negative in the C plateau and mainly attributed to free carrier absorption. By analyzing the time-dependant reflectivity (differential reflectance or variation in reflectivity) more information can be obtained on the different physical process of the excess carriers.

By varying the time delay between the pump laser pulse and the probe laser pulse (f.e. from sub-picosecond to hundreds of picoseconds), different instants of the carrier dynamics may be probed. This is important as there is a time window of about 1 to 30 picoseconds where the excess carrier distribution probes the active doping profile without being affected by the carrier generation and recombination transients .

According to embodiments the pump-probe time delay, i.e. the time delay between the pump laser pulse and the probe laser pulse is smaller than the time wherein diffusive and/or Auger recombination effects of the excess carriers start to dominate. For example for silicon the pump-probe time delay is preferably smaller than 10 picoseconds. According to embodiments the (series of) pump laser pulse (s) and the (series of) probe laser pulse (s) may be generated from a common source laser. The common laser may comprise a pulsed laser oscillator, generate a pulsed laser beam which may be split into a high power pulsed pump laser beam and into a low power probe beam.

According to embodiments the (series of) pump laser pulse (s) and the probe beam pulse may be generated from different source lasers.

The pulsed pump laser beam may be generated by a pump laser and the pulsed probe laser beam may be generated by a probe laser which is different from the pump laser.

A second inventive aspect of the present invention discloses the use of multiple wavelength probe laser for an optical measurement technique to determine the carrier profile of a semiconductor structure. This concept may be combined with any of the other aspects disclosed in embodiments of the present invention.

In order to reconstruct arbitrary profiles (which may be monotonic or non-monotonic) , a multiple wavelength approach is used according to embodiments of another inventive aspect. Since the differential reflectance is the cosine

Fourier transform of the excess carrier derivative (see equation (I)), the excess carrier derivative may be reconstructed by performing the inverse cosine transform on the reflected intensity spectrum. By applying different wavelengths for the probe laser this reconstruction may be done .

According to embodiments of the present invention the probe wavelengths may range from deep UV (100 nm) to near infrared (1000 nm) . The probe wavelength may also range up to 2000 nm.

The multiple wavelength detection may be done using a supercontinuum generator in the probe path. A supercontinuum generator converts the laser light into a light with a very broad spectral bandwidth. The supercontinuum generator is preferably based on a photonic crystal fiber. Alternatively also bulk crystals may be considered.

According to embodiments a so called wave spectrum may be measured, i.e. the reflected signal may be measured for a probe wavelength in the range of 400 to 1000 nm. A spectral resolution of about 1 nm may be achieved.

According to a second inventive aspect a system for measuring a bulk property of a semiconductor substrate is disclosed. The system comprises a pulsed pump laser for generating an amount of excess carriers in the semiconductor substrate, a probe laser for impinging a laser beam on the semiconductor substrate, the probe laser beam at least partially reflected by the excess carriers, thus generating a reflection signal, a means for measuring the reflection signal due to the generated amount of excess carriers, e.g. a detector in function of a predetermined parameter, and a means for determining at least part of the carrier profile based on the measured reflection signal. The lasers may be lasers as described above. The means for measuring the reflection signal may be an optical detector such as for example a PIN Silicon photodiode. The means for determining at least part of the carrier profile may be one or more processors, programmed for determining the carrier concentration and/or the junction depth based on a predetermined algorithm or on a neural network. Determining at least part of the carrier profile may comprise determining at least a doping concentration and determining an electrical junction position.

According to embodiments of the second inventive aspect the probe laser beam is a pulsed probe laser beam. The pulsed pump laser beam and the pulsed probe laser beam may be generated from a common laser.

A pulsed laser oscillator may generate a pulsed laser beam which may be split into a high power pulsed pump laser beam and into a low power probe beam.

According to embodiments of the second inventive aspect the pulsed probe laser has a time delay with respect to the pulsed pump laser. The pump-probe time delay is preferably smaller than the time wherein diffusive and/or recombination effects of the excess carriers start to dominate. Advantageously, the system also comprises a controller for controlling the pulsed pump laser beam, the probe laser beam and the means for measuring the reflection. Such controlling may comprise synchronizing the pulsed pump laser and the probe laser beam and/or the means for measuring the reflection. For silicon the pump-probe time delay is preferably about 30 picoseconds or less. Figure 11 gives a schematic representation of a system according to embodiments of the present invention. A pulsed laser oscillator 1101 generates a picosecond pulsed laser beam 1107. The laser beam 1107 is split into a high power pump laser beam 1108 and a low power probe laser beam 1109. The pump laser beam 1108 passes through a variable attenuator 1105 before impacting the sample. The probe laser beam 1109 is time-delayed with respect to the pump laser beam 1108 using a variable delay line 1103 and passes through a supercontinuum generator 1104 before impacting the sample 1100 at an angle (preferably 45 degrees) and is detected for each wavelength by a detector 1102. An example of a general setup according to embodiments of the present invention is shown in Figure 12. A high power (1 μJ) pulsed laser beam (300 ps, 50 kHz), generated by a low cost and quasi maintenance free fiber optic laser 1201, first passes through a grating pair 1202 to compress the pulse down to 1 picosecond. A polarizing beam splitter 1203 separates (about 1/30 probe/pump ratio) the pulse beam in a p and s polarized pump 1210 and probe 1209 beam respectively. An optical delay line 1204 is introduced on the path of one of the beam in order to create a tunable time delay between the probe 1209 and the pump beams 1210. The pump beam 1210 then passes through a (optional) crystal doubler (1060 nm to 530 nm) 1205 and a variable attenuator 1206 (for power curve acquisition mode) . The pump beam may be mechanically chopped by a beam chopper 1207 at about 1 to 5 kHz for subsequent homodyne detection and is focused on the sample 1200 with a spot radius of about 4 to 8 μm (large enough so that it will completely overlap the probe spot) .

The probe beam 1209 is either directly focused on the sample 1200 at an incident angle of typical 45 degrees or first passes through a supercontinuum generator 1208 before reaching the sample 1200. The pump 1210 and probe 1209 beam can laterally be displaced from each other by sub-μm steps (typical 125 nm) over a few centimeter range. The reflected probe beam 1211 passes through a linear polarizer 1212 in order to eliminate scattered light from the pump. The probe signal is measured by a photodiode (when using single wavelength) or a spectrometer (when using multiple wavelength) and the differential reflectance is extracted by means of a lock-in amplifier set at the chopper frequency .

Different acquisition modes are available on the system. Power curves may be extracted. By stepping the attenuation factor of the pump attenuator, the differential reflectance can be measured as function of the pump power. Time curves may be extracted. By stepping the optical delay line, the differential reflectance can be measured as function of the probe-pump time delay. Offset curves may be determined. By stepping the pump beam laterally with respect to the probe beam, the differential reflectance can be measured as function of the probe-pump spot distance.

A (sub-) picosecond fiber laser may be used which is based on a master oscillator power amplification architecture. The picosecond laser may be based on a mode-locked fiber oscillator (1060 nm, 300 ps pulse, 1 μJ per pulse) with an integrated pulse picker 1213 in order to decrease the pulse repetition rate (to less than 50 kHz) . The pulse may then be amplified by a double stage fiber amplifier 1214, based on a Ytterbium single mode fiber or a single mode double clad fiber. The pulse (frequency chirped) may then be time- compressed to obtain a 1 ps pulse width or less, for example by using grating pair. The picosecond laser may provide sufficient excess carriers (more than Iel8 /cm³, more preferably more than le20 /cm ) .

The methods as described above are some examples of methods that may be used for determining the carrier concentration level and the electrical junction depth in a semiconductor substrate having an arbitrary doped profile. For determining at least part of the excess carrier concentration profile, a particular method may be used for processing the obtained data. A particular example of such a data processing method is described below, by way of illustration .

The data processing method may for example comprise receiving a power curve expressing a reflection signal of a probe laser reflected by an amount of excess carrier concentration in a semiconductor substrate as function of a pump laser power, the pump laser used for generating the amount of excess carriers in the semiconductor substrate. The latter may for example be obtained using a method as described above based on pulsed laser action, although embodiments of the present invention are not limited thereto. The method also comprises determining from at least a first part of the power curve the measured reflectance signal at the highest pump laser power, which represents an active doping concentration N_dop which is equal to the maximum excess carrier concentration for this first part of the power curve. The data processing method also comprises determining the junction depth X_D for said doping concentration N_dop by comparing a simulated reflectance signal with the measured reflectance signal at the highest pump laser power. The data processing method further comprises repeating the steps of determining the measured reflectance and determining the junction depth for at least another part of the power curve and determining based on said obtained junction depths and said obtained doping concentrations the carrier profile.

The basic underlying idea for determining a carrier profile of a semiconductor substrate from an optical measurement according to embodiments of the present invention is shown in Figure 13A to 13D and referred to as the peeling algorithm. Figure 13A to 13D illustrate probing of the doping profile using a PMOR technique, such as for example Carrier Illumination, as function of the carrier injection. The full line 5 shows the doping concentration profile. The arrows 51, 52, 53, 54, 55 represent variations of reflectance which may be detected using the probe laser. When excess carriers are injected, generated by a pump laser beam, regions under high injection (i.e. low doping concentration) show a smooth and flat carrier distribution equal to the excess carrier concentration while regions under low injection (i.e. high doping concentration) essentially retain their equilibrium carrier concentration. This property holds when the flat quasi-fermi level approximation can be assumed, hence when the carriers have redistributed around the electrical field. This results in a carrier distribution, which probes different parts of the doping profile as the carrier concentration (i.e. the pump power applied) is varied over the active doping concentration range. When a single wavelength probe laser is used, a doping profile may be reconstructed by a so- called adapted layer-peeling algorithm. At low power (near equilibrium) , only the low concentration part 71 of the doping profile 5 is probed (Fig. 13B) shown by the white arrows 51, 52. The amount of excess carriers injected at this low power is shown by 61, i.e comparable to or equal to the peak concentration of the low concentration part 71. As the pump power increases (FIG. 13C) , and consequently the excess carrier injection 62, a larger part of the doping profile 72 is progressively probed shown by the white arrows 51, 52, 53. At the end at full pump power (Fig. 13c) , i.e. the highest pump power applied which should ideally generate an amount of excess carriers 63 which is equal to the peak doping concentration of the doping profile 5, the full doping profile 73 is probed as shown by the white arrows 51, 52, 53, 54, 55. In one particular example, the peeling algorithm may comprise more specifically a) receiving a power curve expressing a reflection signal of a probe laser reflected by an amount of excess carrier concentration in a semiconductor substrate as function of a pump laser power of a pump laser used for generating the amount of excess carriers in the semiconductor substrate, b) determining a number of steps k=N; c) determining from the power curve a doping concentration N * of the carriers which is equal to a maximum or minimum excess carrier concentration N_bk=N

at a pump laser power P^k=N ; d) determining the junction depth X^i=Λrfor said doping concentration N^ by comparing a simulated reflection signal with the measured reflection signal at the pump laser power P^k=N ; e) for N-I subsequent steps, performing f) selecting a pump laser power P^k~l lower or higher than the previously considered pump laser power P^k , g) determining from the power curve at the selected pump laser power P^k~l a doping concentration N^ of the carriers which is equal to the maximum excess carrier concentration Λ^^"1 , , , , P^k~l ;

* at the pump laser power h) determining the junction depth X ;^k~l for said doping concentration N_dop by comparing a simulated reflection signal, based on the obtained results, with the measured reflection signal at the pump laser power

i) determining based on said obtained junction depths and said obtained doping concentrations a carrier profile. According to embodiments of the present invention, a method for determining a (unknown) carrier profile of a semiconductor substrate based on an optical measurement is disclosed by deconvolution of a measured power curve. The deconvolution may be performed point by point, where at least two subsequent iterations are necessary. One to reconstruct point by point the unknown profile, which is a fast iteration and a second one to repeat the first process to take into account the eventual non negligible interactions between different parts of the profile. If the differential reflectance (ΔR/R in the range le-5 and le-3) can be measured with enough accuracy during the pulse duration, a method for deconvoluting non-retrograde monotonic graded profiles can be proposed.

A method for determining a (unknown) carrier profile of a semiconductor substrate based on an optical measurement according to embodiments of the present invention is illustrated in Fig. 14A (for a first iteration) and FIG.6B (for a second or further iteration) . The working principle is as follows. The pump laser power P^k is first set to achieve the maximum injection level (k = N) , say N^N _b = 3e20/cm3. The profile is then approximated by a single box, whose doping concentration is assumed to be equal to the injection level, i.e. N^N _dop = N^N _b and whose junction depth Xj^N is optimized such that the differential reflectance (at the current power PN) of the simulated signal matches the measured value. The laser power is then lowered by one step (k = N-I), corresponding to a lower substrate injection level, e.g. N^N"1 _b = le20 /cm3, and a second box (deeper) is appended to the first box of the doping profile, with N^N"1 _dop = N^N"1 _b and Xj^¹ is optimized such that the differential reflectance (at the current power P¹*^"1) of the simulated signal matches the measured value. The procedure is repeated until the pump power is reduced to zero (k = 1) . Practically, the smaller the step, the better the depth and doping concentration resolutions of the extracted profile. At this stage, we have a first estimation of the doping profile by means of a series of N boxes, each being characterized by the couple {^Nkdo_P, X_D ^k} • This first approximation is usually quite good, though the surface doping concentration is usually overestimated. In order to refine the doping profile, a second iteration can be carried out (FIG. 14B) . This second iteration is identical to the first one, except that the previously obtained profile is taken into account during the optimization of X_D ^k, corresponding to power k. Practically, this second iteration mainly affects the doping concentration close to the surface, while the deeper part of the profile (of lower doping concentration) is less affected.

Consequently, the method advantageously may be adapted for combining the features of this aspect of the invention with features and advantages of other aspects of the present invention, as described above, e.g. the aspect of suing different probe wavelengths.

In one aspect, the present invention also relates to a processor for performing a method for determining an active concentration profile. Such a processor may be implemented as hardware or as software implemented or implementable on a computing machine. The processor may comprise a set of commands resulting in performing a method for determining an active concentration profile when the set of commands is run on the processor. The processor may comprises processing portions providing the functionality of the steps in the method for determining an active concentration profile based on a power curve, as described above. In further aspects, embodiments of the present invention also relate to computer-implemented methods for performing determination of a bulk property or performing the data processing performed in such methods as indicated above or as can be obtained by the functionality of the system described above. Embodiments of the present invention also relate to corresponding computing program products. Such methods may be implemented in a computing system, such as for example a general purpose computer. The computing system may comprise an input means and a data processor, which may be set up as a single data processor or as a plurality of processors. The computing system may include a processor, a memory system including for example ROM or RAM, an output system such as for example a CD-rom or DVD drive or means for outputting information over a network. Conventional computer components such as for example a keybord, display, pointing device, input and output ports, etc also may be included. Data transport may be provided based on data busses. The memory of the computing system may comprise a set of instructions, which, when implemented on the computing system, result in implementation of the standard steps of the method as set out above and optionally of the optional steps as set out above. Therefore, a computing system including instructions for implementing the method of obtaining an active dopant profile or for the data processing performed therein is not part of the prior art.

Further aspect of embodiments of the present invention encompass computer program products embodied in a carrier medium carrying machine readable code for execution on a computing device, the computer program products as such as well as the data carrier such as dvd or cd-rom or memory device. Aspects of embodiments furthermore encompass the transmitting of a computer program product over a network, such as for example a local network or a wide area network, as well as the transmission signals corresponding therewith .

Claims

1. An optical measurement method to determine an active doping profile of a semiconductor substrate, the method comprising

- directing at least one pump laser pulse of a pump laser beam on the semiconductor substrate,

- with at least one pump laser pulse having a pump laser pulse power generating an amount of excess carrier concentration in the semiconductor substrate which is at least comparable to the peak concentration of the active doping profile;

- directing a probe laser beam on the semiconductor substrate such that the probe laser beam is at least partially reflected by the generated amount of excess carrier concentration;

- measuring a reflection signal of the probe laser beam, the reflection signal being induced by the generated amount of excess carrier concentration, as a function of a predetermined parameter;

- determining at least part of the active doping profile from the reflection signal as function of the predetermined parameter.

2. An optical measurement method according to claim 1 wherein the amount of excess carrier concentration induces an increase in local temperature in the semiconductor substrate, the increase in local temperature being lower than 30% of the melting point of the semiconductor material .

3. An optical measurement method according to claim 1 or 2 wherein the at least one pump laser pulse is defined by an average pump laser irradiance I, being defined by a pump laser pulse duration D, the pump laser pulse duration D being smaller than 10 picoseconds.

4. An optical measurement method according to claim 3 wherein the at least one pump laser pulse is defined by an average pump laser pulse irradiance I_puise, the average pump laser pulse irradiance I_puise during the pulse duration D being higher than Ie7 W/cm².

5. An optical measurement method according to any of the preceding claims wherein directing a probe laser beam comprises directing at least one probe laser pulse, wherein directing the at least one probe laser pulse is performed after a predetermined pump-probe time delay ΔT with respect to directing the at least one pump laser pulse.

6. An optical measurement method according to claim 5 wherein the pump-probe time delay ΔT is smaller than the time wherein recombination effects of the excess carriers start to dominate.

7. An optical measurement method according to any of the preceding claims wherein the probe laser beam comprises a plurality of different wavelengths.

8. An optical measurement method according to claim 7, wherein the predetermined parameter comprises a probe wavelength from the plurality of different wavelengths, thus forming a wavelength relation expressing the reflection signal as function of the probe wavelength.

9. An optical measurement method according to any of the claims 1 to 8 wherein the predetermined parameter comprises the pump laser pulse power, thus forming a power curve .

10. An optical measurement method according to any of the claims 5 to 9 wherein the predetermined parameter comprises the pump-probe time delay, thus forming a time delay relation expressing the reflection signal as function of the pump-probe time delay.

11. An optical measurement method according to any of the preceding claims wherein the pump laser beam and the probe laser beam are generated from a common laser.

12. A system for measuring a bulk property of a semiconductor substrate, comprising:

- means for generating a pulsed pump laser beam, the pulsed pump laser beam generating an amount of excess carriers in the semiconductor substrate;

- means for generating a probe laser beam for impinging on the semiconductor substrate, the probe laser beam at least partially reflected by the excess carriers, thus generating a reflection signal;

- means for measuring the reflection signal, induced by the amount of excess carriers, as function of a predetermined parameter;

- means for determining at least part of the active doping profile based on the measured reflection signal .

13. A system for measuring a bulk property of a semiconductor substrate according to claim 13, wherein the means for generating the probe laser beam comprises a means for generating a pulsed probe laser beam.

14. A system for measuring a bulk property of a semiconductor according to claim 14, the system further comprising a controller for synchronising the generated pulsed pump laser beam and the generated pulsed probe laser beam with a predetermined pump- probe time delay ΔT.

15. A system for measuring a bulk property of a semiconductor substrate according to claim 13 to 15, the means for generating a probe laser beam being a means for generating a plurality of wavelengths for the generated probe laser beam.