Total-reflection X-ray Fluorescence (TXRF)

Total-reflection X-ray Fluorescence (TXRF) 

This grazing incidence x-ray technique excites elemental contamination on the surface of a wafer.

Principles of Total Reflection X-ray Fluorescence Analysis

1. General

2. XRF

3. Total reflection of X-rays

4. TXRF

5 X-ray source

6. X-ray monochromator

7. Silicon Drifted Detector (SDD)

8. Qualitative analysis

9. Quantitative analysis

10. VPD pretreatment

      FEATURES OF VPD-TXRF

      OUTLINE OF VPD PRETREATMENT

11. PRETREATMENT PROCESS

1. General

The approximately 50A penetration depth of primary X-rays in the TXRF method allows measurements to be made with substantially lower background intensity. This results in higher sensitivity for the elements on a flat, preferably mirror-finish surface, such as silicon wafers or the thin film surface on silicon wafers.

The comparison between Total Reflection X-ray Fluorescence analysis (hereafter referred to as TXRF) and the conventional X-ray Fluorescence (hereafter referred to as XRF) method is described.

2. XRF

 

Optical System of XRF

Fig. 2-1 Optical System of XRF (Energy dispersive system)

In XRF, the X-rays irradiate the surface of the sample material, and travel into the material. The elastic and inelastic X-ray scattering and fluorescent X-rays corresponding to the elements contained in the sample are observed in its spectra. The fluorescent X-rays from each element in the sample have an energy (or wavelength) which is unique to that element, and an intensity (the rate of X-ray photon emission) which is proportional to the concentration of the element in the sample.

An analysis of energy and intensity of the fluorescent X-rays reveals the composition of the material, that is, what elements are present (qualitative analysis) and what amount of each element is present (quantitative analysis) in the sample material.

Notes:

  1. Fluorescent X-rays : When an inner shell electron is excited and ejected from an atom as a photo-electron, the vacant orbital will be filled by an outer shell electron having a higher energy state. An Auger electron or fluorescent X-ray will be emitted which has an energy equal to the level difference between inner and outer electron shell energies. Each atom has its own electron shell binding energy, so the emitted X-ray (fluorescent X-ray) energy is also unique to the atom.
  2. Scattering (of X-rays) : A phenomenon involving a change in direction of X-rays caused by collision with atoms, sometimes accompanied by loss of energy.
  3. Intensity of X-rays : The number of X-ray photons measured in a unit time and expressed in counts per second (cps).

3. Total reflection of X-rays

When X-rays collide with a sample at a sufficiently small incident angle, they are totally reflected at the surface.

To illustrate this phenomenon, consider an x-ray beam incident on a sample surface at an angle of several degrees. The beam is refracted and penetrates the sample surface at an angle somewhat less than the incident angle.

As the incident angle between the X-rays and sample surface is decreased, the X-ray direction in the sample moves nearer to the plane of the sample surface, because the X-ray refractive index is slightly lower than 1. The angle where the incident X-ray cannot enter the object, and all rays are reflected, is called the critical angle for total X-ray reflection.

When the angle of incidence is greater than the critical angle ϕ, the X-rays travel into the sample to a depth of between 10 and 1000 µm (10,000A to 1,000,000A).

Incident Angle

Fig. 3-1 Incident angle

When the angle incidence is smaller than the critical angle ϕ, more than 90% of X-rays are reflected. The percentage of reflection depends on surface roughness. The 1/e depth of refracted X-ray penetration is about 50A.

 

Total reflection of an Incident X-ray

Fig. 3-2 Total Reflection of an Incident X-ray

The critical angle is calculated as below :

In the case of W-Lb X-rays of 9.67 keV energy used for measurement of a Si wafer of density 2.33 g/cm3, for example, the critical angle will be about 0.18 degrees.

4. TXRF

Optical Systems TXRF

Fig. 4-1 Optical System of TXRF

In TXRF analysis, a monochromatic X-ray beam irradiates a mirror finished-sample such as a Si wafer with incident angle set to 0.05 to 0.12 degrees. The incident X-rays are totally reflected, and scattered X-rays (which cause the background of a measurement) are minimized. As a result, the intensities of fluorescent X-rays arising from surface contaminant elements are enhanced, thus improving signal to background ratios (S/B).

The double excitation by primary and reflected X-rays, combined with the lower background, results in superior sensitivity for trace impurities adhering to the mirror-finish sample surface.

5. X-ray source

When an accelerated electron is involved in a collision (or a sudden stop) with a target, the electron emits part of its energy in the form of an X-ray. A few 10’s of kV electron acceleration voltage is required in XRF.

Fluorescence analysis to excite the inner shell electron of atoms in a sample. An X-ray tube is fundamentally a two-electrode vacuum tube. One electrode is a filament (cathode) to generate electrons which are accelerated by the applied voltage, and the other is a target (anode) to which the accelerated electrons are directed and with which they ultimately collide.

In TXRF analysis, a rotating anode X-ray tube operating at high power (e.g. 9 kW) and with high-efficiency target cooling is usually required for microanalysis.

Construction of x-ray tube

Fig. 5-1 Construction of X-ray Tube (open bulb)

6. X-ray monochromator

 X-rays produced by the collision of electrons on the target include the characteristic X-rays 1) of the target material and the white (or continuum) X-rays 1). In TXRF analysis, monochromatic X-rays are used, because the scattering of white X-rays produces a higher background over the entire measured energy range. The characteristic X-rays of the target material are usually used, because of their higher intensity.

To produce monochromatic X-rays, a synthetic multi-layer crystal is used because of its excellent reflection (more than 70 % expected).

X-ray MonochromationFig. 6-1 X-ray Monochromation

Note :

1) Characteristic X-rays and white X-rays :

An electron emits its energy in the form of X-rays (electromagnetic energy) when it collides with a target. The electron is subjected to repeated collisions and energy is lost at each step, generating X-rays over a continuous range of energies (or wavelengths), commonly known as an X-ray continuum. In contrast, when the incident electron excites an atom of the target element, a characteristic X-ray is emitted. The energy of the characteristic X-ray is the same as that of the fluorescent X-ray. The X-ray spectrum from an X-ray tube has the following energy distribution:

Characteristics X-ray and continuous x-ray

Fig. 6-2 Characteristic X-rays and Continuous X-rays

7. Silicon Drifted Detector (SDD)

A Silicon Drifted Detector (SDD) is an X-ray detector which utilizes the ionization effect of Silicon crystal by X-rays. An electron-hole (electric charge) pair is created when an X-ray photon goes into the Si crystal. The number of ionic pairs is calculated by the following formula:

TXRF Formula details

The total electric charge (q) will be:TXRF Formula results

When high voltage (several hundred volts) is applied to the electrode on the Si crystal, electric charge generated is collected to the electrode. As shown by formulas (1) and (2), the total electric charge collected to the electrode is proportional to the energy of X-ray photon. The electric charge integrated is converted into a voltage pulse and its height is also proportional to the energy of the X-ray photon. Pulse height is converted into digital value by means of analog-digital conversion electronics. A multi-channel analyzer (MCA) adds one count to the channel corresponding with the digital value. The above process is repeated for each X-ray photon, and then a spectra chart is created of the number of counts stored in the channel No. corresponding with the digital value versus energy covering whole energy range.

The crystal of the SDD and transistors of the preamplifier are refrigerated at around40°C with a Peltier device to remove thermal noise.

8. Qualitative analysis

Computer analysis of the stored energy spectrum is employed to locate and identify spectral lines (peaks) corresponding to elements in the measured sample. Known intensity ratios for major and minor lines of each identified element are used to resolve less intense peaks, which are overlapped and obscured by adjacent strong peaks.

The following table shows typical elements and their fluorescent X-ray energies. An energy table covering all elements is available in the TXRF-V310 analytical software.

Element chart

 

Spectral peak search and peak identification are performed automatically in TXRF Routine Analysis and Auto Analysis modes while they can be applied manually in Manual Measurement/Data Processing mode.

A spectrum chart reported by Automatic Analysis is shown below.

Spectral Chart

Fig. 8-1 Spectral Chart (by Auto Analysis)

9. Quantitative analysis

Concentration (Quantitative analysis result, atoms/cm2) is calculated using a calibration curve based on the relationship between the intensity of fluorescent X-rays and the chemically analyzed concentrations of standard samples.

In XRF analysis, the “Fundamental Parameter” data correction method is often applied to correct for absorption and enhancement (secondary excitation) effects on a measured fluorescent X-ray intensity by variable concentrations of elements in the sample matrix. These so-called “matrix effects” are absent in TXRF owing to the thinness of the analyzed contaminant layer on the sample surface. In effect, the analyzed layer is “infinitely thin” and hence there is a simple linear relationship between fluorescent X-ray intensity and the concentration of an element in the analyzed layer. Furthermore, the sensitivity ratio of each element can be calculated by the FP method, which enables the analyst to apply the calibration curve for one element to the analysis of another.

For example, the calibration curve for Ni contamination on Si wafer can be applied to all other elements detected in the spectrum by using relative sensitivity factors 2).

                               C (atoms / cm2) =  k x b x l + c   (3)

where

C        : elemental concentration

b, c     : calibration constants of e.g. Ni   b=slope, c=intercept=0 in most cases

k        : sensitivity ratio of analyzed element to e.g. Ni

 

Calibartion Curve

Fig. 9-1 Calibration Curve

The instrument is usually pre-calibrated before shipment from the factory. The calibration curves are based on the standard samples supplied by Rigaku with the equipment. These can be renewed by the user as required, or supplemented by additional standards. A new calibration for any element is easily added to the system.

 Notes : 1) The X-ray intensity determined from the integrated ±2s area of an observed peak.

2) Factors for each element’s sensitivity ratio to Ni are available in the TXRF software. To determine any quantitative value for elements other than Ni, determine the analytical curve constant by multiplying the relative sensitivity factor of the relevant element (i) by the calibration curve constant (slope) for Ni i.e. ki x bNi in equation (3).

10. VPD pretreatment

FEATURES OF VPD-TXRF

The VPD (Vapor Phase Decomposition) method is one of the wafer pretreatment methods for TXRF measurement. The hydrofluoric acid droplet recovers small amount of metal contamination on the very surface of the wafer. By applying the VPD method to a total reflection X-ray fluorescence spectrometer (TXRF), metal contaminants all over a wafer surface can be condensed and collected to one point and the sensitivity of analysis can be improved by approximately two orders of magnitude. In the ordinary TXRF, lower limits of detection for transition metals such as Fe, Ni, Cu and Zn are 109 atoms/cm2 or so. In the VPD-TXRF, which applies the VPD method to the pretreatment, those lower limits of detection are improved to approximately 107 atoms/cm2 and the management of contaminants at 108 atoms/cm2 or so is possible.

In case of the VPD-TXRF, a spectrometer can handle the contamination level of 1010 atoms/cm2 or less.

OUTLINE OF VPD PRETREATMENT

The VPD pretreatment process consists of three processes: vapor phase decomposition, contaminant recovery and drying.

  • Vapor Phase Decomposition: The vapor phase decomposition of a wafer surface is carried out using hydrogen fluoride. A wafer is exposed to a hydrogen fluoride atmosphere to dissolve contaminants together with the oxide film on the wafer surface. Contaminants are dissolved in hydrogen fluoride and stay on the wafer surface as minute drops. The wafer surface after the vapor phase decomposition is hydrophobic and recovery solution is easy to retain in the following contaminant recovery process. If an oxide film exists, recovery solution moves of itself and difficult to retain.
  • Contaminant Recovery: Contaminants are collected using a small quantity (100µL) of hydrofluoric acid solution (recovery solution). A small quantity of recovery solution is dripped on the wafer after the vapor phase decomposition and moved on the wafer surface while pressing the recovery solution from the upside using a rod. By moving the recovery solution all over the wafer surface, contaminants on it can be dissolved and collected.
  • Drying: The recovery solution is dried. Since the TXRF spectrometer cannot analyze liquid, the heating and drying process for the recovery solution is necessary. The photograph of a typical drying trace and a TXRF analysis result are shown below. After drying under suitable conditions, the size of a drying trace is 2mm or less. The dependence of X-ray intensity on an incident angle also shows a pattern characteristic of particle contamination, as shown in Fig. 10-2.

Driving Trace

Fig. 10-1. Photo of Drying Trace

Dependence X-ray Intensity of Incident angle

Fig. 10-2 Dependence of X-Ray Intensity on Incident Angle

11. PRETREATMENT PROCESS

  1. An oxide film on a wafer surface is removed using HF vapor.

Vapor phase

2. Recovery solution is dripped on the wafer surface from which the oxide film has been removed.

Recovery Solution

3. The recovery solution held in a nozzle scans the wafer surface to its center while collecting contaminants.

nozzle hold

4. After the scanning, the drop is dried using a halogen lamp.

recovery solution

 

 

 

VIEW RESOURCES HERE

RELATED PRODUCTS

TXRF 3800e

TXRF 3800e

WAFER SURFACE CONTAMINATION METROLOGY BY TXRF FOR ≤200 mm WAFERS

Explore TXRF 3800e

RELATED APPLICATIONS

Mechanochemical effect of caffeine by DSC

Mechanochemical effect of caffeine by DSC

EXPLORE

TXRF 3760

TXRF 3760

WAFER SURFACE CONTAMINATION METROLOGY BY TXRF FOR ≤200 mm WAFERS

Explore TXRF 3760

TXRF310Fab

TXRF 310Fab

WAFER SURFACE CONTAMINATION METROLOGY BY TXRF FOR ≤300 mm WAFERS

Explore TXRF310Fab

TXRF-V310

Image - TXRF-V310 - 2018-12-18 800x610

WAFER SURFACE CONTAMINATION METROLOGY BY VPD-TXRF FOR  ≤300 mm WAFERS

Explore TXRF-V310