TXRF300/3750

Application Manual

Rigaku Industrial Corp.

Application Lab. Section2
The purpose of this Manual
　 TXRF has been widely used in environmental analyses such as water quality analysis.
TXRF is also used in electronic device fields such as Semiconductor, TFT displays and
Magnetic disks. In fact, 90% of TXRF applications is for contamination analysis in
semiconductor processes. For this reason, TXRF applications have been developed for
semiconductor manufacturing process. This application manual is going to describe
measuring methods and key points of typical thin film measurements in the semiconductor
process. This manual will also including hints for optimization of the measuring conditions.
This manual will be most helpful to the TXRF operator who will be using TXRF in the
semiconductor industry. 　

TXRF-300 TXRF-V300 TXRF-3750

Contents
1. Excitation X-rays and Measurable Elements
Outline of this Chapter
1.1. Relationship between excitation X-rays and measurable elements
1.2. Detection limit and quantification limit
2. Calibration Curve and Equipment Standardization
Outline
2.1. Standard sample and calibration curve
2.2. Equipment standardization sample
2.3. Frequency of updating standardization and calibration curve
3. Measuring Condition
Outline of this Chapter
3.1. Relationship of Measuring condition, Group condition, and Recipe.
3.2. Modify measurement time
3.3. Modify incident angle
3.4. Modify other condition (Incident direction and attenuator)
3.5. Incident angle of TXRF
3.6. Incident angle adjustment sequence
Contents
4. Thin Film Measurement Information
4.1. Bare silicon wafer
4.2. Silicon based thin films (SiO2, Si3N4, SiON, SiOF, Poly-Si, SiC)
4.3. Barrier metal (TiN, TaN)
4.4. Interconnect film (Al, Al-Si, Al-Si-Cu, Cu)
4.5. Low-k (Organic type, Silicon type)
4.6. 強誘電体 , 高誘電体
4.7. High-k gate (HfO2, Al2O3, Al2O3, ZrO2)
4.8. Silicide film (Ti, Co, Ni, W)
4.9. Measuring condition summarized table

5. How to Setup the New thin film on TXRF Software

1. Excitation X-rays and Measurable
Elements
Outline of this chapter
　 The relationship between excitation X-rays and measurable elements will be described in this chapter.
For high sensitivity analysis, excitation X-rays must be chosen following the conditions listed below. This is
to optimize the use of the monochromating crystal used in TXRF.
* The energy of the excitation X-ray should be higher than the X-ray absorption edge energy of
the elements being measured.
* The energy of the excitation X-ray should be as close to the X-ray absorption edge energy of
the elements being measures as possible.
For example, though the W-M (1.77keV) X-ray is able to measure Na, Mg and Al, W-M X-ray can not
measure heavier element than these light elements because they do not have enough energy to cause
fluorescence. On the other hand, though H.E. (24keV) X-rays excite heavy elements such as W, As, Zr and
Mo, H.E. x-rays do not have enough sensitivity for transition metals such as Fe and Cu because H.E.(24keV)
energy x-rays have so much energy that they pass right through these transition elements without interacting
with them. W-Lb therefore should be used for transition metal analysis.
Optimal detection limits can be achieved by selecting the most suitable excitation source. TXRF equipment
has a built-in function to select the optimal excitation source for a given element.
1.1. 　 Relationship between excitation X-rays and measurable elements

(1).W-M (Red Boxes) The best source for light elements on the periodic table are measured by W-M.
(2).W-L(Green Boxes) The best source for most elements on the periodic table are measured by W-L
(3).H.E. (24keV) (Blue Boxes)The best source for heavy elements on the periodic table are measured by H.E.
X-rays
H He

Li Be B C N O F Ne

Na Mg Al Si P S Cl Ar

K Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se Br Kr

Rb Sr Y Zr Nb Mo Tc Ru Rh Pd Ag Cd In Sn Sb Te I Xe

Cs Ba La Hf Ta W Re Os Ir Pt Au Hg Tl Pb Bi Po At Rn

Fr Ra Ac

Ｌａ Ｃｅ Ｐｒ Ｎｄ Ｐｍ Ｓｍ Ｅｕ Ｇｄ Ｔｂ Ｄｙ Ｈｏ Ｅｒ Ｔｍ Ｙｂ Ｌｕ

Ac Th Pa U Np Pu Am Cm Bk Cf Es Fm Md No Lr

Ｗ ―Ｍ α Ｗ ―Ｌ β Ｈ ．Ｅ ．
1.2. 　 Detection limit and Quantification limit

The typical detection limits of TXRF 3750/300 is shown in following table.

TXRF300/3750 TXRF300W/3750W TXRF300/3750 TXRF300W/3750W
ELEMENT W-M W-L H.E. W-L Option H.E. H.E. ELEMENT W-M W-L H.E. W-L Option H.E. H.E.
24KeV W-L 24KeV 24KeV W-L 24KeV
Na 500(50) Sr 19 22
Mg 300(30) Y 18 20
Al 400(40) Zr 18 18
Si Nb 17 18
S 60 130 Hf 3 4
Cl 35 80 Ta 32 20 42
Ar 22 50 W 28 19 37
K 15 32 Pt 20 26
Ca 10 20 Au 19 25
Sc 7.0 15 Hg 18 23
Ti 5.0 10 Pb 14 19
V 3.0 7.5 Bi 14 19
Cr 2.0 5.0 Mo 18 18
Mn 1.5 4.0 Ru 23 23
Fe 1.0 2.6 Pd 22 38
Co 1.0 2.5 Ag 19 33
Ni 1.0 2.5 Cd 16 27
Cu 1.5 2.9 In 13 23
Zn 2.0 3.8 Sn 11 19
Ga 41 35 54 Sb 10 16
Ge 36 43 47 I 8 12
As 30 39 Ba 5 8
Se 26 34 La 4 7
Br 24 32 Ce 4 6

( ) value is particle type contamination , Measuring time: 1000sec

TXRF 3003750 is a rotating anode model, TXRF300W3750W is a sealed tube model
Normally, the quantification limit is from 3 to 5 times of these value.
2. Calibration Curve and
Equipment Standardization
Outline of this chapter
　 Calibration curve updating and equipment standardization will be described in this chapter. Calibration
curve updating and standardization should be carried out regularly.
Standard sample is silicon wafer which was contaminated with known amount of metal element. Usually, a
Ni contaminated standard wafer is used for TXRF standard sample because Ni is most stable element on a
silicon wafer. Also sometimes Al contaminated standard wafer is used for light element calibration curve.
A function of Calibration curve is a equation that transfer from X-ray intensity of detected element to
concentration of contaminants. The calibration curve is prepared using the relationship between the X-ray
intensity of a contaminant element and the known amount of the contaminant element.
A quantification result is influenced by a change of X-ray intensity, however it is possible to minimize the
quantification error by updating the equipment standardization and calibration curve.
 2.1. Standard sample and Calibration curve
　 A Standard sample is a wafer which was contaminated with a known amount of metal solution on a
wafer surface. There are three typical methods for making a standard sample, such as Spin coating,
IAP(Dip) and Micro droplet method.
The standard samples which are made by using these methods have unique contamination types (See
“Principle of TXRF”). A semiconductor process has various contamination types such as particles, film and
bulk types. The type of contamination causes quantitative analysis error ( up to 5 times difference).
Generally the Spin Coated type standard has been the most widely used in the TXRF world.

Basically standard samples do not require other elements except the Ni, because a Ni one point calibration
curve is used in the TXRF ( The Al standard is required in Light element analysis by W-M). 　 The standard
sample can be used a long time if it is kept in a cleanroom environment (from 2 to 3 years).
　
The calibration curve can be shown with a graph that shows the relationship between the X-ray intensity
of the contaminant element and the it’s chemical concentration. The Concentration of an unknown sample
is determined with this graph (equation is used actual).

The Ni one point calibration curve is normally used in TXRF, other elemental concentrations are
determined by Ni based on the relative sensitivity which is calculated by fundamental parameter (See
“Principe of TXRF”).
　
If the incident angle is changed, the calibration curve must be prepared at the same angle because the
calibration curve is dependand on the incident angle.
2.2. Standardization sample
　 X-ray intensity of TXRF is not always constant, it gradually changes with time. One cause of this change
in x-ray intensity is small changes in the filament emission. Also, it is caused by small changes of the
incident angle. In the case of bare silicon, the change of incident angle influences the Silicon X-ray intensity.
Therefore it is possible to standardize the TXRF by using the x-ray intensity of silicon on a standardization
sample.
The standard wafer should be kept in a cleanroom environment to prevent contamination of the standard.
Also, the wafer used as a Silicon standard should be replaced every six months. This is to prevent damaging
the standardization wafer making it unreliable as a standard. Bare silicon wafers purchased form a wafer
vendor are most suitable for use as a standard, however it is important that the same grade of wafer always
be purchase. The surface of the wafer is dependant on the grade of wafer purchased, an the surface
condition can greatly affect the measured Si x-ray intensity. If the same grade of wafer cannot be purchased,
the “Standard Value” must be updated with the new grade of wafer. 　　　
2.3. Frequency of updating the calibration curve and standardization
　 Rigaku recommends that standardization and updating of the calibration curve be carried out once a
week. If TXRF has not been used in a long time, standardization and calibration update should be carried
out before use. Also when the exchanging the filament, exchanging the target or tube maintenance is done,
standardization and calibration updating is necessary.
Good stability results can be achieved from the TXRF if standardization and calibration are performed
weekly.
3.Measuring condition
Outline of this chapter
　 This chapter will describe modification of the measuring condition. Details of the measuring condition such
as Measuring time, Incident angle, Incident direction and Exciting X-rays are entered into the measuring
condition file used by the TXRF software.
Measuring Time
　 Measuring time influences the detection limit directly. Though the detection limit will be improved with
extending measurement time, it should be selected with consideration of throughput, because the improving
ratio follows the square root rules (as explained later).

Incident Angle
　 The incident angle influences the penetration depth of the excitation X-ray. Using this effect, contamination
types can be determined with angle scanning.

Incident direction
　 The function of incident direction is to avoid diffraction of the excitation X-rays by the substrate and to
keep the background intensity to a minimum. In the case of W-Lb excitation x-rays, 35 degrees to the
Flat/Notch is the most suitable direction on <100> cut silicon wafers because W-Lb is not diffracted at 35
degrees.

Attenuator
　 Usually an attenuator is not used. However if a measurement is carried out beyond the critical angle of total
reflection, the attenuator should be used.
3.1. Relations between Measuring condition, Group condition and Recipe

Measurement Recipe
(Set up in the screen of Routine Analysis or Auto Analysis)

Group Condition

Measuring
(Quantitative Unit)
Condition Drople
(Output format) Measuring
(Excitation Source) Substrate SWEEPING VPD t
Calibration (Element line set) Point
Search
(Incident Angle)
Curve)
(Measuring Time) (Overlapping
(Azimuth) correction)
(Attenuator)

　 Relationship of each setting file is shown above. The Group condition , Mapping points, Substrate info
and Droplet search conditions are included in the Measuring recipe which is set up from the Routine
analysis or Auto analysis menus. In these files, the Group condition has the strongest effect on the
measurement because it includes the Measuring conditions and the Calibration curve.
3.2. Modify measuring time
　 The measuring time should be chosen with consideration of the minimum detection limits for
contamination control in the manufacturing process, while providing maximum throughput. 500sec is the
most commonly used time, but 300sec or 100sec can be used if require high throughput. The Measuring
time is most factor which most directly effects the detection limit, as shown in following equation.

BG  T C ( Ni ) 1
LLD ( E10atoms / cm2)  3   
T I ( Ni ) R

BG: Background intensity(cps)

T:Measuring time(sec)
k ： Calibration curve constant
　　 It is understanding from the above equation, that the detection limit is affected by the square root of the
measuring time. For example, measuring time is doubled, the detection limit will be improved by 1/2.
Also, if the measuring time is reduced by half, the detection limit will decrease by 2.
It is not necessary to update the calibration curve when changing the incident angle because the unit of X-
ray intensity is cps (counts per second). A sequence of measuring time changing is shown on the next page
video file. (If the file is not shown, run “*.avi”file directly from the CD-ROM.)
In case of modifying name to original condition
(Movie file : Time1.avi)
In case of creating the new condition
(Movie file name: Time2.avi)
3.3. Modify incident angle

　 Some thin films cannot be measured under the standard incident angle such as 0.09 degrees using W-Lb.
In such a case the incident angle should be changed. If the incident angle is changed a calibration curve must
be prepared with same angle to be used on the measured sample. A detailed description of the effect of
incident angle is written in this chapter. The sequence to modify of the incident angle is shown on next page.
In case of creating the new condition
(Movie file name: Angle1.avi)
3.4. Modify other condition (Incident direction and Attenuator)

　 * Modification of incident direction

　 35 degrees is default value for the angle of incidence. It is optimized for the <100> cut Si wafer most
commonly used. If a different type of wafer such as <111> cut Si is measured, a step scan should be carried out
at the wafer center prior to measurement. The angle which shows the minimum intensity of scattering should
be used as the angle of incidence for all scans.
　
* Modification attenuator
　 An attenuator should be used if strong fluorescence X-rays enter the SSD. The attenuator should not used if
the strong X-ray intensity is caused by the main element of the thin film In this case the incident angle should
be reduced for the measurement because the attenuator lowers the incident X-ray intensity by 1/50, which
lowers the sensitivity of the TXRF. The attenuator is normally used for measurements at angles above the
critical angle. If the attenuator is used in a measurement, the TXRF software will multiply the x-ray intensity by
50 times automatically, therefore it is not necessary to update the calibration curve.
3.5. Incident angle of TXRF
*Mechanical zero degree and Optical zero degree
When the TXRF sets the incident angle for a scan, it is done by tilting the sample stage. The sample stage
is designed so that the x-ray beam is not perfectly parallel to the sample when the stage is set to mechanical
zero. The optical zero degree is when the sample stage is parallel to the x-ray beam. The optical zero and
the mechanical zero are not always at the same angle due to the interaction of the monochromator. The
optical zero is determined by performing a phi-scan when the TXRF is first installed. The measurement
angle using in normal operations is determined based on the optical zero degree location. Updating of the
optical zero degree location is not required because it is done automatically when a TXRF standardization
is performed. 　

* Relationship between incident angle and X-ray intensity of thin film main element
The surface of a wafer or thin film will have a uniform number of atoms. This makes it possible to
accurately determine the angle adjustment for a sample by measuring the x-ray intensity of the main
element of the sample surface. The angle adjustment function of TXRF3750/300 determines the optimum
angle by comparing the measured intensity at a local position with the reference intensity of the sample. If
these values do not match, the incident angle is adjusted until the local and reference intensities are equal.
　

* Relationship between incident angle and penetration depth

The depth of penetration of the incident x-rays is directly affected by the incident angle. If the angle is
increased, the penetration depth increases, if the angle is reduced, the penetration depth is also reduced.
The theoretical penetration depth for the incident beam is shown in “Principle of TXRF”. When
performing a calibration, the incident angle should be the same as the angle used for a normal
measurement because changes in the angle of incidence affect the dead time of the SSD, which in turn
affects the measurement. 　
3.6. Incident angle adjustment sequence
Measurement * Reference intensity measurement
start Angle adjustment is carried out using the intensity of the
A warp correction main element which comprises the thin film of a sample.
for wafer center The reference intensity is determined at the wafer center,
at that time a warp correction for wafer center position is
Angle setup executed because accurate angle is required. The thin film
type must be entered accurately because this information
Pre measurement of Main element is entered in the “substrate” item in the measurement
(Reference intensity)
intensity at wafer center recipe used when the incident angle is adjusted.
Move to measuring * Angle adjustment
position
The wafer is moved to its measuring position and the
Setup incident angle incident angle setup is performed at this location
(Initial angle)
mechanically mechanically. The initial intensity is measured and
Main element intensity compared to the reference intensity. If the intensities do
measurement not match, the incident angle will then be increased or
decreased. This step will be repeated until the local and
Angle adjustment
reference intensities converge. 　
(Increase or decrease)

Intensity
comparison * main measurement
After the angle is selected, the main measurement will
Main measurement start commence.

This procedure is performed at every measurement

End
location and for each excitation beam used.
4. Thin Film Measurement
Outline of this chapter
TXRF first entered the market in the late 1980’s. At that time, most measurement were performed on bare
silicon and on oxide films because metal contamination analysis was important to FOEL (Front End of Line)
processes and for wafer vendors. Since then, metal contamination analysis needs have become much more
diverse as processes evolved in semiconductor manufacturing. The capabilities of TXRF have been extended
to meet the requirements of many of these processes, including BEOL (Back End of Line) processes. In
addition, modifications have been made to accommodate thin films now being used by many semiconductor
manufacturing equipment vendors.
These advances have made flexibility in TXRF important so that it can be applied to all types of thin films. It
is important to note that if the measured thin film is changed, the TXRF must be optimized for the new film.

This chapter will describe how to optimize TXRF for measuring thin films.
4.1. 　 Bare silicon wafer
(1) Details on measurement conditions
* Diffraction and Incident direction 　
When a <100> cut silicon wafer is measured using the W-L source, if W-Lb is radiated from a parallel or
right angle direction from the Flat/Notch ,W-L x-ray will be diffracted to the SSD. If a measurement is
carried out at these angles, a large escape peak from the W-L appears. This escape peak will interfere
with Cu analysis. To prevent this from occurring, the direction of the exciting x-ray must be kept at 35
degrees to the Flat/Notch. 35 degrees is a safe angle which avoids diffraction.

* Sum peak of Silicon K- 　

　 When a strong silicon K-a line appears on spectrum chart, it also produces a Sum peak for Si-Ka at
3.48keV (twice the normal value of 1.74 keV). This sum peak exists between K (3.31keV) and Ca
(3.69keV), so it will interfere with detection of trace amounts of K and Ca. This must be taken into
account when when a trace amount of K and Ca (E10atoms/cm2 levels) have been detected.

Escape peak of W-
Sum peak of Si- L （ 7.93keV)
K (3.48keV)
e.g. : Clean bare silicon wafer
4.1. 　 Bare silicon wafer

(2) Measuring condition

　 Standard measuring condition which are installed to TXRF PC as a default can be applied to measure
bare silicon wafer. Standard measuring conditions are shown as following. (B1, B2, B3 group condition)
Substrate ； Si-Wafer ）
Incident Angle ;
W-M (Beam1) = 0.50degree (0.1degree)
W-Leam0.09degree (0.01degree)
H.E.(Beam3) = 0.05degree(0.01degree)
　 Measuring time ;
All beams set to 500sec 　
（ Measuring time can be changed based on throughput and by contamination control level.)
　 A detection limit is improved with extending measuring time. For example, 1.4E9atoms/cm2 detection
limit is obtained at 500sec measurement, if measuring time extend to 1000sec, detection limit is improved e
as following. 1.4÷√ （ 1000/500 ）＝ 1.0E9atoms/cm2

　 Incident direction ;
35degree 　　 ( 35degree is safe angle for avoiding diffraction at <100> cut wafer)
4.2. 　 Silicon based thin films
　 Silicon based thin films such as SiO2, Si3N4, SiON, SiOF, Poly-Si and SiC are the most frequently
measured after bare silicon wafer. Standard measuring condition can be applied for these silicon based thin
films.
(1) 　 Attention on measurement
* Shape of background pattern for light element analysis by W-M line 　
　 In the case of light element analysis such as Na, Mg and Al measured by W-M line, the background
pattern is changed based on the thickness of thin film and the kind of film being measured. A background
pattern optimized for each kind of film should be prepared before taking a measurement. A group
condition which has the new background pattern stored in it should be prepared for application to the
new films. A thickness of greater than 50 nm will not affect the background pattern due to saturation.

•About silicon X-ray intensity

Si x-ray intensity is stronger when measuring Poly-Si films when compared to bare Si. This is usually
due to the crystal structure of Poly-Si as well as wafer warp. This can affect the detection of trace
amounts of K or Ca.

When SiO2 or Si3N4 is measured, the Si intensity will be much weaker than bare Si because there is 50-
70% less Si in the sample.
(2) 　 Measuring condition
Substrate ； Si-Wafer
Standard condition can be applied to measurement. 　（ can apply B1,B2,B3 group conditions)
4.3. 　 Barrier metal (TiN film)
(1) Attention on measurement
•About incident angle
When a TiN film is measured using the W-L source, a strong Ti-Ka peak is produced. This will increase
the dead time of the SSD. For this reason, the incident angle should be reduced to around 0.04 degrees. A
new calibration curve should also be prepared using Ti as the substrate.

•Sum peak and escape peak

An escape peak for Ti-Ka (at 4.51 keV) appears at 2.77 keV. Though this is close to the Cl- Ka (2.62 keV)
and Ar-Ka (2.96 keV), it can be resolved by peak separation with the TXRF software. A sum peak for Ti-Ka
(at 6.25 keV) appears at 9.02 keV but does not interfere with any elements.
If the TiN film is very thin, the sum peak of Ti-Ka (4.51 keV) and Si-Ka (1.74 keV) will appear at 6.25 keV.
Though this would appear to interfere with the Fe-Ka peak at 6.40 keV, it can be resolved using the TXRF
software.

* Background rising on 2keV-3keV (Low energy side from Ti main peak)

Background intensity from 2keV to 4keV will be increased due to the strong Ti-K signal. Therefore, the
detection limit of K(3.31keV) and Ca (3.69keV) will be lowered.
4.3. 　 Barrier metal(TiN film)
e.g. ： Contamination of Ca, Fe and Cu on TiN film
Fe-K
Ti-K Ti-K
Ti-K
Ti-K Sum peak
Esc peak Ti-K
Esc peak
Ca-K

Cu-K

Strong background
is produced in
around this area
(2) 　 Measuring condition
Substrate ； Ti
Incident angle ；
W-M 0.20degree(0.1degree)
W-L 0.04degree(0.01degree)
H.E. 0.04degree(0.01degree)
Incident direction ；
35degree ( Diffraction line not appears)
4.3. 　 Barrier metal (TaN film)
(1) Attention on measurement
*About incident angle
　 Strong Ta-La which is excited from TaN film appears when TaN film measured by W-Lline. If a TaN
film is measured with standard angle (0.09degree), the dead time of the SSD will be almost 50%. Of
course , TaN film can be measured at this dead time, but measurement throughput becomes lower. If high
throughput is required for TaN film measurement, the incident angle should be reduced to around 0.07
degrees. The standard sample must also be measured at 0.07degree because the calibration curve should
have the same angle.
The “Ta” must be chosen to the “substrate” which is shown in measurement recipe program when
measure the TiN film.

* The Escape peak of the Ta-L

Escape peak of Ta-L(8.15keV) emerge at 6.41keV. This escape peak overlaps Fe-K(6.40keV)
perfectly. Though in this case there is a method of using Fe-kb for analysis line, the x-ray intensity of Fe-
Kb is weak. Also Ta-Ll(7.17keV) will interfere with the Fe-ksignal. Overlap correction by calculation
is applied for solving this problem. If overlap-correction is used for this problem, the overlap factor
should be calculated with the following sequence in advance.
(1) Plot the relationship of X-ray intensity between Ta-L and Ta-L escape on a graph from Angle
scanning measurement (Angle scan point are 3or 4point enough).
(2) Calculate a slope of graph. (this slope is the same as the intensity ratio between Ta-L and escape)
(3) Set the intensity ratio which is calculated in (2) to Group condition.(*See how to overlap correction)
(4) When Group condition which is made in (3) is used, overlap correction of Ta escape is carried out
automatically.
4.3. 　 Barrier metal (TaN film)

(1) Attention on measurement

*Cu analysis on TaN film
　 The Cu-K 8.04keV) which is used for analysis line is overlapped by strong Ta-La peak. Therefore the Cu-
K(8.90keV) line is used for analysis line to avoid this situation. When Cu-K is used , the detection
limit will be changed to almost 5e10atoms/cm2 because the s-ray intensity of Cu-K is 1/10 of Cu-Ka
and background is raised with strong Ta-L. The modification of analysis line is possible in “Group
Condition”.(see the modification of analysis line)
* About Ta-M series line.
　 A lot of Ta-M series line appears around 1.77keV.Therefore Si, P, and S analysis is impossible.
4.3. 　 Barrier metal (TaN film)
　 e.g. ： Cu contamination on TaN film 　

Ta-L
Ta--M
Ta-Ll

Cu-
Ta-M K
(2) 　 Measuring condition Ta-M Ta-L の Esc
Substrate ； Ta
Incident angle ；
W-M line 0.20degree(0.1degree) 　
Mg analysis is impossible to analyze because Ta-Mpeak (1.35keV) interferes to Mg. Also pay
attention of Al result because it is interfered by the foot of Ta-M
W-Lline 0.09degree(0.01degree)
0.05degree (0.01degree) should be recommended if throughput has high priority because dead
time is higher than normal situation.
H.E.line 0.05degree(0.01degree)
Incident direction ；
35degree ( Diffraction line does not appear)
4.4. 　 Interconnect film (Al, Al-Si film)
(1) Attention on measurement
* About incident angle
　 Standard incident angle ( same as Bare Silicon) can be used for measurement except for the W-M
line. “Al” must be chosen as the “substrate” which is shown in measurement recipe program when
measuring these films.
* Light element analysis use by W-M line
　 It is impossible to measure light element use by W-M line because very a strong Al appears from
the substrate.
( It is impossible to measure even if the incident angle is reduced)

(2) 　 Measuring condition

Substrate ； Al
Incident angle ；
W-M line unable to Measure
W-Lline 0.09degree (0.01degree)
H.E. line 0.05degree (0.01degree)
Incident direction ；
35degree ( Diffraction line does not appears)
4.4. 　 Interconnect film(Al-Si-Cu film)
(1) Attention on measurement
* About incident angle
Though a standard incident angle can be used for Al-Si-Cu film, sometimes the dead time increases due to
the Cu intensity which is dependant on concentration. An increased dead time will affect the throughput of
the TXRF. In such a case a incident angle should be reduced to around 0.07 degree. “Al” must be chosen as
the “substrate” as shown in measurement recipe program when measure the these film.
* Light element analysis use by W-M line
　 It is impossible to measure light element use by W-M line because a very strong Al appears from substrate.
( It is impossible to measure even if the incident angle is reduced)
* Escape peak of Cu-Ka
　 Though an escape peak of Cu-K appears at 6.30keV, it is possible to avoid with peak separation.
However trace Cu contamination from 1 to 5E10atoms/cm2 is difficult to
measure.
(2) Measuring condition
Substrate ； Al
Incident angle ；
W-M line unable to Measure
W-Lline 0.09degree (0.01degree) If dead time is high, reduce incident angle to around 0.07
H.E. line 0.05degree (0.01degree)
Incident direction ；
35degree ( Diffraction line does not appear)
4.4. 　 Interconnect film (Cu film)
(1) Attention on measurement
* About incident angle
　 The incident angle for W-L must be reduced to 0.03 degrees because a strong Cu X-ray intensity
emerges from the Cu film. The “Cu” must be chosen as the “substrate” which is shown in the
measurement recipe program when measuring the Cu film.

* Light element analysis use by W-M line

　 It is impossible to measure Na and Mg because Cu-L series line emerges near the Na and Mg. But
Aluminum analysis is possible.

* Escape peak of Cu-Ka

　 Though an escape peak of Cu-K appears at 6.30keV, it is possible to avoid with peak separation.
However trace Cu contamination from 1 to 5E10atoms/cm2 is difficult to measure.
4.4. 　 Interconnect film (Cu film)

e.g. : Spectrum chart of Cu film 　 e.g. : Al contamination after Cu-CMP of Cu film 　

(E12 levels) (Measured by W-M line) 　
(Measured by W-L)
Cu-K Cu--L W-M
Cu-K
Excitation

Cu-K
Al
Cu-L Si-K Esc

(2) Measuring condition

Substrate ； Cu
Incident angle ；
W-M line 0.20degree (0.1degree)
W-L line 0.02degree (0.01degree)
H.E.line 0.03degree (0.01degree)
Incident direction ；
35degree ( Diffraction line not appears)
4.5. 　 Low-k (Organic porous type film)
(1) Attention on measurement
* Critical angle of total reflection
The density of typical organic porous type Low-k film(such as a Silk) is 1.16g/cm3. In this case, a critical
angle of the W-L (9.67keV) for total reflection is shown at 0.13 degrees. The H.E.(24keV) is around 0.05
degree. Also W-M(1.78keV) is 0.70 degrees. The low density of this film will cause a low critical angle.

* X-ray intensity of scattering

　 Scattering X-ray intensity shows a tendency of increasing when a organic sample is measured. Also the
shape of scattering X-ray widens at low energy side ( increasing Compton scattering).

* Incident angle
　 It is difficult to decide a optimum angle because a film density is dependant on porous size. Therefore
first measure with the standard angle( Normally the standard angle can be applied to this film). As a result,
if high dead-time or strong background intensity are shown in a spectrum, then reduce the incident angle.
“Si-wafer” must be chosen as the “substrate” which is shown in measurement recipe program when
measure the this low-k film.
 4.5. 　 Low-k (Organic porous type film)
e.g. : Organic porous type low-k film

Strong scattering of W-L

Scattering X-ray wider to

low energy side

Cu

(2) Measuring condition

Substrate ； Si-Wafer
Incident angle ；（ Make sure with standard condition (B1, B2,B3)in advance.)
W-M line 0.50degree (0.10degree) Recommends 0.20-0.30degree if background intensity is high
W-L line 0.09degree (0.01degree) Recommends 0.05-0.07degree if background intensity is high
H.E. line 0.05degree (0.01degree) Recommends 0.03-0.04degree if background intensity is high
Incident direction ；
35degree ( Diffraction line does not appears)
4.5. 　 Low-k (Silicon based low-k film (Carbon contained or SiOF)

(1) Attention on measurement

* Incident angle
　 Carbon contained SiO2 film or porous SiO2 film and SiOF film are included in this group. Theses type
of low-k film can be measured with standard angle which is included B1, B2 and B3. Substrate condition
which is shown in recipe must be selected to “Si-wafer”
4.5. 　 Low-k (Silicon based low-k film (Carbon contained or SiOF))
e.g : Silicon based low-k film

Standard measuring condition can be

Si Pile up Cu applied to this type of low-k film

Ca

(2) Measuring condition

Substrate ； Si-Wafer
Incident angle ；
W-M line 0.50degree (0.10degree)
W-L line 0.09degree (0.01degree)
H.E. line 0.05degree (0.01degree)
Incident direction ；
35degree ( Diffraction line does not appears)
4.6. 　 Ferro-electric and high dielectric film

Preparing
4.7. 　 High-k gate dielectric film (HfO2 film)

(1) Attention on measurement

　 The incident angle must be reduced because a strong Hf-L(7.898keV) appears. In the case of the W-
L line, the optimum incident angle will be around 0.04 degrees. W-M is around 0.20 degrees. Also H.E.
optimum angle will be 0.04 degrees. “Hf” must be chosen as the “substrate” which is shown in the
measurement recipe program when measuring the HfO2 film.

* Hf-M line appearance

　 Hf-M line series emerge on around 1.6keV when HfO2 is measured with W-LAlso Hf-Mwill
interfere with the Mg. Therefore Mg can not be measured.

* Influence of strong Hf-L

　 Trace Ni and Cu are not able to be measured because a strong Hf-L appears near these elements.
Especially both of the Copper spectrums( Cu Ka and Kb) are interfered by Hf-L. Therefore Cu is
impossible to measure.
4.7. 　 High-k gate dielectric film (HfO2 film)
e.g. : HfO2 high-k gate film 　 by W-Lb excitation

Hf-Ll Hf-L
Hf-M Cu(8.04keV) and Ni (7.47keV) can
not be measured

Hf-M Hf-M Hf-L Esc

(2) Measuring condition

Substrate ； Hf
Incident angle ；
W-M line 0.20degree(0.10degrees)
W-Lline 0.04degree (0.01degrees)
H.E. line 0.04degree(0.01degrees)
Incident direction ；
35degrees
4.7. 　 High-k gate dielectric film (Al2O3 film)

(1) Attention on measurement

* Incident Angle
　 Standard incident angle ( same as Bare Silicon) can be used for measurement except a W-M line.
“Al” must be chosen as the “substrate” which is shown in the measurement recipe program when
measuring the these films.
* Light element analysis use by W-M line
It is impossible to measure light element use with W-M line because a very strong Al line appears
from the substrate. ( It is impossible to measure even if a incident angle is reduced)

(2) Measuring condition

Substrate ； Al
Incident angle ；
W-M line unable to Measure
W-Lline 0.09degree (0.01degrees)
H.E. line 0.05degree (0.01degrees)
Incident angle ；
35degrees
4.7. 　 High-k gate dielectric film (Al2O3+HfO2 film)
(1) Attention on measurement
* Incident angle
　 Same as HfO2 film

* Spectrum situation
　 Same as HfO2 film

*Light element analysis by W-M

　 Same as Al2O3 ( can not be measured by W-M)

(2) Measuring condition

Incident angle ；
W-M line unable to Measure
W-Lline 0.04degree (0.01degrees)
H.E. line 0.04degree (0.01degrees)
Incident direction ；
35degrees
4.7. 　 High-k gate dielectric film (ZrO2 film)

(5) Attention point of measurement of ZrO2 film 　

Preparing
(6) Measuring condition for ZrO2 film
Preparing
4.8. 　 Silicide film

(1) Attention on measurement of Ti silicide film

Same as TiN film (See barrier metal)
(2) Measuring condition of Ti silicide film
Same as TiN film (See barrier metal)
(3) Attention on measurement of Co silicide film
preparing
(4) Measuring condition of Co silicide film
preparing
(5) Attention on measurement of Ni silicide film
preparing
(6) Measuring condition of Ni silicide film
preparing
 For => to
4.9. Measuring condition summarized tables (1) “a” remove
Exciting Incident angle for incident
Film type Substrate Incident angle (deg)
source improve a throughput direction (deg)
W-M 0.5(+-0.1)
Bare Silicon Si Si-Wafer W-L 0.09(+-0.01)
H.E.(24keV) 0.05(+-0.01)
W-M 0.5(+-0.1)
SiO2 Si-Wafer W-L 0.09(+-0.01)
H.E.(24keV) 0.05(+-0.01)
W-M 0.5(+-0.1)
Si3N4 Si-Wafer W-L 0.09(+-0.01)
H.E.(24keV) 0.05(+-0.01)
W-M 0.5(+-0.1)
SiON Si-Wafer W-L 0.09(+-0.01)
H.E.(24keV) 0.05(+-0.01)
W-M 0.5(+-0.1)
Silicon based Film Poly-Si Si-Wafer W-L 0.09(+-0.01)
H.E.(24keV) 0.05(+-0.01)
35
W-M 0.5(+-0.1)
SiC Si-Wafer W-L 0.09(+-0.01)
H.E.(24keV) 0.05(+-0.01)
W-M 0.5(+-0.1)
BPSG BPSG W-L 0.09(+-0.01)
H.E.(24keV) 0.05(+-0.01)
W-M 0.5(+-0.1)
PSG PSG W-L 0.09(+-0.01)
H.E.(24keV) 0.05(+-0.01)
W-M 0.2(+-0.1)
TiN Ti W-L 0.04(+-0.01)
H.E.(24keV) 0.04(+-0.01)
Barrier metal
W-M 0.2(+-0.1)
TaN Ta W-L 0.09(+-0.01) 0.07(+-0.01)
H.E.(24keV) 0.05(+-0.01)
4.9. Measuring condition summarized table (2)
Exciting Incident angle for incident
Film type Substrate Incident angle (deg)
source improve throughput direction (deg)
W-M
Al Al W-L 0.09(+-0.01)
H.E.(24keV) 0.05(+-0.01)
W-M
Al-Si Al W-L 0.09(+-0.01)
H.E.(24keV) 0.05(+-0.01)
Inter connect Film
W-M
Al-Si-Cu Al W-L 0.09(+-0.01) 0.07(+-0.01)
H.E.(24keV) 0.05(+-0.01)
W-M 0.2(+-0.1)
Cu Cu W-L 0.03(+-0.01)
2
H.E.(24keV) 0.03(+-0.01)
W-M 0.5(+-0.1) 0.20-0.30
Organic porous
Si-Wafer W-L 0.09(+-0.01) 0.05-0.07 35
Low-K
H.E.(24keV) 0.05(+-0.01) 0.03-0.04
Low-k
W-M 0.5(+-0.1)
Silicon based Low-k Si-Wafer W-L 0.09(+-0.01)
H.E.(24keV) 0.05(+-0.01)
W-M
BST W-L
H.E.(24keV)
W-M
Ferro, High
SBT W-L
dielectric
H.E.(24keV)
W-M
PZT W-L
H.E.(24keV)
4.9. Measuring condition summarized table (3)

Exciting Incident angle for incident

Film type Substrate Incident angle (deg)
source improve throughput direction (deg)
W-M 0.2(+-0.1)
HfO2 Hf W-L 0.04(+-0.01)
H.E.(24keV) 0.04(+-0.01)
W-M
Al2O3 Al W-L 0.09(+-0.01)
H.E.(24keV) 0.05(+-0.01)
High-k gate
W-M
Al2O3+HfO2 Hf W-L 0.04(+-0.01)
H.E.(24keV) 0.04(+-0.01)
W-M
ZrO2 W-L 35
H.E.(24keV)
W-M 0.2(+-0.1)
Ti, Ti silicide Ti W-L 0.04(+-0.01)
H.E.(24keV) 0.04(+-0.01)
W-M 0.2(+-0.1)
Meal, Silicide film Co, Co silicide Co W-L 0.03(+-0.01)
H.E.(24keV) 0.03(+-0.01)
W-M 0.2(+-0.1)
Ni, Ni silicide (Ni) W-L 0.03(+-0.01)
H.E.(24keV) 0.03(+-0.01)

Meal=> Metal
5. How to setup a new thin film in the
TXRF software
5. How to setup a new thin film in the TXRF software
　 Substrate information (Main element of thin film) which is shown in following table is stored in the
TXRF software by default. It can store up to sixteen of substrates (Main element). If need be, unnecessary
substrate (Main element) should be deleted if the number of substrates becomes too large. The substrate
(main element) X-ray intensity is measured at the wafer center in advance, and the angle adjustment is
carried out using this intensity. Contents of stored items in substrate information are “name of thin
film”,”main element of substrate”, “density”,” region of integral (ROI)” and “pre measurement time”.

1 Si-Wafer
2 Al
3 Ti
4 Ta
5 W
6 PSG
7 BPSG
8 GaAs
9 Cu
10 BST
11 Ru
12 Ge
13 not yet stored
14 not yet stored
15 not yet stored
16 not yet stored
5. How to setup a new thin film in the TXRF software
　 For storing a new thin film to substrate information, substrate parameters are determined with pre
measurement. The TXRF software always carries out the angle correction, thus this situation must be released
to determine the substrate parameters for the new substrate. For releasing this situation it can do with following
procedure.

push simultaneously
Ctrl+Alt+ key on
main menu. Start
“initial setting” from
maintenance menu.

Chose “Measurement” in initial setting

Give check mark to

“Select enable”. Then push
“OK”.
Restart TXRF software
5. How to setup new thin film to TXRF software (continuation)

Start the manual measurement without angle adjustment.

At that time measuring time can be short (e.g. 50sec).
After measurement, save the spectrum chart.
Some of the thin films which generate strong x-rays from
the main element can not use the standard incident angle
because of a dead time increase.
If this situation occurs, the measurement must be
repeated with a low incident angle. The optimum incident
angle should be decided after determining the substrate
parameters.

Read the measured spectrum file with “Quantitative

analysis” for determining the substrate parameters.
Open the “Calculation” in upper part of menu, start the
“Integration” in pull down menu.
5. How to setup new thin film to TXRF software (continuation)

The regional parameters will be shown at the lower

left in a screen as X1 and X2 when the region of
spectrum peak is chosen use by a mouse (The left
figure shows a Hf-La). Take notes of these
parameters because parameters are used for
integration area in angle adjustment.

Start the Substrate in main menu. At that time substrate

information appears. Parameter setting window will be
shown in screen when the “NEW” is pushed.
5. How to setup new thin film to TXRF software (continuation)

Name of thin film 　 Set each parameters with referencing left figure.
After set these parameters, substrate information is stored
Main elements to thin
film in TXRF software.
start the measurement program and check to see new
substrate.
Density (Always
Unknown) Return TXRF software to initial situation (situation of can
not select angle adjustment mode).
* The parameters of other X-ray beam can set up with
same sequence.

X1 X2 Pre measurement time for angle

adjustment
W-M 10sec
W-Lb 10sec
H.E. (10sec –40sec)