Full text of DOI:10.1093/chromsci/41.7.367

Journal of Chromatographic Science, Vol. 41, August 2003
Measurement of As, P, and S in the Waste Gases and
Water Emitted from Semiconductor Processes by
High-Temperature Hydrogen Reduction Gas
Chromatography
Wen Ruimei, Deng Shouquan, Zhang Yafeng, and Fan Wei
Tongji University, Shanghai 200092, China
liquid samples and must exist in an ionized state. The latter
requires gaseous samples, and if there are solid samples, they must
Abstract
be pyrolyzed into a gaseous state before measurement. Thus, it is
significant to propose a method for measuring As, P, and S directly,
quickly, and sensitively in the waste gases and wastewater emitted
from semiconductor processes without pretreatment. In this
work, we report a quick, sensitive, and accurate method for mea-
suring As, P, and S in the waste gases and water emitted from semi-
conductor processes. A high-temperature hydrogen reduction
system is designed. The concentrations of As, P, and S are deter-
mined by the high-temperature hydrogen reduction GC.
A quick, sensitive, and accurate method, high-temperature
hydrogen reduction gas chromatography (GC) (1,2), for measuring
arsenic (As), phosphorus (P), and sulfur (S) in the waste gases and
water emitted from semiconductor processes is proposed in this
paper. A high-temperature hydrogen reduction system that changes
As, P, S, and their compounds into hydrides by atomic hydrogen has
been designed. It is convenient to detect these elements in solid,
liquid, and gaseous samples by high-temperature hydrogen
reduction GC without pretreating samples. The lower detection
limits of As, P, and S by this method are 0.01, 0.003, 0.02 mg/L,
respectively, and the values of relative standard deviation are 6.2%,
8.6%, and 0.3%, respectively. Results determined by high-
temperature hydrogen reduction GC are primarily accordant to
those by conventional methods such as colorimetry and ion
chromatography. The error statistics of this analysis method also
show that high-temperature hydrogen reduction GC can be
successfully used to determine trace As, P, and S in waste gases and
wastewater emitted from semiconductor processes.
Principle
In the waste gases and wastewater emitted from the semicon-
ductor processes, only As, P, S, and their compounds can be
reduced to hydrides by atomic hydrogen under high temperature
in the reduction furnace. These hydrides can be determined
directly with GC. However, other impurities in the waste gases
and wastewater cannot be reduced to hydrides. Therefore, the
interference with analysis can be ignored.
Take phosphate for an example:
Introduction
M3PO4 + 8H0 = P0 + 4H2O + 3M
P0 + 3H0 = PH3
Eq. 1
Eq. 2
It is well known that the semiconductor industry is a new multi
subject trade. In the waste gases and wastewater emitted from
many factories and labs, there are lots of poisonous materials con-
taining arsenic (As), phosphorus (P), and sulfur (S) elements and
their other compounds such as AsH3, PH3, SO2, SO3, H2SO4,
H3PO4, and H3AsO4.
Conventionally, titration, colorimetry, atomic emission spec-
troscopy, atomic absorption spectrometry, inductively coupled
plasma-mass spectroscopy, ion chromatography (IC), gas chro-
matography (GC), and so on have been adopted for measuring As,
P, and S. Of these methods, colorimetry (3), IC, and GC are com-
monly used. However, colorimetry requires a long process of
sample pretreating. As for IC (4–6) and GC (7), the former needs
where P0 refers to the atomic P, H0 refers to the atomic hydrogen,
and M refers metal atoms.
Through the calculation of thermodynamics of P-H2-KBH4 (2),
the dependence of partial pressure of PH3, P2, P4, B, and K on
temperature in the system of P-H2-KBH4 is shown in Figure 1.
Figure 1 shows that PPH →PP →PP under high temperature, so
3
4
the concentration of PH3 is far2higher than that of P2 or P4, and
the greatest part of P can be changed into PH3. Moreover,
PPH3→PK or PB, thus boron (B) or potassium (K), also scarcely
interfered with measuring P. Meanwhile, it is safe and cheap to
use KBH4 as the reagent producing atomic hydrogen.
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher’s permission.
367

Journal of Chromatographic Science, Vol. 41, August 2003
experiment is shown in Table I.
Experimental
Apparatus and operating condition
The high-temperature hydrogen reduction GC system used in
this experiment is shown in Figure 2.
Standard curves
Diluted to 100 mL with high-purity hydrogen as the standard
gas were 1 mL 2.5% AsH3 and PH3, respectively. Respectively, 0.1,
0.2, 0.4, 0.6, 0.8, and 1.0 mL standard gases were sampled and
placed into the analytic system and plotted with the peak height
and concentration of AsH3 or PH3 so that the standard curves
were obtained. For high concentrations of P, 2.0, 4.0, 6.0, 8.0, and
10.0 mL standard gas of PH3 were sampled, and the standard
curve with peak height and concentration of PH3 was plotted.
With hydrogen, 1 mL 90% H2S, obtained by the reaction of Na2S
and H2SO4, was diluted to 100 mL as standard gas, and the stan-
dard curve with the square root of peak height (h1/2) and concen-
tration of H2S was plotted.
The operating condition of the determining system in this
Sampling procedure
There are two different ways to sample, according to phases of
species containing As, P, and S in waste gases. If there are several
different phases of species containing As, P, and S in waste gases,
they can all be sampled by absorbing with solutions. A sampling
tube containing absorbing solution (1% NaOH+H2O2) was con-
nected to a sampling apparatus for sampling waste gases emitted
from semiconductor processes. The gas flow rate was 0.3 L/min
and the sampling time was 20 min. The absorbing solution was
then transferred into a quartz crucible and evaporated under an
IR lamp in a ventilator. If there is only gaseous arsenide, phos-
phide, and sulphide in the waste gases, the best way to sample
them is to concentrate under low temperature. A U-pipe con-
taining chromatography silica gel was sunk into a cooling trap
(liquid nitrogen was added into ethanol and stirred until the tem-
perature went down to –80ºC), and its one side was connected to
the air-sampling apparatus. The gases were taken into the sam-
pling tube from its other side. Thus, AsH3, PH3, SO2, H2S, etc.
were enriched. The gas flow rate was 0.4 L/min, and the sampling
time was 30 min. Wastewater can be sampled directly.
Figure 1. Dependence of partial pressure of PH3, P2, P4, B, and K on temper-
ature in the system of P-H2-KBH.
Analysis procedure
Figure 2. The high-temperature hydrogen reduction GC system: (1) H2, (2) N2,
(3) air, (4) sample entrance, (5) counter mixer, (6) reduction tube, (7) separator
with NaOH, (8) dehydrator, (9) six-way valve, (10) collector (–80°C or 90°C),
(11) separator, and (12) double flame photometric detector. Gaseous samples
can be entered into the system through (4). Liquid samples should be evapo-
rated to dry, then mixed with KBH4, and put into (6). Solid samples that are
directly mixed with KBH4, put into (6).
If the ananlyte is solid, it must be milled and mixed with dried
KBH4 (or NaBH4). The mixture was transferred into a quartz cru-
cible, and then the quartz crucible was put into the high temper-
ature reduction furnace, in which the reduction reaction
happened. If the analyte is liquid, it can be evaporated to dry
under an IR lamp. After that, the analyte was mixed with KBH4
and then measured with the same procedure as P. If the analyte is
a gas, it can be detected directly by GC.
The preceding samples do not need chemical
pretreatment, and they can be directly reduced
with atomic hydrogen decomposed from KBH4 in
the reduction furnace. The As, P, and S impurities
in the samples were reduced to AsH3, PH3, and
H2S, respectively. They were carried into the col-
lector in a cooling trap (–80ºC) by H2. The sam-
pling tube was connected to the GC; synch-
ronously, the six-way valve was located at the
“Sampling” position. After the carrying gas, N2,
purged for approximately 1–2 min, the six-way
valve was located at the “Feeding” position. The
Table I. Operating Condition for High-Temperature Hydrogen Reduction
GC System
GC system
Reduction system
Flow rate of flame H2 (mL/min) 180
Flow rate of reduction H2 (mL/min) 200
Flow rate of air (1) (mL/min)
80
Temperature of furnace (°C)
950 (As), 1000 (P),
850 (S)
Flow rate of air (2) (mL/min)
Flow rate of N2 (mL/min)
170
50
Time of sampling (min)
Flow rate of effluent N2 (mL/min)
8
80
Temperature of separator (°C)
Temperature of detector (°C)
130
100
368

Journal of Chromatographic Science, Vol. 41, August 2003
sampling tube was then put into boiling water. The absorbed gas
was deabsorbed and entered into the GC. Signals were detected by
a double flame photometric detector at 526 nm, amplified, and
recorded. Thus, the concentrations of As, P, or S in analyte can be
obtained according to the standard curves or extrapolation of
standard curves.
which are small enough to satisfy the requirement for trace anal-
ysis and demonstrate that this method has good repeatability.
Recovery experiment
In order to check the authenticity of results, certain amounts of
AsH3, PH3, and H2S (calculated as SO42– ) were added to the flask
–
–
absorbing waste gases, and certain amounts of AsO3 , HPO3 , and
SO422– were added to wastewater. The recovery was determined
according to the aforementioned analysis procedure. The results
are shown in Table V.
Error statistics
The detectability of high-temperature hydrogen
reduction GC
The relationship D = Rn/S is used to calculate the detectability,
where Rn refers to the noise and S refers to the size of the signal
produced by a detector when the smallest amount of analyte
enters the detector. The lower detection limits of AsH3, PH3, H2S,
and solid or liquid As and P are 0.013, 0.0031, 0.02, 0.0088, and
0.0031 mg/L, respectively.
It is clear from Table V that recovery is high and the results are
accurate and authentic when analyzing trace impurities of As, P,
and S emitted from the semiconductor processes by the method
of high-temperature hydrogen reduction GC.
Ten parallel experiments with 0.025% AsH3 and PH3 and 0.9%
H2S were carried out to check the accuracy, precision, and
repeatability of this method. The results are shown in Tables II,
III, and IV.
Tables II, III, and IV show that the values of the relative standard
deviation are 6.2%, 8.6%, and 0.3% for As, P, and S, respectively,
Results and Discussion
Results of measuring waste gases
Results of measuring waste gases discharged after treatment by
a scrubber for As, P, and S pollutants are shown in Table VI (sam-
pled at outlets of the 42-m high emitting gas
pipes) and Table VII (sampled at 20-m away from
Table II. Repetition of Analysis of AsH3 (0.025% AsH3 as Standard Gas)
outlets of emitting gas pipes).
It can be seen from Tables VI and VII that the
concentrations of As, P, and S in gaseous samples
discharged after treatment by a scrubber for As, P,
and S pollutants are very low, normally < 0.1
mg/m3, which is less than those of discharged
standards.
Peak height
(h, mm)
Mean
deviation
Relative
deviation
Standard
deviation
Relative
standard deviation
40.0, 44.0, 41.5, 44.0, 48.0,
46.0, 48.0, 47.0, 45.0
0.00119
4.8%
0.00155
6.20%
Results of measuring wastewater
Table III. Repetition of Analysis of PH3 (0.025% PH3 as Standard Gas)
Results of measuring wastewater discharged
after treatment by methods of coagulation and
settlement are shown in Table VIII.
Peak height
(h, mm)
Mean
deviation
Relative
deviation
Standard
deviation
Relative
standard deviation
From Table VIII, it can be seen that the concen-
trations of As, P, and S in wastewater samples
discharged after treatment by methods of coagula-
tion and settlement in the process of the semicon-
ductor are approximately 0.08, 0.05, and 0.34
mg/L, respectively, which are also less than those
of discharge standards.
Tables VI, VII, and VIII show waste gases and
wastewater treated by using the preceding
methods in the process of the semiconductor; As,
P, and S are in trace quantities, but they could be
determined by this high-temperature hydrogen
reduction GC directly and accurately.
39.5, 36.0, 42.0, 40.0, 44.0,
45.0, 41.5, 44.0, 48.0, 46.0
0.0017
6.9%
0.00214
8.56%
Table IV. Repetition of Analysis of H2S (0.9% H2S as Standard Gas)
Square root of peak
height (h1/2, mm1/2
Mean
deviation
Relative
deviation
Standard
deviation
Relative
standard deviation
)
125.73, 126.49, 126.49, 126.49,
125.98, 126.49, 126.74, 126.24,
126.49, 126.75
0.002
0.2%
0.0027
0.3%
Comparison of different methods for
determining As, P, and S
Table V. Recovery Experiment of Waste Gases and Wastewater
In order to check further the veracity of results
measured by the method of high-temperature
hydrogen reduction GC, As, P, and S in different
samples were analyzed using different analytic
methods such as colorimetry and IC, shown in
Tables IX and X.
Waste gases
PH3
Wastewater
2–
–
–
2–
Tested impurities
AsH3
SO4
AsO3
HPO3
SO4
Added quantity (ng)
Recovery (%)
10.0
95.8
10.0
109.2
6.0
99.0
5.0
98.2
10.0
96.0
10.0
108.7
369

Journal of Chromatographic Science, Vol. 41, August 2003
Table VI. Concentration of Gaseous Impurities after
Absorbed by a Scrubber for As, P, and S Pollutants
Table IX. Comparison of Different Methods of
Determining As and P
2–
Impurities
Arsenide
Phosphide
SO4
High-temperature
hydrogen reduction GC
Colorimetry
Concentration (mg/m3)
State standards
0.080
0.3*
0.045
0.3*
0.082
70†
Number of samples
As
P
As
P
1. Gas sample (mg/m3)
2. Solid sample (mg/L)
3. Liquid sample (mg/L)
0.02
0.12
0.18
0.03
1.50
0.14
0.02
0.12
0.20
0.04
1.20
0.15
* TJ36-79 (Hygienic Standards for the Design of Industrial Enterprises).
† GB16297-96 (integrated emission standard of air pollutants).
Table VII. Concentrations of Gaseous Impurities Sampled
at 20-m Away from Outlets of Emitting Pipes
Table X. Comparison of Different Methods of
Determining S
Impurities
As (mg/m3)
P (mg/m3)
SO42– (mg/m3)
High-temperature
Hydrogen reduction GC
Sample I
Sample II
Sample III
Sample IV
0.016
0.040
0.011
0.025
0.003
0.014
0.006
0.004
0.090
0.092
0.080
0.082
IC
Number of samples
SO42– (mg/m3)
SO42– (mg/m3)
1. Sample A (gas)
2. Sample B (gas)
3. Sample C (gas)
0.10
0.10
1.00
0.10
0.12
0.90
Table VIII. Concentrations of Impurities in Wastewater
Treated by Coagulation and Settlement
The error of analysis satisfies the requirement for trace ele-
ments; the recovery is also high. The lower detection limits of As,
P, and S by this method are 0.01, 0.003, and 0.02 mg/L, respec-
tively, and the values of relative standard deviation are 6.2%,
8.6%, and 0.3%, respectively.
2–
Impurities
AsO43– (mg/L)
HPO3– (mg/L)
SO4
Treated sample A
Treated sample B
State standards
0.08
0.09
0.5*
0.04
0.05
2.0*
0.39
0.24
1.0†
* GB8978-88 (integrated wastewater discharge standard).
† CJ18-86 (discharged standards for municipal wastewater).
References
It can be seen from Tables IX and X that the results determined
by these methods agree with each other very well, which shows
that the results of determining trace As, P, and S by high-temper-
ature hydrogen reduction GC are accurate.
1. W. Ruimei. Measurement of trace phosphorus in dichlorosilane by
high-temperature hydrogen reduction gas chromatography.
J. Chromatogr. A 757(1-2): 319–23 (1997).
2. W. Ruimei. Thermodynamic analysis and measurement and control
of phosphorus in the semiconductors process. Chin. J. Electronics
(4): 93–95 (1995) (In English).
3. C. Bingke, Ed. “The determinatino of granules”. In Inspecting and
Analytical Methods for Air and Waste Gases; the State Environmental
Protection Agency of China. Beijing: Chinese Environmental Science
Publisher, 1990, pp. 257–368, Chapter 4 (In Chinese).
4. Mo Shifen and Liu Kailu. “The application of ion chromatography”.
In Ion Chromatography. Beijing: Chinese Science and Technology
Publisher, Beijing, China, 1986, pp. 226–313 (In Chinese).
5. E. Dabek-Zlotorzynska and F. Douly. Automatic simultaneous deter-
mination of anions and cations in atmospheric aerosols by ion chro-
matography. J. Chromatogr. 640: 217–26 (1993).
6. W.T. Frankenberger, H.C. Mehra, and D.T. Gjerde. Environmental
application of ion chromatography. J. Chromatogr. 504: 211–45
(1990).
7. D.J. David. “The thermoionic detector”. In Gas Chromatographic
Detectors. John Willy & Son, New York, NY, 1974, pp. 114–44.
Conclusion
The high-temperature hydrogen reduction furnace has been
designed, and the method of low temperature (–80ºC) collecting
and sampling is introduced in this experiment. Only As, P, S, and
their compounds can normally be reduced to hydrides by atomic
hydrogen at high temperatures, and hydrides can be determined
directly with GC. Though, other impurities cannot be reduced to
hydrides. So, the interference with analysis would be ignored.
KBH4 was chosen as the reducer and used as the ideal material of
producing atomic hydrogen. High-temperature hydrogen reduc-
tion GC is a new method for determining trace As, P, S, and their
compounds in the solid, liquid, or gaseous states quickly, sensi-
tively, accurately, and directly without pretreating samples.
Manuscript accepted June 3, 2003.
370