Helium detector, baseline drift corrector, inlet sample splitter design and quantitative analysis

Helium detector, baseline drift corrector, inlet sample splitter

design and quantitative analysis

Citation for published version (APA):

Postema, G. A. (1965). Helium detector, baseline drift corrector, inlet sample splitter design and quantitative analysis. Technische Hogeschool Eindhoven.

Document status and date: Published: 01/01/1965

Document Version:

Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.

• The final author version and the galley proof are versions of the publication after peer review.

• The final published version features the final layout of the paper including the volume, issue and page numbers.

Link to publication

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

• You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:

www.tue.nl/taverne Take down policy

If you believe that this document breaches copyright please contact us at: openaccess@tue.nl

providing details and we will investigate your claim.

I

Report of Work Accomplished at

Wilkens Instrwnent & Research, Inc.

on

HELIUM DETECTOR, BASELINE DRIFT CORRECTOR,

INLET SAMPLE SPLITTER D.ESIGN AND QUANTITATIVE ANALYSIS ,

by

G. A. Postema

LABORATORIUM VOOR INSTRUMENTELE ANALYSE TECHNISCHE HOGESCHOOL

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

Report of Work Accomplished at

Wilkens Instrument

&

Research, Inc. Walnut Creek, California

on

HELIUM DETECTOR, BASELINE DRIFT CORRECTOR,

INLET SAMPLE SPLITTER DESIGN AND QUANTITATIVE ANALYSIS

by

Gerke A. Postema

LABORATORIUM VOOR INSTRUMENTELE ANALYSES TECHNISCHE HOGESCHOOL

---

---...

---~---I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

Section I. I I. III. IV. TABLE OF CONTENTS INTRODUCTION . . .

. .

. . .

.

.

.

. .

HELIUM DETECTOR

. . . .

. .

.

. .

. . .

BASELINE DRIFT CORRECTOR

.

. .

.

.

.

. . .

.

A. Introduetion

. . .

B. Sunnnary

c.

Instrument Controls D. Instrumentation

. .

. . .

.

E. Experiments

.

.

. .

.

.

. .

F. Resu1ts

. . . .

.

. .

.

.

G. Discussion

. . .

.

. . .

.

.

H. Conc1usion

.

. . .

.

. .

. .

.

. .

. .

.

SAMPLE SPLITTER DEVICE AND QUANTITATIVE ANALYSIS .

A. B.

c.

D. E. F. Introduetion Sunnnary

Experiments and Resu1ts Discussion • • Conc1usions Literature References i i .

. .

1 2 3 3 4 4 5 5 7 8 11 21 21 21 22 29 30 30

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I. INTRODUCTION

This paper will discuss the work I have accomplished during my six and one-half months stay at Wilkens Instrument and Research, Inc., in Walnut Creek, California.

Wilkeus Instrument

&

Research bas grown rapidly during its eight-year

life and now bas about 300 employees. Wilkeus Instruments was founded in 1956 by Dr. K. P. Dimick, who at this time was working on the analysis of

strawberry flavors. Dr. Dimick was also impressed by a paper publisbed in

1952, "Analysis of Fatty Acids and Antines by Gas Chromatography," by Drs. James and Martin.

Presently, the company manufactures thirty different instruments.

Their uses range from smog analysis to preparative G.C. The greatest

emphasis, however, bas been placed on the general research instrument which includes great versatility for varied applications.

Wilkeus Instrument

&

Research is the largest independent company in

the field of G.C. and produces only G.C. instruments. However, they have

a well equipped analytica! research laboratory which is used to broaden

the companies interest-areas in the future. This is important since the

market may become, at some future date, saturated with G.C. instruments.

During my stay at Wilkeus Instrument

&

Research, I worked on three

projects. I will discuss each project independently in the following

sections of this report. Briefly, the projects are as follows: My first

project consisted of completing an experiment with the helium detector. This new type detector exhibits a very high sensitivity toward specific molecules which are difficult to detect; e.g., nitrogen. At the Pittsburgh Conference on Analytica! Chemistry and Applied Spectroscopy, Mr. H. Hartmann,

an employee of Wilkeus Instrument & Research, presented a paper about this

detector.

My second project consisted of working with the baseline drift cor-rector. A newly developed instrument, the drift corrector makes possible

the continuous correction of a drifting baseline. I made a number of

chromatograms in an attempt to explore the potential capabilities of this instrument.

My third project consisted of evaluating a new splitter device. I

tested the splitter's performancefora combination of concentratien dif-ferences in the injected sample, temperature sensitivity, sample size, different split ratios, and flow rates through the splitter. All the quantitative results were obtained with an Infotronies Integrator. Also, a comparison was made between the Infotronies Integrator and the Disc electro-mechanical integrator.

-I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

II. HELIUM DETECTOR

The Helium detector was introduced in March 1965 at the Pittsburgh Conference on Analytica! Chemistry and Applied Spectroscopy, by

Mr. Hal Hartmann and Dr. K. P. Dimick. Their paper is given in this report.

The experiments I performed with this detector are discussed in this paper. I have measured the linear dynamic range, the calibration linearity curve for propane with the helium detector and the flame ionization detector, and finally the relative sensitivity and linearity of fourteen different gasses.

-I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

HELIUM DETECTOR FOR PERMANENT GASES

*

C. Harold Hartmann and Keene P. Dirnick Wilkens Instrument & Research, lnc. P. 0. Box 313, Walnut Creek, California

ABSTRACT

A new, simple and ultra sensitive gas chromatographic detection device has been developed for the analysis of permanent gases. The detector consists of two electrodes closely spaeed (approx. 1 mm) either in a concentric or

parallel geometry. The interDal detector volume is 150 p.l. A 250 me tritium

foil serves as one electrode. A constant potential of negative 400 volts is

applied to one. electrode with the other electrode lead going to an electrometer

capable of measuring small (10 -ll amps) changes in currelit. Helium passing

from a chromatographic column is.. excited to the metastable state (energy level

=

19.8 ev). All permanent gases except Neon are in turn ionized and produce

a positive increase in detector current. Neon shows a negative peak.

Sensitivity as low as 10 ppb is demonstrated with chromatograms for H

2 ,

o

2, A, N2, CO, and

co

2• The linear response is shown over a range of 10,000.

• Presenled by C.H. Hartma""- Pitlsburgb Co"ference o" A"alytical Cbemistry

I

I

I

I

I

I

I

I

I

I

I

I

I

,

I

I

I

I

I

I

HELWM DETECTOR FOR PERMANENT GASES

In 1958 when Lovelock1 first described the Argon Detector, he suggested that pure helium would be more advantageous than argon as a carrier gas because of its high ionization potential. This is particularly important for the detection of fixed gases where all of these ionization potentials are higher than argon and hence cannot be ionized by it. Several papers have been publisbed using He as a carrier gas with ionization

detec~ors,

2, 3, 4, 5, 6, 7. Bourke 7 and co-authors last year reported the best results with the detection of fixed gases in the low ppm range.

This paper describes an ionization detector which has parts per billion sensitivity to all gas and vapors. It is about 50 times more sensitive to hydro-_, carbons than the best flame detector and comparable sensitivity to fixed or permanent gases. Also included is a discussion of the operating parameters to achieve this sensitivity, techniques for parts per billion calibration, and the demonstration of detector performance with linearity curves for thirteen gas es.

Fig. 1 shows a schematic representation of the detector's geometry. The detector consiste of two parallel plate electrodes spaeed about 1 mm apart. The bottorn electrode is a 1/2" x 1/2" 250 millicurie tritium foil. This forms a·

Tritium Foil

t

Gas Flow

FIGURE I - SCHEMATIC OF HELIUM DETECTOR

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

detector volumn of about 160 J.tl. The carrier gas, helium, passes from under-neath the tritium foil electrode, into the small volume betwe.en the two electrodes, and then exits through a hole in the top electrode. A high negative voltage is applied to this upper electrode to obtain high field gradients of about 4, 000 volts per centimeter.

"' Q. E 0

-

.:::: 4)

...

_... ~ u .:::: .2

-

2 ~~~~----R----~·~~ 'i: ~ ~---~~ 0~---~---r---r---~--~ 0 1000 2000 3000 4000

Field Strength, volts/cm.

FIGURE 2 - lONtZATION CURRENT vs FIELD STRENGTH.

Fig. 2 shows the familiar relationship between detector ionization current and field strength for a simple ionization chamber where the gas has low ion density. The eperating range of field strength R would classically be recommended for gocxl detector performance. However, the eperating point for the detector described in this paper is shown by the dotted lines. As might be suspected this fast changing slope requires a very stabie voltage control. This disadvantage is not difficult to overcome and the resulting benefit in sensitivity increase is about 250 fold. In practice the field strength is increased until the noise level within the detector is just passed usability on the baseline at maximum sensitivity. This eperating point is close to the spark-ing potentlal as is indicated audibly and by rapid surges of high current. Total background current at the eperating voltage is about 7 x 10-8 amps. It is

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

Page 4

believed that the combination of triturn beta radiation and applied voltage raises some of the helium carrier gas to the metastable state with an ionization potentlal of 19.8 eV. If the atmosphere within the detector were absolutely pure helium, then all other fixed gases except Neon would produce an increase in current. Since Neon bas an tonization potentlal greater than 19.8 eV it will not be ionized and will give a negative response. The effect of applying extremely high field gradients apparently improves the efficiency of ionization by either increasing the concentration of mestastable helium atoms within the tonization chamber, or by increasing the energy available for ionization. In practice commercially available helium is not absolutely pure and consequently a given sample to be analyzed may yield a positive, negative, or zero response depending on the

relative concentration in the carrier gas. For example, suppose a tank of helium used as carrier gas contained 1 ppm oxygen, then oxygen samples of higher con- .

centration would yield positive peaks; oxygen concentrations less than 1 ppm would yield negative peaks and the oxygen sample of 1 ppm would yield no peak at all.

lt became quickly apparent in our study that the two most difficult problems were obtaining pure helium and calibrating the detector in the ppb range. The pure helium problem was partially circumvented in two ways. The easiest was to test a number of helium cylinders to determine the lowest concentration of the individual component of interest. The tank of helium with the lowest concentration would yield the largest negative response and therefore would be selected as the carrier gas. The second metbod was to purify the helium using a molecular sieve trap at liquid Nitrogen temperatures. All of the results reported in this paper have been obtained using pre-selected Helium as the carrier gas.

Fig. 3 shows the exponentlal dilution flask used to make up concentrations in the partsper billion range. This is similar to the apparatus reported by Loveloek 8• The apparatus was first evaluated using the hydrogen flame detector with propane as a sample. A typical recorder response is shownat the left with the ll}.athematical expression for determining the concentration of sample in the detector at any given time.

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

C "' Concentration Co= lnitKII Concentrotion

Q = flow Rate thru flosk

V = Volume of flask

T = Time !Iapse after Co

FIGURE 3 - EXPONENTlAL DILUTION FLASK

Fig. 4 shows the results of this experiment: a linear dynamic range from

0. 06 ppm to 1. 5%- a range of 250,000. 1,000,000 )( @ 100,000 ..!! 0 u 10,000 IJ) :;) LL N ₁₀₀₀ Ql "' c 0 a.. "' Ql 100 a:: .... Ql ~ 0 ₁₀ u Ql a::

L

V

-~~--~ ~----

-V

-- -- --~

/

i

/

--··~ ~---

--~--/

/

-- - ·

----/

I

I

/

Linear Range of 250,000

~~-~1

-·· ~~

~--V

I

,/

i

i

1 .01 .1 10 100 1000 10,000 100,000

Propane Concentrat ion, p p m

FIGURE 4- EXPONENTlAL DILUTION Fb.ASK CALIBRATION

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

To more realisitically evaluate the helium detector in chromatographic application, a gas sampling loop and column were inserted between the dilution flask and the detector. And as a further interDal check the flame detector was joined in series with the helium detector and the response from each separately amplified and recorded. A schematic of this apparatus is shown below.

Signall

,,, 100,000 ·6-_{Flame Detector} Elec. :. _{Helium Detector} @J 10,000 11 Column: 1/8" x 5' 8 _Activaled

"'

1,000 _Aluminum :; V•nt ... N

Go• Sampling Volve

..,

"' ïi J: 10 .... 0 " a.. .01 .1 10 100 1000 Propane Concentration, ppm.

FIGURE 5 - CALIBRATION FLASK WITH CHROMATOGRAPH

Various concentrations were made up with the dilution flask and introduced

to the chromatograph at the desired time by means of the gas sample valve. The

resulting calibration linearity for propaneis shown to the leftof Fig. 5. The standard deviation for this metbod of sample dilution was found to be about 10%.

In an effort to put the flame on a par with the Helium detector, the

standard flame sensitivity was made eight times greater by using pure oxygen insteadof air to support the flame cernbustion and the hydrogen flow was in-creased to make a hotter flame.

As seen from Fig. 5, the Helium detector is 50 times more sensitive to

propane than the flame; minimum detectability with the Helium detector is about 1 ppb. A linear dynamic range of about 10, 000 fold is shown. Thirteen gases

I

I

I

I

I

I

I

I

I

I

I

I

I

'

I

I

I

I

I

Page 7

were compared using the straight dilution flask method, i.e. , no column as in Fig. 3. The relative sensitivities and further demonstration of linearity is shown

in Fig. 6. Because Neon gives a negative peak it is plotted on a negative slope.

100 90 80 70 60 ~ " 0 a- 50 ~ Gi "'D 40 0 u

..

"' ₃₀ 20 10 0 0 2 3 Concentration, ppm. 4 HELIUM DETECTOR

Relative Sensitivity and Linearity

using the exponentiol dilution flosk

FIGURE 6- HELIUM DETECTOR RESPONSES

In termsof sensitivity comparison in ppm, the Helium detector is compared

in Table 1 with three other standard ionization detectors for fixed gases. A comparison in sensitivity to grams per second showed a similar result.

I

I

I

I

I

I

co

₂

co

I

02 CH 4 N2 H2

I

I

I

I

I

I

I

I

I

I

I

I

Helium Detector TABLE 1

LOWER DETECTABLE LIMITS

(In parts per billion)

Argon Electron

Detector with Drift

He Carrier Velocitl:

9

by Hartmann, by Bourke, by Smith,

T

I

1 mv.

I

Dirnick .8 3 3 3.5 15 20 2.5 ppm. H2 Argon Dawson, Fidiam Denton 20 600 60 600 70 2000 50 1000 200 200 500 1000

l

____ ... ...,

~ '"-~

"

..lA

r----2x _I 02 I I 0 3 6 Ne Column: 20' x 1/8" Mol. Sieve SA Temp. : 40"C. Sample : 5 mi. He 9 min. H₂ m He Analysis Coaxial10 Micro-ionization by Shahin, Lipsky 500 1000 700 300 3000 1600 STANDARD DEYlATION of H2 Peak 63 68 67 69 68.5 69.5 <T=1.7% 69 70.5 70 64.5 70.5 68.5

FIGURE 7 - A TYP I CAL HELIUM ANAL YSIS

I

I

I

I

I

I

I

I

I

I

I

I

I

I

,

I

I

I

I

I

Fig. 7 shows a typical analysis of cylinder helium. The Neon reponse goes negative because it cannot be ionized by the metastable helium. Neon will show a positive response only when the concentratien in the sample is less than that of the carrier gas. The hydrogen peak represented is 2-1/2 ppm and

was measured eleven consecutive times with a standard deviation of 1. 7%.

Oxygen and nitrogen respond negatively again because their concentrations are less in the sample helium than in the carrier gas helium.

Carrier Gas: He A Sample: Hes 0 3 6 I A 9 min. I 0 A I 3 I 6

Carrier Gas: Hes Sample: HeA

FIGURE 8 - EFFECT OF RELATIVE SAMPLE CONCENTRATIONS

9 min.

Fig. 8 further demonstratea this effect of relative sample concentrations.

Page 9

The trace to the left was run with Helium A as carrier and Helium Bas a sample. The trace to the right was run with Helium B as a carrier and Helium A as a sample. Note that each peak has reversed itself.

There is more application for this detector than just measuring impurities in helium cylinders.

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

·

I

I

I

H, I IN• A I N2 , 10ppm. CH,, 10 ppm. Argon Sample 160 x 12 min.

N:z Column: 20' x 1/8" Mol. Sieve SA

Temperafure : 40• C.

Flow : 50 mi./ min. He

CH,, 5 ppm.

Oxygen Sample

1280 x

12 min.

FIGURE 9 - ARGOI'f AND OXYGEN ANAL YSIS

Fig. 9 shows 10 ppm concentration of hydrogen, nitrogen and methane, in

a standard Argon sample. This is the left-hand trace. These concentrations of

Page 10

H

2 , N2, CH4 at 160X are all encro~ching on the upper limit of linearity of the

detector. Of course, the tail of the Argon seriously limits the minimum detectable

quantity of N

2 and CH 4• To the right of this figure is shown an analysis of a

tank of oxygen at an attenuation of 1280 X with nitrogen too large to measure and

5 ppm methane.

IO,N, I

Column: 5' 1t 1/8" Mol. Sieve SA

Temperafure : 40•c.

Sample : 3 mi. Breoth

12 15 mK1. I 02N2 0

r

16.5ppm. CO (Smoker) 12

FIGURE 10 - CARBON MONOXIDE ANAL YSIS

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

.

I

I

Page 11

Fig. 10 shows the comparison of CO concentration between a nonsmoker and a smoker. Although 0

2 and N2 in the 3 ml breath sample seriously overloads

the detector, CO is easily measured at 2. 5 ppm at an attenuation of 16X.

The extreme sensitivity of this device requires an excessively clean chromato-graphic system free from water vapor, column bleed, and back diffusion of air. The detector is also moderately sensitive to flow and temperature change. Even with these limitations the helium ionization detector can satisfy, at long last, the requirement for ultra high sensitivity for fixed gases and has at least some application for hydrocarbon analysis.

References:

1. ].B. Lovelock, ]· Cbromatograpby,. 1, 35 (1958)

2. V. Willis, Nature, 184,894 (1959)

3. R. Berry, Nature, 188,579 (1960)

4. ]. F. Bllis and C. W. Forrest, Anal. Cbim. Acta, 24, 329 (1961)

5. A. Karmen, L. Giuffrida, R.L. Bowan, ]. Chromatography, 9, 13 (1962)

6. C.M. Boyd and A.S. Meyer, Report No. ABC ORNL 3619 Reprint file #309, 360 (1964)

7. P.J. Bourke, R. W. Dawson and W.H. Denton, ]. Chromatography, 14, 387 (1964)

8. ].B. Lovelock, Gas Chromatography 1960, 26

9. V.N. Smith and ].F. Fidiam, Anal. Chem., 36, 1739 (1964)

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

_

I

III. BASELINE DRIFT CORRECTOR

A. INTRODUCTION

Since the problem of a drifting baseline exists in G.C., efforts

have been made to correct this problem. Often, baseline drift is caused by

the column when it is used for temperature programming. As the column's

temperature increases, the standing current (detector) increases. The

current increase is caused by some of the column's stationary phase entering the gas flow where it is carried to the detector causing baseline drifts

(column bleeding). It is especially necessary to correct for this drift

when performing quantitative experiments.

A previously developed method for minimizing baseline drift was

the'dual column system. This system comprised two columns operated under

the same conditions. However, the sample was injected into one of the columns

while the other column was used as a bleed reference for the first column. The problem with this system is that it is almost impossible to make two columns exactly alike with the same packing density and distribution of partiele size. The little gain in minimizing baseline drift was, however, not satisfactory.

This newly developed baseline drift corrector electronically

corrects baseline drift. But it also has to distinguish between peaks and

drift. This corrector is placed between the chromatograph and the recorder,

thus presenting to the recorder the corrected signal.

The instrument has three circuits: Slope detector, logic circuits, and corrector circuits. The slope detector measures the slope of the signal and when the slope reaches a certain preset value, it will signal the logic

circuit that a peak is emerging. Also, the slope detector contains a noise

reject circuit which filters out baseline noise.

The logic cireuit indicates when a peak is emer.ging and generates the switching signals for the corrector circuit.

This corrector circuit corrects the baseline during the time that no peak is emer.ging and it

when the corrector circuit gets correct during that time at one a linear programmed correction. the baseline to zero.

maintains the baseline at zero. However,

a signal that a peak is emerging, it will constant rate or, if desired, it will give

After emergence of the peak, it will return

The baseline drift corrector has to sense several conditions

during a chromatographic run. It has to distinguish between noise and drift

at one side and peaks at the other side. It has to distinguish the valley

between two unresolved peaks and the baseline and to distinguish between

peak shoulders and s shifting baseline. To get a good chromatogram, the drift

corrector has to handle all mentioned conditions properly.

-I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

Several experiments have been run with this baseline drift corrector to demonstrate the capabilities of this instrument. Accurate comparisons of results were made on the uncorrected as the corrected traces by using a dual pen recorder.

B. SUMMARY

Chromatograms were made with the Model 450 to demonstrate its capabilities. However, several problems occurred while using this

instru-ment. Not all of these problems were eliminated. Therefore, applications

for the Model 450 were not as good as was expected nor was the reliability too good.

C. INSTRUMENT CONTROLS

1. Correction Switch

With the correction switch it is possible to choose the

kind of correction desired. It is not possible to make baseline correction

in the none-mode. There are two different modes of correction: Flat

cor-rection and slope corcor-rection modes. The flat correction mode functions

during the time no peak is emerging and maintains the baseline at zero. When a peak is detected, it stops correcting and the drift and the peak are

recorded together. The slope correction mode maintains the baseline also

at zero as long as no peak is emerging. But, when a peak is detected there

is still a correction for baseline drift. In this way, only the peak would

be recorded. The amount of correction increases linearly and is representative

of the average of the slope prior to the peak detection.

At the end of a peak, the correction unit re-zeros the base-line in both modes of correction.

2. Damping Control

This control, usable only in the slope correction mode, determines the time over which the average of the drifting baseline slope

is taken. This average is the amount of linear increase of the correction

in the slope mode during an emerging peak.

The average drift is continuously stored in a memory circuit and is recalled as soon as a peak is detected.

3. Peak Sensitivity

This control determines what rate of slope will he detected

as a peak. The baseline will be kept at zero for peaks with small slope rates

and thus he eliminated. Greater slope rates will he seen as a chromatographic

peak. It is possible to adjust for drift from 1 to 64 percent (divisions of

the recorder scale per minute).

-I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

·

I

I

I

4. Correction Delay

With a small potentiometer it is possible to adjust the time it will take before the corrector re-zeros the signal at the end of a

peak. In this way it is possible to avoid re-zeroing the baseline between

two unresolved peaks or at a flat shoulder of a peak. Otherwise, this

would give a useless chromatogram.

5. Noise Rejection

This control makes it possible to filter baseline noise

as high as 4% scale deflection without triggering the slope corrector. It

is possible to adjust from 0.5 to 4% of the scale deflection.

6. Reset Button

By pushing this button it is possible to re-zero the signal during a peak detection or at times when you wish to re-zero the signal.

7. Recorder-Zero

This control is used to electrically align the corrector zero with the recorder zero.

8. Meter Adjusts

It is necessary to keep the needles of the meters of the

corrector and detector adjusts on scale. The corrector meter adjust indicates

how much correction of the baseline has been made. Full scale of this meter

represents about 70% of the recorder scale or 0.70 mv.

D.

E.

INSTRUMENTATION

1) Aerograph 1520

2) Aerograph Baseline Drift Corrector, Model 450

3) Texas Instruments Servo/Riter II, dual pen recorder (for

directly camparing a corrected and uncorrected trace) EXPERIMENTS

Three different mixtures were chromatographed.

1. Normal Alkanes Ca-C12 in CS2

It was not necessary to use the baseline drift corrector with this experiment since the baseline was straight enough without correction.

-I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

2. Methyl Esters of Fatty Acids

c

2=

8

-c

18 in Hexane

It was necessary to use temperature programming. The

resulting baseline drift was corrected by the Model 450. Chromatograms were

made of this sample: One in the flat correction mode (Figure 1), two in slope correction mode from which one was with a damping adjustment of 20 (Figure 2) and the other one with a damping adjustment of 40 (Figure 3). The conditions were as follows:

Injector Temperature: 265°C

Column Temperature: 150-235°C (10 C/min) 0

Detector Temperature: 270°C

Column: 101

x 1/8" 20% FFAP on Chrom W

Flow Rates: Column: 40 ml/min· _N2

Flame: 30 ml/min

_1'2

300 ml/min Air Attenuation: 1 x 16 Sample Size: 0.3 ul Peak Sensitivity: 16% 3. Fatti Alcohols

c

1

=f

18 in Hexane

This mixture exbibits a very wide boiling range and

re-quired temperature progrannning. This was necessary to obtain optimum separation

in the shortest time. The baseline drift was corrected with the Aerograph 450. Several chromatograms were made with different adjustments of the baseline drift corrector. One in the flat mode correction (Figure 4) and three in the slope correction mode with damping adjustment of 10 (Figure 5}, 20 (Figure 6), and 40 (Figure 7).

The conditions were as follows: Injector Temperature: 265°C

Column Temperature: 75-245°C (10°C/min)

Detector Temperature: 270°C

Column: 101 _{x 1/8" 20% FFAP on Chrom}

W

Flow Rates: Column:

Flame: Attenuation: Sample Size: Peak Sensitivity: 6 -40 ml/min N 2 30 ml/min H 2 300 ml/min Air 1 x 16 0.4 ul 8%

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

Also demonstrated is the way of correcting a noisy baseline

(Figure 8). The noise reject adjustment was 2.

F. RESULTS

1. Methyl Esters of Fatty Acids

ca:e

2=·

18 in Hexane

a. Flat Correction Mode (Figure 1)

The operation of this correction is demonstrated nicely. Although for some peaks, the corrector detects these peaks a little late so that some corrected peaks are about a scale division smaller than

the uncorrected peaks. Also, the time delay,adjusted to 10 secouds is not

sufficient for the last three peaks. Small peaks with a slope smaller than

16% are filtered out (peaks sb and se).

b. Slope Correction

A damping of 20 (Figure 2} caused too much correction

for several peaks. This was caused by the first part of the peak not being

detected soon enough (peaks 1, 2, 4). Peaks 6, 7, and 8 also have an over correction, while the time delay between 7 and 8 is too short. Peak sa bas not been detected because the slope was not sufficient.

Damping 40 (Figure 3} is usually a good adjustment

as can be seen with peaks 1, 2, 6, and 7. Peak 8 is over-corrected because

of the influence of the slope before the peak detection of peak 6. Also,

the time delay was not long enough for the valley between peaks 7 and 8.

2. Fatty Alc~hols

c

1

-c

18 in Hexane

a. Flat Correction (Figure 4)

The solvent peak bas been kept smaller than full

scale by attenuation. By pushing the reset button, the baseline was brought

back to zero. As a consequence, peaks 2a, 3, and 3a ended below zero. Peaks 3a, 16a, 16b, and 16c were not sensed because their maximum slope was

smaller than 8%. This chromatogram is a good illustration of the possibilities

of the flat correction mode.

b. Slope Correction/Damping 10 (Figure S)

This damping adjustment causes an over-correction

for peaks 4, S, 6, 7, 9, 11, 16, and 17. The reasou is the large influence

of the first part of the peak which is not seen by the peak detector. The

increasing of the slope is the reasou for the under correction of peaks 13 and 14.

Damping 20 (Figure 6) is a better adjustment. The

influence of the first part of a peak is much smaller. Peaks 4, S, 6, and 7 are above the zero baseline because of the changing of the original base-line slope from negative to positive. Peaks 10 to lS are above the basebase-line because of the increasing of the baseline slope.

-I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

Damping 40 (Figure 7) gives a better correction for almost all the peaks. Peaks 4, 5, 6, and 7 are very well corrected. For peaks 8 to 15 there has been some correction but not enough. The reason is that the baseline drift is not linear during most of the peaks, while in this particular case a small horizontal part in front of a peak causes too little correction.

G. DISCUSSION

1. Flat Correction Mode

The flat correction mode is the mode which has given the best results. It is a mode which has some disadvantages when working quantitatively. You have to correct for the area under a peak which is caused by the baseline drift during the peak elution time. But the advan-tage is that you are not dependent from other adjustments of the baseline

drift corrector. It is now only the peak sensitivity which bas to be adjusted. This adjustment has to be just a little larger than the maximum slope of the baseline drift. In this way, it is possible to get almost all the peaks. When your peak sensitivity adjustment is too high, you filter out several

smaller peaks and also you cut off the lowest part of a peak. This introduces a large error in the peak area and so in the percentage of·that particular component of the sample. Because of this compensation, it is fairly difficult to work quantitatively with the correction mode. Qualitatively it is usable.

2. Slope Correction Mode

The slope correction mode has been introduced especially for quantitative work. It corrects the drifting baseline also during an emerging peak. However, several problems occurred while using the correction mode. As long as the baseline is drifting linearly, it will work fine. But as soon as the drift is not linear, it stops working properly. The reason is clear. It cannot see the changing drift slope during a peak and it cor-rects just before the peak was detected. As soon as the slope of the baseline, during a peak, is decreasing, the corrector corrects too much because the

memory gives too high a slope. As soon as the slope of the baseline during a peak is increasing, there will not be enough correction. Also, the reason for this is easy to understand. The memory circuit remembers too small a slope and the result is that the corrector does not correct enough.

However, there also seems to be some reasons why the base-line drift corrector does not work exactly as described above. In some cases, you would expect to obtain good correction, but didn't. While in other cases you would not expect to get good corrections but did. But there is an explanation for this occurrence. The slope of the baseline is put into a memory circuit. As soon as a peak is detected, the slope has already reached a certain value, the preset value of the peak sensitivity. But the signal with a slope just smaller than the preset value passes thru the memory circuit and this slope is higher than than the drift. It increases

-I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

- - - -- -- - -

-from the drift slope to the preset value of the peak sensitivity. The result of this will he that the influence of the first part of a peak, although the peak is still not detected, is very high. This results in a correction larger than you would have wisbed or expected. To make the influence smaller, you have to adjust your damping control. The longer the time over which the average of the slope bas been taken, the smaller the influence of the slope of the peak before the peak bas been detected. At the other side, too long a time will also introduce a mistake in cases where the baseline drift is not linear. The memory circuit will also remember that part of the drifting baseline of which the slope was too high or too small compared to the slope

just before peak is emerging. It seems to he that the best adjustment for the damping control will he 20 or 40, depending on the peak sensitivity ad-justment and the kind of baseline drift you have.

3. Slope Detector Circuit

Of the several instruments I have used, there was some trouble with the peak sensitivity circuit. Sametimes a peak was detected

when its slope was already higher than the preset value at the peak sensitivity limit. Also, the detector didn't see the end of a peak. Using the correction mode, this presents a problem because there is no correction. The reason probably is the ammeter inside the instrument. This meter works with two photo cells at both sides of the zero point of the meter. The distance between these two cells was very small. As soon as the zero adjustment is not .correct this will result in peak detecting at the wrong time and also

in detecting the end of the peak at the wrong time. The ammeter measures the slope of the signal. The needle passes one photo cell as soon as there is an increasing slope, then there is peak detection. When the slope is negative the needle covers the other photo cell. This cell is placed at the other side qf the ammeter zero. As soon as the needle doesn't cover this photo-cell any longer, the ~ak detector stops detecting the peak.

The slope detector circuit reacts slowly. This is to demonstrate with some very fast peaks. The detector is too late with de-teèting these peaks. This will result in large errors for fast peaks.

4. No i se

Figure 8 shows the way of filtering the noise of a noisy, drifting baseline. Usually it is better to set the noise reject adjustment to 0.5. Using a larger adjustment will result in an error of the peak-area.

5. Time Delay

The time delay of 1-10 seconds is too short in some cases. It would he better to have a time delay of 2-20 seconds. This would take care of the fact that the pen will not come back to zero between the last peaks of the methyl esters of the fatty acids. However, this would neces-sitate a larger correction for single peaks when using the flat correction mode.

-I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

·

I

I

6. Saturation Effect

When a peak was larger than full scale, there is a difference between the corrected and the uncorrected trace. The slope of the corrected one is larger. This is the reason why the solvent peak has been kept smaller than 1 mv with the accampanying chromatograms. With later models of the 450,

this will be changed. It is a saturation problem of the electronica in the

drift corrector.

7. Attenuation Possibilities

Proper attenuation will not always be possible when using the Model 450 because most of the time a sudden change in baseline disturbs the slope detector circuit and there will be correction when it is not necessary. After such a correction, it will not be possible to change the attenuation of the electrameter. Attenuation is only possible when the corrector has time between two peaks to came back to the baseline after attenuation.

8. T.C. Detector

The T.C. detector seemed to disturb the slope detector at times. The reason for this may be the very high frequency noise of the detector which cannot be sensed by the recorder, but which can be sensed by the peak detector as a peak. The flame does not exhibit this by reason of the fact that there is a condensation system between the flame and the Model 450.

9. Camparison with the Infotronies Baseline Drift Corrector

There are certain differences between the Model 450 and the Infotronies baseline drift correctors. The Infotronies had only the flat mode correction which is fairly slow, while the Aerograph has the slope

correction mode too. The Model 450 uses an electronic system while Infotronies

uses an electromechanical system for correction. Used was the Infotronies CRS-lOH.

Figure 9 shows a chromatagram using the Infotronies base-line drift corrector. Therefore, the Infotronies integrator is used for

integration. The corrected signal goes to the printer and not to the recorder.

What has occurred during the run is noted on the recorder trace.

Between the horizontal signs there is a slow baseline correction. The vertical signs shows if the correction is positive or negative. The printer tape indicates that for the later peaks there is too

much correction. Signals below the baseline of the corrector are not counted

by the Infotronies printer. So from the printer tape it is possible to see where the baseline was corrected too much and where the corrector didn't

have the opportunity to return to the baseline. There is only baseline

-I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

correction when no peak is detected. This results in the fact that also

the peak detection system of the Infotronies is still a weak point. Also,

baseline drift correction can give large errors in quantitative work. This

would not have happened with the Model 450 because of the time delay. When it does not go properly, a combination of the Model 450 and the Infotronies baseline drift corrector could result in a better instrument. For example, the time delay of the Model 450, together with a noise reject and a proper working slope correction mode, would be a good feature for a new instru-ment.

This is partly realized as the new Aerograph 471 which

is built by Infotronies for Wilkens Instrument

&

Research.

H. CONCLUSION

To obtain the chromatograms which demonstratea the capabilities of the Model 450, I had to first explore the features of the instrument.

This is because the instrument didn1_{t workas was expected. It seemed}

that there are several things which have to be changed before the instrument

is reliable. The time delay has to be changed to 2-20 seconds. The peak

detection system is very weak. The flat correction mode functions as

expected but the slope correction mode did not. It is still a question if

the slope correction mode has good application, as it now works. A large disadvantage is that there is no possibility of attenuating the electrometer, when you have the Model 450 between the

electrometer ànd the recorder. An application for this instrument would

have been to use it together.with a disc electromechanical Integrator.

However, for good accuracy, you need to attenuate when using the hall and disc integrator.

An application of this instrument in a field which doesn't

have demanding requirements is in preparative G.C. In preparative work

you often have a drifting baseline which can be corrected by the Model 450. Quantitative results from several runs in succession are not read from your recorder. Often the ratios between several components of the mixture are known and the injection is automatic.

---~~ --~--~~~~ --~ ~ ~~---.

---

.-I

··

~

__

_...,

_____

_

0 1

..

.

.

,

)

•

) , I ::. ~ )

---·

----~

-1

J

~~

I

·

~~

i

'~

~/"~

""~~

!'lil> ..

r-~~~

~~~

.. _i · I I

""

I

~

-· -t Î -· j · ·

-~"

I

I

~

~

I

_~>

~~

I"' I

~

L ~

~

i

_~,.

""'

~~~

1\ ~~ '

~

I

~~

~

!

·

_~~

e

I ! ~ I .... . 1 -'

.

-...

.

-I

I

~l~

~IJ

~~

·~

,_. ~'-~"·

t-~

1 -~

~~~~

~v

.-jÇ~

4 !Wi 1.1

I

..

-.. .

~~

: ;_ ~ -. . I

I

I

I

I

~ .... f-"

....

~~

~r I

-1-...

~~

.

.

-·

-·

-if

·+

. ~ ~ .. ' • . I I~ .. .

f

.

~~

I I ~ < lj I

.

ï

I i !

I

I.,... I" ~-"v

I

I

~

~~

I

I

~~

•

I

l

I i I I

I

.. ·~"-1"-• I '

!

I

~-t

.. --.

-· · 1 -~ -_1-~

I

~

•

•

~I

'

('

i f41'!1

•

.-' I•""

-I· . I .. . t<l ! 11>

~;

I

i

.

l

i , IE~

-I' Cll ~ 1 1

I

I~

I I '

I

~

i I

---I I t 1:

I

I I

. l TEXAS INSTRUMENTS INCORPORATED,HOUSTON,TEXAS,U

-r I Jl ' , .:

-'

-

~--~~~--+--~~---- - ~ " ~~~-'~~~··H-~~~-+---~~~----~~~--~~~~---~ ~ --- --f.+-1-- l-11-1 -- -'-:I 1---+--1-H+--1--HJ+--H.-1~---H--~ +Ht--H-1f-H-1fl- ----· - +- -. \ \ \ \

'·

~---~-+++--~---D-~·~f----1-t+~-~---~~f---~----~~ +---4---+++---+---~---~+----HH~--HI---4-~~----4---~~-41----. ~ i

-

....

...

•

..

L

.

,

.-.--

,.,

....

...

- .-

~"'1

-~ (i ·-l

...

--•

~

.

" I \ Y I '-

tt--

---*

----HY' ' 1---1-~H--t---1----+-.:...--i~l\---1---- '· n :1

..

ll ~~--~---~ ---~~-:1 " H---+--- ,..-\. '1+J1~ - --t---t---•+t--t--ti-+---1--+-~-f-1-llrl--..l· 1 -+ · :L.'J

•

..

·

....

--~r--+--Hf--- =--+---l---~~-4--+~--~--~~~---

..

4--.-- ·----r--- ---~ .. . -~ -'IJ rJ3iM:Jr_) ~ --, , -li-41:;~-- _- _- - _-_ - • 1111"' ._. . ~ . , · f- _-_-++ -l-l~f+ _-_-_ - _--l+ _ - _- _-_- -+:~f-f- _-_-_ .... -__ -~ ...

1'1

-l

~ A.~-• -llli. 't ., A.-• ...,...----l~ --~"

_,

- ,..ll+44+·= - +- --0 i

----..._.

.

I

1

1

I

I

I

I

I

I

I

I

I

I

I

I

I

• I ~ - --·

.

.

,

_

--~

-~--~---+~

~

--~

~

--~~

~

-++---+-+-+---~~

---~~--~

-~

~~--++~~~~--+~

~~iw~--~~'~

,_-1----

---1--

---1--~--~~~---~~---~~---~

~~---4----==~~~,~~

==

~=tt:~rt:::=t~==tt==j:j:f:

:::t

t===~~=

-·

-=t

-~~

==tJ::±l:±:::t=:t:

_~~~{

~-l-1-

----

~---~1

---~---l

~-4

----;

---

- t · -- -. ---~-l--- l (,,

----~

;

H~r++~H~~~+*~~~~~~~4~

-"'·

'-~~:

··

~

1 -~-r :...+HH+_J -~ _-_ _ __ _ --+ .:_t-.~t · ·_-_J_J ·-- --..r---+~---_-_--lt-1.tl----_-_ ... +--+c.:_-_ -_+ +="_: ·· _-_ :.... _ _ _ r-~-4-r-~.r-_-_j - --· - -_ i j~

.

---+

. --:. - · - - ·- - -+--+---- 1 ----~---h:-~

- ---t---r---+-it-...

+--1----:

~

~~

==.

t

+-++----Ht---~---+

--.:.

·

--++--

-

~_--_

--:~::~~·~"'

~

-4 . ...

~

..

=

..

x~~f

::::;::~:t::::::::

:!:

::::::::t=~!..r'~/~

~

-~~..,..~~

...

;

.

*t.

~'

~

-.."~-~

7

~=::-~~~~t~·C;'f,_:.,~~·~{.,~.:-;

7

~·~-Io....~T~

==

t

=

=

==

±

===

==-=

.

+·

·

=1

·========t=~~~,~

==t====:::±========t====~:j

==~==tt======

~-~:3~

!=

~

~_~L

--

=-

·

--

=

±

===i=:==i

1

====

====tlt

7~~~4±=±j

__

~_,

:-

Ct,

=====~~

~

.

!ti

7~J~

·

-~~?t1~

--~~-~~

---=

:~~~

~

-~~-~t~~

-~i~l:~~:~~

=~--=

~-t~

·

~

---H:L--~

I -·

-+

··

~:L

-

~:J

-+~

~.

~

~·~~~1-~

~

~~~~H~+

r~

~~~

==

~-=

jj~

~

~

---~~

!~~

~1

--~

~~

-~~

t_

-Tr-~

-~~ii~~~~~~~~~~~~~~~~~~~Fr~~~

--~

~~~~

---~~

~~,~,\~~~,~~~~~~~

~~~~

i

\

\

i-'

.:!.: · - ~ H-!--1-J. - 1+-lf - --- -tf+-f H- I · -1-+ H-4- · ---- -- - 1--+-1 ll _, i ~::..__ 11>----t--111---!--f t -

.-1

--P .. "----1 ~1 -- -H-+1 ~--1-1 +1--1-1 ~- --- 11-· +- l+ --l---1 f-t --~ I-+ --1 \ -, \ -/

~

-~----~

-Jir

~

-r

--·

--~,

---

u

-~

-~

---1

~

-41

~

~

~

~~

---~~

----:

~

-"·

-

+-~

---·

·

-~

-==-~

ff~~~~~

~~~\~~~~~~,~

-~~~~~~~~

~

~~~~~~~~~~\~~~

~~

~~~~~~~~

ILI

4

---

-

---!-11--~-+

----

fi*ftt----''-""''-..-tl-t-1.3i:J;__~---+-I

r;LIIl-1

fr

-,-

~

,_

-':_~-

:_

f_J~-

~---U-

-14

W--

--+-I

+-ti--

-

--I

-

-I-+--

~-'J-

+

-

~""'

~

-

---

·

+--

·

·'

I-+

'"---+-'

---f<-'""'

4 - -- ..f

-~

-~~~~~~---4-~~

-~

-

----~~

-+~l

-~

-~---14

~

1

--1

~

--lt

+~=

--J

-~

[

J1~~

--JË

~

-~~~~~~~~~~~~~~~~

-~~~~~

~~~~~~~~~

~~~

~~~~~~~~~~~~~

~

. 1-

r-

-l -- - -+-1+- - --f-.J.-..f -- + +- -- ---- - .,- -.. - -. -.. ---~~ ~-~~ ~~ --+--+ ~-- ~~-+-~ ~~ -+~ -~~ fL -~

-·

,

:·

,

~-1 .• --il - '-~.U • I ' _[ --..._.. a ,.- H- - -1+--1 - -+11---U 1

·

:

x

-f - --4-~1-- 1 ~ - -+ -- -1 -- H ---- -- , t-1 lJf/'# - fl...: l ---~f---_:_-·--; -, 1

-~~~~====~ft·~~~~~======~======t=====~=======t======

j

====

-· .. · r~ -- 1 · i --- ~ -H - - 1 --H- - 1 -·1 -- -r ---1 ---·1 -- - -- 1..._._ - 1 -- -4-U ..._-- . - . +-- .. ,,- ·-, · --...".="'

.:

:

__

:::,~,::..t~~

:;;;;L.

~

t~'+AJ~:fii\I~V:...~~-.·7'

...

J""'

-;.:·:

==-~=-t

-~==j~t

----··

-_

+t

=

't

==

==

~tt

==~

t

~t=

==t-t

-}!

~

-

---.

-1

rElil=~=====~~t:t::!t:..

+

~

... ;;:=..._===:t======±

=·=--==

==±=====±=====

±:

==:

~~~~======~~====~==~~==+=======4===

i~-~~-~~

-

--~~

,j ;::::--:~'! · = ~==-·----+l ilf-' HtJ - --- --~-::_ l i --_ · .. , _ 1 __ _ _ !· -++---H--1---Htii---H +44U -~' . r ~ ~

----: -· ---! ---· - ---- ;-~} ~r-~ -, I - ---HII-H r'--- ---·-l --! :, :.-, ~_-:- -Î-~ ~:.

\

.1Î I ~ ~V ilrï I Ît.trÎ :·~ : --: --··

~-r

~1f;

~~~tt~~

~~~~~~~~~~

~~~~~~

-~

~~

--~

~L~-~

~~-~~~~~~·~~~

~~~

~~Jr~i

T~T~~~

~~~

J-~~~f

I f

"

,...

~,

f

' _: ': -~ - - .. - --1 4-U---'-W-- U-- ---U~--+I ---+-HI -· ---;---I ---tff--H-- '"

_.",~'-t-~

• 7 "'- ...,.. al :...a ,_.I . . ir ~ ; .. ,· - • . -t-..' #--~ ~~~ - -4-H- IH--- .-a+- - ;. Hif---~H-· - - ·-- f ·. ._ -· - !~~ ~- - ::_-~- --· H-·- - +---tt- ----

-':

'

~

·--

---~

•

..

-.

..

..

..

__

:~:.--:._;:···~+-~ ~

..

,..,

~

••

n:

... r '

--,

\l."

! -~ \

=-==;

!

, r ~

!

• p ~~ : ; . ,_ , ; - f-- ~ . -'--~ -t--":s-.--fll-- 11 -1 - ----=-_. -1 -41-___; -U --l----1+ 1---l- 11 - -+11-J.----''

H--·

; .

~:

-·

La_ PI I T ----cH-- ----i:- ---+-- -- ~ +--- ~ -+ =- .-.. .-. , 7-l ~ r,; -~ . -'-, - ·. lf' -1 --_;,

:

~

=

:

·

--q

~

;_

..

· 7j't r..;t_;

P:t

.LI :, 7 J

~

-l/;

.

I ./ .... ·-\ T · --f ---...!...-11-· -H--- + - - ~--D---1--+----H +- ---+--H-=--····-Ji i_:::.:... --:-h~

..

;~ ~ r;J~~r~ :· ! "

'"

-l"

..1

··_

.,..1

1

-.

_..,

-· --1 - - --f-v -'-'-1 4...}+ - 1- ---· : -, _

--tr

,_

-,..~

.,--f--H

~:_

-

_.~

-

--·-··

-·

+::==::-::_-_-__

+-IJ----I-l--

-+~-

-l----l-ll---1~

h

=--=-~

;

~_:~

·~

~~

.;:

1

-~

['

d l'-' -"'l -_ "" -, --1 ---j llv..A--I, -::: -1 ,:..l -···· ---il -'

~~~~--~~

~-

~

-~

-4--~~-ill--~-~~~~

~

~~L

~~~~~ ~~

~~~

~

-

,~-~

-~

--

--~---+----~

-·

--~

---,--i--

-

---~--

--+

-

---

-+--

-·----

:-

--

x~ Ç

-

:!A

n I

0

u~

; ...

:f

:i'

'8:

~-~

J'H

·

_.!I

-- _

..

-

-

-

~~~

~

,

~

-

--=

-=-=--

-l:~tll--

:.=:~=-

j;

~

-

_

=

:.

-

t_

~~~

=~=::~-='t~=::~j

-~~

~::::

-1-t~

_

,

l-

-_·

_·

_-_:_-

_

-

-_

_

::

:::

-::t:::

---

-

i-li

·

-

~::---

t

'--~....,or:::J

.;:. .,,~~ 1 '!Jj ,--~ l·

.

.

-.-J _ I ~

.-i

.

.

·

~ _. -n ~ -t -~ II ~ ~r,.-1,., --f"!. -· r -· -r11r.a ~

... , :::..__ -fV - --- ~ -- :.._ · ""' ,.. ~-- I J

"!

'

~ ' ·• fT!"' ' ..,. ·i- ~ - i-' !!tJ'1--j r _. . -. ": • t:::_"' "/ .. Io..Ltta- - JI~ ----ljH• 1 ----...!..- l ---- .'-- ---· - ---- 1 -~--- --- - + ---.. ·

-' r._ __ __ 1 _______ 1---+---.

-'

t-

---..

4111+ - --'" · '-,

;:r-c • ..." . _ -. 11 .)1 "J . \. -.,n fA,, ,~ - ---~ --- --- ~- ---1--~---l-- - -l--- - ~--- - r--f----

_

_

..,

r

_

•

, . " . _ , .... , ", L r-"' .i . • . . . · -·-·· · -: __ - _- · -l ..j.. -_-_ _ -_- -fl~~~ -- - .:.._ - ..j.. -t - --- ... _ _.;;.._ .:... -- ---- · -- - -! -- - ---

_

-:_-·

_

-

-

- +~---:_ - ---~---- --t +r---- ----~ -..--

-+

-1---4---

--t---

:.-AJ

;;t

_:

. ,.

~

..

T,.

---

~

-

,

(')

,

•

r

~-r--

i ·

= .·

. -- - - : - -

---

--

1

--

#-

--

-1

-,

~,--_

---:

-

1 - - -- - -- l --- · -- -' -- l- __ __ .:._ , __ _

-

,l.r

~ M ., "' "' ~ \ ~ ö ~r--" ; ·- - - -- - - ._ _.l__ ~ ·----~-+-,.,.._

_____

-+- --]. - ·-1 -· -'· . '··i -I ·· - --•r ~-r-f · 'l'" ·-"" -~"---j---l---t---l --__:

-__!~L I ._:_,...--~....,._-f---\ · ~ .... --- + - ,- ---. . · -· .· .._~-~ 1 -• -• -• ~fi, .. L ._ ,,_I _--- f- - --1---! __! ~ 1 · - --: -. -~ .

-r:.

-,

u · r - --~ lfiii'Jiw'_' h 1

-~'1-1"

·

~

=

-..,--J--

---

----· -- , -1 --1 ---l !f ---+ - 1

-

-=::--·

~"

:_

.

~

;

.

L I

r

-1

-trV I 1 -'. _ i 7 L I (I.. '7 _ ' Jl')f'lr _ •:J"-'ti ,_ __ _ I'•::.#"''• · ...W ~ ~ J -TL.ooii7.~:T"' · -z-f-AT . . 7'l.,.~~ ;:__ • ~ ...

1

-_! __ ,; • ·-;r , è -" ' "" I F1/l 1 J . l.J-1/"1 _ --t-T ( .~_ -= ~ -L 1 - J'r-_ -- t- -- -f- -- I--,II-,L..-.!.I r--+l_· -- "'- :_ "' :__:_: ' -_; -...::r - -- -· -t - -- - -- - _-7--t ---; 1 7 ' I t · •y -~ · -· (')' ~ ~ G~~~\~ • 1 -I 1' -t-1f 91Wt-/\Vl.,ll"fo 'MI" . 7- u , - 7 J "' L . -- ·. ~ :-- ·I- 7 I' , ~--~---1~ -~ -~--~-~~ - -- ~~· - - -

--__

--j~ -~~ -~~-.~~~ - ~~ - ---~~~ ~~--~-~~~~ - ---:~~~ -jT

__

~ _~ · ---· ~~ ~ r - --, ~ -~ ~ - ~ - ----~ ---l--- -~----~ 1-·--- 0 T 0 () _ I /' ' • ! -.

~)&;~~~-~-~i

;

• ~ f . .. , l , , ,._ " ... ~- •

1

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I