Final Project:



Final Project:

The Determination of the Iron Content of Hemoglobin

Group R3

Noel Ang

Lisa Romito

Karim Sadak

Ofer Sagiv

4/ 30/ 98

Abstract

The objective of this experiment is to determine the actual weight percent of iron (Fe) in a bovine hemoglobin molecule. Through the use of Atomic Absorption Spectrophotometry (AAS), the weight percent of Fe was calculated to be about 0.331 ( 0.021%. This value can be compared to the literature value of Fe weight percent of 0.333%, which is based on a 67,000 grams/mole molecular weight of hemoglobin. This presents an approximate error of 0.6%. Furthermore, the actual number of Fe atoms within a molecule of hemoglobin protein was obtained; this value was 3.97 ( 0.27 Fe atoms per hemoglobin molecule, a 0.75% error when compared to the actual presence of 4 Fe atoms per hemoglobin molecule. To determine the Fe content in hemoglobin, the hemoglobin molecule was kept intact. To prove that the rest of the protein did not interfere with the absorption readings measured by the AAS, the method of additions was utilized in which iron was added to hemoglobin in increasing amounts. This method showed that the rest of the hemoglobin protein surrounding the iron atom did not interfere with the absorbance readings, which therefore validates the above results.

Table of Contents

I. Abstract 1

II. Background 3

Hemoglobin 3

Atomic Absorption Spectrophotometry 4

Biomedical Relevance 5

III. Apparatus and Materials 6

IV. Procedure 7

Create the Calibration Curve for Iron (Fe) 7

Determine the Effect of Hemoglobin on AAS 7 Calculate the Weight Percent of Iron (Fe) in Hemoglobin 9

Determine the Number of Iron (Fe) Atoms

per Hemoglobin Molecule 9

V. Results 10

Create the Calibration Curve for Iron (Fe) 10

Determine the Effect of Hemoglobin on AAS 11

Calculate the Weight Percent of Iron (Fe) in Hemoglobin 13

Determine the Number of Iron (Fe) Atoms

per Hemoglobin Molecule 13

VI. Discussion 15

Suggestions 16

VII. Error Analysis 17 Error in Hemoglobin Concentrations 17

Error in Iron (Fe) Concentrations 17

Error in the Weight Percent of Iron (Fe) 18

Error in the Number of Iron (Fe) Atoms 20

VIII. References 21

IX. Appendix 22

Background

Hemoglobin

The average human body contains about four to six liters of blood. In this blood, there are 25 trillion tiny erythrocytes, or red blood cells. Each red blood cell contains about 250 million molecules of hemoglobin. (5) This translates into the astonishing fact that there are 6.25 x 1021 molecules of hemoglobin in the human body. Due to its impressive quantities alone, it is clear that hemoglobin is a primary player in the overall function of the human body. More significantly, hemoglobin’s main function is to bind oxygen in organs of all vertebrates and some invertebrates. It also serves to pick up carbon dioxide that has been secreted from organs.

Hemoglobin is a globular protein with four polypeptide chains, each of which contains a non-polypeptide component called a heme. A heme is a nitrogen containing, disc-shaped organic molecule that is centered around an iron (Fe) molecule. The Fe is in a ferrous state (Fe2+).

Hemoglobin is probably most significant when associated to the function of respiratory organs. When blood passes through the alveoli in the lungs, oxygen is chemically bound to the protein. This bond is due to the sharing of electrons between iron and oxygen. The product of this reaction is oxyhemoglobin.

Hb + O2 ( HbO2 (Eq. 1)

Equation 1. Chemically bound blood which contains hemoglobin (Hb) and oxygen results in oxyhemglobin.

The oxygen is carried to the entire body via the circulatory system. When the oxyhemoglobin passes through cells with excess CO2, which at times has been produced via glycolysis, the O2 is encouraged to release into the subsequent organs. At this point the oxyhemoglobin becomes deoxyhemoglobin.

HbO2 ( Hb + O2 (Eq. 2)

Equation 2. This equation states the return of hemoglobin from oxyhemoglobin, the reverse of Equation 1.

The iron remains reduced and consequently, the CO2 attaches itself to the hemoglobin for transportation back to the lungs. The reverse happens in the lungs, where the oxygen facilitates the release of CO2. (7)

Atomic Absorption Spectrophotometry

The principal instrument used in this experiment was the Perkin-Elmer Model AA 4000 Atomic Absorption Spectrophotometer. AAS is an analytical method used to determine the amount of substance in a sample through radiation emission and absorbance. In this process, an electron of the element to be tested enters an excited state. This is brought about by the change in its energy, and it consequently emits a wavelength characteristic of that specific electron. This wavelength of light is passed through a flame containing the same element, causing the electron to gain energy through absorption and subsequently return to its initial state. Since more light is absorbed by substances in which more atoms are present (or in which the concentration is higher), one can use the data obtained from a spectrophotometer to determine the concentration of ions of a specific element in a substance. (1)

In effect, absorbance is related to concentration through the concept of transmittance. Transmittance relates the energy of the light beam before and after passing through the sample. This relationship is shown by the Beer-Lambert law:

I = I0 e(-k( C b) (Eq. 3)

Equation 3. This equation relates inlet (I) and outlet (I0) intensities through variables of the proportionality constant, molar concentration of absorbed atom and light path length.

ln T = ln (I/I0) = -k( C b (Eq. 4)

Equation 4. Through Equation 3, the log of the transmittance T, can be calculated.

I and I0 represent the intensity of the light beam before and after it goes through the sample of gas. T is the fraction transmitted or transmittance, k( is a proportionality constant which depends on the wavelength of the initial light beam, (. The length of the light path is b and C is the molar concentration of absorbing atoms. This shows that absorbance is related to concentration through the concept of transmittance. Using base ten instead of base e logarithms, the relationship becomes Equation 5:

A = - log T = (( C b (Eq. 5)

Equation 5. The Beer-Lambert Law: in which A is the absorbance, T is the transmittance, ελ is the molar absorption coefficient, C is the concentration in moles, and b is the thickness of the light path. From this equation, one can see that the relationship between absorption and concentration is directly proportional. (1)

This relationship between concentration and absorbance will play a central role in this project as seen in the analysis of the resulting data.

Biomedical Relevance

The AAS is frequently used in the biomedical field in forensic, toxicological, and clinical analysis. It is often necessary to determine the concentration of a certain element in a living system to conclude whether the amount present is too toxic, or conversely, if it is not concentrated enough. With this method, one can quickly and efficiently determine concentrations of common or toxic substances in the body. With the aid of AAS, doctors can more easily and accurately make a diagnosis, monitor a patient’s progress, and determine a proper treatment.

It is also important to note that the iron content in hemoglobin has great biomedical relevance. Those who have an iron deficiency in the erythrocytes of their blood face a grave danger. This condition, which causes anemia, is the most common nutritional deficiency in the world. Anemia only occurs towards the end of a process of decreasing iron in hemoglobin, which in some cases may have been in progress for many years. In terms of clinical medicine, it is of great importance to be able to determine how much iron is present in blood. A person may have vague symptoms like lethargy and dizziness; thus there must be a quick and efficient way to detect Fe amounts in blood. Once the iron depletion is diagnosed, therapeutic intervention such as dietary manipulation can be implemented in the life of an afflicted patient. This way the amount of iron in their blood can be raised to a healthy and productive level.

Apparatus and Materials

The following materials will be found in the BE laboratory in order to conduct this experiment:

• Hemoglobin (H2500) Bovine (Lyophilized powder form)

• Perkin-Elmer Model AA 4000 Atomic Absorption Spectrophotometer

• Standard solution of Fe (at 1000 ppm) to make the calibration curve

• Iron nitrate (Manufacturer: Fisher. Product number: I110-100 CAS number: 7782-61-8)

• Several disposable 50mL test tubes with blue caps to hold dilutions

• Various beakers to hold the deionized water (including 250mL, 300mL, and 1000mL)

• Automatic pipettes (only 20 – 200 μL, 100 – 1000 μL, and 2 – 20 μL pipettes in order to strengthen accuracy)

• Analytic balance (accuracy of 0.0005 gram)

• Magnetic stirrer

• Metal spatula to obtain the hemoglobin

• Test tube racks

• Funnel

Procedure

Create the Calibration Curve for Iron (Fe)

For each of the three weeks of experimentation, a calibration curve for the Fe stock solution was constructed. In Week 1, the 1000 ppm stock of iron nitrate (FeNO3) was diluted with deionized water to make 13 solutions. The table below illustrates the calibration solutions made for the construction of the first calibration curve.

When creating the diluted solutions, micro-pipettes were used to add the desired volumes of stock and some of the water (the greater volumes of water were added with the air suction pipettes). However, the analytic balance was employed to measure exact amounts of the Fe stock and water. Table 1 was subsequently adjusted with these precise weight measurements, and absorbance readings of the above solutions were then recorded. The calibration curve was then plotted using Microsoft Excel. This same procedure was performed for the two other calibration curves for Weeks 2 and 3. The calibration curve of Week 2 was constructed over absorbances that ranged from 0.005 to 0.02, and the calibration curve for Week 3 was constructed over absorbances that ranged from 0.01 to 0.2.

Determine the Effect of the Hemoglobin on the AAS

Next, the effect of hemoglobin protein on the AAS was determined by using the method of additions (see Appendix). Three trials were conducted in which iron was added to solutions containing constant hemoglobin concentrations. In order to achieve this constant hemoglobin concentration, a stock hemoglobin solution was made for each trial.

Leaving one solution as the control, an increasing amount of Fe was added to each of the remaining solutions. In trial 1 for example, Fe was added to the remaining solutions according to Table 3.

As can be seen, the solutions were made according to their predicted absorbances so that a linear plot of equally spaced data points would result.

The iron from the Fe stock, hemoglobin from the hemoglobin stock, and water were added in that precise order for each of the trials. All additions were made on the analytic balance for greater accuracy. The amount of Fe stock was added first using the micro-pipettes, and the concentrations were adjusted accordingly as done with the calibration curve. The amount of hemoglobin stock added in trial 1 was 23 mL. In trials 2 and 3, 60 mL of hemoglobin stock were added to each tube as seen in Table 2. These amounts were added using the air suction pipettes and micro-pipettes as to achieve almost constant hemoglobin in all solutions (note: the balance changes it’s accuracy at 60 mL from 0.0005 to 0.005 mL or mg). The water was added through a funnel and a beaker. When the total volume of the solution approached about 175 mL, micro-pipettes were employed to achieve a constant total volume of approximately 180 mL. This prevented accidental over-pouring when using just the funnel and beaker. Finally, all solutions were mixed using the magnetic stirrer and the absorbances were taken. This was determined to be the best method in making the solutions for the method of additions.

Using Microsoft Excel, a linear relationship of only the Fe concentration vs. the absorbances of the solutions was plotted. The absorbance of the control and the y-intercept of that line were then compared.

Calculate the Weight Percent of Iron (Fe) in Hemoglobin

To determine the weight percent of iron, 16 solution of varying hemoglobin concentrations were made over the 3 weeks. To make these solutions, each beaker was placed on the analytic balance and the balance was tarred. Hemoglobin was added and the measured amount was recorded. The amount of hemoglobin in each of the solutions made was approximately 35 mg in 180 mL of water. A concentration of 200 ppm of hemoglobin was not exceeded to ensure that the AAS was not damaged. Because the balance readings fluctuated, the glass walls were put up to eliminate any interference from the surrounding air. After the solutions were mixed thoroughly with the magnetic stirrer, four absorbance readings were taken and then averaged. The corresponding calibration curves for each week were used to determine the Fe concentration in the solutions from the recorded absorbances. The ratio between the concentration of the iron and the concentration of the hemoglobin (both in ppm’s) was calculated, giving the appropriate weight percent of Fe in hemoglobin.

Determine the Number of Iron (Fe) Atoms per Hemoglobin Molecule

Using this experimental weight percent, various calculations (see Results) were carried out to find the number of Fe atoms per hemoglobin molecule. All values were average and analyzed for error.

Results

Create the Calibration Curve for Iron (Fe)

Figure 2 is one of the three calibration curves that were made for each of the three weeks of experimentation.

As seen in the figure above, the calibration curve becomes non-linear at absorbency values of approximately 0.3A. For this reason only the first five data points were used to find a linear relationship between the concentration of iron and the absorbance of the Fe solutions. In order to convert absorption units (A), y, to concentrations (ppm), x, for the hemoglobin solution of Week 1, Equation 6a below was used. The relationship for Week 2 and Week 3 is represented by Equation 6b and Equation 6c, respectively. These relationships were used to find the iron concentrations in the hemoglobin solutions made in Week 2 and Week 3.

y = 0.0302 x (Eq. 6a)

y = 0.0366 x (Eq. 6b)

y = 0.0285 x (Eq. 6c)

Determine the Effect of Hemoglobin on AAS

As seen in Table 2 (see Procedure), three trials were conducted in which iron was added to solutions containing constant hemoglobin concentrations. In each case the y-intercept of the best-fit line relating the Fe concentrations with the absorbances was compared to the absorbance of the control in which no Fe was added.

In trial 2 for example, seven mixtures of Fe and constant hemoglobin were made and their absorbances were taken. The control contained only hemoglobin and absorbed at 0.022 absorbance units (A). It can be seen from Table 4 that the measured absorbance values for these solutions deviated significantly from the predicted absorption starting at mixtures of greater absorbance than 0.1A (the predicted absorbance values were calculated using the calibration curve for Week 2).

The graphs, which illustrate the method of additions, are shown for trial 2 below. Figure 3 relates the added Fe concentrations of the solutions and their measured absorbances. It can be seen that the y-intercept is 0.0271A. However, the control absorbed at 0.022A. This value contains approximately 23% error. Excluding the mixtures which absorb at values higher than 0.1A and subsequently fitting a line through those points below 0.1A results in a y-intercept of 0.0217 which has approximately 1.3% error from the control (see Figure 4).

The same trend can be seen in trials 1 and 3. In trial 1, the control absorbed at 0.024A and the y-intercept of the corresponding curve to Figure 3 was 0.0258A. Because only few mixtures which absorb above 0.1A where made, an error of 7% was seen when all the data was included. Excluding the two mixtures which did have an absorbance above 0.1A resulted in a y-intercept of 0.0245A, an error of only 2%. In trial 3, including all solutions in the linear plot gave an intercept with an error of 31%; again, excluding all solutions which absorb above 0.1A resulted in an intercept with an error of 3%. Table 5 summarizes the results for all three trials.

Calculate the Weight Percent of Iron (Fe) in Hemoglobin

As said previously, 16 hemoglobin solutions were made and using their absorbencies, their iron concentrations were calculated using equations 6a, 6b, and 6c (see Table 6). The weight percent of the iron in hemoglobin could then be found through the ratio of the iron concentration of each solution to their respective Hemoglobin Concentration. This can be seen through the following ratio:

Equation 7. The weight percent of iron, Wt%Fe, was found through the ratio of iron concentration, [C]Fe, to the hemoglobin concentration, [C]hemo, which were both in ppm’s.

After averaging the ratios for all 16 solutions the resulting value for the weight % in Iron was 0.331% ( 0.021. This value can be compared to the literature value of Fe weight percent of 0.333%, which is based on a 67,000 grams/mole molecular weight of hemoglobin. This presents an approximate error of 0.6%.

Determine the Number of Iron (Fe) Atoms in a Hemoglobin Molecule

With the weight percentages known for all the 16 solutions, the actual number of iron atoms in a hemoglobin molecule can then be calculated. This can be fulfilled through a simple series of equations. The calculations are illustrated in Equations 8a to 8c:

Equation 8a. The grams of hemoglobin in each sample, which was carefully measured on the balance, can be multiplied by the experimental weight percent of hemoglobin found previously to achieve the mass of iron.

Equation 8b. The moles of iron can be found by dividing the mass of the iron by the molecular weight of iron, which is 55.85 grams/mole.

Equation 8c. Avogadro’s Law is then employed that basically states that a mole is a mole, no matter the material. Therefore, the moles of iron over the moles of measured hemoglobin will give a molar ratio, specifically the number of iron atoms per hemoglobin molecule.

This group of calculations yielded a value of 3.97 ( 0.27 Fe atoms per hemoglobin molecule when taken averaging the values derived for all 16 solutions. The percent error, when compared to the actual number of 4 atoms per molecule, was 0.75%. (see Table 7 for derivation of this average).

Discussion

As seen in Figure 2, the curve at lower concentrations exhibits a linear behavior, but as the concentration increases, the slope of this calibration curve decreases and approaches zero. This behavior is analogous to enzyme activity; at some maximum point of substrate concentration, the reactions of the enzymes with the substrate will reach saturation; any following increase in substrate concentration will not increase the rate of the reaction. Light transmitted by the spectrophotometer is analogous to this enzyme activity. The reason for the nonlinearity at higher concentrations (above A > 0.3) of the pure Fe calibration curve is due to the effect called coincidence. As more iron atoms are added to the solution and the concentration increases, there is only so much light that can be absorbed, therefore the curve tapers off to the nonlinear region. This nonlinear region differs for each element. For example, sodium’s (Na) calibration curve becomes nonlinear around absorbances of 0.6A. (4) The reason for this difference may be due to the size of the atom. Fe is larger then Na and therefore absorbs more light with the same increase in concentration.

In carrying out the method of additions it is important to note a few key points regarding minimization of error. Instead of weighing a constant solid hemoglobin and placing it in each of the beakers in the three trials, a hemoglobin stock solution was made. This eliminated the large error that would result in the weighing of such small solid quantities in the order of milligrams; instead, hemoglobin was added by volume in milliliters. As seen in Table 2 (see Procedure), enough water was added to the hemoglobin so the stock would not run out when adding a constant amount of hemoglobin to each tube. The exact amount of hemoglobin that was added to each tube within each trial was not important; what is important, however, is that the hemoglobin concentration is constant within each trial. In addition, as can also be seen from Table 2, the hemoglobin concentrations were made so the maximum concentration would not exceed 220 ppm. Higher hemoglobin concentration than this limit might damage the AAS and produce inaccurate readings.

To see a broader picture of the phenomenon showed in the method of additions (see Results), a comparison between the calibration curve of pure Fe and the curve of total iron concentration of all mixtures made in this experiment (with their absorbances) is shown as Figure 4. It is seen from the above figure that the mixture line deviates from the calibration curve at absorptions above 0.1A. In addition, the mixture line changes linearity much faster than the calibration curve. If any of the hemoglobin protein affected the AAS readings, a distinct shift along the x-axis by the mixture line would be observed; in turn, the two curves would not converge. The fact that these two lines do converge at their respective regions of linearity shows that the deviation is a result of the physical presence of the protein, which increases the viscosity of the solutions, and not because the hemoglobin interfered with the Fe absorptions. In this case, the calibration is affected by only coincidence while the mixture line is effected by both coincidence and the viscosity. Hence, the mixture line becomes nonlinear at a lower absorbance of 0.1A compared to 0.3A for the calibration curve.

The specific aim of this project focuses on finding the weight percent of iron in a hemoglobin molecule. To minimize error in determining this quantity several key points should be noted. Different calibration curves were made each week in order to account for the intrinsic drift within the AAS. Each corresponding calibration curve was then used to find the iron concentration in the 16 solutions over the three weeks of experimentation. Because it was shown that the viscosity of the iron and hemoglobin mixtures affected the absorption readings at A > 0.1, the 16 hemoglobin solutions were diluted to absorptions of approximately 0.01 < A < 0.02. This was carried out in order to ensure that the viscosity of the solutions did not interfere with the AAS readings.

Because of the high dilutions, error was a major factor in these solutions. To minimize this error, the volume of the solutions was 180 mL in each case, which reduces the error in the water. However, there was still a considerable error in the amount of hemoglobin put into the solutions due to the balance. (see Error Analysis)

Suggestions

A great value to this experiment that would minimize most of the error in the amount of solid hemoglobin would be the use of a more accurate balance. The balance currently present in the BE lab has an accuracy of 0.0005 g (and only reads to 0.001 g). Other materials such as actual human hemoglobin, rather than bovine hemoglobin, would more greatly signify the importance of this study.

Error Analysis

As in the results, there are two components that determine the weight percent of iron in hemoglobin. Those are the iron (Fe) and hemoglobin concentrations. The following shows how the error in these two components was determined. Furthermore, the analysis illustrates the determination of the error in the weight percent of Fe in hemoglobin as well as the number of Fe atoms in each molecule of protein.

Error in Hemoglobin Concentrations

The uncertainty of the hemoglobin concentrations throughout this experiment can be attributed to error in the weighing of the hemoglobin and the error in the addition of water to make the hemoglobin solutions. The analytic balance, which had an error of approximately 0.0005 g, or 0.5 mg, was the only factor which caused error in these values. As said in the procedure, the balance was relied on to measure all volumes and weights. It should be noted though, that error in the 180 mL of water used was negligible because it was such a large volume in comparison to the hemoglobin used, therefore it was discounted. The following equations resulted in order to find the upper and lower limits for hemoglobin concentration:

Equation 9a. The upper limit is found by dividing the maximum error in the mass of the hemoglobin by the total volume. In this case, the maximum error is approximately 1 mg, which is the accuracy of the scale.

Equation 9b. The lower limit is the minimum value divided by the total volume. These two equations account for the 0.001 g uncertainty of the analytic balance.

Error in Iron (Fe) Concentrations

The error in the Fe concentrations was due to the AAS display readings, which gave values to the nearest 0.001A (absorbance units). This error was accounted for and further determined graphically using the corresponding calibration curve of pure Fe for each solution as seen in Figure 6. The upper and lower limits were found by analyzing the upper and lower 95% confidence limits of the slope of the calibration curve and finding the maximum deviations.

Error in Weight Percent of Iron (Fe)

Using the previously described upper and lower limits of Fe concentration and the upper and lower limit of the hemoglobin concentrations, the upper and lower limits of the weight percent of iron in hemoglobin could be determined. Equations 10a and 10b show these calculations. Table 8 below shows the derived range in weight % from all the 16 solutions.

Equation 10a. The upper limit of the weight percent of Fe can be found by dividing the upper limit of the Fe concentration by the lower limit of the hemoglobin concentration.

Equation 10b. Conversely, the lower limit can be found by dividing the lower limit of the iron concentration by the upper limit of the hemoglobin concentration.

Error in the Number of Iron (Fe) Atoms

Error in the number of Iron atoms per hemoglobin molecule was calculated using the upper and lower limits of the weight percent for all 16 solutions. Equations 8a, 8b, and 8c were utilized once again. Table 9 shows the resulting error.

References

1. BE210 Bioengineering Laboratory II Laboratory Manual. Spring 1998;

Experiment 3.

2. Beaty, Richard D. “Concepts, Instrumentation and Techniques in Atomic Absorption

Spectrophotometry.” Perkin-Elmer; 1982.

3. perkin-

4. Ang, Noel, et. al. (Group R3) “Atomic Absorption Spectrophotometry.” Spring 1998.

5. Campbell, Neil A. Biology; Fourth Edition. Menlo Park, California: Benjamin / Cummings Publishing Company, Inc., 1996.

6. Bell, George Wall, et. al. “Absorption Spectra of Equine Hemoglobin.” Spring 1997.

7. Lemberg, & Legge. Hematin Compounds and Bile Pigments. New York. 1949.

Appendix A

Method of Additions

The Method of Additions was the procedure used to determine the effect of the Hemoglobin protein on the AAS. This method involves one control solution that contains a constant amount of substance A (the hemoglobin for this experiment). All other solutions contain the same constant amount of substance “A” as well as equally increasing increments of substance “B” (the iron). A graph of the concentration of the increasing amount of substance “B” vs. the corresponding absorbency values is then analyzed, and the y-intercept will correspond to the absorbency of a solution with only substance “A” and none of substance “B”. In this experiment, if the protein has no effect on the absorbency readings, then this intercept absorbency value and that of the control should be the same.

-----------------------

[pic]

Figure 1. A hemoglobin molecule consists of four hemoglobin hemes, one shown here. As can be seen from the figure, iron (Fe) is the centralized atom of this protein component. (6)

Table 1. This table shows the calculations to determine the correct amounts of iron stock and water needed to create the calibration solutions. As can be seen from the above table, calibration solutions were made according to the predicted absorbances so that a range of concentrations of low and high aborbances would be present.

Table 2. The above shows how the hemoglobin stock was made for each trial. Also shown is how much of the stock was added to each tube in each trial. As shown in the chart, less than the total available hemoglobin stock solution was used. The amount of hemoglobin added to each solution was calculated using the ratio of Volume added to each solution to Total volume of stock. ([C] of Hemo is based on the final volume of 180 mL.)

Table 3. The table shows how much Fe was added to each of the four solutions containing a constant amount of hemoglobin in trial 1. The desired absorbances do not include the absorbances of the hemoglobin, so the resulted absorption values would be higher. After the hemoglobin was added from the stock as described before, each solution was added to a volume of 180mL of water.

Figure 2. The above figure shows the calibration curve of Fe for the first week of experimentation. It can be seen that the curve become non-linear at absorbency values of approximately 0.3A.

Equations 6a, 6b, and 6c. Absorption units (y) can be directly related to concentrations (x) through a single coefficient. These three equations were used to determine the iron concentrations in the 16 hemoglobin solutions that were made. These values were used to determine the weight percent of Fe in hemoglobin.

Table 4. The above chart shows the predicted absorption of the Fe and hemoglobin mixture of trial 2 based on the calibration curve of Week 2. It can be seen that the predicted absorbance deviated from the measured absorbance with mixtures of A > 0.1.

Figure 3. The above graph shows the relationship between the added Fe concentrations and the measured absorbance for the mixtures made in trial 2. The control, whose absorbance was 0.022A, should have been close to the y-intercept, 0.0271A, of this linear plot. Instead a 23% error is evident.

Figure 4. Excluding all mixtures which absorb considerably above 0.1A results in a y-intercept of 0.0217A, which is similar to the control whose absorbance was 0.022A. This y-intercept contains only 1.3% error

Table 5. A summary of the results from the method of additions.

[pic]

[pic]

Table 6. The above table illustrates the absorbencies of the 5 hemoglobin solutions made in week 2. Their iron concentrations(calculated from Eq.6b) are shown. Also shown is the Hemoglobin concentration and the corresponding weight percent.

[pic]

[pic]

[pic]

[pic]

[pic]

[pic]

Table 7. The table above shows how the number of Iron atoms per molecule of hemoglobin was calculated for each one of the 16 solutions. The average yielded 3.97 Iron atoms per hemoglobin molecule

Figure 5. A clear deviation of the mixture line from the pure iron calibration line is seen. This graph relates the approximate concentrations of all mixtures and their absorbances.

[pic]

[pic]

Figure 6. This graphical method was used in which the upper and lower 95% confidence limits of the slope of the calibration curve was found. Because each hemoglobin solution had a certain absorption and corresponding concentration, two other lines could be found intersecting the point indicated by the yellow dot on the graph. Since the error in the AAS is known to be 0.001A, four more concentration values can be found, two from each new line. The highest and lowest values of these four will determine the upper and lower limit of the Fe concentration, respectively.

Table 8. This table shows the weight percent of iron in hemoglobin as well as the upper and lower limits for all 16 solutions created in this experiment. The averages were calculated and the resulting weight percent was 0.331% ( 0.021% of iron content in a hemoglobin molecule.

Table 9. This table displays the arrived value for the number of Fe atoms within a hemoglobin molecule as well as the upper and lower limits. Again, all 16 solutions were taken into consideration and the resultant average was 3.97 ( 0.27 Fe atoms per hemoglobin molecule.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download