Ultramicroscopy 3 (1978) 185-189 0 North-Holland Publishing Company
RADIATlON DAMAGE DUE TO KNOCK-ON LIQUID HELIUM TEMPERATURE
PROCESSES
ON CARBON
FOILS
COOLED
TO
I. DIETRICH *, F. FOX *, H.G. HEIDE +, E. KNAPEK * and R. WEYL * * Forschw~gslaboratorien der Siemens AG, Miinchen, Fed. Rep. Germany + Institut fiir Eiektronenmikroskopie am Fritz-Haber-Institrct der Max-Planck-Gesellscltaft,
Berlin,
Fed. Rep.
Germany
Received 6 February 1978
Radiation damage on a holey carbon foil was investigated in an electron microscope with a superconducting lens system, where the temperature of the specimen and its environment initially was 4 K. Due to an electron dose of 2 X lo4 As/cm2 the diameter of a hole increased 5 nm. Rough calculations show that this increase can be ascribed to knock-on processes. Estimates of the rise in specimen temperature during the irradiation are given.
of the crystal structure as observedby electron diffraction damage,the improvement factor is about five at best [3]. However, for a number of biological specimensthe masslossshouldbe more representative of the lossin information due to electron microscopicalimaging.In the latter casethe situation is much better,becausethis masslosscan indeed be stopped by cooling the specimen[4-71, and thus it shouldbe possiblethat the molecular architecture remainsmore or lessunchanged.But this pertains only to ionization-induced massloss, and not to nuclear displacementprocesses. One of our goalswith the superconducting lens systemis atomic resolution in decorated organic specimens.If imagesof heavy atomsgive sufficient information, and if the specimenand its surroundingsare kept at 4 K, the radiation damage shouldbe essentiallycausedby knock-on processes.For this reasonwe are very much interested in methods for measuringthe knock-on effect. For investigating the masslossdue to the knock-on effect, certain requirementshave 61 be fulfilled. The low temperature of the specimenmust prevent evaporation, and the residualgaspressurein the environment of the specimenmust be solow that neither contamination nor etching can take place. As pointed out in former papers [8], theseconditions are met in our 400 kV microscopeequipped with a superconductinglenssystem.Therefore one
1. Introduction
It is well known that the main problem in high resolution
electron microscopy is radiation damage, especiallyfor the investigation of biological speci-
mens.The damageis initiated by ionization and breaking of chemicalbonds due to the energy transfer by inelastically scattered electrons, and the mechanisms of the ensuing reactions are very com-
plicated. Another type of damageis causedby knock-on processes.In this casenuclei in the specemenare displaceddue to direct impacts of high energy electrons. The influence of the beamvoltage on the radiation damage,however, will not be discussedhere. Severalmethods for the determination of these radiation damageshave been used,such asmassloss and lossof specific elements;and the various changesin the image,in the electron diffraction pattern, and in the infrared, visible, ultra violet or electron energy lossspectra.The resultsof the experiments point out that the electron rate permitted for imagingspecimensat room temperature without radiation damageis of the order of 1O-4-1O-2 As/cm* for aliphatic compoundsand lo-‘-10” As/cm’ for aromatic compounds[ 1,2]. These resultswere achievedon undecorated specimens. Cooling the object to liquid helium temperature will improve the radiation resistance.When the collapse 185
186
I. Dietrich
et al. /Radiation
is alble for the first time to observedirectly in the ITliC :roscopethe masslossdue to knock-on collisions. Thr: experiment we carried out wasto determine the increasein the diameterof a hole in a carbon foil due to strong electron irradiation.
damage
on carbon
foils
2. Experiments After adjustingthe microscope,a focus series was taken with a relatively smallillumination aperture. Subsequentlythe specimenwasirradiate:d
(b)
Fig. 1. Carbon foil (a) before and (b) after irradiation.
I. Dietrich et al. /Radiation
with a higher electron current density for 1.5 h. At the end of this period a further focus serieswas photographed. A micrograph of each seriesis reproduced in figs. la and lb. The electron microscopic data relevant for the evaluation of the pictures are: (a) beamvoltage 230 kV; (b) electron optical magnification 500 000 : 1; (c) current density during exposure of photographic emulsioni = 0.6 A/cm*; (d) (illumination aperture 2 X 10s4rad) case1; (e) (illumination aperture 6 X 10m4rad) case11; (f) current density during strong irradiation i = 4 A/cm* ; and (g) exposuretime 30 s. To carry out an appropriate comparisonof the micrographstaken before and after strong irradiation, we had to be surethat no changein magnification occurred. Since the magnifying lensesare superconductingand are driven in the persistent current mode, a drift in focal length can be excluded. Beamvoltage variation could be eliminated by choosinga picture with the samedefocus distance, establishedby the ring systemin the light optical diffractograms. In addition, it waspossibleto test the absolutevalue of the magnification because,by chance,crystalline gold wasdepositedon top of the carbon fiim so that the 0.204 nm lattice distance appearedin the image. As shown in figs. la and lb, the hole diameter increased5 nm during a radiation exposure of 2.2 X lo4 As/cm*.
damage on carbon foils
181
electron current density i, and the effective cross section for atomic displacementu, which can be calculated from well known relations [9-l 11. For elementswith low atomic number 2 0=2.5x
lo-*5 xzz-
1 -p* p’
x
[
X-
2x(X1/2 137
1-P*lnX+fl
- 1 -a InX)](l)
(a in cm*) is valid, /I = u/c can be determined from p* = U(U-+ 1.02)&Y + 0.51)2 taking the electron energy U in MeV. The parameterX = E,/Ed with the maximum transferrable energy E, E,=2147XU(Ut
1.02)/A
in eV, if U in MeV (A atomic weight). The displacementthreshold energy Ed, i.e. the minimum kinetic energy to be transferred to a lattice atom for displacingit, is not explicitly given in the literature for amorphouscarbon fdms. In table 1 somelikely valuesof Ed and the effective crosssection cr for three different electron energies are listed. The most probable value of the threshold energy seemsto be Ed = 10 eV according to Scherzer [ 121. We alsoconsideredsecondary knock-on processes,and for this reasonwe had to multiply u ascalculated from eq. (1) with 0.31 t 1nX [13]. For our experimental conditions the effective knock-on crosssection of carbon should be 1 X 10-** cm*, and consequently an irradiation for 1.5 h
3. Discussionof results For calculating the enlargementof such a hole due to knock-on processes,we start from a simplified model. We assumethat the carbon film consistsof monoatomic layers piled up, one above the other, and take it for granted that only in the surface layer are displacedatoms detachedfrom the film after the impact. According to theseconsiderations, which are appliedto the atoms displacedin the surface layer of the edgeof the hole, the changein diameter causedby knock-on processesshould be independentof the film thickness. The displacementrate is given by the product of
Table 1 Cross section (in barns) for displacement of C-atoms by fast electrons. Secondary knock-on processes are also considered. U (keV)
100 250 500
Ed
’
Em @VI
5 eV
1OeV
20 eV
30 eV
312 341 364
56 105 133
24 42
I 19
1 barn = 1O-24 cm*.
20.1 56.9 136.2
188
I. Dietrich
et al. /Radiation
with i = 2 A/cm2 shouldremoveelevenmonolayers from the surface.Assuminga density of the carbon. foil 2 g/cm3, an increasein hole diameter of about 4.5 nm results(experimental value 5 run). Taking into account the crudenessof the model, the agreement betweentheory and experiment is excellent. Although the knock-on effect cannot beoverlooked in practical high resolution electron microscopy, the damageis negligiblein general.After an irradiation with a doseof .20 As/cm2 at a beamvoltage of 230 kV only one of 90 carbon atomsis displaced. The relation usedaboveis not applicablefor heavy atomswithout modifications. Yet if the beamvoltage is not too high, the maximum transferred energyE, becomessosmall [for atomswith Z > 70, E, (250 kV)
RADIATlON DAMAGE DUE TO KNOCK-ON LIQUID HELIUM TEMPERATURE
PROCESSES
ON CARBON
FOILS
COOLED
TO
I. DIETRICH *, F. FOX *, H.G. HEIDE +, E. KNAPEK * and R. WEYL * * Forschw~gslaboratorien der Siemens AG, Miinchen, Fed. Rep. Germany + Institut fiir Eiektronenmikroskopie am Fritz-Haber-Institrct der Max-Planck-Gesellscltaft,
Berlin,
Fed. Rep.
Germany
Received 6 February 1978
Radiation damage on a holey carbon foil was investigated in an electron microscope with a superconducting lens system, where the temperature of the specimen and its environment initially was 4 K. Due to an electron dose of 2 X lo4 As/cm2 the diameter of a hole increased 5 nm. Rough calculations show that this increase can be ascribed to knock-on processes. Estimates of the rise in specimen temperature during the irradiation are given.
of the crystal structure as observedby electron diffraction damage,the improvement factor is about five at best [3]. However, for a number of biological specimensthe masslossshouldbe more representative of the lossin information due to electron microscopicalimaging.In the latter casethe situation is much better,becausethis masslosscan indeed be stopped by cooling the specimen[4-71, and thus it shouldbe possiblethat the molecular architecture remainsmore or lessunchanged.But this pertains only to ionization-induced massloss, and not to nuclear displacementprocesses. One of our goalswith the superconducting lens systemis atomic resolution in decorated organic specimens.If imagesof heavy atomsgive sufficient information, and if the specimenand its surroundingsare kept at 4 K, the radiation damage shouldbe essentiallycausedby knock-on processes.For this reasonwe are very much interested in methods for measuringthe knock-on effect. For investigating the masslossdue to the knock-on effect, certain requirementshave 61 be fulfilled. The low temperature of the specimenmust prevent evaporation, and the residualgaspressurein the environment of the specimenmust be solow that neither contamination nor etching can take place. As pointed out in former papers [8], theseconditions are met in our 400 kV microscopeequipped with a superconductinglenssystem.Therefore one
1. Introduction
It is well known that the main problem in high resolution
electron microscopy is radiation damage, especiallyfor the investigation of biological speci-
mens.The damageis initiated by ionization and breaking of chemicalbonds due to the energy transfer by inelastically scattered electrons, and the mechanisms of the ensuing reactions are very com-
plicated. Another type of damageis causedby knock-on processes.In this casenuclei in the specemenare displaceddue to direct impacts of high energy electrons. The influence of the beamvoltage on the radiation damage,however, will not be discussedhere. Severalmethods for the determination of these radiation damageshave been used,such asmassloss and lossof specific elements;and the various changesin the image,in the electron diffraction pattern, and in the infrared, visible, ultra violet or electron energy lossspectra.The resultsof the experiments point out that the electron rate permitted for imagingspecimensat room temperature without radiation damageis of the order of 1O-4-1O-2 As/cm* for aliphatic compoundsand lo-‘-10” As/cm’ for aromatic compounds[ 1,2]. These resultswere achievedon undecorated specimens. Cooling the object to liquid helium temperature will improve the radiation resistance.When the collapse 185
186
I. Dietrich
et al. /Radiation
is alble for the first time to observedirectly in the ITliC :roscopethe masslossdue to knock-on collisions. Thr: experiment we carried out wasto determine the increasein the diameterof a hole in a carbon foil due to strong electron irradiation.
damage
on carbon
foils
2. Experiments After adjustingthe microscope,a focus series was taken with a relatively smallillumination aperture. Subsequentlythe specimenwasirradiate:d
(b)
Fig. 1. Carbon foil (a) before and (b) after irradiation.
I. Dietrich et al. /Radiation
with a higher electron current density for 1.5 h. At the end of this period a further focus serieswas photographed. A micrograph of each seriesis reproduced in figs. la and lb. The electron microscopic data relevant for the evaluation of the pictures are: (a) beamvoltage 230 kV; (b) electron optical magnification 500 000 : 1; (c) current density during exposure of photographic emulsioni = 0.6 A/cm*; (d) (illumination aperture 2 X 10s4rad) case1; (e) (illumination aperture 6 X 10m4rad) case11; (f) current density during strong irradiation i = 4 A/cm* ; and (g) exposuretime 30 s. To carry out an appropriate comparisonof the micrographstaken before and after strong irradiation, we had to be surethat no changein magnification occurred. Since the magnifying lensesare superconductingand are driven in the persistent current mode, a drift in focal length can be excluded. Beamvoltage variation could be eliminated by choosinga picture with the samedefocus distance, establishedby the ring systemin the light optical diffractograms. In addition, it waspossibleto test the absolutevalue of the magnification because,by chance,crystalline gold wasdepositedon top of the carbon fiim so that the 0.204 nm lattice distance appearedin the image. As shown in figs. la and lb, the hole diameter increased5 nm during a radiation exposure of 2.2 X lo4 As/cm*.
damage on carbon foils
181
electron current density i, and the effective cross section for atomic displacementu, which can be calculated from well known relations [9-l 11. For elementswith low atomic number 2 0=2.5x
lo-*5 xzz-
1 -p* p’
x
[
X-
2x(X1/2 137
1-P*lnX+fl
- 1 -a InX)](l)
(a in cm*) is valid, /I = u/c can be determined from p* = U(U-+ 1.02)&Y + 0.51)2 taking the electron energy U in MeV. The parameterX = E,/Ed with the maximum transferrable energy E, E,=2147XU(Ut
1.02)/A
in eV, if U in MeV (A atomic weight). The displacementthreshold energy Ed, i.e. the minimum kinetic energy to be transferred to a lattice atom for displacingit, is not explicitly given in the literature for amorphouscarbon fdms. In table 1 somelikely valuesof Ed and the effective crosssection cr for three different electron energies are listed. The most probable value of the threshold energy seemsto be Ed = 10 eV according to Scherzer [ 121. We alsoconsideredsecondary knock-on processes,and for this reasonwe had to multiply u ascalculated from eq. (1) with 0.31 t 1nX [13]. For our experimental conditions the effective knock-on crosssection of carbon should be 1 X 10-** cm*, and consequently an irradiation for 1.5 h
3. Discussionof results For calculating the enlargementof such a hole due to knock-on processes,we start from a simplified model. We assumethat the carbon film consistsof monoatomic layers piled up, one above the other, and take it for granted that only in the surface layer are displacedatoms detachedfrom the film after the impact. According to theseconsiderations, which are appliedto the atoms displacedin the surface layer of the edgeof the hole, the changein diameter causedby knock-on processesshould be independentof the film thickness. The displacementrate is given by the product of
Table 1 Cross section (in barns) for displacement of C-atoms by fast electrons. Secondary knock-on processes are also considered. U (keV)
100 250 500
Ed
’
Em @VI
5 eV
1OeV
20 eV
30 eV
312 341 364
56 105 133
24 42
I 19
1 barn = 1O-24 cm*.
20.1 56.9 136.2
188
I. Dietrich
et al. /Radiation
with i = 2 A/cm2 shouldremoveelevenmonolayers from the surface.Assuminga density of the carbon. foil 2 g/cm3, an increasein hole diameter of about 4.5 nm results(experimental value 5 run). Taking into account the crudenessof the model, the agreement betweentheory and experiment is excellent. Although the knock-on effect cannot beoverlooked in practical high resolution electron microscopy, the damageis negligiblein general.After an irradiation with a doseof .20 As/cm2 at a beamvoltage of 230 kV only one of 90 carbon atomsis displaced. The relation usedaboveis not applicablefor heavy atomswithout modifications. Yet if the beamvoltage is not too high, the maximum transferred energyE, becomessosmall [for atomswith Z > 70, E, (250 kV)
