Electron-Beam Evaporator

1. Theory

• Definition: An electron-beam evaporator melts the evaporant locally with a focused electron beam in a water-cooled copper hearth, so it reaches above 3000 ${}^{\circ }$C and can evaporate refractory metals and oxides (Darling; TU Wien).
• Emission: The vapour leaves in a lobe narrower than a cosine, ${\cos }^n$ with $n\approx 3.5$ to 4 for a 5 kW gun (TU Wien).
• Radiation: The high-voltage beam produces secondary electrons and x-rays that reach the substrate (Darling; Volmer et al.).
• Oxide damage: Bremsstrahlung from the Al target creates electron traps in gate oxides that a forming-gas anneal reduces (Ning).

2. Working Principle

• Gun: A filament at -10 kV emits electrons that a magnetic field bends through 270 degrees into the hearth, keeping the filament out of the flux; typical voltages are 8 to 10 kV (Darling).
• Hearth: Crucibles or cones sit in a water-cooled copper hearth, often a 4-pocket rotary one for multilayers in one pump-down (Darling).
• Sweep: X-Y sweeping enlarges the melt, raises the rate and allows a soak preheat (Darling).
• Rate: Typically 10 to 100 angstrom/s, with lift-off metallization often far slower (Darling; Volmer et al.).

3. Applications

• Refractory metals and oxides: W, Ta, Mo, Pt, Ti, Zr and oxides such as Al${}_2$O${}_3$, SiO${}_2$ and TiO${}_2$ (Darling; TU Wien).
• Lift-off metallization: Cr/Au contacts over PMMA patterns from the Electron-Beam Lithography Writer page (Volmer et al.).
• MOS gate metal: Al gates, with an anneal to remove radiation traps (Ning).

Gun, Hearth and Rate

Name	Formula	Related Formulae	Variables/Constants
Electron gun	$V_{fil}=-10\mathrm{kV},V_{hearth}=0$	$V_{acc}=8\text{to}10\mathrm{kV}\text{(typical)},4\text{to}20\mathrm{kV}\text{(range)}$	$\begin{array}{r}Darling:270degreebentbeam,filament\\ shieldedfromtheflux,magneticfocusand\\ X-Ysweep;Volmeretal.:4to20\mathrm{kV}in\\ use,systemsat4and8\mathrm{kV}\end{array}$ $\begin{array}{r}(computed:1\mathrm{kW}at10\mathrm{kV}needs0.1Aof\\ beamcurrent)\end{array}$
Source temperature and rate	$T_{source}>3000{}^{\circ }C$	$R_{dep}=10\text{to}100\mathring{\mathrm{A}}/s$	$\begin{array}{r}Darling:againstabout1800{}^{\circ }Cand1to20\\ \mathring{\mathrm{A}}/sforresistivesources;Volmeretal.:\\ 0.02and0.4nm/sforlift-off\\ metallization\end{array}$ $\begin{array}{r}(computed:0.02to0.4nm/sis0.2to4\\ \mathring{\mathrm{A}}/s)\end{array}$
Hearth	$\text{water-cooled Cu hearth, 4-pocket rotary}$	$\text{reactive and refractory metals: Ta, Ti, Zr, W, Mo, Pt, Rh}$	$\begin{array}{r}Darling:crucibles,conesandpyrolytic\\ graphiteliners,sequentiallayersinone\\ pump-down,oxidesA{l}_2{O}_3,SiO,Si{O}_2,Sn{O}_2,\\ Ti{O}_2,Zr{O}_2;TUWien:cooledCucruciblefor\\ reactiveandrefractorymetals\end{array}$ $\begin{array}{r}(myreading:onlythemeltishot,sothe\\ evaporantneveralloyswithaboat)\end{array}$

Emission and Uniformity

Name	Formula	Related Formulae	Variables/Constants
Electron-gun emission lobe	$\frac{dm}{d\Omega }\propto {\cos }^n\varphi ,n\approx 3.5\text{to}4$	$\varphi \lesssim 30^{\circ }$	$\begin{array}{r}TUWien:5\mathrm{kW}gun,Lambertsurfacesource\\ hasn=1\end{array}$ $\begin{array}{r}(myreading:adeepmeltpoolcollimates\\ thevapour)\end{array}$
Flat-wafer uniformity	$\frac{d_{edge}}{d_{center}}={\cos }^{n+3}\theta$	$\cos \theta =\frac{r_0}{\sqrt{r_0^2+r_w^2}}$	$r_0:sourceheight$ $\begin{array}{r}{r}_w:waferradius;fromthe{\cos }^{n}lobeofTU\\ WienandthegeometryofDarling,which\\ gives{\cos }^{4}forn=1;ThermalEvaporatorpage\end{array}$ $\begin{array}{r}(computed:a3inchwaferat18inches\\ gives0.976forn=4against0.986forn\\ =1)\end{array}$
Sweep	$A_{melt}\uparrow \Rightarrow R_{dep}\uparrow$	$\text{soak preheat of the charge}$	$\begin{array}{r}Darling:X-Ysweepenlargestheevaporant\\ surface,allowspreheatingandfulleruse\\ ofthesource\end{array}$ $\begin{array}{r}(myreading:sweepingalsoavoidsdrilling\\ aholeindielectricsthatconductheat\\ poorly)\end{array}$

Radiation Effects

Name	Formula	Related Formulae	Variables/Constants
Secondary electrons	$E_{SE}<50\mathrm{eV}$	$T_{g,PMMA}\approx 105{}^{\circ }C$	$\begin{array}{r}Volmeretal.:secondaryelectrons\\ cross-linkPMMAandAZ5214E,cause\\ bubblesandlift-offfailure;heating\\ alonenotresponsible;Electron-Matter\\ Interactionpage\end{array}$ $\begin{array}{r}(myreading:theresistseesanunintended\\ electronexposureduringmetallization)\end{array}$
Bremsstrahlung damage in MOS	${\sigma }_n={\sigma }_0\exp \left(-b{E}_{ox}\right)$	${\sigma }_0=1.6\times 10^{-15}\mathrm{c}{\mathrm{m}}^2,b=7.35\times 10^{-7}cm/V$	$\begin{array}{r}Ning:neutraltrapsfromx-raysattheAl\\ target,positivetrapsabove10^{-13}\mathrm{c}{\mathrm{m}}^2\\ varyingas{E}_{ox}^{-3},bothreducedbyforminggas\\ at550{}^{\circ }Cfor20\min ;Darling:x-raysfrom\\ thebeam\end{array}$ $(computed:{\sigma }_n=4.1\times 10^{-17}\mathrm{c}{\mathrm{m}}^{2}at5MV/cm)$
Mitigation	$V_{electrode}=1500\mathrm{V},V_{sample}=200\mathrm{V}$	$\text{magnets, grounded plates, higher rate and voltage, tight focus}$	$\begin{array}{r}Volmeretal.:deflectionelectrode\\ removesthefence-likeshadow;charged\\ metalatomsfavouredbyhighcurrent,\\ smallhearthandnoliner\end{array}$ $\begin{array}{r}(myreading:positivebiasesrepelions\\ andcollectelectronsbeforetheyreach\\ theresist)\end{array}$

Important Questions

Ques: How does an electron-beam evaporator work?

By melting a small spot of evaporant with a steered electron beam. Darling describes a cathode filament at -10 $\mathrm{kV}$ with the hearth at ground; the beam is bent through 270 degrees by a magnetic field, which keeps the filament out of the evaporant flux and also sweeps the spot in X and Y. The evaporant sits in crucibles or cones in a water-cooled copper hearth, often a 4-pocket rotary hearth so multilayers are deposited in one pump-down. The source exceeds 3000 ${}^{\circ }\mathrm{C}$ and deposits at 10 to 100 angstrom/s, against about 1800 ${}^{\circ }\mathrm{C}$ and 1 to 20 angstrom/s for resistive boats (Darling; Thermal Evaporator page). Sweeping enlarges the melt, raising the rate and allowing a soak preheat (Darling). Reference: R. B. Darling, ‘Physical Vapor Deposition’, EE-527 MicroFabrication lecture slides, University of Washington: 270 degree bent electron-beam gun with the filament out of the evaporant flux, magnetic focusing and X-Y sweep, cathode filament at -10000 V and hearth at 0 V, typical emission voltage 8 to 10 kV, evaporation systems typically 1 to 10 kW, water-cooled copper hearth with crucibles or cones and pyrolytic graphite liner, 4-pocket rotary hearth for sequential layers in one pump-down, temperatures above 3000 C, rates 10 to 100 angstrom/s, sweeping enlarging the melt and allowing soak preheating, materials including all resistance-heated ones plus Ni, Pt, Ir, Rh, Ti, V, Zr, W, Ta, Mo and the oxides Al2O3, SiO, SiO2, SnO2, TiO2, ZrO2, substrates exposed to secondary electrons and x-rays from the high-voltage beam; resistive sources reach about 1800 C at 1 to 20 angstrom/s (fetch-tool summary); Thermal Evaporator page of this vault.

Ques: Why can electron-beam evaporation handle refractory metals and oxides?

Because only the melt is hot. Darling lists everything resistive sources can evaporate plus Ni, Pt, Ir, Rh, Ti, V, Zr, W, Ta and Mo, and oxides including Al${}_2$O${}_3$, SiO, SiO${}_2$, SnO${}_2$, TiO${}_2$ and ZrO${}_2$. TU Wien explains that a water-cooled Cu crucible allows reactive metals such as Ta, Ti and Zr and refractory metals such as W, Mo, Pt and Rh. A resistive boat must be hotter than its charge and can alloy with it, whereas the cooled copper stays cold, so the charge forms its own crucible (my reading). Temperatures above 3000 ${}^{\circ }\mathrm{C}$ against about 1800 ${}^{\circ }\mathrm{C}$ for resistive heating set the boundary between the two tools (Darling). Reference: R. B. Darling, ‘Physical Vapor Deposition’, EE-527 MicroFabrication lecture slides, University of Washington: 270 degree bent electron-beam gun with the filament out of the evaporant flux, magnetic focusing and X-Y sweep, cathode filament at -10000 V and hearth at 0 V, typical emission voltage 8 to 10 kV, evaporation systems typically 1 to 10 kW, water-cooled copper hearth with crucibles or cones and pyrolytic graphite liner, 4-pocket rotary hearth for sequential layers in one pump-down, temperatures above 3000 C, rates 10 to 100 angstrom/s, sweeping enlarging the melt and allowing soak preheating, materials including all resistance-heated ones plus Ni, Pt, Ir, Rh, Ti, V, Zr, W, Ta, Mo and the oxides Al2O3, SiO, SiO2, SnO2, TiO2, ZrO2, substrates exposed to secondary electrons and x-rays from the high-voltage beam; resistive sources reach about 1800 C at 1 to 20 angstrom/s (fetch-tool summary); TU Wien Institute of Solid State Physics, ‘Physik und Technologie Duenner Schichten / Thin Film Technology’, chapter 2 (vacuum evaporation): Lambert cosine emission and plane-substrate cos^4 law, electron-gun emission lobes cos^n with n about 3.5 to 4 for a 5 kW gun, valid to about 30 degrees, water-cooled Cu crucible allowing reactive metals (Ta, Ti, Zr) and refractory metals (W, Mo, Pt, Rh) (fetch-tool summary).

Ques: What voltage and rate are typical, and which figures are more likely?

About 8 to 10 $\mathrm{kV}$ and 10 to 100 angstrom/s for production; slower rates are used for lift-off. Darling gives a typical emission voltage of 8 to 10 $\mathrm{kV}$, while Volmer et al. state that 4 to 20 $\mathrm{kV}$ is used and describe systems running at 4 and 8 $\mathrm{kV}$. The two agree: 8 to 10 $\mathrm{kV}$ is the common case inside a wider range. Rates differ more. Darling quotes 10 to 100 angstrom/s, while Volmer et al. deposited at 0.02 and 0.4 $nm/s$, or 0.2 to 4 angstrom/s (computed), for thin contacts and micromagnets. Darling’s figure is more likely typical for thick films. Volmer et al. themselves recommend higher rates to limit radiation damage, which supports it (my reading). Reference: R. B. Darling, ‘Physical Vapor Deposition’, EE-527 MicroFabrication lecture slides, University of Washington: 270 degree bent electron-beam gun with the filament out of the evaporant flux, magnetic focusing and X-Y sweep, cathode filament at -10000 V and hearth at 0 V, typical emission voltage 8 to 10 kV, evaporation systems typically 1 to 10 kW, water-cooled copper hearth with crucibles or cones and pyrolytic graphite liner, 4-pocket rotary hearth for sequential layers in one pump-down, temperatures above 3000 C, rates 10 to 100 angstrom/s, sweeping enlarging the melt and allowing soak preheating, materials including all resistance-heated ones plus Ni, Pt, Ir, Rh, Ti, V, Zr, W, Ta, Mo and the oxides Al2O3, SiO, SiO2, SnO2, TiO2, ZrO2, substrates exposed to secondary electrons and x-rays from the high-voltage beam; resistive sources reach about 1800 C at 1 to 20 angstrom/s (fetch-tool summary); F. Volmer, I. Seidler, T. Bisswanger, J.-S. Tu, L. R. Schreiber, C. Stampfer and B. Beschoten, ‘How to solve problems in micro- and nanofabrication caused by the emission of electrons and charged metal atoms during e-beam evaporation’, arXiv:2010.06459 (2020, v2 2021): acceleration voltages typically 4 to 20 kV, systems at 4 kV and 8 kV, secondary electrons usually below 50 eV causing PMMA and AZ 5214 E resist cross-linking, bubbles and blisters and lift-off failure, heating alone not responsible (bake 2 h at 180 C, vacuum heating to 500 C without cross-linking), PMMA glass transition about 105 C, charged metal atoms favoured by high beam current, small hearth and no liner, bremsstrahlung x-rays, fixes by a deflection electrode with positive electrode (1500 V) and sample (200 V) potentials, permanent magnets, grounded plates, higher rate, higher voltage and tighter focus, 0.02 nm/s Co and 0.4 nm/s micromagnet depositions in the lower 1e-9 mbar range, 5 nm Cr and 50 nm Au on 285 nm SiO2 (fetch-tool summary).

Ques: How does thickness uniformity compare with a thermal evaporator?

It is slightly worse at the same throw. TU Wien gives the electron-gun emission as ${\cos }^n\varphi$ with $n\approx 3.5$ to 4 for a 5 $\mathrm{kW}$ gun, valid to about 30 degrees, while a Lambert source has $n=1$. Combining that lobe with the flat-wafer geometry Darling uses gives $d_{edge}/d_{center}={\cos }^{n+3}\theta$ (computed), which reduces to the ${\cos }^4$ law of the Thermal Evaporator page for $n=1$. For Darling’s 3 inch wafer at 18 inches this is 0.976 for $n=4$, against 0.986 for $n=1$ (computed): 2.4 per cent instead of 1.4 per cent non-uniformity. In my reading, the narrower lobe is why electron-beam systems use domes or long throws, and it also favours directional deposition for lift-off. Reference: TU Wien Institute of Solid State Physics, ‘Physik und Technologie Duenner Schichten / Thin Film Technology’, chapter 2 (vacuum evaporation): Lambert cosine emission and plane-substrate cos^4 law, electron-gun emission lobes cos^n with n about 3.5 to 4 for a 5 kW gun, valid to about 30 degrees, water-cooled Cu crucible allowing reactive metals (Ta, Ti, Zr) and refractory metals (W, Mo, Pt, Rh) (fetch-tool summary); R. B. Darling, ‘Physical Vapor Deposition’, EE-527 MicroFabrication lecture slides, University of Washington: 270 degree bent electron-beam gun with the filament out of the evaporant flux, magnetic focusing and X-Y sweep, cathode filament at -10000 V and hearth at 0 V, typical emission voltage 8 to 10 kV, evaporation systems typically 1 to 10 kW, water-cooled copper hearth with crucibles or cones and pyrolytic graphite liner, 4-pocket rotary hearth for sequential layers in one pump-down, temperatures above 3000 C, rates 10 to 100 angstrom/s, sweeping enlarging the melt and allowing soak preheating, materials including all resistance-heated ones plus Ni, Pt, Ir, Rh, Ti, V, Zr, W, Ta, Mo and the oxides Al2O3, SiO, SiO2, SnO2, TiO2, ZrO2, substrates exposed to secondary electrons and x-rays from the high-voltage beam; resistive sources reach about 1800 C at 1 to 20 angstrom/s (fetch-tool summary); Thermal Evaporator page of this vault.

Ques: How do stray electrons and ions damage resist during lift-off metallization?

By exposing the resist during deposition. Volmer et al. report that secondary electrons, usually below 50 $\mathrm{eV}$, cross-link PMMA and AZ 5214 E and create bubbles and blisters, leading to lift-off failure and residues. Heating alone was ruled out: a 2 h bake at 180 ${}^{\circ }\mathrm{C}$ and vacuum heating to 500 ${}^{\circ }\mathrm{C}$ caused no cross-linking, though PMMA structures below 1 $\mu \mathrm{m}$ were destroyed above its 105 ${}^{\circ }\mathrm{C}$ glass transition. Charged metal atoms are favoured by high beam current, a small hearth and no liner. Their fixes are a deflection electrode with positive electrode (1500 $\mathrm{V}$) and sample (200 $\mathrm{V}$) potentials, permanent magnets, grounded plates, and higher rate, higher voltage and tighter focus. This matters for patterns from the Electron-Beam Lithography Writer page (my reading). Reference: F. Volmer, I. Seidler, T. Bisswanger, J.-S. Tu, L. R. Schreiber, C. Stampfer and B. Beschoten, ‘How to solve problems in micro- and nanofabrication caused by the emission of electrons and charged metal atoms during e-beam evaporation’, arXiv:2010.06459 (2020, v2 2021): acceleration voltages typically 4 to 20 kV, systems at 4 kV and 8 kV, secondary electrons usually below 50 eV causing PMMA and AZ 5214 E resist cross-linking, bubbles and blisters and lift-off failure, heating alone not responsible (bake 2 h at 180 C, vacuum heating to 500 C without cross-linking), PMMA glass transition about 105 C, charged metal atoms favoured by high beam current, small hearth and no liner, bremsstrahlung x-rays, fixes by a deflection electrode with positive electrode (1500 V) and sample (200 V) potentials, permanent magnets, grounded plates, higher rate, higher voltage and tighter focus, 0.02 nm/s Co and 0.4 nm/s micromagnet depositions in the lower 1e-9 mbar range, 5 nm Cr and 50 nm Au on 285 nm SiO2 (fetch-tool summary); Electron-Beam Lithography Writer page of this vault.

Ques: How does electron-beam evaporation damage MOS gate oxides?

Through x-rays generated at the target. Darling notes that the high-voltage beam generates x-rays and Volmer et al. identify them as bremsstrahlung. Ning found increased electron trapping in the SiO${}_2$ of n-channel polysilicon-gate transistors whose aluminium was electron-beam evaporated, and attributed it to x-rays from the beam striking the Al target. He identified positively charged traps with capture cross-sections above $10^{-13}$ $\mathrm{c}{\mathrm{m}}^2$, varying roughly as $E_{ox}^{-3}$, and neutral traps of about $10^{-15}$ $\mathrm{c}{\mathrm{m}}^2$ following ${\sigma }_0\exp (-b{E}_{ox})$ with ${\sigma }_0=1.6\times 10^{-15}$ $\mathrm{c}{\mathrm{m}}^2$ and $b=7.35\times 10^{-7}$ $cm/V$. At 5 $MV/cm$ this is $4.1\times 10^{-17}$ $\mathrm{c}{\mathrm{m}}^2$ (computed). A dry forming-gas anneal at 550 ${}^{\circ }\mathrm{C}$ for 20 $\min$ reduced both trap types. The x-ray mechanism is described on the Electron-Matter Interaction page. Reference: T. H. Ning, ‘Electron trapping in SiO2 due to electron-beam deposition of aluminum’, Journal of Applied Physics, doi:10.1063/1.325368 (IBM Research listing): increased electron trapping in the SiO2 of n-channel polysilicon-gate transistors with e-beam evaporated Al, attributed to x-rays generated where the beam strikes the Al target, positively charged traps with low-field capture cross-section above 1e-13 cm2 and field dependence about Eox^-3, neutral traps about 1e-15 cm2 with sigma = sigma0 exp(-b Eox), sigma0 = 1.6e-15 cm2 and b = 7.35e-7 cm/V, both reduced by a dry forming-gas anneal at 550 C for 20 min (fetch-tool summary); R. B. Darling, ‘Physical Vapor Deposition’, EE-527 MicroFabrication lecture slides, University of Washington: 270 degree bent electron-beam gun with the filament out of the evaporant flux, magnetic focusing and X-Y sweep, cathode filament at -10000 V and hearth at 0 V, typical emission voltage 8 to 10 kV, evaporation systems typically 1 to 10 kW, water-cooled copper hearth with crucibles or cones and pyrolytic graphite liner, 4-pocket rotary hearth for sequential layers in one pump-down, temperatures above 3000 C, rates 10 to 100 angstrom/s, sweeping enlarging the melt and allowing soak preheating, materials including all resistance-heated ones plus Ni, Pt, Ir, Rh, Ti, V, Zr, W, Ta, Mo and the oxides Al2O3, SiO, SiO2, SnO2, TiO2, ZrO2, substrates exposed to secondary electrons and x-rays from the high-voltage beam; resistive sources reach about 1800 C at 1 to 20 angstrom/s (fetch-tool summary); F. Volmer, I. Seidler, T. Bisswanger, J.-S. Tu, L. R. Schreiber, C. Stampfer and B. Beschoten, ‘How to solve problems in micro- and nanofabrication caused by the emission of electrons and charged metal atoms during e-beam evaporation’, arXiv:2010.06459 (2020, v2 2021): acceleration voltages typically 4 to 20 kV, systems at 4 kV and 8 kV, secondary electrons usually below 50 eV causing PMMA and AZ 5214 E resist cross-linking, bubbles and blisters and lift-off failure, heating alone not responsible (bake 2 h at 180 C, vacuum heating to 500 C without cross-linking), PMMA glass transition about 105 C, charged metal atoms favoured by high beam current, small hearth and no liner, bremsstrahlung x-rays, fixes by a deflection electrode with positive electrode (1500 V) and sample (200 V) potentials, permanent magnets, grounded plates, higher rate, higher voltage and tighter focus, 0.02 nm/s Co and 0.4 nm/s micromagnet depositions in the lower 1e-9 mbar range, 5 nm Cr and 50 nm Au on 285 nm SiO2 (fetch-tool summary); Electron-Matter Interaction page of this vault.

References

1. R. B. Darling, “Physical vapor deposition,” EE-527 MicroFabrication lecture notes, University of Washington. Available: https://WeCanFigureThisOut.org/NANO/lecture_notes/Microfabrication_Supporting_materials_files/R_B_Darling - Washington U - PhysicalVaporDeposition.pdf
2. TU Wien Institute of Solid State Physics, “Physik und Technologie duenner Schichten (Thin film technology),” chapter 2, vacuum evaporation. Available: https://static.ifp.tuwien.ac.at/homepages/Personen/duenne_schichten/pdf/t_p_tf_chapter2.pdf
3. F. Volmer, I. Seidler, T. Bisswanger, J.-S. Tu, L. R. Schreiber, C. Stampfer and B. Beschoten, “How to solve problems in micro- and nanofabrication caused by the emission of electrons and charged metal atoms during e-beam evaporation,” arXiv:2010.06459, 2020. Available: https://arxiv.org/abs/2010.06459
4. T. H. Ning, “Electron trapping in SiO2 due to electron-beam deposition of aluminum,” Journal of Applied Physics, doi:10.1063/1.325368. Available: https://research.ibm.com/publications/electron-trapping-in-siolessinfgreater2lessinfgreater-due-to-electron-beam-deposition-of-aluminum
5. Thermal Evaporator page of this vault.
6. Electron-Matter Interaction page of this vault.
7. Electron-Beam Lithography Writer page of this vault.

Related pages in this vault: Thermal Evaporator, Electron-Matter Interaction, Electron-Beam Lithography Writer, Quartz Crystal Microbalance (QCM), DC and RF Magnetron Sputter System.