Heavy-ion-induced desorption yields of cryogenic surfaces bombarded with 1.4 MeV/u xenon ions

Abstract

Heavy-ion-induced desorption of two different cryogenic targets was studied with a new experimental setup installed at the GSI High Charge State Injector. One gold-coated and one amorphous-carbon-coated copper target, bombarded under perpendicular impact with 1.4 MeV/u Xe18+ ions, were tested. Partial pressure rises of H2, CO, CO2, and CH4 and effective desorption yields were measured at 300, 77, and 8 K using continuous heavy-ion bombardment. We found that the desorption yields decrease with decreasing target temperature and measured the yield rises as a function of CO gas cryosorbed at 8 K. In this paper we describe the experimental system comprising a new cryogenic target assembly, the preparation of the targets, the test procedure, and the evaluation of the effective pumping speed of the setup. Pressure rise and gas adsorption experiments are described; the obtained results are discussed and compared with literature data.

Full text

Heavy-ion-induced desorption yields of cryogenic surfaces bombarded
with 1:4 MeV=uxenon ions
Donat Philipp Holzer*and Edgar Mahner
CERN, 1211 Geneva 23, Switzerland
Holger Kollmus, Markus Bender, Daniel Severin, and Marc Wengenroth
GSI, 64291 Darmstadt, Germany
(Received 28 February 2012; revised manuscript received 17 July 2013; published 23 August 2013)
Heavy-ion-induced desorption of two different cryogenic targets was studied with a new experimental
setup installed at the GSI High Charge State Injector. One gold-coated and one amorphous-carbon-coated
copper target, bombarded under perpendicular impact with 1:4 MeV=uXe
18þions, were tested. Partial
pressure rises of H2, CO, CO2, and CH4and effective desorption yields were measured at 300, 77, and 8 K
using continuous heavy-ion bombardment. We found that the desorption yields decrease with decreasing
target temperature and measured the yield rises as a function of CO gas cryosorbed at 8 K. In this paper we
describe the experimental system comprising a new cryogenic target assembly, the preparation of the
targets, the test procedure, and the evaluation of the effective pumping speed of the setup. Pressure rise
and gas adsorption experiments are described; the obtained results are discussed and compared with
literature data.
DOI: 10.1103/PhysRevSTAB.16.083201 PACS numbers: 29.27.a, 41.75.Ak, 79.20.Rf, 07.30.Kf
I. INTRODUCTION
Heavy-ion-induced molecular desorption can be a seri-
ous limitation for synchrotrons operating with low charge-
state heavy ions due to the observed large pressure rises
that often reduce the obtainable beam intensity and lifetime
of the accelerator [1]. Since 2000, several laboratories have
studied this phenomenon [1–3] to find appropriate mitiga-
tion methods. So far, most of the experiments investigated
a variety of ambient temperature targets, which were bom-
barded with heavy ions to measure their desorption yields.
First experiments with cryogenic targets were performed at
CERN; accelerator relevant materials as bare and gold-
coated copper targets were used there [4,5].
Desorption yield measurements at cryogenic tempera-
tures are motivated by the design of future heavy-ion
accelerators, for example, SIS100, which is part of the
GSI FAIR (Facility for Antiproton and Ion Research)
project [6], and the operation of the Large Hadron
Collider (LHC). Two different targets were chosen for
our studies at HLI; the first material is gold-coated copper,
which has already been investigated at ambient and cryo-
genic temperatures [5,7]. Such gold coatings are of special
interest because they are part of the baseline design for the
SIS100 cryocatcher [8]. As a second target an amorphous
carbon thin film, sputter coated onto a copper substrate,
was chosen. The dynamic vacuum properties of such films
are important to investigate since amorphous carbon coat-
ings are foreseen to mitigate the electron cloud effect [9]in
the CERN Super Proton Synchrotron [10]. Ambient tem-
perature desorption yield measurements were reported
recently [11].
II. EXPERIMENT
A. Experimental setup
The accelerator providing the ion beam for this experi-
ment is the GSI High Charge State Injector (HLI). The HLI
is capable of delivering ions at a fixed energy of
1:4 MeV=u, with pulse lengths between 0.49 and 5.5 ms
at a repetition rate between 0.1 and 50 Hz. The accelerator
is described in detail elsewhere [12]. The experimental
layout is shown in Fig. 1.
A gate valve separates the HLI from the experimental
setup such that samples can be changed without disturbing
accelerator operation. Downstream the gate valve, adjust-
able beam collimators are installed to control the beam
size. The collimator jaws are electrically isolated and
can be used to measure the beam current. To separate the
desorption experiment from the high pressure (107mbar)
of the HLI, a differential pumping system comprising two
separate vacuum chambers are installed between the colli-
mators and the sample chamber (see Fig. 1).
The first vacuum chamber contains a turbo molecular
pumping (TMP) group and a baffle, which is cooled
to liquid nitrogen temperature. The cold baffle retains
hydrocarbons and water vapor originated from the linac.
*Also at Vienna University of Technology, Vienna, Austria.
donat.philipp.holzer@cern.ch
Published by the American Physical Society under the terms of
the Creative Commons Attribution 3.0 License. Further distri-
bution of this work must maintain attribution to the author(s) and
the published article’s title, journal citation, and DOI.
PHYSICAL REVIEW SPECIAL TOPICS - ACCELERATORS AND BEAMS 16, 083201 (2013)
1098-4402=13=16(8)=083201(9) 083201-1 Published by the American Physical Society

A retractable Faraday cup is installed and used to measure
the beam current.
The second vacuum chamber is equipped with a sputter
ion pump (SIP) and an integrated titanium sublimation
pump (TSP) providing a total pumping speed of about
1000 l=sfor nitrogen. It also houses a retractable
Chromolux (Cr doped Al2O3) beam position monitor,
which is used to align the beam.
At the end of the experimental line, the installed sample
chamber is equipped with another TMP group, which is
used during pump-down and bakeout of the entire setup.
The group is separated from the setup (valve closed) during
heavy-ion desorption experiments. Pumping is provided
through a conductance (¼63 mm,l¼228 mm) result-
ing in an effective pumping speed of about 82 l=s
(N2equivalent). A gas injection system is also connected
to the sample chamber. It comprises a variable leak valve
and a membrane pump.
For desorption experiments at cryogenic temperatures
a dedicated cold-head assembly was designed, built,
and installed at the end of the experimental setup (see
Fig. 2).
The target used for heavy-ion-induced desorption mea-
surements is mounted on a dedicated sample holder, which
is attached to the second stage of the cold head (Leybold
Coolpower 7=25). It is equipped with two temperature
sensors (LakeShore diode DT 470-CU-12A). One diode
is mounted on the backside of the test sample, the second
one is connected onto the copper radiation shield of the
cold-head assembly. In order to control the sample tem-
perature, the target holder is equipped with a resistive DC
heater, which provides a maximum heating power of
62.5 W. Using this setup, the sample temperature during
the experiments, with impinging beam, was found to be
stable within 0.3 K or better.
For partial and total pressure measurements, the sample
vacuum chamber is equipped with a Balzers QMA 125
residual gas analyzer (RGA) and a calibrated Leybold IE
514 extractor gauge (extr.). The RGA was calibrated in situ
with respect to the total pressure gauge.
B. Fabrication and preparation of targets
Fabrication, cleaning, and coating of the tested samples
were performed at CERN. Two discs were machined from
oxygen-free electronic grade copper with a thickness of
7 mm and a diameter of 50 mm. After standard cleaning,
FIG. 2. Top: schematic view of the cold-head assembly used
for heavy-ion-induced desorption measurements at cryogenic
temperatures. Bottom left: frontal view of the cold head with
installed gold-coated target and inner thermal radiation shield,
thermally anchored at the first stage of the cold head. Bottom
right: installed target and outer thermal radiation shield anchored
at ambient temperature.
FIG. 1. Schematic view of the experimental setup, which has been attached at the GSI HLI for heavy-ion-induced desorption
measurements at 1:4 MeV=u.
DONAT PHILIPP HOLZER et al. Phys. Rev. ST Accel. Beams 16, 083201 (2013)
083201-2

one of the targets was coated with amorphous carbon by
magnetron sputtering. The used deposition system, the
sputtering parameters, and resulting sample topography
are described in detail elsewhere [11,13]. The film thick-
ness was measured by scanning electron microscopy to be
360 nm. The second copper target was galvanically coated
with a 7mthin gold film and a 2mthin nickel under-
layer acting as a diffusion barrier during bakeout of the
experimental setup. This target was gold coated simulta-
neously with another copper sample, which was previously
tested for low-temperature heavy-ion-induced desorption
at LINAC 3 [5].
C. Experimental procedure
After sample installation and pump-down of the experi-
mental line, the setup was baked at 150 Cfor 48 hours.
Heavy-ion-induced desorption yields were measured at a
fixed projectile impact angle of ¼0(perpendicular
impact). The HLI beam consisted of Xe18þions with an
energy of 1:4 MeV=u, a pulse length of 3.18 ms, and a
repetition rate of 4.8 Hz. Both targets were bombarded with
an average ion flux of 2109ions per second. Prior to
each individual desorption experiment, the Faraday cup
was used to measure the beam current from which the
ion flux is calculated. The collimator current measure-
ments allowed continuous monitoring of the beam current.
The collimator jaws were adjusted to result in a beam spot
size of 11cm
2on the target. As described above, the
amorphous-carbon-coated copper target (aC=Cu sample)
and the gold-coated copper target (Au=Cu sample) were
tested for desorption at 300, 77, and 8 K.
The pressure rise method [14] was used to determine the
yields of the studied samples. The effective heavy-ion-
induced desorption yield eff is calculated as
eff ¼PS
_
NXekbT;(1)
where Pis the total pressure increase, Sis the effective
pumping speed, _
NXe is the average number of impacting
Xe ions per second, kbis the Boltzmann constant, and Tis
the temperature of the released gas at the position of the
pressure gauge (300 K). PSis defined as the gas flux
Q0, which in this case is the gas flux of desorbed gas
molecules produced by the impinging ions on a target.
D. Pumping speed of cryogenic targets
Since the cold head acts as a cryopump at cryogenic
temperatures, the effective total pumping speed Sincreases
when the target is cooled down. We define
S¼Seff þScryoðÞ;(2)
where Seff is the ambient temperature effective pumping
speed of the TSP and SIP in the experimental chamber and
Scryo is the effective pumping speed generated by the cold
surface of the target, which depends on the sticking proba-
bility .
At this point we want to underline that in our previous
heavy-ion-induced desorption experiment of cryogenic
targets at CERN-LINAC 3 [5], the so-called single-shot
technique [14] was used to determine Scryo. Unfortunately
this was not possible for the GSI-HLI study presented here.
In order to quantify the influence of cryopumping in our
system we have chosen Monte Carlo simulations as well as
analytical calculations, which are described next.
The gas flux Q0, generated by the ion-induced desorp-
tion at the target, is reduced by a factor that depends on
the target sticking probability .
The radiation shield of the cold-head assembly, as sche-
matically shown in Fig. 3, forms a short cylindrical duct in
front of the target introducing a certain transmission proba-
bility and backscattering probability ¼1for re-
leased gas molecules.
The gas flux arriving in the measurement vacuum cham-
ber, Q1, is therefore equal to Q0times the fraction of gas
that is not backscattered and pumped by the target. We find
Q1ðÞ¼Q0½þð1Þþ2ð1Þ2þ...(3)
¼Q0X
1
n¼0
nð1Þn(4)
¼Q01
1ð1Þ:(5)
Therefore, the reduction factor is defined as
ðÞ¼1ð1Þ
(6)
and finally Q0can be calculated as
Q0ðÞ¼Q1ðÞ(7)
¼PSðÞ:(8)
FIG. 3. Schematic display of the gas flux, pumping speeds, and
geometry relevant for the heavy-ion-induced desorption of a
cryogenic target installed in the experimental vacuum chamber.
HEAVY-ION-INDUCED DESORPTION YIELDS OF ... Phys. Rev. ST Accel. Beams 16, 083201 (2013)
083201-3

We have calculated the product of the pumping speed S
and the reduction factor as a function of the sticking
probability . For the specific geometry of our experimen-
tal setup with S0¼1000 l=sfor CO, we find ¼0:582
and ¼0:418 [15]. In addition, the measurement setup
has been simulated using the Monte Carlo flow analysis
software MolFlowþ[16]. The obtained results are shown
in Fig. 4.
We find that the analytical calculation is in good agree-
ment with the Monte Carlo simulation (see Fig. 4). A slight
difference is only observed for values close to 1. This
result was expected, since the Monte Carlo simulation
takes into account that the relevant desorbing surface is
only the beam spot in the middle of the target, and not the
entire surface, which is the implicit assumption in the
analytical case. Because the calculation of the effective
desorption yields eff , the values from the Monte Carlo
simulation were used in this work.
So far we have derived a relation for Sas a function
of the sticking probability . In the following we will
determine the sticking probability at the required tempera-
tures of 77 and 8 K.
We used the measured pumping speeds Scryo of our
previous LINAC 3 heavy-ion-induced desorption studies
[5] to obtain ð77 KÞand ð8KÞ. This approach is
well justified because the bombarded target surfaces
(Au=Cu,aC=Cu) and the target temperatures (300, 77,
8 K) were identical with the LINAC 3 experiment. Under
such conditions the sticking probabilities , which are
independent of the sample size and the experimental setup,
can be derived from our previous experiment. We used the
relation
¼Smax
Scryo
¼Aka
Scryo
;(9)
where Smax is the maximum possible pumping speed of a
given surface, Ais the surface area, and kais the incident
volume per area per unit time, directly related to the
average speed vaof molecules, ka¼1
4va.
III. RESULTS
A. Desorption yield measurements at 300 K
The gold-coated copper and amorphous-carbon-coated
targets were continuously bombarded for 15 min with
1:4 MeV=uXe
18þions impacting under perpendicular
incidence. Total and partial pressure rises were measured
with the extractor gauge and the residual gas analyzer. The
obtained results are shown in Fig. 5.
For the aC=Cu sample it has to be noted that about 600 s
after the start of the continuous heavy-ion bombardment,
the RGA was switched to scan mode to acquire a full mass
FIG. 5. Heavy-ion-induced pressure rises at 300 K target tem-
perature. Partial pressure rises (P)ofH2, CO, CO2,CH4, and
total pressure rises measured for the Au=Cu and aC=Cu samples
continuously bombarded under ¼0with 1:4 MeV=uXe
18þ
ions are shown.
FIG. 4. Product of pumping speed Sand reduction factor as
a function of sticking probability for the used experimental
setup calculated; the shown curves were calculated for CO.
DONAT PHILIPP HOLZER et al. Phys. Rev. ST Accel. Beams 16, 083201 (2013)
083201-4

spectrum. Therefore, no time-dependent data points are
available for about 200 s (see Fig. 5).
For both targets the heavy-ion-induced desorption at
300 K is dominated by CO and CO2. Compared to H2,
the CO and CO2partial pressure increase is a factor of
3 higher for the Au=Cu sample and more than 1 order of
magnitude higher for the aC=Cu sample.
Using the pumping speed relation shown in Fig. 4and
assuming a sticking probability of below 0.01, the ef-
fective desorption yields were calculated from the total
pressure rises. We find eff 5400 molecules=ion for
Au=Cu and eff 127000 molecules=ion for aC=Cu.
The relative error of the calculated desorption yields
is dominated from uncertainties in the pressure measure-
ments and is estimated to 30%. An additional variation
of the desorption yield, even for a similar target material,
cannot be excluded due to the potential influence of surface
treatment, cleaning, storage, and bakeout. This variation
has not been taken into account in this analysis, and
therefore the given values are called effective desorption
yields.
B. Desorption yield measurements at 77 K
After the desorption measurements at ambient target
temperature, each sample was cooled down from 300 to
77 K where the temperature was stabilized using the con-
trol system comprising the DC heater and the temperature
sensors attached to the samples.
The measured pressure rises at 77 K are compared in
Fig. 6.ForAu=Cu, the partial pressures of H2, CO, and
CO2contribute roughly equal to the pressure rise. H2is
now the dominating measured gas for aC=Cu. The hydro-
gen partial pressure is more than a factor of 10 higher than
for CO and CO2.
At 77 K we obtain for a sticking probability of
0:037 the following heavy-ion-induced desorption
yields: eff 620 molecules=ion (Au=Cu) and eff 
3100 molecules=ion (aC=Cu).
C. Desorption yield measurements at 8 K
Following the desorption experiments at 77 K, both
samples were further cooled down to 8 K, which is the
limit temperature that could be reached in this measure-
ment setup.
The straight line visible in Fig. 7for the aC=Cu sample is
an artefact, most probably introduced by the electronics of
the total pressure gauge. To some extent, the same artefact
is also visible in Fig. 6for the Au=Cu target. The erroneous
data points were not taken into account for the desorption
yield evaluation.
The measured pressure rises at 8 K are compared in
Fig. 7. With a sticking probability of 0:1the resulting
effective desorption yields are eff 430 molecules=ion
for Au=Cu and eff 1000 molecules=ion for aC=Cu.
D. Desorption yields for different gas species
A comparison of the effective desorption yields eff ,
calculated from the total and partial pressures, is shown
in Table Iand II.
The data from total and partial pressure measurements
are in reasonable agreement for both targets, with the
exception of the aC=Cu sample for 77 K. Presently we
have no sound explanation for the yield difference ob-
served for this measurement.
E. Temperature dependence of the desorption yields
The results obtained from the above described heavy-
ion-induced desorption measurements of cryogenic sur-
faces are summarized in Fig. 8, which also includes a
comparison with earlier desorption yield measurements
performed at CERN-LINAC 3 [5]. In this study we confirm
that the heavy-ion-induced desorption yield is temperature
FIG. 6. Heavy-ion-induced pressure rises at 77 K target tem-
perature. Partial pressure rises (P)ofH2, CO, CO2,CH4, and
total pressure rises measured for the Au=Cu and aC=Cu samples
continuously bombarded under ¼0with 1:4 MeV=uXe
18þ
ions are shown.
HEAVY-ION-INDUCED DESORPTION YIELDS OF ... Phys. Rev. ST Accel. Beams 16, 083201 (2013)
083201-5

dependent; see Fig. 8(left). For the amorphous-carbon-
coated copper target, bombarded at GSI HLI under
perpendicular impact with 1:4MeV=uXe
18þions, we
measured a desorption yield of eff 1000 molecules=ion
at 8 K, which is a factor of 100 lower compared with the
300 K value. Since for the gold-coated copper sample the
yield reduction at 8 K is a factor of 12,wefindeff 
430 molecules=ion.
1. Comparison and discussion of results
In Fig. 8(right) the HLI results of this study are com-
pared with earlier LINAC 3 results. Both experiments had
in common that all investigated targets were bombarded
under perpendicular impact (¼0). The chosen target
temperatures were nearly identical: 300 K, 77 K, 8 K
(HLI), and 6.3 K (LINAC 3). In addition, both Au=Cu
targets were fabricated from the same Cu raw material,
cleaned, and finally gold coated at CERN using an
identical procedure. Significant differences between the
HLI and LINAC 3 experiments are the projectile energies
(1:4 MeV=u,4:2 MeV=u) and the ion species (Xe18þ,
Pb54þ). The experimental setups are also not identical
but the measurement principle (pressure rise method)
was the same.
We observe that the gold-coated copper targets have
a similar desorption yield value at 300 K, i.e.
6200 molecules=Pb54þion and 5400 molecules=Xe18þ
ion. Therefore, the yield measured with impacting Xe18þ
ions onto identically prepared targets is 12% lower than
the value obtained for Pb54þ. In previous studies [2,17]we
found that the ambient temperature desorption yield is
related to the electronic energy loss ðdE=dxÞel of the
projectile as follows:
¼kdE
dxn
el
;
where kis a scaling factor and ncan vary between 1.5 and
3. Calculating the electronic energy loss values for Pb
(4:2 MeV=u) and Xe (1:4 MeV=u) ions using SRIM [18]
and assuming n¼2, we obtain for k¼1a value of
½ðdE=dxÞXe=ðdE=dxÞPb 20:30. This number compares
with the experimentally found desorption yield ratio of
0:88. Since the kvalue is not known, we cannot explain
the yield differences measured for the gold-coated copper
FIG. 7. Heavy-ion-induced pressure rises at 8 K target tem-
perature. Partial pressure rises (P)ofH2, CO, CO2,CH4, and
total pressure rises measured for the Au=Cu and aC=Cu samples
continuously bombarded under ¼0with 1:4 MeV=uXe
18þ
ions are shown.
TABLE II. Heavy-ion-induced desorption yields calculated
from the total and partial pressure rises for the aC=Cu sample,
continuously bombarded under ¼0with 1:4 MeV=uXe
18þ
ions.
Xe18þð1:4 MeV=uÞ!aC=Cu
Gas eff ð300 KÞeff ð77 KÞeff ð8KÞ
Ptot 129 000 3100 1000
H212 000 8400 1300
CO 53 000 210 400
CO243 000 180 170
CH43000 44 39
TABLE I. Heavy-ion-induced desorption yields calculated
from the total and partial pressure rises for the Au=Cu sample,
continuously bombarded under ¼0with 1:4 MeV=uXe
18þ
ions.
Xe18þð1:4 MeV=uÞ!Au=Cu
Gas eff ð300 KÞeff ð77 KÞeff ð8KÞ
Ptot 5500 620 430
H21100 560 230
CO 1500 150 98
CO2840 150 120
CH4140 16 15
DONAT PHILIPP HOLZER et al. Phys. Rev. ST Accel. Beams 16, 083201 (2013)
083201-6

targets. On the other hand, this discrepancy between the
HLI and LINAC 3 desorption results is not surprising,
since it is very well known that the amount of surface
impurities, especially adsorbed carbon and oxygen, plays
a significant role on the absolute values of eff [11]. This
argument may be supported by the different experimental
conditions, e.g., the sample storage time in air and the
bakeout temperature difference between the two experi-
ments, which both can influence the amount of adsorbed
carbon and oxygen on gold surfaces [7].
At 77 K the desorption yield difference for the two
Au=Cu targets, bombarded with 1:4 MeV=uXe
18þions
and 4:2 MeV=uPb
54þions, increases to 61% (see
Fig. 8). The observed yield difference between the two
experiments may also be due to the amount of rest gas
molecules adsorbed onto the target during the cool down
from 300 to 77 K. Compared to the LINAC 3 experiment,
the base pressure of the HLI setup, measured at ambient
target temperature, was a factor of 12 higher than at
LINAC 3. We also have to mention the additional uncer-
tainty due to the above described determination of the
sticking coefficient.
For the HLI experiment at 8 K and the LINAC 3 experi-
ment at 6.3 K, the heavy-ion-induced desorption rates of
both gold-coated copper surfaces differ by 27% [see
Fig. 8(right)]. It is interesting to note that for 3 out of 4
samples tested in two different experimental setups at GSI
HLI and CERN LINAC 3, the heavy-ion-induced
desorption yields of cryogenic targets, cooled down to
the temperature range of about 6–8 K, are quite similar
although the used projectile energies and charge states
were significantly different. This holds for bare and
gold-coated copper, which have a yield of effð6–8KÞ
400 60 molecules=ion. Only the copper target coated
with amorphous carbon has a factor of 2:5higher de-
sorption yield at 8 K.
F. CO gas adsorption experiments at 8 K
The above described gas injection system, connected
close to the samples placed in the experimental vacuum
chamber, was used to inject CO gas onto the cryogenic
target surfaces. The same experimental procedure as pre-
viously described in Ref. [5] was used in this study. CO gas
was injected into the vacuum system, and a known amount
of monolayers were adsorbed on the samples. In a second
step, the gas was pumped out until the limit pressure of the
system was reached again. Finally, each sample was con-
tinuously bombarded with 1:4 MeV=uXe
18þions under
perpendicular impact. The bombardment period was at
least 5 min or until a stable pressure rise was obtained.
Pressure data were taken continuously with the RGA and
the extractor gauge. After the end of the ion bombardment,
the vacuum system was pumped again until the limit
pressure was obtained. Afterwards the same procedure
was repeated with a higher number of adsorbed CO mono-
layers on the cryogenic surface. The results of these
measurements, performed for the amorphous-carbon and
gold-coated copper targets at 8 K are summarized in Fig. 9,
which also includes a comparison with earlier gas adsorp-
tion measurements of the same type performed at CERN
LINAC 3 [5].
The lowest desorption yields of eff 
430 molecules=Xe18þion (Au=Cu) and eff 
520 molecules=Xe18þion (aC=Cu) were measured at 8 K
FIG. 8. Left: temperature dependence of the heavy-ion-induced desorption yield of gold-coated and amorphous-carbon-coated
copper surfaces bombarded under ¼0with 1:4 MeV=uXe
18þions. Right: comparison of LINAC 3 desorption data obtained for
4:2 MeV=uPb
54þions bombarding bare and gold-coated copper targets under ¼0[5].
HEAVY-ION-INDUCED DESORPTION YIELDS OF ... Phys. Rev. ST Accel. Beams 16, 083201 (2013)
083201-7

without additional CO gas adsorption onto the cryogenic
surfaces; see Fig. 9(left). For the amorphous-carbon-
coated target the 8 K desorption yield increases faster after
the cryosorption of CO monolayers than the gold-coated
target. After the adsorption of 1 ML of CO gas the yield
increases by a factor of 4:2for carbon but remains nearly
unchanged for the gold surface. The 77 K desorption
yield value of both samples is reached with 1–2ML
of CO gas adsorbed at 8 K. The ambient temperature
desorption rate value of Au=Cu is exceeded after the
cryosorption of more than 25 ML of CO gas. With 
45 ML, the heavy-ion-induced desorption yield is 
43 000 molecules=ion for aC=Cu, still a factor of 3
below the measured 300 K value. For both 8 K targets
we find that with 45 ML cryosorbed CO gas the
heavy-ion-induced desorption yield is not yet constant;
the effect is more pronounced for the gold-coated copper
sample.
1. Comparison and discussion of results
In Fig. 9(right) the above described HLI measurements
are compared with earlier LINAC 3 results. It is remark-
able that the bare and gold-coated copper targets, bom-
barded with 4:2 MeV=uPb
54þions at LINAC 3, as well as
the amorphous-carbon-coated copper target, bombarded
with 1:4 MeV=uXe
18þions at HLI, show the same de-
sorption behavior at cryogenic temperature, i.e., the change
in desorption yield as a function of cryosorbed CO mono-
layers is very similar for perpendicular heavy-ion impact.
The LINAC 3 experiment has demonstrated that eff con-
tinues to increase until 100 ML and remains constant up to
300 ML of cryosorbed CO gas. The gold-coated target,
bombarded with Xe18þions, shows a different behavior;
see Fig. 9(right). At present we have no sound explanation
for this difference.
IV. SUMMARY AND CONCLUSION
Heavy-ion-induced desorption of gold-coated and
amorphous-carbon-coated copper targets, bombarded at
GSI HLI with 1:4 MeV=uXe
18þions under ¼0per-
pendicular incidence, was studied at 300, 77, and 8 K.
Desorption yields of eff ð8KÞ1000 molecules=ion
were measured for the carbon-coated and eff ð8KÞ
430 molecules=ion for the gold-coated copper target.
These 8 K yields are a factor of 127 (aC=Cu) and
13 (Au=Cu) lower than the 300 K values. The low-
temperature heavy-ion-induced desorption yields in-
creased for both targets after a dedicated cryosorption of
up to 45 monolayers of CO gas. A comparison with our
previous LINAC 3 studies confirms qualitatively the cor-
relation between eff and the target temperature as well as
the rise of eff with adsorbed CO gas.
At present there is no reliable theory available that
describes the heavy-ion-induced desorption phenomena
and the measured desorption rates, neither for ambient
nor for cryogenic target temperatures. Some progress has
been recently made for desorption measurements with
ambient temperature targets [19], but a deeper understand-
ing of the underlying desorption mechanism is still lacking.
Therefore, we suggest continuing the development of the
inelastic thermal spike model and its extension to cryo-
genic target temperatures. We also want to highlight the
practical impact of such a development, which would be
very beneficial for the design, commissioning, and opera-
tion of future heavy-ion synchrotrons, for example, SIS100
of FAIR.
FIG. 9. Left: Heavy-ion-induced desorption yields as a function of adsorbed CO monolayers on gold-coated and amorphous-carbon-
coated copper surfaces bombarded at 8 K with 1:4 MeV=uXe
18þions under ¼0. Right: comparison with LINAC 3 desorption data
obtained for 4:2 MeV=uPb
54þions bombarding under ¼0bare and gold-coated copper targets at 6.3 K [5].
DONAT PHILIPP HOLZER et al. Phys. Rev. ST Accel. Beams 16, 083201 (2013)
083201-8

ACKNOWLEDGMENTS
We would like to thank the CERN workshop for the
fabrication and Marina Malabaila, Pedro Costa Pinto, and
Paul Victor Edwards for the coatings of the targets. We
acknowledge the work of the GSI accelerator department,
delivering excellent stable beam conditions during our
experiments.
[1] E. Mahner, Phys. Rev. ST Accel. Beams 11, 104801
(2008), and references therein.
[2] H. Kollmus, A. Kra
¨mer, M. Bender, M.C. Bellachioma,
H. Reich-Sprenger, E. Mahner, E. Hedlund, L.
Westerberg, O. B. Malyshev, M. Leandersson, and E.
Edqvist, J. Vac. Sci. Technol. A 27, 245 (2009).
[3] E. Hedlund, L. Westerberg, O. B. Malyshev, E. Edqvist,
M. Leandersson, H. Kollmus, M. C. Bellachioma, M.
Bender, A. Kra
¨mer, H. Reich-Sprenger, B. Zajec, and A.
Krasnov, Nucl. Instrum. Methods Phys. Res., Sect. A 599,
1 (2009).
[4] E. Mahner, M. Bender, and H. Kollmus, in Proceedings of
High Intensity and High Brightness Beams, edited by I.
Hofmann, J.-M. Lagniel, and R. W. Haase (Melville, New
York, 2005), pp. 219–221; CERN Vacuum Technical Note
05-04, 2005.
[5] E. Mahner, L. Evans, D. Ku
¨chler, R. Scrivens, M. Bender,
H. Kollmus, D. Severin, and M. Wengenroth, Phys. Rev.
ST Accel. Beams 14, 050102 (2011).
[6] ‘‘Facility for Antiproton and Ion Research in Europe
GmbH,’’ 2011, http://www.fair-center.de.
[7] E. Mahner, J. Hansen, D. Ku
¨chler, M. Malabaila, and M.
Taborelli, Phys. Rev. ST Accel. Beams 8, 053201 (2005).
[8] L. Bozyk, D. H. H. Hoffmann, P. Spiller, and H. Kollmus,
in Proceedings of the 2nd International Particle
Accelerator Conference, San Sebastia
`n, Spain (EPS-AG,
Spain, 2011), pp. 1527–1529.
[9] F. Caspers, G. Rumolo, W. Scandale, and F. Zimmermann,
CERN Report No. CERN-BE-2009-005, 2008.
[10] C. Yin Vallgren, G. Arduini, J. Bauche, S. Calatroni, P.
Chiggiato, K. Cornelis, P. Costa Pinto, B. Henrist, E.
Me
`tral, H. Neupert, G. Rumolo, E. N. Shaposhnikova,
and M. Taborelli, Phys. Rev. ST Accel. Beams 14,
071001 (2011).
[11] E. Mahner, D. Holzer, D. Ku
¨chler, R. Scrivens, P. Costa
Pinto, C. Yin Vallgren, and M. Bender, Phys. Rev. ST
Accel. Beams 14, 101001 (2011).
[12] K. Tinschert, D. M. Rueck, H. Emig, K. Leible, and N.
Angert, Nucl. Instrum. Methods Phys. Res., Sect. B 113,
59 (1996).
[13] P. Chiggiato and P. Costa Pinto, Thin Solid Films 515, 382
(2006).
[14] E. Mahner, J. Hansen, J.-M. Laurent, and N. Madsen,
Phys. Rev. ST Accel. Beams 6, 013201 (2003).
[15] Foundations of Vacuum Science and Technology, edited by
J. M. Lafferty (Wiley-Interscience, New York, 1998).
[16] R. Kersevan and J.-L. Pons, J. Vac. Sci. Technol. A 27,
1017 (2009).
[17] A. W. Molvik, H. Kollmus, E. Mahner, M. K. Covo, M. C.
Bellachioma, M. Bender, F. M. Bieniosek, E. Hedlund, A.
Kra
¨mer, J. Kwan, O. B. Malyshev, L. Prost, P. A. Seidl, G.
Westenskow, and L. Westerberg, Phys. Rev. Lett. 98,
064801 (2007).
[18] J. F. Ziegler, M. D. Ziegler, and J. P. Biersack, Nucl.
Instrum. Methods Phys. Res., Sect. B 268, 1818 (2010).
[19] M. Bender, H. Kollmus, H. Reich-Sprenger, M.
Toulemonde, and W. Assmann, Nucl. Instrum. Methods
Phys. Res., Sect. B 267, 885 (2009).
HEAVY-ION-INDUCED DESORPTION YIELDS OF ... Phys. Rev. ST Accel. Beams 16, 083201 (2013)
083201-9